text
stringlengths 66
377k
|
|---|
Assessing Stability of Crutch Users by Non-Contact Methods
# Introduction
Estimates from the World Health Organization (WHO) show that over one billion people need one or more assistive devices, with an expectation that these figures will double to beyond two billion by 2050. Among those assistive devices, the two most popular are walking and visual aids. Walking aids accounted for 46.3% of all assistive devicesand were found to be more common in rural areas and among males while visual aids were more common in urban areas and among female users. One explanation for the prevalence of visual and mobility aids is that both are relatively inexpensive and can be used effectively with little training or support while having the potential to make a big functional difference.
People who use assistive devices to walk are already unstable and in danger of further injury if experiencing a fall. The purpose of this paper is to evaluate one method with two parameters of measuring the stability of a person walking with the most common walking aid, a crutch.
Stability is generally associated with gait. Gait analysis is the assessment of the attributes or manner of walking done by observing a person as they walk in a straight line. According to some gait models, the symmetry and stability of gait constitute a gait appraisal model.
Typically, there are two ways to perform gait evaluation. The first is an evidencebased approach, while the second measures body kinematics, muscle activity, and body mechanics using sophisticated instrumentation. A number of studies into measuring gait have been conducted using high-end tools including force platforms, 3D cameras, and optical markers. These systems require installation in areas with sufficient space and are very expensive. A specialized technician is also required to properly operate the system. The main interests in motion analysis currently revolve around detecting a specified activity or event (e.g., falls). Most research in this area uses linear acceleration and gyroscopes to detect falls by applying thresholds to velocities, accelerations, and angles. It has been shown that inertial measurement systems (IMU) that use a combination of accelerometers and gyroscopes can reliably replace camera-based systems for clinical body motion and gait analyses.
In recent years, researchers have developed software that is able to detect and track human limbs using 135 key-points in a single image. OpenPose, one of the first realtime multi-person tracking systems, was developed by researchers at Carnegie Mellon University in 2014 . It was released in the form of Python code, C++ implementation, and a Unity Plugin. This multi-person 2D pose estimation method uses a specific nonparametric representation, Part Affinity Fields (PAFs), to learn to associate body parts with individuals in the image.
The OpenPose algorithm has shown to be capable of learning to associate human body parts with identifiable individuals within frames of a video, track and identify the action of doing bilateral squatsand be used to assess the movements of preschool children to diagnose gross motor action. Since the creation of OpenPose, many pose-estimation systems with similar capabilities have been created. Multi-person pose estimation is expected to be applied to many fields, such as fitness training, pedestrian recognition, and military training. To our knowledge, OpenPose has not been used to access stability dynamically and initial research regarding non-contact methods for remote assessment has only been explored recently.
Human joint stability is a parameter that is difficult to define accurately, or to quantify, as its assessment requires spatial and temporal measurements. Studies of gait look into movement parameters of individual steps such as flight time or stance but further insight can be gained by studying the higher vector derivatives of displacement such as acceleration, which relates to the force experienced by the participants. Measurements of kinematic parameters of gait such as acceleration and jerk, the rate of change of acceleration, have been proposed to that effect, with abrupt accelerations or jerk related to poor control or stability. For instance, jerk has been tested as the parameter of the sway in participants with Parkinson's disease. Postural sway can be distinguished by calculating jerk on patients' lower back. Also, the jerk has been used as a parameter for measuring smoothness of movement. For example, a study found a correlation between jerk and proficiency of movement between experienced and novice dancers.
Fall prevention and stability is also a major concern in the field of rehabilitation and gait. There are several studiesthat use jerk as a measure of smoothness in movement. One model estimates the Cartesian kinematic jerk of the hips' orientation during a three-dimensional movement by using a smartphone to collect gyroscopic data. The study finds that the Cartesian kinematic jerk grows in value as the analyzed motion becomes less fluent. The other study proposes the movement fluency variables: hesitation, coordination and smoothness. The smoothness variable is based on the jerk of the total body center of mass. These parameters allow for the quantitative identification in changes to motion control due to age.
As the signals of such parameters tend to carry a lot of noise, differ even in people of the same age group and health status, and are affected by the measurement methods, additional processing is required to extract meaningful information from them. There are many studies in medical imaging systems that use Signal-to-Noise Ratio (SNR) for measuring the picture quality of imagesin order to achieve improvements in penetration depth in medical ultrasound. Increased SNR can also be used to allow imaging at higher frequencies and thereby increase spatial resolution without any loss of penetration. SNR has been used extensively in signal processing, where there is noise interfering in the signal, and large values of SNR are considered good outcomes since it represents a strong signal with low noise. In other words, we calculate SNR as a measure of useful information to false information, as the acceleration of the participants in the present study incorporates noise. The use of SNR in our context is different and it has been calculated as the inverse of the coefficient of variation, and this is the first time that SNR has been used to assess the smoothness of movement. Because of this, the SNR that we calculate is dimensionless.
To our knowledge, there has been only one preliminary study assessing stability by measuring the acceleration of the people who walk with crutches, and that study looked at periodic patterns while using sensors and optical motion capturing of participants. In this paper, our objective is to eliminate expensive setups and simplify the measurements by using video analysis. The goal is to propose a simple, inexpensive, non-contact, remote evaluation of stability in crutch users.
The next section details methods and materials used in the experiment and measurements.
# Materials and methods
This section discusses the experimental setup and methods applied to collect data. It also provides information about the participants and data collected from them.
## Experimental setup and data
Video data collection was done at 30 frames per second using two 8-megapixel iPhone 6s set up on tripods. One camera recorded a side-view of participants at a distance far enough away to capture three to five steps when using swing-through gait. The other camera captured a front-view and was placed at an angle 90 degrees off from the side-view camera. This setup allowed for the capture of a stereoscopic view of the participants. A large checkerboard pattern was placed on the wall in view of both cameras to acquire the stereoscopic transformation matrix. This matrix was then used to obtain 3D positional data of objects within the recorded area.
The two videos were manually synchronized by finding the frame where a major physical event occurred. This was typically a first step where either the foot of the subject or the crutch first left the ground or first contacted the ground that was visible in both videos. Since the framerates had been set to be equal, a certain number of frames was taken after that point to record the entire walk. This way the entire video sequence analyzed was synchronized and viewable in both videos and no extraneous material was run through the software.
The layout of the walkway is shown in. The 2D pixel location values of the joints were extracted from OpenPose and exported to the MATLAB software for 3D localization using stereo vision techniques. Furthermore, MATLAB extrapolated the 3D velocities, accelerations, and jerk from those positions. The 2D pixel location values of the joints were extracted from OpenPose and exported to the MATLAB software for 3D localization using stereo vision techniques. Furthermore, MATLAB extrapolated the 3D velocities, accelerations, and jerk from those positions.
After processing with OpenPose software, the visual data is presented in. The 2D pixel location values of the joints were extracted from OpenPose and exported to the MATLAB software for 3D localization using stereo vision techniques. Furthermore, MATLAB extrapolated the 3D velocities, accelerations, and jerk from those positions.
After processing with OpenPose software, the visual data is presented in. To evaluate the movement of the participants the following time series values have been calculated from each video frame: acceleration (a) represented by the equation a = Δv/Δt, where v is the velocity and t is the time, and jerk (j), which is the second derivative of velocity, represented by the equation j = Δa/Δt. For each video frame , the instantaneous acceleration in space (i.e., in the x, y, and z coordinates) was calculated for the ankle of the weight-bearing leg for the whole run. The acceleration (a) has been calculated by combining the three acceleration components (αx, αy, αz in the Cartesian coordinates) with the Equation (1).
[formula] = + +(1) [/formula]
Additionally, the data is processed to calculate Signal-to-Noise-Ratio (SNR) of acceleration and jerk. In practice, SNR is equivalent to the inverse of the coefficient of variation, or it is approximated by the ratio of the mean of all data ( ) divided by the standard deviation of all data (s) as shown in Equation (2), as variables are always non-negative. To evaluate the movement of the participants the following time series values have been calculated from each video frame: acceleration (a) represented by the equation a = ∆v/∆t, where v is the velocity and t is the time, and jerk (j), which is the second derivative of velocity, represented by the equation j = ∆a/∆t. For each video frame, the instantaneous acceleration in space (i.e., in the x, y, and z coordinates) was calculated for the ankle of the weight-bearing leg for the whole run. The acceleration (a) has been calculated by combining the three acceleration components (α x , α y , α z in the Cartesian coordinates) with the Equation (1).
[formula] a = a 2 x + a 2 y + a 2 z(1) [/formula]
Additionally, the data is processed to calculate Signal-to-Noise-Ratio (SNR) of acceleration and jerk. In practice, SNR is equivalent to the inverse of the coefficient of variation, or it is approximated by the ratio of the mean of all data (x) divided by the standard deviation of all data (s) as shown in Equation (2) The SNR metric was selected in this stud movement includes a large component of nois tically to the signal. This reflects on the displ which are velocity, acceleration and jerk. In o mation from the analysis of the video regard data was used for processing. As the SNR compares the level of the des abrupt changes in movement with the resultin by the SNR value. A low SNR value relates to changes in movement, which will make it diff nal accurately. A high SNR value relates to " relate to a steady walk. Therefore, the SNR of as a single measurement providing informat user compares to a natural system variation.
A common problem with data collection the acceleration signal. The extraneous stocha tive of the technique used, and it is present e attached to users while walking. This using techniques like moving average, where set are calculated to smooth out short-term flu or cycles, or employing low pass filters, which selected cutoff frequency and attenuate signa frequency, in order to interpret them. Follow identified and parameters related to gait are e For this particular study an approach wa of every step of the participants' entire walk. ation and jerk of the ankle for each individua the acceleration and the SNR for the jerk were culated characterize all recorded steps, and c
[formula] [ X 2 ] = σ 2 + µ 2 {\displaystyle \operatorname {E} \left[Xˆ{2}\right] = \sigmaˆ{2}+\muˆ{2}}. SNR = x s(2) [/formula]
The SNR metric was selected in this study because while walking with crutches, the movement includes a large component of noise or abrupt changes superimposed stochastically to the signal. This reflects on the displacement data and the time derivatives of it which are velocity, acceleration and jerk. In order to capture all of the important information from the analysis of the video regarding gait and stability, the raw acceleration data was used for processing. As the SNR compares the level of the desired signal to the level of background noise, abrupt changes in movement with the resulting changes in acceleration will be registered by the SNR value. A low SNR value relates to "high" noise in the signal, or very abrupt changes in movement, which will make it difficult to determine the magnitude of the signal accurately. A high SNR value relates to "low" noise, or a smoother gait, which we relate to a steady walk. Therefore, the SNR of both acceleration and jerk can be regarded as a single measurement providing information on how the acceleration of the current user compares to a natural system variation.
A common problem with data collection for gait analysis is the existence of noise in the acceleration signal. The extraneous stochastic component in the signal holds irrespective of the technique used, and it is present even in measurements with accelerometers attached to users while walking. This requires the processing of time series data using techniques like moving average, where averages of different subsets of the full data set are calculated to smooth out short-term fluctuations and highlight longer-term trends or cycles, or employing low pass filters, which pass signals with a frequency lower than a selected cutoff frequency and attenuate signals with frequencies higher than the cutoff frequency, in order to interpret them. Following this, steps in the walk are individually identified and parameters related to gait are extracted.
For this particular study an approach was implemented that processed the raw data of every step of the participants' entire walk. After calculating the instantaneous acceleration and jerk of the ankle for each individual time point of the whole run, the SNR for the acceleration and the SNR for the jerk were calculated. Both SNR values that were calculated characterize all recorded steps, and combine both the phases and parameters of gait such as heel strike, stance, and push off in one single parameter.
Our approach is based on an assumption that the values of acceleration and jerk are adequate measures of stability. The justification is that these two parameters are being used in vehicles, elevators, and roller coaster design where it is necessary to limit the exposure of passengers to both the maximum force (acceleration) and maximum jerk since time is required to adjust muscle tension and get used to even modest stress changes. It has been found that excessive jerking is associated with an uncomfortable ride, even at levels of acceleration that do not cause injury.
The next section describes the participants in the study and the protocol for data collection.
## Participants
A total of 10 participants were evaluated. Five disabled users and five non-disabled users, and one experienced temporary crutch user. The data presented in this paper is based on the following: (i) 15 videos taken of disabled users, who had been using crutches for many years because of their disabilities incurred by poliomyelitis, (ii) 46 videos taken of the five non-disabled users, with four using crutches sporadically for the needs of these measurements for periods ranging between one day to a few weeks, and (iii) 7 videos taken with an experienced temporary user who had been using crutches continuously for a period of six weeks due to multiple fractures of his leg. The summary of data is given in. The data have been collected over the period of 8 months, from June 2020 to January 2021. The IRB approval for the study has been obtained from SUNY Korea on 7 August 2020, and it follows the Declaration of Helsinki and its set of ethical principles.
The next section presents the results obtained from the analysis of video files from the 10 participants.
# Results
For the purpose of analysis, the participants are grouped in three categories: novice users, experienced temporary user, and disabled users.presents the analysis of motion measurements for each group calculated from raw data. The left ankle acceleration of each user of the three groups was measured using the video analysis, with the left side chosen because it was weight bearing for the majority of users. When a comparison is made, using the box and whisker chart, see, between the accelerations of typical runs of each of the three groups studied in this work, the skewness in a typical dataset is visualized. The median line which cuts the box is not in the middle of it and the 25% quartile side, the bottom part, is very small, while the top part of the box, the 75% quartile, is quite larger as accelerations higher than the median show a wider range in that section. The upper and lower whiskers extent out to the extreme values of acceleration, while the outliers are not shown for clarity. However, once the results are further analyzed and processed a different picture emerges, one where differences in gait characteristics are not related to disability. Our results show that the acceleration SNR of the disabled users is the lowest at 2.1 m/s 2 , and the novice users at 3.9 m/s 2 , twice as that of the disabled users, and the experienced user recording the highest value of the three user groups at 8.8 m/s 2 . Although the walk of the disabled users was periodic, it was characterized by abrupt changes in movement due to their disability, which was verified by the actual videos. Novice crutch users showed a steady gait but lacked confidence and experience which resulted in intermediate acceleration SNR values when compared to the temporarily disabled user.
A similar observation was made in the calculated jerk SNR. Jerk, taken from engineering system studies, measures the rapid changes in acceleration that the body experiences. In the previous biometric studies, it has been established that Parkinson's disease patients showed larger jerk values, which authors used to measure the relative smoothness of postural sway and interpreted as a measure of dynamic instability. Others used acceleration-based parameters to characterize the joint stability and jerk was defined as a potential stability measure, since anomalous jerk could be a symptom of inadequate body control. However, once the results are further analyzed and processed a different picture emerges, one where differences in gait characteristics are not related to disability. Our results show that the acceleration SNR of the disabled users is the lowest at 2.1 m/s 2 , and the novice users at 3.9 m/s 2 , twice as that of the disabled users, and the experienced user recording the highest value of the three user groups at 8.8 m/s 2 . Although the walk of the disabled users was periodic, it was characterized by abrupt changes in movement due to their disability, which was verified by the actual videos. Novice crutch users showed a steady gait but lacked confidence and experience which resulted in intermediate acceleration SNR values when compared to the temporarily disabled user.
A similar observation was made in the calculated jerk SNR. Jerk, taken from engineering system studies, measures the rapid changes in acceleration that the body experiences. In the previous biometric studies, it has been established that Parkinson's disease patients showed larger jerk values, which authors used to measure the relative smoothness of postural sway and interpreted as a measure of dynamic instability. Others used acceleration-based parameters to characterize the joint stability and jerk was defined as a potential stability measure, since anomalous jerk could be a symptom of inadequate body control.
However, keep in mind that in the present study, high values of SNR jerk represent a smoother walk. Our results show that the disabled users' group had the lowest jerk SNR of 0.8 m/s 3 , the novice users an intermediate value at 2.4 m/s 3 and the experienced user the highest value of jerk SNR at 3.5 m/s 3 .
The SNR of acceleration and jerk for all novice users runs over this 8-month period are shown in, and for all experienced (disabled and one temporary user) participants are shown in. The changes in acceleration displayed by the experienced users are accompanied by similar changes in jerk in practically every case, with the jerk being consistently lower than the acceleration. It suggests that the data for crutch users with more experience reflects a more confident walk, irrespective of the differences in gait. This consistent behavior observed in the disabled and temporary experienced users is not observed in the novice users. In some runs, the acceleration and jerk do not follow each other at all, while there are many runs where they match in values. This last point is interesting as it indicates a walk with highly abrupt changes in acceleration and movement. The values of acceleration and jerk appear to be far higher in most cases in the nov- This consistent behavior observed in the disabled and temporary experienced users is not observed in the novice users. In some runs, the acceleration and jerk do not follow each other at all, while there are many runs where they match in values. This last point is interesting as it indicates a walk with highly abrupt changes in acceleration and move- This consistent behavior observed in the disabled and temporary experienced users is not observed in the novice users. In some runs, the acceleration and jerk do not follow each other at all, while there are many runs where they match in values. This last point is interesting as it indicates a walk with highly abrupt changes in acceleration and movement. The values of acceleration and jerk appear to be far higher in most cases in the novice users' group.
The experienced temporary user data is shown in. It can be seen clearly that both the acceleration and jerk SNR follow each other consistently. The acceleration SNR shows a much higher value while the jerk SNR is consistently low. This is interpreted as a stable and confident walk with crutches where the user follows a clear path in a controlled manner. The difference in acceleration and jerk between the disabled users and experienced temporary user is related to the complex movement due to the disability level.
# Discussion
The proposed interpretation of the acceleration data allows us to differentiate between each of the three groups by the movement their legs perform in space. The disabled users have a low, i.e., noisy, acceleration SNR, as their legs perform a complex movement in three dimensions due to their disabilities. The novice users, who are not hindered by any disability, show an improved acceleration SNR twice that of the disabled users, while the experienced temporary user shows an acceleration SNR nearly four times almost that of the disabled users and twice that of the novice users. In effect, a low acceleration SNR value indicates a far more complex 3-dimensional movement of the leg of the user, which can be evidenced by the videos. A high acceleration SNR value points to a near 2-dimensional movement of the leg along a plane, something which the novice users perform to a certain extent while walking with crutches, and certainly, the experienced temporary user does confidently and steadily.
The interpretation of the jerk data allows for a similar observation regarding stability and movement control between each group. The low jerk SNR value measured for the disabled users, which is related to very high rates of change of acceleration, matches previous observations of a higher jerk for neurological patients. An unstable walk registers as high rates of change of acceleration, which is associated with a more abrupt gait.
Intermediate jerk values are encountered for the novice users, where jerk SNR values measured three times higher than those of the disabled users, an effect of a steadier movement of the leg on a 2-dimensional plane. The experienced temporary user has a jerk value for his walks which is four times that of the disabled users and 50% higher than that of the novice users. This value reflects the steady and controlled movement with crutches that appeared subjectively in the videos.
When the acceleration SNR data are studied together with the jerk SNR data, there appears to be a differentiation between groups depending on their experience with crutches. The acceleration SNR data for all novice users do not show a consistent trend or value over a long period of time, but it appears that for the majority of runs of different users, the acceleration and the jerk SNR have similar and comparable values. This observation does not hold true for the experienced temporary user and the disabled users
# Discussion
The proposed interpretation of the acceleration data allows us to differentiate between each of the three groups by the movement their legs perform in space. The disabled users have a low, i.e., noisy, acceleration SNR, as their legs perform a complex movement in three dimensions due to their disabilities. The novice users, who are not hindered by any disability, show an improved acceleration SNR twice that of the disabled users, while the experienced temporary user shows an acceleration SNR nearly four times almost that of the disabled users and twice that of the novice users. In effect, a low acceleration SNR value indicates a far more complex 3-dimensional movement of the leg of the user, which can be evidenced by the videos. A high acceleration SNR value points to a near 2-dimensional movement of the leg along a plane, something which the novice users perform to a certain extent while walking with crutches, and certainly, the experienced temporary user does confidently and steadily.
The interpretation of the jerk data allows for a similar observation regarding stability and movement control between each group. The low jerk SNR value measured for the disabled users, which is related to very high rates of change of acceleration, matches previous observations of a higher jerk for neurological patients. An unstable walk registers as high rates of change of acceleration, which is associated with a more abrupt gait. Intermediate jerk values are encountered for the novice users, where jerk SNR values measured three times higher than those of the disabled users, an effect of a steadier movement of the leg on a 2-dimensional plane. The experienced temporary user has a jerk value for his walks which is four times that of the disabled users and 50% higher than that of the novice users. This value reflects the steady and controlled movement with crutches that appeared subjectively in the videos.
When the acceleration SNR data are studied together with the jerk SNR data, there appears to be a differentiation between groups depending on their experience with crutches.
The acceleration SNR data for all novice users do not show a consistent trend or value over a long period of time, but it appears that for the majority of runs of different users, the acceleration and the jerk SNR have similar and comparable values. This observation does not hold true for the experienced temporary user and the disabled users where the acceleration SNR and the jerk SNR show the same trend without registering the same values. This is clearly confirmed with the experienced temporary user for all of his runs. Therefore, it appears that experience with using crutches allows for a much higher acceleration SNR value than a jerk SNR value. This observation was confirmed irrespective of the varying gait characteristics of all the experienced users, both disabled and temporary. Confidence and control of the user are demonstrated by relative steady and high acceleration values with low jerk SNR values regardless of their disability.
# Conclusions
In this study we present preliminary results derived from the data of ten people walking with crutches. Three groups have been analyzed: novice users, experienced temporary user, and disabled users. The SNR metric was selected in this study because while walking with crutches, the movement includes a large component of noise or abrupt changes superimposed stochastically to the signal. The SNR metric takes that into account by incorporating standard deviation in the calculation. When the acceleration SNR data are studied together with the jerk SNR data, there appears to be a differentiation between groups depending on their experience with crutches.
The most interesting result is that the experienced temporary user data shows both the acceleration and jerk SNR follow its other consistently. This is interpreted as a stable and confident walk with crutches where the user follows a clear path in a controlled manner. The experienced temporary user has a jerk value for his walks which is 4 times that of the disabled users and 50% higher than that of the novice users. This value reflects the steady and controlled movement with crutches that appeared subjectively in the videos.
The approach is promising because it allows us to differentiate between confident users and ones with unstable movement. This differentiation is independent of the degree of disability. This approach could improve the assessment of crutch user stability. Improving assessment would allow correcting the gait of individuals and preventing unnecessary falls.
Future work will involve expanding the sample size of experienced users, focusing on those temporarily using assistive devices due to their smoother movements that are easier to study. Furthermore, future studies will seek to define a more precise function of acceleration to use in the calculations. That will involve analyzing individual walk cycles for more granularity in the data. Informed Consent Statement: Written informed consent was obtained from all subjects involved in the study.
# Data availability statement:
The data from this study is open source and can be found here: https://www.mobilityaid.org/publications/gaitdata.
## Conflicts of interest:
The authors declare no conflict of interest. |
Diplopia due to acquired Brown syndrome after COVID-19 infection
## Case report
A 31-year-old woman, an intensive care unit nurse, was admitted to the ophthalmology service, Ankara University, School of Medicine, with a complaint of double vision and pain in up-and-left gaze for 4 days. No history of systemic or ocular disease was noted other than dyschromatopsia. However, 3 weeks prior to admission, she tested positive for COVID-19 on real-time polymerase chain reaction testing. Complete blood count analysis showed a white blood cell count of 3340/mm 3 , indicating leukopenia. Laboratory blood analysis revealed that ferritin, fibrinogen, D-dimer, erythrocyte sedimentation rate, and C-reactive protein levels were within normal limits. The patient received combined treatment of favipiravir, enoxaparine, and dexamethasone as a standardized therapy for COVID-19. Prior to admission, however, only acetylsalicylic acid (81 mg/ day) had been given to the patient. On examination, visual acuity was 20/20 in each eye. Biomicroscopic and fundus findings were within normal limits, and intraocular pressure was 15 mm Hg in each eye. The patient was orthophoric in primary position but had a right hypotropia of 5 D in up-and-left gaze and a right hypotropia of 2 D -3 D with right head tilt. The patient had limited elevation in adduction . Standard forced duction testing 1 revealed minimal restriction of elevation in adduction. In addition, the patient experienced pain on palpation over the trochlea. Orbital magnetic resonance imaging (MRI) revealed thickening and contrast enhancement in the distal part of superior oblique muscle-tendon-trochlea complex .
The patient complained of headache and was evaluated by the neurologist at our institution. She underwent laboratory tests and cranial MRI. Rheumatoid factor, antinuclear antibody, antinuclear cytoplasmic antibody, and anti-double stranded DNA were all negative. MRI revealed nonspecific hyperintensities in the periventricular white matter of the parietal lobes, indicating vasculitis . The patient received 1 gm of intravenous methylprednisolone per day for 3 days, followed by 64 mg of oral methylprednisolone per day, tapering 16 mg every 2 days. Because the patient did not have diplopia in primary position, no additional ocular treatment was required. By 2 months' follow-up, diplopia had resolved. There was mild limitation to elevation in adduction ; repeat MRI of the orbits showed resolution of the previously observed hyperintense thickening of the trochlea-superior oblique complex .
# Discussion
Ocular motility disorders associated with oculomotor nerve dysfunction in patients with COVID-19 have been reported previously. 2, [bib_ref] Cranial nerve III palsy in the setting of COVID 19 infection, Fitzpatrick [/bib_ref] We present a case of myositis/ trochleitis that manifested with diplopia and limited elevation in adduction, consistent with a Brown syndrome secondary to COVID-19.
COVID-19 is known to affect muscles, tendons, and cartilage, and cases of myositis involving paraspinal, facial, bulbar, proximal limb, and obturator muscles caused by SARS-CoV-2 have been reported. [bib_ref] COVID-19-associated myositis with severe proximal and bulbar weakness, Zhang [/bib_ref] [bib_ref] Paraspinal Myositis in Patients with COVID-19 Infection, Mehan [/bib_ref] In addition, arthritis, enthesitis, and dactylitis associated with SARS-CoV-2 infection have also been reported. [bib_ref] Can SARS-CoV-2 trigger reactive arthritis?, Wendling [/bib_ref] [bib_ref] Reactive arthritis after COVID-19 infection, Ono [/bib_ref] The exact mechanism of musculoskeletal system involvement in patients with COVID-19 remains unclear, however. Various viral pathogens, such as hepatitis C, human immunodeficiency virus, Middle East respiratory syndrome coronavirus, and influenza A and B are known to cause myositis. Furthermore, SARS-CoV-2 has a high affinity for angiotensin-converting enzyme 2 (ACE2), which skeletal muscles express. [bib_ref] The interaction between SARS-CoV-2 and ACE2 may have consequences for skeletal muscle..., Ferrandi [/bib_ref] Thus, direct viral invasion of skeletal muscles by SARS-CoV-2 is possible.
Another mechanism of musculoskeletal system involvement may be the immune-mediated pathway. Viral pathogens are known to cause immune-mediated myositis by triggering or exacerbating autoimmunity. [bib_ref] Autoimmunity as the comet tail of COVID-19 pandemic, Talotta [/bib_ref] In addition, many inflammatory cells are stimulated, and cytokines are released in patients with COVID-19. This inflammatory response may trigger immunemediated pathways or may be myotoxic.
In the present case, acquired Brown syndrome was associated with trochlear tendon complex involvement because orbital MRI revealed thickening and enhancement in the distal part of the superior oblique muscletendon-trochlea complex. In addition, the onset of ocular findings 3 weeks after COVID-19 infection suggests that this condition may be a reactive response rather than a direct viral invasion.
Noncongenital juvenileonset bilateral lamellar cataract in 1p36 deletion syndrome Carla Danese, MD, a Silvia Pignatto, MD, a and Paolo Lanzetta, MD a,b
We report the case of a 16-year-old girl with 1p36 deletion syndrome, who experienced visual loss in both eyes for 2 months because of lamellar cataracts. Mutations on some 1p36 genes in both experimental models and humans may be associated with cataract. This is the first detailed description of acquired juvenileonset bilateral cataract with 1p36 deletion. M onosomy 1p36, affecting 1 in 5,000 newborns, represents the most common terminal deletion syndrome. [bib_ref] The ocular motor examination, Wright [/bib_ref] Deletions, 95% of which are de novo, occur equally in males and females and in all ethnic groups. It is difficult to identify the genes responsible for the phenotypic manifestations, because the distal end of the short arm of the chromosome 1 is very rich in genes. [bib_ref] Acute abducens nerve palsy in a patient with the novel coronavirus disease..., Falcone [/bib_ref] There is no common breakpoint or deletion size in 1p36 monosomy. [bib_ref] The ocular motor examination, Wright [/bib_ref] [bib_ref] Acute abducens nerve palsy in a patient with the novel coronavirus disease..., Falcone [/bib_ref] [bib_ref] Cranial nerve III palsy in the setting of COVID 19 infection, Fitzpatrick [/bib_ref] Affected individuals are phenotypically heterogeneous, because the size of the deletion varies widely. The features are not specific for this syndrome, and the cytogenetic identification of the deletion is often difficult. Therefore, some individuals may be misdiagnosed. [bib_ref] Cranial nerve III palsy in the setting of COVID 19 infection, Fitzpatrick [/bib_ref] The systemic findings are summarized in . 1,2,4 Ocular malformations or functional visual problems are present in more than half the cases. They include strabismus (35%-67%), hyperopia (41%-67%), myopia (17%-40%), astigmatism (23%), nystagmus (13%-23%), unilateral cataract (5.9%), retinal albinism (5.9%), and unilateral optic nerve coloboma (2.9%). [bib_ref] The ocular motor examination, Wright [/bib_ref] [bib_ref] Acute abducens nerve palsy in a patient with the novel coronavirus disease..., Falcone [/bib_ref] [bib_ref] Cranial nerve III palsy in the setting of COVID 19 infection, Fitzpatrick [/bib_ref] |
MFN1-dependent alteration of mitochondrial dynamics drives hepatocellular carcinoma metastasis by glucose metabolic reprogramming
BACKGROUND: Mitochondrial dynamics plays an important role in tumour progression. However, how these dynamics integrate tumour metabolism in hepatocellular carcinoma (HCC) metastasis is still unclear. METHODS: The mitochondrial fusion protein mitofusin-1 (MFN1) expression and its prognostic value are detected in HCC. The effects and underlying mechanisms of MFN1 on HCC metastasis and metabolic reprogramming are analysed both in vitro and in vivo. RESULTS: Mitochondrial dynamics, represented by constant fission and fusion, are found to be associated with HCC metastasis. High metastatic HCC displays excessive mitochondrial fission. Among genes involved in mitochondrial dynamics, MFN1 is identified as a leading downregulated candidate that is closely associated with HCC metastasis and poor prognosis. While promoting mitochondrial fusion, MFN1 inhibits cell proliferation, invasion and migration capacity both in vitro and in vivo. Mechanistically, disruption of mitochondrial dynamics by depletion of MFN1 triggers the epithelial-to-mesenchymal transition (EMT) of HCC. Moreover, MFN1 modulates HCC metastasis by metabolic shift from aerobic glycolysis to oxidative phosphorylation. Treatment with glycolytic inhibitor 2-Deoxy-D-glucose (2-DG) significantly suppresses the effects induced by depletion of MFN1. CONCLUSIONS: Our results reveal a critical involvement of mitochondrial dynamics in HCC metastasis via modulating glucose metabolic reprogramming. MFN1 may serve as a novel potential therapeutic target for HCC.
# Introduction
Hepatocellular carcinoma (HCC) is one of the most common human malignancies and the second leading cause of cancerrelated mortality worldwide. Almost 50% of newly diagnosed HCC cases occur in China.Despite significant progress in HCC treatment, the prognosis of HCC patients is still far from satisfactory, mainly resulting from early metastasis and recurrence.Therefore, a new perspective of tumour metastasis mechanism in HCC is urgently needed, which may help to identify potential targets and develop effective therapies for HCC patients.
Accumulative evidence suggests that mitochondria is an important mediator of tumorigenesis and tumour progression.Mitochondria are bioenergetic, biosynthetic and signalling organelles that are critical for multiple cellular functions including growth, division, calcium homoeostasis, energy metabolism and apoptosis.Mitochondria exist as a dynamic network that constantly changes the morphology to meet the energy demand of cells and to adapt to the extracellular environment. Morphologically, the mitochondrial network is characterised by a mixed structure of long interconnected tubules with short isolated dot-like spheres, which is precisely regulated by two opposite processes: fusion and fission.Mitochondrial dysfunction is closely related to tumorigenesis and tumour progression, where mitochondrial dynamics plays a pivotal role.Several studies have reported dysregulated expression of mitochondrial dynamic proteins such as dynamin 1 like (DNM1L), mitofusin-1 (MFN1) and mitofusin-2 (MFN2) in lung, colon and breast cancers.Recent studies showed that lung cancer exhibited excessive mitochondrial fission and impaired mitochondrial fusion due to an imbalance of DNM1L/MFN expression, which is important for cell cycle progression.In addition, marked upregulation of DNM1L expression was reported to promote breast cancer metastasis via enhancing mitochondria fission.Increased mitochondrial fission was reported to promote the survival of HCC cells by facilitating autophagy and inhibiting mitochondria-dependent apoptosis.For HCC cells, mitochondrial fission was positively related to their metastatic capacity via modulating calcium homoeostasis.Cancer cells exhibit aberrant metabolism characterised by high glycolysis in the presence of abundant oxygen. This phenomenon, known as the Warburg effect or aerobic glycolysis, facilitates tumour growth with elevated glucose uptake and lactate production.The Warburg effect has now been widely accepted as a hallmark of cancer and gives rise to the development of corresponding cancer therapeutic agents. More than ten genes encoding glycolytic enzymes are found to be directly responsible for the Warburg effect.In the present study, we found notable downregulation of MFN1 expression in high metastatic human HCC cell lines and HCC patients with vascular invasions. Overexpression of MFN1 in HCC cells promotes mitochondrial fusion, and inhibits cell proliferation, invasion and migration via modulating metabolic shift from aerobic glycolysis to oxidative phosphorylation.
# Materials and methods
## Cell lines
Human embryonic kidney cell lines HEK and 293T, human HCC cell lines with genetically identical backgrounds and increasingly metastatic potential, MHCC-LM3, MHCC97-H and MHCC97-L, were obtained from the Liver Cancer Institute, Fudan University.The HCC cell lines, SMMC-7721, HepG2 and Hep3B were purchased from the Institute of Biochemistry and Cell Biology, Chinese Academy of Science (Shanghai, China). All these cell lines were cultured in Dulbecco's modified Eagle's medium (DMEM) (Hyclone) containing 10% foetal bovine serum (FBS) (Gibco) routinely at 37.0°C in a humidified incubator with 5% CO 2 .
Vectors and cell transfections Expression vector mediated by lentivirus for human MFN1 was established. The sequence of MFN1 was amplified from cDNA library via a specific sense primer CCGGAATTCATGGCAGAA CCTGTTTCTCCACT, antisense primer CGCGGATCCTTAGGATTCTTC ATTGCTTGAAGGTA. Then harvested DNA was inserted into pCDH-GFP expression vector (System Biosciences).
Expression vector for MFN1 and shRNA for MFN1 were constructed as previously described. Cell transfections was also performed as previously described.Metabolism assays Oxygen consumption rate (OCR) and extracellular acidification rate (ECAR) were measured by using a 96-well XF or XFe extracellular flux analyzer (EFA) (Seahorse Bioscience).Cell proliferation assays Cell proliferation assay was carried out by using Cell Counting Kit-8 (CCK-8; Dojindo, Kumamoto, Japan). Cancer cells were seeded into 96-well plates at a density of 3 × 10 3 /100 μl cells per well (n = 5 for each time point) and incubated for 12, 24, 36, 48 and 72 h. After that the medium was replaced with 10% CCK-8 in complete medium. The absorbance at 450 nm was measured after incubation for 2 h. Each experiment was repeated three times on each condition.
Colony formation assays Cancer cells were seeded into six-well plates at a density of 1 × 10 3 cells in each well and incubated for 2 weeks. After that, cells were fixed in 4% formalin for 20 min and stained with 1% crystal violet for 20 min. The number of colonies was counted after the plates were being stained. Each experiment was repeated three times.
Wound-healing assays Monolayers of cells were plated in 12-well plates and then scratched with the tip of a 20-μL pipette and washed several times with PBS (Hyclone) until dislodged cells were removed clearly. Tumour cells were cultured in DMEM medium free from FBS. The healing status of wound area was photographed at 0 h and 24th/48th/72nd hour after scratching.
Transwell assays Transwell assays with and without Matrigel were carried out by using 24-well plate with Transwell chambers with 8-µm pores (BD Pharmingen). The bottom chamber was filled with DMEM medium with 20% FBS as a chemoattractant. Fifty thousand cells for migration assay and 10 × 10 4 cells for invasion assay were seeded into the upper chamber maintained in medium free from serum. Cells migrated to the underside of the membrane were fixed with 4% formaldehyde, stained with 1% crystal violet and counted with light microscope (×100, Leica).
Animal studies All experimental procedures involving animals were approved by The Animal Care and Use Committee of Huashan Hospital of Fudan University (Shanghai, China). All mice (BALB/c nu/nu, 18.69 ± 0.69 g), under the age of 5 weeks (±3 days), were obtained from Shanghai Slac Laboratory Animal Co. and fed in a specific pathogen-free (SPF) vivarium under standard conditions. Briefly, mice were housed in a 12-h light/dark cycle, temperature-(22 ± 1°C) and humidity-(55 ± 5%) controlled room and were allowed free access to water and a maintenance diet. All cages contained sterilised wood shavings. Prior to any surgical procedure, all mice were given intraperitoneal (i.p.) injection of Pentobarbital Sodium (40 mg/kg; Sigma Aldrich, St. Louis, MO, USA) as anaesthesia because of its effectiveness and safety. All experimental procedures were conducted in the SPF laboratory during light phases. Mice were finally killed by cervical dislocation.
To construct subcutaneous xenograft tumour model, four HCC cell lines including EV MHCC97-H, OE-MFN1 MHCC97-H, sh-Ctrl HepG2 and sh-MFN1 HepG2 (1 × 10 6 cells suspended in 100 μl of Matrigel and PBS) were injected into the right flanks of mice that were randomly allocated into four groups (n = 5 mice/group) after anaesthesia. Tumour volume was monitored weekly with a calliper until mice were killed after 6 weeks. Tumour volume was calculated by the formula: a × b 2 /2 (where a and b separately represent the largest and smallest tumour diameters).All the tumours were collected. RNA was extracted from fresh tumour tissues and the rest of the tumours were fixed in 4% formalin, and then embedded in paraffin. Consecutive sections of tumour tissues were applied by immunohistochemistry staining.
For orthotopic model, tumours from subcutaneous xenograft tumour model of four groups were minced into 1-2 mm 3 sections and inoculated into the left liver lobes of mice that were randomly allocated into four groups (n = 5 mice/group) after anaesthesia. Mice were killed after 7 weeks.Tumours, livers and lungs of all mice were collected and fixed in 4% formalin, and then embedded in paraffin. Consecutive sections of lung tissue were stained with haematoxylin and eosin. The number of lung metastasis was calculated and evaluated independently by two pathologists.
Clinical samples, tissue microarray (TMA) and immunohistochemistry (IHC) We applied IHC staining analysis on a cohort consisting of 87 HCC patients, by dividing them into two groups, high MFN1 expression (H-MFN1) and low MFN1 expression (L-MFN1). Clinical samples from patients were obtained after acquiring their consent in accordance with the protocol approved by the Ethics Boards of Huashan Hospital of Fudan University (Shanghai, China).
Formalin-fixed and paraffin-embedded tissues were used to construct TMA as previously described.Four-micron core biopsies from the donor blocks were taken and transferred to the recipient paraffin block at predefined array positions and constructed 87 cases of TMA blocks in this study. IHC staining was performed as described previously.After deparaffinisation, rehydrating and antigen retrieval, primary antibodies were added to slides, incubated at 4°C overnight, and followed with incubation with secondary antibody (Dako Cytomation, Glostrup, Denmark) at 37°C for 30 min. Staining was performed with DAB and counterstaining was operated with haematoxylin.
Evaluation of IHC scores Scoring for MFN1 staining was conducted as previous described,briefly using percentage score × staining intensity score. The percentage of positive-staining cells: 0-5% scored 0, 6-25% scored 1, 26-50% scored 2, 51-75% scored 3 and more than 75% scored 4; staining intensity: no staining scored 0, weakly staining scored 1, moderately staining scored 2 and strongly staining scored 3.
Immunofluorescence (IF) After incubating with 100 nM MitoTracker Deep Red in DMEM medium with 10% serum, 5 × 10 3 HCC cells were plated into 24-well plate containing slide glasses of 12 mm in diameter and then fixed in 4% paraformaldehyde for 20 min. After washing with PBS three times, cells were incubated in 1% BSA at room temperature for 1 h. After washing with PBS three times, samples were counterstained with DAPI 100 μg/ml in PBS for 10 min. Finally, the slides were sealed on a slide mounted in antifade reagent (Bevotime, China). Fluorescence were detected by using confocal fluorescent microscope (Nikon, Tokyo, Japan).
## Mitochondrial length mitochondrial length was assessed by staining with mitotracker
Deep Red and performed as described before. Mitochondrial length was measured by tracing the mitochondria using ImageJ software. Mitochondrial length was either binned into different categories (<1 mm, fragmented; 1-2 mm, intermediated; >2 mm, elongated) or taken as an average.
## Western blotting
Western blotting was carried out as described before.Total protein was extracted by RIPA buffer containing protease cocktail inhibitor. All protein samples were quantified. Protein samples were separated by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) and then transferred onto polyvinylidene fluoride (PVDF) membranes. After blocking with 5% skim milk in TBS-T (BD Pharmingen), PVDF membranes were incubated with the primary antibody overnight at 4°C and then with the secondary antibody for 1 h at room temperature. The following antibodies were used: anti-MFN1 (1:1000, Abcam), anti-MFN2 RNA isolation, reverse transcription and quantitative real-time polymerase chain reaction (qRT-PCR) RNA of cell lines and frozen tissue was isolated using Trizol reagent (Invitrogen). After quantification, complementary DNA synthesis was performed using PrimeScript reverse transcriptase reagent kit (Takara) according to the manufacturer's directions.
Quantified real-time PCR was performed using SYBR Green PCR Master Mix (Takara) and then detected by ABI PRISM 7900 Sequence Detection System (Applied Biosystems). The results were normalised to β-actin for mRNA measurement. All the primers are listed in . Each experiment was repeated six times on each condition.
# Statistical analysis
Statistical analyses were carried out using Statistical Package for Social Sciences Version 21.0 (SPSS 21.0) and Graphpad Prism ® 7.0.
The analysis of variance (ANOVA) test was used to compare mean values among three or more groups, whereas independentsample two-sided Student's t test was used to compare two groups with normal distribution data. The correlation analysis was determined by Pearson. Kaplan-Meier survival analyses were used to estimate the overall survival and disease-free survival, and the log-rank test was used to assess the differences. All statistics were two sided and p < 0.05 was considered statistically significant.
# Results
MFN1 is identified as a leading downregulated gene in mitochondria dynamics closely associated with HCC metastasis.
To evaluate mitochondrial dynamics in human HCC cell lines with different metastatic potential, we applied IF technique to manifest mitochondrial morphology. The MitoTracker Deep Red showed an explicit phenomenon that HCC cell lines with high metastatic potential, like MHCC97-H and HCC-LM3, tended to have more fragmented mitochondria, while those with low metastatic potential, like HepG2 and Hep3B, shared much more elongated mitochondria. Similar results were also observed by fluorescence in human HCC cell lines with different metastatic potential. To further identify the key regulators of mitochondrial dynamics, we performed qRT-PCRand western blotfor such candidates including MFN1, MFN2, Opa1 and MFF. MFN1 showed significant variations among HCC cell lines with different metastatic potential, in both mRNA and protein levels. High metastatic HCC cell lines like MHCC97-H and HCC-LM3, had significantly decreased expressions of MFN1, which is consistent with their more fragmented mitochondria. These results suggested that mitochondria in high metastatic HCC cell lines prefer a dynamics imbalance towards fission, which is accompanied by MFN1 downregulation.
Downregulation of MFN1 is associated with vascular invasion and poor outcome of HCC patients To further understand the regulatory role of mitochondrial dynamics in HCC, we compared expression levels of mitochondrial dynamics-related genes between tumour and peritumour mesenchyme tissues by qRT-PCR in 34 HCC samples . MFN1 expression was greatly decreased in tumour tissues compared with para-tumour tissues. Correlation analysis between patients' clinical information and MFN1 expression levels elaborated that MFN1 expression was negatively related to TNM stage and portal vein tumour thrombosis (PVTT) frequency , , b). PVTT from HCC was apparently linked to high recurrence and intrahepatic metastasis. To define the specific changes in candidate mitochondrial dynamics regulators associated with metastatic progression, we checked their expression using ten pairs of HCC samples by qRT-PCR. The expression levels of three genes, including MFN1, DNM1L and MFF, were significantly different between HCC with and without PVTT, among which the difference of MFN1 was the greatest . In these patients with PVTT, we observed that HCC cells with low staining degrees particularly tend to invade into vascular . In addition, Kaplan-Meier analysis showed that patients with low MFN1 expression had obviously a poorer prognosis in both disease-free survival (DFS) and overall survival (OS) than those with high MFN1 expression (log rank P < 0.05 and < 0.01, respectively, .
## Mfn1 inhibits hcc proliferation and metastasis via mitochondria fusion in vitro and in vivo
To explore the functions of mitochondrial fusion and MFN1 in HCC cells, we performed both in vitro and in vivo gain-and loss-offunction studies. We established MHCC97-H cell line with stable overexpression of MFN1 (OE-MFN1) and MFN1-knockdown HepG2 cell line (sh-MFN1). Overexpression of MFN1 led to increased fusion of mitochondria (elongated mitochondria). On the contrary, knockdown of MFN1 resulted in more mitochondrial fragmentation. Furthermore, overexpression of MFN1 significantly inhibited proliferation, clone formation, migration and invasionof MHCC97-H cells in vitro, while knockdown of MFN1 exhibited an opposite effect in HepG2. As epithelial-tomesenchymal transition (EMT) is suggested to be closely associated with acquisition of pro-invasive capacities in tumour,we wondered whether MFN1 plays important roles in EMT of HCC cells. As expected, MFN1 induced an increased expression of the epithelial marker E-cadherin, and reduced expression of mesenchymal markers including N-cadherin, vimentin as well as snail. Consistently, knockdown of MFN1 in HepG2 led to the opposite effect.
To verify the role of MFN1 in vivo, we constructed subcutaneous and orthotopic xenograft models. In subcutaneous xenograft models, the average volume of MHCC97-H tumour in the MFN1 transfection group was significantly smaller than the control group. On the other hand, downregulation of MFN1 could significantly promote in vivo tumour growth with much larger tumour sizes than control in xenograft models of HepG2 cells. In orthotopic xenograft models, MHCC97-H or HepG2 subcutaneous xenografts were isolated and implanted into the liver to establish orthotopic xenografts. The percentage of nude mice with lung metastasis and the total numbers of lung metastatic lesions in the MFN1 overexpression group were much lower than those in the control group. However, an increase in the percentage of lung metastasis was observed in nude mice bearing orthotopic xenografts from sh-MFN1-HepG2 cells, though the total numbers of lung metastatic lesions in MFN1knockdown group were not significantly increased. Low MFN1 level in human HCC was correlated with vascular invasion and poor prognosis. a qRT-PCR analysis for MFN1 expression levels in 34 HCC samples between tumour and peritumour mesenchyme tissues indicated that MFN1 expression was greatly inhibited in tumour tissues compared with para-tumour tissues. n = 6, mean ± SEM, ***P < 0.001, **P < 0.01, *P < 0.05. b qRT-PCR analysis for MFN1 expression levels in 20 HCC patients with and without portal vein tumour thrombosis (PVTT) indicated that for patients with PVTT MFN1 expression decreased. n = 6, mean ± SEM, ***P < 0.001, **P < 0.01, *P < 0.05. c Representative immunohistochemical (IHC) staining images of MFN1 in HCC tissue microarrays. Scale bar: 100 µm. d, e Kaplan-Meier analysis of overall survival (OS) and disease-free survival (DFS) in HCC patients classified by expression levels of MFN1 showed that MFN1 low expression contributes to poor prognosis.
that MFN1 inhibits HCC progression via controlling mitochondria fusion.
MFN1 modulated glucose metabolic reprogramming in HCC Since mitochondria play an essential role in cell metabolism and energy production, we hypothesised that disrupted mitochondria dynamics might influence metabolic reprogramming in HCC cells. We used Seahorse platform to monitor the metabolism state of OE-MFN1 MHCC97-H cells and sh-MFN1 HepG2 cells . We found that MFN1 could induce a glucose metabolic shift from aerobic glycolysis to oxidative phosphorylation , while aerobic glycolysis was promoted when MFN1 was knocked down in HepG2 cells . To better understand this metabolic shift induced by MFN1 in HCC, we measured the expression of metabolic enzymes in glucose metabolism by qRT-PCR. The representative enzymes crucial for oxidative phosphorylation were found highly expressed in OE-MFN1 MHCC97-H cells, while the expression levels of key metabolic enzymes involved in aerobic glycolysis significantly increased in sh-MFN1 HepG2 cells . These results suggest that MFN1 modulates the glucose metabolic reprogramming in HCC cells.
A decrease in anabolic glycolysis is required for the anti-metastatic effect of MFN1 In order to ascertain whether glucose metabolic shift induced by depletion of MFN1 directly affects the function of HCC cells, we used 2-Deoxy-D-glucose (2-DG) to selectively block aerobic glycolysis in HCC cells. 2-DG is a glycolysis inhibitor and competitively inhibits the production of glucose-6-phosphate. HepG2 cells treated with aerobic glycolysis inhibitor 2-DG no longer exhibited high proliferation , clone formation , migration and invasion induced by depletion of MFN1. Furthermore, 2-DG could inhibit EMT enhanced by MFN1 knockdown in HCC cells . Instead, treatment of 2-DG showed no effects on the function of HCC cells with overexpression of MFN1 . Together, we concluded that metabolic shift against aerobic glycolysis would restrain HCC proliferation, invasion and migration. This shift was achieved by MFN1-mediated alteration of mitochondria dynamics.
# Discussion
As a milestone in cancer metabolism research, aerobic glycolysis is first described by Otto Warburg. Warburg effect had long been interpreted as an adaptation for tumour under stringent conditions with impaired mitochondrial respiration.However, growing evidence begins to reshape the view on cancer metabolism reprogramming. It is known that mitochondria are actively functioning in some types of cancer and contribute to cancer progression.Currently, metabolic reprogramming has been regarded as a general hallmark of malignant tumour. Mitochondria play a central role in coordinating the physiology and metabolism of most eukaryotic cells. Mitochondria are morphologically heterogeneous organelles and undergo constant fission and fusion that is closely related to differed functional states.Mitochondrial dynamics has been recognised as an important biological process in carcinogenesis and cancer progression.Mitochondrial dynamics was reported to regulate the migration and invasion of breast cancer.Also, mitochondrial fission promotes autophagy, preventing cancer from apoptosis.In this study, we demonstrated that mitochondrial dynamics was associated with HCC metastasis. High metastatic HCC displayed excessive mitochondrial fission with more fragmented mitochondria.
It is widely known that mitochondrial fission and fusion balance is regulated by several key regulators, including MFN2, DNM1L, OPA1 and MFF.In our current research, we found that these regulators of mitochondrial dynamics showed differed metastatic potential in HCC, and that MFN1 is a leading downregulated gene in mitochondria dynamics closely associated with HCC metastasis. Furthermore, MFN1 was identified as a tumour-suppressor gene, promoting the mitochondrial fusion to suppress the invasion and metastasis both in vitro and in a xenograft model in vivo. The clinical data support the in vitro and xenograft findings showing that downregulation of MFN1 is associated with vascular invasion and poor prognosis of HCC patients. The data would be more intact and convincing to reveal the underlying connection between MFN1 expression and HCC metastasis if we could compare expression of MFN1 in primary and lung metastatic tissues from patients. According to the current clinical practice guidelines, surgical treatment is not recommended to HCC patients with lung metastasis. Thus, it is not feasible for us to acquire samples from these patients. We hope to collect such samples for deeper-insight research in our future study. Previous studies demonstrated that increased DNM1L expression upregulated mitochondrial fission to promote breast cancer progression.In our study, DNM1L was also slightly upregulated in HCC patients with PVTT but not in HCC cell lines with high metastatic potential. Beyond the above, the function of DNM1L on regulating mitochondrial fission relies on phosphorylation of two sites (serine 616 and serine 637);it requires further solid evidence to elucidate the actual function that DNM1L carries in HCC.
Cancer metastasis is a multistep process, influenced by different factors. EMT usually represents early steps of metastasis such as circulating tumour cells (CTC) formation, which is often linked to a poor prognosis in many types of malignancies including cholangiocarcinoma, gastric cancer, breast cancer and so on.Previous studies have demonstrated that EMT is vital as a metastatic driver in different cancers including HCC. In the present study, we found that HCC cells with MFN1 depletion showed Ecadherin downregulation and increased mesenchymal markers, all of which are critical regulators of EMT. These results were further confirmed by in vivo evidences that MFN1 strongly decreased the metastatic potential of HCC cells along with changes in their expression pattern of E-cadherin. Furthermore, we found the correlation between mitochondria dynamics-mediated metabolic reprogramming and EMT activation. We hope that this regulatory mechanism can be elaborated in the future.
Cancer cells exhibit a featured glucose metabolic shift to increased aerobic glycolysis compared with normal tissues. Evidence has been accumulated that key metabolic enzymes involved in aerobic glycolysis and their modifications regulate cancer metastasis.Furthermore, metabolic reprogramming resulted from other processes also influences the metastasis.As mitochondria take the central role in energy metabolism including glucose metabolism, metabolic shift must be closely connected with mitochondrial function, which is largely dependent on mitochondrial dynamics. In the present study, we demonstrated that MFN1 modulated HCC metastasis by metabolic shift from aerobic glycolysis to oxidative phosphorylation, which was attributed to mitochondrial fusion. Treatment with aerobic Quantified analysis on mitochondrial length and ratio analysis based on quantified mitochondria length OE-MFN1 97H cells and sh-MFN1 HepG2 cells confirmed the morphologic changes in IF. n = 3, mean ± SEM, ***P < 0.001, **P < 0.01, *P < 0.05. c, d Cell proliferation assay was assessed by CCK-8 assay in OE-MFN1 97 H cells and sh-MFN1 HepG2 cells. Overexpression of MFN1 inhibited 97H cell proliferation, while knockdown of MFN1 promoted HepG2 cells proliferation. OD, absorbance degrees. n = 3, mean ± SEM, ***P < 0.001, **P < 0.01, *P < 0.05. e Images on plate clone formation assay in OE-MFN1 97H cells and sh-MFN1 HepG2 cells. The results were consistent with cell proliferation assay. f, g Images and quantified column figures on Transwell assay in OE-MFN1 97H cells and sh-MFN1 HepG2 cells. Overexpression of MFN1 inhibited migration and invasion of 97H cells, while knockdown of MFN1 promoted migration and invasion of HepG2 cells. n = 3, mean ± SEM, ***P < 0.001, **P < 0.01, *P < 0.05. glycolysis inhibitor 2-DG could significantly suppress the effects induced by MFN1 depletion. We also confirmed that the expressions of glycolysis-related genes were influenced by mitochondrial dynamics. Though we did not clarify the specific mechanism on how mitochondrial dynamics influenced expression of glycolysis-related genes, according to multiple researches published previously, MFN1/2 can modulate Ca 2+ buffering through ER-mitochondria tethering activity, thereby regulating nuclear translocation and transcriptional activity of multiple genes.This may possibly explain potential modulation mechanism on glycolysis-related genes induced by MFN1. Overall, these findings may provide innovative ideas for integrating how metabolic reprogram influences the metastatic process.
In conclusion, we found that MFN1 suppressed HCC malignancy via driving the balance of mitochondrial dynamics from fission to fusion, which mediated metabolic shift from aerobic glycolysis to . n = 3, mean ± SEM, ***P < 0.001, **P < 0.01, *P < 0.05. d-i Transwell assay (d, e, f, g) and wound-healing assays (h, i) confirmed that 2-DG could reverse the effect of sh-MFN1 to retrain migration and invasion of HepG2 cells. n = 3, mean ± SEM, ***P < 0.001, **P < 0.01, *P < 0.05.
Consent to publish No consent was involved in this publication.
Data availability All data and materials generated during and/or analysed during this study are available from the corresponding author on reasonable request. |
Phase II study of weekly irinotecan and capecitabine treatment in metastatic colorectal cancer patients
Background: The purpose of this phase II study was to evaluate the safety and efficacy of weekly irinotecan and capecitabine (wXELIRI) treatment in patients with metastatic colorectal cancer, specifically the rate of severe diarrhea. Methods: Patients with unresectable histologically confirmed metastatic colorectal cancer with measurable disease received weekly irinotecan 90 mg/m 2 on day 1 and capecitabine 1200 mg/m 2 twice daily on days 1-5. Patients naïve to systemic chemotherapy for metastatic disease or who had failed FOLFOX (infusional 5-fluorouracil [5-FU], leucovorin, and oxaliplatin) or XELOX (capecitabine plus oxaliplatin) as first-line treatment were eligible. The primary endpoint was the rate of grade 3/4 diarrhea. Secondary endpoints included progression-free survival (PFS), overall survival (OS), and safety.Results: A total of 52 patients were enrolled, 30 of whom received wXELIRI as first-line treatment and 22 as second-line treatment. Grade 4 diarrhea was observed in one patient and the rate of grade 3/4 diarrhea was 7.7%. The other common grade 3/4 toxicities included leukopenia (9.6%), neutropenia (17.3%), nausea (3.8%), vomiting (3.8%), fatigue (1.9%), and hand-foot syndrome (1.9%). The median progression-free survival and overall survival for the 30 patients treated in the first-line setting was 8.5 and 16.3 months, while those for the 22 patients treated in the second-line setting was 5.0 and 10.7 months, respectively.Conclusions:The wXELIRI regimen resulted in a low rate of severe diarrhea with an acceptable toxicity profile. This study provides a basis for a subsequent randomized controlled study of wXELIRI versus FOLFIRI (irinotecan, 5-FU, and folinic acid) to further explore the efficacy and safety of this regimen.Trial registration: ClinicalTrials.gov: NCT01322152.
# Background
Colorectal cancer, the second leading cause of cancer death in the USA, has increased in frequency in China in recent years. Based on a registry in Shanghai, a city with a population of 23 million, colorectal cancer has become the third most prevalent malignancy. Approximately 50-60% of patients diagnosed with colorectal cancer will develop metastases [bib_ref] Towards a pan-European consensus on the treatment of patients with colorectal liver..., Van Cutsem [/bib_ref] [bib_ref] Liver resection for metastatic colorectal cancer in the age of neoadjuvant chemotherapy..., Yoo [/bib_ref] , for whom systematic chemotherapy is the standard treatment.
Irinotecan combined with continuous infusion of 5fluorouracil (5-FU) and folinic acid (the FOLFIRI regimen) is the established option for first-and/or second-line treatment of metastatic colorectal cancer. Capecitabine, an oral fluoropyrimidine that mimics continuous 5-FU infusion by generating 5-FU preferentially in the tumor tissues [bib_ref] Design of a novel oral fluoropyrimidine carbamate, capecitabine, which generates 5-fluorouracil selectively..., Miwa [/bib_ref] , has been shown to have comparable efficacy to 5-FU/folinic acid as first-line treatment of metastatic colorectal cancer, with an additional benefit of convenient administration without hospitalization [bib_ref] Verweij J: Capecitabine, an oral fluoropyrimidine carbamate with substantial activity in advanced..., Van Cutsem [/bib_ref]. The combination of irinotecan and capecitabine (XELIRI) has been assessed, but the associated gastrointestinal toxicity, especially the incidence of severe diarrhea, affected the feasibility of this regimen [bib_ref] Randomized, controlled trial of irinotecan plus infusional, bolus, or oral fluoropyrimidines in..., Fuchs [/bib_ref] [bib_ref] Van Cutsem E: Irinotecan combined with infusional 5-fluorouracil/ folinic acid or capecitabine..., Kohne [/bib_ref]. The incidence of grade 3/4 diarrhea was higher (17-36% vs 12-15%) with a 3-week XELIRI regimen than with the FOLFIRI regimen [bib_ref] A randomized phase II trial of capecitabine and two different schedules of..., Borner [/bib_ref] [bib_ref] Randomized multicenter Phase II trial of two different schedules of irinotecan combined..., Bajetta [/bib_ref] [bib_ref] Results of a phase II open-label study of capecitabine in combination with..., Cartwright [/bib_ref] [bib_ref] Capecitabine plus 3-weekly irinotecan (XELIRI regimen) as first-line chemotherapy for metastatic colorectal..., Patt [/bib_ref] , and non-superior time to progression (TTP; 6-9 months vs 6.7-8.5 months) and overall survival (OS; 13-20 months vs 14.8-17.4 months) were observed [bib_ref] Irinotecan plus fluorouracil and leucovorin for metastatic colorectal cancer. Irinotecan Study Group, Saltz [/bib_ref] [bib_ref] Irinotecan combined with fluorouracil compared with fluorouracil alone as first-line treatment for..., Douillard [/bib_ref] [bib_ref] CPT-11 (irinotecan) addition to bimonthly, high-dose leucovorin and bolus and continuousinfusion 5-fluorouracil..., Andre [/bib_ref]. Toxicityinduced dose reduction and treatment delay weakened the efficacy of the XELIRI regimen.
To reduce the side effect of diarrhea, a 2-week XELIRI regimen was tried recently. The 2-week regimen (irinotecan on day 1, capecitabine on days 2-8 or days 1-5 and 8-12) exhibited promising activity (TTP, 8-10 months, OS, 15-19 months) with improved tolerability (grade 3/4 diarrhea, 8.1-15.0%) [bib_ref] administered as a 2-weekly schedule, as first-line chemotherapy for patients with metastatic..., Garcia-Alfonso [/bib_ref] [bib_ref] Efficacy of combination chemotherapy with irinotecan (CPT-11) plus capecitabine in patients with..., Choi [/bib_ref] , but an increased rate of severe diarrhea was noted in elderly patients, which resulted in dose reduction [bib_ref] administered as a 2-weekly schedule, as first-line chemotherapy for patients with metastatic..., Garcia-Alfonso [/bib_ref].
This study aimed to evaluate the tolerability of a weekly XELIRI regimen as first-or second-line treatment in patients with metastatic colorectal cancer. The dose and schedule was chosen based on the assumption that 5 days on/2 days off administration of capecitabine may better mimic continuous 5-FU infusion [bib_ref] Efficacy of combination chemotherapy with irinotecan (CPT-11) plus capecitabine in patients with..., Choi [/bib_ref]. The dose of irinotecan was calculated as a weekly dose according to the FOLFIRI regimen [bib_ref] CPT-11 (irinotecan) addition to bimonthly, high-dose leucovorin and bolus and continuousinfusion 5-fluorouracil..., Andre [/bib_ref]. The study investigated the possibility of a further reduction of the rate of severe diarrhea with weekly XELIRI treatment, and evaluated the safety and efficacy of this schedule in Chinese patients with metastatic colorectal cancer.
# Methods
## Patients
Patients aged 18-70 years with histologically or cytologically confirmed advanced colorectal adenocarcinoma, with an Eastern Cooperative Oncology Group performance status of 0 to 1 and life expectancy of ≥ 3 months were eligible. Patients who were chemotherapy-naive or who had failed first-line treatment with either XELOX or FOLFOX were enrolled if they had at least one measurable disease lesion according to the Response Evaluation Criteria in Solid Tumors (RECIST) criteria, version 1.1. Previous (neo) adjuvant chemotherapy was permitted if it had been completed ≥ 6 months before enrollment, although prior irinotecan therapy was not allowed. Patients were required to have adequate bone marrow, hepatic, and renal function. Patients with previous chronic inflammatory bowel disease, chronic diarrhea or recurrent bowel obstruction, pelvic radiotherapy within 6 months, or symptomatic brain metastases were excluded. The study was approved by the independent ethics committee of Fudan University Shanghai Cancer Center, Shanghai, China and registered at ClinicalTrials. gov (NCT01322152). The study was carried out in accordance with the Declaration of Helsinki. All patients provided written informed consent before study entry.
## Treatment
All enrolled patients received a weekly regimen of irinotecan and capecitabine (wXELIRI), as follows: irinotecan (Campto®, Pfizer) 90 mg/m 2 given intravenously on day 1; capecitabine (Xeloda®, Roche) 1200 mg/m 2 given orally twice daily on days 1-5. The treatment cycles were administered every week until disease progression or unacceptable toxicity, or consent withdrawal. For patients with poor tolerance to toxicities, treatment delay was permitted for no more than 2 weeks.
Unless contraindicated, atropine could be given to prevent the cholinergic adverse effects (including early diarrhea). Loperamide was recommended as the standard anti-diarrhea treatment, and other symptomatic treatment could be given according to the institution's practice guidelines.
## Dose modification
Dose adjustments were made based on the worst grade of toxicity encountered during the previous cycle. For hematological toxicities, the dose of chemotherapeutic drugs was reduced in the following cases: grade 4 neutropenia or leukopenia; grade 3 or greater febrile neutropenia; grade 3 or greater thrombocytopenia. For nonhematological toxicity, the dose of related drug was reduced when grade 3 or greater toxicities occurred (except for alopecia). The dose of irinotecan or capecitabine was reduced by 25% of the starting dose. If a patient required more than two successive dose reductions, therapy was dicontinued.
Treatment was delayed until the absolute neutrophil count was ≥ 1.5 × 10 9 /L and platelet count was ≥ 80 × 10 9 /L, and recovery to grade ≤ 1 for mucositis, diarrhea and other toxicities (with exception of alopecia).The maximum authorized delay is of 2 weeks.
## Assessments
Pretreatment assessment included a detailed medical history, physical examination, routine laboratory tests, and performance status. Laboratory evaluation included a routine blood count, urinalysis, and electrolyte, renal, and liver function tests. Adverse events and concomitant medications were recorded at the end of each cycle. Toxicity was evaluated and graded according to the National Cancer Institute Common Terminology Criteria for Adverse Events, version 4.0.
Radiographic scans (computed tomography or magnetic resonance imaging) for efficacy evaluation were conducted at baseline and every 2 months thereafter according to the RECIST 1.1 guidelines. The best overall response was reported. Survival status was assessed every 3 months after discontinuation of study treatment.
# Statistical analysis
This phase II study was designed to assess the rate of severe diarrhea (calculated as the percentage of patients with grade 3 and grade 4 diarrhea) with the wXELIRI regimen. The rate of severe diarrhea with the 2-week XELIRI or the FOLFIRI regimens was reported as 15% [bib_ref] administered as a 2-weekly schedule, as first-line chemotherapy for patients with metastatic..., Garcia-Alfonso [/bib_ref] [bib_ref] A: FOLFIRI followed by FOLFOX6 or the reverse sequence in advanced colorectal..., Tournigand [/bib_ref] , and we supposed that the rate with the wXE-LIRI regimen was, at most, 5%. A one-stage Fleming design, with an exact significance level of p = 0.05 and a power of 80%, was used to test the hypothesis. With a sample size of 52 evaluable patients, the regimen would be declared promising if less than 6 patients with severe diarrhea were observed.
The efficacy analysis included all patients who received at least one dose of study medication and had at least one efficacy evaluation after baseline. For the safety analysis, all patients who received one dose of study medication were included. The primary study end point was the rate of severe diarrhea. Secondary end points were PFS (defined as the time between the first dose of study medication and first documented disease progression or death), ORR, DCR (defined as the percentage of patients with CR, PR, and SD for at least 8 weeks), OS (defined as the time between the first dose of study medication and death), and safety.
Survival function of time-to-event end points was estimated by using the Kaplan-Meier method. Chi-square test was performed for enumeration data on response rate and clinical benefit rate.
# Results
## Patients
From March 2011 to October 2012, a total of 52 patients (25 men and 27 women), aged from 26 to 70 years (median, 60 years) with advanced colorectal cancer received wXELIRI treatment. Previous chemotherapy regimens for the patients receiving second-line treatment included FOLFOX (infusional 5-FU, leucovorin, and oxaliplatin) and XELOX (capecitabine plus oxaliplatin). [fig_ref] Table 1: Clinical and demographic characteristics at enrollment [/fig_ref] shows the patients' demographic and clinical characteristics at enrollment.
## Treatment
Up to 31 April 2013, the 52 enrolled patients had received a total of 644 cycles of wXELIRI. The median number of treatment cycles was 12 (range, 1-50). Median dose
## Safety
The overall incidence of grade 3/4 adverse events was 55.5%. The most common adverse events (≥20%) were neutropenia, leucopenia, diarrhea, nausea, vomiting, anorexia, and fatigue [fig_ref] Table 2: Incidence of adverse reactions caused by weekly irinotecan and capecitabine [/fig_ref]. The rate of severe diarrhea, the primary study end point, was 7.7% (6.7% and 9.0% for the first-and second-line settings, respectively), with one case of grade 4 diarrhea in the second-line setting . Grade 3 diarrhea occurred in three patients, one each during the first, third, and fifth treatment cycles, and was relieved by symptomatic treatment for diarrhea and dehydration. Subsequent re-treatment did not result in diarrhea greater than grade 2. The incidence of grade 1/2 diarrhea was 40.4%, and the symptoms were relieved after standard use of loperamide and symptomatic supportive treatment in most patients. Nausea and vomiting was another frequent gastrointestinal reaction, with an incidence of 38.5-42.3%. After prophylactic administration of an antiemetic drug (5-hydroxytryptamin 3 [5-HT 3] receptor antagonists), nausea and vomiting was tolerable in most cases. One patient had dose reduction and delayed treatment due to intolerable grade 3 vomiting.
The major hematological toxicities were leucopenia and neutropenia. Three patients developed grade 4 neutropenia and one developed grade 3 febrile neutropenia. After symptomatic therapy and prophylactic antibiotic treatment, all patients recovered. The incidence of mild to moderate anemia was 38.5%. One patient discontinued treatment because of grade 3 anemia complicated by grade 3 fatigue after 5 cycles of wXELIRI, and one patient had the capecitabine dose reduced due to the grade 3 hand-foot syndrome after 16 cycles.shows the adverse reactions.
## Efficacy
After a median follow-up of 13.9 months (range, 1-24 months), 25 of 30 patients (83.3%) treated with wXELIRI in the first-line setting experienced disease progression, and 16 patients (53.3%) died [fig_ref] Figure 1: Overall survival and progression-free survival in the first-line setting [/fig_ref]. Ten patients experienced partial response (PR) and 11 patients had stable disease (SD), whereas no complete response (CR) was observed and 3 patients did not have response evaluation due to withdrawal of consent after two cycles. The objective response rate (ORR) was 37% (10/27 patients) and the disease control rate (DCR) was 77.7% (21/27 patients). Secondary R0 metastasectomy was performed in 2 patients after 6 cycles. In patients pretreated with FOLFOX or XELOX (n = 22), the median PFS was 5.0 months (95% CI: 1.74-8.26 months) and the median OS was 10.7 months (95% CI: 5.80-15.60 months), with a median follow-up period of 13.8 months (range, 1-17 months) [fig_ref] Figure 2: Overall survival and progression-free survival in the second-line setting [/fig_ref]. Among the 21 evaluable patients in the second-line setting, three (14.3%) achieved PR, and 13 (61.9%) had stable disease, whereas five (23.8%) had disease progression at the first efficacy evaluation. One patient withdrew from the study due to personal reasons after one cycle.
# Discussion
This study demonstrated that patients with colorectal cancer experienced a relatively low incidence of severe diarrhea with the wXELIRI regimen as first-or secondline treatment. The incidence of grade 3 and grade 4 diarrhea was 5.8% and 1.9%, respectively. Among the 30 patients who received the study treatment in the firstline setting, the rate of grade 3 diarrhea was only 6.7% and no grade 4 diarrhea was observed. This was much lower than that observed with the 3-week or 2-week XELIRI [bib_ref] A randomized phase II trial of capecitabine and two different schedules of..., Borner [/bib_ref] [bib_ref] Results of a phase II open-label study of capecitabine in combination with..., Cartwright [/bib_ref] [bib_ref] Capecitabine plus 3-weekly irinotecan (XELIRI regimen) as first-line chemotherapy for metastatic colorectal..., Patt [/bib_ref] [bib_ref] administered as a 2-weekly schedule, as first-line chemotherapy for patients with metastatic..., Garcia-Alfonso [/bib_ref] and FOLFIRI [bib_ref] Irinotecan plus fluorouracil and leucovorin for metastatic colorectal cancer. Irinotecan Study Group, Saltz [/bib_ref] [bib_ref] Irinotecan combined with fluorouracil compared with fluorouracil alone as first-line treatment for..., Douillard [/bib_ref] regimens. In Incidence of diarrhea after first-and second-line weekly irinotecan and capecitabine the MAC-6 study [bib_ref] Efficacy of combination chemotherapy with irinotecan (CPT-11) plus capecitabine in patients with..., Choi [/bib_ref] , the 5 days on/2 days off administration of capecitabine was introduced to ensure the absence of a completely drug-free interval during treatment, while maintaining a reasonable dose intensity, and a rate of grade 3 diarrhea of 8.1% was obtained. We have also demonstrated a similar low incidence of severe diarrhea in this series when using the same dosing schedule of capecitabine and dividing the dose of irinotecan from the FOLFIRI regimen into once weekly administration.
In the literature, the median time for delayed diarrhea caused by irinotecan was 6-14 days after administration. Diarrhea may be aggravated in the second week when using the traditional regimen of capecitabine from days 1 to 14 every 21 days, which may result in discontinuation of the oral fluoropyrimidine causing under-dosage.
Our results indicate that weekly use of capecitabine is feasible and tolerable with less drug interruption when combined with irinotecan. We analyzed the efficacy and survival data according to the treatment setting. Survival in this study was similar to that in previous studies in patients undergoing first-line treatment with irinotecan and capecitabine [fig_ref] Table 5: Summary of clinical trials of capecitabine and irinotecan in the first-line setting... [/fig_ref]. About 10% of the patients in our study accepted targeted therapy as subsequent treatment. The median OS for first-line treatment was comparable with the published data for patients undergoing chemotherapy [bib_ref] XELIRI-bevacizumab versus FOLFIRI-bevacizumab as first-line treatment in patients with metastatic colorectal cancer:..., Pectasides [/bib_ref] [bib_ref] Cetuximab and chemotherapy as initial treatment for metastatic colorectal cancer, Van Cutsem [/bib_ref] [bib_ref] Bevacizumab plus irinotecan, fluorouracil, and leucovorin for metastatic colorectal cancer, Hurwitz [/bib_ref] [bib_ref] Randomized, phase III trial of panitumumab with infusional fluorouracil, leucovorin, and oxaliplatin..., Douillard [/bib_ref]. In the second-line setting, wXELIRI showed promising efficacy. Among 22 patients who had failed prior oxaliplatin, the ORR was 14.3% and the median PFS was 5.0 months, which were significantly superior to the ORR of 4% and PFS of 2.5 months with FOLFIRI after failure of FOLFOX in the GERGOR study [bib_ref] A: FOLFIRI followed by FOLFOX6 or the reverse sequence in advanced colorectal..., Tournigand [/bib_ref] , and the median OS exceeded 10 months. Given these data from a limited number of patients, it would be interesting to ascertain the survival data for second-line treatment of colorectal cancer in a greater number of patients.
The low rate of severe diarrhea enabled a greater intensity of chemotherapy to be achieved, thus ensuring improved efficacy; a treatment delay of only 4.7% was observed in this study. Patients experienced more grade 3/4 neutrophilia than in the MAC-6 study [bib_ref] Efficacy of combination chemotherapy with irinotecan (CPT-11) plus capecitabine in patients with..., Choi [/bib_ref] , but hematological suppression was easily controlled with granulocyte colony-stimulating factor support.
It should be admitted that the doses of the weekly regimen of irinotecan and capecitabine in our study were not established through a formal phase I study. However, the tolerable dose for weely-used irinotecan was evaluated in the previous phase I trial, of which irinotecan with a dose of 100 mg/m 2 was given on days 1 and 8, and capecitabine with 2000 mg/m 2 on days 1-14 of a 3-week cycle was recommended. In our study, the dose of irinotecan was chosed based on that in the FOLFIRI regimen with 180 mg/m 2 for 2 weeks [bib_ref] A: FOLFIRI followed by FOLFOX6 or the reverse sequence in advanced colorectal..., Tournigand [/bib_ref] , equal to 90 mg/m 2 weekly. Although the planned dose intensities of irinotecan and capecitabine with wXELIRI were higher than those with the biweekly [bib_ref] administered as a 2-weekly schedule, as first-line chemotherapy for patients with metastatic..., Garcia-Alfonso [/bib_ref] or every three weeks [bib_ref] Results of a phase II open-label study of capecitabine in combination with..., Cartwright [/bib_ref] [bib_ref] Capecitabine plus 3-weekly irinotecan (XELIRI regimen) as first-line chemotherapy for metastatic colorectal..., Patt [/bib_ref] irinotecan/capecitabine regimen, the rate of diarrhea decreased while the survival data was similar, suggesting the feasibility of weekly used regimen.
Additional use of targeted therapy based on irinotecan and capecitabine is a new direction [bib_ref] XELIRI-bevacizumab versus FOLFIRI-bevacizumab as first-line treatment in patients with metastatic colorectal cancer:..., Pectasides [/bib_ref] [bib_ref] Efficacy and safety of bevacizumab-based combination regimens in patients with previously untreated..., Ducreux [/bib_ref] [bib_ref] Randomised phase-II trial of CAPIRI (capecitabine, irinotecan) plus bevacizumab vs FOLFIRI (folinic..., Souglakos [/bib_ref] [bib_ref] Cetuximab plus capecitabine and irinotecan compared with cetuximab plus capecitabine and oxaliplatin..., Moosmann [/bib_ref]. Recent study data indicated that the 3-week XELIRI regimen combined with bevacizumab had comparable efficacy to FOLFIRI combined with bevacizumab, with incidences of grade 3/4 diarrhea and agranulocytosis of 19% and 13%, respectively [bib_ref] XELIRI-bevacizumab versus FOLFIRI-bevacizumab as first-line treatment in patients with metastatic colorectal cancer:..., Pectasides [/bib_ref]. The XELIRI plus cetuximab regimen also achieved good efficacy with an OS of more than 20 months [bib_ref] Cetuximab plus capecitabine and irinotecan compared with cetuximab plus capecitabine and oxaliplatin..., Moosmann [/bib_ref] , but poor tolerance was a concern as 32% of patients required dose reduction [bib_ref] Results of a phase II trial of cetuximab plus capecitabine/irinotecan as first-line..., Cartwright [/bib_ref]. Whether the wXELIRI regimen could become an alternative to combination therapy with targeted drugs needs further investigation.
# Conclusions
The weekly irinotecan and capecitabine combination is associated with a low incidence of severe diarrhea and has an acceptable toxicity profile. wXELIRI could be an alternative regimen for patients with metastatic colorectal cancer, especially in the second-line setting. Further randomized controlled studies are needed to evaluate the efficacy and safety of this regimen in a larger sample size.
## Competing interests
The authors declare that they have no competing interests.
Authors' contributions WL participated in acquisition of data, analysis and interpretation of data, and drafting of the manuscript. JX, LS, TL, WZ, WG, ZC, XZ carried out the study during clinical observation, follow up, collected the clinical data for analysis. JL conceived of the study, participated in its design and coordination and revised the final manuscript. All authors have read and approved the final manuscript.
Authors' information JL: MD, PhD, Professor, Director of Department of Medical Oncology, Fudan University Shanghai Cancer Center.
## Acknowledgements
The study was supported by CSCO-Tonghui 2011 Fund (No: Y-H2010-013), awarded by Chinese Society of Clinical Oncology.
[fig] Figure 1: Overall survival and progression-free survival in the first-line setting. (a) progression-free survival; (b) overall survival. [/fig]
[fig] Figure 2: Overall survival and progression-free survival in the second-line setting. (a) progression-free survival; (b) overall survival. [/fig]
[table] Table 1: Clinical and demographic characteristics at enrollment (N = 52) ECOG: Eastern Cooperative Oncology Group. [/table]
[table] Table 2: Incidence of adverse reactions caused by weekly irinotecan and capecitabine (N = 52) [/table]
[table] Table 4: Analysis of efficacy of weekly irinotecan and capecitabine as first-or second line treatment [/table]
[table] Table 5: Summary of clinical trials of capecitabine and irinotecan in the first-line setting CAPIRI: capecitabine and irinotecan; ORR: objective response rate; OS: overall survival; PFS: progression-free survival; TTP: time to progression; XELIRI: irinotecan and capecitabine. [/table]
|
A proof-of-concept clinical study examining the NRF2 activator sulforaphane against neutrophilic airway inflammation
Sulforaphane (SFN), a naturally occurring isothiocyanate found in cruciferous vegetables, is implicated as a possible therapy for airway inflammation via induction of the transcription factor NF-E2-related factor 2 (NRF2). In this proof-of-concept clinical study, we show that supplementation of SFN with broccoli sprout homogenate in healthy human subjects did not induce expression of antioxidant genes or protect against neutrophilic airway inflammation in an ozone-exposure model. Therefore, dietary sulforaphane supplementation is not a promising candidate for larger scale clinical trials targeting airway inflammation.
# Introduction
Dear Editor, Asthma is a heterogeneous chronic disease that can be stratified based on features such as eosinophil or neutrophil predominance, and responsiveness to corticosteroids. Current available therapies including corticosteroids are not as effective for certain forms of the disease, particularly neutrophil-predominant asthma. Furthermore, due to negative perceptions of corticosteroids, the use of complementary and alternative medicine and nutritional interventions for asthma is increasing in the U.S [bib_ref] Use of complementary and alternative medicine in children with asthma, Torres-Llenza [/bib_ref]. The naturally occurring isothiocyanate, sulforaphane (SFN), is found in cruciferous vegetables and has been implicated as a possible therapy for airway inflammation via induction of the transcription factor NF-E2-related factor 2 (NRF2), which regulates expression of cytoprotective phase II antioxidant enzymes. The relevance of targeting antioxidant gene expression extends to other airway diseases as well, such as COPD, which is characterized by oxidative stress and dysregulation of antioxidant gene expression [bib_ref] Oxidative stress in COPD, Kirkham [/bib_ref]. However, there are conflicting reports concerning the ability of SFN to induce antioxidant gene expression, and its effectiveness against airway inflammation [bib_ref] A Randomized Controlled Trial of the Effect of Broccoli Sprouts on Antioxidant..., Sudini [/bib_ref] [bib_ref] Oral sulforaphane increases Phase II antioxidant enzymes in the human upper airway, Riedl [/bib_ref] [bib_ref] Sulforaphane-rich broccoli sprout extract attenuates nasal allergic response to diesel exhaust particles, Heber [/bib_ref]. In this brief communication, we report our findings from a proof-of-concept study examining if in vivo supplementation with SFN with broccoli sprout homogenate (BSH) is an effective intervention for ozone (O 3 )induced airway inflammation, a model of neutrophilic airway inflammation. Oxidative injury is especially relevant for those with asthma, as antioxidant reserve may be impaired in this population. O 3 inhalation causes significant airway neutrophilia in healthy non-asthmatic persons [bib_ref] Comparative airway inflammatory response of normal volunteers to ozone and lipopolysaccharide challenge, Hernandez [/bib_ref] , making this a useful model for neutrophilic airway disease.
# Methods
For this randomized, placebo-controlled, crossover study we recruited 16 non-atopic, non-smoking healthy volunteers between the ages of 18-50 years. The O 3 study protocol was approved by the University of North Carolina's Institutional Review Board, and written informed consent was obtained from all study subjects. All subjects underwent a standardized screening protocol including allergy skin testing and methacholine challenge as previously described [bib_ref] Vitamin E, gamma-tocopherol, reduces airway neutrophil recruitment after inhaled endotoxin challenge in..., Hernandez [/bib_ref] [bib_ref] Atopic asthmatic patients have reduced airway inflammatory cell recruitment after inhaled endotoxin..., Hernandez [/bib_ref]. Volunteers were randomized in a 1:1 ratio to consume either 200 g of BSH (equivalent to 111 g of commercially available Broccospr-outs® (Brassica Protection Products LLC)), or 200 g of alfalfa sprout homogenate (ASH), which lacks SFN. The dose of BSH and ASH were chosen based on results of a prior study that found maximal induction of NRF2dependent gene expression by BSH with a 3 day 200-g dosing regimen [bib_ref] Oral sulforaphane increases Phase II antioxidant enzymes in the human upper airway, Riedl [/bib_ref]. Subjects received supplements once daily for 3 days during the initial study period, and the alternate treatment during the crossover period. On the third day of supplementation, each subject was exposed to O 3 (0.4 ppm) for 2 h while performing four 15 min sessions of intermittent moderate exercise (defined as minute ventilation or VE min = 30-40 L/min) on a treadmill, separated by 15 min of seated rest. Induced sputum was obtained at screening and at 4 h post-O 3 exposure and processed for measurement of cytokines and cell counts as previously described [bib_ref] Vitamin E, gamma-tocopherol, reduces airway neutrophil recruitment after inhaled endotoxin challenge in..., Hernandez [/bib_ref] [bib_ref] Atopic asthmatic patients have reduced airway inflammatory cell recruitment after inhaled endotoxin..., Hernandez [/bib_ref]. Blood was collected at screening and post-O 3 for determination of SFN and SFN-conjugate levels by mass spectroscopy. Additionally, blood and nasal epithelial cells (NECs) were collected 4 h post-O3 to measure NRF2-regulated gene expression (HO-1, NQO-1, GSTM-1). There was a minimum washout period of 14 days between treatment periods.
# Results
Our primary hypothesis was that NRF2 activation with SFN would decrease %PMNs in induced sputum compared to placebo after O 3 exposure. The primary endpoint for this study was the effect of SFN compared to placebo on the O 3 -induced change (post-O 3 minus pre-O 3 ) in %PMNs in airway sputum. To analyze the treatment effect on sputum cellularity, we compared active (BSH) to placebo (ASH) treatment using a linear mixed model approach [bib_ref] Design and Analysis of Crossover Trials, Jones [/bib_ref]. Comparisons between post-O 3 active or placebo treatment to baseline values were carried out using Wilcoxon-Signed rank tests. Criterion for significance was taken to be p ≤ 0.05.
Sixteen subjects were randomized, and fifteen subjects completed all visits. There were no serious adverse events during the course of the study. Three days of supplementation with BSH significantly increased levels of SFN (p = 0.001) and its major metabolites, SFN-N-acetyl-Lcysteine (p = 0.002) and SFN-glutathione (p < 0.001) compared to placebo [fig_ref] Figure 1: Despite significantly increased plasma levels of SFN in the BSH group, post-O... [/fig_ref]. O 3 exposure significantly increased the quantity of neutrophils in sputum (expressed as neutrophils/mg and %PMN) in both the placebo and BSH treatment groups [fig_ref] Figure 1: Despite significantly increased plasma levels of SFN in the BSH group, post-O... [/fig_ref] , but the BSH supplementation group showed no significant difference in sputum neutrophilia compared to placebo. [fig_ref] Table 1 %: Change antioxidant gene expression in healthy volunteers following O3 exposure Data are... [/fig_ref].
# Discussion
SFN has received significant attention in recent years as a possible intervention for oxidant-induced airway inflammation through induction of NRF2-regulated antioxidant genes, but reports concerning its ability to induce antioxidant gene expression and protect against airway inflammation are conflicting. Possible explanations for these contradictory results include variable dosing, dosage forms, and differential biological responses in diseased versus healthy populations. Our study utilized a similar BSH preparation and dosing schedule as Reidl et al., in which 200 g of BSH was ingested daily for three days by healthy volunteers [bib_ref] Oral sulforaphane increases Phase II antioxidant enzymes in the human upper airway, Riedl [/bib_ref]. This preparation reportedly delivered 102 μmol SFN per dose. In contrast to Reidl et al., we saw no differences in phase II enzyme expression in NECs or peripheral blood. Furthermore, BSH supplementation had no impact on lower airway inflammation, as determined by O 3 -induced changes in sputum neutrophilia. Our results are in agreement with Sudini et al., in which ingestion of 100 g of whole broccoli sprouts daily by allergic asthmatics for 3 days had no effect on either NRF2-dependent gene expression in NECs and PBMCs, or eosinophilic lower airway inflammation (measured by FENO) [bib_ref] A Randomized Controlled Trial of the Effect of Broccoli Sprouts on Antioxidant..., Sudini [/bib_ref]. On the other hand, supplementation of SFN using a standardized dose of broccoli sprout extract inhibited nasal inflammatory responses to diesel exhaust in cat-allergic subjects [bib_ref] Sulforaphane-rich broccoli sprout extract attenuates nasal allergic response to diesel exhaust particles, Heber [/bib_ref]. These contradictory results may be due to differing systemic levels of SFN achieved with dietary supplementation with BSH, which is not standardized. However, because variability exists in the timing and methods of detection for SFN conjugate levels, it is difficult to compare systemic SFN levels across studies.
Similar to our study, other groups have demonstrated marked increases in SFN conjugate levels following in vivo supplementation with BSH with minimal effects on antioxidant gene expression [bib_ref] A Randomized Controlled Trial of the Effect of Broccoli Sprouts on Antioxidant..., Sudini [/bib_ref] [bib_ref] Absorption and chemopreventive targets of sulforaphane in humans following consumption of broccoli..., Atwell [/bib_ref]. It is possible that the plasma levels of SFN achieved with our BSH supplementation regimen were not sufficiently high to be biologically active. A dosing study using fresh broccoli sprouts that achieved significantly higher peak plasma levels of SFN metabolites found no significant increases in antioxidant gene expression in whole blood [bib_ref] Absorption and chemopreventive targets of sulforaphane in humans following consumption of broccoli..., Atwell [/bib_ref]. Although several in vitro studies report induction of NRF2 genes with SFN treatment, it is important to note that many of these studies utilize concentrations of SFN in the micromolar range [bib_ref] Temporal changes in gene expression induced by sulforaphane in human prostate cancer..., Bhamre [/bib_ref] [bib_ref] Sulforaphane and phenylethyl isothiocyanate protect human skin against UVR-induced oxidative stress and..., Kleszczynski [/bib_ref] [bib_ref] Potent induction of phase 2 enzymes in human prostate cells by sulforaphane, Brooks [/bib_ref] [bib_ref] Differential responses from seven mammalian cell lines to the treatments of detoxifying..., Jiang [/bib_ref]. Plasma levels of SFN achieved in our study are several orders of magnitude lower than those used in vitro; furthermore, levels achieved in target tissues are likely less than those achieved in plasma. Therefore, doses of BSH that can be reasonably consumed by an adult may exhibit little biologic activity.
# Conclusions
In summary, dietary supplementation of SFN with BSH did not induce expression of NRF2-regulated genes, or have protective effects with O 3 exposure, a model of neutrophilic airway inflammation. Collectively, these findings suggest that SFN supplementation with BSH is not a promising candidate for larger scale clinical trials targeting airway inflammation.
Abbreviations ASH, alfalfa sprout homogenate; BSH, broccoli sprout homogenate; NECs, nasal epithelial cells; NRF2, NF-E2-related factor 2; SFN, sulforaphane Funding MLH is supported by NIEHS K23-ES021745. AJB is supported by 5T32GM086330. This project was supported in part by grants R01ES012706 and P30ES010126 from the National Institute of Environmental Health Sciences. This work was also funded by CR 83578501 from the US Environmental Protection Agency.
## Availability of data and materials
The datasets supporting the conclusions of this article are included within the article.
Authors' contributions MLH, DBP, and MJK contributed to the conception and design of the study. KHM, MMA, CGD, and HRD were involved in acquisition of data. YP, HZ, MLH, CGD, AJB, and MMA were involved in the analysis and interpretation of the data. All listed authors were involved in drafting and revising the manuscript critically for important intellectual content. All authors read and approved the final manuscript.
## Competing interests
The authors declare that they have no competing interests.
[fig] Figure 1: Despite significantly increased plasma levels of SFN in the BSH group, post-O 3 gene expression of NRF2 and Plasma levels of SFN and its major metabolites SFN-N-acetylcysteine and SFN-glutathione following BSH supplementation (a). O 3 -induced changes in sputum neutrophil counts with placebo or SFN supplementation (N = 14) (b). Changes are presented as PMNs/mg sputum and %PMNs/mg sputum phase II antioxidant defense genes in NECs and peripheral blood were not significantly different from placebo [/fig]
[table] Table 1 %: Change antioxidant gene expression in healthy volunteers following O3 exposure Data are shown as mean ± SEM Changes in GSTM1 expression were performed only on GSTM1 sufficient subjects. For all genes, N = 6-14 *Comparisons between Placebo and SFN groups were carried out using paired Wilcoxon-Signed rank tests [/table]
|
Indoleamine 2,3‐dioxygenase 1 (IDO1): an up‐to‐date overview of an eclectic immunoregulatory enzyme
Indoleamine 2,3-dioxygenase 1 (IDO1) catalyzes the initial rate-limiting step in the degradation of the essential amino acid tryptophan along the kynurenine pathway. When discovered more than 50 years ago, IDO1 was thought to be an effector molecule capable of mediating a survival strategy based on the deprivation of bacteria and tumor cells of the essential amino acid tryptophan. Since 1998, when tryptophan catabolism was discovered to be crucially involved in the maintenance of maternal T-cell tolerance, IDO1 has become the focus of several laboratories around the world. Indeed, IDO1 is now considered as an authentic immune regulator not only in pregnancy, but also in autoimmune diseases, chronic inflammation, and tumor immunity. However, in the last years, a bulk of new information-including structural, biological, and functional evidence-on IDO1 has come to light. For instance, we now know that IDO1 has a peculiar conformational plasticity and, in addition to a complex and highly regulated catalytic activity, is capable of performing a nonenzymic function that reprograms the expression profile of immune cells toward a highly immunoregulatory phenotype. With this state-of-the-art review, we aimed at gathering the most recent information obtained for this eclectic protein as well as at highlighting the major unresolved questions.
# Introduction
Indoleamine 2,3-dioxygenase 1 (IDO1) is a cytosolic, monomeric, heme-containing enzyme that catalyzes the initial rate-limiting step in the degradation of the essential amino acid L-tryptophan (L-Trp) along a pathway known as the kynurenine pathway (KP) . This pathway is a cascade of enzymatic steps that produces several biologically active molecules, such as L-kynurenine (L-Kyn), and can lead to the endogenous production of nicotinamide adenine dinucleotide (NAD + ). IDO1 is normally an inducible enzyme and its most important inducer is the cytokine interferon-c (IFN-c).
IDO1 was discovered in the 60' by a Japanese group led by Osamu Hayaishi, along a series of studies on oxygenases in rabbit intestine [bib_ref] Tryptophan pyrrolase of rabbit intestine. D-and L-tryptophan-cleaving enzyme or enzymes, Yamamoto [/bib_ref]. However, the year that changed the perception of IDO1 by the scientific community is 1998, when David Munn and Andrew Mellor performed a pioneering experiment demonstrating that IDO1 is crucially involved in the maintenance of maternal T-cell tolerance [bib_ref] Prevention of allogeneic fetal rejection by tryptophan catabolism, Munn [/bib_ref]. These data raised the interest on this enzyme to such an extent that the number of total publications on IDO1 has increased enormously since the early '90s.
Recent evidence indicates that the IDO1 biology is more complex than initially assumed. For instance, in human monocyte-derived macrophages and tumor cells either as such or stimulated with IFN-c, the majority of IDO1 protein molecules do not bind heme, and therefore, a high percentage of IDO1 is in the apo form, whose biology and function are currently unknown [bib_ref] Structural insights into substrate and inhibitor binding sites in human indoleamine, Lewis-Ballester [/bib_ref] [bib_ref] Antioxidants inhibit indoleamine 2,3-dioxygenase in IFN-gammaactivated human macrophages: posttranslational regulation by pyrrolidine..., Thomas [/bib_ref] [bib_ref] The versatile biochemistry of iron in macrophage effector functions, Behmoaras [/bib_ref] [bib_ref] Immune-modulating enzyme indoleamine 2,3-dioxygenase is effectively inhibited by targeting its apo-form, Nelp [/bib_ref]. Moreover, IDO1 does not only reside in cytosol but can have a distinct intra-and extracellular topology, depending on the cell microenvironment [bib_ref] Class IA PI3Ks regulate subcellular and functional dynamics of IDO1, Iacono [/bib_ref]. Perhaps particularly intriguing is the fact that IDO1 is not just an enzyme and, in fact, after phosphorylation of critical tyrosine residues, can act as a mediator of a signaling pathway that profoundly changes the functional phenotype of specific immune cells [bib_ref] Indoleamine 2,3-dioxygenase is a signaling protein in long-term tolerance by dendritic cells, Pallotta [/bib_ref]. Structural studies suggest that the distinct functions of IDO1, that is, catalytic versus signaling, may be associated with distinct protein conformations [bib_ref] Tracking hidden binding pockets along the molecular recognition path of l-Trp to..., Greco [/bib_ref].
Therefore, IDO1 should be considered a moonlighting protein , that is, capable of mediating distinct functions in response to distinct cellular needs, and as such should be taken into account as drug target for a more effective immunotherapy.
## Insights into ido1 key structural elements
Several crystallographic structures of human IDO1 are available in different ligand-bound and unbound forms . Overall, they show a conserved fold of the enzymic primary sequence that is composed of a large domain containing the catalytic cleft and a small domain featuring two functional immunoreceptor tyrosine-based inhibitory motifs (ITIMs; Y111, Y249) that are associated with the noncatalytic signaling function of the enzyme [fig_ref] Figure 2: Crystal structure of substrate-bound IDO1 [/fig_ref] [bib_ref] Indoleamine 2,3-dioxygenase is a signaling protein in long-term tolerance by dendritic cells, Pallotta [/bib_ref]. A further . The KP in mammals. The first and rate-limiting step in the KP is catalyzed by IDO1, IDO2, and TDO. N-formylkynurenine is rapidly metabolized by AFMID into Lkynurenine, which can be transformed by KATs, KYNU, and KMO into kynurenic acid, anthranilic acid, and 3-hydroxykynurenine, respectively. 3-hydroxykynurenine can be then transformed by KATs into xanthurenic acid or KYNU into 3-hydroxyanthranilic acid, which can be converted by 3-HAO into 2amino-3-carboxymuconic-6-semialdehyde that spontaneously transforms into quinolinic acid. NAD + , the final product of the pathway, is then obtained by several enzymatic reactions. IDO1, indoleamine 2,3dioxygenase 1; TDO, tryptophan dioxygenase; IDO2, indoleamine 2,3dioxygenase 2; AFMID, kynurenine formamidase; KATs, kynurenine aminotransferases; KYNU, kynureninase; KMO, kynurenine monooxygenase; 3-HAO, 3-hydroxyamino oxidase; NAD+, nicotinamide adenine dinucleotide; QPRT, quinolinate phosphoribosyltransferase. 6100 conserved post-transcriptional regulatory site is located at a YENM motif (residues [bib_ref] Combining immune checkpoint inhibitors: established and emerging targets and strategies to improve..., Khair [/bib_ref] [bib_ref] IDO1 inhibition synergizes with radiation and PD-1 blockade to durably increase survival..., Ladomersky [/bib_ref] [bib_ref] Prognostic value of indoleamine 2,3-dioxygenase expression in colorectal cancer: effect on tumorinfiltrating..., Brandacher [/bib_ref] [bib_ref] Role of the immunosuppressive enzyme indoleamine 2,3-dioxygenase in the progression of ovarian..., Inaba [/bib_ref] , which extends from the small domain toward the large domain. Upon phosphorylation, it triggers the interaction with class IA phosphoinositide 3-kinases (PI3Ks) that, once activated, promote a localization shift of IDO1 from the cytosol to early endosomes (see below) [bib_ref] Class IA PI3Ks regulate subcellular and functional dynamics of IDO1, Iacono [/bib_ref].
Large and small domains are connected by a loop (DE-loop; residues 260-268) above the sixth coordination site of the heme cofactor. A large flexible loop (JK-loop; residues 360-382) controls the access to the catalytic site, adopting distinct closed, intermediate, and open conformations that are involved in mediating substrate recognition [bib_ref] Tracking hidden binding pockets along the molecular recognition path of l-Trp to..., Greco [/bib_ref]. Another structural element of IDO1 is a narrow channel formed by a-helices E and F that extends from the solvent-exposed surface of the protein to the heme group [bib_ref] Structural insights into substrate and inhibitor binding sites in human indoleamine, Lewis-Ballester [/bib_ref]. This channel is relevant for the catalytic activity as it is thought to mediate the shuttling of the co-substrate oxygen to the catalytic cleft for promoting the oxidative cleavage of the indole ring of L-Trp [bib_ref] Structural insights into substrate and inhibitor binding sites in human indoleamine, Lewis-Ballester [/bib_ref]. At odds with its functionally related enzyme tryptophan dioxygenase (TDO), IDO1 shows a wide recognition of substrates beyond L-Trp, including many indole-bearing compounds such as serotonin, melatonin, and tryptamine , suggesting that IDO1 activity may translate into additional functions as compared to the L-Trp-specific enzyme TDO.
Recent crystallographic studies of human IDO1 in complex with the substrate cast a light on the binding modes of L-Trp into the catalytic cleft and an accessory site, with the latter being associated with the inhibition by substrate phenomenon [bib_ref] Structural insights into substrate and inhibitor binding sites in human indoleamine, Lewis-Ballester [/bib_ref]. Upon oxygen binding to the ferrous active form of IDO1 [21-24], L-Trp binds to the catalytic pocket on the sixth coordination site of the heme cofactor, yielding the catalytically active ternary complex [fig_ref] Figure 3: Binding mode of L-Trp [/fig_ref]. According to this binding mode, the indole ring engages S167 with a water-mediated hydrogen bond and establishes face-toedge p-stacking and hydrophobic interactions with F163 and Y126. The ammonium group of L-Trp is hydrogen-bonded to T379 and the 7-propionate group of the heme, while the carboxylic moiety makes three hydrogen bonds with R231 and T379. This kind of interactions is frequently observed for primary amine and carboxylic moieties in the crystal structures of ligand-bound protein complexes .
High concentration of L-Trp inhibits the catalytic activity of IDO1, causing the inhibition by substrate phenomenon . This is ascribed to the interaction of L-Trp to an accessory site that is located below the fifth coordination site of the heme cofactor as observed in the crystal structure of the F270G variant of IDO1 [fig_ref] Figure 4: Binding mode of L-Trp [/fig_ref] [19]. According to this binding mode, L-Trp engages the side chains of V170, V269, L339, and L342 with extensive hydrophobic contacts, yielding a poor dissociation constant, that is, in the millimolar range of potency (K d = 26 mM) . Noteworthy, positive allosteric modulators (PAMs) of IDO1 such as indoleamine-ethanol (IDE) and Nacetylserotonin (NAS), as well as some uncompetitive inhibitors such as mitomycin C, have been proposed to bind into this accessory site of the enzyme [bib_ref] Structural insights into substrate and inhibitor binding sites in human indoleamine, Lewis-Ballester [/bib_ref].
Beside the canonical catalytic cleft and the substrate accessory site, other cryptic binding pockets of L-Trp have been conjectured in the structure of IDO1. However, their functional relevance is still elusive. In this regard, computational and biophysical studies have suggested the presence of a metastable binding pocket that is located on K238 and promotes a first interaction of L-Trp to the enzyme along the substrate approaching pathway from the solvent to the catalytic cleft [bib_ref] Tracking hidden binding pockets along the molecular recognition path of l-Trp to..., Greco [/bib_ref]. In the same study, another cryptic site was suggested using microscale thermophoresis to investigate binding interactions between L-Trp and IDO1. Specifically, multiple binding events of the substrate to the enzyme were observed in both the heme-bound and apo states of IDO1, yielding high and low constants. Although the high dissociation constants were in agreement with previous studies on the binding interaction of L-Trp to the catalytic active site , the low dissociation constants were clue for the existence of an additional site of L-Trp in the structure of IDO1, which has still to be characterized.
Overall, these findings suggest a high conformational plasticity of IDO1, with at least two distinct conformations that may be stabilized by the interaction of the substrate and/or ligands to distinct pockets of the enzyme, including the catalytic site. The stabilization of each of these conformations may affect the ability of IDO1 to recruit protein partners, thus driving enzymic or nonenzymic functions.
## Cellular mechanisms of enzymic and nonenzymic functions of ido1
Over the years, the central role of IDO1 in orchestrating immune responses has acquired interesting aspects. L-Trp starvation and generation of immunoactive kynurenines by IDO1 + dendritic cells (DCs) contribute to create an immunosuppressive microenvironment, characterized by impaired T effector cell functions and enhanced activity of regulatory T (Treg) cells . In particular, the degradation and thus starvation in microenvironments of L-Trp (i.e., an essential amino acid) activates general control nonderepressible 2 (GCN2) [31], a kinase that shuts down the expression of several genes, including the pro-inflammatory cytokine interleukin-6 (IL-6), via the inactivation of the eukaryotic translation initiation factor 2A (eIF2A). This effect leads to anergy in effector T cells and also blocks the conversion of Treg cells into proinflammatory T helper type 17 (Th17) cells [bib_ref] Indoleamine 2,3-dioxygenase controls conversion of Foxp3+ Tregs to TH17-like cells in tumor-draining..., Sharma [/bib_ref]. Moreover, L-Kyn, the main product of IDO1 catalytic activity, is an endogenous agonist of the aryl hydrocarbon receptor (AhR; [fig_ref] Figure 5: Intracellular dynamics of IDO1 in DCs [/fig_ref] [bib_ref] An endogenous tumour-promoting ligand of the human aryl hydrocarbon receptor, Opitz [/bib_ref] , a ligand-activated transcription factor that, besides induction of xenobiotic metabolism genes in the liver, can promote the differentiation of effector T cells into Treg cells [bib_ref] An interaction between kynurenine and the aryl hydrocarbon receptor can generate regulatory..., Mezrich [/bib_ref] and also upregulation of IDO1 in DCs [bib_ref] Aryl hydrocarbon receptor negatively regulates dendritic cell immunogenicity via a kynurenine-dependent mechanism, Nguyen [/bib_ref] , rendering these potent antigen-presenting cells immunosuppressive. Very recently, blockade of the AhR was demonstrated to restrict an immunosuppressive axis mediated by Treg cells and tumor-associated macrophages (TAM) [bib_ref] Blockade of the AHR restricts a Treg-macrophage suppressive axis induced by L-Kynurenine, Campesato [/bib_ref]. Approximately 10 years ago, a novel function of IDO1 was discovered [bib_ref] Indoleamine 2,3-dioxygenase is a signaling protein in long-term tolerance by dendritic cells, Pallotta [/bib_ref]. The nonenzymic function of IDO1 depends on the presence of two ITIMs, namely ITIM1 and ITIM2, located in the small, noncatalytic domain of the enzyme. The ITIM sequence, I/V/L/ SxYxxL/V/F (where x indicates any amino acids), when tyrosine is phosphorylated, acts as a docking site for diverse molecular partners containing Src homology 2 (SH2) domains [bib_ref] Myeloid specific human CD33 is an inhibitory receptor with differential ITIM function..., Paul [/bib_ref] [bib_ref] Inhibitory receptors, ITIM sequences and phosphatases, Unkeless [/bib_ref] , thus fulfilling diverse immunological needs. In fact, when DCs face the immunosuppressive cytokine transforming growth factor b (TGF-b), Src homology 2 domain phosphatase-1 (SHP1) and SHP2 are upregulated and preferentially associate with ITIM1 rather than ITIM2 [fig_ref] Figure 5: Intracellular dynamics of IDO1 in DCs [/fig_ref] [bib_ref] Indoleamine 2,3-dioxygenase is a signaling protein in long-term tolerance by dendritic cells, Pallotta [/bib_ref]. This event translates into SHPs' interaction with the interleukin-1 receptor-associated kinase (IRAK), Pro-inflammatory contexts driven by IL-6 are characterized by the upregulation of SOCS3, which associates with phosphorylated ITIM2 via its SH2 domain, recruits the E3 ubiquitin ligase complex, and directs cytosolic IDO1 to proteasomal degradation. (C) In a microenvironment dominated by immunoregulatory TGF-b, IDO1 is phosphorylated in ITIM (ITIM1 and ITIM2) and YENM domains, which bind and activate SHPs and PI3Ks, respectively. PI3K activation implies IDO1 direct binding to the p85 regulatory subunit that in turn promotes activation of p110, the PI3K catalytic subunit. PI3K binding favors IDO1 anchoring to EE and signaling function. SHP binding promotes IKKa-dependent activation of the noncanonical pathway of NF-jB that, via the p52/RelB heterodimer translocating to the nucleus, upregulates the expression of Ido1 and Tgfb1 genes. This mechanism establishes a positive immunoregulatory circuitry that ensures IDO1 long-term expression in dendritic cells. IDO1, indoleamine 2,3dioxygenase 1; IFN-c, interferon c; L-Trp, L-tryptophan; L-Kyn, L-kynurenine; AhR, aryl hydrocarbon receptor; IL-6, interleukin-6; SOCS3, suppressor of cytokine signaling 3; ITIM, immune tyrosine-based inhibitory motif; TGF-b, transforming growth factor b; PI3Ks, phosphoinositide 3-kinases; EE, early endosome; SHPs, Src homology 2 domain phosphatases; IKKa, inhibitory-jB kinase a. thus favoring the activation of noncanonical, antiinflammatory NF-jB rather than canonical, proinflammatory NF-jB [bib_ref] Indoleamine 2,3-dioxygenase is a signaling protein in long-term tolerance by dendritic cells, Pallotta [/bib_ref]. In turn, noncanonical NF-jB, driven by activated inhibitory-jB kinase a (IKKa) and in the form of p52-RelB heterodimer, translocates to the nucleus and induces the expression of the Ido1 gene itself [bib_ref] Indoleamine 2,3-dioxygenase is a signaling protein in long-term tolerance by dendritic cells, Pallotta [/bib_ref] [bib_ref] Allosteric modulation of metabotropic glutamate receptor 4 activates IDO1-dependent, immunoregulatory signaling in..., Volpi [/bib_ref] [bib_ref] IDO and regulatory T cells: a role for reverse signalling and non-canonical..., Puccetti [/bib_ref] [bib_ref] Noncanonical NF-kappaB signaling in dendritic cells is required for indoleamine 2,3-dioxygenase (IDO)..., Tas [/bib_ref]. Because the IDO1 induction is accompanied by upregulation of Tgfb1 [bib_ref] Class IA PI3Ks regulate subcellular and functional dynamics of IDO1, Iacono [/bib_ref] [bib_ref] Indoleamine 2,3-dioxygenase is a signaling protein in long-term tolerance by dendritic cells, Pallotta [/bib_ref] [bib_ref] Forced IDO1 expression in dendritic cells restores immunoregulatory signalling in autoimmune diabetes, Pallotta [/bib_ref] [bib_ref] IDO: more than an enzyme, Chen [/bib_ref] , this mechanism will ultimately establish a positive feedback loop that confers a long-term immunoregulatory phenotype on both conventional and plasmacytoid mouse DCs (cDCs and pDCs, respectively) [bib_ref] Indoleamine 2,3-dioxygenase is a signaling protein in long-term tolerance by dendritic cells, Pallotta [/bib_ref] [bib_ref] A relay pathway between arginine and tryptophan metabolism confers immunosuppressive properties on..., Mondanelli [/bib_ref].
Suppressor of cytokine signaling (SOCS) proteins are critical modulators of immune responses (11) that possess an SH2 domain, binding phosphotyrosinecontaining peptides, and a SOCS box. The latter domain participates in the formation of an E3 ubiquitin ligase complex that targets several signaling proteins for proteasomal degradation [bib_ref] SOCS-3 negatively regulates innate and adaptive immune mechanisms in acute IL-1-dependent inflammatory..., Wong [/bib_ref] [bib_ref] CD33 responses are blocked by SOCS3 through accelerated proteasomal-mediated turnover, Orr [/bib_ref]. In the presence of a microenvironment dominated by proinflammatory interleukin-6 (IL-6), upregulated SOCS3 preferentially associates with phosphorylated IDO1 ITIM2, recruits the E3 ubiquitin ligase complex, and targets IDO1 to proteasomal degradation [bib_ref] SOCS3 drives proteasomal degradation of indoleamine 2,3-dioxygenase (IDO) and antagonizes IDO-dependent tolerogenesis, Orabona [/bib_ref]. Therefore, SOCS3 interaction with phosphorylated ITIM2 will reduce IDO1 half-life, interrupt tolerogenic mechanisms, and favor a pro-inflammatory phenotype in the DCs [fig_ref] Figure 5: Intracellular dynamics of IDO1 in DCs [/fig_ref].
A crucial point in these processes is the tyrosine kinase responsible for the phosphorylation of IDO1 ITIMs. A meta-analysis of pDC gene expression public data sets indicated that the most widely represented is Fyn, a kinase belonging to the Src family, and that Fyn inhibition decreases IDO1 ITIM phosphorylation by TGF-b [bib_ref] Indoleamine 2,3-dioxygenase is a signaling protein in long-term tolerance by dendritic cells, Pallotta [/bib_ref]. In cDCs, the most important kinase involved in this mechanism is instead Src [bib_ref] Allosteric modulation of metabotropic glutamate receptor 4 activates IDO1-dependent, immunoregulatory signaling in..., Volpi [/bib_ref] [bib_ref] A relay pathway between arginine and tryptophan metabolism confers immunosuppressive properties on..., Mondanelli [/bib_ref] [bib_ref] Aryl hydrocarbon receptor control of a disease tolerance defence pathway, Bessede [/bib_ref]. However, the possible existence of a preferential tyrosine phosphorylation of IDO1 ITIM1 versus ITIM2 by these kinases in distinct conditions has not been clarified yet.
Therefore, on the basis of these observations, IDO1 may be considered a moonlighting protein , that is, endowed with an additional function besides catalytic activity. Moonlighting proteins are very peculiar in that they will switch between functions by changing their conformational state, as may occur in response to altered environmental conditions . Those include the redox state of the cell, temperature, but also posttranslational modifications (such as phosphorylation), changes in cellular localization, and interactions with other polypeptides. The additional function of IDO1 is compatible with that of a signal transducer leading to genomic effects and reprogramming DC functions.
The interesting thing is that phosphorylable ITIMs can either prolong or reduce IDO1 expression, depending on specific cellular microenvironments and distinct partnerships [bib_ref] Different partners, opposite outcomes: a new perspective of the immunobiology of indoleamine..., Orabona [/bib_ref] , thus underlining the plasticity of IDO1 biology.
## Modulation of ido1 expression
In humans, IDO1 is constitutively expressed in a restricted set of cells, including placental and pulmonary endothelial cells, epithelial cells inside the female genital tract, and antigen-presenting cells, such as mature DCs in secondary lymphoid organs [bib_ref] Extensive profiling of the expression of the indoleamine 2,3-dioxygenase 1 protein in..., Theate [/bib_ref] [bib_ref] Immunosuppressive IDO in cancer: mechanisms of action, animal models, and targeting strategies, Zhai [/bib_ref]. IDO1 is also highly expressed in b-cells of pancreatic islet of healthy individuals but is absent in the residual b-cells of patients with autoimmune diabetes [bib_ref] Loss of IDO1 expression from human pancreatic beta-cells precedes their destruction during..., Anquetil [/bib_ref]. Several human tumors constitutively express IDO1. In this regard, the IDO1 protein can be found in neoplastic cells themselves but also in fibroblasts and myeloid and endothelial cells present in the tumor bed [bib_ref] Tryptophandegrading enzymes in tumoral immune resistance, Van Baren [/bib_ref]. In some tumor cells, prostaglandin E2 (PGE2) and interleukin-1b (IL-1b) contribute to the maintenance of IDO1 basal expression. Additionally, constitutive IDO1 expression in human cancer may be sustained by an autocrine signaling loop involving IL-6 and AhR [bib_ref] Constitutive IDO expression in human cancer is sustained by an autocrine signaling..., Litzenburger [/bib_ref] , suggesting that IL-6 in tumors may behave differently from normal DCs in which the cytokine promotes IDO1 proteasomal degradation [bib_ref] SOCS3 drives proteasomal degradation of indoleamine 2,3-dioxygenase (IDO) and antagonizes IDO-dependent tolerogenesis, Orabona [/bib_ref].
One of the major features of IDO1 is its inducibility [bib_ref] Relationship between interferon-gamma, indoleamine 2,3-dioxygenase, and tryptophan catabolism, Taylor [/bib_ref]. IFN-c is considered the main inducer of IDO1 in several types of cells, including but not limited to immunocytes and fibroblasts. Indeed, the promoter region of the IDO1 gene (present in human chromosome 8p22) contains several IFN-stimulated elements (ISREs) and gamma activation sequences (GAS) [bib_ref] Importance of the two interferon-stimulated response element (ISRE) sequences in the regulation..., Konan [/bib_ref] that respond to transcription factors such as signal transducer and activator of transcription 1 (STAT1) and interferon-regulatory factors (IRFs). IDO1 induction by IFN-c may represent a counteracting response in inflammatory conditions [bib_ref] Tolerance, DCs and tryptophan: much ado about IDO, Grohmann [/bib_ref] [bib_ref] IDO expression by dendritic cells: tolerance and tryptophan catabolism, Mellor [/bib_ref] [bib_ref] IDO in the tumor microenvironment: inflammation, counter-regulation, and tolerance, Munn [/bib_ref] , a mechanism also exploited by tumors to favor their immune escape [bib_ref] IDO in the tumor microenvironment: inflammation, counter-regulation, and tolerance, Munn [/bib_ref]. Indeed, some neoplastic cells downregulate the expression of Bin1 (encoding a Myc-interacting protein with features of immunosuppressor), with consequent elevation of STAT1-dependent expression of IDO1 [bib_ref] Inhibition of indoleamine 2,3-dioxygenase, an immunoregulatory target of the cancer suppression gene..., Muller [/bib_ref]. Type I IFNs, that is, IFN-a and IFN-b, also represent IDO1 inducers, although, at variance with IFN-c (type II IFN), activate a pathway that leads to transcription factors that bind ISREs but not GAS and are thus less effective than IFN-c [bib_ref] On watching the watchers: IDO and type I/II IFN, Puccetti [/bib_ref]. Involvement of type I IFNs is mainly observed in pDCs, which are the cells that produce the highest levels of these cytokines. Over the years, several other signals have been identified as IDO1 inducers. These include the cytokine TGF-b that, however, favors the signaling rather than catalytic function of IDO1 [bib_ref] Indoleamine 2,3-dioxygenase is a signaling protein in long-term tolerance by dendritic cells, Pallotta [/bib_ref] [bib_ref] Different partners, opposite outcomes: a new perspective of the immunobiology of indoleamine..., Orabona [/bib_ref]. In cDCs, activation of the Src kinase necessary for IDO1 ITIM phosphorylation can be directly promoted by spermidine, a polyamine produced downstream arginase 1 (ARG1; induced by TGF-b earlier than IDO1) [bib_ref] A relay pathway between arginine and tryptophan metabolism confers immunosuppressive properties on..., Mondanelli [/bib_ref] , a finding that linked for the first time two potent immunosuppressors such as IDO1 and ARG1 [bib_ref] Immunoregulatory interplay between arginine and tryptophan metabolism in health and disease, Mondanelli [/bib_ref] [bib_ref] Control of immune response by amino acid metabolism, Grohmann [/bib_ref] as well as polyamines and kynurenines [bib_ref] Polyamines and kynurenines at the intersection of immune modulation, Proietti [/bib_ref]. The signaling function of IDO1 and consequent immunosuppressive effects can also be triggered by incubation of cDCs with ADX188, a positive allosteric modulator (PAM) of metabotropic glutamate receptor 4 (mGluR4), via potentiation of protein G-independent mechanisms [bib_ref] Allosteric modulation of metabotropic glutamate receptor 4 activates IDO1-dependent, immunoregulatory signaling in..., Volpi [/bib_ref]. Other cytokines, that is, tumor necrosis factor a (TNF-a), IL-6, and IL-10, can synergize with each other to increase IDO1 expression [bib_ref] Molecular pathways: targeting IDO1 and other tryptophan dioxygenases for cancer immunotherapy, Zhai [/bib_ref].
Pro-inflammatory signals capable of upregulating IDO1 include pathogen-associated molecular patterns (PAMPs) [bib_ref] PAMPs and DAMPs: signal 0s that spur autophagy and immunity, Tang [/bib_ref]. PAMPs are present in diverse organisms but absent in the host and provide exogenous signals that alert the immune system to the presence of pathogens, thereby promoting immunity. They are recognized by and activate Toll-like receptors (TLRs), normally contributing to inflammatory processes but also immune tolerance [bib_ref] Pathogen recognition by the innate immune system, Kumar [/bib_ref]. Lipopolysaccharide (LPS) and oligonucleotides containing unmethylated CpG motifs (CpG-DNA) are PAMPs that can modulate IDO1 activity via activation of TLR4 and TLR9, respectively. TLR4 is an intriguing receptor that, by virtue of its surface membrane or early endosome (EE) localization, can recruit either MyD88 or TRIF signaling adaptors, mediators of inflammatory or tolerogenic effects, respectively [bib_ref] The p110delta isoform of the kinase PI(3)K controls the subcellular compartmentalization of..., Aksoy [/bib_ref]. After repeated stimulation, LPS can in fact generate a state of immune suppression, known as endotoxin tolerance. Bessede et al. showed that IDO1 is highly expressed in cDCs repeatedly stimulated with low doses of LPS in vitro and that administration of LPS-treated DCs or L-Kyn alone can confer protection in recipient mice from endotoxic shock via activation of AhR and IDO1 upregulation in cDCs [bib_ref] Aryl hydrocarbon receptor control of a disease tolerance defence pathway, Bessede [/bib_ref]. Human DCs conditioned with LPS exhibit upregulation of anti-inflammatory IL-10 that in turn activates the noncanonical NF-jB pathway, with overexpression of RelB in an IDO1dependent manner [bib_ref] The role of indoleamine 2,3-dioxygenase-aryl hydrocarbon receptor pathway in the TLR4-induced tolerogenic..., Salazar [/bib_ref]. Activation of TLR9 by CpGoligodeoxynucleotides induces the expression of IDO1 in splenic CD19 + DC, which acquire suppressive functions [bib_ref] Cutting edge: CpG oligonucleotides induce splenic CD19+ dendritic cells to acquire potent..., Mellor [/bib_ref]. Moreover, in pDCs, the tolerogenic signaling via TLR9 that leads to IDO1 expression is dependent on the dosage of CpG-ODN administration (i.e., low dose being immunostimulatory and high dose immunoregulatory) and requires the involvement of the TRIF adaptor [bib_ref] High doses of CpG oligodeoxynucleotides stimulate a tolerogenic TLR9-TRIF pathway, Volpi [/bib_ref].
An intriguing issue is the capacity of membraneanchored molecules to induce IDO1 in DCs via 'reverse signaling'. This phenomenon occurs when a molecule normally operating as a receptor and expressed on the surface membrane of another cell activates a signaling pathway via engagement of a membrane molecular counterpart normally acting as a ligand [bib_ref] IDO and regulatory T cells: a role for reverse signalling and non-canonical..., Puccetti [/bib_ref]. By binding to CD80 and CD86 on DCs, cytotoxic T lymphocyte antigen 4 (CTLA-4) engenders an IFN-c-dependent induction of IDO1 [bib_ref] CTLA-4-Ig regulates tryptophan catabolism in vivo, Grohmann [/bib_ref]. Because Treg cells constitutively express CTLA-4, this mechanism represents an important functional bridge between Treg cells and regulatory DCs in immune tolerance [bib_ref] Modulation of tryptophan catabolism by regulatory T cells, Fallarino [/bib_ref]. In addition to CTLA-4, Treg cells possess the glucocorticoid-inducible TNF receptor (GITR) [bib_ref] Exemplifying complexity of immune suppression by a "canonical" speech: A glimpse into..., Ayroldi [/bib_ref]. Engagement of its ligand, GITRL, by a soluble form of GITR activates a reverse signaling in pDCs resulting in the activation of the noncanonical pathway of NF-jB, increased IDO1 expression, and IDO1dependent immunosuppressive effects in vivo [bib_ref] Reverse signaling through GITR ligand enables dexamethasone to activate IDO in allergy, Grohmann [/bib_ref]. Because GITR and GITRL are under the control of glucocorticoids, these data unveiled an important mechanism of action of such potent compounds.
Therefore, IDO1 expression and functions appear to be under the control of a great variety of stimuli that can act either directly or indirectly and can sometimes behave very differently, depending on the circumstances.
## Intra-and extracellular localization of ido1 and role of pi3ks
The bulk of literature has repeatedly indicated a cytosolic localization for the IDO1 enzyme in several types of cells [bib_ref] Indoleamine 2,3-dioxygenase (IDO) expression in invasive extravillous trophoblast supports role of the..., Honig [/bib_ref] [bib_ref] Asp274 and his346 are essential for heme binding and catalytic function of..., Littlejohn [/bib_ref]. This localization may favor IDO1 intense and transient catalytic activity, possibly because of the easier access to substrate, that is, L-Trp. However, until recently, no information was available regarding the intracellular topology of IDO1 as a signaling molecule. A critical discovery was that, in addition to ITIMs, IDO1 contains a consensus binding site for the p85 regulatory subunit of PI3K [bib_ref] Class IA PI3Ks regulate subcellular and functional dynamics of IDO1, Iacono [/bib_ref]. This motif, namely YENM, is characterized by a YxxM sequence (where 'x' indicates any amino acid). Alignment of the amino acid sequences reveals that the YENM sequence of IDO1 in mammals is completely conserved [bib_ref] Class IA PI3Ks regulate subcellular and functional dynamics of IDO1, Iacono [/bib_ref] but is not present in mammalian indoleamine 2,3dioxygenase 2 (IDO2), the IDO1 paralogue whose gene is considered an ancient version of IDO1 itself (see next paragraph) [bib_ref] Characterization of an indoleamine 2,3-dioxygenase-like protein found in humans and mice, Ball [/bib_ref]. When the tyrosine residue within YENM becomes phosphorylated, IDO1 associates with p85 that in turn interacts with the p110 catalytic subunits of PI3Ks [bib_ref] Class IA PI3Ks regulate subcellular and functional dynamics of IDO1, Iacono [/bib_ref]. PI3Ks are a family of enzymes subdivided into three classes. Among them, class IA PI3Ks are heterodimeric proteins consisting of a regulatory subunit (p85a, p85b, or p55c) and a catalytic subunit (p110a, p110b, or p110d) [bib_ref] PI3K in lymphocyte development, differentiation and activation, Okkenhaug [/bib_ref]. Although p110a and p110b subunits are ubiquitously expressed, p110d expression is generally restricted to cells of the immune system [bib_ref] PI3K in lymphocyte development, differentiation and activation, Okkenhaug [/bib_ref] and promotes anti-inflammatory effects in DCs and macrophages [bib_ref] The p110delta isoform of the kinase PI(3)K controls the subcellular compartmentalization of..., Aksoy [/bib_ref]. As a consequence of IDO1 binding to p85, all class IA p110 catalytic subunits can be recruited, although the greater extent is observed for p110d in cell transfectants expressing mouse IDO1 [bib_ref] Class IA PI3Ks regulate subcellular and functional dynamics of IDO1, Iacono [/bib_ref]. Class I PI3Ks possess important roles in several physiologic conditions as well as in disease [bib_ref] Structural basis for regulation of phosphoinositide kinases and their involvement in human..., Burke [/bib_ref] [bib_ref] The PI3K Pathway in Human Disease, Fruman [/bib_ref]. By activating multiple signaling pathways and involving a plethora of downstream effectors, they can orchestrate both pro-inflammatory and anti-inflammatory mechanisms to maintain effective immunity while protecting host tissues. Interestingly, among their multiple functions, members of the PI3K family can also affect intracellular trafficking of vesicles and proteins [bib_ref] The emerging mechanisms of isoform-specific PI3K signalling, Vanhaesebroeck [/bib_ref].
Several activated receptors and signaling proteins accumulate in EE, which therefore can be defined as a platform hosting intracellular signaling [bib_ref] Endosomes: a legitimate platform for the signaling train, Murphy [/bib_ref]. Signals deriving from receptors and other proteins located in EE have different roles, such as control of growth, differentiation, survival, inflammation, and immunity. Efficient TGF-b signaling-also responsible for IDO1 signaling activation-requires internalization of the TGF-b receptor in EE [bib_ref] TGF beta receptor internalization into EEA1-enriched early endosomes: role in signaling to..., Hayes [/bib_ref] [bib_ref] Internalizationdependent and -independent requirements for transforming growth factor beta receptor signaling via..., Penheiter [/bib_ref]. In a very recent study, treatment of pDCs with IFN-c or TGF-b determined a dominant localization of the IDO1 protein in cytosol or EE, respectively [bib_ref] Class IA PI3Ks regulate subcellular and functional dynamics of IDO1, Iacono [/bib_ref]. Moreover, IDO1 anchoring to EE was mediated by binding of class I PI3Ks [bib_ref] Class IA PI3Ks regulate subcellular and functional dynamics of IDO1, Iacono [/bib_ref] , an effect favoring IDO1 signaling rather catalytic activity in pDCs. These observations indicated, for the first time, an IDO1 localization distinct from cytosol and how the intracellular switch can occur. It is interesting to note that, in the absence of any stimulus, in both primary fresh cultures of pDCs and tumor transfectants highly expressing IDO1, IDO1 is mainly present in the cytosol but discrete quantity of the protein can be detected in EE as well, suggesting the a priori existence in IDO1 + cells of a pool of IDO1 proteins with distinct topology and possibly ready for distinct functions. Furthermore, although not demonstrated yet, IDO1 in EE may not contain heme. In fact, in unstimulated human ovarian cancer cells, IDO1 is dynamically bound to its heme cofactor and, perhaps most importantly, at least 85% of IDO1 exists in the apo form, that is, catalytically inactive, that can be nevertheless activated by exogenous heme added in the form of hemin [bib_ref] Immune-modulating enzyme indoleamine 2,3-dioxygenase is effectively inhibited by targeting its apo-form, Nelp [/bib_ref]. Likewise, hemin increases IDO1 activity 2-fold in IFN-c-activated human monocyte-derived macrophages [bib_ref] Antioxidants inhibit indoleamine 2,3-dioxygenase in IFN-gammaactivated human macrophages: posttranslational regulation by pyrrolidine..., Thomas [/bib_ref] , suggesting that cellular heme is limiting for IDO1 activity and that a significant portion of IDO1 in these cells is present as apoprotein. Along this line of evidence, transfected mouse IDO1 with tyrosine in ITIM1 substituted with glutamate (i.e., mimicking phosphorylated tyrosine) does not show any catalytic activity .
IDO1 is not confined into the cells but can be secreted in extracellular vesicles (EVs), including exosomes. In fact, stem cells from amniotic fluid treated with the main IDO1 inducer IFN-c produce immunoregulatory EVs containing IDO1 [bib_ref] Stem cells from human amniotic fluid exert immunoregulatory function via secreted indoleamine..., Romani [/bib_ref]. EVs from human semen express high levels of IDO1 transcripts, suggesting a mechanism whereby these vesicles induce tolerance [bib_ref] Extracellular vesicles in human semen modulate antigen-presenting cell function and decrease downstream..., Vojtech [/bib_ref]. Also, malignant glioblastoma cells generate immunosuppressive EVs containing IDO1, suggesting that tumors can exploit both intracellular and extracellular IDO1 as immune escape mechanism. Exosomes derived from IDO1overexpressing rat bone marrow mesenchymal stem cells promote immunotolerance of cardiac allografts [bib_ref] Exosomes derived from IDO1-overexpressing rat bone marrow mesenchymal stem cells promote immunotolerance..., He [/bib_ref]. Although we do not know how and when IDO1 is transferred outside the cells via vesicles yet, these observations further confirm the complexity and dynamics of IDO1 in physiologic and pathologic conditions.
As a whole, IDO1 expression in cytosol, EE, and EVs-and, perhaps most importantly, the distinct IDO1 functions in cytosol versus EE-further confirm the hypothesis of IDO1 being a moonlighting protein [bib_ref] Moonlighting proteins are important players in cancer immunology, Adamo [/bib_ref]. In fact, the localization of a protein in a new cellular microenvironment could promote unpredicted molecular interactions by many potential new binding partners and contribute to generate new functions, as occurring for IDO1 [bib_ref] Class IA PI3Ks regulate subcellular and functional dynamics of IDO1, Iacono [/bib_ref]. Distinct activities mediated by STAT3 and Tgase2 are associated with different localization, including exosomes [bib_ref] Moonlighting proteins: an intriguing mode of multitasking, Huberts [/bib_ref] [bib_ref] Transglutaminase type 2-dependent selective recruitment of proteins into exosomes under stressful cellular..., Diaz-Hidalgo [/bib_ref].
In conclusion, the IDO1 catalytic and signaling functions are spatially segregated. Further studies are necessary to define the role of the IDO1 signaling function and its subcellular localization in pathological conditions, such as cancer, in order to optimize the development of new and more effective therapies [bib_ref] Is there a clinical future for IDO1 inhibitors after the failure of..., Eynde [/bib_ref].
## Ido1, ido2, and tdo: similar or distinct enzymes?
In addition to IDO1, there are two other mammalian enzymes capable to convert L-Trp to L-Kyn, that is, IDO2 and TDO. Since its discovery in 2007 [bib_ref] Characterization of an indoleamine 2,3-dioxygenase-like protein found in humans and mice, Ball [/bib_ref] , IDO2 represents the 'enigmatic version' of IDO1. In fact, the two IDOs are encoded by genes located in tandem on chromosome 8, in both humans and mice [bib_ref] Evolution of vertebrate indoleamine 2,3-dioxygenases, Yuasa [/bib_ref] , and share a high level of sequence similarity at the amino acid level. Nevertheless, IDO2 catalytic activity is negligible in vitro [bib_ref] Characterization of an indoleamine 2,3-dioxygenase-like protein found in humans and mice, Ball [/bib_ref]. Moreover, IDO2 knockout does not contribute to the reduction in L-Kyn sera levels in vivo as instead shown by IDO1 depletion [bib_ref] IDO2 is critical for IDO1-mediated T-cell regulation and exerts a non-redundant function..., Metz [/bib_ref] and the pattern of the expression of the two molecules is consistently different, being IDO2 mainly expressed in the liver under physiologic conditions. Interestingly, lower vertebrates generally possess only IDO2, and many fish species have both IDO1 and IDO2; as a possible explanation, molecular phylogenetic analyses showed that the gene duplication occurred before the divergence of vertebrates, with IDO1 having been lost in a number of lower vertebrate lineages [bib_ref] Low efficiency IDO2 enzymes are conserved in lower vertebrates, whereas higher efficiency..., Yuasa [/bib_ref]. Sequence analysis of the human IDO2 showed a peculiar and wide distribution of two single nucleotide polymorphisms (SNPs), which are rs10109853, leading to a > 90% reduction in IDO2 catalytic activity (R248W), and rs4503083, generating a premature stop codon (Y359X) [bib_ref] Novel tryptophan catabolic enzyme IDO2 is the preferred biochemical target of the..., Metz [/bib_ref]. However, the exact functional significance of the two common IDO2 variants is still far from being deciphered, as is the understanding of IDO2 physiological role. In fact, the putative structural analogies between IDO1 and IDO2, suggested by their high amino acid sequence homology, seem to not correspond to similar functions.
In vivo studies demonstrated a contrasting role for IDO2, with experiments in preclinical models of autoimmune arthritis suggesting a pro-inflammatory role driving the disease [bib_ref] Differential roles of IDO1 and IDO2 in T and B cell inflammatory..., Merlo [/bib_ref]. In neoplastic diseases, IDO2 is much less expressed in tumor cells than IDO1 and its function is still matter of debate. In humans, pancreatic ductal adenocarcinoma (PDAC) [bib_ref] Genotyping and expression analysis of IDO2 in human pancreatic cancer: a novel,..., Witkiewicz [/bib_ref] and non-small-cell lung cancer (NSCLC) [bib_ref] Indoleamine 2,3-dioxygenase 2 immunohistochemical expression in resected human non-small cell lung cancer:..., Mandarano [/bib_ref] show particularly high levels of IDO2 expression. Nevertheless, the presence of a genotype compatible with an inactive form of IDO2 significantly associates with an improved disease-free survival in patients with PDAC [bib_ref] Host IDO2 gene status influences tumor progression and radiotherapy response in KRAS-driven..., Nevler [/bib_ref] but increases the risk of developing NSCLC [bib_ref] Grohmann U & Volpi C (2021) Current challenges for IDO2 as target..., Mondanelli [/bib_ref].
Many studies have demonstrated that the knockout of the Ido2 gene can inhibit tumor growth in multiple animal models [bib_ref] The emerging roles of IDO2 in cancer and its potential as a..., Li [/bib_ref]. However, the use of IDO2 specific inhibitors could be a failure strategy, as it was in case of IDO1 [bib_ref] Is there a clinical future for IDO1 inhibitors after the failure of..., Eynde [/bib_ref]. In fact, the inhibition of IDO1 enzyme activity only may be not sufficient to exploit an effective therapeutic function, considering the complex functional dynamics of IDO1 and its additional signaling function. Regarding IDO2, we might hypothesize that it could possess a completely different function other than the poor enzymatic one, presumably relying on some still unidentified partnership with other proteins and consequent signaling activity. The presence of one ITIM motif (i.e., ITIM2) in both human and mouse IDO2 amino acid sequences pleads for a potential role of this molecule as a signaling molecule, although its function is still unknown.
Unlike IDO2, the physiologic role of TDO is better defined. TDO is mainly located in the liver and uses L-Trp as substrate, thus showing a higher substrate specificity compared to IDO1 [bib_ref] Substrate stereo-specificity in tryptophan dioxygenase and indoleamine 2,3-dioxygenase, Capece [/bib_ref]. In the liver, TDO degrades excess dietary L-Trp in order to maintain systemic stable levels of the amino acid [bib_ref] Tryptophan 2,3-dioxygenase is a key modulator of physiological neurogenesis and anxiety-related behavior..., Kanai [/bib_ref]. Interestingly, seric L-Kyn, the main product of both TDO and IDO1, is increased in TDO knockout mice but is totally absent in mice lacking IDO1, indicating that circulating L-Kyn is mainly produced by IDO1. This is likely because the L-Kyn produced by TDO is further degraded in the liver, which expresses all the enzymes of the L-Kyn pathway, whereas inflammatory and immune cells that express IDO1 in the periphery may not express these enzymes [bib_ref] Inhibition of tryptophan-dioxygenase activity increases the antitumor efficacy of immune checkpoint inhibitors, Schramme [/bib_ref].
TDO expression is also detected in the decidualized endometrium, where its function is unknown [bib_ref] Differential expression and regulation of Tdo2 during mouse decidualization, Li [/bib_ref] , and in the brain, contributing to the synthesis of neuroactive compounds [bib_ref] TDO as a therapeutic target in brain diseases, Yu [/bib_ref]. TDO is thought to participate in the pathogenesis of several neurodegenerative diseases, such as Alzheimer, Parkinson, and Huntington disease (AD, PD, and HD, respectively), via regulation of proteotoxic events. Significantly high TDO immunoreactivity has been observed in the AD patients' hippocampus [bib_ref] Expression of tryptophan 2,3-dioxygenase and production of kynurenine pathway metabolites in triple..., Wu [/bib_ref]. Moreover, in fly models of AD and PD, TDO inhibition improves motor performance and helps survival, and increased kynurenic acid relative to 3-hydroxykynurenine alleviates HD, indicating that shifting the KP to kynurenic acid rather than 3-hydroxykynurenine synthesis can reduce neurodegeneration [bib_ref] Tryptophan-2,3-dioxygenase (TDO) inhibition ameliorates neurodegeneration by modulation of kynurenine pathway metabolites, Breda [/bib_ref].
TDO2 mRNA can be also found in a number of human tumor samples, including hepatocarcinomas, glioblastomas, melanomas, and bladder carcinomas [bib_ref] Tryptophan 2,3-dioxygenase expression identified in human hepatocellular carcinoma cells and in intratumoral..., Hoffmann [/bib_ref] and in many human tumor cell lines, including glioblastoma, colorectal, head and neck, lung, and gall bladder carcinoma cell lines, where its protein expression and activity have been confirmed [bib_ref] Reversal of tumoral immune resistance by inhibition of tryptophan 2,3-dioxygenase, Pilotte [/bib_ref]. Like IDO1, TDO plays an immunoregulatory role in tumors via L-Kyn activation of AhR [bib_ref] An endogenous tumour-promoting ligand of the human aryl hydrocarbon receptor, Opitz [/bib_ref] [bib_ref] (TDO) inhibitors. 3-(2-(pyridyl)ethenyl) indoles as potential anticancer immunomodulators, Dolusic [/bib_ref]. TDOtransfected mouse mastocytoma P815 cells resist immune rejection by mice immunized against P1A, the dominant antigen expressed in these tumors, and tumor rejection is restored upon treatment of mice
## Physiologic versus pathologic effects of ido1
Physiologic roles of IDO1
The IDO1 protein is a physiologic checkpoint that guarantees immune homeostasis in our organism by finely modulating a balanced immune response orchestrated by antigen-presenting cells, such as DCs and macrophages. It plays a key role in supporting immune privilege [bib_ref] The kynurenine pathway is a double-edged sword in immune-privileged sites and in..., Routy [/bib_ref] , including the feto-maternal interface in allogeneic pregnancy to protect the fetus from the maternal immune system [bib_ref] Prevention of allogeneic fetal rejection by tryptophan catabolism, Munn [/bib_ref] and further immune niches as cornea [bib_ref] Function of indoleamine 2,3-dioxygenase in corneal allograft rejection and prolongation of allograft..., Beutelspacher [/bib_ref] and brain [bib_ref] Kynurenine pathway metabolism in human astrocytes: a paradox for neuronal protection, Guillemin [/bib_ref]. L-Trp metabolites generated by IDO1 along the KP serve as signaling molecules for fine-tuning the host immune response, mainly via AhR engagement of AhR [bib_ref] Tolerogenic phenotype of IFN-gamma-induced IDO+ dendritic cells is maintained via an autocrine..., Li [/bib_ref] [bib_ref] Engagement of nuclear coactivator 7 by 3-hydroxyanthranilic acid enhances activation of Aryl..., Gargaro [/bib_ref]. The large evidence of a nonenzymic function of IDO1 that co-operates with its catalytic activity renders more intricate the mechanisms of IDO1-based immune tolerance [bib_ref] Indoleamine 2,3-dioxygenase is a signaling protein in long-term tolerance by dendritic cells, Pallotta [/bib_ref] [bib_ref] Different partners, opposite outcomes: a new perspective of the immunobiology of indoleamine..., Orabona [/bib_ref] that, nevertheless, opens the generation of innovative strategies targeting IDO1.
IDO1 catalytic activity also controls the de novo synthesis of NAD + , in addition to the salvage pathway that supplies the cell with the highest levels of NAD + cofactor. The metabolic relevance of the kynurenine in NAD + supplementation in both health and disease is not well defined, although the contribution of L-Trp metabolism to NAD + formation in the central nervous system (CNS) and in the tumor microenvironment (TME) is more evident than in other tissues [bib_ref] Effects of kynurenine pathway inhibition on NAD metabolism and cell viability in..., Braidy [/bib_ref] [bib_ref] Kynurenine pathway metabolism is involved in the maintenance of the intracellular NAD..., Grant [/bib_ref].
A growing body of evidence has demonstrated that IDO1 becomes a pathogenic factor overarching a wide range of human pathologies, including infections, neoplasia, autoimmunity, and neurodegeneration [bib_ref] Tryptophandegrading enzymes in tumoral immune resistance, Van Baren [/bib_ref] [bib_ref] Immune checkpoint molecules, personalized immunotherapy, and autoimmune diabetes, Orabona [/bib_ref] [bib_ref] Tryptophan feeding of the IDO1-AhR axis in hostmicrobial symbiosis, Zelante [/bib_ref] [bib_ref] Tryptophan metabolism as a common therapeutic target in cancer, neurodegeneration and beyond, Platten [/bib_ref] , as outlined below.
## Ido1 in infectious diseases
L-Trp metabolism is an important regulator of immune host-pathogen interactions. Several PAMPs can induce IDO1 expression by acting as TLR ligands [bib_ref] High doses of CpG oligodeoxynucleotides stimulate a tolerogenic TLR9-TRIF pathway, Volpi [/bib_ref] [bib_ref] New Insights into IDO Biology in Bacterial and Viral Infections, Schmidt [/bib_ref].
In infectious diseases, L-Trp depletion caused by IDO1 activity has long been known to induce antimicrobial actions for auxotroph pathogens [bib_ref] Tryptophan depletion as a mechanism of gamma interferon-mediated chlamydial persistence, Beatty [/bib_ref] [bib_ref] Feast or famine: the host-pathogen battle over amino acids, Zhang [/bib_ref]. Several in vitro studies reported that measles, influenza, cytomegalovirus, and herpes simplex viral infections are susceptible to L-Trp levels [bib_ref] New Insights into IDO Biology in Bacterial and Viral Infections, Schmidt [/bib_ref] [bib_ref] Feast or famine: the host-pathogen battle over amino acids, Zhang [/bib_ref]. However, L-Trp starvation can reprogram several auxotroph invaders to re-acquire the capacity to synthesize this essential amino acid, as in the case of Mycobacterium tuberculosis [bib_ref] Tryptophan biosynthesis protects mycobacteria from CD4 T-cellmediated killing, Zhang [/bib_ref]. At the same time, generation of downstream kynurenines by IDO1 acts as a double-edge sword in infectious diseases, creating a state of immunosuppression that may impair clearance of the microorganism.
Combined antiviral therapy with IDO1 blockade significantly reduces the simian immunodeficiency virus (SIV) load in plasma and lymph nodes of treated rhesus macaques [bib_ref] Combined effect of antiretroviral therapy and blockade of IDO in SIV-infected rhesus..., Boasso [/bib_ref] , indicating that, in this case, immunoregulatory IDO1 is at work in the disease and its inhibition favors recovery of infected animals. Recently, Gautam et al. [bib_ref] In vivo inhibition of tryptophan catabolism reorganizes the tuberculoma and augments immunemediated..., Gautam [/bib_ref] demonstrated in macaques affected by active tuberculosis that inhibition of IDO1 activity leads to a reduced M. tuberculosis burden, pathology, and improved animal survival. Currently, it is unclear yet whether elevated IDO1 is causative of the progression to active infection or a compensatory response to the microbe. However, these observations suggest that a potential for using IDO1 inhibitors as an effective anti-microbial therapy may exist.
In addition to the control of pathogen load, IDO1 activation can restrain pathogenic immune activation that would ultimately worsen the infection. Indeed, in experimental fungal infections, IDO1 blockade greatly exacerbates the disease and the associated inflammatory pathology, as a result of dysregulated innate and adaptive immune responses to the fungi [bib_ref] Protective tolerance to fungi: the role of IL-10 and tryptophan catabolism, Romani [/bib_ref]. Interestingly, fungi express IDO1 and produce kynurenines that are involved in fungal but also host fitness [bib_ref] Aspergillus fumigatus tryptophan metabolic route differently affects host immunity, Zelante [/bib_ref].
## Ido1 in cancer
In tumors, activation of L-Trp catabolism by any of the L-Trp-degrading enzymes leads to a local generation of immunosuppressive kynurenines and L-Trp depletion that determines a poor immune response. Based on this postulate, a fervid drug discovery activity in the development of anticancer small molecules targeting both IDO1 and TDO enzymes has characterized the last two decades [bib_ref] Cancer immunotherapy by targeting IDO1/TDO and their downstream effectors, Platten [/bib_ref] [bib_ref] Discovery of IDO1 inhibitors: from bench to bedside, Prendergast [/bib_ref] [bib_ref] Trial watch: IDO inhibitors in cancer therapy, Vacchelli [/bib_ref]. A variety of catalytic inhibitors of IDO1 have been developed with favorable properties for a drug candidate (i.e., higher selectivity, potency, oral bioavailability, and safe profile) and with different mechanism of inhibition (i.e., competitive, noncompetitive, and heme-displacing molecules) [bib_ref] Molecular pathways: targeting IDO1 and other tryptophan dioxygenases for cancer immunotherapy, Zhai [/bib_ref] [bib_ref] Advances in the discovery and development of selective heme-displacing IDO1 inhibitors, Sun [/bib_ref]. An enormous attention has developed around targeting IDO1 in cancer immunotherapy, encouraged by promising results coming from preclinical studies that demonstrated, in a number of animal models, the antitumor effect of IDO1 blockade in monotherapy or in combination with other immune checkpoint inhibitors [bib_ref] Combining immune checkpoint inhibitors: established and emerging targets and strategies to improve..., Khair [/bib_ref] [bib_ref] IDO1 inhibition synergizes with radiation and PD-1 blockade to durably increase survival..., Ladomersky [/bib_ref]. Based on the correlation of IDO1 expression with worse prognosis in a wide range of human tumors [bib_ref] Prognostic value of indoleamine 2,3-dioxygenase expression in colorectal cancer: effect on tumorinfiltrating..., Brandacher [/bib_ref] [bib_ref] Role of the immunosuppressive enzyme indoleamine 2,3-dioxygenase in the progression of ovarian..., Inaba [/bib_ref] , the most promising compound, that is, epacadostat, reached clinical trial phase 3. Disappointingly, the largest phase 3 clinical trial (ECHO-301) in advanced melanoma with epacadostat and pembrolizumab (an anti-PD1 antibody) failed. The trial was suspended because the therapeutic strategy did not show any increased benefit compared with placebo plus pembrolizumab in melanoma patients [bib_ref] Epacadostat plus pembrolizumab versus placebo plus pembrolizumab in patients with unresectable or..., Long [/bib_ref]. To date, none of the IDO1 inhibitors has been approved for anticancer therapy yet.
Despite the failed experience with epacadostat, a strong translational rationale still exists for targeting IDO1 and, more widely, L-Trp metabolism in cancer immunotherapy. Since it became evident that standalone IDO1 blockade is not effective in tumor disease progression, numerous novel molecules with dual inhibitory activity have been developed. Dual TDO/IDO1 inhibitors have been designed for tumors where IDO1 inhibition can play the by-pass to L-Kyn generation by TDO. TDO might provide compensatory mechanisms to L-Trp depletion and production of L-Kyn metabolites. Therefore, dual or even pan-inhibition of L-Trpdegrading enzymes (i.e., also targeting IDO2) might be beneficial and complementary for efficacy improvement in immunotherapy [bib_ref] Reversal of tumoral immune resistance by inhibition of tryptophan 2,3-dioxygenase, Pilotte [/bib_ref].
Novel L-Trp-L-Kyn-AhR axis inhibitors, as L-Kyndegrading enzymes, direct AhR antagonists, and L-Trp mimetics are also advancing in preclinical development. Although multitargeting of L-Trp metabolism in cancer appears a rational strategy, potential side effects should of course be considered. Lastly, the nonenzymatic function of IDO1 that promotes in DCs a longterm immunosuppressive phenotype [bib_ref] Indoleamine 2,3-dioxygenase is a signaling protein in long-term tolerance by dendritic cells, Pallotta [/bib_ref] [bib_ref] A relay pathway between arginine and tryptophan metabolism confers immunosuppressive properties on..., Mondanelli [/bib_ref] could contribute to restrain the antitumor immune response in TME. Conformations that mediate both enzymatic and nonenzymatic (i.e., signaling) activities of IDO1 are in a dynamic balance in the cell . Computational studies performed by Mammoli et al. on the crystal structures of IDO1, with or without bound inhibitors, suggest that diverse inhibitors can induce different conformational changes in the small domain of IDO1 and thus affect not only its catalytic but also the signaling function. In patients with noncurable glioblastoma, the advanced age was associated with increased IDO1 expression, decreased immunotherapeutic efficacy, and was not reversed by IDO1 enzyme inhibition [bib_ref] IDO1 inhibition synergizes with radiation and PD-1 blockade to durably increase survival..., Ladomersky [/bib_ref] , proposing that targeting the signaling activity of IDO1 in combination with its catalytic function may improve IDO1-targeting immunotherapy. Indeed, the nonenzymatic function could provide either an interpretation of the IDO1 inhibitors' failure in the trials or suggest new strategies for targeting IDO1 more effectively.
## Ido1 in autoimmunity
The pleiotropic mechanisms of IDO1 are deeply involved in the pathogenesis of autoimmune conditions featured by an aberrant immune response against selfantigens. Contrary to what happens in cancer, a poor activity of IDO1 facilitates the over-activation of immune effectors in autoimmunity [bib_ref] Amino acid metabolism as drug target in autoimmune diseases, Mondanelli [/bib_ref]. A critical role is played by IDO1 + DCs that, in physiological conditions, may control immune responses by inhibiting effector T cells and promoting Treg lymphocytes in an antigen-specific manner [bib_ref] IDO and regulatory T cells: a role for reverse signalling and non-canonical..., Puccetti [/bib_ref] [bib_ref] Tolerance, DCs and tryptophan: much ado about IDO, Grohmann [/bib_ref] [bib_ref] IDO expression by dendritic cells: tolerance and tryptophan catabolism, Mellor [/bib_ref].
In type 1 diabetes (T1D), a defective IDO1 activity has been observed in DCs from nonobese diabetic (NOD) mice [bib_ref] Forced IDO1 expression in dendritic cells restores immunoregulatory signalling in autoimmune diabetes, Pallotta [/bib_ref] [bib_ref] A defect in tryptophan catabolism impairs tolerance in nonobese diabetic mice, Grohmann [/bib_ref] [bib_ref] The proteasome inhibitor bortezomib controls indoleamine 2,3-dioxygenase 1 breakdown and restores immune..., Mondanelli [/bib_ref] , in a majority of peripheral blood mononuclear cells (PBMCs) isolated from pediatric diabetic patients [bib_ref] Deficiency of immunoregulatory indoleamine 2,3-dioxygenase 1in juvenile diabetes, Orabona [/bib_ref] , and in b-cells of human pancreas from adult diabetic donors [bib_ref] Loss of IDO1 expression from human pancreatic beta-cells precedes their destruction during..., Anquetil [/bib_ref]. In both DCs from NOD mice and PBMCs from diabetic children, poor IDO1 activity is caused by a rapid turnover of the protein triggered by an IL-6-driven proinflammatory milieu. Indeed, the IDO1 defect can be corrected by blocking the IL-6 receptor (IL-6R) by means of tocilizumab (TCZ). The drug inhibits IDO1 degradation and restores normoglycemia in diabetic NOD mice via an IDO1-dependent mechanism [bib_ref] Blood and islet phenotypes indicate immunological heterogeneity in type 1 diabetes, Arif [/bib_ref]. Similarly, TCZ can restore L-Trp metabolism in PBMCs from a subgroup of T1D children characterized by a deregulated IL-6R-SOCS3 axis and a distinct frequency of a single nucleotide polymorphism (SNP), that is, IDO1 rs7820268. These observations are consistent with the data reported by Antequil et al. that describe a gradual decay of IDO1 in pancreatic b-cells during T1D pathogenesis, ranging from autoantibodypositive donors to recent-onset T1D patients, and a complete absence of IDO1 in insulin-deficient islets [bib_ref] Loss of IDO1 expression from human pancreatic beta-cells precedes their destruction during..., Anquetil [/bib_ref].
Recently, Mondanelli et al.
[28] proposed an innovative therapeutic indication for NAS, a L-Trp-derived metabolite produced along the serotonin pathway, in experimental autoimmune encephalomyelitis (EAE; an animal model of human multiple sclerosis, MS). For the first time, NAS has been described as a PAM of the IDO1 enzyme, capable of implementing the production of L-Kyn that in turn activates AhR. As a consequence, NAS reinforces the immunoregulatory IDO1-L-Kyn-AhR axis in DCs and restrains the pathogenic neuroinflammation that characterizes MS. NAS protects mice from EAE in an IDO1-and AhRdependent manner. In PBMCs isolated from patients with relapsing-remitting MS [bib_ref] Amino acid catabolism in multiple sclerosis affects immune homeostasis, Negrotto [/bib_ref] , NAS increases L-Kyn release in IFN-c-stimulated conditions without modulating IDO1 expression.
The analysis of IDO1 SNPs in both autoimmune diabetes and MS cohorts displayed a different distribution of the rs7820268 SNP among autoimmune patients and healthy subjects. Although several SNPs in IDO1 have been associated with the occurrence of autoimmune/chronic inflammatory diseases, including T1D, Crohn's disease, systemic sclerosis, and MS [bib_ref] Amino acid metabolism as drug target in autoimmune diseases, Mondanelli [/bib_ref] [bib_ref] Genetic alterations affecting the genes encoding the enzymes of the kynurenine pathway..., Boros [/bib_ref] [bib_ref] Indoleamine 2,3 dioxygenase gene polymorphisms correlate with CD8+ Treg impairment in systemic..., Tardito [/bib_ref] , the majority of them affects intronic regions or the IDO1 promoter and lacks a cause-effect relationship. A bottleneck in developing IDO1 as a predictive biomarker and promising therapeutic target in both T1D and MS is in fact the lack of a genotypeto-protein association. To date, a unique rs751360195 single variant in the IDO1 coding sequence has been identified in a 30-year-old woman affected by coeliac disorder, selective IgA deficiency, thyroiditis, and thrombocytopenia [bib_ref] A novel mutation of indoleamine 2,3-dioxygenase 1 causes a rapid proteasomal degradation..., Mondanelli [/bib_ref]. Interestingly, the single nucleotide variant (SNV) rs751360195 encodes a missense mutation (K257E) very closed to the ITIM2 motif that shortens IDO1 half-life in PBMCs. The recent identification of a specific IDO1 genotype encoding for a short-lived protein IDO1 in a patient affected by a broader dysimmune disorder extends the investigative field of defective IDO1 in further autoimmune conditions and reinforce the perspective to develop IDO1 as a predictive biomarker that will contribute to precision medicine in the autoimmunity field.
## Ido1 in neurodegeneration
KP metabolites have been closely linked to the pathogenesis of several neurodegenerative diseases, including AD, PD, and HD [bib_ref] Tryptophan metabolism as a common therapeutic target in cancer, neurodegeneration and beyond, Platten [/bib_ref] [bib_ref] The kynurenine pathway and neurodegenerative disease, Maddison [/bib_ref]. In AD, the amyloid peptide upregulates IDO1 expression and increases the production of quinolinic acid in human macrophages and microglia [bib_ref] A beta 1-42 induces production of quinolinic acid by human macrophages and..., Guillemin [/bib_ref]. Furthermore, human neurons treated with quinolinic acid upregulate genes involved in tau phosphorylation, the mechanism crucially involved in the formation of pathogenetic neurofibrillary tangles [bib_ref] The excitotoxin quinolinic acid induces tau phosphorylation in human neurons, Rahman [/bib_ref]. In HD, high levels of L-Kyn lead to high production of neurotoxic quinolinic acid and very low levels of neuroprotective kynurenic acid [bib_ref] Dysfunction of brain kynurenic acid metabolism in Huntington's disease: focus on kynurenine..., Jauch [/bib_ref] , thus suggesting the existence of an imbalance between distinct L-Trp metabolites that may play a causative role in HD pathogenesis. Such imbalance has also been observed in PD patients in dopaminergic neurons and surrounding microglia [bib_ref] The involvement of neuroinflammation and kynurenine pathway in Parkinson's disease, Zinger [/bib_ref].
As compared to other pathological conditions, targeting IDO1 and other L-Trp-metabolizing enzymes in neurodegenerative diseases is still in its infancy. Nevertheless, it appears that efforts in this direction might very likely be worth making.
## Concluding remarks
By comparing the effects of carbon monoxide on the conformations of human IDO1 versus human TDO, Syun-Ru Yeh and colleagues very recently found that IDO1 exhibits remarkable conformational plasticity as compared to the more rigid TDO [bib_ref] Conformational plasticity in human heme-based dioxygenases, Pham [/bib_ref]. The authors hypothesized that this structural fingerprint may be relevant for IDO1 being a multitasking protein under a harsh inflammatory environment [bib_ref] Conformational plasticity in human heme-based dioxygenases, Pham [/bib_ref]. Lacking completely in TDO, the small domain and interconnecting loop may thus represent the main structural parts of IDO1 modulating its conformational as well as functional plasticity. As a matter of fact, all identified motifs allowing IDO1 to establish a partnership with other molecules, that is, SHPs, PI3Ks, and SOCS3, have been localized in its small domain. Therefore, a peculiar overall architecture could imprint IDO1 with the flexible capacity to promote Trp metabolism, its expression, or its proteasomal degradation, meeting specific cellular needs in specific microenvironmental conditions. In this regard, it should be mentioned that several multitasking proteins undergo structural transformation when performing distinct functions [bib_ref] Multitasking immune Sp185/333 protein, Rsptransformer-E1, and Its recombinant fragments undergo secondary structural..., Lun [/bib_ref] [bib_ref] The many structural faces of calmodulin: a multitasking molecular jackknife, Kursula [/bib_ref]. An additional level of complexity derives from the IDO1 ability to catalyze other catalytic reactions in addition to its dioxygenase activity, including peroxygenase [bib_ref] Indole peroxygenase activity of indoleamine 2,3-dioxygenase, Kuo [/bib_ref] , peroxidase [bib_ref] Human indoleamine 2,3-dioxygenase is a catalyst of physiological heme peroxidase reactions: implications..., Freewan [/bib_ref] , and nitrite reductase [bib_ref] Human indoleamine 2,3-dioxygenase 1 is an efficient mammalian nitrite reductase, Lim [/bib_ref] activities as well as the generation of singlet oxygen [bib_ref] Singlet molecular oxygen regulates vascular tone and blood pressure in inflammation, Stanley [/bib_ref]. Of note, the production of singlet oxygen by IDO1 leads to the formation of a tricyclic hydroperoxide that decreases blood pressure [bib_ref] Singlet molecular oxygen regulates vascular tone and blood pressure in inflammation, Stanley [/bib_ref]. Therefore, the versatility of IDO1 likely extends to the catalysis of several biologically relevant enzyme reactions.
Although we do not know where binding of SOCS3 and ubiquitination may occur yet (but we may hypothesized the cytosol, given the cytosolic localization of the proteasome complex), IDO1 enzymic and nonenzymic activities take place in the cytosol and EE, respectively, at least in DCs and IDO1-transfected P815 mouse mastocytoma cells [bib_ref] Class IA PI3Ks regulate subcellular and functional dynamics of IDO1, Iacono [/bib_ref]. An important question would be how the apo form of IDO1 influences responsiveness to specific cellular microenvironments (i.e., cytokines, TLR ligands, Trp, and O 2 concentration, pH), and thus, the choice of a specific function or fate. Vice versa, specific microenvironments may change the balance between holo-and apo-IDO1. Moreover, which is the intra-(and possibly extra-) cellular topology of apo-IDO1? Presumably, because tumor transfectants expressing a mouse IDO1 mutant mimicking tyrosine-phosphorylated ITIM1 perform IDO1 signaling activity but do not bind heme and thus do not produce L-Kyn , the apoprotein may be associated with EE. Therefore, a systematic study of the percentages of apo-IDO1, dynamics of cellular localization, and performance of a specific function, either in basal conditions or in response to specific stimuli, should be undertaken in both primary cells (including DCs but also macrophages, fibroblasts, and epithelial cells known to express IDO1) and tumors, of mouse and human origin. The delineation of the biologic and functional profile of IDO1 in tumor contexts may lead to the generation of novel drugs targeting this important protein in neoplastic patients and, perhaps, most importantly, could be more effective than the currently available catalytic inhibitors. In this regard, it would be very interesting to know whether the catalytic inhibitors targeting apo-IDO1 [bib_ref] Immune-modulating enzyme indoleamine 2,3-dioxygenase is effectively inhibited by targeting its apo-form, Nelp [/bib_ref] -and preventing heme binding -do also influence the IDO1 signaling activity, which may be more important in a chronic disease like neoplasia. In chronic inflammation and autoimmunity, drug targeting of IDO1 (and enzymes in general) appears to be a more difficult task. In fact, in these conditions, IDO1 function/s should be potentiated rather than inhibited. However, the very recent discovery of VIS351 [bib_ref] Identification of a 2-propanol analogue modulating the non-enzymatic function of indoleamine 2,3-dioxygenase..., Albini [/bib_ref] , an inducer of IDO1 signaling, and NAS [28], an enhancer of IDO1 catalytic activity protecting from neuroinflammation, may bring some fresh air in IDO1-based immunotherapies.
Therefore, the great amount of work that has been performed on IDO1 in the most recent years raises new biologic questions but also opens new exciting perspectives for therapeutic advances in several disease areas.
[fig] Figure 2: Crystal structure of substrate-bound IDO1 (pdb code: 5WMU). Key structural elements are highlighted and labeled: ITIMs (Tyr111, Tyr249, orange CPK atoms); YENM motif (orange CPK atom sticks); narrow channel (cyan surface). Heme cofactor (green carbon atoms) and L-Trp (orange carbon atom sticks).6101 The FEBS Journal 289 (2022) 6099-6118 ª 2021 The Authors. The FEBS Journal published by John Wiley & Sons Ltd on behalf of Federation of European Biochemical Societies [/fig]
[fig] Figure 3: Binding mode of L-Trp (orange carbon atom sticks) into the catalytic site of IDO1 (pdb code: 5WMU). Hydrogen bond interactions between L-Trp and binding site residues are shown in black dashed lines. The EF loop is highlighted with a purple cartoon. [/fig]
[fig] Figure 4: Binding mode of L-Trp (orange carbon atom sticks) into the accessory site of the Phe270Gly variant of IDO1 (pdb code: 5WMW). L-Trp engages a cluster of nonpolar residues (Val170, Val269, Leu342) in hydrophobic contacts, whereas no polar interactions are observed. 6102 The FEBS Journal 289 (2022) 6099-6118 ª 2021 The Authors. The FEBS Journal published by John Wiley & Sons Ltd on behalf of Federation of European Biochemical Societies [/fig]
[fig] Figure 5: Intracellular dynamics of IDO1 in DCs. Depending on the microenvironment, IDO1 shapes its conformation and acquires functions suitable for cellular needs. (A) In acute inflammation, the pro-inflammatory cytokine IFN-c induces IDO1 enzymic activity, promoting transformation of L-Trp into L-Kyn, an agonist of AhR, a ligand-activated transcription factor that moves from cytosol to the nucleus and, via a positive feedback loop, upregulates Ido1 gene expression. (B) [/fig]
[fig] 6103 The: FEBS Journal 289 (2022) 6099-6118 ª 2021 The Authors. The FEBS Journal published by John Wiley & Sons Ltd on behalf of Federation of European Biochemical Societies [/fig]
[fig] 6104 The: FEBS Journal 289 (2022) 6099-6118 ª 2021 The Authors. The FEBS Journal published by John Wiley & Sons Ltd on behalf of Federation of European Biochemical Societies [/fig]
[fig] 6105 The: FEBS Journal 289 (2022) 6099-6118 ª 2021 The Authors. The FEBS Journal published by John Wiley & Sons Ltd on behalf of Federation of European Biochemical Societies [/fig]
[fig] 6106 The: FEBS Journal 289 (2022) 6099-6118 ª 2021 The Authors. The FEBS Journal published by John Wiley & Sons Ltd on behalf of Federation of European Biochemical Societies [/fig]
[fig] 6107 The: FEBS Journal 289 (2022) 6099-6118 ª 2021 The Authors. The FEBS Journal published by John Wiley & Sons Ltd on behalf of Federation of European Biochemical Societieswith a TDO inhibitor[119], thus pinpointing TDO as a potential target for the development of selective antitumoral drugs. [/fig]
[fig] 6108 The: FEBS Journal 289 (2022) 6099-6118 ª 2021 The Authors. The FEBS Journal published by John Wiley & Sons Ltd on behalf of Federation of European Biochemical Societies [/fig]
[fig] 6109 The: FEBS Journal 289 (2022) 6099-6118 ª 2021 The Authors. The FEBS Journal published by John Wiley & Sons Ltd on behalf of Federation of European Biochemical Societies [/fig]
[fig] 6110 The: FEBS Journal 289 (2022) 6099-6118 ª 2021 The Authors. The FEBS Journal published by John Wiley & Sons Ltd on behalf of Federation of European Biochemical Societies [/fig]
|
Differential Pharmacological Actions of Methadone and Buprenorphine in Human Embryonic Kidney 293 Cells Coexpressing Human μ-Opioid and Opioid Receptor-Like 1 Receptors
Methadone and buprenorphine are used in maintenance therapy for heroin addicts. In this study, we compared their effects on adenylate cyclase (AC) activity in human embryonic kidney (HEK) 293 cells stably overexpressing human l-opioid receptor (MOR) and nociceptin/opioid receptor-like 1 receptor (ORL1) simultaneously. After acute exposure, methadone inhibited AC activity; however, buprenorphine induced compromised AC inhibition. When naloxone was introduced after 30 min incubation with methadone, the AC activity was enhanced. This was not observed in the case of buprenorphine. Enhancement of the AC activity was more significant when the incubation lasted for 4 h, and prolonged exposure to buprenorphine elevated the AC activity as well. The removal of methadone and buprenorphine by washing also obtained similar AC superactivation as that revealed by naloxone challenge. The study demonstrated that methadone and buprenorphine exert initially different yet eventually convergent adaptive changes of AC activity in cells coexpressing human MOR and ORL1 receptors.
# Introduction
Three conventional opioid receptors-l (MOR), d (DOR), and j (KOR)-have been characterized based on their pharmacological, anatomical, and molecular properties [bib_ref] International union of pharmacology. XII. Classification of opioid receptors, Dhawan [/bib_ref] [bib_ref] Molecular pharmacology of the opioid receptors, Satoh [/bib_ref] [bib_ref] Modulation of opioid receptor function by protein-protein interactions, Alfaras-Melainis [/bib_ref]. A non-opioid branch of opioid receptors has also been identified and named as opioid receptor-like orphan receptor (ORL1) [bib_ref] ORL1, a novel member of the opioid receptor family. Cloning, functional expression..., Mollereau [/bib_ref]. This receptor family was renamed after identification of its endogenous peptidergic agonist, nociceptin or orphanin FQ, as the nociceptin/orphanin FQ peptide (NOP) receptor; it displays pharmacology distinct from those of conventional opioid receptors [bib_ref] Nociceptin/orphanin FQ peptide receptors: pharmacology and clinical implications, Chiou [/bib_ref] [bib_ref] Opioid receptor signalling mechanisms, Connor [/bib_ref]. Activation of l-, d-, jor ORL1 receptor produces common cellular actions by regulating the same secondary messengers, including inhibition of adenylate cyclase (AC) activity [bib_ref] Cloning of a delta opioid receptor by functional expression, Evans [/bib_ref] [bib_ref] Cloning and pharmacological characterization of a rat kappa opioid receptor, Meng [/bib_ref] and N-type [bib_ref] The cloned kappa opioid receptor couples to an N-type calcium current in..., Tallent [/bib_ref] and L-type Ca 2? channels [bib_ref] Voltage-dependent inhibition of Ca 2? channels in GH3 cells by cloned mu-and..., Piros [/bib_ref]. Activation of opioid receptors also increases phospholipase C activity, causes a transient increase in intracellular Ca 2? [bib_ref] Mobilization of Ca 2? from intracellular stores in transfected neuro2a cells by..., Spencer [/bib_ref] , and activates inwardly rectifying K ? channels [bib_ref] Functional coupling of the nociceptin/orphanin FQ receptor with the G-protein-activated K ?..., Ikeda [/bib_ref] and mitogen-activated protein kinases (MAPK) [bib_ref] Functional coupling of the delta-, mu-, and kappa-opioid receptors to mitogen-activated protein..., Fukuda [/bib_ref].
Methadone and buprenorphine are currently used in maintenance treatment programs for heroin addicts [bib_ref] Pharmacotherapy of addictions, Kreek [/bib_ref] [bib_ref] Innovations in agonist maintenance treatment of opioid-dependent patients, Haasen [/bib_ref]. Methadone is an orally available synthetic opioid functioning as a full agonist of MOR. First tested as a treatment for heroin addicts in early 1964 at The Rockefeller University, it provided a ''blockade'' against the effects of superimposed heroin through the mechanism of opioid cross-tolerance [bib_ref] Pharmacotherapy of addictions, Kreek [/bib_ref]. Buprenorphine, a derivative of thebaine, differs from standard opioid agonists in two aspects. The first is the slow receptor dissociation kinetics featured with the biphasic (''bell''-or ''inverted U''-shaped) dose-response relation [bib_ref] Buprenorphine-induced antinociception is mediated by mu-opioid receptors and compromised by concomitant activation..., Lutfy [/bib_ref] [bib_ref] Broad analgesic profile of buprenorphine in rodent models of acute and chronic..., Christoph [/bib_ref]. The second is that buprenorphine has a ceiling effect on respiratory depression in humans, suggesting a greater safety margin of buprenorphine relative to other clinical opioids [bib_ref] Buprenorphine induces ceiling in respiratory depression but not in analgesia, Dahan [/bib_ref]. Methadone is a long-acting MOR agonist with pharmacological properties qualitatively similar to those of morphine, whereas buprenorphine is a MOR partial agonist and a potent j-opioid receptor antagonistas well as an ORL1 agonist [bib_ref] Buprenorphine-induced antinociception is mediated by mu-opioid receptors and compromised by concomitant activation..., Lutfy [/bib_ref] [bib_ref] Agonistic effects of the opioid buprenorphine on the nociceptin/OFQ receptor, Bloms-Funke [/bib_ref].
Adaptive changes in neurons underlie altered behaviors associated with opioid dependence and withdrawal syndrome [bib_ref] Neurobiology of craving, conditioned reward and relapse, Weiss [/bib_ref]. Adaptations affect neuronal excitability, synaptic transmission, transcription factors, and MAPK [bib_ref] Opioids: cellular mechanisms of tolerance and physical dependence, Bailey [/bib_ref]. Prolonged exposure of NG108-15 neuroblastoma x glioma hybrid cells (expressing mainly d-opioid receptors) to morphine leads to increased AC activity [bib_ref] Opiate-dependent modulation of adenylate cyclase, Sharma [/bib_ref] , suggesting this phenomenon may underlie the withdrawal state. Withdrawal of the agonist by washing (natural withdrawal) or by adding the antagonist naloxone (precipitated withdrawal), which relieves the inhibition of AC exerted by the agonist, revealed the phenomenon of AC superactivation or overshoot. Such regulation of AC could be a general means of cellular adaption to the alteration of opioid receptors [bib_ref] Chronic morphine augments adenylyl cyclase phosphorylation: relevance to altered signaling during tolerance/dependence, Chakrabarti [/bib_ref] [bib_ref] Chronic opioid treatment induces adenylyl cyclase V superactivation. Involvement of Gbetagamma, Avidor-Reiss [/bib_ref].
In several subpopulations of CNS neurons involved in pain regulation, MOR and ORL1 are coexpressed [bib_ref] Distribution of nociceptin/orphanin FQ precursor protein and receptor in brain and spinal..., Houtani [/bib_ref] [bib_ref] Dimerization of morphine and orphanin FQ/nociceptin receptors: generation of a novel opioid..., Pan [/bib_ref]. Furthermore, heterodimerization of MOR and ORL1 impairs the potency of MOR agonist [bib_ref] Heterodimerization of opioid receptor-like 1 and mu-opioid receptors impairs the potency of..., Wang [/bib_ref] and attenuates ORL1-mediated inhibition of N-type channels [bib_ref] Heterodimerization of ORL1 and opioid receptors and its consequences for N-type calcium..., Evans [/bib_ref]. Interestingly, mice lacking the ORL1 gene partially lose tolerance liability to morphine analgesia [bib_ref] Partial loss of tolerance liability to morphine analgesia in mice lacking the..., Ueda [/bib_ref] and show marked attenuation of morphine-induced physical dependence, manifested as naloxone-precipitated withdrawal symptoms after repeated morphine treatments [bib_ref] Enhanced spinal nociceptin receptor expression develops morphine tolerance and dependence, Ueda [/bib_ref]. Hence, coexpressed MOR?ORL1 may reflect the native opioid receptors in some CNS regions and could provide the insight on the development of dependence on opioid drugs.
The human embryonic kidney (HEK) 293 cell line is a widely-distributed mammalian cell expression system and shares similar protein expression profiles with human neuronal cells [bib_ref] Preferential transformation of human neuronal cells by human adenoviruses and the origin..., Shaw [/bib_ref]. In this study, we have established an in vitro cell model overexpressing human MOR and ORL1 individually or simultaneously in human embryonic kidney (HEK) 293 cells. Alterations of AC activity after acute or prolonged exposure to methadone and buprenorphine were compared in this model. Naloxone was utilized to demonstrate the specificity of the drugs and elicit the precipitated withdrawal. Effects of morphine and Ro 64-6198 were also examined as positive control for MOR and ORL1, respectively.
## Experimental procedure
## Molecular cloning of human l-opioid and orl1 receptors
The full-length cDNA clones encoding the human MOR (Clone ID 30915262) and ORL1 receptor (Clone ID 5164017) were purchased from Open Biosystems (Huntsville, AL, USA). HA epitope (YPYDVPDYA) was added to the N-terminus of MOR with the aid of PCR amplification. Subsequently, cDNA of HA-tagged MOR was subcloned into a mammalian expression vector, pcDNA4/V5-His C (Invitrogen, Carlsbad, CA, USA), which is a zeocin-selectable vector [bib_ref] Heterodimerization of opioid receptor-like 1 and mu-opioid receptors impairs the potency of..., Wang [/bib_ref]. The cDNA encoding ORL1 was subcloned into a mammalian expression vector, pCMV-Tag3 (Stratagene, La Jolla, CA, USA), which is a geneticin-selectable vector providing a myc tag (EQKLISEEDL) to the N-terminus. All sequences were verified by DNA sequence analysis.
Stable Expression of Human l-Opioid and ORL1 Receptors in HEK 293 Cells HEK 293 cells were grown in minimal essential medium (Invitrogen) supplemented with 10% fetal bovine serum, 100 units/ml penicillin, and 100 lg/ml streptomycin. Cell cultures were maintained at 37°C in a humidified 5% CO 2 incubator. The pcDNA4/V5-His C vector containing cDNA of HA-tagged human MOR or the pCMV-Tag3 vector containing cDNA of myc-tagged human ORL1 was transfected to HEK 293 cells by lipofection using FuGENE HD (Roche, Mannheim, Germany). Cell lines stably expressing HA-tagged MOR and myc-tagged ORL1 were selected by adding zeocin (0.5 mg/ml) and geneticin (0.5 mg/ml) to the culture medium, respectively. Surface expression of HA-tagged human MOR or myc-tagged human ORL1 was confirmed by measuring agonist-mediated inhibition of forskolin-induced cAMP accumulation [bib_ref] Heterodimerization of opioid receptor-like 1 and mu-opioid receptors impairs the potency of..., Wang [/bib_ref].
## Receptor deglycosylation
HEK cells stably expressing MOR or ORL1 were grown to near confluence in 10 cm dishes. Cell extracts were prepared by incubating the cells in 0.4 ml of lysis buffer-composed of 150 mM Tris-HCl (pH 7.4), 300 mM NaCl, 1 mM MgCl 2 , 1 mM CaCl 2 , 1% Triton X-100, 10% glycerol, and 1% protease inhibitor mixture (containing aprotinin, leupeptin, pepstatin A, 4-(2-aminoethyl)benzenesulfonyl fluoride hydrochloride, bestatin, and E-64; Sigma-Aldrich, St. Louis, MO, USA)-for 1 h on ice. Cell debris was precipitated by centrifugation at 14 000 g for 10 min at 4°C and the supernatant was used for the analysis. Protein concentration of the supernatant was determined using the BCA assay (Pierce Biotechnology, Rockford, IL, USA) with bovine serum albumin as standard. To investigate receptor glycosylation, 20 lg protein was incubated with 1 unit N-glycosidase F (Roche) in deglycosylation buffer-consisting of 25 mM sodium phosphate buffer (pH 7.2), 25 mM EDTA, 0.1% SDS, and 1% (v/v) 2-mercaptoethanol-at 37°C for 3 h [bib_ref] Proteasome involvement in agonist-induced down-regulation of mu and delta opioid receptors, Chaturvedi [/bib_ref]. For immunoblotting, treated and untreated lysates were diluted with 69 gel loading buffer (300 mM Tris-Cl (pH 6.8), 12% (w/v) SDS, 0.3% (w/v) bromophenol blue, 60% (v/ v) glycerol, and 600 mM b-mercaptoethanol); and proteins were resolved using 10% SDS polyacrylamide gel and then transferred to polyvinylidene difluoride membranes (Immobilon; Millipore, Billerica, MA, USA). Membranes were incubated with monoclonal anti-HA or anti-myc antibody (1:1000 dilution; Cell Signaling Technology, Danvers, MA, USA) overnight at 4°C. After being washed, membranes were incubated with sheep anti-mouse horseradish peroxidase-linked secondary antibody (GE Healthcare Life Sciences, Piscataway, NJ, USA). Subsequently, immunoreactive proteins on the membrane were visualized by enhanced chemiluminescence (SuperSignal West Pico chemiluminescent substrate kit; Pierce Biotechnology, Rockford, IL, USA). Molecular weights were determined using ImageQuant TL (GE Healthcare Life Sciences).
## Radioligand binding assays
HEK 293 cells were washed twice and harvested on ice in Versene solution containing 0.2 g/L EDTA- 4Na in phosphate-buffered saline (Invitrogen) and centrifuged at 500 g for 3 min at 4°C. The cell pellet was suspended in buffer A-consisting of 5 mM Tris-Cl (pH 7.4), 5 mM EDTA, 5 mM EGTA, and 0.1 mM phenylmethylsulfonyl fluoride-passed through a 26-gauge 1/2 needle 5 times, and then centrifuged at 48 000 g for 30 min. The membrane pellet was resuspended using a Polytron homogenizer in buffer Bcomposed of 50 mM Tris-Cl (pH 7.0) and 0.32 mM sucrose-aliquoted, frozen in dry ice/ethanol, and stored at -80°C. Protein concentration of the membrane preparation was measured by the Bradford method (Bio-Rad protein assay kit, Bio-Rad Laboratories, Hercules, CA, USA).
Saturation radioligand binding assay was performed using opaque white 96-well filter plates with FB glass fiber filters (model MSFB N6B, Multiscreen Assay System; Millipore). Cell membranes (8 * 12 and 50 lg of protein/ well for nociceptin and DAMGO binding, respectively) were incubated with various concentrations of [ 3 H]-nociceptin (PerkinElmer Life Analytical Sciences, Boston, MA, USA) or [ 3 H]-DAMGO (PerkinElmer) in binding buffer consisting of 50 mM Tris-Cl (pH 7.4) and 1 mM EGTA for 1 h at 25°C. Non-specific binding was determined by adding 3 lM nociceptin (Tocris Bioscience, Ellisville, Missouri, USA) or DAMGO (Tocris) to the reaction mixture. The reaction was terminated by rapid filtration, and the filters were washed 3 times with ice-cold binding buffer and dried at room temperature, overnight. After adding MicroScint-20 cocktail (PerkinElmer), bound radioactivity was measured using the TopCount NXT microplate scintillation and luminescence counter (Perkin-Elmer). Prism (GraphPad Software, La Jolla, CA, USA) was used to analyze the data derived from the saturation binding assay to obtain B max and K D values.
## Confocal microscopy and image analysis
Cells were grown on microscope cover glasses (Fisher Scientific, Pittsburgh, PA, USA) and incubated for 2-3 days prior to immunocytochemistry. Immunostaining was performed by incubating the cells at 37°C with 1:100 dilution of monoclonal anti-HA (Cell Signaling Technology) or polyclonal anti-ORL1 (raised against the N-terminus of the human OPRL1 receptor, MEPLFPAPFWEVIYGSHL, and affinity-purified by Pro-Sci Incorporated, Poway, CA, USA) antibody in complete medium for 1.5 h. After three washes in complete medium and two washes in PBS ? (1 9 phosphate-buffered saline containing 1 mM MgCl 2 and 0.1 mM CaCl 2 ), cells were fixed with 4% paraformaldehyde at room temperature for 15 min. After quenching with two washes with 50 mM NH 4 Cl in PBS ? and washed once with PBS ? , cells were permeabilized with 0.1% Triton X-100 in PBS ? at room temperature for 15 min. To remove excess Triton X-100, cells were washed 5 times with PBS ? at room temperature. Nonspecific binding was then blocked by incubating the cells with 10% BSA in PBS ? at room temperature for 30 min. The secondary antibody (1:500 dilution of Alexa Fluor 488 goat anti-mouse or Alexa Fluor 647 anti-rabbit antibody, Invitrogen) was applied at 4°C overnight. Cells were then washed 3 times with PBS ? and mounted (Pro-Long Gold Antifade Kit, Invitrogen) for imaging. Images were acquired using a Leica TCS SP5 confocal microscope (Leica Microsystems CMS GmbH, Mannheim, Germany) with a 63 9 1.4 NA oil immersion objective in the inverted configuration. Quantitative analysis of colocalization rate was performed using Leica LAS-AF software based on 30% background subtractions in both receptors and 30% threshold for determining colocalization.
## Htrf camp assays
The cAMP quantification was performed using a homogeneous time-resolved fluorescence (HTRF) cAMP detection kit (cAMP HiRange; Cisbio, Bagnols/Cèze Cedex, France). HEK 293 cells were dispensed with 25 ll of compound buffer consisted of minimal essential medium supplemented with 0.5 mM isobutylmethylxanthine (Sigma-Aldrich), 0.2% fatty acid-free bovine serum albumin (Sigma-Aldrich), 0.5 mg/ml zeocin, and/or 0.5 mg/ml geneticin at 2-6 9 10 4 cells/well in 96 half-well plates (Costar, Corning, NY, USA) on the day of the experiment. After an incubation of 1 h at 37°C in a humidified 5% CO 2 incubator, 25 ll of compound buffer containing 10 lM forskolin and desired concentrations of methadone (United States Pharmacopeia, Rockville, MD, USA) or buprenorphine (Sigma-Aldrich) were added to the cells, followed by 30 min incubation at room temperature. Morphine (National Bureau of Controlled Drugs, Taipei, Taiwan) and Ro 64-6198 (a gift from F. Hoffmann-La Roche Ltd., Basel, Switzerland)-which are a MOR agonist and an ORL1 agonist, respectively-were included as the positive control. Subsequently, 25 ll of cAMP-d2 and 25 ll of anti-cAMP Cryptate conjugate were added to each well. After 1 h incubation at room temperature, the plate was read on a FlexStation 3 microplate reader (Molecular Devices, Silicon Valley, CA, USA) with emission wavelength at 615 and 665 nm.
To verify the specificity of the drugs, 1 lM naloxone (Sigma-Aldrich) was included in the compound buffer containing 10 lM forskolin and 0.1 lM opioids, followed by 30 min co-incubation at room temperature. For evaluation of AC superactivation, desired concentrations of drugs were added to the compound buffer and incubated at 37°C for 30 min or 4 h; the compound buffer was then replaced by either 10 lM forskolin only or in combination with 1 lM naloxone. Afterward, the incubation with cAMP-d2 and anti-cAMP Cryptate conjugate was immediately carried out as stated above. The cAMP concentrations were calculated by nonlinear regression analysis with SoftMax Pro (Molecular Devices, Sunnyvale, CA, USA). Concentration-response curves of cAMP accumulation, potency (pIC 50 ) and efficacy (E max ) for inhibition of forskolin-stimulated cAMP formation by morphine, methadone, buprenorphine, and Ro 64-6198 were analyzed using Prism (GraphPad Software) [bib_ref] Quantitative highthroughput screening using a live-cell cAMP assay identifies small-molecule agonists of..., Titus [/bib_ref] [bib_ref] High throughput screening technologies for direct cyclic AMP measurement, Gabriel [/bib_ref].
## Statistical analyses
All results are expressed as the mean ± SE value of n experiments. Paired/unpaired t test (two-tailed) or oneway/two-way ANOVA followed by Bonferroni's test was used to determine whether the difference is statistically significant (P \ 0.05).
# Results
## Establishment of hek 293 cells stably expressing ha-tagged mor and myc-tagged orl1
We have established an in vitro cell model by overexpressing epitope-tagged human MOR and ORL1 in HEK 293 cells. Plasmids harboring HA-tagged MOR or myctagged ORL1 were individually or simultaneously transfected in HEK 293 cells, and the stable clones were selected by appropriate antibiotics. Since adenylate cyclase (AC) activity is the major endpoint measured in this study, the three stable clones presented here (MOR-, ORL1-, and MOR?ORL1-expressing cells) were chosen based on the strongest adenylate AC inhibition by acute treatment of 1 lM of DAMGO and/or nociceptin compared to other stable clones during the initial screening. Subsequent experiments also demonstrated similar AC inhibition levels between the MOR-and MOR?ORL1-expressing cells or those between the ORL1-and MOR?ORL1-expressing cells elicited by MOR or ORL1 agonists and. Western blot analysis using monoclonal anti-HA antibody revealed that MOR was expressed as two major heterogeneous forms, with apparent molecular masses of *43-50 and 72-80 kilodaltons (kDa) [fig_ref] Figure 1: MOR and ORL1 expressed in HEK 293 cells are N-linked glycoproteins [/fig_ref] , Left panel). Due to its smaller mass, the ORL1 revealed by antimyc antibody migrated more quickly as two major diffused bands with apparent molecular masses of *46-50 and 67-75 kDa. The minor band at *55 kDa is supposedly a non-specific protein elicited by overexpression of the plasmid harboring the myc-ORL1 receptor [fig_ref] Figure 1: MOR and ORL1 expressed in HEK 293 cells are N-linked glycoproteins [/fig_ref].
Since the apparent molecular weights of the overexpressed MOR and ORL1 were larger than the expected values, we hypothesized that the overexpressed human MOR and ORL1 might undergo glycosylation [bib_ref] Molecular pharmacology of the opioid receptors, Satoh [/bib_ref]. Evidence that MOR and ORL1 are glycoproteins was provided by digestion with N-glycosidase F, an amidase that cleaves nearly all types of N-glycan chains from the asparagines in N-linked glycoproteins. N-Glycosidase F treatment of MOR and ORL1 increased the mobility of both bands to species of apparent molecular masses of 43 kDa for MOR, and 39 and 41 kDa for ORL1. The predicted molecular masses for the HA-tagged MOR and myc-tagged ORL1 are *45 and 43 kDa, respectively. The reason for the aberrant electrophoretic mobility of the receptors, even after deglycosylation, is unknown at present [bib_ref] Proteasome involvement in agonist-induced down-regulation of mu and delta opioid receptors, Chaturvedi [/bib_ref]. The immunoreactive bands with slower mobility may represent opioid receptor dimers described previously [bib_ref] Modulation of opioid receptor function by protein-protein interactions, Alfaras-Melainis [/bib_ref] [bib_ref] G-protein-coupled receptor heterodimerization modulates receptor function, Jordan [/bib_ref].
In saturation radioligand binding studies, [ 3 H]-DAMGO displayed a similar affinity (K D ) for both stably transfected cell lines harboring MOR; but cells expressing MOR?ORL1 possess more DAMGO-binding sites than those expressing MOR alone, as reflected by the significantly higher B max Neurochem Res (2011) value [fig_ref] Table 1 B: max [/fig_ref]. In contrast, [ 3 H]-nociceptin showed significantly higher affinity (lower K D value) for cells expressing both MOR and ORL1 than cells expressing ORL1 alone; yet cells co-expressing MOR and ORL1 possess fewer nociceptin-binding sites than those expressing only ORL1, as demonstrated by the lower B max value [fig_ref] Table 1 B: max [/fig_ref]. We next examined whether MOR colocalizes with coexpressed ORL1. Shown in are confocal fluorescence images of HEK 293 cells expressing HA-tagged MOR and myc-tagged ORL1. ORL1 was clearly present in vesicles distributed throughout the cytoplasm and also on the plasma membrane. MOR localizes to the cell surface as well as vesicular structures, and prominently colocalizes with ORL1 (colocalization rate: 83.65 ± 4.33%).
## Ac inhibition after acute opioid exposure
## Methadone inhibited ac as a mor agonist
Effects of acute exposure to morphine, methadone, buprenorphine, and Ro 64-6198 on MOR-or ORL1-mediated Ga i/o -coupled adenylate cyclase (AC) inhibition were examined. Morphine and methadone concentration-dependently inhibited forskolin-stimulated cAMP accumulation in HEK 293 cells expressing MOR only and b, open circles;as well as MOR?ORL1 and b, filled squares;. Methadone was somewhat more potent than morphine in cells expressing only MOR ( 7.024 ± 0.193), although the difference is not statistically significant. In HEK 293 cells stably expressing ORL1, morphine and methadone did not inhibit forskolin-stimulated AC activity . These results suggest that methadone acts as a potent MOR agonist comparable to morphine and has no effect on ORL1.
Buprenorphine Acted as an ORL1 Agonist and a Partial MOR Agonist Unlike methadone, buprenorphine displayed a flat concentration-inhibition curve on cAMP accumulation in MOR-expressing cells, with efficacy of only 33.52 ± 3.38% inhibition , open circles;, supporting its partial agonist characteristic at MOR. In ORL1-expressing cells, buprenorphine at higher concentrations ([30 nM) inhibited cAMP accumulation in a concentration-dependent manner , open triangles;, as did Ro 64-6198 , open triangles;, a non-peptide ORL1 agonist [bib_ref] A synthetic agonist at the orphanin FQ/nociceptin receptor ORL1: anxiolytic profile in..., Jenck [/bib_ref]. This suggests that buprenorphine acts as an ORL1 agonist at higher concentrations. Interestingly, buprenorphine showed a significantly greater potency (higher pIC 50 )but lower efficacy (lower E max )in MOR?ORL1-coexpressing cells than in cells expressing ORL1 only ;. However, the AC inhibition curves elicited by Ro 64-6198 were comparable in ORL1and MOR?ORL1-expressing cells . This suggests that the effect of buprenorphine on MOR?ORL1expressing cells is not solely owing to its agonistic property on ORL1 receptor.
## Interactions of naloxone
Naloxone, a generic opioid receptor antagonist, was co-incubated with 0.1 lM morphine, methadone and buprenorphine to verify if their effects were mediated by MOR. At 1 lM, naloxone did not affect forskolin-stimulated cAMP accumulation per se (nalox). However, it completely reversed the inhibitory effects of morphine and methadone; but it only slightly reversed the inhibitory effect of buprenorphine in cells expressing MOR [fig_ref] Figure 4: Blockade by naloxone of the morphine, methadone, and buprenorphine inhibition on forskolin-stimulated... [/fig_ref] and MOR?ORL1 [fig_ref] Figure 4: Blockade by naloxone of the morphine, methadone, and buprenorphine inhibition on forskolin-stimulated... [/fig_ref]. Additionally, naloxone did not affect the Ro 64-6198-induced inhibition on cAMP accumulation [fig_ref] Figure 4: Blockade by naloxone of the morphine, methadone, and buprenorphine inhibition on forskolin-stimulated... [/fig_ref] , suggesting that naloxone at 1 lM specifically targets MOR and leaves ORL1 unaffected. Naloxone, if added at the end of 30 min opioid exposure, relieved AC inhibition by morphine in cells expressing MOR or MOR?ORL1 and even caused a rebound facilitation of AC activity in the lM range [fig_ref] Figure 5: Effects of naloxone on acute exposures to morphine, methadone, buprenorphine, or Ro... [/fig_ref]. In cells treated with methadone, the AC activity was changed in a profile reminiscent of those exposed to morphine after naloxone challenge [fig_ref] Figure 5: Effects of naloxone on acute exposures to morphine, methadone, buprenorphine, or Ro... [/fig_ref]. Naloxone completely relieved buprenorphineinduced AC inhibition but did not induce rebound AC activity in cells expressing MOR and MOR?ORL1 [fig_ref] Figure 5: Effects of naloxone on acute exposures to morphine, methadone, buprenorphine, or Ro... [/fig_ref]. Interestingly, naloxone also significantly relieved the AC inhibition induced by buprenorphine at higher concentrations [fig_ref] Figure 5: Effects of naloxone on acute exposures to morphine, methadone, buprenorphine, or Ro... [/fig_ref] as well as that induced by Ro 64-6198 [fig_ref] Figure 5: Effects of naloxone on acute exposures to morphine, methadone, buprenorphine, or Ro... [/fig_ref] in ORL1-expressing cells. These results demonstrate that naloxone relieved AC inhibition induced by MOR agonists (morphine, methadone, and buprenorphine) and even induced rebound AC activity after 30 min exposure to higher concentrations of MOR agonists (morphine and methadone) but not after similar exposure to a partial agonist (buprenorphine). Naloxone also partially relieved the AC inhibition induced by ORL1 agonists, buprenorphine [fig_ref] Figure 5: Effects of naloxone on acute exposures to morphine, methadone, buprenorphine, or Ro... [/fig_ref] and Ro 64-6198 [fig_ref] Figure 5: Effects of naloxone on acute exposures to morphine, methadone, buprenorphine, or Ro... [/fig_ref] , at higher concentrations in ORL1-expressing cells. Values represent the mean ± SE of four experiments performed in duplicate as described in * indicates a significant difference (P \ 0.05) between the pIC 50 value of ORL1-expressing cells and that of MOR?ORL1-expressing cells, according to two-way ANOVA with Bonferroni correction. N/C = not converged. Lowercase letters (a-d) denote the points of comparisons described in the text AC Superactivation After Chronic Opioid Treatment
## Ac superactivation after naloxone challenge following long-term opioid exposure
The above results show that rebound AC activity can be induced by naloxone in cells exposed to MOR agonists for only 30 min. Accordingly, we extended our investigation by adding naloxone to cells exposed to opioids for 4 h, an incubation period reported to show a prominent overshoot in forskolin-stimulated cAMP accumulation [bib_ref] Adenylylcyclase supersensitization in mu-opioid receptor-transfected Chinese hamster ovary cells following chronic opioid..., Avidor-Reiss [/bib_ref]. When HEK 293 cells expressing MOR and MOR? ORL1 were exposed to morphine or methadone for 4 h, addition of naloxone caused AC superactivation, as revealed by the overshoot (up to 400%) of cAMP accumulation [fig_ref] Figure 6: Naloxone precipitation on the effects of chronic exposures to morphine, methadone, buprenorphine,... [/fig_ref]. However, no AC superactivation was observed in MOR-expressing cells chronically exposed to buprenorphine [fig_ref] Figure 6: Naloxone precipitation on the effects of chronic exposures to morphine, methadone, buprenorphine,... [/fig_ref] or Ro 64-6198 (except at [ 0.3 lM) [fig_ref] Figure 6: Naloxone precipitation on the effects of chronic exposures to morphine, methadone, buprenorphine,... [/fig_ref] after naloxone challenge. Interestingly, AC superactivation did occur in ORL1-and MOR?ORL1-expressing cells chronically exposed to buprenorphine, and ORL1-expressing cells even exhibited larger magnitude of AC superactivation than MOR? ORL1-expressing cells, as shown in the higher slope value in linear regression analysis [fig_ref] Figure 6: Naloxone precipitation on the effects of chronic exposures to morphine, methadone, buprenorphine,... [/fig_ref]. Chronic treatment with Ro 64-6198 also responded to naloxone ''precipitation'' and resulted in AC superactivation in cells expressing ORL1, both individually and simultaneously, with intriguingly bell-shaped concentration-response curves and higher efficacy in cells coexpressing MOR and ORL1 [fig_ref] Figure 6: Naloxone precipitation on the effects of chronic exposures to morphine, methadone, buprenorphine,... [/fig_ref].
## Ac superactivation after agonist removal following long-term opioid exposure
Since naloxone does not act on ORL1 [fig_ref] Figure 4: Blockade by naloxone of the morphine, methadone, and buprenorphine inhibition on forskolin-stimulated... [/fig_ref] yet cells stably expressing ORL1 showed AC superactivation upon addition of naloxone [fig_ref] Figure 6: Naloxone precipitation on the effects of chronic exposures to morphine, methadone, buprenorphine,... [/fig_ref] , we investigated whether removal of agonist is sufficient to reveal AC superactivation after prolonged agonist exposure. Indeed, both buprenorphine and Ro 64-6198 resulted in AC superactivation (sixfold and 2.5-fold increases for buprenorphine and Ro 64-6198, respectively) after agonist removal in cells expressing solely ORL1 [fig_ref] Figure 7 ': 'Natural withdrawal'' on the effects of chronic exposures to morphine, methadone, buprenorphine,... [/fig_ref]. Not surprisingly, morphine and methadone were still able to stimulate AC superactivation, yet to a lesser extent, in cells expressing MOR and MOR?ORL1 in the absence of naloxone challenge [fig_ref] Figure 7 ': 'Natural withdrawal'' on the effects of chronic exposures to morphine, methadone, buprenorphine,... [/fig_ref]. Hence, we demonstrated that ''natural withdrawal'' by washout [fig_ref] Figure 7 ': 'Natural withdrawal'' on the effects of chronic exposures to morphine, methadone, buprenorphine,... [/fig_ref] and ''precipitated withdrawal'' by naloxone [fig_ref] Figure 6: Naloxone precipitation on the effects of chronic exposures to morphine, methadone, buprenorphine,... [/fig_ref] have similar effects on cAMP accumulation; additionally, chronic exposure to buprenorphine and Ro 64-6198 could also contribute to the ''withdrawal syndrome'' (i.e. AC
# Discussion
## Heterodimerization of mor-orl1
Heterodimerization has been reported for MOR and various receptors, such as the d opioid receptor [bib_ref] Heterodimerization of mu and delta opioid receptors: a role in opiate synergy, Gomes [/bib_ref] , ORL1 receptor [bib_ref] Heterodimerization of opioid receptor-like 1 and mu-opioid receptors impairs the potency of..., Wang [/bib_ref] , sst 2A somatostatin receptor [bib_ref] Heterodimerization of somatostatin and opioid receptors cross-modulates phosphorylation, internalization, and desensitization, Pfeiffer [/bib_ref] , substance P receptor [bib_ref] Heterodimerization of substance P and mu-opioid receptors regulates receptor trafficking and resensitization, Pfeiffer [/bib_ref] , cannabinoid CB 1 receptor [bib_ref] mu-Opioid receptor forms a functional heterodimer with cannabinoid CB1 receptor: electrophysiological and..., Hojo [/bib_ref] , and metabotropic glutamate receptor 5 [bib_ref] Allosteric modulation of metabotropic glutamate receptor 5 affects phosphorylation, internalization, and desensitization..., Schroder [/bib_ref]. Receptor heterodimerization usually leads to alterations in MOR phosphorylation, internalization, desensitization, MAPK activation, and coupling to voltagedependent Ca 2? channels [bib_ref] Delta receptors are required for full inhibitory coupling of mu-receptors to voltagedependent..., Walwyn [/bib_ref]. We demonstrated the colocalization of coexpressed human MOR and ORL1 receptors in HEK 293 cells . Although the colocalization rate did not reach 100%, indicating the presence of homomers in the coexpressed condition, the high colocalization rate suggests the formation of heterodimerized MOR-ORL1 as recently reported in tsA-201 cell [bib_ref] Heterodimerization of ORL1 and opioid receptors and its consequences for N-type calcium..., Evans [/bib_ref]. We attempted to co-immunoprecipitate the coexpressed human MOR and ORL1 using anti-HA and anti-myc antibodies, but were unable to detect the direct association of MOR and ORL1 by pulling down the receptors together. Hence, human MOR and ORL1 may indeed heterodimerize but cannot be co-immunoprecipitated, or co-immunoprecipitation with the anti-N-terminal epitope tags is unable to detect the association of human MOR and ORL1. Our saturation binding assay using [ 3 H]-nociceptin [fig_ref] Table 1 B: max [/fig_ref] -which showed that co-expressing ORL1 with MOR reduced the number of nociceptin-binding sites (lower B max ) yet significantly increased the nociceptin affinity of the receptor (lower K D )also implies the novel properties of coexpressed human MOR-ORL1 receptors. This interesting phenomenon was not seen in HEK 293 cells co-expressing rat MOR and ORL1 [bib_ref] Heterodimerization of opioid receptor-like 1 and mu-opioid receptors impairs the potency of..., Wang [/bib_ref]. Since the expression levels of myc-tagged ORL1 are not drastically different in cells expressing ORL1 alone and MOR?ORL1 as revealed by immunoblotting [fig_ref] Figure 1: MOR and ORL1 expressed in HEK 293 cells are N-linked glycoproteins [/fig_ref] , the reduction of B max might be due to the decreased number of ORL1 transported from ER-Golgi to the plasma membrane [bib_ref] Heterodimerization of mu-and delta-opioid receptors occurs at the cell surface only and..., Law [/bib_ref]. Another possibility is that coexpressed human MOR and ORL1, perhaps forming heteromers, adopted a different conformation of the nociceptin-binding site from ORL1, rendering the binding affinity higher than ORL1 homomers [bib_ref] Allosteric modulation of metabotropic glutamate receptor 5 affects phosphorylation, internalization, and desensitization..., Schroder [/bib_ref] [bib_ref] A day in the life of a G protein-coupled receptor: the contribution..., Milligan [/bib_ref]. A potential caveat of the overexpression system is that the interaction between two overexpressed receptors might be mass action-induced and does not exist in endogenous system where the receptor abundance is much lower. Therefore, our in vitro cell model is a simplified system to address the possible heterodimerization between MOR and ORL1, and may not truthfully reflect the native condition of these two opioid receptors in the in vivo system.
## Acute activation of mor and orl1 inhibits ac activity
Acute agonist exposure inhibits forskolin-induced accumulation of cAMP in recombinant HEK 293 cells expressing cloned MOR [bib_ref] Activation of type II adenylyl cyclase by the cloned mu-opioid receptor: coupling..., Chan [/bib_ref] or ORL1 [bib_ref] Identification of dynorphins as endogenous ligands for an opioid receptor-like orphan receptor, Zhang [/bib_ref] ; this effect is mediated by inhibition of AC activity upon opioid receptor activation [bib_ref] Opioid receptor-coupled second messenger systems, Childers [/bib_ref] [bib_ref] Nociceptin/orphanin FQ and the opioid receptor-like ORL1 receptor, Meunier [/bib_ref]. In our cell model, acute treatment with two MOR agonists, morphine and methadone, specifically repressed the AC activity in cells expressing recombinant MOR. The ORL1 agonist, Ro 64-6198, acutely inhibited AC in cells expressing recombinant ORL1 but not MOR alone. Buprenorphine, which acted as a partial agonist at MOR and as a full agonist at ORL1, exhibited an intermediate potency (pIC 50 ) and efficacy (E max ) in cells coexpressing MOR and ORL1 in comparison to MOR or ORL1 alone. This intermediate response might result from the heterodimerization of MOR and ORL1, or the simultaneous regulation of common secondary messengers by MOR and ORL1. Changes in the cAMP system in the locus coeruleus (LC) play a role in mediating acute opioid action and underlying the development of opioid dependence and withdrawal [bib_ref] Acute and chronic opiate-regulation of adenylate cyclase in brain: specific effects in..., Duman [/bib_ref]. Morphine acutely inhibited AC in vitro in the LC, dorsal raphe, frontal cortex, and neostriatum; and the inhibition was blocked by naloxone. This response is mediated by a pertussis toxin-sensitive G-protein (i.e. G i/o ) [bib_ref] Acute and chronic opiate-regulation of adenylate cyclase in brain: specific effects in..., Duman [/bib_ref]. We successfully replicated the phenomenon in the in ; and the inhibition was also blocked by naloxone . Acute exposure to agonists of G i/o -coupled receptors-such as MOR, m 4 (muscarinic type 4), D 2 (dopaminergic type 2), and CB 1 (cannabinoid type 1) receptors-inhibits the activity of AC types I, V, VI, and VIII, while stimulating the activity of AC types II, IV, and VII [bib_ref] Chronic opioid treatment induces adenylyl cyclase V superactivation. Involvement of Gbetagamma, Avidor-Reiss [/bib_ref] [bib_ref] G protein regulation of adenylate cyclase, Simonds [/bib_ref]. Furthermore, acute activation of G i/o -coupled receptors leads to inhibition of AC-VIII-A and -B, but not of the AC-VIII-C splice variant in COS-7 cells transfected with MOR [bib_ref] Regulation of adenylate cyclase type VIII splice variants by acute and chronic..., Steiner [/bib_ref]. It would be interesting to investigate whether the cells coexpressing MOR and ORL1 utilize the same mechanism to regulate AC activity as well.
## Ac superactivation after chronic agonist treatment
Opioid-induced AC superactivation has been widely used as an indicator of cellular dependence [bib_ref] Drug addiction: a model for the molecular basis of neural plasticity, Nestler [/bib_ref]. Cell lines expressing either MOR or ORL1 have been successfully utilized to investigate the overshoot of AC activity following chronic treatment of agonists [bib_ref] Adenylylcyclase supersensitization in mu-opioid receptor-transfected Chinese hamster ovary cells following chronic opioid..., Avidor-Reiss [/bib_ref] [bib_ref] Chronic activation of ORL1 receptor induces supersensitization of adenylyl cyclase, Chan [/bib_ref]. The present study further used cells co-expressing MOR and ORL1 and observed unique responses of the coexpressed MOR?ORL1 receptors, which suggests that coexpressed MOR?ORL1 receptors perhaps represent a distinct population of the opioid receptors: ones that bear pharmacological profiles different from either MOR or ORL1 alone. Our results demonstrated that the AC superactivation could start as early as 30 min after morphine or methadone treatment [fig_ref] Figure 5: Effects of naloxone on acute exposures to morphine, methadone, buprenorphine, or Ro... [/fig_ref] ; and 4 h incubation with the MOR agonist resulted in profound AC superactivation either in the presence of naloxone challenge [fig_ref] Figure 6: Naloxone precipitation on the effects of chronic exposures to morphine, methadone, buprenorphine,... [/fig_ref] or simply after agonist washout [fig_ref] Figure 7 ': 'Natural withdrawal'' on the effects of chronic exposures to morphine, methadone, buprenorphine,... [/fig_ref]. The LC nucleus possesses a high density of opioid receptors, particularly MOR [bib_ref] Neuroanatomical patterns of the mu, delta, and kappa opioid receptors of rat..., Tempel [/bib_ref] and ORL1 [bib_ref] cDNA cloning and regional distribution of a novel member of the opioid..., Fukuda [/bib_ref]. Chronic morphine administration elevates levels of AC-I and AC-VIII in the LC [bib_ref] CREB (cAMP response element-binding protein) in the locus coeruleus: biochemical, physiological, and..., Lane-Ladd [/bib_ref] and knockout of either calmodulindependent isoform attenuates the ability of chronic morphine exposure to increase LC neuronal excitability and behavioral features of opioid withdrawal [bib_ref] Distinct roles of adenylyl cyclases 1 and 8 in opiate dependence: behavioral,..., Zachariou [/bib_ref]. Blockade of cAMP response-element binding protein (CREB) in the LC prevents the opioid-induced AC-VIII up-regulation [bib_ref] CREB (cAMP response element-binding protein) in the locus coeruleus: biochemical, physiological, and..., Lane-Ladd [/bib_ref]. Such blockade also diminishes the ability of chronic morphine treatment to increase LC neuronal excitability and to induce dependence and withdrawal [bib_ref] CREB (cAMP response element-binding protein) in the locus coeruleus: biochemical, physiological, and..., Lane-Ladd [/bib_ref] [bib_ref] Role of cAMP response element-binding protein in the rat locus ceruleus: regulation..., Han [/bib_ref]. An ex vivo LC slice culture system combined with viral-mediated gene transfer and genetic mutant mice provides direct evidence supporting that prolonged morphine exposure induces homeostatic adaptations intrinsic to LC neurons, involving up-regulation of cAMP-CREB pathway, which enhances LC neuronal excitability [bib_ref] Essential role of the cAMP-cAMP response-element binding protein pathway in opiate-induced homeostatic..., Cao [/bib_ref]. Given the colocalization of MOR and ORL1 in the LC and our HEK 293 cells , the present in vitro cell model expressing MOR?ORL1 offers the opportunity to dissect, in a simplified and more accessible manner, the relationship among MOR, ORL1, AC-I, AC-VIII, cAMP-CREB, and cell excitability during chronic exposure to morphine, methadone, and buprenorphine.
## Ac superactivation induced by buprenorphine and ro 64-6198 at orl1
Although buprenorphine did not elicit AC superactivation at the 30 min time point [fig_ref] Figure 5: Effects of naloxone on acute exposures to morphine, methadone, buprenorphine, or Ro... [/fig_ref] , 4 h buprenorphine exposure induced prominent forskolin-induced cAMP accumulation (*3.5 fold) in MOR?ORL1-expressing cells and drastically high cAMP accumulation (*sixfold) in ORL1-expressing cells [fig_ref] Figure 6: Naloxone precipitation on the effects of chronic exposures to morphine, methadone, buprenorphine,... [/fig_ref]. Another striking observation is the bell-shaped curve of Ro 64-6198-induced AC superactivation [fig_ref] Figure 6: Naloxone precipitation on the effects of chronic exposures to morphine, methadone, buprenorphine,... [/fig_ref]. This biphasic response might reflect an intrinsic ORL1 property that receptor activation by higher concentration (above 0.1 lM) of agonists would generate a reduction, not an enhancement, of AC superactivation, thus contributing to its role in modulating opioid antinociception [bib_ref] Orphanin FQ is a functional anti-opioid peptide, Mogil [/bib_ref] and blocking the rewarding effects of several abused drugs, including morphine [bib_ref] The effect of a systemically active ORL-1 agonist, Ro 64-6198, on the..., Shoblock [/bib_ref] , cocaine [bib_ref] The endogenous OFQ/N/ORL-1 receptor system regulates the rewarding effects of acute cocaine, Marquez [/bib_ref] , and amphetamine [bib_ref] Nociceptin inhibits acquisition of amphetamine-induced place preference and sensitization to stereotypy in..., Kotlinska [/bib_ref]. The difference between buprenorphine and Ro 64-6198 remains to be elucidated if it is due to the difference between the partial agonist (buprenorphine) and full agonist (Ro 64-6198) for ORL1.
## Clinical implications
When the responses to long-term treatment of methadone and buprenorphine are compared, cells expressing both MOR and ORL1 display matching concentration-response curves [fig_ref] Figure 6: Naloxone precipitation on the effects of chronic exposures to morphine, methadone, buprenorphine,... [/fig_ref] and c, filled squares). In contrast, cells expressing MOR alone concentrationdependently responded to methadone, not to buprenorphine [fig_ref] Figure 6: Naloxone precipitation on the effects of chronic exposures to morphine, methadone, buprenorphine,... [/fig_ref] and c, open circles), whereas ORL1-expressing cells exhibited concentration-dependent response to buprenorphine but not to methadone [fig_ref] Figure 6: Naloxone precipitation on the effects of chronic exposures to morphine, methadone, buprenorphine,... [/fig_ref]. This implies that chronic methadone and buprenorphine treatments induce differential effects on cells expressing either MOR or ORL1 alone, yet lead to similar cellular responses in the context of coexpressed MOR and ORL1. Therefore, our cellular model could mimic the physiological responses of patients, expressing both MOR and ORL1, under maintenance therapy.
The temporal difference between spontaneous and precipitated withdrawal might be explained by our cell model, revealing that naloxone-precipitated withdrawal [fig_ref] Figure 6: Naloxone precipitation on the effects of chronic exposures to morphine, methadone, buprenorphine,... [/fig_ref] elicited more prominent AC superactivation than natural withdrawal [fig_ref] Figure 7 ': 'Natural withdrawal'' on the effects of chronic exposures to morphine, methadone, buprenorphine,... [/fig_ref] after chronic opioid exposure. Moreover, naloxone produces ''overshoot'' phenomena suggestive of early acute physical dependence 6-24 h after a single dose of a MOR agonist [bib_ref] Acute opioid physical dependence in postaddict humans: naloxone dose effects after brief..., Heishman [/bib_ref]. This acute physical dependence is reflected in the elevated AC activity precipitated by naloxone in MOR-expressing cells after 30 min exposure to morphine and methadone [fig_ref] Figure 5: Effects of naloxone on acute exposures to morphine, methadone, buprenorphine, or Ro... [/fig_ref]. Thus, our study provides a clue to the cellular mechanism of the opioid withdrawal precipitated by naloxone.
In summary, our study suggests that methadone and buprenorphine exert different adaptive changes on the secondary messengers. While methadone and morphine bear almost indistinguishable pharmacological profiles in AC superactivation after chronic treatment, buprenorphine carries a dissimilar pharmacological portrait, presumably originating from its agonistic function to ORL1. The in vitro cell model of coexpressed human MOR?ORL1 receptors provides insight into cross-talk between opioid receptors following prolonged opioid exposure, and could provide a new approach to examining novel drugs prior to their clinical use and an uncomplicated tool for investigating related signaling pathways of opioids at the cellular level.
[fig] Figure 1: MOR and ORL1 expressed in HEK 293 cells are N-linked glycoproteins. Cell lysates were prepared by extracting monolayers of HEK 293 cells expressing HA-tagged MOR and/or myc-tagged ORL1 receptors in lysis buffer for 1 h on ice. Cellular debris was pelleted by centrifugation; and the supernatants were treated with or without N-glycosidase F (protease-free, 50 units/mg of membrane protein) at 37°C for 3 h, then resolved using 10% SDS-PAGE. HA-or myctagged receptors were assayed by immunoblotting using the monoclonal anti-HA (Left panel) or anti-myc (Right panel) antibody. The mobilities of molecular mass standards (in kDa) are indicated to the left. -, glycosylated (untreated) receptors; ?, deglycosylated (treated with N-glycosidase F) receptors [/fig]
[fig] Figure 2, Figure 3: Representative confocal images from cells expressing MOR (Upper left panel), ORL1 (Upper right panel), and MOR?ORL1 (Lower panels). HA-tagged MOR was detected with anti-HA mouse monoclonal antibody and visualized by Alexa Fluor 488 goat antimouse antibody (green); myc-tagged ORL1 was detected with anti-ORL1 rabbit polyclonal antibody and visualized by Alexa Fluor 647 goat anti-rabbit antibody (red); the colocalization of MOR and ORL1 is depicted in yellow in the merged picture. Scale bars are equal Effects of acute exposures to morphine, methadone, buprenorphine, or Ro 64-6198 on forskolin-stimulated cAMP accumulation in HEK 293 cells expressing MOR and/or ORL1. HEK 293 cells expressing MOR (open circle), ORL1 (open triangle) or both MOR and ORL1 (filled square) were treated with morphine (a), methadone (b), buprenorphine (c), or Ro 64-6198 (d) for 30 min at room temperature in the presence of 10 lM forskolin prior to HTRF cAMP assays. Each point represents the mean ± SE value of four experiments performed in duplicate using different batches of cells. 100% defines forskolin-stimulated cAMP accumulation in cells not treated with aforementioned drugs. Asterisks in (c) indicate the very significant difference (P \ 0.01) between curves of ORL1-expressing cells and MOR?ORL1-expressing cells according to paired t test (two-tailed) analysis [/fig]
[fig] Figure 4: Blockade by naloxone of the morphine, methadone, and buprenorphine inhibition on forskolin-stimulated cAMP accumulation in HEK 293 cells expressing MOR and/or ORL1. HEK 293 cells expressing MOR (a), MOR?ORL1 (b), or ORL1 (c) were treated with 0.1 lM morphine (morph), methadone (methad), buprenorphine (bupren), or Ro 64-6198 (Ro) in the presence or absence of 1 lM naloxone (nalox) for 30 min at room temperature with 10 lM forskolin prior to HTRF cAMP assays. Each point represents the mean ± SE value of four experiments performed in duplicate using different batches of cells. 100% defines forskolin-stimulated cAMP accumulation in cells not treated with aforementioned drugs. The significance of differences between without and with naloxone treatment were determined by one-way ANOVA followed by Bonferroni's test, *P \ 0.05, ***P \ 0in cells expressing ORL1 in our in vitro cell model. [/fig]
[fig] Figure 5: Effects of naloxone on acute exposures to morphine, methadone, buprenorphine, or Ro 64-6198 on forskolin-induced cAMP accumulation in HEK 293 cells expressing MOR and/or ORL1. After exposure to morphine (a), methadone (b), buprenorphine (c), or Ro 64-6198 (d) for 30 min at room temperature, the incubation media were subsequently removed, and HEK 293 cells expressing MOR (open circle), ORL1 (open triangle) or both MOR and ORL1 (filled square) were treated with 1 lM naloxone accompanied by 10 lMforskolin immediately prior to HTRF cAMP assays. Each point represents the mean ± SE value of four experiments performed in duplicate using different batches of cells. 100% defines forskolinstimulated cAMP accumulation in cells treated with none of the aforementioned drugs but naloxone. Asterisks indicate the significant differences (**P \ 0.01, ***P \ 0.001) between curves of MORexpressing cells and MOR?ORL1-expressing cells according to paired t test (two-tailed) analysis [/fig]
[fig] Figure 6: Naloxone precipitation on the effects of chronic exposures to morphine, methadone, buprenorphine, or Ro 64-6198 on forskolinstimulated cAMP accumulation in HEK 293 cells expressing MOR and/or ORL1. HEK 293 cells expressing MOR (open circle), ORL1 (open triangle) or both MOR and ORL1 (filled square) were treated with morphine (a), methadone (b), buprenorphine (c), or Ro 64-6198 (d) for 4 h at 37°C. The incubation medium was subsequently replaced by 1 lM naloxone and 10 lM forskolin in compound buffer prior to HTRF cAMP assays. Each point represents the mean ± SE value of four experiments performed in duplicate using different batches of cells. 100% defines forskolin-stimulated cAMP accumulation in cells not treated with aforementioned drugs. Asterisks indicate the extremely significant differences (P \ 0.001) between curves of cells expressing MOR alone (open circle) and both MOR and ORL1 (filled square). Pound signs indicate the extremely significant difference (P \ 0.001) between curves of cells expressing only ORL1 (open triangle) and both MOR and ORL1 (filled square) according to paired t test (two-tailed) analysis. Dashed and dotted lines represents the linear regression fitted to the 6 data points at the high concentration end of cells expressing ORL1 (open triangle; slope = 186.4 ± 30.82) and both MOR and ORL1 (filled square; slope = 97.80 ± 17.31), respectively. Arrows indicate the significant difference (P \ 0.05) between the slopes of the model: morphine acutely inhibited AC in HEK 293 cells overexpressing MOR and MOR?ORL1 (Fig [/fig]
[fig] Figure 7 ': 'Natural withdrawal'' on the effects of chronic exposures to morphine, methadone, buprenorphine, or Ro 64-6198 on forskolinstimulated cAMP accumulation in HEK 293 cells expressing ORL1 and/or MOR. HEK 293 cells expressing MOR (open circle), ORL1 (open triangle) or both MOR and ORL1 (filled square) were treated with morphine (a), methadone (b), buprenorphine (c), or Ro 64-6198 (d) for 4 h at 37°C. The incubation medium was subsequently replaced by 10 lM forskolin in compound buffer prior to HTRF cAMP assays. Each point represents the mean ± SE value of four experiments performed in duplicate using different batches of cells. 100% defines forskolin-stimulated cAMP accumulation in cells not treated with aforementioned drugs. Dashed and dotted lines represents the linear regression fitted to the six data points at the high concentration end of cells expressing ORL1 (open triangle; slope = 235.8 ± 32.62) and both MOR and ORL1 (filled square; slope = 101.9 ± 13.89), respectively. Arrows indicate the extremely significant difference ( P \ 0.001) between the slopes of the linear regression. Pound signs indicate the significant difference (P \ 0.05) between curves of cells expressing only ORL1 (open triangle) and both MOR and ORL1 (filled square) according to paired t test (twotailed) analysis [/fig]
[table] Table 1 B: max (pmol/mg protein) and K D (nM) values of l-opioid receptors (MOR) and opioid receptor-like 1 receptors (ORL1) expressed in HEK 293 cells [/table]
|
Follicle Stimulating Hormone is an accurate predictor of azoospermia in childhood cancer survivors
This is the Supplementary Information file, listing the full-text excluded articles together with reasons for exclusion.WHOFull texts excluded for non-specification of WHO guidelines for assessment of azoospermia: [[BenArush2000] M W Ben Arush, I Solt, A Lightman, S Linn, and A Kuten. Male gonadal function in survivors of childhood Hodgkin and non-Hodgkin lymphoma.Abstract: The aim of this study was to investigate the impact of therapy on long-term gonadal function of young people cured of childhood lymphomas and to assess whether a prepubertal state during the treatment protects the gonads from chemotherapy and/or radiotherapy late effects. Clinical evaluation, semen analysis, and endocrine status were studied in 20 survivors of childhood lymphomas. Five patients received Inverted Y radiotherapy, 2320 cGy (1550-4000); all 20 received chemotherapy as follows: MOPP/ABVD protocol, 9 patients; COMP protocol, 5 patients; MOPP protocol, 3 patients; other protocols, 3 patients. Semen analysis results were as follows: normal values, 4/20 patients; oligospermia, 8/20 patients; azoospermia, 8/20 patients; FSH above normal level, 10/20 patients; 4/5 who received Inverted Y irradiation were azoospermic and 1 was severely oligospermic. Treatment damage to the testis involves tubular germinal elements. Radiotherapy and chemotherapy combinations that included nitrogen mustard or cyclophosphamide were associated with high rates of oligospermia and azoospermia. MOPP/ABVD combination did not have a significant better outcome of sperm counts compared to MOPP alone. Age at chemotherapy did not correlate with the sperm count; hence a prepubertal state did not protect the gonad from the late effects of treatment.[Bohlen2001] D Bohlen, F C Burkhard, R Mills, R W Sonntag, and U E Studer. Fertility and sexual function following orchiectomy and 2 cycles of chemotherapy for stage I high risk nonseminomatous germ cell cancer. Journal of Urology, 165(2):441-444, 2001.Abstract: PURPOSE: We investigate fertility and sexual function in patients following orchiectomy and adjuvant cisplatin based chemother-
apy for high risk, stage I nonseminomatous germ cell tumor of the testis. MATERIALS AND METH-ODS: Between 1985 and 1994, 59 patients with stage I nonseminomatous germ cell tumor and poor prognostic factors were treated with 2 cycles of cisplatin, vinblastine and bleomycin, or bleomycin, etoposide and cisplatin after orchiectomy. At least 32 months following treatment all patients were contacted and asked to complete a questionnaire regarding fertility and sexual activity, and to volunteer for a semen and hormonal analysis. RESULTS: Of the 59 patients 49 (83%) completed the questionnaire. Before chemotherapy 18 (37%) patients had fathered children, 6 (12%) were involuntarily childless and none had a major sexual dysfunction. After treatment 11 (22%) patients fathered children, and 5 (10%) were involuntarily childless, with 4 involuntarily childless before chemotherapy. There were no significant alterations in sexual function. Semen analysis in 27 patients was normal in 23, and revealed mild oligospermia in 2 and azoospermia in 2. In 18 patients with hormone analysis median values for luteinizing hormone and free testosterone were normal but median value for follicle-stimulating hormone was slightly increased. CONCLUSIONS: Two cycles of cisplatin based adjuvant chemotherapy do not seem to affect adversely fertility or sexual activity. Abstract: BACKGROUND: The objective of this study was to assess markers of spermatogenesis in long-term survivors of testicular cancer (TC) according to treatment, and to explore correlations between the markers and associations with achieved paternity following TC treatment. METHODS: In 1191 TC survivors diagnosed between 1980 and 1994, serum-follicle stimulating hormone (s-FSH; n=1191), s-inhibin B (n=441), and sperm counts (millions per ml; n=342) were analysed in a national Paternity was assessed by a questionnaire. RESULTS: At median 11 years follow-up, 44% had oligo-(<15 millions per ml; 29%) or azoospermia (15%). Sperm counts and s-inhibin B were significantly lower and s-FSH was higher after chemotherapy, but not after radiotherapy (RT), when compared with surgery only. All measures were significantly more abnormal following high doses of chemotherapy (cisplatin (Cis)>850 mg, absolute cumulative dose) compared with lower doses (Cis < 850 mg). Sperm counts were moderately correlated with s-FSH (-0.500), s-inhibin B (0.455), and s-inhibin B : FSH ratio (-0.524; all P<0.001).
All markers differed significantly between those who had achieved post-treatment fatherhood and those with unsuccessful attempts. Abstract: PURPOSE: Gonadal functions were evaluated in 26 male patients with Hodgkin's disease (HD), who were in continuous unmaintained remission following combination chemotherapy consisting of COPP/MOPP. MA-TERIALS AND METHODS: These patients had received chemotherapy during the prepubertal phase. The median duration after termination of chemotherapy was 72 months. RESULTS: Semen analysis of 18 patients showed azoospermia. Hormonal analysis showed elevated mean levels follicle-stimulating hormone (FSH) and inhibin as compared to age-matched controls, whereas luteinizing hormone levels were only marginally elevated. CONCLUSIONS: These results suggest that COPP/MOPP causes severe damage to germinal epithelium even when given during prepubertal age. Sertoli cells, which are responsible for secretion of inhibin, are resistant to these cytotoxic agents. Our data emphasize the lack of gross dysfunction of Leydig cells. It is possible that an alternative chemotherapy protocol (ABVD) may be used in young patients to minimize the gonadal damage.
[Fossa1980] S D Fossa, O Klepp, A Aakvaag, and K Molne. Testicular function after combined chemotherapy for metastatic testicular cancer. International Journal of Andrology, 3(1):59-65, 1980.
# Abstract:
In 10 patients cured for metastatic testicular cancer by combination chemotherapy serum hormone levels and serum agglutinating antibodies were analysed 12 to 35 months after discontinuation of the treatment. Together with these examinations sperm analysis was done. All patients had increased levels of follicle stimulating hormone (FSH). Serum testosterone was usually in the lower part of the normal range (below 20 nmol/l). In eight patients the serum agglutinating antibodies were normal, while two patients had increased levels. In all patients azoospermia was observed, indicating long-lasting infertility in patients testicular cancer treated by combination chemotherapy. The possible importance of cryopreservation prior to start of chemotherapy is discussed.
[Heikens1996] J Heikens, H Behrendt, R Adriaanse, and A Berghout. Irreversible gonadal damage in male survivors of pediatric Hodgkin's disease. Cancer, 78(9):2020-2024, 1996.
Abstract: BACKGROUND: Gonadal damage in adult patients after chemotherapy for Hodgkin's disease is well documented, but data of patients treated before adulthood are scarce. METHODS: Gonadal and hormonal function were studied in 19 male long term survivors of Hodgkin's disease who were treated with mechlorethamine, vincristine, procarbazine, and prednisone (MOPP chemotherapy) before (n = 15) or during puberty (n = 4). The studies were performed a median of 10 years after treatment and repeated in the majority of the patients at the time of yearly visits. RESULTS: Germ cell damage was present in all patients. Semen analysis revealed azoospermia in 12 patients and oligospermia in 6; no recovery of spermatogenesis was seen at followup. Testicular size was small in all but one patient. Follicle-stimulating hormone levels were elevated (mean, 14.4 +/-7.8 U/l) and increased over time (mean, 21.1 +/-10.5 U/l, P < 0.001).
In seven patients, luteinizing hormone (LH) was elevated, indicating Leydig cell dysfunction; also in four of those patients, plasma testosterone was decreased. In three other patients, the response of LH to gonadotropin-releasing hormone was exaggerated with a normal basal LH and testosterone. Comparing testicular function of prepubescent versus pubescent state at time of treatment appears to show a trend for improved outcome in the younger patients. CON-CLUSIONS: Gonadal function of long term survivors of pediatric Hodgkin's disease treated with MOPP chemotherapy is severely impaired permanently.
[Humpl1999] T Humpl, P Schramm, and P Gutjahr. Male fertility in long-term survivors of childhood ALL. Archives of Andrology, 43(2):123-129, 1999.
Abstract: A study of fertility was conducted in postpubertal male patients who had been treated for acute lymphoblastic leukemia (ALL) during childhood or adolescence between 1970 and 1980. Thirteen men (age 18 to 35 years) participated on a volunteer basis. Their age at diagnosis was between 2 and 15 years. Therapy followed the protocol "Memphis VII (Pinkel)." Interview, physical examination, andrological studies (ejaculate), and hormone status (luteinizing hormone, follicle-stimulating hormone, and testosterone) were performed at least 5 years after completion of therapy. No normozoospermia was achieved; 10 patients were identified with asthenozoospermia and 3 patients with azoospermia. With respect to these data, patients treated for ALL between 1970 and 1980 have more significantly impaired spermatogenesis than expected.
[Ishikawa2004] T Ishikawa, S Kamidono, and M Fujisawa. Fertility after high-dose chemotherapy for testicular cancer. Urology, 63(1): .
Abstract: OBJECTIVES: To describe long-term gonadal function after high-dose chemotherapy (HDC). HDC for testicular cancer was recently developed. The evaluation of testicular function after chemotherapy for testicular cancer is an important part of overall care, especially in young patients. METHODS: Between 1994 and 2001, 27 patients underwent HDC (1250 mg/m2 carboplatin, 1500 mg/m2 etoposide, and 7.5 g/m2 ifosfamide) at Kobe University Hospital. Information on gonadal function during follow-up was available for 10 of these patients. The mean patient age +/-SD at treatment was 32.2 +/-8.4 years. The relationships among age at treatment, semen analysis, serum hormone levels (follicle-stimulating hormone, luteinizing hormone, testosterone, prolactin, and estradiol), cumulative dose of cisplatin and carboplatin, and length of follow-up were determined. RESULTS: Spermatogenesis recovered after cessation of HDC in 5 of 10 patients. Semen analysis in these patients showed the mean sperm concentration and motility at 42.4 +/-10.4 million/mL and 67.2% +/-17.0%, respectively. The patients were divided into azoospermic and nonazoospermic groups. The age of the nonazoospermic and azoospermic patients was 28.2 +/-8.7 and 36.2 +/-6.5 years, respectively. Follicle-stimulating hormone levels in the nonazoospermic group (11.7 +/-3.4 mIU/mL) were significantly lower than in the azoospermic group (32.8 +/-14.4 mIU/mL; P = 0.0472). No other statistically significant difference was observed in the other hormone levels or the cumulative dose of cisplatin and carboplatin between the azoospermic and nonazoospermic groups. CONCLUSIONS: Spermatogenesis recovers after HDC in some patients. Patients should be informed that they may or may not be fertile after HDC.
[Kader1991] H A Kader and A Y Rostom. Follicle stimulating hormone levels as a predictor of recovery of spermatogenesis fol-lowing cancer therapy. Clinical Oncology (Royal College of Radiologists), 3(1):37-40, 1991.
Abstract: Infertility, both temporary and permanent, is a well-recognized complication of certain cancer treatments. The main objective of this study was to determine whether recovery of fertility in male patients, could be predicted by monitoring changes in serum follicle stimulating hormone (FSH) levels. Twenty male patients participated in the study. Sperm counts and serum FSH levels were measured before, during and after treatment. Azoospermia was universal in all 20 patients during the first year, with significantly raised FSH in almost all the patients. Reduction of FSH levels during the second year was frequently followed by recovery of spermatogenesis. Patients in whom the FSH did not fall during the second year were highly unlikely to regain fertility.
[Kliesch1997] S Kliesch, M Bergmann, L Hertle, E Nieschlag, and H M Behre. Semen parameters and testicular pathology in men with testicular cancer and contralateral carcinoma in situ or bilateral testicular malignancies. Human Reproduction, 12 (12):2830-2835, 1997.
# Abstract:
We evaluated 14 patients with bilateral testicular tumour, one-sided tumour and contralateral carcinoma in situ (CIS) of the testis or testis tumour in single testis with respect to their fertility. We analysed semen parameters, serum hormones [follicle-stimulating hormone (FSH), luteinizing hormone (LH) and testosterone], testicular sonography, testicular volumes and testicular histology prior to further anti-cancer treatment. Ten out of 14 patients showed normal or reduced sperm concentrations, while 4/14 patients were azoospermic. Serum FSH levels showed a significant negative correlation with sperm concentrations in patients with testicular malignancies (r = -0.64, P = 0.025). Testicular volumes revealed a significant positive correlation with semen parameters in patients with testes that were affected by CIS (r = 0.733, P = 0.038). We conclude that even bilateral testicular cancer and/or CIS do not pre-clude fertility and, therefore, patients should be offered andrological investigation and therapy, including possibly surveillance strategy or the chance for cryopreservation of the semen prior to further treatment in order to preserve their chances for paternity.
[ Abstract: Testicular and ovarian functions were assessed in 33 patients with Hodgkin's disease 1 to 17 years after cessation of COPP chemotherapy with cyclophosphamide, vincristine, procarbazine, prednisone. Diagnostic procedures consisted of hormone measurements, interviews, and semen analyses. In women serum levels of follicle-stimulating hormone (FSH), luteinizing hormone (LH), 17 beta-estradiol, progesterone, prolactin, and in men FSH, LH, 17 betaestradiol, testosterone, and prolactin were determined. Semen analyses were performed in all men. Information concerning pregnancies, pregnancy outcome, future fertility wishes, sexual functions, menstrual pattern, and incidence of premature menopausal symptoms was ascertained by interview and questionnaire. Nineteen of 19 (100%) men showed elevated serum FSH levels between 715 and 1910 (median 1095) ng/ml and azoospermia, 1 to 11 years after therapy. Serum levels of testosterone were within normal limits in 18/19 (95%) of the men, and LH values were normal in all men. Permanent ovarian failure occurred in 8/14 (57%) women, causing infertility and premature menopausal symptoms. The incidence of ovarian failure in women over 24 years was 86% (6/7) versus 28% (2/7) in those under 24 years at the time of treatment. In women receiving estrogen replacement, incidence and severity of these symptoms were significantly reduced. Of 14 women 3 (21%) became pregnant and delivered 5 healthy children after treatment. Our results suggest irreversible sterility and normal Leydig cell function after COPP chemotherapy in all men. Drug-induced ovarian failure was age-related and caused premature menopausal symptoms, detracting from the quality of the patient's life. To reduce premature menopausal symptoms and to prevent adverse cardiovascular and metabolic late sequelae, hormonal replacement is indicated. Pregnancies ending in normal live births can be achieved after COPP chemotherapy in young women. In both men and women, serum FSH and LH levels proved to be feasible markers to determine degree and duration of endocrine and reproductive gonadal injury after chemotherapy.
[ Abstract: Only limited data is currently available on long-term gonadal toxicity and its impact on bone mineralization in men and women treated for Hodgkin's disease. The present study was therefore conducted to evaluate gonadal toxicity and bone loss in 49 patients with Hodgkin's disease 2-10 (median 5.37) years after chemotherapy. Most patients were treated with the COPP/ABVD regimen +/-irradiation according to the protocols of the German Hodgkin Study Group. Blood samples were tested for gonadotropins (FSH, LH), gonadal steroids, parathyroid hormone, osteocalcin, and calcitonin. Bone mineral density was measured using single-and dual-energy quantitative computed tomography as well as single-photon absorptiometry. FSH serum levels were significantly increased in 21/27 (80%) men demonstrating germ-cell aplasia. 13/15 (86%) men showed azoospermia after the COPP/ABVD regimen. In contrast, testosterone levels were within normal limits in all men tested, suggesting normal Leydig-cell function. 17/22 (77%) women exhibited increased FSH and LH levels, indicating premature ovarian failure. Women with therapy-induced ovarian failure had a significantly lower trabecular (98 +/-34) and cortical (292 +/-48 mg/cm3) spinal bone density than those with normal ovarian function. Men showed no evidence of bone loss after therapy. These data suggest severe gonadal toxicity in both men and women treated with the COPP/ABVD regimen. In female patients, druginduced ovarian failure has a significant impact on bone mineralization. Abstract: This study evaluated male gonadal function in long-term survivors of childhood cancer and assessed the suitability of offering sperm analysis to all those patients independently of the diagnosis and treatment received. A total of 43 survivors of acute lymphoblastic leukemia (21), acute myeloid leukemia (1), neuroblastoma (8), ganglioneuroblastoma (1), ganglioneuroma (2), Wilms' tumor (9), and mesoblastic nephroma (1) underwent sperm analysis at a mean age of 20.2 years, after a mean time off treatment of 13.6 years. Eight of the patients (19%) were azoospermic, 2 (5%) were severely oligo-asthenozoospermic, and only 16 (37%) were normozoospermic. A control group of healthy volunteers aged < or = 30 years included no azoospermic subjects, 7% severely oligo-asthenozoospermic, and 67% normozoospermic. Comparisons were also made with patients treated at our Human Reproductive Unit aged < or = 30 years (n = 373) whose percentages for the above parameters were 4, 9, and 42%, respectively. Cumulated cyclophosphamide dose and basal follicle-stimulating hormone (FSH) levels were identified as independent factors associated with azoospermia or severe oligo-asthenozoospermia. Azoospermic and severely oligo-asthenozoospermic survivors had significantly smaller mean testicular volume and higher basal FSH levels than the other survivors, but small testicles (sum of both testicular volume < or = 20 mL) and/or abnormally high basal FSH (> 10 mIU/mL) were present in only half of the azoospermic survivors. Male long-term survivors of childhood cancer constitute a highrisk subpopulation for altered sperm analysis. It seems justified to offer sperm analysis to all long-term survivors.
[Muller1996] H L Muller, M Klinkhammer-Schalke, B Seelbach-Gobel, A A Hartmann, and J Kuhl. Gonadal function of young adults after therapy of malignancies during childhood or adolescence. European Journal of Pediatrics, 155(9):763-769, 1996.
Abstract: UNLABELLED: As the survival rate of children with malignancies has increased over past decades, the follow up of adult long-term survivors (LTS) of childhood cancer should focus on late effects of disease and treatment. Gonadal function was therefore studied in 54 LTS (aged 17-29 years; 33 male, 21 female) 2-18 years after treatment for malignancies during childhood or adolescence. To analyse the sensitivity of different diagnostic methods, tests of endocrine function (n = 52), spermiograms (n = 14), gynaecological status (n = 20) and ultrasonography of the gonads (n = 53) were compared with the results of equivalent tests in 23 age-matched normal controls (12 male, 11 female). There were no differences between male and female LTS concerning age at diagnosis, gonadal dose of irradiation (XRT) and doses of applied chemotherapeutic agents. Whereas male LTS had elevated levels of luteinizing hormone (LH) and folliclestimulating hormone (FSH) before (P < 0.05; P < 0.001) and after (P < 0.01; P < 0.001) stimulation with gonadotropin releasing hormone, female LTS exhibited normal endocrine function. Accordingly, male patients exhibited lower testicular volumes than normal controls, as measured with a Prader orchidometer (P < 0.01) or by ultrasonography (P < 0.001). Gynaecological status and ultrasonography of the gonads were normal in female LTS and controls. Whereas all spermiograms of normal controls (n = 8) showed a normal sperm cell density (SCD), only 2 of 14 male LTS exhibited a normal SCD (P < 0.001). Azoospermic LTS (n = 9) had been treated more often with alkylating agents and had received higher (P < 0.05) gonadal doses of XRT. All male LTS with testicular volumes below the normal range (< 13 ml) and basal FSH levels above the normal range (> 10 IU/l) exhibited azoospermia, whereas LTS with normal values for testicular volume and basal FSH had a normal SCD. CON-CLUSION: A sex-specific susceptibility for gonadal damage after treatment for malignancies might be responsible, in part, for the impaired gonadal function of male LTS. Therapy with alkylating agents and/or high gonadal doses of XRT were important risk factors for azoospermia. A simple method to estimate potential fertility in individual LTS is to measure testicular volume, using a Prader orchidometer, and basal FSH serum levels.
[Ortin1990] T T Ortin, C A Shostak, and S S Donaldson. Gonadal status and reproductive function following treatment for Hodgkin's disease in childhood: the Stanford experience. International Journal of Radiation Oncology, .
# Abstract:
To ascertain the impact of therapy on gonadal function and reproductive outcome among children treated for Hodgkin's disease, we reviewed the experience at Stanford University Medical Center during the years 1965-1986. There were 240 children 15 years of age or younger, 92 girls and 148 boys; with median follow-up of 9 years, maximum follow-up was 26 years. Of this cohort, data on gonadal function were available on 20 boys, 5 of whom were considered prepubescent; they had no clinical evidence of sexual maturation and were less than 13 years of age. Evaluation of the boys included testicular biopsy, semen analyses and the ability to procreate. Serum gonadotropin hormone levels (FSH, LH) were studied in 11 boys who also had semen analyses. Sexual maturation was attained in all boys without the need for androgen replacement. Among the eight boys treated with radiation alone, four were able to father a child (3 following 40-45 Gy pelvic radiation dose, 1 without pelvic radiation) from 3-19 years following treatment. Three others who received 30-44 Gy pelvic radiation were oligospermic when tested at 10 to 15 years post-treatment. Semen analyses in 10 of 12 (83%) boys who had been treated with six cycles of MOPP with or without pelvic radiation revealed absolute azoospermia with no evidence of recovery as along as 11 years of follow-up. Following prolonged azoospermia, 2 of the 12 boys (17%) had recovery of fertility, with normalization of sperm count and/or ability to procreate at 12 and 15 years following treatment. There was no correlation with serum gonadotropin levels and sterility. Data on menstrual history, pregnancy and offspring were available in 86 (92%) of the girls. Seventyfive of the 86 girls (87%) have normal menstrual function. However, none of the females who underwent pelvic radiation without prior oophoropexy has maintained ovarian function. Both the prepubescent and postpubescent boys were affected by 6 cycles of MOPP whether or not pelvic radiation was administered. On the other hand, in girls similarly treated, ovarian injury was directly related to both the number of cycles of chemotherapy and the ovarian radiation dose. The chances of maintaining gonadal function following combined modality treatment are significantly greater among girls than boys. The progeny of patients treated for Hodgkin's disease appear normal and no excess fetal wastage has been noted.
[Pelliccione2011] F Pelliccione, V Verratti, A D'Angeli, A Micillo, C Doria, A Pezzella, G Iacutone, F Francavilla, C Di Giulio, and S Francavilla. Physical exercise at high altitude is associated with a testicular dysfunction leading to reduced sperm concentration but healthy sperm quality. Abstract: As a follow-up to the pilot study of semen quality of soldiers with various military assignments a larger, more complete study was conducted. Soldiers were recruited at Fort Hood, Texas. Thirty-three men were exposed to radar as part of their duty assignment in the Signal Corps, 57 men were involved with firing the 155 mm howitzer (potential lead exposure), and 103 soldiers had neither lead nor radar exposure and served as the comparison control group. Both serum and urinary follicle-stimulating hormone and luteinizing hormone and serum, salivary, and urine testosterone levels were determined in all men. A complete semen analysis was conducted on each soldier. For statistical analysis, the primary study variables were: sperm concentration, sperm/ejaculate, semen volume, percent normal morphology, percent motile, percent viable (both vital stain and hypoosmotic swelling), curvilinear velocity, straight-line velocity, linearity, sperm head length, width, area, and perimeter. Variables were adjusted for significant confounders (e.g., abstinence, sample age, race). No statistical differences (P < 0.05) were observed in any measurement. While these results are in agreement with two previous studies assessing soldiers firing the 155-mm howitzer, they contradict our previous report indicating that radar exposure caused a significant decrease in sperm numbers. A possible explanation is that the radar exposure in this study was that used in Signal Corps operations while the men in the previous study were using different radar as part of military intelligence operations. The data presented here in men firing the 155mm howitzer combined with the results from the previous studies confirms that there are no deficits in semen quality in these men. The contradiction between the results of the radar exposure studies indicates that more data are needed to evaluate the relationship of military radar and male reproductive health.
[Shamberger1981] R C Shamberger, R J Sherins, and S A Rosenberg. The effects of postoperative adjuvant chemotherapy and radiotherapy on testicular function in men undergoing treatment for soft tissue sarcoma. Cancer, 47 (10): .
Abstract: Testicular function was studied in 26 men with sarcoma who received adjuvant treatment with doxorubicin, cyclophosphamide, and high-dose methotrexate (with or without radiotherapy). Testicular size, sperm output, and serum FSH, LH and testosterone levels were assessed after treatment. Five of 17 men who received chemotherapy or chemotherapy with radiotherapy to the neck, arm, chest, or leg, had normal testicular function. Eight of the remaining 12 men who provided ejaculates were oligospermic or azoospermic and serum FSH was increased threefold and LH twofold; testosterone levels were normal. In the five men with normal testicular function, FSH was increased fourfold during therapy but returned to normal six to 21 months after treatment. In men less than 40 years old, the mean FSH was less than that of men over 40 years of age (P = to 0.05), suggesting that recovery from the injury was age-related. By contrast, all nine men who received chemotherapy plus radiotherapy to the abdomen or thigh had decreased testicular size, azoospermia, fourfold increase in FSH, and twofold increase in LH levels; but testosterone concentration was normal. These data increase in FSH, and reversible testicular injury occurs after treatment with doxorubicin, cyclophosphamide, and high-dose methotrexate; recovery is age-related. However, these agents in combination with use of adjuvant radiotherapy to the thigh or abdomen may produce permanent testicular injury even in young patients.
[ Abstract: BACKGROUND: Infertility is one of the most significant side-effects in long-term survivors of successfully treated Hodgkin's lymphoma (HL). PATIENTS AND METHODS: The fertility status was assessed in male HL patients enrolled into trials of the German Hodgkin Study Group from 1988 to 2003. RESULTS: In pre-treatment analysis (n = 202), 20% of patients had normozoospermia, 11% azoospermia and 69% had other dyspermia. In posttreatment analysis (n = 112), 64% of patients had azoospermia, 30% other dyspermia and 6% normozoospermia (P < 0.001). Azoospermia was observed in 90% of patients treated with chemotherapy alone, 67% of those treated with combined modality and 11% of those treated with radiotherapy alone (P < 0.001). Azoospermia was more frequent after 4x cyclophosphamide, vincristine, procarbazine, prednisone, doxorubicin, bleomycin, vinblastine, dacarbazine (COPP/ABVD) (91%), 8x bleomycin, etoposide, doxorubicin, cyclophosphamide, vincristine, procarbazine, prednisone (BEACOPP) baseline (93%) and 8x BEACOPP escalated (87%) compared with 2x COPP/ABVD (56%; P = 0.003).
There was a statistically significant difference in post-treatment follicle-stimulating hormone (FSH) levels between patients with azoospermia and those with preserved spermatogenesis (P = 0.001). CONCLUSIONS: Depending on the treatment received, male HL patients are at high risk of infertility after treatment. FSH might be used as surrogate parameter for male fertility in future studies. One hundred forty-five reproductive age couples without known risk factors for infertility and who had discontinued contraception to achieve pregnancy completed this component of this study. Each couple was followed for < or =12 menstrual cycles while they attempted to conceive. INTERVENTION(S): Semen quality measures for the first ejaculates were obtained at the start of the study along with a single blood sample. Levels of FSH, bioactive FSH, inhibin B, LH, and T were measured for each man. MAIN OUTCOME MEASURE(S): Semen analysis, FSH, inhibin B, LH, T, and clinical pregnancy. RESULTS: Significant positive relationships were observed between the two measures of FSH as well as between both of the FSH measures and LH. Follicle-stimulating hormone as measured by RIA was significantly negatively correlated with inhibin B. Inhibin B showed a marginally significant negative correlation with LH, and LH and T had a marginally significant positive correlation. Inhibin B increased significantly, and both measures of FSH activity showed significant decreases, with increasing levels in several semen quality measures. There was no significant relationship between
[fig] [: Brydoy2012] M Brydoy, S D Fossa, O Klepp, R M Bremnes, E A Wist, T Bjoro, T Wentzel-Larsen, O Dahl, and I I I study group Norwegian Urology Cancer Group. Sperm counts and endocrinological markers of spermatogenesis in long-term survivors of testicular cancer. British Journal of Cancer, 107(11):1833-1839, 2012. [/fig]
|
Air-Stable NaxTMO2 Cathodes for Sodium Storage
Sodium-ion batteries are considered to be the most promising alternative to lithium-ion batteries for large-scale stationary energy storage applications due to the abundant sodium resource in the Earth' crust and as a result, relatively low cost. Sodium layered transition metal oxides (Na x TMO 2 ) are proper Na-ion cathode materials because of low cost and high theoretical capacity. Currently most researchers focus on the improvement of electrochemical performance such as high rate capability and long cycling stability. However, for Na x TMO 2 , the structure stability against humid atmosphere is essentially important since most of them are instable in air, which is not favorable for practical application. Here we provide a comprehensive review of recent progresses on air-stable Na x TMO 2 oxides. Several effective strategies are discussed, and further investigations on the air-stable cathodes are prospected.Influence of Air on Na x TMO 2So far, many researches have proven that the water and CO 2 molecules from air can react with Na x TMO 2 , bringing FIGURE 3 | The schematic of the crystal structure for layered sodium TM oxides, left is P2 type and right is O 3 type.FIGURE 4 | (A)Insertion of water molecules in Na + layer, (B) Ion exchange between H + and Na + in Na + layer.Frontiers in Chemistry | www.frontiersin.org
# Introduction
The growing demand for large-scale energy storage applications has driven the research interest into new energy storage systems with low cost. Although lithium-ion battery (LIB) can deliver high energy and power density, the limited resource and the rising cost of lithium may restrict their application in grid scale energy storage. Recently, sodium-ion battery (SIB), which owns a similar chemical storage mechanism to LIB, has been rapidly developed as a complementary technology. As the second lightest alkali metal, sodium resource is inexpensive and almost globally available. The common abundant sodium salt such as Na 2 SO 4 , NaCl, and Na 2 CO 3 could be obtained from marine or mineral. In addition, copper foil can be replaced by cheaper aluminum foil for anode current collector since sodium has no reaction with aluminum. Therefore, SIB has received considerable attention as a promising alternative to LIB [bib_ref] Electrical energy storage for the grid: a battery of choices, Dunn [/bib_ref] [bib_ref] Electrochemical energy storage for green grid, Yang [/bib_ref] [bib_ref] Update on Na-based battery materials. A growing research path, Palomares [/bib_ref] [bib_ref] Room-temperature stationary sodiumion batteries for large-scale electric energy storage, Pan [/bib_ref] [bib_ref] Research development on sodium-ion batteries, Yabuuchi [/bib_ref] [bib_ref] A comprehensive review of sodium layered oxides: powerful cathodes for Na-ion batteries, Han [/bib_ref] [bib_ref] Review-practical issues and future perspective for Na-Ion batteries, Kubota [/bib_ref] [bib_ref] The emerging chemistry of sodium ion batteries for electrochemical energy storage, Kundu [/bib_ref] [bib_ref] Recent advances and prospects of cathode materials for sodium-ion batteries, Xiang [/bib_ref] [bib_ref] Sodium-ion batteries: present and future, Hwang [/bib_ref] [bib_ref] Surface and interface engineering of silicon-based anode materials for lithium-ion batteries, Luo [/bib_ref] [bib_ref] From lithium-ion to sodium-ion batteries: advantages, challenges, and surprises, Nayak [/bib_ref] [bib_ref] Engineering the distribution of carbon in silicon oxide nanospheres at atomic level..., Zhu [/bib_ref].
The SIB system consists of five parts: cathode, anode, membrane, electrolyte and current collector, which has the same structure as LIB. shows typical configuration of a SIB coin cell, in which sodium layered transition metal oxide (Na x TMO 2 ) and hard carbon are employed as cathode and anode, respectively. During the charge process, the Na + and e − migrates to hard carbon anode with voltage increasing. During the discharge process, Na + and e − return to Na x TMO 2 cathode reversibly with voltage decreasing. The overall reaction can be described as:
[formula] Na x TMO 2 + C ↔ TMO 2 + Na x C(1) [/formula]
Numerous cathode materials such as polyanion compounds [bib_ref] Na-ion mobility in layered Na 2 FePO 4 F and olivine Na[Fe,Mn]PO..., Tripathi [/bib_ref] , layered transition metal (TM) oxides [bib_ref] Patterning of sodium ions and the control of electrons in sodium cobaltate, Roger [/bib_ref] [bib_ref] Electrochemical investigation of the P2-NaxCoO 2 phase diagram, Berthelot [/bib_ref] [bib_ref] The P2-Na 2/3 Co 2/3 Mn 1/3 O 2 phase: structure, physical..., Carlier [/bib_ref] and Prussian blue or Metal-Organic compounds [bib_ref] A metalorganic compound as cathode material with superhigh capacity achieved by reversible..., Fang [/bib_ref] [bib_ref] Prussian blue nanocubes with an open framework structure coated with PEDOT as..., Su [/bib_ref] [bib_ref] Sodiumion batteries: prussian blue cathode materials for sodium-ion batteries and other ion..., Qian [/bib_ref] have been applied as Na + host materials. Layered TM oxides show a high theoretical capacity among these cathode materials [bib_ref] Layered oxide cathodes for sodium-ion batteries: phase transition, air stability, and performance, Wang [/bib_ref]. In addition, taking the preparation process and cost into consideration, the layered transition metal oxides are the optimal choice for practical application because they can be easily obtained by calcining the precursors in air. As a result, the layered transition metal oxides with general formula Na x TMO 2 have attracted more and more attention since the first report by Delmas' group in the 1980s [bib_ref] Structural classification and properties of the layered oxides, -X Delmas [/bib_ref].
Most of researches about Na x TMO 2 focused on the improvement of electrochemical properties, such as: (i) eliminating Na + vacancy ordering to improve rate capability; [bib_ref] Na( + )/vacancy disordering promises high-rate Na-ion batteries, Wang [/bib_ref] (ii) suppressing phase transition or surface coating to achieve long cycling life; [bib_ref] Suppressing the P2-O2 phase transition of Na 0.67 Mn 0.67 Ni 0.33..., Wang [/bib_ref] [bib_ref] A Honeycomb-layered oxide cathode for sodium-ion batteries with suppressed P3-O1 phase transition, You [/bib_ref] [bib_ref] A chemical approach to raise cell voltage and suppress phase transition in..., Sathiya [/bib_ref] (iii) exploring oxygen ion redox mechanism FIGURE 1 | The configuration of sodium-ion battery.
FIGURE 2 | The redox couples in Na x TMO 2 compounds with their corresponding potential range. to achieve high energy density [bib_ref] Anionic redox chemistry in Na-rich Na 2 Ru 1−y Sn y O..., Rozier [/bib_ref] [bib_ref] Rational design of Na(Li1/3Mn2/3)O2 operated by anionic redox reactions for advanced sodium-ion..., Kim [/bib_ref] [bib_ref] Anionic redox activity in a newly Zn-doped sodium layered oxide P2-Na2/3Mn1-yZnyO2 (0..., Bai [/bib_ref] [bib_ref] Oxygen redox chemistry without excess alkali-metal ions in Na 2/3, Maitra [/bib_ref] [bib_ref] Reversible anionic redox activity in Na 3 RuO 4 cathodes: a prototype..., Qiao [/bib_ref] , and so on. However, most Na x TMO 2 materials are hygroscopic and air-instable, which limit their practical applications because huge cost will be spent on materials' storage and transportation [bib_ref] Chemistry and electrochemistry of low-temperature manganese oxides as lithium intercalation compounds, Franger [/bib_ref] [bib_ref] Intercalation of water in P2, T2 and O2 structure A z, Lu [/bib_ref] [bib_ref] Synthesis and characterization of high-temperature hexagonal P2-Na 0.6 MnO 2 and its..., Caballero [/bib_ref] [bib_ref] Unique properties of α-NaFeO 2 : De-intercalation of sodium via hydrolysis and..., Monyoncho [/bib_ref] [bib_ref] Uptake of CO 2 in layered P2-Na 0.67 Mn 0.5 Fe 0.5..., Duffort [/bib_ref] [bib_ref] Review-practical issues and future perspective for Na-Ion batteries, Kubota [/bib_ref] [bib_ref] Charge storage mechanism and degradation of P2-type sodium transition metal oxides in..., Boyd [/bib_ref] [bib_ref] Understanding the air-exposure degradation chemistry at a nanoscale of layered oxide cathodes..., You [/bib_ref]. So in recent years, the design and synthesis of air-stable Na x TMO 2 materials have become a hot topic. In this review, we summarize the recent progress on air-stable Na x TMO 2 materials from structure understanding to corresponding solutions, and at the same time we address the remaining problems and challenges for further development.
## Structure of na x tmo 2
In Na x TMO 2 compounds, TM layers are usually occupied by Ti,V, [bib_ref] Na x VO 2 as possible electrode for Na-ion batteries, Hamani [/bib_ref] [bib_ref] P2-Na x VO 2 system as electrodes for batteries and electron-correlated materials, Guignard [/bib_ref] [bib_ref] Research progress on vanadium-based cathode materials for sodium ion batteries, Wang [/bib_ref] Cr, [bib_ref] Etude par desintercalation electrochimique des systemes Na x CrO 2 et Na..., Braconnier [/bib_ref] [bib_ref] NaCrO 2 cathode for high-rate sodium-ion batteries, Yu [/bib_ref] Mn, [bib_ref] Electrochemical properties of monoclinic NaMnO 2, Ma [/bib_ref] Fe, [bib_ref] Crystal structures and electrode performance of alpha-NaFeO 2 for rechargeable sodium batteries, Yabuuchi [/bib_ref] Co, [bib_ref] Electrochemical properties of Na x CoO 2 (x∼0.71) cathode for rechargeable sodium-ion..., Rai [/bib_ref] Ni, [bib_ref] Electrochemical properties of monoclinic NaNiO 2, Vassilaras [/bib_ref] [bib_ref] An abnormal 3.7 Volt O3-type sodium-ion battery cathode, Wang [/bib_ref] Cu, [bib_ref] Electrochemical properties of NaCuO 2 for sodium-ion secondary batteries, Ono [/bib_ref] [bib_ref] A phasetransition-free cathode for sodium-ion batteries with ultralong cycle life, Jiang [/bib_ref] [bib_ref] Structural analysis of NaCuO 2 cathode at various charged/discharged stages and its..., Ono [/bib_ref] a mixture of two [bib_ref] On the Na x Ni 0.6 Co 0.4 o 2 system: physical..., Saadoune [/bib_ref] [bib_ref] P2-type Na x [Fe 1/2 Mn 1/2 ]O 2 made from earthabundant..., Yabuuchi [/bib_ref] [bib_ref] Structural and electrochemical characterizations of P2 and new O3-Na x Mn 1−y..., Mortemard De Boisse [/bib_ref] [bib_ref] Synthesis and characterization of pure P2-and O3-Na 2/3 Fe 2/3 Mn 1/3..., Gonzalo [/bib_ref] [bib_ref] A novel tunnel Na 0.61 Ti 0.48 Mn 0.52 O 2 cathode..., Guo [/bib_ref] [bib_ref] Free-standing Na 2/3 Fe 1/2 Mn 1/2 O 2 @graphene film for..., Zhu [/bib_ref] [bib_ref] A P2-Na 0.67 Co 0.5 Mn 0.5 O 2 cathode material with..., Zhu [/bib_ref] [bib_ref] Stable layered P3/P2 Na 0.66 Co 0.5 Mn 0.5 O 2 cathode..., Chen [/bib_ref] [bib_ref] Tunnelstructured Na 0.54 Mn 0.50 Ti 0.51 O 2 and Na 0.54..., Jiang [/bib_ref] [bib_ref] Copper substituted P2-type Na 0.67 Cu x Mn 1−x O 2 :..., Kang [/bib_ref] [bib_ref] P2-Na 0.6 [Cr 0.6 Ti 0.4 ]O 2 cation-disordered electrode for high-rate..., Wang [/bib_ref] [bib_ref] An O3-type NaNi 0.5 Mn 0.5 O 2 cathode for sodium-ion batteries..., Wang [/bib_ref] [bib_ref] P2-Na x Co y Mn 1−y O 2 (y = 0, 0.1)..., Bucher [/bib_ref] [bib_ref] Investigation of Al-doping effects on the NaFe 0.5 Mn 0.5 O 2..., Kee [/bib_ref] [bib_ref] Layered P2-Na 2/3 [Ni 1/3 Mn 2/3 ]O 2 as high-voltage cathode..., Liu [/bib_ref] [bib_ref] Structural characterization of layered Na 0.5 Co 0.5 Mn 0.5 O 2..., Manikandan [/bib_ref] [bib_ref] Layered P2-Na 2/3 Co 1/2 Ti 1/2 O 2 as a high-performance..., Sabi [/bib_ref] [bib_ref] Na-rich layered Na 2 Ti 1−x Cr x O 3−x/2 (x =..., Song [/bib_ref] or more elements [bib_ref] Intercalation of water in P2, T2 and O2 structure A z, Lu [/bib_ref] [bib_ref] Water sensitivity of layered P2/P3-Na x Ni 0.22 Co 0.11 Mn 0.66..., Buchholz [/bib_ref] [bib_ref] O3-type Na(Mn 0.25 Fe 0.25 Co 0.25 Ni 0.25 )O 2 :..., Li [/bib_ref] [bib_ref] A quinary layer transition metal oxide of NaNi 1/4 Co 1/4 Fe..., Yue [/bib_ref] [bib_ref] High-performance P2-phase Na 2/3 Mn 0.8 Fe 0.1 Ti 0.1 O 2..., Han [/bib_ref] [bib_ref] P2-Type Na x Cu 0.15 Ni 0.20 Mn 0.65 O 2 cathodes..., Kang [/bib_ref] [bib_ref] Synthesis, structural and electrochemical study of O3-NaNi 0.4 Mn 0.4 Co 0.2..., Satyanarayana [/bib_ref] [bib_ref] Sodium insertion cathode material Na 0.67 Ni 0.4 Co 0.2 Mn 0.4..., Sun [/bib_ref] [bib_ref] P2-Na 2/3 Ni 1/3 Mn 5/9 A l1/9 O 2 microparticles as..., Zhang [/bib_ref] The corresponding Frontiers in Chemistry | www.frontiersin.org redox potential ranges of these TM are presented in . Na x TiO 2 compound is usually used as anode material due to its low redox potential range. Na x (Ni y Mn 1−y )O 2 compound has been thoroughly investigated as cathode material because of the relatively high redox potential and theoretical capacity. [bib_ref] In situ X-ray diffraction study of P2-Na 2/3 [Ni 1/3 Mn 2/3..., Lu [/bib_ref] [bib_ref] Investigation of the NaNi x Mn 1−x O 2 (0 <= x..., Fielden [/bib_ref] V, Cr and Co substitution also shows a proper potential range for cathode but it may not suitable for practical application since V, Cr, and Co are expensive and toxic. Although Fe and Cu are almost electrochemical inactive when used as LiTMO 2 for LIB system, [bib_ref] Preparation of LiFeO 2 with Alpha-NaFeO 2 -type structure using a mixed-alkaline..., Ado [/bib_ref] [bib_ref] Reversible electrochemical reaction of CuO with Li in the LiCuO 2 system, Arachi [/bib_ref] these two elements are proven to be highly active in Na x TMO 2 as Na-ion host. [bib_ref] P2-type Na x [Fe 1/2 Mn 1/2 ]O 2 made from earthabundant..., Yabuuchi [/bib_ref] [bib_ref] Structural analysis of NaCuO 2 cathode at various charged/discharged stages and its..., Ono [/bib_ref] Since Ni and Co resources are mostly consumed by LIB system, the abundant Fe and Cu resources with low price are suitable for Na x TMO 2 as sodium storage materials. [bib_ref] Prototype sodium-ion batteries using an air-stable and Co/Ni-Free O3-layered metal oxide cathode, Mu [/bib_ref] In addition, electrochemical inactive metal ions such as Li + , Mg 2+ and Zn 2+ could also be introduced into the TM layer for the improvement of electrochemical performance [bib_ref] Identifying the critical role of Li substitution in P2-Na x [Li y..., Xu [/bib_ref] [bib_ref] Suppressing the P2-O2 phase transition of Na 0.67 Mn 0.67 Ni 0.33..., Wang [/bib_ref]. [fig_ref] TABLE 1 |: Ionic radius of metal ions [/fig_ref] lists the most common metal ions for the construction of TM layers and their corresponding ionic radii with coordination number of six [bib_ref] Revised effective ionic-radii and systematic studies of interatomic distances in halides and..., Shannon [/bib_ref]. Cations with similar ionic radius can partially substitute each other to form solid solutions, and hence various Na x TMO 2 compounds could be designed by choosing two or more proper metal ions for the TM layer to improve electrochemical performance.ta The crystal structure of Na x TMO 2 can be usually classified into two types, P2 and O3 . The symbols of "P" and "O" are from the abbreviation of "prismatic" and "octahedral, " "2" and "3" represents the stacking arrangement per unit of O ions. For P2 type structure (usually x = 2/3), Na ions occupy two different prismatic sites, one shares faces between TMO 6 octahedra called Na f sites and another shares edges between TMO 6 octahedra called Na e sites. TM ions are surrounded by oxygen frameworks with a stacking mode of ABBA. For O3 structure (usually x = 1), all Na ions share one edge and one face with TMO 6 octahedra. The oxygen frameworks are arranged in ABCABC pattern [bib_ref] Structural classification and properties of the layered oxides, -X Delmas [/bib_ref] [bib_ref] Sodium-ion diffusion and ordering in singlecrystal P2-Na x CoO 2, Shu [/bib_ref] [bib_ref] Crystal-to-stripe reordering of sodium ions in Na x CoO 2, Morris [/bib_ref] [bib_ref] Vacancy ordering in O3-type layered metal oxide sodium-ion battery cathodes, Toumar [/bib_ref] [bib_ref] Effects of transition-metal mixing on Na ordering and kinetics in layered P2..., Zheng [/bib_ref]. negative influence on its morphology, crystal structure and electrochemical performance. The water molecules are easy to react with air-instable Na x TMO 2 by inserting into the Na layer (Le [bib_ref] Structural and electrochemical characteristics of a lamellar sodium manganese oxide synthesized via..., Goff [/bib_ref] [bib_ref] Studies of the layered manganese bronzes, Na 2/3 [Mn 1−x M x..., Paulsen [/bib_ref] [bib_ref] Chemistry and electrochemistry of low-temperature manganese oxides as lithium intercalation compounds, Franger [/bib_ref] [bib_ref] Intercalation of water in P2, T2 and O2 structure A z, Lu [/bib_ref] [bib_ref] Synthesis and characterization of high-temperature hexagonal P2-Na 0.6 MnO 2 and its..., Caballero [/bib_ref] [bib_ref] Uptake of CO 2 in layered P2-Na 0.67 Mn 0.5 Fe 0.5..., Duffort [/bib_ref] [bib_ref] Charge storage mechanism and degradation of P2-type sodium transition metal oxides in..., Boyd [/bib_ref] or exchanging Na + with H + , [bib_ref] Unique properties of α-NaFeO 2 : De-intercalation of sodium via hydrolysis and..., Monyoncho [/bib_ref] [bib_ref] Review-practical issues and future perspective for Na-Ion batteries, Kubota [/bib_ref] [bib_ref] Moisture exposed layered oxide electrodes as Na-ion battery cathodes, Han [/bib_ref] [bib_ref] Designing air-stable O3-type cathode materials by combined structure modulation for Na-Ion batteries, Yao [/bib_ref] leading to the expansion of interlayer spacing and the formation of impure phase . While CO 2 can transform to CO 2− 3 on the surface of Na x TMO 2 [bib_ref] Understanding the air-exposure degradation chemistry at a nanoscale of layered oxide cathodes..., You [/bib_ref]. These air exposed Na x TMO 2 usually show serious capacity decay and large polarization because of: (i) the side reaction between water and electrolyte [bib_ref] Decomposition reaction of LiPF 6 -based electrolytes for lithium ion cells, Kawamura [/bib_ref] [bib_ref] The mechanism of HF formation in LiPF 6 based organic carbonate electrolytes, Lux [/bib_ref] [bib_ref] Moisture exposed layered oxide electrodes as Na-ion battery cathodes, Han [/bib_ref] ; (ii) the active-materials' surface dissolution triggered by the acid attack of proton, which is released by water molecules [bib_ref] Self-discharge of LiMn 2 O 4 /C Li-Ion cells in their discharged..., Blyr [/bib_ref] [bib_ref] Structural fatigue in spinel electrodes in high voltage (4 V) Li/Li x..., Thackeray [/bib_ref] [bib_ref] Free energies for acid attack reactions of lithium cobaltate, Benedek [/bib_ref] ; (iii) capacity and electronic conductivity decrease caused by inactive Na 2 CO 3 layer [bib_ref] Understanding the air-exposure degradation chemistry at a nanoscale of layered oxide cathodes..., You [/bib_ref]. Therefore, the air-instable Na x TMO 2 cannot maintain its original crystal structure and electrochemical property under moisture atmosphere condition.
## Reaction mechanisms of water on na x tmo 2
Water molecules can insert into Na layer to form a Na x TMO 2 ·yH 2 O hydrate phase, which usually occurs in P2 type structure (Le [bib_ref] Structural and electrochemical characteristics of a lamellar sodium manganese oxide synthesized via..., Goff [/bib_ref] [bib_ref] Chemistry and electrochemistry of low-temperature manganese oxides as lithium intercalation compounds, Franger [/bib_ref] [bib_ref] Intercalation of water in P2, T2 and O2 structure A z, Lu [/bib_ref] [bib_ref] Synthesis and characterization of high-temperature hexagonal P2-Na 0.6 MnO 2 and its..., Caballero [/bib_ref] [bib_ref] Uptake of CO 2 in layered P2-Na 0.67 Mn 0.5 Fe 0.5..., Duffort [/bib_ref] FIGURE 5 | (A) XRD patterns of pristine and air-exposed Na 2/3 Co 1/6 Ni 1/6 Mn 2/3 O 2 samples, (B) XRD patterns of pristine and air-exposed Na 2/3 Co 1/3 Mn 2/3 O 2 samples, (C) Rietveld refinement of (Na·H 2 O) 2/3 Co 1/3 Mn 2/3 O 2 , (D) Crystal structure of (Na·H 2 O) 2/3 Co 1/3 Mn 2/3 O 2 . Reproduced with permission [bib_ref] Intercalation of water in P2, T2 and O2 structure A z, Lu [/bib_ref]. Copyright 2001, American Chemical Society. (E) Water molecules insert into the Na layer of Na 0.7 MnO 2 and the change of XRD patterns with the increasing time of water soaking. Reproduced with permission [bib_ref] Chemistry and electrochemistry of low-temperature manganese oxides as lithium intercalation compounds, Franger [/bib_ref]. Elsevier. (B) XRD patterns of pristine, air exposed and water soaked NaNi 0.5 Mn 0.5 O 2 samples. Reproduced with permission . Copyright 2017, American Chemical Society.
FIGURE 7 | SEM images of (a) pristine NaNi 1/3 Mn 1/3 Co 1/3 O 2 , (b) NaNi 1/3 Mn 1/3 Co 1/3 O 2 after 15 days air-exposure, (c) NaNi 1/3 Mn 1/3 Co 1/3 O 2 after 30 days air-exposure. (d) Infrared spectrum of NaNi 1/3 Mn 1/3 Co 1/3 O 2 after 30 days air-exposure. Reproduced with permission [bib_ref] Synthesis, structure, and electrochemical properties of the layered sodium insertion cathode material:..., Sathiya [/bib_ref]. Copyright 2012, American Chemical Society. (e) SEM image of extracted sample, the sample is prepared by drying the aqueous solution of water-NaFeO 2 mixture, (f) XRD patterns of the extracted sample and commercial Na 2 CO 3 . Reproduced with permission [bib_ref] Unique properties of α-NaFeO 2 : De-intercalation of sodium via hydrolysis and..., Monyoncho [/bib_ref]. Copyright 2013, Elsevier.
Frontiers in Chemistry | www.frontiersin.org 2015; [bib_ref] Charge storage mechanism and degradation of P2-type sodium transition metal oxides in..., Boyd [/bib_ref]. In 2001, [bib_ref] Intercalation of water in P2, T2 and O2 structure A z, Lu [/bib_ref] studied the water insertion reaction mechanism for the first time in P2-Na 2/3 Co x Ni 1/3−x Mn 2/3 O 2 compound (x = 1/6 or 1/3). Compared with the XRD patterns of pristine Na 2/3 Co x Ni 1/3−x Mn 2/3 O 2 , two new peaks around 14 - and 28 - were observed after exposing Na 2/3 Co x Ni 1/3−x Mn 2/3 O 2 samples in humid air environment . These two peaks were assigned as hydrate phase due to the insertion of water molecules in Na layers. Rietveld refinement of hydrate Na 2/3 Co 1/3 Mn 2/3 O 2 ·yH 2 O indicated that the ratio of water/Na is close to 1:1 and the oxygen atoms from water was in the 2c site of the crystal structure . [bib_ref] Chemistry and electrochemistry of low-temperature manganese oxides as lithium intercalation compounds, Franger [/bib_ref] investigated the influence of water soaking on α-Na 0.7 MnO 2 . With the increasing of water soaking time, the two peaks around 8 - and 16 - were vanished and four new peaks around 6.5 - , 13 - , 19 - and 21 - appeared gradually. The α-Na 0.7 MnO 2 was totally transformed to Na 0.45 MnO 2 ·0.6H 2 O after 60 min of water soaking treatment (NaNMCu) and P2-Na 0.64 Mn 0.62 Cu 0.31 O 2 (NaMCu) samples. After air-exposure treatment of these four samples for 8 days, the XRD patterns of NaNMCu and NaMCu samples remained unchanged, while two new peaks around 12.5 - and 25 - appeared in the patterns of NaNMFe and NaNMCo samples, indicating that water can insert in the interlayer spacing of NaNMFe and NaNMCo samples . From the STEM images of these four samples before and after air-exposure, an obvious extension in interlayer spacing could be seen after air-exposure, proving the insertion of water molecules in the interlayer spacing . Although water molecules can insert into the interlayer spacing of P2-Na x TMO 2 to form a P2-Na x TMO 2 ·yH 2 O hydrate phase, Na x TMO 2 phase can be regenerated by heat treatment at 200 - C to remove the water molecules [bib_ref] Intercalation of water in P2, T2 and O2 structure A z, Lu [/bib_ref].
For most O3-type NaTMO 2 , water molecules can release H + to exchange the Na + , [bib_ref] Unique properties of α-NaFeO 2 : De-intercalation of sodium via hydrolysis and..., Monyoncho [/bib_ref] [bib_ref] Review-practical issues and future perspective for Na-Ion batteries, Kubota [/bib_ref] [bib_ref] Moisture exposed layered oxide electrodes as Na-ion battery cathodes, Han [/bib_ref] [bib_ref] Designing air-stable O3-type cathode materials by combined structure modulation for Na-Ion batteries, Yao [/bib_ref] which could be regarded as hydrolysis reaction:
[formula] NaTMO 2 + xH 2 O → Na 1−x H x TMO 2 + NaOH(2) [/formula]
Specially, if the TM layers contain a certain amount of Ni 2+ ions, NiO would be emerged during the air exposure treatment:
[formula] NaNi y TM 1−y O 2 + xH 2 O → Na 1−x H x Ni y−z TM 1−y O 2−z + xNaOH + zNiO(3) [/formula]
This hydrolysis phenomenon has been confirmed in NaNi 0.5 Mn 0.5 O 2 , NaNi 0.7 Mn 0.15 Co 0.15 O 2 and NaFeO 2 compounds [bib_ref] Unique properties of α-NaFeO 2 : De-intercalation of sodium via hydrolysis and..., Monyoncho [/bib_ref] [bib_ref] Review-practical issues and future perspective for Na-Ion batteries, Kubota [/bib_ref] [bib_ref] Understanding the air-exposure degradation chemistry at a nanoscale of layered oxide cathodes..., You [/bib_ref]. A simple way to verify this hydrolysis reaction is to analyze the change of pH value after soaking NaTMO 2 in deionized water due to the release of NaOH. In 2013, Bissessur reported that the pH of aqueous solution was higher than 12 after mixing O3-NaFeO 2 sample with deionized water [bib_ref] Unique properties of α-NaFeO 2 : De-intercalation of sodium via hydrolysis and..., Monyoncho [/bib_ref]. In addition, compared with the XRD patterns of pristine NaFeO 2 sample, the 003 peak of hydrolysis Na 1−x H x FeO 2 sample became broader and shifted to low angle, indicating the formation of a disordered crystal structure. Wang et al. investigated the hydrolysis reaction of O3-NaNi 0.5 Mn 0.5 O 2 sample by testing the temperature after water soaking because this reaction can release heats . After NaNi 0.5 Mn 0.5 O 2 was added into water, the temperature of the water was increased from 24.4 to 30.8 - C. In contrast to the XRD pattern of as-synthesized sample, the 003 and 006 peaks of water soaked NaNi 0.5 Mn 0.5 O 2 shifted to low angle with the generation of NiO impurity phase. Importantly, unlike P2 type, the hydrolysis reaction between O3 type NaTMO 2 and water is irreversible. Reaction Mechanisms of CO 2 on Na x TMO 2 As mentioned above, NaOH is generated on the surface of O3 type Na x TMO 2 during the air exposure process, then CO 2 can further react with NaOH to form Na 2 CO 3 [bib_ref] Synthesis, structure, and electrochemical properties of the layered sodium insertion cathode material:..., Sathiya [/bib_ref] [bib_ref] Unique properties of α-NaFeO 2 : De-intercalation of sodium via hydrolysis and..., Monyoncho [/bib_ref] [bib_ref] Understanding the air-exposure degradation chemistry at a nanoscale of layered oxide cathodes..., You [/bib_ref]. This reaction can be described as:
NaTMO 2 + xH 2 O + CO 2 → Na 1−x H x TMO 2 + Na 2 CO 3 (4) [bib_ref] Synthesis, structure, and electrochemical properties of the layered sodium insertion cathode material:..., Sathiya [/bib_ref] proved the formation of Na 2 CO 3 on the surface of NaNi 1/3 Mn 1/3 Co 1/3 O 2 particles. Compared to the pristine sample , the surface showed no obvious change after 15 days air-exposure but became quite rough after 30 days air-exposure . IR spectrum revealed the bands of CO 2− 3 at 1,450 and 863 cm −1 , suggesting the existence of sodium carbonates on NaNi 1/3 Mn 1/3 Co 1/3 O 2 particles' surface . [bib_ref] Unique properties of α-NaFeO 2 : De-intercalation of sodium via hydrolysis and..., Monyoncho [/bib_ref] extracted the aqueous solution from the mixture of NaFeO 2 and water . The XRD pattern of extracted sample matched very well to commercial Na 2 CO 3 , proving the reaction of CO 2 and NaFeO 2 compounds .
After the formation of Na 2 CO 3 on the surface, the CO 2− 3 can even be inserted into the TM layer, forming a "CO 4 " tetrahedron. Duffort et al. elucidated the mechanism of CO 2− 3 insertion in Na 0.67 Mn 0.5 Fe 0.5 O 2 crystal . With increasing the time of air exposure, ribbonlike particles start to appear and grow longer gradually. In addition, the corresponding XRD patterns of Na 0.67 Mn 0.5 Fe 0.5 O 2 are also changed. New peak is observed around 13 - after a month air exposure, indicating the formation of hydrate phase (phase 3) and sodiumdepleted P2 phase (phase 2). In Fourier difference map, the existence of large residual nuclear density is caused by the carbonate ionsbecause the insertion of CO 2− 3 leading to the changing of the nuclear density distribution [fig_ref] FIGURE 9 |: SEM images of [/fig_ref]. In the TM layer, the CO 2− 3 is combined with one C-O bond to form a CO 4 tetrahedron structure.
Except for Na 2 CO 3 , other surface components are also observed. You et al. studied the surface reaction between NaNi 0.7 Mn 0.15 Co 0.15 O 2 and air systematically by using time-offlight secondary ion mass spectroscopy (TOF-SIMS) [bib_ref] Understanding the air-exposure degradation chemistry at a nanoscale of layered oxide cathodes..., You [/bib_ref]. The pristine sample shows a microsphere morphology, which is consisted of nanosized particles [fig_ref] FIGURE 9 |: SEM images of [/fig_ref]. After 24 h air exposure treatment, the surface of this microsphere becomes smooth with the absence of nano-particles [fig_ref] FIGURE 9 |: SEM images of [/fig_ref]. Finally, a thick layer of impurities is formed on the surface after 7 days air exposure [fig_ref] FIGURE 9 |: SEM images of [/fig_ref]. Elements of Na, Ni, and C are distributed uniformly on the impurity surface [fig_ref] FIGURE 9 |: SEM images of [/fig_ref]. The existence of NaNi + , NiO, NaC 2 O − 2 , C 3 H − 2 , Na 2 F + , and F − composition are confirmed by TOF-SIMS, indicating the surface degradation of NaNi 0.7 Mn 0.15 Co 0.15 O 2 as well as the reaction between sodium carbonates and PVDF [fig_ref] FIGURE 9 |: SEM images of [/fig_ref]. The 003 peak shifts to lower angles gradually because of the migration of Na + to the surface while the 104 peak becomes weak and vanishes after 48 h air exposure [fig_ref] FIGURE 9 |: SEM images of [/fig_ref]. According to the XRD patterns in [fig_ref] FIGURE 9 |: SEM images of [/fig_ref] , impurities' peaks such as NiO and Na 2 CO 3 are observed, indicating the surface reaction when NaNi 0.7 Mn 0.15 Co 0.15 O 2 contacts to CO 2 and H 2 O. Since the CO 2 and H 2 O are absorbed and reacted with NaNi 0.7 Mn 0.15 Co 0.15 O 2 , the air-exposed sample loses more weight than the fresh sample [fig_ref] FIGURE 9 |: SEM images of [/fig_ref]. All the results above can prove hat NaNi 0.7 Mn 0.15 Co 0.15 O 2 can react with the water and carbon dioxide in the air and the impurities of NaNi + , NiO, NaC 2 O − 2 , C 3 H − 2 , Na 2 F + and F − are generated on the surface.
## Air-stable na x tmo 2 compounds
As mentioned above, NaTMO 2 compounds can react with water and CO 2 in air, which lead to: (i) the formation of impure phase on the surface; (ii) the insertion of H 2 O and CO 2− 3 into interlayer spacing and TM layers, respectively. On one side, the formed NaOH and Na 2 CO 3 impure phase are electrochemical inactive and have low conductivity hence the rate capability of Na x TMO 2 suffer serious decrease. On the other side, the water molecules can bring side reaction with electrolyte while the insertion of CO 2− 3 FIGURE 10 | (A) XRD pattern of pristine Na 2/3 Ni 1/3 Mn 2/3 O 2 , (B) XRD pattern of Na 2/3 Ni 1/3 Mn 2/3 O 2 after 10 days air-exposure, (C) the Ni/Mn cationic ordering arrangement. Reproduced with permission [bib_ref] Intercalation of water in P2, T2 and O2 structure A z, Lu [/bib_ref]. Copyright 2001, American Chemical Society.
Frontiers in Chemistry | www.frontiersin.org affects the valence state of TM ions, leading to severe capacity decay. Therefore, more and more researchers are focusing on strategies to address this air-instable problems.
One strategy is to prevent the materials from contacting moisture. During the materials preparation process, once the high-temperature treatment is done, the Na x TMO 2 products are transferred to drying room [bib_ref] An O3-type NaNi 0.5 Mn 0.3 Ti 0.2 O 2 compound as..., Wang [/bib_ref] or argonfilled glove box [bib_ref] P2-type Na x [Fe 1/2 Mn 1/2 ]O 2 made from earthabundant..., Yabuuchi [/bib_ref] [bib_ref] Electrochemical properties of NaNi 1/3 Co 1/3 Fe 1/3 O 2 as..., Vassilaras [/bib_ref] immediately for the cooling process and subsequent cell assembling. However, this strategy may not be suitable for the large-scale application because huge cost will rise for materials' storage. Another strategy is to design air-stable Na x TMO 2 material. Recently, several P2 and O3 type Na x TMO 2 materials with high stability against moisture have been reported [bib_ref] Intercalation of water in P2, T2 and O2 structure A z, Lu [/bib_ref] [bib_ref] Prototype sodium-ion batteries using an air-stable and Co/Ni-Free O3-layered metal oxide cathode, Mu [/bib_ref] [bib_ref] Environmentally stable interface of layered oxide cathodes for sodium-ion batteries, Guo [/bib_ref] [bib_ref] Designing air-stable O3-type cathode materials by combined structure modulation for Na-Ion batteries, Yao [/bib_ref] [bib_ref] Cu( 2+ ) dual-doped layer-tunnel hybrid Na 0.6 Mn 1−x Cu x..., Chen [/bib_ref] [bib_ref] High energy density sodium-ion battery with industrially feasible and air-stable O3-type layered..., Deng [/bib_ref]. Under the treatment of air exposure and water soaking, these air-stable cathodes can maintain their original crystal structure and electrochemical performance. In this part, several effective strategies for air-stable Na x TMO 2 designing are summarized.
## Constructing tm cationic ordering arrangement
Lu et al. first reported an air-stable P2-Na 2/3 Ni 1/3 Mn 2/3 O 2 with high stability under moisture condition [bib_ref] Intercalation of water in P2, T2 and O2 structure A z, Lu [/bib_ref]. For the P2-Na 2/3 Ni 1/3 Mn 2/3 O 2 sample, after undergoing a 10 days air-exposure treatment, no peaks shift or new peaks formation were observed in the XRD pattern, indicating that water could not be inserted into the interlayer spacing . According to neutron diffraction analysis, the Ni 2+ and Mn 4+ ions formed an honeycomb structure with an √ 3a × √ 3a ordering arrangement in the TM layers [bib_ref] Superlattice ordering of Mn, Ni, and Co in layered alkali transition metal..., Lu [/bib_ref]. This Ni/Mn ordering arrangement in TM FIGURE 13 | (A) XRD patterns of Na 2/3 Cu x Ni 1/3−x Mn 2/3 O 2 samples (x = 0, 1/12, 1/6 and 1/4) before/after air-exposure, (B) the corresponding charge/discharge profiles. Reproduced with permission . Copyright 2017, American Chemical Society.
layers was supposed to induce a strong interlayer interaction to prevent the water insertion. When this ordering arrangement was suppressed by the substitution of Co or Fe for Ni, water molecules could be inserted into the interlayer spacing . However, this "interlayer interaction" between adjacent ordering TM layer has not been confirmed yet.
## Coating protective layer
Coating a protective layer on the surface of Na x TMO 2 is an effective method to prevent the Na x TMO 2 from air contacting. The most common way is to coat high-voltage metal oxides with high stability against moisture. In 2017, Zhou and co-workers . designed an efficient spinel-like titanium (III) oxides protective interface to improve the structure/electrochemical stability of NaMnTi 0.1 Ni 0.1 O 2 . The sample surface was covered by a high Ti concentration layer with thickness of 2 nm, as shown in the electron energyloss spectroscopy image. Two distinct phases could be observed from the high-angle annular dark field scanning transmission electron microscopy (HAADF-STEM) image. The bulk phase was a typical layered structurewhile the surface phase was spinellike structure. After exposing the naked bulk phase in humid air, two new peaks appeared around 12 - and 25 - , indicating the insertion of water molecules. Compared to the naked bulk phase, the XRD pattern of NaMnTi 0.1 Ni 0.1 O 2 sample showed no peak change since the spinel-like titanium (III) oxides interface can act as shield to protected the bulk phase from water attacking and the electrochemical performance of bulk phase can be maintained. In half cell system, the naked bulk phase showed a dramatic decrease after 50 cycles whereas NMTN sample only showed a slight decay after 100 cycles. [bib_ref] Understanding the air-exposure degradation chemistry at a nanoscale of layered oxide cathodes..., You [/bib_ref] Except for Ti and Zr oxides, we suppose that other metal oxides such as MgO, ZnO, and Al 2 O 3 also have the ability to work as protective layer because these high-voltage metal oxides all have high tolerance for moisture.
## Cu 2+ substitution
The Cu 2+ substitution is the simplest way to obtain air-stable Na x TMO 2 compounds. The success of Cu 2+ substitution to achieve high stability against moisture has been proven by many reports [bib_ref] Prototype sodium-ion batteries using an air-stable and Co/Ni-Free O3-layered metal oxide cathode, Mu [/bib_ref] [bib_ref] Designing air-stable O3-type cathode materials by combined structure modulation for Na-Ion batteries, Yao [/bib_ref] [bib_ref] Cu( 2+ ) dual-doped layer-tunnel hybrid Na 0.6 Mn 1−x Cu x..., Chen [/bib_ref] [bib_ref] High energy density sodium-ion battery with industrially feasible and air-stable O3-type layered..., Deng [/bib_ref] , few references give the working mechanisms of Cu 2+ in these air-stable Na x TMO 2 compounds.
In, 2017, Yao et al. designed an air-stable O3-NaNi 0.45 Cu 0.05 Mn 0.4 Ti 0.1 O 2 (NaNCMT) cathode though cosubstitution of Cu 2+ and Ti 4+ in O3-NaNi 0.5 Mn 0.5 O 2 (NaNM) compound. This strategy could decrease the Na + interlayer distance and increase the valence state of TM ions. According to the refined crystal structure of NaNM and NaNCMT, the interlayer distance was reduced from 3.45 Å to 3.37 Å, respectively, which was in favor of preventing the insertion of water molecules. DFT calculation revealed that Cu/Ti cosubstitution facilitated the increasing in valence state of Ni. Compared with NaNM compound, the XRD pattern of NaNCMT sample showed no obvious peaks change after air-exposure or water soaking. During charge/discharge process, only slight capacity decay was observed after aging experiments. However, the explanation about the relationship between valence state of TM ions and air-stability was not mentioned.
Zheng et al. investigated the structure stability of Na 2/3 Cu x Ni 1/3−x Mn 2/3 O 2 compounds (0 ≤ x ≤ 1/4) by airexposure treatment. Compared to the XRD patterns of pristine samples, neither peaks position change nor new peaks formation were observed after exposing Na 2/3 Cu x Ni 1/3−x Mn 2/3 O 2 in air condition for 21 days . According to charge/discharge profiles, all exposed electrodes had a little higher open circuit voltage than the un-exposed electrodes, but the average voltage and reversible capacity of the exposed electrodes showed no change or decay, indicating the air stability of these electrodes . Since the radii of Cu 2+ (0.73 Å) and Ni 2+ (0.69 Å) were similar, replacing Ni 2+ in Na 2/3 Ni 1/3 Mn 2/3 O 2 by Cu 2+ had no influence on the Ni/Mn cationic ordering arrangement. Therefore, the existence of Cu/Ni/Mn ordering arrangement could prevent the insertion of water molecules into the Na 2/3 Cu x Ni 1/3−x Mn 2/3 O 2 interlayer spacing because of the interlayer interaction between the adjacent Cu/Ni/Mn layer. However, no evidence was provided to prove the Cu/Ni/Mn ordering arrangement.
Other compounds such as O3-Na 0.9 Cu 0.22 Fe 0.30 Mn 0.48 O 2 , P2-Na 7/9 Cu 2/9 Fe 1/9 Mn 2/3 O 2 , O3-NaLi 0.05 Mn 0.5 Ni 0.3 Cu 0.1 Mg 0.05 O 2, and P2-Na 0.6 Mn 0.9 Cu 0.1 O 2 have been proved to be air-stable because all of their XRD patterns remained unchanged after air-exposure and water soaking treatment [bib_ref] Prototype sodium-ion batteries using an air-stable and Co/Ni-Free O3-layered metal oxide cathode, Mu [/bib_ref] [bib_ref] Cu( 2+ ) dual-doped layer-tunnel hybrid Na 0.6 Mn 1−x Cu x..., Chen [/bib_ref] [bib_ref] High energy density sodium-ion battery with industrially feasible and air-stable O3-type layered..., Deng [/bib_ref].
It seems that the Cu 2+ plays an important role in maintaining the structure stability of these compounds under moisture. The reported Cu 2+ substituted Na x TMO 2 compounds, such as O3-NaNi 0.45 Cu 0.05 Mn 0.4 Ti 0.1 O 2 , P2-Na 2/3 Cu x Ni 1/3−x Mn 2/3 O 2 , O3-Na 0.9 Cu 0.22 Fe 0.30 Mn 0.48 O 2 , O3-NaLi 0.05 Mn 0.5 Ni 0.3 Cu 0.1 Mg 0.05 O 2 , P2-Na 7/9 Cu 2/9 Fe 1/9 Mn 2/3 O 2, and P2-Na 0.6 Mn 0.9 Cu 0.1 O 2 , all show excellent structure stability under moisture condition. However, few investigations explain the working mechanism of Cu 2+ substitution in these air-stable compounds. In O3-NaNi 0.45 Cu 0.05 Mn 0.4 Ti 0.1 O 2 compound system, the working mechanism of Cu 2+ is attributed to the increase of the Ni valence state by DFT calculation, but the relationship between valence state of Ni and air-stability is not clear so far. In addition, how to explain the Ni free compound systems for their air stability after Cu 2+ substitution still remains problems. [fig_ref] TABLE 2 |: A summary of current air-stable Na [/fig_ref] lists most of the air-stable Na x TMO 2 compounds published to date, including the design strategies and the corresponding electrochemical performance.
## Summary and prospects
Air-stability is one of the key issues for practical application of Na x TMO 2 SIB cathode materials. In recent years, with understanding the structure of Na x TMO 2 , several strategies have been developed to obtain air-stable Na x TMO 2 compounds, including constructing TM ordering arrangement, coating protective layer and Cu 2+ substitution. However, there still remain some challenges. For example, the reaction mechanism of the "strong interlayer interaction" for TM ordering arrangement as well as the substitution of Cu 2+ and other cations should be further understood. In any case, we believe that the air-stable Na x TMO 2 materials with low cost and high theoretical capacity are highly competitive as SIB cathode materials in the large-scale energy storage application.
# Author contributions
RZ contributed conception and design of the manuscript. YZ organized the reference and wrote the first draft of the manuscript. All authors contributed to manuscript revision, read and approved the submitted version.
[fig] FIGURE 6 |: (A) XRD patterns of pristine and water soaked NaFeO 2 samples. Reproduced with permission(Monyoncho and Bissessur, 2013). Copyright 2013, [/fig]
[fig] FIGURE 8 |: (a) SEM images of Na 0.67 Mn 0.5 Fe 0.5 O 2 after 1 day air-exposure, (b) Na 0.67 Mn 0.5 Fe 0.5 O 2 after a couple of weeks' air-exposure, (c) Na 0.67 Mn 0.5 Fe 0.5 O 2 after a couple of months' air-exposure, (d) XRD patterns of Na 0.67 Mn 0.5 Fe 0.5 O 2 samples under different air-exposure time, (e) Nuclear Fourier difference map of Na 0.67 Mn 0.5 Fe 0.5 O 2 , (f) The local environment of CO, (g) The nuclear density around the carbon element position. Reproduced with permission. Copyright 2015, American Chemical Society. [/fig]
[fig] FIGURE 9 |: SEM images of (a) pristine NaNi 0.7 Mn 0.15 Co 0.15 O 2 , (b) NaNi 0.7 Mn 0.15 Co 0.15 O 2 after 24 h air-exposure, (c) NaNi 0.7 Mn 0.15 Co 0.15 O 2 after 7 days air-exposure, (d) dendrite-like impurity in part b with the corresponding elemental mappings of Na, C and Ni, (e) TOF-SIMS chemical mapping of NaNi 0.7 Mn 0.15 Co 0.15 O 2 after 24 h exposure, (f) XRD patterns of NaNi 0.7 Mn 0.15 Co 0.15 O 2 samples under different air exposed time, (g) the enlarged peaks from part f, (h) TGA profiles of pristine NaNi 0.7 Mn 0.15 Co 0.15 O 2 and 7 days exposed NaNi 0.7 Mn 0.15 Co 0.15 O 2 samples. Reproduced with permission (You et al., 2018). Copyright 2018, American Chemical Society.Frontiers in Chemistry | www.frontiersin.org [/fig]
[fig] FIGURE 11 |: (a) EELS chemical mapping of Ti element in NMTN sample, (b) STEM-HADDF image of NMTN sample, (c) Enlarged image extracted from the blue rectangle of part b, (d) Enlarged image extracted from the red rectangle of part b, (e) vulnerable mechanism of naked samples and protective mechanism of NMTN sample under moisture condition. (f) XRD patterns of naked sample (left) and NMTN sample (right) before/after air-exposure, (g) cycling performance of naked sample and NMTN sample. Reproduced with permission (Guo et al., 2017). Copyright 2017, Nature Publishing Group. [/fig]
[fig] FIGURE 12 |: (A) refined crystal structure of NaNM and NaNCMT samples, (B) Electronic density of states on Ni ion of NaNM, NaNCM, and NaNMT samples, respectively. (C) XRD patterns of pristine, air-exposure and water soaking NaNCMT samples. (D) charge/discharge curves of pristine, air-exposure and water soaking NaNCMT samples. Reproduced with permission. Copyright 2017, American Chemical Society. [/fig]
[table] TABLE 1 |: Ionic radius of metal ions. [/table]
[table] TABLE 2 |: A summary of current air-stable Na [/table]
|
Screening‐Level Estimates of Environmental Release Rates, Predicted Exposures, and Toxic Pressures of Currently Used Chemicals
We describe a procedure to quantify emissions of chemicals for environmental protection, assessment, and management purposes. The procedure uses production and use volumes from registration dossiers and combines these with Specific Environmental Release Category data. The procedure was applied in a case study. Emission estimations were made for chemicals registered under the European Union chemicals regulations for industrial chemicals (Registration, Evaluation, Authorisation and Restriction of Chemicals [REACH]) and for the active ingredients of medicines and crop protection products. Emissions themselves cannot be validated. Instead, emission estimates were followed by multimedia fate modeling and mixture toxic pressure modeling to arrive at predicted environmental concentrations (PECs) and toxic pressures for a typical European water body at steady state, which were compared with other such data. The results show that screeninglevel assessments could be performed, and yielded estimates of emissions, PECs, and mixture toxic pressures of chemicals used in Europe. Steady-state PECs agreed fairly well with measured concentrations. The mixture toxic pressure at steady state suggests the presence of effects in aquatic species assemblages, whereby few compounds dominate the predicted impact. The study shows that our screening-level emission estimation procedure is sufficiently accurate and precise to serve as a basis for assessment of chemical pollution in aquatic ecosystems at the scale of river catchments. Given a recognized societal need to develop methods for realistic, cumulative exposures, the emission assessment procedure can assist in the prioritization of chemicals in safety policies (such as the European Union REACH regulation), where "possibility to be used safely" needs to be demonstrated, and environmental quality policies (such as the European Union Water Framework Directive), where "good environmental quality" needs to be reached.
# Introduction
Human activity inevitably results in possibly non-negligible emissions of thousands of chemical substances. A recent inventory showed that more than 350 000 chemicals are currently registered globally [bib_ref] Toward a global understanding of chemical pollution: A first comprehensive analysis of..., Wang [/bib_ref] , and numbers and amounts of chemicals in use are expected to increase (United Nations Environment Programme 2013; European Chemicals Agency 2016b; [bib_ref] Synthetic chemicals as agents of global change, Bernhardt [/bib_ref]. This resulted in a call for comprehensive risk assessments, which would require quantification of emissions, exposures, and risks for large numbers of chemicals and their mixtures. Freshwater ecosystems are among the most human-impacted habitats, and chemical pollution is one of the main drivers of deterioration of freshwater biodiversity [bib_ref] Global threats to human water security and river biodiversity, Vörösmarty [/bib_ref]. This has triggered the development of preventive and curative regulatory measures. In terms of prevention, chemical substances can be registered and marketed only after an ex ante chemical safety assessment has demonstrated the possibility that (the registered tonnage of) the chemical can be used safely. Under the European Union Registration, Evaluation, Authorisation and Restriction of Chemicals (REACH) regulation , for example, exposure concentrations (predicted environmental concentration are compared with critical effect concentrations (predicted no-effect concentration ; the quotients (PEC/PNEC) are to remain below 1 for all possible uses of chemicals, commonly assessed individually.
In the present study, we expand on the ex ante assessment of net chemical pollution by considering the derivation of PECs for a large number of compounds from 1) amounts of chemicals used, and 2) fractions thereof released into the environment. This enables estimation of PECs by means of multimedia fate modeling and, subsequently, by a prospective assessment of the risk of net chemical pollution. The latter results are useful to evaluate whether the chemical safety policies are sufficient, or whether the environmental quality is affected by mixtures. For the aquatic compartment, the European Union Water Framework Directive is the regulatory framework that has been implemented to protect or restore water quality [bib_ref] establishing a framework for community action in the field of water policy, European Commission [/bib_ref]. We used the Water Framework Directive definition of good water quality status as benchmark to evaluate the PECs and the cumulation of risks of chemical pollution. Calls to account for mixtures for all policy frameworks have been voiced by scholars and regulatory bodies alike (European Commission 2017; [bib_ref] Regulate to reduce chemical mixture risk, Kortenkamp [/bib_ref] [bib_ref] Statement paper on advancing the assessment of chemical mixtures and their risks..., Drakvik [/bib_ref].
The assessment of emissions is currently the least developed step in the comprehensive assessment of chemical pollution risks. The main driver for the research of the present study was to develop a practice-oriented procedure by which the environmental release rates of large numbers of chemicals can be estimated, such that they can be used in chemical safety assessments and environmental quality assessments, for prioritization of risk-preventing and/or curative management efforts. Prospective modeling of ecological impacts of chemical substances starts with estimating the amounts of chemicals that are released to the environment as a consequence of production and downstream use in the production-use-waste chain. Emitted masses, combined with characteristics of the compounds and the environment (e.g., the breakdown rates and the volume of environmental compartments, respectively), result in exposure concentrations, either in steady state (derived from multimedia fate modeling;or for local water bodies (derived from a dedicated model; [bib_ref] iSTREEM ® : An approach for broad-scale in-stream exposure assessment of "down-the-drain"..., Kapo [/bib_ref]. Combining these data with knowledge of the toxicities of individual chemicals subsequently allows us to quantify the impacts of chemicals, alone or in mixtures. The information gained helps us to prioritize water bodies regarding the expected pollution pressures as well as those chemicals contributing most to the pollution, and to focus environmental protection and management [bib_ref] The status of the European waters in 2015: A review, Tsakiris [/bib_ref] [bib_ref] Water and health: From environmental pressures to integrated responses, Boelee [/bib_ref].
We developed a new emission estimation procedure by which release rates (into air, water, and soil) of chemicals are estimated from registered use volumes, using additional information submitted by registrants in registration dossiers (i.e., information about the intended uses and about the physical and chemical properties of the chemical). Commonly, available data only provide gross insights in potentially emitted fractions of chemicals. In a case study, we succeeded in using those data for 6409 chemicals and we applied further exposure and impact assessment modeling for the 4757 chemicals with sufficient data for all analysis steps. We developed an estimation procedure capable of delivering the release rates necessary for comprehensive (and practicable) impact assessments of chemicals produced and used in Europe. Because predicted emissions themselves cannot be validated at the European scale, we tested the resulting release rates for plausibility and functionality by comparing predicted and observed concentrations and toxic pressure data obtained from other studies.
The principal research question addressed in the present study was whether emission rates of chemicals into the environment can be estimated with sufficient accuracy and precision for environmental risk assessments of chemicals and their mixtures and for prioritization of chemicals, as a basis for regulation and management. Our aims were to: 1) describe a generic emission estimation approach that was developed to estimate emission rates of chemicals; 2) apply and test this emission estimation approach in a case study that involves a large fraction of the chemicals used in the European Union (REACH chemicals, pesticides, and pharmaceuticals), by comparing the outcomes of the emission model with data from other studies (involving both modeled and monitored data); 3) analyze the outcomes of the case study regarding the issue of predicted impacts of chemical pollution (as total mixture); and 4) evaluate the results regarding potential utility for the prioritization of chemicals for management (chemical and environmental quality policies).
The results of the use of the outcomes of the present study for water body-specific chemical pollution assessments and the implications for water quality protection and management have been published elsewhere [bib_ref] Species sensitivity distributions for use in environmental protection, assessment, and management of..., Posthuma [/bib_ref].
# Materials and methods
The analyses consisted of 3 main steps: 1) emission modeling, starting from data on masses of chemicals in trade in Europe; 2) exposure modeling, resulting in predicted mixture exposure concentrations of studied chemicals in a "typical European Union water body"; and 3) the assessment of the (mixture) toxic pressure of the predicted exposure concentrations.
## Emission modeling
We developed a new substance-flow estimation procedure (emission model) that enables accounting for the differences in release rates patterns covering all possible uses in the entire life cycle of chemicals. Release rates differ across compounds. That is, crop protection products used in open field applications are released almost entirely to crops and soils, "down-the-drain" household product chemicals find their ways to the environment via sewage collection and treatment systems, and other chemicals are made for use as parts of durable products, from which very variable amounts are released into the environment.
The new emission model is based on the generic emission estimation method developed in the 1980s by [bib_ref] Emissions of chemicals, Van Der Poel [/bib_ref]. This method was used in the first European Union System for the Evaluation of Substances (EUSES) model for regulatory chemical substances evaluation [bib_ref] European Union system for the evaluation of substances (EUSES). Principles and structure, Vermeire [/bib_ref]. It was later reworked into the format of Specific Environmental Release Category (spERC) tables (European Chemical Industry Council 2012), which are currently used in chemical safety assessments under REACH.
The new emission model considers all life stages of a chemical and the generic characteristics of the stages, coded and defined as follows.
Life cycle stages I -1 and I -2 (production and transport). Small, but significant amounts of chemical may be released to air, (waste)water, or soil directly from the first life cycle stage, during manufacturing or transport, although exceptions with specific point-source releases may occur (e.g., [bib_ref] Estimating emissions of PFOS and PFOA to the Danube River catchment and..., Lindim [/bib_ref] [bib_ref] Europe-wide estuarine export and surface water concentrations of PFOS and PFOA, Lindim [/bib_ref]. Manufacturing may or may not take place in Europe. Fractions of the overall tonnage, manufactured in Europe, were obtained from various dossiers and are accounted for in the emission estimation.
Life cycle stage I -3 (processing). Chemicals are rarely used without further processing. Generally, chemicals are formulated into products and distributed to other places in or outside Europe.
Life cycle stage II (use). During service life, chemicals are used in different ways and released to different degrees. For most chemicals, the largest releases take place during service life.
Life cycle stage III (recycling). Small but increasing fractions of chemicals produced and used are recycled; these data were combined with modeling chemical release from sewage treatment plants for chemicals passing through the plants, as appropriate. Total emission into the environment was thus modeled as the product of use volume, release fraction, and fraction not retained in the process of waste treatment, using literature-based estimated emission fractions [fig_ref] TABLE 1: Estimated emission fractions [/fig_ref] :
[formula] ∑ = ( × )× × E UseVol ActCat RF Fstp i j k i j k i i j j k i k , ,, , , , , (1) [/formula]
where E i,j,k denotes the European Union-wide emission rate [M/T] of substance i from use j into environmental medium k,
[formula] ( × ) UseVol ActCat i i j [/formula]
, represents the volume of substance i, used in activity category j [M], RF j,k is the fraction [−] of this use that is released into medium k, and Fstp i,k [T −1 ] denotes the fraction of substance i released to medium k after sewage treatment, as predicted by the sewage treatment plant model SimpleTreat, recommended in the REACH Guidance (European Chemicals Agency 2016a). Equation 1 was initially applied to 14 000 + chemical substances known to be currently used in Europe (REACH chemicals, pesticides, and pharmaceuticals), but due to missing data needed to derive PECs, the environmental-chemical analyses proceeded with 6409 compounds. Even for these relatively data-rich compounds, neither total amounts used, nor fractions applied in specific uses are available to allow precise emission estimation according to Equation 1. Knowledge of amounts produced and used is generally confidential and unavailable for research. Fractions applied in specific uses are usually unknown, even to those who file registration dossiers. Therefore, the full emission model of Equation 1 could mainly be applied to chemical substances of which 1) the total amounts produced and used, and 2) the fractions of this use, applied in specific uses, are known or assumed, that is, for pharmaceuticals and pesticides. For the REACH substances it was necessary to categorize the 169 spERCs (Supplemental Data, Table SI-1), claimed to be "best possible estimates" (European Chemical Industry Council 2012), into the 12 composite main uses listed in [fig_ref] TABLE 1: Estimated emission fractions [/fig_ref] , and to assign release-category averages to each chemical that, according to our expert judgement, belonged predominantly to this release category.
## Exposure modeling
European Union risk assessment under REACH prescribes that registrants of new chemicals provide an explicit estimation of what environmental exposure concentrations are to be expected after the chemical is brought onto the European Union market. We have adopted this line of reasoning, including the use of terminology that comes with it, particularly of the use of the word "expected." We use this word in the present study in its statistical meaning: the (statistical) expectation of the true, but unknown value that a random variable may take in a specific situation. In using the word "expected," we further follow [bib_ref] Normal species sensitivity distributions in probabilistic ecological risk assessment, Aldenberg [/bib_ref] who named the outcome of the Van Straalen-Aldenberg convolution integral of Equation 2 as "expected risk." In European Union risk assessment, such "predictions" or "expectations" relate to standardized emission/exposure/effect circumstances (termed in the present study "typical European Union water body"), which is based on the 'Unit World' concept that was introduced in the late 1970s by Baughman and Lassiter [bib_ref] Spatial concentration distributions, Mackay [/bib_ref]. Expected exposure concentrations (PECs) of chemicals in a typical European Union water body' (sensu REACH) were calculated from estimated emission rates using the multimedia mass balance model SimpleBoxTreat. Specifically, SimpleBox-Treat4Solutions was constructed as a simplified (spatially and temporally invariable) version of the spatially explicit model for European water bodies as created for the European Union project SOLUTIONS [bib_ref] The SOLUTIONS project: Challenges and responses for present and future emerging pollutants..., Brack [/bib_ref] [bib_ref] Increase coherence, cooperation and cross-compliance of regulations on chemicals and water quality, Munthe [/bib_ref]. It is a combination of the emission estimation model just described with the environmental fate simulation models SimpleBox Ver 4 [bib_ref] SimpleBox 4.0: Improving the model while keeping it simple, Hollander [/bib_ref] and Simple-Treat Ver 4 [bib_ref] Adapting SimpleTreat for simulating behaviour of chemical substances during industrial sewage treatment, Struijs [/bib_ref] [bib_ref] Evaluation of SimpleTreat 4.0: Simulations of pharmaceutical removal in wastewater treatment plant..., Lautz [/bib_ref] , and a further simplification to a 3-compartment (air/water/soil) version as applied in the KnowSEC decision support system developed at the German Federal Environment Agency (UBA;. We used integrated emission + exposure modeling to test the usefulness and plausibility of the outcomes of the emission model.
Using emission estimates (Equation 1), we used Simple-BoxTreat4Solutions to simulate emissions of chemicals at socalled local and regional spatial scales, and calculated expected steady-state concentrations of chemicals in a "generic receiving environment." This offered the opportunity to validate the emission + exposure model results against concentrations reported in other studies, because emission estimates themselves cannot be validated directly. Subsequently, by combination with expected critical effect concentrations, expected toxic pressures (defined in the following section, Toxic pressure modeling) were derived, which offered us the opportunity to validate predicted (mixture) toxic pressures against such data from other studies.
## Toxic pressure modeling
We adopted the method of toxic pressure calculation as described by Van Straalen (2002) and [bib_ref] Normal species sensitivity distributions in probabilistic ecological risk assessment, Aldenberg [/bib_ref] , who quantified "toxic pressure" as the probability that ambient exposure concentrations would exceed concentration levels that are considered to be "riskful" [bib_ref] Probabilistic risk assessment using species sensitivity distributions, Solomon [/bib_ref]. In the present study, we derived toxic pressure as the probability that expected concentrations of one or more chemicals in typical European Union water would be greater than a selected critical effect concentration. We chose to use the hazardous concentration for 50% of the tested species (HC50), based on median effect concentration (EC50) data for tested subsets of species (i.e., HC50-EC50). This toxic pressure parameter empirically relates to impacts on ecological status in aquatic ecosystems [bib_ref] Mixtures of chemicals are important drivers of impacts on ecological status in..., Posthuma [/bib_ref].
Following Van Straalen (2002) and [bib_ref] Normal species sensitivity distributions in probabilistic ecological risk assessment, Aldenberg [/bib_ref] , we derived toxic pressure from the convolution integral of 2 distributions, that is, the probability density function of exposure concentrations (from the exposure modeling) and the cumulative probability function of critical effect concentrations (from [bib_ref] Species sensitivity distributions for use in environmental protection, assessment, and management of..., Posthuma [/bib_ref] , known as the Van Straalen-Aldenberg convolution integral:
[formula] ∫ ∞ ∞ = { > } = ( ) × ( ) − + C pdf zC CDF z dz Toxic pressure Pr w EC50 w EC50(2) [/formula]
with:
[formula] ( ) = − ( ) ( ) z C C sd w log w av logEC50 logEC50(3) [/formula]
and
[formula] ( )= ( )− ( ) ( ) z sd EC50 log EC50 av logEC50 logEC50(4) [/formula]
where ( ) pdf zC w and ( ) CDF zEC50 represent the probability density function of toxicologically standardized exposure concentrations (in water) and the cumulative distribution function of toxicologically standardized acute EC50 values, respectively [fig_ref] FIGURE 1: Graphical illustration of the Van Straalen-Aldenberg convolution integral [/fig_ref].
Values of sd(logEC50), needed in Equation 4, are often not known with great precision for many chemicals, because of insufficient numbers of experimental toxicity data [bib_ref] Species sensitivity distributions for use in environmental protection, assessment, and management of..., Posthuma [/bib_ref]. Following the practice of [bib_ref] Species sensitivity distributions for use in environmental protection, assessment, and management of..., Posthuma [/bib_ref] , we assigned the value 0.7 to across-species standard deviations of all individual chemicals, for the purpose of toxic pressure calculation.
## Case study
The complete model (covering emissions, fate, and mixture impacts) could be completed for 4757 chemical substances, representative of the substances currently used in the European Union as described in [bib_ref] Increase coherence, cooperation and cross-compliance of regulations on chemicals and water quality, Munthe [/bib_ref].
Data on the amounts of substances used were collected from various sources, as follows.
REACH substances. Amounts produced and marketed to regulatory authorities are part of the obligatory submission of registration information. The information is confidential, to be used only for registration purposes (e.g., European Chemicals Agency 2019). We obtained so-called European Union tonnages (defined as = + EUtonnage Production − Imports Exports) from registration dossier data for use as research data from the registration dossiers (submitted until April 2015).
Active ingredients of medicines. Amounts of more than 1000 individual pharmaceuticals sold in Sweden and the United Kingdom were obtained from public sources (combined data from [bib_ref] Pharmaceuticals and personal care products in the environment: What are the big..., Boxall [/bib_ref].
Additional sales data for a lesser numbers of pharmaceuticals, sold in Austria, Switzerland, Germany, France, and the Netherlands, were obtained from various other public reports [bib_ref] Ecotoxicology of human pharmaceuticals-Review, Fent [/bib_ref] ; International Commission for the Protection of the Rhine against Pollution 2009). No more sufficiently detailed or more recent use data were publicly available for our study.
Active ingredients of crop protection products. Amounts of active ingredients of crop protection products are monitored by individual European Union member states and are reported to the European Statistics Agency EUROSTAT. This agency combines the confidential original data into main categories of active ingredients. Original sales data of pesticides could not be made available for the present study. Instead, estimations made by the Joint Research Center of the European Commission by means of the so-called harvested area approach [bib_ref] Integrated assessment of environmental impact of Europe in 2010: Data sources and..., Sala [/bib_ref] were used. Use volumes of over 400 active ingredients of crop protection products, used in different European Union countries, were thus obtained.
Releases to air, water, and soil were estimated according to the procedure described previously (in the Toxic pressure modeling section, Equation 1). Expected exposure concentrations in European Union air, water, and soil were modeled with the exposure model mentioned previously in the Toxic pressure modeling section. Expected toxic pressures, for each of the substances and of all substances together, were calculated with the Van Straalen and Aldenberg's method, as described previously (in the Toxic pressure modeling section, Equations 2-4). Specific steps for the 3 compound groups were needed, as follows:
REACH substances. Fractions released to air, water, and soil in the different life cycle stages, aggregated across the 12 composite main uses of Pharmaceuticals. Pharmaceuticals are assumed to be used entirely according to stage II, use 5 (use in medicine). Based on a publication by [bib_ref] Evaluation of human pharmaceutical emissions and concentrations in Swedish river basins, Lindim [/bib_ref] , who studied releases of 56 pharmaceuticals in Sweden, we have assigned a release fraction of 12% of the human consumption rate, which is assumed to be released entirely to sewage treatment plants.
Crop protection products. Pesticides are assumed to be used entirely according to stage II, use 4 (use in agriculture). Based on the studies of [bib_ref] Integrated assessment of environmental impact of Europe in 2010: Data sources and..., Sala [/bib_ref] , who applied the so-called harvested area approach to estimate releases to air, water, plants, and soil, we assigned the emission fractions listed in [fig_ref] TABLE 2: Numbers and volumes of chemical substances used in Europe, and their expected... [/fig_ref] : 100% release to the environment (of which 15% to air, 1% to water, and 84% to soil), without wastewater treatment.
In the present case study, (expected) steady-state concentrations were modeled for all chemicals (REACH substances, pharmaceuticals, and pesticides). In the real-world situation, steady states are not expected for all chemicals (i.e., particularly not for agriculturally used pesticides). We interpret steady-state solutions of the multimedia model as predictors of the usually uncertain (or unknown, even) space-and timeaverage value of all possible concentrations in the hypothetical European Union water body. Finally, the case study results were compared with data sets for aquatic exposure concentrations and for toxic pressure, to evaluate whether the emission model (which itself can hardly be validated) could provide useful assessment information.
# Results
## Overview
The numbers of chemicals in various subgroups, the amounts of chemicals used and their estimated emission rates, the expected exposure concentrations, and the expected toxic pressures are summarized in [fig_ref] TABLE 2: Numbers and volumes of chemical substances used in Europe, and their expected... [/fig_ref] (for further details, see Supplemental Data, .
European Union-wide emission rates to (continental) air, water, and soil were estimated for 4757 compounds with complete data (3940 REACH chemicals, 397 pharmaceuticals, and 420 pesticides with data for emission, concentration, and impact modeling). European Union-wide use volumes and expected emission masses, exposure concentrations in water, and expected per-chemical toxic pressures on aquatic life in the typical European Union water body all varied by several orders of magnitude. The predicted mixture toxic pressure, expressed as the multisubstance potentially affected fraction (msPAF)-EC50 was estimated as 3.7% in the typical European Union water body, with a dominant contribution of the REACH chemicals (which also represented by far the largest relative mass). The relative contributions of the different chemicals to the mixture toxic pressure at the EC50 level were calculated, and expressed as Pareto ratios for the total mixture and for the mixtures for each studied compound group. Clearly, for each group of chemicals (REACH substances, pharmaceuticals, and pesticides), a relatively small fraction of the chemicals was responsible for a large fraction of the toxic pressure. Similar observations have been made in other fields of science [bib_ref] Power laws and air pollution, Muller [/bib_ref] and have been termed the 80-20 rule, or (more correctly), the Pareto principle (Wikipedia 2020).
## Amounts used and rates of emission
For reasons of confidentiality, European Union tonnages of individual chemical substances cannot be shared for REACH chemicals. Information on amounts of chemicals in trade is shown in [fig_ref] FIGURE 2: Registered amounts of chemical substances used in Europe [/fig_ref] as a histogram. Emission rates were estimated for all chemical substances represented in [fig_ref] FIGURE 2: Registered amounts of chemical substances used in Europe [/fig_ref] and are listed in Supplemental Data, [fig_ref] TABLE 2: Numbers and volumes of chemical substances used in Europe, and their expected... [/fig_ref] REACH substances. According to the REACH registrations per April 2015, a total mass of 3.14 × 10 9 tons of chemical substances is marketed annually in the European Union. European Union tonnages were obtained for 5592 different chemical substances (representing 2.2 × 10 9 tons, or 70% of the European Union total mass). Of these, 3940 substances (2.7 × 10 8 tons or 8.5% of European Union total mass) were relatively simple monoconstituent organic substances (in REACH terminology), of which expected emissions and steadystate concentrations in a typical European Union water body could be predicted by means of the combined emission and exposure model.
Pharmaceuticals. Active ingredients of medicines are produced and used in much lower numbers and masses compared with REACH substances. Use volumes of 397 different pharmaceuticals, representing 2.5 × 10 5 tons (0.01% of European Union total mass) were available, and used in the combined emission and exposure model.
## Pesticides. the same (lower numbers and masses compared
with REACH) also holds for active ingredients of crop protection products. Use volumes for 420 chemicals, representing 1.8 × 10 5 tons (0.006% of European Union total mass) were available, and used in the combined emission and exposure model. Predicted emissions encompass narrower tonnage ranges than use volumes, as shown by comparing [fig_ref] FIGURE 2: Registered amounts of chemical substances used in Europe [/fig_ref] and B. Estimated emission rates span a range of roughly 10 orders of magnitude: from less than 1 kg/yr to over 1 million tons/yr.
## Exposure concentrations
Expected steady-state concentrations for the typical European Union water body are shown in [fig_ref] FIGURE 3: Expected steady-state concentrations of chemical substances in a "typical European Union water... [/fig_ref] , together with the information on species sensitivities (at the level of acute EC50s) as obtained from [bib_ref] Species sensitivity distributions for use in environmental protection, assessment, and management of..., Posthuma [/bib_ref] and shown in [fig_ref] FIGURE 3: Expected steady-state concentrations of chemical substances in a "typical European Union water... [/fig_ref]. Expected steady-state concentrations in the European Union water for the studied 4757 compounds span roughly the same range as emission masses. Apparently, variance in "fate" (as modeled by multimedia environmental fate modeling) contributes little to variance in exposure, compared with the contributions of variances in tonnages and emissions. Expected exposure concentrations range from femtograms/liter to milligrams/liter [fig_ref] FIGURE 3: Expected steady-state concentrations of chemical substances in a "typical European Union water... [/fig_ref]. This is in agreement with earlier observations of [bib_ref] Definition and applications of a versatile chemical pollution footprint methodology, Zijp [/bib_ref] , who predicted similar concentration distributions in river Rhine water, and found these to match available concentration measurements. It is also in agreement with recent findings from the SOLUTIONS project [bib_ref] Increase coherence, cooperation and cross-compliance of regulations on chemicals and water quality, Munthe [/bib_ref] , which reported similar concentration distributions for currently used organic chemicals in various other river catchments. Note the small overlap between the distribution of exposure concentrations in [fig_ref] FIGURE 3: Expected steady-state concentrations of chemical substances in a "typical European Union water... [/fig_ref] and the distribution of critical effect concentrations in [fig_ref] FIGURE 3: Expected steady-state concentrations of chemical substances in a "typical European Union water... [/fig_ref] , which provides a visual indication of the probability of occurrence of harmful concentrations of chemicals.
## Toxic pressures
Expected probabilities that critical effect concentrations of chemicals would be exceeded by concentrations of chemicals in a typical European Union water body appear to be below 5%, in the present study relating to exceedance of the median EC50 of tested species (Table 2 and [fig_ref] FIGURE 4: Distribution of expected toxic pressures [/fig_ref].
Earlier referred to as "ecological risk" (Van Straalen 2002) or "expected risk" [bib_ref] Normal species sensitivity distributions in probabilistic ecological risk assessment, Aldenberg [/bib_ref] , we refer to this outcome as "(mixture) toxic pressure," as suggested by [bib_ref] Mapping risks of heavy metals to birds and mammals using species sensitivity..., Traas [/bib_ref]. Toxic pressure is a metric with which compounds or water bodies can be ranked regarding toxicity. The mixture toxic pressure of the typical European water body at steady state in an exposure assessment that considers approximately 8.5% of the mass of chemicals used in Europe was calculated as 3.7%, that is, the probability that steady-state exposures would exceed the acute EC50 of a median-sensitive species is 3.7%. Estimated toxic pressures at steady state are, logically, lower for separate chemicals, and higher if all chemicals would have been modeled. Most chemicals have exposure concentrations in water that result in negligible "expected risk" at the EC50 level. Toxic pressures at the EC50 level of over 99% of the chemicals tested were below the value of 10 −6 [fig_ref] TABLE 2: Numbers and volumes of chemical substances used in Europe, and their expected... [/fig_ref] : it is less than 0.0001% probable that concentrations greater than the EC50 of a median-sensitive species will be found, which is the same as saying that it is 0.0001% probable to encounter a (A) (B) species for which the EC50 would be exceeded in the typical European Union water body. The outcome of the calculations is that higher toxic pressures at the EC50 level and steady state are to be expected for less than 1% of the chemicals that are currently used in the European Union. It is worth noting that, out of the 4757 chemicals considered, only 23 individual chemicals (17 REACH chemicals and 6 pesticides) contributed substantially to the 3.7% EC50-based mixture toxic pressure found in the present case study. [fig_ref] TABLE 2: Numbers and volumes of chemical substances used in Europe, and their expected... [/fig_ref] lists the associated Pareto ratios found in our study for REACH chemicals (99.6/0.4), pharmaceuticals (99.8/0.2), and pesticides (98/2).
# Discussion
## Overview
The present study is the first to provide an approach to and an analysis of the emissions of chemical pollutants at the European scale, evaluated from calculated steady-state concentrations in a typical European water body for chemicals currently in trade. The results are based on a combination of techniques that are all well founded in the scientific literature and used in risk assessment practices (spERC data multimedia, effect, and mixture models). The steps performed resemble those of regulatory prospective chemical safety assessment (here, e.g., REACH), and expand on that by attempting to consider chemical pollution as a whole, as most recently emphasized by [bib_ref] Toward a global understanding of chemical pollution: A first comprehensive analysis of..., Wang [/bib_ref]. The steps can also be expanded with emission locator mapping and hydrological modeling to predict spatiotemporally explicit compound concentrations and (mixture) toxic pressures, like higher tier assessments reported elsewhere [bib_ref] Species sensitivity distributions for use in environmental protection, assessment, and management of..., Posthuma [/bib_ref] [bib_ref] Increase coherence, cooperation and cross-compliance of regulations on chemicals and water quality, Munthe [/bib_ref]. Outcomes can help in assessing whether "safe use of chemicals" (as meant by the European Union REACH regulation, for example) can be achieved and whether the environment goal of "good ecological status" (as meant by the European Union Water Framework Directive) can be maintained or reached. We specifically aimed at practical utility for both types of assessments. We recognize that each model step, and all required data, can be refined for any individual chemical and region, as has been done by others for specific compounds or compound groups [bib_ref] Europe-wide estuarine export and surface water concentrations of PFOS and PFOA, Lindim [/bib_ref] [bib_ref] Evaluation of human pharmaceutical emissions and concentrations in Swedish river basins, Lindim [/bib_ref] [bib_ref] Aquatic risks from human pharmaceuticals-Modelling temporal trends of carbamazepine and ciprofloxacin at..., Oldenkamp [/bib_ref]. The aim of achieving insights into the chemical pollution problem as a whole called for some simplifying steps, and the use of a relatively simple approach to emission estimation. Nonetheless, the results allow for some key observations and some conclusions on various aspects of our study, as well as some uncertainties that need further improvements.
## Comparing outcomes with regulatory goals
Our case study yielded results of the kind needed for comprehensive assessment of chemical pollution for a region, based on a relevant fraction of chemicals used in a region. Such results allow one to assess the mixture toxic pressure in a typical European water body, and to rank the chemicals contributing most to such pressure. The results [fig_ref] TABLE 2: Numbers and volumes of chemical substances used in Europe, and their expected... [/fig_ref] indicate that the current use of industrial chemicals, plant protection products, biocides, and pharmaceuticals leads to exposure concentrations that could cause significant effects on aquatic ecosystems, because the probability of exceeding the median EC50 of tested species was 3.7% (msPAF-EC50 = 3.7%, for all studied chemicals). This is far higher than the toxic pressure threshold value utilized in chemical safety assessment/chemical, which is defined as msPAF-no-observed-effect concentration = 5% and is used as a science-based threshold to define the PNEC. Studies by [bib_ref] Organic chemicals jeopardize the health of freshwater ecosystems on the continental scale, Malaj [/bib_ref] and [bib_ref] Species sensitivity distributions for use in environmental protection, assessment, and management of..., Posthuma [/bib_ref] confirm that water body-or area-specific exposures of single chemicals and mixtures indeed suggest that a nontoxic environment (European Commission 2014; [bib_ref] Increase coherence, cooperation and cross-compliance of regulations on chemicals and water quality, Munthe [/bib_ref] or toxic-free environment (European Commission 2019) seems far away. Thus, the comprehensive assessment of emissions, fate, and mixture impacts serves to provide the kinds of outputs needed to provide a basis for prospective decision support and prioritization of actions to forward the goal of the zero-pollution policy ambition. Such an assessment is a potentially valuable addition to current practices, which often neglect mixtureseven while mixtures form the common real-life exposure situation in the field . With the outcomes in kind being potentially useful, there is still a need to consider issues such as precision and accuracy, dependent on the context of use of the outcomes.
Accuracy, precision, and uncertainty in toxic pressure calculation Exposure metrics are often used in the practice of chemical safety assessments, in the format of PECs being compared with PNECs, as the PEC/PNEC ratio (European Commission 2006) of individual chemicals. The accuracy of predicted exposures is thus key for decisions to allow chemicals on the market. We acknowledge that emissions cannot be estimated with great certainty, by virtue of basing the modeling on available principles and data, because key information on the uses of chemicals is unavailable especially in REACH registration dossiers. As a result, the precision of the emission estimation is necessarily poor. However, estimated emission rates may be accurate, in the sense that the emission model yields outcomes that are right on average. First, an indirect judgment of emission model accuracy can be made by comparing predicted with measured environmental concentrations. Such an assessment has been made using a much more detailed hydrologic model by [bib_ref] Increase coherence, cooperation and cross-compliance of regulations on chemicals and water quality, Munthe [/bib_ref] , who showed that PECs-obtained with integrated emission + exposure modeling-were on average similar to measured environmental concentrations (MECs) for 226 substance/basin combinations for which both PECs and MECs were available, whereas for 65 and 90% of cases, the under-or overprediction of PECs for individual substances was within 1 and 2 orders of magnitude, respectively.
It should be noted that the MECs themselves may not reflect true (i.e., accurate) exposure due to temporal variability. That is, the MEC for a chemical in a water body can vary widely when determined before or after pesticide spraying or before or after a rain event causing city runoff. The accuracy of the PEC estimation of the present study is further similar to independently obtained results for pharmaceuticals [bib_ref] A high-resolution spatial model to predict exposure to pharmaceuticals in European surface..., Oldenkamp [/bib_ref]. Earlier studies [bib_ref] Definition and applications of a versatile chemical pollution footprint methodology, Zijp [/bib_ref] , and preliminary assessments with Monte Carlo simulations (data not shown) suggested that the PECs for individual chemicals may range over nearly 1 order of magnitude as a result of uncertain emission estimations. Given the many orders of magnitude difference in exposures across chemicals [fig_ref] FIGURE 3: Expected steady-state concentrations of chemical substances in a "typical European Union water... [/fig_ref] , these outcomes show that, although the emission model outcomes are of low precision, they are accurate enough to serve as a starting point for toxic pressure calculation.
Second, it is important to note that uncertain emission estimation does result in the shown uncertainty in exposure assessment, but not in uncertain (mixture) toxic pressures. When calculated from overlap of distributions of exposure concentrations and critical effect concentrations, as in Equation 2, decreased precision (i.e., increased uncertainty) of emission estimation and hence, of expected exposure concentrations, results in increased toxic pressure, rather than in uncertain toxic pressure. This can be seen from [fig_ref] FIGURE 1: Graphical illustration of the Van Straalen-Aldenberg convolution integral [/fig_ref] : greater uncertainty in emissions leads to wider concentration distributions, with enhanced probability of having greater-than-average exposure concentrations. Use of this method of calculation (Equation 2) directly translates uncertainty in emissions into enhanced probability of exceeding the EC50, and thus in greater (mixture) toxic pressure. Although neither Van Straalen (2002) nor [bib_ref] Normal species sensitivity distributions in probabilistic ecological risk assessment, Aldenberg [/bib_ref] have mentioned this factor in their studies, the original procedure of toxic pressure calculation automatically accounts for uncertainty in exposure estimation.
Third, the core aim of environmental protection and management-a nontoxic or toxic-free environment (European Commission 2014, 2019)-is the absence of or negligible net impacts (in the present study, in mixtures). Comparison of mixture toxic pressures derived from PECs or MECs from other studies provides another way to indirectly judge the emission modeling approach. This should be done by acknowledging the Pareto ratios found in the present study and when based on MECs (e.g., [bib_ref] Screening level mixture risk assessment of pharmaceuticals in STP effluents, Backhaus [/bib_ref] [bib_ref] Use of the maximum cumulative ratio as an approach for prioritizing aquatic..., Vallotton [/bib_ref]. That is, considering greater numbers of chemicals does not always result in finding greater (mixture) toxic pressures. When the Pareto principle is applied, and when the 20% most important chemicals have been looked at already, not much additional toxic pressure knowledge is to be expected from considering the remaining 80%.
Fourth, the outcomes of the present study were compared with other toxic pressure assessments [fig_ref] TABLE 3: Comparison between mixture toxic pressure for the "typical European Union water body" [/fig_ref] ; expressed as msPAF-EC50; as fraction, varying between 0 and 1). The present study yielded a mixture toxic pressure for 4757 compounds of 3.7% for a typical European Union water body (based on msPAF-EC50). That compares well with the mixture toxic pressures found for Europe and specific European basins, which are oin the same order of magnitude, or lower when considering the 95th percentile estimate of daily predicted values over 1 yr (i.e., mixture toxic pressures in a basin are higher than the 50th percentile of msPAF-EC50 over sites for 18 of 365 d). Note that accounting for spatiotemporal variation also suggests the presence of lower and higher mixture toxic pressures [fig_ref] TABLE 3: Comparison between mixture toxic pressure for the "typical European Union water body" [/fig_ref] , columns with 5th and 95th percentile of msPAF-EC50 over sites). The outcomes are also in line with a large data set of measured environmental concentrations for the Netherlands, despite the large variability in number of compounds measured/sample (between 1 and 263). In summary, the comparison indicates that the emission model is accurate in helping to obtain the necessary insights into the expected level of the ambient mixture toxic pressure and into compounds likely contributing the least and the most.
Given the findings summarized in [fig_ref] TABLE 3: Comparison between mixture toxic pressure for the "typical European Union water body" [/fig_ref] , it may be concluded that there is indirect evidence that the emission estimation procedure helps to provide insights into the average values of possible real-life emissions and associated effects. That is, despite large uncertainties in the model and the input data, it appears that the accuracy of the emission prediction can yet be considered sufficient for screening-level calculation of mixture toxic pressures of ambient mixtures of chemicals.
## Prioritization of chemicals for management
The key problems of chemical safety assessment policies have always been 1) how to judge "the universe of chemicals" [bib_ref] Toward a global understanding of chemical pollution: A first comprehensive analysis of..., Wang [/bib_ref] , and 2) how to judge sufficiently safe use of all chemicals. The former problem was addressed by introducing prioritization principles, where the potentially worst chemicals are identified, grouped, and tested for hazardous properties first. The present approach addresses both questions, and this resulted in a corroboration of the finding of a suite of other studies in which it was often found that mixture impacts can commonly be attributed to a relatively low number of chemicals (see, e.g., Backhaus and Karlsson 2014 for pharmaceuticals in European coastal waters, Vallotton and Price 2016 for pesticides in the United States, and Posthuma et al. 2016 for a suite for chemicals in Dutch surface waters). Our case study also reconfirms one of the earliest findings of this kind [bib_ref] Estimating the impact of high-production-volume chemicals on remote ecosystems by toxic pressure..., Harbers [/bib_ref] , that generally only small fractions of the chemicals produced, used, and emitted are responsible for the larger part of the direct ecological impacts for a particular situation. The results provide both a signal of insufficient protection (see preceding discussion) as well as a very "sharp" prioritization of chemicals regarding potential effects that occur via direct exposure to species assemblages. As discussed also by [bib_ref] Species sensitivity distributions for use in environmental protection, assessment, and management of..., Posthuma [/bib_ref] , the prioritization should as yet not be interpreted as absolute ranking of the 4757 compounds. Instead, as derived from preliminary Monte Carlo simulations (data not shown), the absolute rank order of chemicals may vary in this respect, but chemicals with toxic pressures greater than 10 −6 appear to remain above this negligibility threshold in all ranking simulations under uncertainty; chemicals with toxic pressures less than 10 −6 appear to be of low priority, regardless of uncertainty in emission estimations.
The outcome of our study clearly illustrates the Pareto principle, namely, that few causes often explain a large part of the result. In toxic pressure calculation, this goes well beyond the classical Pareto 80-20 rule. We observed Pareto ratios [fig_ref] TABLE 2: Numbers and volumes of chemical substances used in Europe, and their expected... [/fig_ref] that were systematically larger: for pesticides a 98-2 rule was found, whereas for the "ordinary" (monoconstituent) REACH chemicals and the pharmaceuticals, even greater ratios were found. This pattern should be conceptually expected from the probability calculation that underlies the toxic pressure derivation by the Van Straalen-Aldenberg integral (overlaps of tails of distributions). This statistical reason has not been recognized so far. However, many studies report such findings for a wide array of contexts [bib_ref] Power laws, Pareto distributions and Zipf's law, Newman [/bib_ref] [bib_ref] Power laws and air pollution, Muller [/bib_ref] [bib_ref] Scale: The Universal Laws of Growth, Innovation, Sustainability, and the Pace of..., West [/bib_ref] , and-in hindsight-it would have been surprising to find nonskewed outcomes. The most unlikely outcome would be that mixtures in the field would be equitoxic, that is: composed such that all chemicals contribute equally to the mixture effect everywhere. For the present results and accuracy level, we suggest using the outcomes in practice by considering 3 classes: even despite some uncertainties, the method can help us to identify a class of chemicals unlikely to pose harm in the European Union water body, a class for which this is possible (and depending on circumstances), and a class that likely poses harm. Such assessments have also been made, based on different degrees of model focus (e.g., emissions, hazard) and approaches, and for other endpoints [bib_ref] European Union regulators and industry agree on improving specific environmental release categories:..., Ahrens [/bib_ref] [bib_ref] Development of a novel scoring system for identifying emerging chemical risks in..., Oltmanns [/bib_ref] [bib_ref] Using REACH registration data to rank the environmental emission potential of persistent..., Schulze [/bib_ref] [bib_ref] Species sensitivity distributions for use in environmental protection, assessment, and management of..., Posthuma [/bib_ref] , given that prioritization is key to managing more than 350 000 chemicals [bib_ref] Toward a global understanding of chemical pollution: A first comprehensive analysis of..., Wang [/bib_ref].
## Suitability for regulatory decision-making
We found that the kinds of outputs of the emission model (after combination with exposure and effect models) are suitable for 2 regulatory contexts, namely, chemical safety assessment and environmental quality assessment and management. We conclude from the present study that the emission estimation procedure is sufficiently accurate and precise for use in screening-level chemical safety assessments of substances and their mixtures (as required in REACH and Life Cycle Impact Assessment) and for screening-level assessment of mixture toxic pressures at local spatial scales (as needed for assessment of "good ecological status" under the European Union Water Framework Directive). Refined approaches may be developed for chemical safety assessment of the assemblage of substances produced and used, or they have already been developed for environmental quality assessments , such that more precise information can be generated for situations where the screening-level assessment suggest non-negligible exposures, risks, or impacts. The main finding of the present study is that it has demonstrated how uncertain emission estimates for a large number of chemicals can be used to assess and understand expected impacts of expected exposure concentrations of currently used chemicals, with emphasis on the novel insights in the consequences of mixtures. We emphasize that toxic pressure calculation owes this to consequent application of the Van Straalen-Aldenberg convolution integral, and that the final interpretation of our results can be understood from 2 considerations: 1) Valerie Forbes' statementthat "although we may be right on average, we may be wrong most of the time"; and 2) George E.P. Box's teaching [bib_ref] Robustness in the strategy of scientific model building, Box [/bib_ref] that "all models are wrong, but some are useful."
# Conclusions
We developed a new model to estimate emission rates ofin principle-all chemicals in commerce. We found that application of the model to the chemicals typically used in Europe resulted in sufficient precision and accuracy to derive screening-level statements on expected impacts of the use of these chemicals on aquatic ecosystems and on ranking the relative contributions of individual chemicals to the net expected impacts. Our study demonstrates how, for the purpose of environmental risk assessment and management of (mixtures of) chemicals, one can afford to be "wrong" (not precise) most of the time, provided that one is "right" (accurate) on average. The outcomes are useful for screening-level chemical safety assessment (REACH) and water quality assessment (Water Framework Directive) purposes at the European scale. The results provide information to help in prioritizing chemicals for chemical safety policies, and for water bodies and chemicals within water bodies for water quality management. Acknowledgment-The present study was supported by, and prepared as result of, the SOLUTIONS project (European Union's Seventh Framework Programme for research, technological development and demonstration under grant agreement 603437). Contributions of the SOLUTIONS research team, and especially the SOLUTIONS-modelers team, have been indispensable to reach the present results. We thank K. Kramer for critical reading and useful comments. Special thanks are due to J. van Gils, who has been leading the integrated SOLUTIONS modeling, and who ran the spatiotemporalspecific model analyses. The data source (Registration, Evaluation, Authorisation and Restriction of Chemicals [REACH]chemicals: European Chemicals Agency) can be found at http:// echa.europa.eu/. Further data providers are gratefully acknowledged. The comments of 2 SOLUTIONS reviewers on an earlier version of this manuscript (I. Cousins and M. Rahmberg) are gratefully acknowledged.
Disclaimer-The present study was performed under the strict confidentiality agreement on production and import/export masses between the Dutch competent authority for the European Union Registration, Evaluation, Authorisation and Restriction of Chemicals regulation and the Dutch National Institute for Public Health and the Environment (RIVM; https:// echa.europa.eu/legal-notice). This agreement states, among other factors, that "[. .] the replication, in whole or in substantial part, of the ECHA databases is prohibited, unless ECHA's prior written permission is given. All requests shall be submitted to ECHA through the information request forms on ECHA's contact page."
Author Contributions Statement-D. van de Meent, D. de Zwart, and L. Posthuma designed the study, D. van de Meent and D. de Zwart ran the data collection and analyses, D. van de Meent wrote the article, D. de Zwart, L. Posthuma checked the article, and all co-authors finalized the manuscript.
Data Availability Statement-Data, associated metadata, and calculation tools are available from the corresponding author ([email protected]).
[fig] FIGURE 1: Graphical illustration of the Van Straalen-Aldenberg convolution integral (Equation 2). Probability of exceedance of critical effect concentrations in water is obtained by evaluating the product of the probability density function of exposure concentrations (dotted line) and the cumulative distribution function of critical effect concentrations (black line) at all possible values of the standardized concentration z. [/fig]
[fig] FIGURE 2: Registered amounts of chemical substances used in Europe (A) and estimated amounts of chemicals emitted to the environment (B). Black = pharmaceuticals; gray = pesticides; white = monoconstituent organics. Use volumes and releases range over 12 orders of magnitude: from as little as kg/yr (pharmaceuticals) to over a million tons/yr (monoconstituent organics). [/fig]
[fig] FIGURE 3: Expected steady-state concentrations of chemical substances in a "typical European Union water body," as defined in European Chemicals Agency (2016; A, present study), compared with the average aquatic log median effect concentration (EC50; acute) for the studied chemicals (B, fromPosthuma et al. 2019b). Note the small overlap between the 2 types of distributions, as inFigure 1. Black = pharmaceuticals; gray = pesticides; white = monoconstituent organics. [/fig]
[fig] FIGURE 4: Distribution of expected toxic pressures (TPs) of chemicals used in Europe, calculated from the overlap of the exposure and (median effect concentration [EC50]) effect distributions established for each chemical according to Figure 3. Black = pharmaceuticals; gray = pesticides; white = monoconstituent organics. msPAF = multisubstance potentially affected fraction. [/fig]
[fig] Supplemental: Data-The Data are available on the Wiley Online Library at https://doi.org/10.1002/etc.4801. [/fig]
[table] TABLE 1: Estimated emission fractions (%) in the various uses and life cycle stages of chemicals, based on published and unpublished reports [/table]
[table] TABLE 2: Numbers and volumes of chemical substances used in Europe, and their expected concentrations and per-chemical and mixture toxic pressures in a "typical European Union water body" at steady state Cw = predicted exposure concentration in water at steady state; EC = effect concentration (here median of the EC50s of the tested species); TP = toxic pressure (multisubstance potentially affected fraction [msPAF]-EC50); REACH = Registration, Evaluation, Authorisation and Restriction of Chemicals. [/table]
[table] TABLE 3: Comparison between mixture toxic pressure for the "typical European Union water body" (present study) and mixture toxic pressures (msPAF-EC50) reported for large numbers of water bodies in major catchments in Europe (PEC-based) and the same for a set of monitoring data for Dutch surface waters (MEC-based) PEC = predicted environmental concentration; MEC = measured environmental concentration; msPAF = multisubstance potentially affected fraction; P5 = 5th percentile; P50 = 50th percentile; P95 = 95th percentile. [/table]
|
Two Cases of Pituitary Stalk Interruption Syndrome in Syrian Children
Pituitary stalk interruption syndrome (PSIS) is an extremely rare cause of growth failure and delayed puberty. It can be diagnosed by magnetic resonance imaging (MRI) of the hypothalamus and pituitary gland, showing an ectopic or absent posterior pituitary, an absent or interrupted pituitary stalk, or small anterior pituitary, in combination with growth hormone or other pituitary hormone deficiencies. e exact etiology of PSIS is unknown. In this article, we describe two cases of PSIS in Syria which are, as far as we know, the first published cases.
# Introduction
Growth failure is one of the most common reasons for referral to the endocrine clinic. Although the most frequent causes of short stature are constitutional and genetic, endocrine disorders, especially growth hormone (GH) deficiency, should always be considered. Pituitary stalk interruption syndrome (PSIS, ORPHA 95496) is an extremely rare disorder, but its exact prevalence is still unknown. is syndrome is a congenital anomaly of the pituitary gland characterized by GH deficiency (with or without other pituitary hormone deficiencies) along with radiological features of a thin or interrupted pituitary stalk, an ectopic or absent posterior pituitary, or a hypoplastic or absent anterior pituitary.
In this article, we describe two cases of PSIS that were diagnosed in boys at different ages from Syria.
## Case 1
A 7-year-old boy was referred to the endocrine clinic for short stature. He was born at full term without complications. His birth weight was 3 kg, and his family history was unremarkable. During childhood, he had all his vaccinations and had not been hospitalized.
On examination, he seemed well without any abnormal facial features, and he was 19 kg in weight and 101 cm in height (body mass index (BMI) 18.62 kg/m 2 ). Compared to his midparental predicted length (119 cm), he was 3.6 standard deviation (SD) points below expected. Heart and breath sounds were normal and so was the abdominal examination. Testicular volume was 3 ml each, measured by using a Prader orchidometer. Pubic hair distribution and penile size (3 by 1 cm) were both consistent with Tanner stage 1.
Routine blood panel and urine and stool tests were within normal range, except for mild anemia. Plain films of the left wrist were ordered and were compatible with a bone age of a three-year-old (>2 SD below chronological age), as shown in. A GH stimulation test conducted with 75 μcg of clonidine was consistent with GH deficiency. Other pituitary hormones were also evaluated and revealed central hypothyroidism and adrenal insufficiency
## Case 2
A 17-year-old boy presented to the endocrine clinic for short stature as well as absence of secondary sexual characteristics.
He was born full term in a breech position without complications and had a birth weight of 2.3 kg. No hypoglycemia or other complications were reported during childhood. On examination, he had no abnormal facial features and looked well. He weighed 43 kg and was 150 cm in height (BMI 19.11 kg/m 2 ). Compared to his midparental predicted length (169 cm), he was 3 SD points below expected. Auscultation of the heart and lungs did not reveal any abnormalities. Abdominal exam was normal. Testicular volume was 4 ml each, measured by using a Prader orchidometer. Pubic hair was consistent with Tanner stage 1. Penile size was 2 by 1 cm, consistent with micropenis. No gynecomastia was noted.
Routine blood panel and urine and stool tests were normal. A plain film X-ray of the left wrist was consistent with a bone age of 13 years. Plain film X-ray of the knee showed open epiphyseal plates. Insulin-like growth factor 1 (IGF1) was low. Gonadal hormone assessment and GH stimulation test, conducted with 150 μcg clonidine after sex hormone priming with 100 mg IM testosterone, showed GH deficiency as well as hypogonadotropic hypogonadism. Other pituitary hormones were also evaluated and revealed central hypothyroidism and adrenal insufficiency. Pituitary MRI was performed (Figures 2(c) and 2(d)) and again confirmed the diagnosis of pituitary stalk interruption syndrome (PSIS). Results of echocardiography, kidney ultrasound, and ophthalmic evaluation were normal for both patients.
## Patients follow-up
Our first patient was treated with hydrocortisone 15 mg daily and then levothyroxine 50 μcg daily, followed by recombinant GH 1.7 IU SC daily. Within three months, he became euthyroid and gained 3 kg weight and 3 cm height.
Our second patient was treated with hydrocortisone 15 mg daily, levothyroxine 75 mcg daily, and testosterone 100 mg monthly. He was later referred to a specialized center for recombinant GH replacement.
# Discussion
Pituitary stalk interruption syndrome (PSIS) was initially described and defined by . PSIS is diagnosed primarily based on the radiologic findings mentioned above along with permanent GH deficiency and one or more other pituitary hormone deficiencies. e exact prevalence of this syndrome remains unknown. Less than 1000 cases were reported in the literature until 2010. Its prevalence as a cause of GH deficiency is estimated to be around 4%. e age at diagnosis differs according to the severity of the hormone deficiency. When PSIS presents at birth, hypoglycemia and failure to thrive are the most common symptoms. In childhood, growth retardation is usually the presenting complaint, whereas delayed puberty is usually the chief complaint when it manifests in adolescence and early adulthood. Although these two patients had multiple hormone deficiencies, diagnosis was established relatively later in life.
Breech delivery, cesarean section, and neonatal hypoxemia were suggested as potential causes of this syndrome. However, many case series found an antenatal origin. Genetic mutations were also detected in familial cases, such as HESX1, LHX4, SOX3, PROKR2, and OTX2 genes. Of note, genetic mutations are thought to contribute to less than 5% of PSIS cases. Zwaveling-Soonawala et al. found additional suggested genes for isolated PSIS (DCHS1, ROBO2, CCDC88C, and KIF14) and one for syndromic PSIS (KAT6A). Presence of many probable pathogenic genes and heritage from normal parents are implicative of polygenic etiology of sporadic cases of PSIS. Similar findings were found in Guo et al.'s study in which a group of gene mutations was identified in 92% of the patients and genes NCOR2, ZIC2, and NKD2 were found to be mostly associated with Notch, Shh, and Wnt signaling pathways, respectively. While our second patient had a breech delivery, our first patient underwent uncomplicated vaginal delivery and had a negative family history. e presence of multiple hormone deficiency suggests a genetic etiology. However, due to lack of availability of genetic testing assays in our country, this was not performed.
In most published series, PSIS exhibited a male predominance, similar to what is observed in our cases. e underlying reason behind this phenomenon is unknown, though there is a selection bias in such patients since boys, more than girls, are brought by parents for growth assessment.
TSH deficiency was the most common deficit (79.5%) alongside GH deficiency in PSIS patients, and ACTH deficiency also was common (67.5%), followed by FSH/LH deficiency in 65.1% of patients.
Our two cases had GH, TSH, and ACTH deficiency. LH/ FSH deficiency may be detected in minipuberty or discovered later in peripuberty. erefore, gonadal assessment should be conducted regularly in patients with PSIS. While hypogonadism was detected only in the second case, gonadal axis was not assessed in the first patient given his young age. e incidence of extrapituitary manifestations in PSIS is high. Midline defects affect mainly the brain and eyes. Extracerebral abnormalities also include defects of the heart, skin, and extremities, which highlights the importance of regular examinations and cardiac, ophthalmologic, and cerebral screening. e lifelong prognosis depends on the time of diagnosis and early treatment of hormonal deficiency.
Close monitoring of a child's growth over time helps in early detection of growth failure and early treatment of the underlying causes. Rare etiologies such as PSIS should be kept in mind especially in multiple hormonal deficiencies, as it carries good long-term prognosis if diagnosed early on and treated adequately.
Finally, these two cases are, as far as we know, the first published cases from Syria.
## Consent
Informed consent was obtained from the patients' parents.
## Conflicts of interest
e authors declare that they have no conflicts of interest. |
Intrinsic refractive index matched 3D dSTORM with two objectives: Comparison of detection techniques
We have built a setup for 3D single molecule localisation microscopy (SMLM) where a very high resolution is achieved by, firstly, the use of two objectives instead of one and, secondly, minimizing optical aberrations by refractive index matching with a glycerol-water mixture as immersion medium in conjunction with glycerol-immersion objectives. Multiple optical paths of the microscope allow to switch between astigmatic and interferometric localisation along the optical axis, thus enabling a direct comparison of the performance of these localisation methods.
Single molecule localisation microscopy techniques as stochastic optical reconstruction microscopy (STORM) [bib_ref] Sub-diffraction-limit imaging by stochastic optical reconstruction microscopy (STORM), Rust [/bib_ref] , direct stochastic optical reconstruction microscopy (dSTORM) 2 , photoactivated localisation microscopy (PALM), fluorescence photoactivation localisation microscopy (FPALM) [bib_ref] Ultra-high resolution imaging by fluorescence photoactivation localization microscopy, Hess [/bib_ref] and ground state depletion followed by individual molecule return (GSDIM) 5 elegantly circumvent the diffraction limit of optical microscopy by imaging single fluorescent molecules which stochastically switch between a fluorescent and a non-fluorescent state. Spatial separation of the fluorescent emitters allows to obtain the dye molecule coordinates, and therefore to resolve the stained structures, with a precision at least ten-fold better than the optical resolution. Originally, this approach was limited to 2D imaging, but subsequently various techniques have been developed to also obtain the axial (z) coordinate and reconstruct 3D images. They all involve some means to make the single molecule images more z-dependent, for example by inserting a cylindrical lens into the detection beam path which adds astigmatic distortion to the images [bib_ref] Three-dimensional super-resolution imaging by stochastic optical reconstruction microscopy, Huang [/bib_ref] , imaging two distinct axial planes simultaneously 7 or by capturing the fluorescent light with two objectives and bringing it to interference [bib_ref] Interferometric fluorescent super-resolution microscopy resolves 3D cellular ultrastructure, Shtengel [/bib_ref] [bib_ref] Two-color nanoscopy of three-dimensional volumes by 4Pi detection of stochastically switched fluorophores, Aquino [/bib_ref] [bib_ref] Single-Molecule Real-Time 3D Imaging of the Transcription Cycle by Modulation Interferometry, Wang [/bib_ref] , which causes z-dependent intensity variations. The localisation precisions achievable with these methods were investigated theoretically [bib_ref] Isotropic 3D Nanoscopy based on single emitter switching, Von Middendorff [/bib_ref] and a direct experimental comparison of biplane and astigmatic detection was performed for a microscope with a single water-immersion objective [bib_ref] Experimental characterization of 3D localization techniques for particle-tracking and super-resolution microscopy, Mlodzianoski [/bib_ref] , but not for two-objective schemes such as interferometric detection.
The localisation precision surpasses the resolution limit, but nevertheless depends on the size of the microscope's point spread function (PSF) and therefore its numerical aperture (NA). However, objective lenses with high NA which are required for good optical resolution mostly use oil as immersion medium due to the resulting high NA, whereas biological samples usually have to be imaged in aqueous buffers. This leads to a refractive index mismatch between cover glass and watery sample, which causes both aberrations and loss of signal [bib_ref] Aberration correction for confocal imaging in refractive-index-mismatched media, Booth [/bib_ref]. For single molecule localisation, this is unfavourable as both effects degrade the localisation precision. Furthermore, wavefront deformations will distort the interference pattern when interferometric detection is used.
One approach to circumvent the degrading effects of refractive index matching is to include a wavefront correction mechanism into the system. This was successfully implemented for astigmatic detection [bib_ref] PSF shaping using adaptive optics for three-dimensional single-molecule super-resolution imaging and tracking, Izeddin [/bib_ref] and interferometric detection [bib_ref] Ultra-High Resolution 3D Imaging of Whole Cells, Huang [/bib_ref]. However, a more straightforward approach is to prevent the refractive index mismatch and the associated degrading effects by matching the refractive indices of immersion medium, cover glass and sample. For example, when water-immersion objectives were employed for 3D STORM with astigmatic detection [bib_ref] Resolution doubling in 3D-STORM imaging through improved buffers, Olivier [/bib_ref] , they yielded better results than oil-immersion objectives. But due to the lower refractive index of water, water-immersion objectives have a lower NA (typically 1.2), which increases the size of the PSF and reduces the number of collected photons, effects that are disadvantageous for the localisation precision. For oil-immersion objectives with their high NA of 1.4 or more, a refractive index match may be achieved by embedding the sample in 2,2'-thiodiethanol (TDE) instead of aqueous buffers [bib_ref] 2′-thiodiethanol: a new water soluble mounting medium for high resolution optical microscopy, Staudt [/bib_ref] [bib_ref] A simple and cost-effective setup for super-resolution localization microscopy, Ma [/bib_ref]. But this method has the disadvantage that for an exact refractive index matching, the aqueous medium has to be replaced stepwise with TDE to avoid artefacts due to osmosis. Glycerol-immersion objectives may be another alternative. With NA = 1.35, their numerical aperture is higher than those of water-immersion objectives, but not as high as with oil immersion. Their advantage is that embedding biological samples in glycerol-containing media is easier feasible than in TDE. No stepwise procedure is required, it is very inert, and many mounting media for fluorescence microscopy contain glycerol anyway. The required refractive index for glycerol immersion objectives is slightly higher than for silicone oil immersion objectives, which have recently been used to enable refractive index matching for live cell confocal spinning disc microscopy.
Here we have investigated the performance of glycerol immersion objectives 20 for 3D localisation microscopy with astigmatic and interferometric detection. These objectives were already successfully employed for superresolution microscopy, but as far as we know they have not yet been used for (d)STORM. Testing different detection methods on the same setup with identical optics enables a direct comparison of their localisation precision under these conditions.
Interferometric detection requires a setup with two objectives. For astigmatic detection, this is advantageous as well [bib_ref] Dual-objective STORM reveals three-dimensional filament organization in the actin cytoskeleton, Xu [/bib_ref] [bib_ref] Actin, spectrin, and associated proteins form a periodic cytoskeletal structure in axons, Xu [/bib_ref] , because twice as many photons are detected, which increases the localisation precision by 2. For such a scheme with two objectives, refractive index matching is especially beneficial. With a mismatch between cover glass and sample buffer, aberrations are more prominent deep inside the sample. Accordingly, in a microscope with two objectives the localisation precision would be degraded for one of the objectives because there is a layer of several micrometers of buffer between sample and objective. However when the refractive index of the buffer is adjusted, the thickness of the buffer film has no impact on the image quality.
# Results
Degradation of localisation precision due to refractive index mismatch. To demonstrate the effect of refractive index mismatch, we performed simulations of the detection PSF generated by an oil immersion objective. As detection PSF model, we employed the distribution of the electromagnetic field generated by a radiating dipole at the focus of a microscope with infinity-corrected objectives [bib_ref] Theoretical study of detection of a dipole emitter through an objective with..., Enderlein [/bib_ref] , which treats the electromagnetic fields as vectors, not as scalar quantities.
A refractive index mismatch between cover glass and sample causes an optical path length offset δ which depends on the refractive indices of cover glass and immersion medium n 1 , on the refractive index of the sample n s and on the distance between focus and cover glass d s
[formula] 13 , d n n [ cos( ) cos( )] (1) s s s 1 1 δ θ θ = − [/formula]
where θ 1 and θ s are the polar angles in the immersion medium and the sample, respectively. This path length difference leads to a phase offset which has to be included into the calculation of the electromagnetic fields. The deflection of the electromagnetic field vectors due to refraction at the glass-sample interface was neglected, as well as the loss of intensity. For a more accurate investigation of the impact of refractive index mismatch on localisation precision, these effects would have to be included; the present example only serves to estimate the influence of the additional optical path length. The resulting distributions of the intensity I without aberration and with d s = 1 μm and d s = 10 μm are shown in [fig_ref] Figure 1: Impact of refractive index mismatch on the PSF, calculated for an oil... [/fig_ref]. They were calculated for an oil-immersion objective with n 1 = 1.52 and an aqueous sample with n s = 1.33. The half-aperture angle α was set to α = 60°, corresponding to NA = 1.32, because larger angles would require taking into account the deflection of the field vectors. From these intensity distributions, we calculated the theoretical limits of the lateral (xy) localisation precision according to,
[formula] ∫ ∫ μ μ = ∂ ∂ −∞ ∞ −∞ ∞ − [/formula]
where a is the quantity to be estimated from the measurement (in this case, the fluorescent molecule's x position), N the number of detected photons and μ the intensity distribution according to the PSF model normalized so that x y a x y ( , ; ) d d 1
[formula] ∫ ∫ μ = −∞ ∞ −∞ ∞ . [/formula]
All computations were carried out in MATLAB ® (The MathWorks, Inc., Natick, Massachusetts, USA). For a fluorescent Alexa568 molecule emitting N = 2826 photons [bib_ref] Evaluation of fluorophores for optimal performance in localization-based super-resolution imaging, Dempsey [/bib_ref] , with a hypothetical background of N/1000 photons per camera pixel of size 160 nm × 160 nm in the object space, we found that the mean of u x on an axial range of 0.5 μm around the focus was u x = 3.1 nm without mismatch (n s = n 1 = 1.52). When a refractive index mismatch was assumed (n s = 1.33, n 1 = 1.52), the limit of localisation precision was u x = 3.5 nm at a depth of d s = 1 μm in the sample and u x = 4.2 nm at d s = 10 μm depth. Accordingly, in this (somewhat simplified) model the localisation precision is degraded due to the mismatch by 13% and 37%, respectively. This degradation can be avoided by matching the sample buffer's refractive index to that of cover glass and immersion medium, as was done in this work. For this, we used glycerol immersion objectives and fused silica cover glasses with n = 1.46 instead of oil immersion objectives and standard cover glasses with n = 1.52. These have an half-aperture angle of α = 68.5°, which is close to the half-aperture angle α = 67.3° of typical oil immersion objectives with NA = 1.40. Accordingly, the numbers of detected photons should be similar for both types of objectives as they cover nearly the same solid angle.
A more detailed discussion of the impact of refractive index mismatch and NA on the localisation precision and on the limitations of our model can be found in the Supplementary Information, as well as additional simulations.
However, with the index matching as described here, a residual refractive index mismatch may still arise between buffer and the cytosol or the nucleus of cells when biological samples are imaged. Accordingly, the approach presented here cannot completely prevent aberrations arising from refractive index inhomogeneities, it merely eliminates a major source of such aberrations.
## Comparison of localisation precisions.
To compare the performance of astigmatic and interferometric detection, localisation precisions for the x, y (lateral plane) and z (optical axis) direction were estimated for various z positions by imaging fluorescent beads.
To measure the localisation precision, a fluorescent bead was moved along the optical axis and at each z position, 20 images were taken. For each z step, the x, y and z positions of the bead and their standard deviations were estimated. Results for z steps where the bead had been successfully localised less than ten times were discarded in order to obtain the standard deviation from a sample of sufficient size. When these beads were illuminated with the 561 nm laser, with the respective AOTF transmission set to very low values of about 1.2%, approximately 17000 photons per frame were detected, with a mean background of 1.4 photons per pixel. The electron multiplying process in the EMCCDs leads to an increase in noise of about 2 which corresponds to cutting the photon number in half [bib_ref] The noise performance of electron multiplying charge-coupled devices, Robbins [/bib_ref]. This effect was included in these calculations.
The relative error of the standard deviation is assumed as 1/(2 × 10) 1/2 ≈ 0.2, the expected uncertainty of the standard deviation for a sample of ten values. As described above, the sample size may range between ten and twenty, but here an upper bound for the error was estimated.
The experimental results were compared to the theoretical limits of localisation precision u x,y,z according to Eq. (2).
When astigmatic detection is employed, the cylindrical lens causes an offset of the focal length between the x and y direction. For our system where the focal length of the cylindrical lenses is 300 mm and their distance to the tube lenses 160 mm, this offset is about 700 nm in the sample space, as may be derived with geometrical optics. This offset causes an asymmetric, quadratic wavefront aberration 11,28 which has to be included into the calculation of the electrical fields arriving at the detectors.
For interferometric detection, the total electrical field arriving at the cameras was computed as the sum of the fields from each interferometer arm with their respective phases [bib_ref] Interferometric fluorescent super-resolution microscopy resolves 3D cellular ultrastructure, Shtengel [/bib_ref] [bib_ref] Two-color nanoscopy of three-dimensional volumes by 4Pi detection of stochastically switched fluorophores, Aquino [/bib_ref] [bib_ref] Isotropic 3D Nanoscopy based on single emitter switching, Von Middendorff [/bib_ref]. The measured localisation precisions are shown in [fig_ref] Figure 2: Localisation precisions u against axial position z with TransFluoSpheres ® for x,... [/fig_ref] , as well as the theoretical limits of localisation precision obtained from simulations of the PSF. The focus position is at z = 0.
When astigmatic detection was employed, localisation precisions u x = (2.0 ± 0.1) nm, u y = (2.8 ± 0.2) nm and u z = (2.6 ± 0.2) nm were measured for z ∈ (−250 nm, 250 nm). Here the theoretical limits were u x = 2.5 nm, u y = 1.3 nm and u z = 2.9 nm. Due to the astigmatic aberration, the x and y localisation precisions are slightly degraded and the value for y lies above the theoretical limit, but on the other hand the localisation precision is nearly isotropic.
In contrast to interferometric detection, the particle positions are determined independently from the data of both cameras when astigmatic detection is used. This allows to compare the localisation precisions of the two-objective-scheme to the case where the images are obtained by only one objective and detector. For classic astigmatic localisation with a single objective lens, the measured localisation precisions were u x = (2.6 ± 0.1) nm, u y = (3.7 ± 0.2) nm and u z = (3.7 ± 0.2) nm. Accordingly, the second objective enhanced the localisation precision by a factor of 1.3, 1.3 and 1.5, respectively. The mean enhancement is 1.4 which agrees well with the predicted value 2 .
The observed values for interferometric detection were u x = (1.2 ± 0.1) nm, u y = (1.4 ± 0.1) nm and u z = (1.2 ± 0.1) nm. The precisions lie slightly above the theoretical limits u x = 1.1 nm, u y = 1.1 nm and u z = 0.5 nm, but reach lower values than astigmatic detection. The deviation of u z from the theoretical value seems to be mainly caused by a reduced modulation depth of the interference pattern. Loss of modulation depth is partly SCIentIFIC RepoRts | (2018) 8:13343 | DOI:10.1038/s41598-018-31595-z induced by polarization effects, which can be avoided using a polarizing beamsplitter 9 , at the cost of necessitating four detection channels instead of three and an additional Babinet-Soleil compensator. Another contribution is reduced coherence of the detected light. Due to its spectral width and low coherence length, coherent detection of fluorescent light turned out to be a major challenge, which makes this detection method susceptible to imperfect adjustment and flaws in optical components, rendering it the technically most demanding 3D localisation modality.
It is noteworthy that interferometric detection reaches better results for all three dimensions. In astigmatic detection, the astigmatic aberration caused by the cylindrical lens slightly degrades the x and y localisation precision. In our microscope, we employed a cylindrical lens with a short focal length in order to obtain a strong astigmatism, enabling good z localisation precision at the cost of x and y localisation precision. In contrast, the x and y localisation precision is not compromised in interferometric detection.
To compare the 3D resolving power of these two detection schemes, we calculated a localisation volume V
[formula] with V u u u x y z 4 3 = π [/formula]
, the volume of a rotational ellipsoid, where the semi-axes u x,y,z are the mean values for z ∈ (−250 nm, 250 nm). The resulting volumes were (60 ± 10) nm 3 for astigmatic detection and (8 ± 2) nm 3 for interferometric detection. This underlines that interferometric detection reaches by far the best 3D localisation precision.
A measurement with less bright emitters, which emit photon numbers comparable to bright dSTORM dyes, can be found in the supporting information.
Resolving power for stained microtubules. To investigate the performance of the setup in dSTORM experiments, microtubules of 3T3 cells were stained with two different commonly used dSTORM dyes, Alexa Fluor 647 (Alexa647) for the red spectral range and Alexa Fluor 568 (Alexa568) for the orange spectral range. As sub-resolution structures of homogeneous, well-known size, microtubules are ideally suited to test superresolution microscopy setups [bib_ref] Three-dimensional super-resolution imaging by stochastic optical reconstruction microscopy, Huang [/bib_ref].
The glycerol which was added to the dSTORM imaging buffer in order to enable refractive index matching did not markedly alter the number of detected photons and the time the molecules spent in the fluorescing state (see supporting information), therefore it was possible to reconstruct superresolved 3D images of microtubules for astigmatic and interferometric detection [fig_ref] Figure 3: dSTORM images of stained microtubules [/fig_ref]. Partly, the microtubules appear not homogeneously stained but dotted. Comparison with 2D STED images (see Supplementary Information) of similarly stained samples showed discontinuous structures as well, which indicates that this is mainly not an artefact of single molecule localisation but a result of rather sparse staining; the distance between dye molecules is larger than the localisation precision. Of course, insufficient detection of dye molecules during dSTORM imaging and data analysis may further contribute to a sparse number of localisations.
Axial reconstructions of the microtubules were created by projecting localisations onto a plane perpendicular to the microtubule [fig_ref] Figure 3: dSTORM images of stained microtubules [/fig_ref] ; the mean depth of these projections was about 200 nm. In these axial slices, astigmatic as well as interferometric detection resolve the layer of dye molecules surrounding the microtubules as hollow tubes (see intensity profiles in the Supporting Information). The cavity inside these tubes corresponds well to the outer diameter of microtubules, which is 25 nm [bib_ref] Cytoplasmic microtubular images in glutaraldehyde-fixed tissue culture cells by electron microscopy and..., Weber [/bib_ref].
In the interferometric reconstructions, weaker ghost images of the structures appear in some regions at a distance of about half the wavelength. These artefacts are caused by false assignments of z positions due to the similarity of the interferometric PSFs at distances of half the wavelength. Their number was low enough not to impair recognisability of the tubulin strands.
The dye Alexa568 is less bright than Alexa647 and therefore features inferior localisation precision. Accordingly, the ring-shaped structures are less clearly discernible for Alexa568 than for Alexa647 when astigmatic detection was used, and could not be resolved without ambiguity. In contrast, for interferometric detection the localisation precision was superior so that even for Alexa568 the rings were clearly visible.
Resolving power several μm deep in the sample. To show that our setup with refractive index matching permits imaging several μm deep in the sample, we also performed 3D dSTORM with astigmatic detection of nuclear pore complexes (NPCs) in HeLa cells. Here, the nucleoporin 358 (Nup358) was labelled with Alexa647. Nup358 is located at the tip of the highly flexible cytoplasmic filaments of the NPC. In the 3D dSTORM images, the NPCs stained in this way are visible as little pores with diameter ~70 nm (see [fig_ref] Figure 4: 3D dSTORM images of nuclear pore complexes in the nuclear membrane, several... [/fig_ref] , which is in the same range as results from electron microscopy measurements for different species [bib_ref] Nuclear pore complex structure and dynamics revealed by cryoelectron tomography, Beck [/bib_ref] [bib_ref] Structure determination of the nuclear pore complex with three-dimensional cryo electron microscopy, Von Appen [/bib_ref]. The pores can be resolved at the nuclear membrane adherent to the cover glass as well as on the opposite membrane at a distance ≳5 μm to the cover glass, which shows that the resolution is not markedly degraded deep in the sample.
A measurement with glycerol-free buffer of refractive index 1.35 proved that index matching is advantageous for 3D localisation with two objectives (see supporting information). Due to the refractive index mismatch between cover glass and sample buffer, the objective at the opposite side of the sample exhibited a pronounced spherical aberration and yielded inferior results than the objective directly facing the cells, especially for axial localisation.
Distance measurements with Nanorulers. The axial resolving power in dSTORM measurements was further examined using custom-made GATTAquant Nanorulers, which are convenient tools to test the spatial resolution of fluorescence microscopes. The Nanorulers are artificial DNA structures where two surfaces at an axial distance of (33 ± 3) nm are labelled with Alexa647. One side of the structure is functionalised with biotin so that it couples to cover glasses coated with BSA-biotin and neutravidin such that the Nanoruler is oriented parallel to the optical axis. To investigate if the detection modalities implemented in our microscope could resolve the spatial separation of both surfaces, their distance was measured in intensity profiles created from dSTORM reconstructions [fig_ref] Figure 5: Distance measurements with 3D DNA structures [/fig_ref].
After image acquisition and reconstruction, axial slices of the structures were analysed with ImageJ 33 . In these images, the DNA constructs appear elongated in direction of the optical axis. Intensity profiles along their extent were created by summing up all image pixel counts on a width of 100 nm using ImageJ's Plot Profile function.
In the second step, the intensity profiles were normalized and MATLAB ® (The MathWorks, Inc., Natick, Massachusetts, USA) was used to fit the sum of two 1D Gaussian functions
[formula] = − − + − − g d d p p p d p p ( ) exp ( ) 2 exp ( ) 2(3) 1 2 4 2 3 2 2 2 5 [/formula]
2 to these intensity distributions by least squares fitting. d is the distance along the 1D intensity plot and p 1 to p 5 are fit parameters. Accordingly, the distance between the two maxima of the plot is d z = |p 2 − p 1 |. Fits g(d) which deviated so far from the data I thatsurpassed the χ 2 value corresponding to a 5% level of significance were excluded from the analysis. N l is the number of data points of the intensity profile. This was done to ensure that only data sets which could be correctly modelled by two distinct peaks were evaluated. 50 data sets acquired with astigmatic detection were analysed. Fits resulted in an axial distance of d z = (38 ± 2) nm. The error is the standard deviation of all d z values divided by the square root of the number of measurements. The result for d is slightly greater than the specified value (33 ± 3) nm. This indicates that the distance was close to the resolution of the astigmatic scheme, but could be resolved by this detection method.
[formula] I g d g d [ ( )] / ( ) i N i i i 2 2 l χ = ∑ − [/formula]
When images of 36 structures obtained with interferometric detection were analysed, an axial distance of d z = (34 ± 2) nm was obtained. Compared to the other z detection method, this result is closest to the true value, which confirms the higher resolving power of interferometric detection.
The deviation of the distance measured by astigmatic detection from the true value is probably a systematic error. This error is caused by the inferior resolution of this method, as distances smaller than the resolution cannot be detected and therefore do not contribute to the computation of the mean distance d mean . If we assume that the measured distances are normally distributed around a mean value d 0 with a standard deviation s, this can be seen from an estimate for d mean : where the integration limits d min and d max correspond to the measurement range. Let us assume that d 0 is the unbiased true value d 0 = 33 nm, and that we can measure distances infinitely long, d max = ∞, but not smaller than the full width at half maximum calculated from the localisation precision, d u 2 2 ln(2) z min = . The distance is a difference of two positions, so we estimate for the width of its distribution approximately two times the localisation precision, s = 2u z . If we interpolate the localisation precisions from the bead measurements to the typical brightness of Alexa647 by multiplying by a factor of 17000/3823 (photon number for beads/photon number for Alexa647 26 ), we obtain d mean = 33 nm for interferometric detection, which means that the result is still unbiased. But for astigmatic detection, the inferior resolution increases the expected distance to d mean = 38 nm, which agrees well with the measured result.
[formula] ∫ ∫ π π = − − − − d q s q d s q s q d s q 2 exp ( ) 2 d 1 2 exp ( ) 2 d ,(4) [/formula]
# Discussion
We showed that glycerol immersion objectives in conjunction with a glycerol-containing imaging buffer are well suited for dSTORM imaging of biological samples. Systematic comparison of theoretical as well as experimental localisation precisions for astigmatic and interferometric detection revealed that astigmatic 3D detection with two glycerol immersion objectives already yields almost isotropic 3D localisation precisions and resolves structures of about 30 nm in dSTORM experiments without necessitating complicated alignment or specialized optics, making it a convenient choice for biological applications. Adding a second objective improved the localisation precision by 2 as predicted. When interferometric detection is employed, the localisation precisions are even better for all three dimensions, but the experimental complexity is increased.
As our setup allows to switch between these detection options, it enables choosing the method for estimation of the axial coordinate according to the required resolution. For many applications, astigmatic detection with two objectives will be sufficient, but when even smaller structures need to be resolved, it is possible to change to interferometric detection.
The microscope presented here can be easily converted into a biplane detection setup, which further enhances the flexibility of the setup. This will allow to analyse the performance of refractive index matching in conjunction with an additional 3D imaging technique (preliminary results not shown).
As we showed by simulations, exact adjustment of the refractive indices improves the localisation precision. Furthermore, it will be especially beneficial for quantitative measurements, for example distance measurements, because it not only reduces distortions of the image field but chromatic errors as well. Furthermore, different refractive indices in sample buffer and immersion medium may lead to erroneous scaling of z localisation results due to different contractions of the optical axes. With identical refractive indices, this problem does not occur.
Because aberrations due to refractive index mismatch are more prominent deep in the sample, index matching enables single molecule localisation microscopy of thicker samples. In our measurements, we could image intracellular structures near the cover glass as well as more than 5 μm deep in the sample. The benefit of index matching for 3D single molecule localisation microscopy was also recently demonstrated for water immersion objectives in conjunction with a double-helix PSF [bib_ref] Three-Dimensional Super-Resolution in Eukaryotic Cells Using the Double-Helix Point Spread Function, Carr [/bib_ref]. In contrast to water immersion, glycerol immersion enables employing objectives with a higher NA. As the detection schemes presented by us only perform well for a depth of field of about 0.5 μm, thicker samples of several μm require capturing many planes at varying z positions. But this is easily feasible by moving the piezoelectric stage with the sample holder, however at the cost of an increased acquisition time. Two glycerol-immersion objective lenses (HCX PL APO 100x/1.35 GLYC CORR, Leica Microsystems, Wetzlar, Germany) capture the fluorescence light emitted by the sample. A glycerol-water mixture with 87% glycerol (G5515-100ML, SIGMA ALDRICH CHEMIE GmbH, Steinheim, Germany) and 13% water (v/v) is a suitable medium for immersion. Objective O1 in [fig_ref] Figure 6: Sketch of the beam path and photo of the detection part [/fig_ref] is adjustable along three axes. Coarse positioning is done via a custom-made three-axis linear adjustment unit (R. Rittmeyer GmbH, Münster, Germany) with counter surfaces of sapphire to minimise friction and undesired shift during motion. This unit carries a three-axis piezo nanopositioner (Nano-F3D with Nano-Drive controller and 20 bit USB interface, Mad City Labs Inc., Madison, Wisconsin, USA) for fine adjustment. The sample is mounted onto a three-axis linear adjustment unit with counter surfaces of sapphire (R. Rittmeyer GmbH, Münster, Germany) as well. Fine adjustment along the optical axis with position feedback is required for focusing and for capturing calibration curves for z position estimation. This Mirrors in removable magnetic mounts (KS1R, Thorlabs Inc, Newton, New Jersey, USA) allow switching between two detection pathways: If the mirrors are inserted, the fluorescence light detected by each objective is directly led to one EMCCD camera (iXon3 897, Andor Technology Ltd., Belfast, UK) where it is focused onto the detector chip by a tube lens with focal length 200 mm (G063205000, Qioptiq Photonics GmbH & Co. KG, Göttingen, Germany). This beam path is employed for astigmatic detection. For this 3D detection method, additional cylindrical lenses with focal length 300 mm (LJ1558RM-A, Thorlabs Inc, Newton, New Jersey, USA) have to be inserted in front of the cameras. They cause an elliptic distortion of the PSF which enables to back-calculate an emitters' axial position from the width of the PSF.
# Methods
If the removable mirrors are absent, the light of both objectives is combined by two 50:50-beamsplitters (BS013, Thorlabs Inc, Newton, New Jersey, USA), one 70:30-beamsplitter (BS019, Thorlabs Inc, Newton, New Jersey, USA) and a prism (RAP100C, Laser Components GmbH, Olching, Germany) which may be translated precisely via a piezoelectric linear actuator (8353 with controller 8742-4-KIT, Newport Spectra-Physics GmbH, Darmstadt, Germany). This configuration leads to sinusoidal, z-dependent variations of the intensities measured at each camera 8 and enables to detect the axial coordinate with interferometric precision. As this requires an exact alignment of the detected beams and precise adjustment of the optical path lengths, both the tilt of the beam exiting objective O1 and the path length in the corresponding arm may be controlled by an adjustable mirror. This mirror can be tilted along two axes via a piezoelectric tilt actuator (Nano-MTA2 with Nano-Drive controller and 20 bit USB interface, Mad City Labs Inc., Madison, Wisconsin, USA) and displaced backward and forward by a single axis nanopositioner (Nano-OP30-M with Nano-Drive controller and 20 bit USB interface, Mad City Labs Inc., Madison, Wisconsin, USA). A two-axis linear translation stage with rotation platform (XYR1/M, Thorlabs Inc, Newton, New Jersey, USA) enables coarse positioning of the mirror. A custom-made dispersion compensator consisting of a pair of glass wedges and a compensator plate (Bernhard Halle Nachfl. GmbH, Berlin, Germany) allows to compensate small differences in glass thickness of both arms of the interferometer which arise from manufacturing tolerances of the optical building parts.
In front of the cameras, the laser light is removed by notch filters (QuadLine Rejectionband ZET405/488/561/640, AHF analysentechnik AG, Tübingen, Germany). A set of three emission filters (700/75 ET Bandpass, 617/73 BrightLine HC, 525/50 ET Bandpass, AHF analysentechnik AG, Tübingen, Germany) permits to choose the spectral range of the detected fluorescence.
Particle localisation algorithm. For obtaining the x and y coordinates (lateral plane), a 2D Gaussian function was fit to regions of interest containing a single fluorescent molecule. Fitting was effected with maximum likelihood estimation.
Coordinates along the optical axis (z) were estimated by comparing properties of the single molecule images to calibration curves. To obtain calibration curves, a sub-resolution fluorescent bead is translated along the optical axis and imaged every 20 nm.
Cylindrical lenses cause an astigmatic distortion. For this detection scheme, a quadratic function with some higher order terms is fit to the width of the bead images.
For interferometric detection, sine curves are fit to the normalized intensities of each detector. To avoid ambiguities, a slight astigmatic aberration present in the system is used for coarse positioning [bib_ref] Superresolution fluorescence imaging of mitochondrial nucleoids reveals their spatial range, limits, and..., Brown [/bib_ref].
The Supplementary Information includes a more detailed description of the particle localisation process. Spatial misalignments between the image stacks captured by each camera were corrected by determining affine transformations between the images from measurements with brightly fluorescent beads prior to imaging. Bead x and y positions were obtained as described above and from these sets of corresponding positions, the affine transformations were calculated using functions from MATLAB ® 's Image Processing Toolbox (The MathWorks, Inc., Natick, Massachusetts, USA).
Spatial drift during measurements was corrected, if necessary, by localising sub-resolution sized fluorescent beads added to the samples and subtracting their deviation from their position at the beginning of the measurement from the x, y and z coordinates.
## Preparation of bead samples.
To prepare a bead sample for estimating the localisation precision, one 30 mm diameter fused silica cover glass (Leica Microsystems, Wetzlar, Germany) was glued into a metal ring (Feinmechanische Werkstatt im Physikalischen Institut, Münster, Germany). First of all, the cover glass in the ring and a second one required as counterpart were treated in a plasma cleaner (Femto with generator 40 kHz, 2000 V, 200 W, Diener electronic GmbH + Co. KG, Elbhausen, Germany). The generator was set to work at medium power (100 W) and processing time (6 min 40 s). After treatment in the plasma cleaner, 50 μl of TransFluoSpheres ® (T8861, Invitrogen Ltd., Paisley, UK) emitting at 605 nm diluted 1:10000 in deionized water were pipetted onto the cover glass and incubated for 2 min. Then the cover glass was washed by carefully applying and removing 150 μl of deionized water and let dry. Finally, 15 μl of immersion medium were added onto the cover glass surface, the second cover glass was put on top and the sample was sealed with a two-component-glue Carl Roth GmbH + Co. KG, Karlsruhe, Germany) in PBS. After this, they were washed three times and incubated for one hour with PBS containing 3% (w/v) bovine serum albumin (BSA, A9085, SIGMA ALDRICH CHEMIE GmbH, Steinheim, Germany) and 0.2% (w/v) saponin (S7900, SIGMA ALDRICH CHEMIE GmbH, Steinheim, Germany), denoted blocking buffer in the following. Next, the primary antibodies (mouse monoclonal α-tubulin, 302 211, Synaptic Systems GmbH, Göttingen, Germany) were applied. They were diluted 1:500 in 250 μl blocking buffer and incubated for one hour.
After removing the antibody solution, the cover glasses were washed five times with PBS with 0.2% (w/v) BSA and 0.05% (w/v) saponin, called washing buffer. Subsequently, the secondary antibodies were applied. For the red spectral range, they were labelled with Alexa647 (goat anti-mouse, A21236, Invitrogen Ltd., Paisley, UK) and for the orange spectral range with Alexa568 (goat anti-mouse, A21207, Invitrogen Ltd., Paisley, UK).
The antibodies were diluted 1:500 in 250 μl of blocking buffer and incubated for one hour. The cover glasses were washed five times with 500 μl washing buffer and one time with PBS, fixed for 10 min with 500 μl of 4% paraformaldehyde and 0.1% glutaraldehyde (glutaraldehyde 25% solution in water, 23114.02, SERVA Electrophoresis GmbH, Heidelberg, Germany) in PBS and washed once again. Finally, the samples were blocked for 20 min with 500 μl of 20 mM glycine in PBS and washed five times with PBS. NPC staining. The primary antibody was polyclonal rabbit anti-Nup358 (a gift from E. Coutavas, Rockefeller University of New York) and the secondary antibody Alexa647-labelled goat anti-rabbit (A32733, Invitrogen Ltd., Paisley, UK), otherwise the staining process was identical to the tubulin staining described above.
Nanoruler sample preparation. To prepare a sample with Nanorulers, a fused silica cover glass (diameter 30 mm, custom-made by XanTec bioanalytics GmbH, Düsseldorf, Germany) was cleaned in the plasma cleaner as Paisley, UK). The beads were diluted 1:1000000 in 500 μl PBS and incubated on the sample for 2 min. Then the sample was washed another three times with 500 μl PBS. In measurements with NPCs, we used TetraSpeck ® Microspheres (T7279, Invitrogen Ltd., Paisley, UK). They were diluted 1:100 in 250 μl PBS and incubated on the sample for 5 min. Prior to imaging, all samples was embedded in dSTORM buffer (see below).
For dSTORM, samples were embedded in imaging buffer based on a buffer containing 200 mM 1,4-piperazinediethanesulfonic acid (A1079, AppliChem GmbH, Darmstadt, Germany), supplied with 40% glucose (6887.1, Carl Roth GmbH + Co. KG, Karlsruhe, Germany) and adjusted to a pH of 7.2. Glycerol (G5515-100ML, SIGMA ALDRICH CHEMIE GmbH, Steinheim, Germany) was added to the buffer until its refractive index was in the range of 1.45. The glycerol concentration then was about 55% of volume. The refractive index was measured with a refractometer (DR201-95, A. Krüss Optronic, Hamburg, Germany). Directly before applying the buffer to the sample, it was supplied with 0.8 mg ml glucose oxidase (G2133-10KU, SIGMA ALDRICH CHEMIE GmbH, Steinheim, Germany), 0.08 mg ml catalase (C1345-1G, SIGMA ALDRICH CHEMIE GmbH, Steinheim, Germany) and 10 mM MEA (also known as cysteamine, 30070-10 G, SIGMA ALDRICH CHEMIE GmbH, Steinheim, Germany).
For imaging, 15 μl of dSTORM imaging buffer were added to the samples and they were mounted with a second cover glass onto a custom metal sample holder in the same way as the bead samples.
For each sample, the correction rings of the objective lenses were adjusted to its refractive index by rotating the rings until the brightness of fluorescent beads in the sample was maximal. Then, calibration curves were obtained. In measurements with Alexa647, the AOTF transmission of the 647 nm laser illumination was set to 100%. This corresponds to an intensity of about [bib_ref] Interferometric fluorescent super-resolution microscopy resolves 3D cellular ultrastructure, Shtengel [/bib_ref] in the focal plane. When Alexa568 was used, the transmission for 561 nm was 100%, corresponding to about [bib_ref] Three-dimensional sub-100 nm resolution fluorescence microscopy of thick samples, Juette [/bib_ref] in the focal plane. For this dye, illumination with the violet diode laser, whose power was reduced to 15 mW, was employed to mediate photoswitching. The respective AOTF transmission was increased from 2% to 15% during measurements in an exponential fashion, augmenting the intensity in the focal plane from about . 0 9 W cm 2 to 24 W cm 2 . For every dSTORM reconstruction, 60000 frames with an acquisition time of 30 ms were acquired.
We found that in experiments with NPCs, the background was higher at the beginning of the image acquisition, so we discarded the first ~16000 frames for data analysis.
## Data availability
The datasets generated during and analysed during the current study are available from the corresponding authors on reasonable request.
[fig] Figure 1: Impact of refractive index mismatch on the PSF, calculated for an oil immersion objective with halfaperture angle α = 60°, focusing into an aqueous sample. (a) No refractive index mismatch (n s = n 1 = 1.52). (b) Refractive index mismatch (n s = 1.33, n 1 = 1.52), depth in sample d s = 1 μm. (c) Refractive index mismatch (n s = 1.33, n 1 = 1.52), depth in sample d s = 10 μm. Note that the focus is shifted substantially due to the path length difference. SCIentIFIC RepoRts | (2018) 8:13343 | DOI:10.1038/s41598-018 [/fig]
[fig] Figure 2: Localisation precisions u against axial position z with TransFluoSpheres ® for x, y and z. (a) Astigmatic detection. (b) Interferometric detection. Insets show magnifications of half a micrometer around the focus. For interferometric detection, simulated localisation precisions are identical for x and y. SCIentIFIC RepoRts | (2018) 8:13343 | DOI:10.1038/s41598-018-31595-z [/fig]
[fig] Figure 3: dSTORM images of stained microtubules. (a) Overview image of Alexa647 staining for astigmatic detection, scale bar 2 μm. (b) Five axial slices through microtubules stained with Alexa647 for astigmatic detection, scale bar 100 nm. (c) Overview image of Alexa647 staining for interferometric detection, scale bar 2 μm. (d) Five axial slices through microtubules stained with Alexa647 for interferometric detection, scale bar 100 nm. (e) Overview image of Alexa568 staining for astigmatic detection, scale bar 2 μm. (f) Five axial slices through microtubules stained with Alexa568 for astigmatic detection, scale bar 100 nm. (g) Overview image of Alexa568 staining for interferometric detection, scale bar 2 μm. (h) Five axial slices through microtubules stained with Alexa568 for interferometric detection, scale bar 100 nm. For each method, the first axial slice is also shown with a dashed ring of diameter 25 nm, indicating the microtubule thickness, and a white arrow denoting the optical axis. [/fig]
[fig] Figure 4: 3D dSTORM images of nuclear pore complexes in the nuclear membrane, several μm deep in HeLa cells. Nucleoporin 358 was stained with Alexa647 and measured with astigmatic detection. (a) Bottom side of the nucleus, near the cover glass. (b) Top side of the same nucleus, at a distance of 6.4 μm to the bottom. (c) Second example for the bottom side of a nucleus, near the cover glass. d: Top side of the same nucleus, at a distance of 5.4 μm to the bottom. Scale bar 1 μm. SCIentIFIC RepoRts | (2018) 8:13343 | DOI:10.1038/s41598-018-31595-z [/fig]
[fig] Figure 5: Distance measurements with 3D DNA structures. dSTORM images with corresponding line fits, for (a) astigmatic detection, (b) interferometric detection. Shown are three examples for each method. Scale bar 100 nm, white arrows indicate the optical axis, d is the position along the intensity profile and I the normalized intensity. SCIentIFIC RepoRts | (2018) 8:13343 | DOI:10.1038/s41598-018-31595-z [/fig]
[fig] Figure 6: Sketch of the beam path and photo of the detection part. Dashed lines indicate the detection beam path with inserted removable mirrors, which is used for astigmatic detection. AM: adjustable mirror, AOTF: acousto-optical tunable filter, BS50: beamsplitter 50:50, BS70: beamsplitter 70:30 (T:R), CCD1: charge-coupled device 1, CCD2: charge-coupled device 2, CCD3: charge-coupled device 3, COL: collimation lens, CP: compensator plate, CYL: cylindrical lens, DC: dispersion compensator, DM: dichroic mirror, EF: emission filter, L405: laser 405 nm, L491: laser 491 nm, L561: laser 561 nm, L647: laser 647 nm, λ/4: quarter wave plate, M: mirror, NF: notch filter, O1: objective 1, O2: objective 2, QBF: quadband beamsplitting filter, P: prism, RM: removable mirror, S: sample, SH1: shutter 1, SH2: shutter 2, T1: telescope lens 1, T2: telescope lens 2, TL: tube lens, WP: wedge prisms. SCIentIFIC RepoRts | (2018) 8:13343 | DOI:10. [/fig]
[fig] SCIentIFIC: RepoRts | (2018) 8:13343 | DOI:10.1038/s41598-018-31595-z [/fig]
|
Randomised Phase II Trial (NCT00637975) Evaluating Activity and Toxicity of Two Different Escalating Strategies for Pregabalin and Oxycodone Combination Therapy for Neuropathic Pain in Cancer Patients
Purpose: Neuropathic pain is commonly associated with cancer. Current treatments include combination opioid and adjuvant therapies, but no guidelines are available for dose escalation strategies. This phase II study compared the efficacy and tolerability of two dose escalation strategies for oxycodone and pregabalin combination therapy.Methods: Patients (N = 75) with oncological neuropathic pain, previously untreated with pregabalin, were recruited in 5 Italian institutions between 2007 and 2010. Patients were randomised to two different dose escalation strategies (arm A; N = 38) oxycodone at a fixed dose with increasing pregabalin doses; (arm B; N = 37) pregabalin at a fixed dose with increasing oxycodone doses. Patients were evaluated from daily diaries and follow-ups at 3, 7, 10, and 14 days after beginning treatment with a numerical rating scale (NRS), neuropathic pain scale (SDN), and well-being scale (ESAS). The primary endpoint was a $1/3 reduction in pain (NRS); secondary endpoints included the time to analgesia and adverse effects. The study had a 90% probability of detecting the best strategy for a true difference of at least 15%.Results: More patients in arm A (76%) than arm B (64%) achieved $1/3 overall pain reduction even after controlling for baseline factors (gender, baseline pain). Group A reported fewer side effects than group B; constipation 52.8% vs. 66.7%; nausea: 27.8% vs. 44.4%; drowsiness: 44.4% vs. 55.6%; confusion: 16.7% vs. 27.8%; itching: 8.3% vs. 19.4%.Conclusions: Both strategies effectively controlled neuropathic pain, but according to the adopted selection design arm A is preferable to arm B for pain control.Trial Registration: ClinicalTrials.gov NCT00637975.
# Introduction
Neuropathic pain is a common symptom in patients with cancer. Among patients with oncological pain, at least 1/3 is diagnosed with neuropathic pain [bib_ref] Contemporary assessment and management of neuropathic pain, Irving [/bib_ref]. In this setting, pain can be caused by a tumour compressing a nerve or it may be a side effect of chemotherapy and radiotherapy. The mainstay of treatment is opioid administration; however, a monotherapy is often insufficient to control neuropathic pain. It is widely recognised that some cancer pain syndromes are only partially responsive to opioids. This has led to the search for new strategies of treatment [bib_ref] Efficacy and safety of opioid agonists in the treatment of neuropathic pain..., Eisenberg [/bib_ref] [bib_ref] Randomized, double-blind, cross-over trial comparing safety and efficacy of oral controlled-release oxycodone..., Bruera [/bib_ref].
Currently, adjuvant drugs, like antidepressants or anticonvulsants, are often employed in combination with the primary therapy [bib_ref] Ajuvant analgesics in pain management, Portenoy [/bib_ref].
Previously, a meta-analysis on the role of opioids in treating benign neuropathic pain showed that some opioids, particularly oxycodone, are more effective than other agents [bib_ref] Efficacy and safety of opioid agonists in the treatment of neuropathic pain..., Eisenberg [/bib_ref] [bib_ref] Opioids for chronic noncancer pain: a meta-analysis of effectiveness and side effects, Furlan [/bib_ref]. In a metaanalysis by Finnerup et al. [bib_ref] Algorithm for neuropathic pain treatment: an evidence based proposal, Finnerup [/bib_ref] , anticonvulsants, particularly gabapentin, showed a favourable trade-off between harm and benefit; thus, these were considered the best candidates for combining with opioid treatments. Moreover, previous studies on the use of combination therapies in patients with cancer and neuropathic pain showed that gabapentin enhanced analgesia [bib_ref] Gabapentin for neuropathic cancer pain: a randomized controlled trial from the Gabapentin..., Caraceni [/bib_ref].
Pregabalin, a new anticonvulsant, appeared to be effective for relieving neuropathic pain, and it acted synergistically with oxycodone, with no additional toxicity [bib_ref] Randomized, double-blind, cross-over trial comparing safety and efficacy of oral controlled-release oxycodone..., Bruera [/bib_ref]. Furthermore, pregabalin was active in patients with a resistance to gabapentin. A recent pharmaeconomics analysis demonstrated that pregabalin was cost-effective for patients with refractory neuropathic pain [bib_ref] A costutility study of the use of pregabalin in treatment-refractory neuropathic pain, Gordon [/bib_ref].
Gatti et al. demonstrated the effectiveness and tolerability of pregabalin combined with oxycodone in treating non-cancer pain in a large cohort of patients [bib_ref] Controlled-release oxycodone and pregabalin in the treatment of neuropathic pain: results of..., Gatti [/bib_ref]. This combination seemed to be effective, particularly in an escalating dose strategy [bib_ref] Long-term controlled-release oxycodone and pregabalin in the treatment of non-cancer pain: an..., Gatti [/bib_ref]. However, the strategy for increasing the dose of opioids or anticonvulsants is entirely empirical; currently, there are no data available to advocate any particular strategy.
Previous studies have demonstrated that genetic variations in pain-related receptors, transporters, and metabolising enzymes were related to opioid efficacy. The most interesting genes discovered were the mu and kappa opioid receptors [bib_ref] Association of ABCB1/ MDR1 and OPRM1 gene polymorphisms with morphine pain relief, Campa [/bib_ref] [bib_ref] The 118 A.G polymorphism in the human mu-opioid receptor gene may increase..., Klepstad [/bib_ref] [bib_ref] Genetic factors in pain and its treatment, Stamer [/bib_ref]. Those studies generated intense interest in whether genetic analyses can be used to guide the choice of opioid treatment. However, as reviewed by Hirschhorn et al. [bib_ref] A comprehensive review of genetic association studies, Hirschhorn [/bib_ref] , most studies that found candidate single nucleotide polymorphisms (SNPs) associated with outcomes could not be replicated [bib_ref] Association between OPRM1 gene polymorphisms and fentanyl sensitivity in patients undergoing painful..., Fukuda [/bib_ref] [bib_ref] Diversity of opioid requirements for postoperative pain control following oral surgery-is it..., Fukuda [/bib_ref].
To our knowledge, no previous studies have investigated different dose escalation strategies in a combination therapy for neuropathic pain. In current clinical practice, this decision is typically made empirically by individual physicians, based on personal experience.
The present prospective study aimed to evaluate two different dose escalation strategies for combining pregabalin with oxycodone in treating patients with neuropathic pain caused by neoplasms. We also examined genetic SNPs as a potential basis for differences in drug responses.
# Methods
The protocol for this trial and supporting CONSORT checklist are available as supporting information; see Protocol S1 and Checklist S1. This research has been approved by the Ethics Committee of Fatebenefratelli and Oftalmico Hospital in Milan and has been conducted according to the Declaration of Helsinki Principles. Five Italian institutions participated in the trial from September 2007 to December 2010. The protocol was approved by the ethics committees of each participating centre and written informed consent was obtained from all participants. The study is registered at Clinicaltrial.gov (NCT00637975, Supplementary File Neuropain Protocol).
## Patient characteristics
Patients with cancer pain were enrolled when they had a clinical diagnosis of cancer and pain with a neuropathic component. Neuropathic pain was identified by the physician as burning pain, shooting or lancinating pain episodes, dysesthesia, or allodynia.
## Inclusion and exclusion criteria
The inclusion criteria were at least 18 years of age; pain that occurred within the 24-h period preceding the screening visit and was rated $4 on a NRS (range, 0-10) [bib_ref] Pain measurement tools and methods in clinical research in palliative care: recommendations..., Caraceni [/bib_ref] ; pain with a neuropathic component caused by malignant infiltration or compression of nervous structures; Performance Status score ,3, according to the toxicity and response criteria of the Eastern cooperative oncology group (ECOG); and written informed consent. Exclusion criteria were serum creatinine .2 mg/ml or creatinine clearance ,40 ml/min; previous or current pregabalin use; mild or severe hepatic insufficiency; iatrogenic neuropathy caused by chemotherapeutic agents; previous allergic reactions to oxycodone or pregabalin; and pregnancy or breastfeeding.
Chemotherapy crossover was not allowed in the 3 weeks before screening or throughout the study. Radiotherapy was not allowed on the lesion that caused the neuropathic pain. Hormone therapy that had been started before trial entry could be continued, but the dose could not be changed during the study.
## Study design
Patients were randomly assigned by a computer-generated scheme to receive one of two dose escalation strategies. (Arm A) Patients received oxycodone controlled release (CR), 20 mg/day, plus increasing doses of pregabalin, starting at 50 mg/day. (Arm B) Patients received pregabalin, 50 mg/day, plus increasing doses of oxycodone, starting at 20 mg/day [fig_ref] Figure 1: Study design [/fig_ref].
On the day of the screening visit, seven days before treatment start (27 days), a titration phase began with the administration of normal-release morphine. On the day that treatment started (day 0), eligible patients were randomly allocated to Arm A or Arm B. After randomisation, patients were observed for 14 days. Pain and tolerability were recorded daily in a patient diary and evaluated by a physician at 3, 7, 10, and 14 days.
When the prior 24-hour pain scores, taken at specified time points, were $4 and no side effects were reported, the patients received a dose escalation, according to the randomisation arm. In Arm A, the dose of pregabalin was increased by 50 mg up to a maximum dose of 300 mg. In Arm B, oxycodone CR was increased in increments less than fifty percent of the previous dose. Rescue doses with normal release morphine were permitted, but they had to be recorded in the patient's daily diary.
## Pain assessment at baseline
During the 24-hour period preceding the randomisation visit, average pain intensity was assessed on a 0 to 10 NRS [bib_ref] Pain measurement tools and methods in clinical research in palliative care: recommendations..., Caraceni [/bib_ref] , where 0 corresponded to ''no pain'' and 10 to ''the worst possible pain''. The neuropathic component was measured with the SDN version validated in Italian [bib_ref] Validation of the Italian version of the ''Neuropathic Pain Scale'' and its..., Negri [/bib_ref] , and ''well-being'' was rated with the ESAS [bib_ref] The Edmonton Symptom Assessment System (ESAS): a simple method for the assessment..., Bruera [/bib_ref].
The somatic component, mainly due to bone metastases, was also clinically evaluated by the physician and recorded on the clinical chart.
Allodynia, defined as pain in response to non-painful stimulation, was assessed by asking the patient to describe the location of a referred pain, and gently stroking the area with a cotton swab. The pain was recorded as present or absent.
## Pain intensity follow-up and pain diary
At the screening visit (day -7), patients were instructed on how to complete a daily pain diary, which included the average intensity of global pain, evaluated with the NRS, and any concomitant medications. The patient was also required to record the number of daily breakthrough pain episodes (BTP) and the use of rescue analgesic doses. At each follow-up visit (days 3, 7, 10, and 14), pain was re-assessed by the physician with the SDNs, and ESAS interviews were performed.
## Efficacy and safety outcomes
The primary endpoint was overall analgesia, defined as a reduction of pain intensity by at least 1/3 (NRS). Secondary endpoints were reductions in BTP episodes and neuropathic pain (SDN) and an improvement in patient well-being (ESAS).
Drug safety was assessed by evaluating the type, frequency, and intensity of any reported adverse events according to the Common Terminology Criteria for Adverse Events Version 4.02 (CTCAE 4.02).
## Ancillary study
Within the main study, there was an optional ancillary study aimed to detect a possible correlation between selected genes and the drug response. A subset of 50 out of 76 patients from five Italian institutions participated in the ancillary study, from September 2007 to December 2010. The candidate genes included, in addition to others, the opioid receptors, kappa 1 (OPRK1) and mu 1 (OPRM1) [bib_ref] Association of ABCB1/ MDR1 and OPRM1 gene polymorphisms with morphine pain relief, Campa [/bib_ref] [bib_ref] The 118 A.G polymorphism in the human mu-opioid receptor gene may increase..., Klepstad [/bib_ref] [bib_ref] Genetic factors in pain and its treatment, Stamer [/bib_ref]. DNA was extracted from blood samples with a Maxwell 16 DNA Purification Kit (Promega, Milan, Italy). Three OPRK1 polymorphisms (rs7815824, rs702764, and rs1051660) and one OPRM1 polymorphism (rs1799971) were genotyped with TaqMan SNP Genotyping assays (Applied Biosystems, Monza, Milan). PCR was carried out in 384-well plates, prepared with an automatic liquid handling system (epMotion 5075, Eppendorf, Milan Italy). The PCR-amplified DNA fragments were analysed with Allelic Discrimination Sequence Detection Software (Applied Biosystems, Monza, Italy).
# Statistical methods
Design and sample size evaluation. The study adopted a randomised selection design described by Simon et al [bib_ref] Randomized phase II clinical trials, Simon [/bib_ref]. The minimum expected response rate was set to 45%, and a minimum of 74 patients was required for a power of 90% probability of correctly detecting the best schedule for a true response difference of at least of 15%. With this sample size, the study also had a $90% power to make the correct choice of the best treatment, even with baseline response rates greater than 45%, under the hypothesis that the true response difference was at least 15%.
Analysis. Analgesia, defined as at least a 1/3 reduction in pain intensity, was the primary endpoint. Analgesia was described in terms of frequency and proportion, with the relative 95% C.I. and median time-to-analgesia derived from Kaplan-Meyer curves.
Changes in well-being (ESAS) and reductions in BTP episodes, allodynia, somatic pain, and side effects were reported in terms of absolute and relative frequencies or medians and ranges, according to the type of data (categorical, nominal, or continuous). Since the randomised selection design is not powered for formal comparisons between arms, the differences between the two groups were analysed only for exploratory purposes with the x 2 test or Wilcoxon non parametric test, according to the type of data. Significance was set at p-values,0.05.
A multivariate analysis was performed and adjusted for the presence of a somatic component.
# Results
From September 2007 to November 2010, 75 patients were randomised in 5 Italian centres. Thirty-eight (50.7%) were assigned to arm A and 37 (49.3%) were assigned to arm B.
## Study profile
The trial implementation profile is shown in [fig_ref] Figure 2: Patient CONSORT Diagram [/fig_ref]
## Patient demographics and baseline clinical characteristics
Demographic characteristics of patients and baseline pain characteristics are reported in [fig_ref] Table 1: Patient demographic characteristics and baseline pain in patients with oncological neuropathic pain [/fig_ref]. Although an imbalance was detected regarding the site of disease, the somatic component of pain was equally distributed in the two arms of treatment. All patients had received previous treatment with either weak or strong opioids.
## Primary activity outcome
Reduction of at least 33% of the baseline pain was achieved in 26 (76.5%) patients in arm A vs. 23 (63.9%) patients in arm B (OR = 1.84; 95% CI: 0.65-5.22; p-value = 0.25). The median NRS score for pain was 6 (range 4-9) in arm A and 5 (range 2-9) in arm B. The result was confirmed in a multivariate analysis, after adjustments for gender, the presence of a somatic pain component, and the initial pain score (OR = 1.76; 95% CI: 0.61-5.05; pvalue = 0.29). Analgesia was achieved in arm A with a mean dose of 100 mg of pregabalin and in arm B with a mean dose of 60 mg of oxycodone. The median time to achieve analgesia was 10 days (95% CI: 5-14 days) in arm A and 11 days (95% CI: 5-18 days) in arm B. Activity outcome is summarised in [fig_ref] Table 2: Activity outcomes [/fig_ref].
## Secondary activity outcomes
The frequency of the use of rescue doses of analgesics for BTP was similar between groups: arm A, 10 patients (29.4%); arm B, 8 patients (28.6%); p = 0.59.
The well-being of patients, based on the ESAS scale and SDN item distributions was not statistically different between the study arms. Toxicity Unexpectedly, distinct toxicity profiles for each study arm could not be found [fig_ref] Table 3: Evaluation of toxicity [/fig_ref]. In fact, there was no difference between the two study arms (constipation, drowsiness, confusion, itching, nausea).
# Analysis of polymorphisms
We analysed DNA samples from 50 patients to determine whether the OPRK1 and OPRM1 genes were correlated to the activity and/or toxicity of the combination of pregabalin and oxycodone. We found no significant correlations. Moreover, the distributions of the different genotypes were similar in the two groups.
# Discussion
The aim of this phase two study was to assess two different dose escalation strategies for the treatment of neuropathic pain in an oncological setting. The study was designed according to the Simon selection design in order to select the best strategy for a subsequent phase 3 trial. Regarding tolerability, we confirmed the data available in the literature, which showed that the combination of oxycodone and pregabalin was effective and safe [bib_ref] Efficacy and safety of opioid agonists in the treatment of neuropathic pain..., Eisenberg [/bib_ref] [bib_ref] Opioids for chronic noncancer pain: a meta-analysis of effectiveness and side effects, Furlan [/bib_ref] [bib_ref] Algorithm for neuropathic pain treatment: an evidence based proposal, Finnerup [/bib_ref]. In fact, adherence to the study protocol was very high: only one patient discontinued treatment based on toxicity, due to severe disorientation (G3); two other patients refused to continue treatment, one for each arm. Our results also showed that, in contrast to the daily clinical practice of many physicians, pain control can be safely achieved with incremental doses of pregabalin, and it does not require increasing doses of opioids, like oxycodone. Studies on pregabalin [bib_ref] Pregabalin: an antiepileptic agent useful for neuropathic pain, Blommel [/bib_ref] have demonstrated that doses of up to 600 mgs daily were well tolerated. In clinical practice, sometimes the dose escalation takes place faster than recommended, and consequently, neurological toxicity occurs, like dizziness, nausea, and general malaise, which can lead to the premature interruption of treatment. Instead, we observed that, when patients gradually increased pregabalin, as in this study, they could benefit from high pregabalin doses in a safe setting.
The response to neuropathic pain treatment and the incidence of side effects are different for each patient. This variability may be explained by the specific characteristics of individual patients. Thus, we selected and evaluated potential polymorphisms in two key genes involved in the opioid response with genetic analyses of the OPRK1 and OPRM1 SNPs. Unfortunately, our data were not sufficiently powered to determine genetic influences on drug response. Nevertheless, the distributions of these genotypes were equivalent between the groups studied. This allowed us to exclude the possibility that the clinical differences observed between the two randomised groups were influenced by differences in mu and kappa receptor genes. Our analysis indicated that the genes identified by rs1051660, rs1799971, and rs7815624 had no interactions with the results. However, the rs702764 polymorphism showed an odds ratio of 1.8; thus, in future large scale studies, it might be interesting to analyse the role of this polymorphism in the response to combination therapy. The treatment of neuropathic pain remains an open issue. Active research should be conducted to find more effective and personalised treatments. This paper could lead to a phase III study that would compare the strategy of oxycodone plus pregabalin at increasing doses with a conventional treatment, like gabapentin plus morphine, as published by Caraceni and colleagues [bib_ref] Gabapentin for neuropathic cancer pain: a randomized controlled trial from the Gabapentin..., Caraceni [/bib_ref].
## Supporting information
Protocol S1 Trial protocol.
## (doc)
Checklist S1 CONSORT checklist. (DOC)
[fig] Figure 1: Study design. doi:10.1371/journal.pone.0059981.g001 [/fig]
[fig] Figure 2: Patient CONSORT Diagram. doi:10.1371/journal.pone.0059981.g002 [/fig]
[table] Table 1: Patient demographic characteristics and baseline pain in patients with oncological neuropathic pain. [/table]
[table] Table 2: Activity outcomes. doi:10.1371/journal.pone.0059981.t002 [/table]
[table] Table 3: Evaluation of toxicity. doi:10.1371/journal.pone.0059981.t003 [/table]
|
IFNβ-producing CX3CR1+ macrophages promote T-regulatory cell expansion and tumor growth in the APCmin/+ / Bacteroides fragilis colon cancer model
Increased T-regulatory cell activity drives tumor progression in the compound APC min/+ /enterotoxic Bacteroides fragilis colon cancer model. At the same time, how microbially-induced inflammation promotes T-regulatory cell expansion in the dysplastic intestine remains poorly described. Analysis of post-infection immune cell kinetics in the colon lamina propria revealed that CD4+ Foxp3+ cell numbers increased by 25-fold between days 3-14. Importantly, T-regulatory cell expansion was preceded by a 12fold spike in lamina propria CD11b + cell numbers between days 0-4; suggesting a link between the myeloid compartment and the T-regulatory cells. Consistent with this notion, in vitro co-culture studies utilizing sorted myeloid cell subsets and CD4 + T-cells demonstrated that the CD11b + CX3CR1 + but not the CD11b + CX3CR1 − subset preferentially induced Foxp3 expression in CD4 + T-cells. Phenotypic analysis revealed that the CD11b + CX3CR1 + subset represented a homogenous CD64 + CD24 − CD103a − macrophage population. Global CX3CR1 knockout or conditional depletion of CX3CR1 + myeloid cells resulted in diminished CD4 + Foxp3 + cell expansion and a 3 to 6-fold reduction in tumor burden establishing CX3CR1 + macrophages as a major driver of the T-regulatory cell-tumor axis. Quantitative analysis of CD11b + myeloid cell subsets for IFNβ mRNA revealed that the CX3CR1 + macrophages expressed 15-fold higher levels of IFNβ in comparison to the CX3CR1 − myeloid subset. Antibody mediated neutralization of IFNβ resulted in the suppression of CD4 + Foxp3 + cell induction and tumor growth, demonstrating the central role of IFNβ in mediating CX3CR1 + macrophage-driven T-regulatory cell expansion. These studies shed new mechanistic light on the cellular ontogeny of pro-tumorigenic T-regulatory cells in the inflamed colon of the APC min/+ mouse.ARTICLE HISTORY
# Introduction
APC min/+ /enterotoxic Bacteroides fragilis (ETBF) murine model of colon cancer 1 has been used extensively to study the role of microbially-induced inflammation in colon tumorigenesis and tumor progression. Specifically, a series of elegant studies by the Sears and the Housseau laboratories have demonstrated the central role of type 17 T-cell immunity in tumor development and growth in this model; and delineated how ETBF promotes type 17 immunity via its pleiotropic activity on gut epithelium and the myeloid cell compartment. [bib_ref] A human colonic commensal promotes colon tumorigenesis via activation of T helper..., Wu [/bib_ref] [bib_ref] Enterotoxigenic bacteroides fragilis (ETBF)-mediated colitis in Min (Apc±) mice: a human commensal-based..., Housseau [/bib_ref] [bib_ref] Redundant innate and adaptive sources of IL17 production drive colon tumorigenesis, Housseau [/bib_ref] [bib_ref] The myeloid immune signature of enterotoxigenic bacteroides fragilis-induced murine colon tumorigenesis, Orberg [/bib_ref] [bib_ref] Bacteroides fragilis toxin coordinates a pro-carcinogenic inflammatory cascade via targeting of colonic..., Chung [/bib_ref] Separately, these and other studies, including those from our laboratory, revealed an equally critical role for T-regulatory cells (Treg) in driving tumor growth in the ETBF-colonized gut. [bib_ref] Regulatory T-cell response to enterotoxigenic bacteroides fragilis colonization triggers IL17-dependent colon carcinogenesis, Geis [/bib_ref] [bib_ref] Oral IL-10 suppresses colon carcinogenesis via elimination of pathogenicCD4(+) T-cells and induction..., Gu [/bib_ref] Importantly, a recent report demonstrated that Treg can preferentially help to expand Th17 cells by acting as a sink for IL-2 in the above model, establishing a novel link between the two pathogenic subsets. [bib_ref] Regulatory T-cell response to enterotoxigenic bacteroides fragilis colonization triggers IL17-dependent colon carcinogenesis, Geis [/bib_ref] While these reports illuminated the cellular origin and the pathogenic role of type 17 inflammation that accompanies and drives colon tumorigenesis, the ontogeny of the ETBFdriven pro-tumorigenic Treg response remains poorly understood. It is known that under steady-state conditions gutresident CD11b + CX3CR1 + myeloid cells are critical to the maintenance of immune homeostasis in the GI tract [bib_ref] Interleukin-10 receptor signaling in innate immune cells regulates mucosal immune tolerance and..., Shouval [/bib_ref] and more specifically that of the lamina propria (LP) Treg compartment. [bib_ref] Critical role for the microbiota in CX3CR1(+) intestinal mononuclear phagocyte regulation of..., Kim [/bib_ref] However, the cellular mechanisms that regulate the generation and expansion of pro-tumorigenic Treg in the inflamed dysplastic colon have not been directly addressed.
Herein we report that in the APC min/+ /ETBF model Treg can drive tumor growth independent of Th17 cell activity and that the ontogeny of pro-tumorigenic Treg is linked to the CD11b + CD64 + CD24 − CD103a − CX3CR1 + gut-resident macrophages, which expand within 48 hours of bacterial colonization. We further show that the unique ability of these macrophages to rapidly expand Treg is associated, at least in part, with production of IFNβ that is required both for Treg induction and tumorigenesis.
# Results
CD4 + Foxp3 + Treg expansion is rapid and is required for tumor growth independent of Th17 cell activity In order to obtain insight into the ontogeny of ETBF-induced Treg we first analyzed the Treg kinetics in the inflamed colon of the APC min/+ mouse. Treg numbers in the colon LP were monitored between days 0 (un-infected mice) and day 21 postcolonization. The data shown in [fig_ref] Figure 1: Effect of CD4 + Foxp3 + Treg on tumorigenesis and T-effector cell... [/fig_ref] demonstrate that Treg expansion commenced between days 3 and 6 post-ETBF (8-fold), reached its peak on day 14 (25-fold) and declined thereafter but remained significantly above baseline . To confirm that this dramatic increase drove tumor growth, we determined whether depletion of Treg affected tumor burden in APC min/+ mice that were heterozygous for Foxp3 DTR/+ . Mice were administered diphtheria toxin (DT) starting on day 7 post-ETBF administration for 3 weeks and tumor burden was analyzed in control vs DT-treated animals. [fig_ref] Figure 1: Effect of CD4 + Foxp3 + Treg on tumorigenesis and T-effector cell... [/fig_ref] shows that depletion of Foxp3 + T-cells (Supplemental [fig_ref] Figure 1: Effect of CD4 + Foxp3 + Treg on tumorigenesis and T-effector cell... [/fig_ref] resulted in a robust 5-fold reduction in tumor number confirming the critical role of Treg alone in tumor growth. Analysis of colon LP T-cell subsets revealed that Treg depletion resulted in global T-effector cell proliferation [fig_ref] Figure 1: Effect of CD4 + Foxp3 + Treg on tumorigenesis and T-effector cell... [/fig_ref]. Specifically, all effector T-cell subset numbers, including Th1, Th17 and CD8 + T-cells increased by~28, 7 and 55-fold, respectively. Suppression of tumor growth in Treg-depleted mice, despite continued Th17 cell expansion, suggested that Treg drove tumor growth independent of Th17 cells. Separately, in addition to dramaticallyincreased numbers, Th1 and CD8 + T-cells displayed enhanced activation as measured by IFNγ + production [fig_ref] Figure 1: Effect of CD4 + Foxp3 + Treg on tumorigenesis and T-effector cell... [/fig_ref] , establishing a link between Treg presence and cytotoxic T-effector function.
Treg expansion is driven by the CD11b + CX3CR1 + myeloid cell subset
The essential role of LP-resident CX3CR1 + myeloid cell-Treg axis in the maintenance of gut immune homeostasis is wellknown. [bib_ref] Interleukin-10 receptor signaling in innate immune cells regulates mucosal immune tolerance and..., Shouval [/bib_ref] [bib_ref] Critical role for the microbiota in CX3CR1(+) intestinal mononuclear phagocyte regulation of..., Kim [/bib_ref] Separately, recent studies established that ETBF colonization results in major quantitative and qualitative changes in myeloid cell populations that are found in the colonic LP. [bib_ref] The myeloid immune signature of enterotoxigenic bacteroides fragilis-induced murine colon tumorigenesis, Orberg [/bib_ref] We thus hypothesized that gut-resident CX3CR1 + myeloid cells may be critical to regulating post-infection Treg activity. To this end, we first analyzed the quantitative changes in CD11b + myeloid cells in the colon during the first week of colonization. [fig_ref] Figure 2: Post-ETBF myeloid cell subset kinetics and function [/fig_ref] shows that ETBF induced rapid quantitative changes in both the CD11b + CX3CR1 + and the CD11b + CX3CR1 − subsets. Specifically, CD11b + CX3CR1 + cell numbers increased by 6.5-fold on day 2 and peaked at 13-fold above background on day 4. CD11b + CX3CR1 − subset prevalence increased with a b c somewhat slower kinetics (~2-fold on day 2) but ultimately caught up (11-fold on day 4). These findings showed that the CD11b + CX3CR1 + subset expansion preceded the Treg response by~2-3 days, consistent with a role for this subset in driving Treg ontogeny. We then focused on the ability of CX3CR1 + vs CX3CR1subsets to promote Foxp3 expression in CD4 + T-cells. For this we utilized an in vitro co-culture system similar to one we used previously. 11 CD11b + CX3CR1 + and CD11b + CX3CR1 − subsets were sorted and co-incubated with purified CD4 + T-cells in culture in the presence of IL-2 for 4 days and analyzed for Foxp3 vs IL-17 expression (corresponding to the two subsets that have been linked to tumorigenesis in this model). The data are shown in [fig_ref] Figure 2: Post-ETBF myeloid cell subset kinetics and function [/fig_ref]. The findings demonstrate that the CX3CR1 + subset preferentially induced Foxp3 expression in CD4 + T-cells that were sorted from ETBF-colonized mice (8-fold increase) vs the CX3CR1 − subset (2-fold increase). In contrast, both myeloid cell subsets suppressed IL-17 expression in CD4 + T-cells, suggesting that post-ETBF Th17-cell expansion, which accompanied Treg expansion with similar kinetics (Supplemental [fig_ref] Figure 1: Effect of CD4 + Foxp3 + Treg on tumorigenesis and T-effector cell... [/fig_ref] was likely driven by another cell type.
Next, we wanted to determine whether the CX3CR1 + subset was essential to Treg expansion in vivo. For this we initially used a global CX3CR1 knockout model in the APC min/+ background. APC min/+ mice were backcrossed to CX3CR1 flstopflDTR/flstopflDTR mice to obtain APC min/+ mice homozygous for the modified DTR gene insertion downstream of the endogenous CX3CR1 promotor, which results in an inactive CX3CR1 gene (Supplemental [fig_ref] Figure 2: Post-ETBF myeloid cell subset kinetics and function [/fig_ref]. These mice were then administered ETBF and tumor development was evaluated 4 weeks post-infection. [fig_ref] Figure 3: Role of CX3CR1 + myeloid cells in Treg induction and T-effector cell... [/fig_ref] demonstrates that global loss of CX3CR1 expression led to a 5.5-fold reduction in colon tumor burden, establishing the critical requirement for CX3CR1 expression during tumorigenesis. Moreover, analysis of colon LP T-cell subsets revealed a significant~2-fold reduction in Treg accompanied with a 2-fold increase in IFNγ-producing CD8 + T-cells consistent with the [fig_ref] Figure 1: Effect of CD4 + Foxp3 + Treg on tumorigenesis and T-effector cell... [/fig_ref] data. In contrast, there was no significant effect on Th1/Th17 populations (Supplemental [fig_ref] Figure 3: Role of CX3CR1 + myeloid cells in Treg induction and T-effector cell... [/fig_ref]. These findings were consistent with our hypothesis that CX3CR1 + myeloid cells were important to Treg expansion and tumor growth in the above model. [fig_ref] Figure 3: Role of CX3CR1 + myeloid cells in Treg induction and T-effector cell... [/fig_ref] data convincingly demonstrated an important role for CX3CR1 in Treg expansion/tumor growth. At the same time, since CX3CR1 can be expressed by multiple leukocyte subsets we could not exclude the possibility of contribution by CX3CR1 + cells that are non-myeloid in origin to Treg expansion/tumor progression. To address this issue directly, we generated a double transgenic mouse in the APC min/+ background, i.e. LysM cre/+ CX3CR1 flstopflDTR/+ APC min/+ mice in which CX3CR1 + myeloid cells could be partially depleted in a lineage-specific manner by DT administration (Supplemental [fig_ref] Figure 2: Post-ETBF myeloid cell subset kinetics and function [/fig_ref]. These mice were administered ETBF and starting one week after bacterial inoculation were either treated with DT or PBS for 3 weeks. The colons were then analyzed for tumor burden and T-cell subset prevalence. the specific role of this subset in tumorigenesis/tumor growth [fig_ref] Figure 3: Role of CX3CR1 + myeloid cells in Treg induction and T-effector cell... [/fig_ref]. Furthermore, loss of CX3CR1 + myeloid cells resulted in a modest but significant decrease in Treg prevalence [fig_ref] Figure 3: Role of CX3CR1 + myeloid cells in Treg induction and T-effector cell... [/fig_ref] consistent with a central role for CD11b + CX3CR1 + myeloid subset in specifically driving the Treg-tumor axis. In this model, partial depletion of CX3CR1 + myeloid cells did not affect CD8 + T-cell activity [fig_ref] Figure 3: Role of CX3CR1 + myeloid cells in Treg induction and T-effector cell... [/fig_ref] or Th1/17 cell prevalence (Supplemental [fig_ref] Figure 3: Role of CX3CR1 + myeloid cells in Treg induction and T-effector cell... [/fig_ref].
CD11b + CX3CR1 + subset represents a homogenous CD64 + CD24 − CD103a − macrophage population
The above data confirmed the critical involvement of gut-resident CD11b + CX3CR1 + myeloid cells in post-ETBF Treg expansion. However, whether this population represented a unique cell type, or a heterogenous mix of dendritic cell (DC) and/or macrophage subsets that are found in the inflamed gut, 12,13 remained undetermined. To this end, further phenotypic analysis of CD11b + CX3CR1 + myeloid cells infiltrating the colonic LP in day 3 post-ETBF mice was undertaken. The data shown in [fig_ref] Figure 4: Phenotypic and morphological characterization of CD11b + CX3CR1 + myeloid cells [/fig_ref] demonstrate that this subset consisted overwhelmingly (≥90%) of CD64 + CD24 − CD103a − cells, consistent with macrophage lineage, as well as morphology. 14,15 Further analysis confirmed that CD64, CX3CR1 and CD103a expression patterns clearly distinguished these macrophages from CD11c + MHCII + CD64 − CD24 + CD103a + DC in both phenotype and morphology (Supplemental [fig_ref] Figure 4: Phenotypic and morphological characterization of CD11b + CX3CR1 + myeloid cells [/fig_ref]. Separately, CX3CR1 + macrophages expressed high levels of CD11c and MHCII, again consistent with that of an activated LP-resident subset [bib_ref] Origin, differentiation, and function of intestinal macrophages, Bain [/bib_ref] (Supplemental [fig_ref] Figure 4: Phenotypic and morphological characterization of CD11b + CX3CR1 + myeloid cells [/fig_ref].
## Preferential expansion of treg by cx3cr1 + macrophages requires ifnβ
Whereas the above findings supported a major role for the CX3CR1 + macrophages in promoting the expansion of Treg in the LP of ETBF-inflamed colon, they did not provide insight into the molecular mechanism that mediates this expansion. A single recent study reported a requirement for IFNβ in the maintenance of Treg in the gut by resident CX3CR1 + myeloid cells. [bib_ref] Apoptotic epithelial cells control the abundance of Treg cells at barrier surfaces, Nakahashi-Oda [/bib_ref] To determine whether a similar mechanism was responsible for the ontogeny of the Treg in the ETBF-infected colon, we first analyzed IFNβ expression in CD11b + cell subsets. CX3CR1positive and negative subsets were sorted and analyzed for IFNβ expression by qPCR. [fig_ref] Figure 5: Role of IFNβ in CX3CR1 + macrophage-mediated Treg expansion [/fig_ref] shows that CD11b + CX3CR1 + macrophages expressed significantly higher levels of IFNβ mRNA in comparison to CX3CR1 − subset both prior to (~5-fold) and subsequent to (~15-fold) bacterial inoculation. In contrast, no significant differences were observed in the relative levels of TNFα mRNA between the two subsets. Importantly, treatment of mice with a neutralizing anti-IFNβ antibody not only reduced Treg prevalence in the colonic LP but also resulted in a significant (~2.5-fold) reduction in tumor burden [fig_ref] Figure 5: Role of IFNβ in CX3CR1 + macrophage-mediated Treg expansion [/fig_ref]. The reduction in Treg/tumor burden was a b accompanied with increased CD8 + T-cell activity similar to that observed in [fig_ref] Figure 1: Effect of CD4 + Foxp3 + Treg on tumorigenesis and T-effector cell... [/fig_ref] , again consistent with a link between IFNβ and Treg activity. IFNβ neutralization resulted in a tendency toward increased Th1 and Th17 activity, which did not reach significance (Supplemental [fig_ref] Figure 5: Role of IFNβ in CX3CR1 + macrophage-mediated Treg expansion [/fig_ref]. Collectively, these findings strongly support a role, at least in part, for IFNβ in mediating the CX3CR1 + macrophage-mediated expansion of colonic Treg.
# Discussion
Our studies, for the first time, demonstrate that the expansion of Treg in the colonic LP of ETBF-colonized APC min/+ mice is driven by CX3CR1 + tissue-resident macrophages. The findings also establish that Treg expansion requires the production of IFNβ by these macrophages. Additionally, our data suggest that the Treg, independent of Th17 cells, are essential to tumor growth in this model. Collectively, this information sheds significant new light on the post-ETBF Treg ontogeny in the APC min/+ mouse colon. The finding that in the inflamed colon CX3CR1 + macrophages promote Treg activity in an IFNβ-dependent manner is consistent with the literature on the central role of the CX3CR1 + myeloid cell-Treg axis in maintaining steady-state tolerance in the gut. [bib_ref] Interleukin-10 receptor signaling in innate immune cells regulates mucosal immune tolerance and..., Shouval [/bib_ref] [bib_ref] Critical role for the microbiota in CX3CR1(+) intestinal mononuclear phagocyte regulation of..., Kim [/bib_ref] Our data confirm and extend the physiological relevance of this pathway to the regulation of microbially-induced inflammation in the dysplastic colon. Whereas we identified gut-resident CX3CR1 + macrophages as the primary driver of Treg expansion, our data cannot distinguish between the ontologically and functionally distinct subsets within this population. [bib_ref] Self-maintaining gut macrophages are essential for intestinal homeostasis, Schepper [/bib_ref] Therefore, whether a unique CX3CR1 + macrophage subpopulation is directly responsible for the activity observed here remains to be determined.
In addition to driving Treg expansion, CX3CR1 + macrophages actively inhibited IL-17 production in CD4 + T-cells. This observation suggests that co-expansion of the Th17 compartment in ETBF-colonized gut is driven by another cell type such as the recently-described inflammatory DC-like antigenpresenting cells that arise from Ly6C hi monocytes. [bib_ref] Inflammation switches the differentiation program of Ly6Chi monocytes from antiinflammatory macrophages to..., Rivollier [/bib_ref] [bib_ref] Ly6C hi monocytes in the inflamed colon give rise to proinflammatory effector..., Zigmond [/bib_ref] [bib_ref] Intestinal monocyte-derived macrophages control commensal-specific Th17 responses, Panea [/bib_ref] In this study we did not attempt to identify the cell type that promotes the expansion of Th17 cells or the mechanism by which CX3CR1 + macrophages suppress IL-17. One potential candidate for the latter is IFNβ, which has been reported to suppress IL-17 production in both murine and human CD4 + T-cells. [bib_ref] Type I IFN promotes IL-10 production from T cells to suppress Th17..., Zhang [/bib_ref] [bib_ref] The role of endogenous IFN-beta in the regulation of Th17 responses in..., Tao [/bib_ref] Previous studies showed that Treg can enhance Th17 expansion by acting as a sink for IL-2 [bib_ref] Regulatory T-cell response to enterotoxigenic bacteroides fragilis colonization triggers IL17-dependent colon carcinogenesis, Geis [/bib_ref]. Our data show that Th17 expansion persisted in the absence of Treg, and thus could occur independent of Treg help. One possible explanation for this seeming contradiction is that Treg act as a secondary enhancer, rather than the primary driver, of Th17 expansion. Alternatively, the timing of the Treg elimination may be important. That is, in the previous report Treg depletion was initiated prior to ETBF administration and performed for one week followed by analysis of the colon for Th17 activity. In contrast, we initiated depletion one week after ETBF colonization and administered DT for 3 weeks (starting DT prior to ETBF and continuing for 4 weeks resulted in significant mortality). It is therefore possible that the one-week period between ETBF infection and Treg depletion was sufficient for Treg to enhance Th17 cell polarization, which then expanded without further Treg help.
In this study, we did not address the mechanism that mediates tumor suppression in Treg-depleted mice. Our data showed that global elimination of CX3CR1 resulted in potent CTL activation, which would be expected to suppress tumor growth. Of note, this effect was strong enough to counter expanding protumorigenic Th17 cell activity, suggesting that the functional balance between CTL and Th17 is important to outcome in this model. Consistent with this notion, we previously reported that suppression of colon tumorigenesis by orally-administered IL-10 formulations was associated with the paradoxical ability of IL-10 to suppress Th17 cell activity while enhancing CD8 + T-cell cytotoxicity in the ETBF-colonized gut. [bib_ref] Oral IL-10 suppresses colon carcinogenesis via elimination of pathogenicCD4(+) T-cells and induction..., Gu [/bib_ref] Finally, our findings may have broader implications for the treatment of colon cancer, which (with the exception of a small microsatellite unstable subset) is unresponsive to immune therapy.Specifically, as ETBF has been associated with colorectal cancer (CRC) as well as familial adenomatous polyposis (FAP), IFNβ and IL-17 may represent relevant immunological targets in CRC and/or FAP patients.
# Materials and methods
Mice and the tumor model For colonization with ETBF, 5-6 week old APC min/+ mice (or transgenic mice in APC min/+ background) were administered clindamycin (0.1g/L) and streptomycin (5g/L) for 3-5 days before oral gavage (~5 × 10 7 bacteria in PBS) essentially as described. [bib_ref] A human colonic commensal promotes colon tumorigenesis via activation of T helper..., Wu [/bib_ref] All experiments were conducted in accordance with guidelines set forth by the Institutional Animal Care and Use Committees at the University of Louisville (Louisville, KY).
## Genetic models
For APC min/+ Foxp3 DTR/DTR mice, APC min/+ male mice were crossed to Foxp3 DTR/DTR female mice to generate APC min+ Foxp3-DTR/+ male mice. Then APC min+ Foxp3 DTR/+ male mice were crossed to Foxp3 DTR/DTR female mice for the second time to generate APC min/+ Foxp3 DTR/DTR mice. For APC min/+ CX3CR1 flstopflDTR/+ LysM cre/+ mice, APC min/+ male mice were initially crossed to CX3CR1 flstopflDTR/flstopflDTR mice to establish APC Min/ + CX3CR1 flstopflDTR/+ male mice. Then APC Min/+ CX3CR1 flstopflDTR/+ male mice were crossed to LysM cre female mice to establish APC min/+ CX3CR1 flstopflDTR/+ LysM cre/+ mice. Genotyping for each strain was performed using the primer sequences and PCR conditions recommended by the supplier (Jackson Laboratories, Bar Harbor, ME).
## Flow cytometry
MLN were processed into single cell suspensions and colons were digested and LP lymphocytes were isolated as described. [bib_ref] Isolation and subsequent analysis of murine lamina propria mononuclear cells from colonic..., Weigmann [/bib_ref] For experiments requiring detection of intracellular antigens, cell suspensions were cultured in the presence of Golgistop (5 μL/mL; BD), phorbol myristate acetate (50 ng/mL; Sigma) and ionomycin (1mg/ mL Sigma). Cells were then permeabilized and fixed using an intracellular staining kit (eBioscience). The following antibodies were used: CD4 (RM4-5, eBioscience), CD8α (53-6.72, Biolegend), CD11b (M1/70, Biolegend), CD45R/B220 (RA3-6B2, BD Pharmingen), CX3CR1 (SA011F11, Biolegend), CD11c (HL3, BD), MHCII (M5/114, Biolegend), CD24 (M1/69, Biolegend), CD64 (X54-5/7.1, Biolegend), CD103a (2E7, Biolegend), IL-17A (TC11-18H10, Biolegend), Foxp3 (FJK-16s, eBioscience) and IFNγ (XMG1.2, BD Pharmingen).
# Cytospin analysis
CD11b + CX3CR1 + or CD11c + CD24 + CD64 − cells were sorted to near homogeneity (>93%, FACSARIA III) from colonic lamina propria of day 3 APC min/+ /B fragilis mice, were stained with Wright-Giemsa, and were evaluated by light microscopy (800x magnification).
## Diphtheria toxin (dt) administration
For Treg depletion, APC min/+ Foxp3 DTR/DTR mice received 1μg of DT (Sigma) dissolved in 100μl of Dulbecco PBS by intraperitoneal injection on days 7, 14, and 21 after ETBF inoculation. [bib_ref] Oral interleukin-10 alleviates polyposis via neutralization of pathogenic T-regulatory cells, Chung [/bib_ref] APC min/+ CX3CR1 flstopflDTR/+ LysM cre/+ mice also received 1μg of DT dissolved in 100μl of Dulbecco PBS by intraperitoneal injection on days 7, 14, and 21 after ETBF inoculation for lineage-specific knockdown of CX3CR1 + myeloid cells.
## Ifnβ neutralization
Anti-mIFN-β mAb (HDβ-4A7, Leinco Technologies) was given intraperitoneally to APC min+ mice (0.25mg in 0.25ml PBS, three times per week for 3 weeks) to neutralize IFN-β (PBS as control). The above treatment was initiated 1 week after ETBF inoculations.
## Quantitative real-time pcr
Total RNA was extracted using Trizol (Invitrogen/ThermoFisher). The Trizol solution was then added to the Zymo-Spin IIC column for further purification per the manufacturer's instructions (Zymo Research, Irvine, CA). DNA was removed by adding rDNase I (Ambion/ThermoFisher) to the RNA solution. RNA was then eluted in DNase/RNase free water and immediately processed for reverse transcription. Complementary DNA (cDNA) was synthesized by using TaqMan reverse Transcription reagents (Applied Biosystems/ThermoFisher). The prepared cDNAs were amplified using iTaq Universal SYBR Green supermix kit (Bio-Rad, Hercules, CA) using the Stratagene Mx3005P system with the following primers: IFN-β Fwd: 5ʹ-CCAGCTCCAAGAAAGGACGA −3ʹ, Rev: 5ʹ-CGCCCTGTAGGTGAG GTTGAT −3ʹ; TNFα Fwd 5ʹ-GAA CTG GCA GAA GAG GCA CT −3ʹ and Rev 5ʹ-AGG GTC TGG GCC ATA GAA CT −3ʹ; β-actin Fwd 5ʹ-TCA CCC ACA CTG GCC CAT CTA CGA −3ʹ and Rev 5ʹ-TGG TGA AGC TGT AGC CAC GCT −3ʹ. Relative quantitative measurement of target gene levels was performed using the 2 −ΔΔΔΔCt method. βactin was used as the endogenous housekeeping control gene.
## In vitro cell culture
Lymphocytes (CD45 + CD4 + , CD45 + CD11b + CX3CR1 + and CD45 + CD11b + CX3CR1 − cells) were sorted to >90% purity from MLN of APC min/+ mice 3 days after ETBF inoculation on a FACSAria III (BD Pharmingen). Foxp3-induction assays were performed by in vitro co-culture of 1 × 10 5 /ml CD45 + CD11b + CX3CR1 + or CD45 + CD11b + CX3CR1 − cells together with 1 × 10 6 /ml CD45 + CD4 + T cells and 40U/ml IL-2 (Peprotech, Inc) for 4 days.
# Statistical analysis
Statistical calculations were performed using Student t test in pairwise comparisons of groups. In experiments with multiple groups homogeneity of inter-group variance was analyzed by one-way ANOVA. P values of 0.05 or less were considered statistically significant.
## Abbreviations
## Disclosure of potential conflicts of interest
No potential conflicts of interest were disclosed.
# Funding
The work was supported by the University of Louisville start-up funds (N.K.E.).
## Anova
[fig] Figure 1: Effect of CD4 + Foxp3 + Treg on tumorigenesis and T-effector cell activity. Panel a. Kinetics of Treg expansion in ETBF-colonized colon. Colons were harvested on indicated days post-ETBF colonization and total LP Treg numbers were determined per colon. Error bars = SEM, n = 3/time point. Days 5-21 were significantly different from Day 0 (p < .01). Panel b. Colon tumor burden in the presence and absence of Treg. Tumor burden was quantified 3 weeks after ETBF administration in control and DT-administered mice. Error bars = SEM, n = 3/time point (** denotes p < .01). Panel c. The colons from control and DT-administered mice were harvested and LP T-effector subsets were quantified. Representative flow panels showing percent cytokine-producing cells and histograms depicting respective T-effector cell numbers/colon are shown. Error bars = SEM, n = 3/time point (* and *** denote p < .05 and 0.001, respectively). Data from one of two repeats with similar results are shown. [/fig]
[fig] Figure 2: Post-ETBF myeloid cell subset kinetics and function. Panel a. CD11b + myeloid cell subset kinetics in the colon. LP CD11b + CX3CR1 + and CD11b + CX3CR1 − myeloid cell numbers were quantified in post-ETBF mice. Error bars = SEM, n = 3-6/time point. Days 2-6 were significantly different than day 0 (p < .05). Panel b. In vitro co-culture assay. Sorted CD11b + cell subsets were incubated with CD4 + T-cells as described in the Methods. Control T-cells were isolated from naïve APC min/+ mice whereas experimental T-cells (Bf) were purified from ETBF-inoculated mice. Representative FACS panels are shown for control (T-cells only, top row) and experimental (T-cells + myeloid cells, bottom row). CD4 + T-cells were gated on and analyzed for Foxp3 vs IL-17 production. Quantitative data are shown for percent Foxp3 + and IL-17 + cells for each well. Error bars = SEM, n = 3 mice per group (** and *** denote p < .01 and < 0.001, respectively). Data from one of two repeats with similar results are shown. [/fig]
[fig] Figure 3: Role of CX3CR1 + myeloid cells in Treg induction and T-effector cell activity in vivo. Panel a. Effect of global knockout of CX3CR1 + on tumor burden and T-cell activity. Control APC min/+ or APC min/+ CX3CR1 flstopflDTR/flstopflDTR mice were inoculated with ETBF and tumor burden/colon LP T-cell activity were analyzed 3 weeks after bacteria administration. Representative FACS panels and quantitative data for Treg and CD8 + T-cells are shown. Error bars = SEM, n = 5-6 mice per group (*, ** and *** denote p < .05, 0.01 and 0.001, respectively). Panel b. Effect of CX3CR1 + myeloid cell depletion on tumor burden and T-cell activity. Control or DT-administered LysM cre/+ CX3CR1 flstopflDTR/+ APC min/+ mouse colons were analyzed 3 weeks post-ETBF for tumor burden and T-cell activity. Representative FACS panels and quantitative data for Treg and CD8 + T-cells are shown. Error bars = SEM, n = 6 mice per group. ** denotes p < .01. [/fig]
[fig] Figure 4: Phenotypic and morphological characterization of CD11b + CX3CR1 + myeloid cells. Panel a. LP CD45.2 + cells that were CD11b + CX3CR1 + were analyzed for CD24 and CD64 expression. CD11b + CX3CR1 + CD64 + macrophages were then assessed for CD103a, CD11c and MHCII (shaded histograms = fluorescence minus one control; open histograms = antibody staining). Panel b. CD11b + CX3CR1 + subset was sorted and analyzed by cytospin. Intact cells demonstrated abundant foamy cytoplasm with prominent cytoplasmic vacuoles typical of macrophage morphology. Magnification = 800x. [/fig]
[fig] Figure 5: Role of IFNβ in CX3CR1 + macrophage-mediated Treg expansion. Panel a. Relative IFNβ expression in CD11b + myeloid cell subsets. CD11b + CX3CR1 + or CD11b + CX3CR1 − cells from the colon LP were sorted, RNA extracted and IFNβ mRNA levels were quantified by qPCR. TNFα levels were also analyzed as an independent variable. Relative ratios of cytokine levels (CD11b + CX3CR1 + / CD11b + CX3CR1 − ) prior to and 3 days after ETBF colonization are shown. Average of 3 technical replicates from pooled samples (5 mice/group) were calculated for CX3CR1 + and CX3CR1 − subsets and relative proportion was determined. Panel b. In vivo IFNβ neutralization. Effect of anti-IFNβ antibody on tumor burden and T-cell activity in ETBF colonized mice was determined. Representative FACS panels and quantitative data (% Foxp3 + CD4 T-cells and number of IFNγ + CD8 T-cells per colon) are shown. Error bars = SEM, n = 4 mice per group. * and ** denote p < .05 and 0.01. [/fig]
[fig] C57BL/ 6: B6), C57BL/6J-Apc Min /J (APC min/+ ), B6.129(Cg)-Foxp3 tm3(DTR/GFP)Ayr /J (Foxp3 DTR/DTR ), B6N.129P2-Cx3cr1 tm3(DTR)Litt /J (CX3CR1 flstopflDTR/flstopflDTR ) and B6.129P2-Lyz2 tm1(cre)Ifo /J (LysM cre ) mice were purchased from Jackson Laboratory. Enterotoxic B fragilis strain 86-5443-2-2 (ETBF) which secretes B. fragilis toxin 2 (BFT-2) was a kind gift from Dr. Cynthia L Sears (Johns Hopkins University School of Medicine, Baltimore, Maryland). ETBF was grown under anaerobic conditions at 37ºC overnight prior to administration to mice. [/fig]
|
Genome survey and SSR analysis of Apocynum venetum
Apocynum venetum is an eco-economic plant that exhibits high stress resistance. In the present paper, we carried out a whole-genome survey of A. venetum in order to provide a foundation for its whole-genome sequencing. High-throughput sequencing technology (Illumina NovaSep) was first used to measure the genome size of A. venetum, and bioinformatics methods were employed for the evaluation of the genome size, heterozygosity ratio, repeated sequences, and GC content in order to provide a foundation for subsequent whole-genome sequencing. The sequencing analysis results indicated that the preliminary estimated genome size of A. venetum was 254.40 Mbp, and its heterozygosity ratio and percentage of repeated sequences were 0.63 and 40.87%, respectively, indicating that it has a complex genome. We used k-mer = 41 to carry out a preliminary assembly and obtained contig N50, which was 3841 bp with a total length of 223949699 bp. We carried out further assembly to obtain scaffold N50, which was 6196 bp with a total length of 227322054 bp. We performed simple sequence repeat (SSR) molecular marker prediction based on the A. venetum genome data and identified a total of 101918 SSRs. The differences between the different types of nucleotide repeats were large, with mononucleotide repeats being most numerous and hexanucleotide repeats being least numerous. We recommend the use of the '2+3' (Illumina+PacBio) sequencing combination to supplement the Hi-C technique and resequencing technique in future whole-genome research in A. venetum.
# Introduction
Apocynum venetum (Apocynaceae), also known as sword-leaf dogbane, is a perennial grass that is a valuable wild plant resource in China. It is also known as 'the king of wildlife fibers' due to the excellent quality of its fiber products. It was first discovered in the Luopu plains in Xinjiang province in the 1950s and was named 'Luobu hemp' [bib_ref] A new advanced textile fiber plant in China-Apocynum, Dong [/bib_ref]. It is an eco-economic plant with high stress resistance and is widely distributed in saline and alkaline deserts, desert boundaries, river banks, alluvial plains, lakes, and the Gobi Desert in China at latitudes of 35 - -45 - N. Its leaves can be used to brew tea; its stems can be made into fibers; and it is also a source of honey. The roots, stems, and flowers of A. venetum are used in medicinal preparations. In 1977, A. venetum was listed in the Pharmacopoeia of the People's Republic of China as a primary treatment for hypertension and hyperlipidemia, and is particularly suitable for treating constipation, obesity, and heart palpitations in middle-aged and elderly people. As a plant with important economic value, research on A. venetum has attracted widespread attention in recent years. Relevant studies have mainly focused on the physiological characteristics [bib_ref] Daily dynamics of photosynthesis and water physiological characteristics of Apocynum venetum and..., Wang [/bib_ref] [bib_ref] The preliminary research on photosynthetic characteristics of Apocynum venetum under different shading, Zhang [/bib_ref] , pollination characteristics [bib_ref] Comparative pollination biology of Apocynum venetum at different desert landscapes, Chen [/bib_ref] , medicinal components [bib_ref] Apocynum venetum leaf extract reverses depressive-like behaviors in chronically stressed rats by..., Li [/bib_ref] , germplasm resources [bib_ref] Review of current research and utilization status of Apocynum venetum germplasm in..., Xu [/bib_ref] , tissue culture [bib_ref] Research on culture in vitro and rapid multiplication techniques of Apocynum venetum, Chen [/bib_ref] [bib_ref] Selection of explants and callus induction of Apocynum venetum, Liu [/bib_ref] , and chalcone synthase (CHS) cloning [bib_ref] Cloning and sequence analysis of chalcone synthase gene (CHS) from Apocynum venetum, Li [/bib_ref] of A. venetum, while very few studies have assessed the genetics of A. venetum. There are only 72 sequences for A. venetum available on the NCBI database (as of 21 July 2018). Limited gene sequence resources have restricted any in-depth research into the molecular biology of A. venetum, and thus the whole-genome sequencing of this economically valuable plant is warranted.
All life phenomena in organisms are associated with genes. The whole-genome sequencing of A. venetum is thus required to elucidate its high stress resistance and the characteristics associated with its high economic value. Omics technologies are constantly developing, particularly in genomics. Third-generation technologies have matured and are presently widely used, providing strong technical support for genome sequencing. In 2000, the whole-genome sequencing of a higher plant, Arabidopsis thaliana, was completed [bib_ref] Analysis of the genome sequence of the flowering plant arabidopsis thaliana, Initiative [/bib_ref] , which provided a foundation for whole-genome sequencing research in plants. Subsequently, whole-genome sequences from rice (Oryza sativa) [bib_ref] A draft sequence of the rice genome (Oryza sativa L. ssp. japonica), Goff [/bib_ref] , sorghum (Sorghum bicolor) [bib_ref] The Sorghum bicolor genome and the diversification of grasses, Paterson [/bib_ref] , maize (Zea mays) [bib_ref] The B73 maize genome: complexity, diversity, and dynamics, Schnable [/bib_ref] and many other plants have been published. This has provided a technical foundation for the genome sequencing of other plants. In order to fully mine molecular biology data from A. venetum, analyze crucial genes in the synthesis of A. venetum fibers, determine stress-related functional genes, and understand the nature of its stress resistance, we have sequenced and analyzed the genome of A. venetum in the present study. This will not only have significance for genomics and evolutionary research, but will also provide a foundation for the exploitation of the economic value of A. venetum. Before carrying out large-scale whole-genome deep sequencing in plants, it is necessary to first conduct low-coverage genome analysis and simple sequence repeat (SSR) molecular marker analysis in order to understand the compositional characteristics and patterns of the entire genome. K-mer analysis of low-depth sequencing data based on fragment libraries can be used to effectively assess genome size, GC content, heterozygosity ratio, and repeated sequences, and enables us to comprehensively evaluate the genome characteristics. This provides a basis for subsequent de novo sequencing and whole-genome assembly studies in A. venetum.
# Materials and methods
# Experimental materials
The A. venetum roots were collected from the A. venetum experimental base in Yinchuan City, Ningxia Province, at the end of March 2018 and were planted in pots. In early May, healthy plants exhibiting good growth were selected, and young leaves and stems at the apex were collected, snap-frozen in liquid nitrogen, and stored at −80 - C until subsequent experiments.
# Experimental methods
## Sample extraction and measurement
The modified cetyltrimethylammonium bromide (CTAB) method was used to extract A. venetum genomic DNA [bib_ref] ESTSSRs development and paternity analysis for Liriodendron spp, Xu [/bib_ref]. UV spectrophotometry was used to measure the concentration of the template (the ratio of absorbance at 260 and 280 nm was used to determine the purity and extraction results of the DNA). Agarose gel electrophoresis was used to determine the integrity of the template.
## Sequencing data generation and quality control
The A. venetum DNA samples deemed to be of suitable quality were randomly sheared into 350-bp fragments using an ultrasonicator (Covaris Inc.). Electrophoresis was used to recover the DNA fragments of required lengths before end-repair, following which poly A-tail and sequencing adapters were added. The obtained fragments were purified before PCR amplification for library preparation. The Illumina NovaSep platform was used for high-throughput paired-end sequencing of the constructed libraries. In order to ensure the quality of the analysis, we filtered reads that would interfere with subsequent information, reads with adapters, reads with an N (unable to determine base information) ratio greater than 10%, and low-quality reads from the raw reads to obtain clean reads. The entire genome sequencing was carried out by Novogene Co. Ltd. (Beijing, China).
17-mer analysis and prediction of genome size, heterozygosity ratio, and repeated sequences K-mer analysis was used to predict the genome size, heterozygosity ratio, and repeated sequences of the clean reads before genome assembly. A K-value of 17 was used for the prediction, analysis, and iterative selection of 17-bp base sequences from the clean reads. We assumed that the K-mer depth frequency distribution followed a Poisson distribution and that all K-mers that were obtained base-by-base from the reads could cover the entire genome. Following this, we tallied the K-mer frequency distribution and calculated the K-mer depth distribution curve and depth product curve. The estimated K-mer depth value was obtained from the curve and used to estimate genome size [bib_ref] The genome of the cucumber, Cucumis sativus L, Huang [/bib_ref] [bib_ref] De novo assembly of human genomes with massively parallel short read sequencing, Li [/bib_ref]. The tailing phenomenon of the K-mer depth distribution curve was used to estimate the repeated sequences in the genome. A heterozygous genome includes two types: heterozygous K-mer and homozygous K-mer. Assuming that each heterozygous site is covered by 2×K K-mers, the expected depth of the heterozygous K-mer is 1 / 2 . Therefore, the number of heterozygous sites can be estimated using 1 / 2 (percentage of heterozygous K-mer types) and n Kspecies (total number of all K-mer types). The heterozygosity ratio was calculated using (eqn 1). The proportion of repeated sequences can be obtained by calculating the proportion of the number of K-mers greater than 1.8-fold of the homozygous peak depth.
[formula] ϕ = a1/2 × nKspecies/ (2 × K ) nKspecies − a1/2 × nKspecies/2 = a1/2 K (2 − a1/2)(1) [/formula]
where a 1/2 is the percentage of heterozygous K-mer types and n Kspecies is the total number of all K-mer types.
## Preliminary genome assembly
SOAPdenovo2 software was used to carry out preliminary genome assembly [bib_ref] The sequence and de novo assembly of the giant panda genome, Li [/bib_ref]. First, low frequency K-mers were corrected before K-mer = 41 was used to cut the corrected fragment library reads into even smaller sequence fragments. The overlap between these reads was used to construct a de Bruijn graph. A simplified de Bruijn graph was obtained from selection, simplification, and merging, and the sequences at every bifurcation locus were truncated to obtain the initial contigs. The reads obtained from sequencing all the libraries were aligned to the initially obtained contigs. The connectivity relationship between the reads and the information of the inserted fragment size were used to further assemble the contig into a scaffold and obtain the primary genomic sequence containing Ns. Following this, the filtered reads were aligned to this assembled sequence using SOAP to obtain the base depth. A window size of 10 kb was used for non-repetitive advancement in the sequence and calculation of the mean depth and GC content of every window to generate a GC depth plot. The graph could be used to examine whether there was significant GC bias or bacterial contamination in the sequences. Further, the stratification of GC clusters could be used to determine the heterozygosity ratio and percentage of repeated sequences of the genome [bib_ref] Genome survey analysis in Siraitia grosvenorii, Tang [/bib_ref].
## Analysis of ssr molecular markers
MicroSatellite identification tool (MISA) software (http://pgrc.ipk-gatersleben.de/misa/misa.html), which was written using Perl, was used for the analysis of SSR molecular markers in the genome, to identify all microsatellite repeat units in the genome sequence, and to calculate the location, length, quantity, start sites, and end site of the SSRs in the scaffold. The parameters were set prior to MISA operation as follows: number of mononucleotide repeats ≥ 10, number of dinucleotide repeats ≥ 6, and number of trinucleotide, tetranucleotide, pentanucleotide, and hexanucleotide repeats ≥ 5.
# Results and discussion
## Sequencing data statistics and quality evaluation of a. venetum
The Illumina NovaSep platform was used for high-throughput paired-end sequencing to obtain 36.62 Gb of A. venetum raw bases. After filtering low-quality data, we obtained 36.61 Gb of clean bases, which accounted for 99.96% of the raw bases. In second-generation sequencing, a corresponding quality value will be obtained by sequencing every base. This quality value is an important marker for measuring sequencing accuracy. The higher the quality value (Q), the lower the probability that the base was incorrectly sequenced. The Illumina platform specifies that the Q20 and Q30 should be at least 90 and 85%, respectively. The sequencing quality evaluation showed that Q20 and Q30 were 96.09 and 89.95%, respectively [fig_ref] Table 1: Data statisticsAbbreviations [/fig_ref]. This indicates that the high-throughput sequencing of A. venetum was highly accurate. [fig_ref] Figure 1: Distribution figure of GC contentThe left half of the dotted line in... [/fig_ref] shows the proportion of single bases, which is used to detect whether AT and GC separation is present. It can be seen that the content of A and G and C and T are close and the N content is almost zero. The results demonstrated that the sequencing quality was good.
## Results of 17-mer analysis and prediction of genome size, heterozygosity ratio, and percentage of repeated sequences
We used 36.61 Gb of valid A. venetum genome data for K-mer analysis using a K-value of 17 and obtained a frequency distribution graph [fig_ref] Figure 2: Distribution curve of K-merIt is an analysis of the genome size prediction... [/fig_ref]. The x-coordinates represent the K-mer depth and the y-coordinates represent the number of K-mers for the corresponding depth. From [fig_ref] Figure 2: Distribution curve of K-merIt is an analysis of the genome size prediction... [/fig_ref] , we can see that a depth = 106 is near the main peak, i.e. the expected depth of K-mers. SOAPdenovo software calculated the total number of K-mers as 27.48 Gb. A formula (genome size = K-mer number/K-mer depth) estimated the genome size to be approximately 259. [bib_ref] Estimation of genome size of Moso Bamboo, Li [/bib_ref] Mbp. After excluding the error effects due to erroneous K-mers, the corrected genome size was found to be 254.40 Mbp. (eqn 1) was used for calculating the proportion of heterozygous sites in the sequence, obtaining a gene heterozygosity ratio of 0.63%. By calculating the percentage of 1.8-times the number of K-mers after the main peak over the total number of K-mers, we obtained a percentage of repeated sequences of 40.87%.
## Preliminary genome assembly results for a. venetum
The 36.61 Gb of clean bases was used for preliminary genome assembly, and a K-mer value of 41 was selected to construct the contig and scaffold, obtaining optimal assembly results. The results are shown in [fig_ref] Table 2: Statistics of the assembled genome sequences in A [/fig_ref]. We obtained a total of 282245 contigs with a total length of 223949699 bp, and the longest assembled sequence had a length of 134265 bp. The length of the N50 contig was 3841 bp. We obtained 239333 scaffolds after further assembly with a total length of 227322054 bp, and the longest sequence assembled was 191270 bp. The length of the N50 scaffold was 6502 bp. The results from [fig_ref] Figure 3: Distribution figure of contig coverage depth and lengthIn the figure, the peak... [/fig_ref] showed significant peaks. It can be determined that the peak value at approximately 82x is a homozygous peak. The peak that is located around half of the x-coordinates in front of the homozygous peak is the heterozygous peak. Therefore, the genome of A. venetum is heterozygous and complex.
## Gc content and distribution status
GC depth analysis [fig_ref] Figure 5: Distribution figure of GC depth [/fig_ref] indicated that the GC content of almost all windows was 20-60%, and the sequencing depth was greater than 20-fold. The A. venetum samples did not show any apparent abnormalities, and there was no obvious GC bias. The A. venetum GC depth distribution was divided into two layers, and low depth was found in one region. The sequences of the low-depth distribution were extracted and Blast software was used to align these sequences to the NCBI Nucleotide (Nt) database. The results showed that these samples do not contain exogenous contamination. The GC clusters were divided into two obvious layers, which may be due to the heterozygosity. This is because heterozygosity causes the two homologous chromosomes at the heterozygous site to assemble into one or two strands. Additionally, the read products of these sites are half of the entire genome product. This causes a lower layer to appear in the GC content graph.
## Whole-genome ssr sequencing analysis of a. venetum
Using the MISA script, a total of 101918 SSRs in the A. venetum genome and the repeat units were used for the classification and calculation of nucleotide repeats. There were 65220 mononucleotide repeats, accounting for 63.99% of the total; 29276 dinucleotide repeats, accounting for 28.73% of the total; and 6632 trinucleotide repeats, accounting for 6.51% of the total. Tetranucleotide, pentanucleotide, and hexanucleotide repeats accounted for 0.52, 0.16, and 0.10% of the total, respectively. Therefore, the mononucleotide repeat was the main form in the A. venetum genome, while the proportion of hexanucleotide repeats was the lowest. Among the mononucleotide SSR repeats, A/T repeats were the most common. Among the dinucleotide SSR repeats, most repeats were AT/TA, accounting for 20.27% of the total, while CG/GC only accounted for 0.02% of the total [fig_ref] Table 3: Type and proportion of SSR [/fig_ref].
# Discussion
## Prediction of genome size
Whole-genome sequencing is a modern tool that enables the examination of the genetic code of plants. The whole-genome sequencing and construction of genome-wide maps in plants has promoted modern life science research. Genome size refers to the DNA content of all biological haploids and is usually described using the C-value. This value is a constant for every species and shows species-specific characteristics [bib_ref] Nuclear DNA content and genome size of troutand human, Dolezel [/bib_ref]. The genome can reflect all the genetic information of a biological species, and increasing studies have found that genome size is related to different biological parameters, such as stress resistance and economic characteristics. Increasingly, studies have found that genome size is related to different biological parameters, such as cell cycle and cell size, and genome size plays an important role in plant evolution and adaptation [bib_ref] The evolutionary mechanism of genome size, Shi [/bib_ref]. Methods for studying genome size range from renaturation kinetics [bib_ref] Herpes simplex virus: genome size and redundancy studied by renaturation kinetics, Frenkel [/bib_ref] , to pulsed-field gel electrophoresis [bib_ref] Genome size of Myxococcus xanthus determined by pulsedfield gel electrophoresis, Chen [/bib_ref] , to flow cytometry [bib_ref] Flow cytometric approach to study genome size variation in eurasiatic green toads..., De Vita [/bib_ref] [bib_ref] Estimation of genome size of Moso Bamboo, Li [/bib_ref] , to modern high-throughput sequencing and K-mer estimation. These methods are becoming increasingly convenient and the results are becoming more accurate. K-mer estimation of genome characteristics can greatly assist in exploring the genomes of unknown species. This method was successfully applied in the prediction of genome sizes for camphor tree (Cinnamomum camphora) [bib_ref] Genome survey in Cinnamomum camphora L, Wu [/bib_ref] , monk fruit (Siraitia grosvenorii) [bib_ref] Genome survey analysis in Siraitia grosvenorii, Tang [/bib_ref] , Chinese tulip tree (Liriodendron chinense), Ammopiptanthus mongolicus [bib_ref] Genomic survey sequencing and estimation of genome size of Ammopiptanthus mongolicus, Wang [/bib_ref] , and many other species. There is great variation in genome sizes among different species. Among angiosperm plants, the species with the smallest genome is Genlisea tuberosa from the family Lentibulariaceae, with a genome size of 61 Mbp [bib_ref] Evolution of genome size and chromosome number in the carnivorous plant genus..., Fleischmann [/bib_ref] ; while the species with the largest genome is the canopy plant (Paris japonica) from the family Melanthiaceae, with a genome size of 150 Gb [bib_ref] The largest eukaryotic genome of them all, Pellicer [/bib_ref]. The difference in genome size between these two species is approximately 2400-times. In the present study, the genome size of A. venetum was determined to be 254.40 Mbp [fig_ref] Table 1: Data statisticsAbbreviations [/fig_ref] , which is close to the genome size of Oropetium thomaeum (245 Mbp) and Kalanchoe fedtschenkoi (260 Mbp). A comparison of the published genome sizes of plants suggests that the A. venetum genome is relatively small, indicating that future genome assembly and annotation should be relatively simple. We have completed the genome survey of A. cannabinum and found its genome to be 239.02 Mbp in size, which is smaller than that of A. venetum [bib_ref] Whole genome sequencing and development of SSR markers in Apocynum cannabinum, Song [/bib_ref].
The genome survey of A. venetum has provided a solid foundation for its whole-genome sequencing, which will greatly accelerate and promote the exploration of this eco-economic plant, particularly its medicinal value. As a traditional medicine of China, A. venetum has long been used to calm the nerves, and promote diuresis, and the tea of roasted A. venetum leaves has been commercialized as a sedative and anti-aging supplement. Modern medical and pharmacological studies have indicated that A. venetum has broad pharmacological activities that include antihypertensive, cardiotonic, hepatoprotective, antioxidant, lipid-lowering, antidepressant, and anxiolytic effects. In additional, as an ethnopharmacological herb in Lop Nor region of China, A. venetum leaves have even been added to tobacco to detoxify nicotine [bib_ref] Botany, traditional uses, phytochemistry and pharmacology of Apocynum venetum L. (Luobuma): a..., Xie [/bib_ref]. One study verified that A. venetum leaf extract inhibits aortic contraction via its superoxide anion scavenging properties and nitric oxide releasing effect, which may account for its use as an antihypertensive treatment in traditional folk medicine. Furthermore, an aqueous extract of A. venetum leaves could inhibit sodium channels, thus affecting neurotransmission and modulating neuronal ion channels, and thus may exert neuropharmacological effects [bib_ref] Apocynum venetum leaf aqueous extract inhibits voltage-gated sodium channels of mouse neuroblastoma..., Kuo [/bib_ref].
## Genome gc content, heterozygosity ratio, and percentage of repeated sequences in a. venetum
The GC content of the A. venetum genome is 32.91% [fig_ref] Table 2: Statistics of the assembled genome sequences in A [/fig_ref] , which is similar to A. mongolicus (36.51%) and A. thaliana . It is lower than that of rice, maize, and sorghum, which have a GC content of more than 40%, but is higher than that of apples (Malus pumila) and alfalfa (Medicago sativa), which have a GC content of less than 30%. Shangguan et al. [bib_ref] Evaluation of genome sequencing in selected plant species using expressed sequence tags, Shangguan [/bib_ref] summarized most of the available data on plant genomes and found that their GC contents mostly range within 30-47%. Another study also showed that excessively high (>65%) or excessively low (<25%) GC contents will result in errors in high-throughput sequencing and affect the accuracy of spliced data [bib_ref] Analyzing and minimizing PCR amplification bias in Illumina sequencing libraries, Aird [/bib_ref]. The heterozygosity ratio and percentage of repeated sequences in the sequencing data have important significance for guiding the assembly and splicing of genomes. In the present study, we calculated the heterozygosity ratio of A. venetum to be 0.63%. In the contig distribution graph [fig_ref] Figure 3: Distribution figure of contig coverage depth and lengthIn the figure, the peak... [/fig_ref] , the peak located halfway in front of the main peak is the heterozygous peak. This also proved the existence of heterozygosity in the A. venetum genome. The size of the heterozygosity ratio can usually be used to divide genomes into low heterozygosity (0.5% ≤ heterozygosity ratio <0.8%), high heterozygosity (heterozygosity ratio ≥0.8%), highly repetitive genome (percentage of repeated sequences ≥50%), and low repetitive genomes (percentage of repeated sequences <50%) [bib_ref] Genome survey in Cinnamomum camphora L, Wu [/bib_ref]. This preliminary analysis determined the A. venetum genome to have low heterozygosity. A K-mer distribution curve [fig_ref] Figure 2: Distribution curve of K-merIt is an analysis of the genome size prediction... [/fig_ref] was calculated, obtaining a repeated sequence percentage of 40.87%. Thus, the genome of A. venetum is complex with low heterozygosity and low repeated sequences. Repeated sequences are one of the major factors that control the recombination and regulation of structural genes and are also important components of non-coding regions. Repeated sequences exist in a state of dynamic variation, and gene expression may be precisely regulated by repeated sequences. Currently, there are three hypotheses to explain the existence of repeated sequences, namely the mutation-drift model, adaptive theory, and transposition mechanisms [bib_ref] The evolutionary mechanism of genome size, Shi [/bib_ref]. However, there is still a lack of understanding of the functions of repeated sequences in A. venetum, and further investigation is required.
## Whole-genome ssr marker characteristics of a. venetum
Molecular markers are an ideal form of genetic marker. In addition to facilitating detection, multiple allele polymorphism, and codominant inheritance, molecular markers also possess advantages that are not found in rapid fragment length polymorphism (RFLP) and amplified fragment length polymorphism (AFLP) markers. We found that mononucleotides were the most common SSR loci, with A/T ratios greater than the G/C ratio, accounting for 57.97% of the total number of repeat units. Among the dinucleotide repeats, AT/TA repeats were highest, accounting for 20.27% of the total number of repeat units, while CG/GC repeats were lowest, accounting for only 0.02% of the total. This may be due to methylation of cytosine into thymidine, resulting in a larger difference between these two nucleotide repeats [bib_ref] Analysis of SSRs in grape genome and development of SSR database, Cai [/bib_ref]. Among the trinucleotides, the levels of AAT/TTA were the highest. This is similar to the distribution of trinucleotide sequences in A. mongolicus and grapes [bib_ref] Genome survey and SSR analysis of Ammopiptanthus mongolicus, Zhou [/bib_ref]. Statistical analysis of the differences in the quantity and types of SSRs in A. venetum and an initial exploration of the genome data have provided a foundation for the further construction of high-density genetic maps and the study of the regulatory mechanisms of A. venetum under stress conditions. This also provides a good reference for future genome and molecular marker research.
# Conclusion
This is the first study to measure the size of the entire A. venetum genome and preliminarily assess the corresponding parameters. The size of the A. venetum genome was estimated to be 259. [bib_ref] Estimation of genome size of Moso Bamboo, Li [/bib_ref] Mbp, which was corrected to 254.40 Mbp.
The heterozygosity ratio was 0.63% and the percentage of repeated sequences was 40.87%. We used K-mer = 41 to carry out a preliminary assembly and determined contig N50 to be 3841 bp, with a total length of 223949699 bp, while scaffold N50 was determined to be 6196 bp, with a total length of 227322054 bp. A total of 101918 SSRs were identified from the A. venetum genome data. There was great variation between the different types of nucleotide repeats, of which mononucleotides were the most abundant and hexanucleotides were the least abundant. Based on an analysis of various markers, we also deduced that the A. venetum genome is complex. In the future, the '2+3' (Illumina+PacBio) sequencing technique combination strategy should be employed to supplement the Hi-C technique and resequencing technique for whole-genome research in A. venetum and the analysis of gene differences between A. venetum from different sources. This can be used to identify genetic variation information, and genome assembly and annotation can be used to analyze key genes in A. venetum for fiber synthesis.
[fig] Figure 1: Distribution figure of GC contentThe left half of the dotted line in this figure is the read-1 GC content distribution, and the right half is the read-2 GC content distribution, different colors represent different base types, which is used to detect whether AT, GC separation is present. [/fig]
[fig] Figure 2: Distribution curve of K-merIt is an analysis of the genome size prediction of Apocynum, which determines the expected depth of K-mer from the position of the main peak. [/fig]
[fig] Figure 3: Distribution figure of contig coverage depth and lengthIn the figure, the peak with the most distribution is the main peak, the heterozygosity of the genome was judged according to the peak of 1/2 position before the main peak. [/fig]
[fig] Figure 4: Distribution figure of contig coverage depth and numberIn the figure, the peak with the most distribution is the main peak, the heterozygosity of the genome was judged according to the peak of 1/2 position before the main peak. [/fig]
[fig] Figure 5: Distribution figure of GC depth [/fig]
[table] Table 1: Data statisticsAbbreviations: Q20, percentage of bases with quality value ≥ 20; Q30, percentage of bases with quality value ≥ 30. [/table]
[table] Table 2: Statistics of the assembled genome sequences in A. venetum [/table]
[table] Table 3: Type and proportion of SSR [/table]
|
A Multi-Strain Probiotic Formulation Improves Intestinal Barrier Function by the Modulation of Tight and Adherent Junction Proteins
# Introduction
In recent years, scientific research in the fields of gut microbiota and probiotic consumption has risen exponentially, revealing the critical role of microbiota in supporting human health. A large number of commensal bacteria colonize the human gut epithelial layer, providing protection against pathogen invasion. An imbalance of gut flora, on the other hand, may contribute to the onset of intestinal dysbiosis, which is thought to exacerbate a variety of pathological conditions, including inflammatory bowel disease (IBD), diabetes, pancreatitis, non-alcoholic fatty liver disease, and neurological disorders [bib_ref] Dysfunction of the Intestinal Microbiome in Inflammatory Bowel Disease and Treatment, Morgan [/bib_ref] [bib_ref] Intestinal Homeostasis and Its Breakdown in Inflammatory Bowel Disease, Maloy [/bib_ref] [bib_ref] Gut Microbiota Dysbiosis Drives and Implies Novel Therapeutic Strategies for Diabetes Mellitus..., Li [/bib_ref] [bib_ref] Current Understanding of the Role of Gut Dysbiosis in Type 1 Diabetes, Abdellatif [/bib_ref] [bib_ref] Dysbiosis of Intestinal Microbiota and Decrease in Paneth Cell Antimicrobial Peptide Level..., Chen [/bib_ref] [bib_ref] Gut Microbiota Dysbiosis Worsens the Severity of Acute Pancreatitis in Patients and..., Zhu [/bib_ref] [bib_ref] Lactobacillus and Pediococcus Ameliorate Progression of Non-Alcoholic Fatty Liver Disease through Modulation..., Lee [/bib_ref] [bib_ref] Intestinal Permeability, Microbial Translocation, Changes in Duodenal and Fecal Microbiota, and Their..., Maccioni [/bib_ref] [bib_ref] Changes in the Gut Microbiota of Children with Autism Spectrum Disorder, Zou [/bib_ref]. Moreover, a suboptimal Western lifestyle is associated with dysbiosis and local endothelial dysfunction [bib_ref] Gut Dysbiosis and Western Diet in the Pathogenesis of Essential Arterial Hypertension:..., Canale [/bib_ref] [bib_ref] Suppression of Gut Dysbiosis Reverses Western Diet-Induced Vascular Dysfunction, Battson [/bib_ref].
Notably, evidence is now accumulating in support of the link between dysbiosis, gut barrier dysfunction, and altered intestinal permeability [bib_ref] Gut Mucosal Barrier Dysfunction, Microbial Dysbiosis, and Their Role in HIV-1 Disease..., Mudd [/bib_ref] [bib_ref] Gut Barrier Dysfunction and Inflammation in Dementia: A Pilot Study, Stadlbauer [/bib_ref]. The gastrointestinal tract serves as a barrier between the internal and external environments and selects nutritive substances from food. Although nutrient absorption is essential for maintaining metabolic homeostasis, a "leaky gut" could create an easy access point for pathogens and pro-inflammatory substances. The gut microbiota, on the other hand, can influence the intestinal barrier integrity, immune response, cardiometabolic functions, and the gut-brain axis.
The intestinal epithelial layer is the limiting hurdle for nutrient and drug permeation through the gastrointestinal tract. As a result, the integrity of the barrier is considered critical.
# Materials and methods
## Chemicals, reagents, and media
DMEM high glucose, fetal bovine serum (FBS), penicillin-streptomycin, trypsin-EDTA, L-glutamine, and Transwell inserts (0.64 mm diameter, 0.4 µm pore size, Costar) were purchased from Euroclone (Milan, Italy). Minimum essential medium non-essential amino acids (MEM-NEAA), SuperSignal TM West Dura Extended Duration Substrate, Pierce ® IP lysis buffer, Pierce TM Coomassie Plus Assay Reagent, LPS from E. coli 026:B6, TRIzol ® , High-Capacity Reverse Transcription Kit, and PowerUp TM SYBR TM Green Master Mix were obtained from Thermo Fisher Scientific (Rockford, IL, USA). Skim milk powder was purchased from Microgem (Pozzuoli, Italy). Acrylamide/bis-acrylamide solution 30% (29:1) was obtained from HiMedia Laboratories (Nashik, India). Primary monoclonal antibodies against CDH1 (4A2), CLDN-1 (D5H1D), CLDN-2 (E1H90), OCLN (E6B4R), and HRP-linked anti-mouse and anti-rabbit IgG were obtained from Cell Signaling Technology (Waltham, MA, USA). The primary antibody against β-Tubulin (E-AB-20095) was purchased from Elabscience (Huston, TX, USA).
## Cell cultures
The Caco-2 cell line was established by Dr. Jorgen Fogh (Sloan Kettering Memorial Cancer Center, Rye, NY, USA) in 1974 from a 72-year-old male patient who had previously been treated with 5-fluorouracil and Cytoxan [bib_ref] Differentiation of human colon cancer cells: A new approach to colon cancer, Zweibaum [/bib_ref]. Caco-2 cells were maintained at 37 - C in humidified atmosphere using DMEM high glucose w/sodium pyruvate, w/L-glutamine supplemented with 10% FBS, 1% MEM-NEAA, 10 U/mL penicillin, and 10 µg/mL streptomycin in 75 cm 2 flasks. Cells were sub-cultured at a 1:3 ratio when 80% of confluence was achieved. In order to obtain the phenotypic stabilization, Caco-2 cells were cultured for at least two weeks after thawing and sub-cultured at least four times before the experiments [bib_ref] Determination of Drug Permeability and Prediction of Drug Absorption in Caco-2 Monolayers, Hubatsch [/bib_ref]. Differentiated cells were used. In brief, Caco-2 cells were seeded at 2 × 10 5 cells/cm 2 density and cultured between 18 and 21 days to achieve full monolayer formation and cell polarization. In particular, Transwell inserts and T25 flasks were used for the TEER measurement and RNA/protein extraction, respectively. The culture medium was replaced six hours after seeding [bib_ref] Determination of Drug Permeability and Prediction of Drug Absorption in Caco-2 Monolayers, Hubatsch [/bib_ref] and then every two to three days.
## Treatment with the probiotic formulation
Caco-2 cells were challenged using four different cell/colony-forming unit (CFU) ratios (i.e., 10:1, 1:1, 1:10, and 1:100) for 24 h, in the absence of or in combination with an inflammatory stimulus (i.e., LPS at a 1 µg/mL concentration) added 4 h after probiotic exposure. Serobioma (Bromatech S.r.l., Milan, Italy) is composed of L. rhamnosus LR 32, B. lactis BL 04, and B. longum BB 536.
Prior to each treatment, the Serobioma was suspended in complete medium without antibiotics. Concentrations employed in the present work were selected based on previous research [bib_ref] Anti-Inflammatory Effect of Multistrain Probiotic Formulation (L. rhamnosus, B. lactis, and B...., Sichetti [/bib_ref].
## Mtt assay
Cell viability was determined by the conventional 3-(4,5-dimethylthiazol-2-yl)-2,5diphenyltetrazolium bromide (MTT) reduction assay, as described by Taticchi et al. [bib_ref] In Vitro Anti-Inflammatory Effects of Phenolic Compounds from Moraiolo Virgin Olive Oil..., Taticchi [/bib_ref]. In brief, Caco-2 cells (4 × 10 4 cells/well) were plated onto 96-well plates until sub-confluence was reached. After treatments, the medium was removed, and the cell monolayers were washed with PBS to eliminate probiotics in suspension. Fresh medium and 10 µL of a 5 mg/mL MTT solution was added to each well. Plates were incubated for 4 h at 37 - C. Finally, the formazan crystals were dissolved in 100 µL DMSO at 37 - C and the absorbance of each well was recorded at 595 nm using a microplate reader. Data were expressed as percentages (%) of reduced MTT, assuming the absorbance of control cells was 100%. Four independent experiments were performed.
## Teer measurements
Cells were treated as previously described. After treatment, cells were washed with sterile PBS and TEER values were registered using a Millicell-ERS Volt-Ohm meter (Milli-poreSigma, Burlington, MA, USA) according to the manufacturer's instructions. In brief, electrodes were equilibrated for two hours in complete medium. The washing step was carried out to remove probiotics adhered to the cell monolayer that could alter the TEER measurements. The measures were repeated three times for each well. The results are obtained from the three independent experiments.
## Rt-pcr
Cells were treated as described above, and total RNA was isolated using TRIzol Reagent according to the manufacturer's protocol. Then, 2 µg of RNA was reverse transcribed with a High-Capacity cDNA Reverse Transcription Kit in a 20 µL reaction mixture. The cDNA obtained was diluted at 1:5 and 2 µL of it was amplified and quantified by RT-PCR. The RT-PCR reaction was carried out in a 25 µL mixture containing 400 nM of forward and reverse primers (2 µL), 12.5 µL of ready-to-use PowerUP SYBR Green Master Mix, and RNase and DNase-free water (8.5 µL). As a reference gene, GAPDH was used. RT-PCR reactions for each sample and gene were run in triplicate. The PCR conditions were 50 - C for 2 min and 95 - C for 10 min, followed by 40 cycles at 95 - C for 15 s and 60 - C for 1 min. The sequences of the primer pairs were as indicated in [fig_ref] Table 1: Primer sequences used for RT-PCR [/fig_ref]. The mRNA relative expression levels were calculated as 2 −∆∆Ct .
# Western blot analysis
After the incubation period, cells were scraped and lysed with ice-cold Pierce IP lysis buffer. Lysates were centrifuged at 12,000× g and the supernatant containing the protein extract was collected. The protein amount in each lysate was determined by the Bradford assay. For the Western immunoblotting, 20 µg of total proteins were resolved on SDS-Page gel 8-15% and transferred onto a nitrocellulose membrane (0.2 µm pores). For the immunodetection, nitrocellulose membranes were exposed to the primary antibodies against CLDN-1 (1:500), CLDN-2 (1:500), OCLN (1:4000), and CDH1 (1:1000) at 4 - C overnight and, after, with HRP-conjugated secondary antibodies (1:10,000) (2 h, room temperature). Protein signals were detected using ImageQuant LAS 500 and the ECL kit. The β-tubulin was used as loading control.
# Statistical analysis
The statistical analysis was carried out using SPSS (SPSS Inc., Chicago, IL, USA). The normal distribution was assessed by the Shapiro-Wilk test and a comparison between the treated and untreated samples was then performed by one-way ANOVA followed by Dunnett's post hoc. Cells co-incubated with Serobioma and LPS were compared with the LPS-treated group. Student's t-test was also used to compare the untreated cells versus the LPS-treated sample at proportions of 10:1 versus 10:1 LPS, 1:1 versus 1:1 LPS, 1:10 versus 1:10 LPS, and 1:100 versus 1:100 LPS. Samples were considered statistically different when the p-value was <0.05.
# Results
## Cell viability
Potential cytotoxic effects of Serobioma in Caco-2 cells were determined by measuring the cell viability using the MTT assay after incubation with Serobioma in the presence or in absence of LPS. The results showed that the pre-treatment with Serobioma at all the cell/CFU ratios was not toxic to the Caco-2 cells after 24 h. Quite the opposite, a statistically significant increase in the cell viability was observed at the 1:1 and 1:10 ratios (p-value = 0.0004 and p-value = 0.0371, respectively). Of note, the pre-treatment of Caco-2 with Serobioma at the ratios of 1:10 and 1:100 was able to induce an increase in the cell viability (p-value = 0.0059 and p-value = 0.0394, respectively), even in the presence of LPS added 4 h after the probiotic administration [fig_ref] Figure 1: Figure 1 [/fig_ref]. MTT results expressed as the relative cell survival. Caco-2 cells were challenged using four different cell/colony-forming unit (CFU) ratios (i.e., 10:1, 1:1, 1:10, and 1:100) for 24 h, in the absence of or in combination with an inflammatory stimulus (i.e., LPS at a 1 µg/mL concentration) added 4 h after probiotic exposure. Statistical analysis: one-way ANOVA followed by Dunnett's post hoc. * p-value < 0.05 with respect to the untreated control. # p < 0.05 with respect to the LPS-treated sample.
## Teer measurement
Our results show that LPS (1 µg/mL) induced a marked reduction in the TEER values in the Caco-2 cell monolayer after 20 h of exposure (p-value = 0.001). Conversely, Serobioma was able to significantly increase the TEER values after the 10:1 cell/CFU ratio exposure (p-value = 0.009). Caco-2 cells pretreated with the multi-strain probiotic mixture at the 10:1, 1:1, and 1:10 cell/CFU ratios and then challenged with LPS showed a significantly higher TEER values if compared to the LPS treated cells (p-value = 0.002; 0.008; and 0.004, respectively) [fig_ref] Figure 2: TEER measurement results [/fig_ref]. Data referring to the 100:1 cell/CFU ratio treatment were not considered, as this concentration led to a rapid acidification of the cell media and affected the TEER measurement (data not shown).
## Effect of serobioma on tj gene expression in caco-2 cells
To investigate the molecular mechanism underlying the protection exerted by the probiotics, we focused on the expression of six genes encoding TJPs (i.e., ZO-1, ZO-2, CLDN1, CLDN2, OCLN, and CDH1). RT-PCR analysis showed that, compared with the untreated cells, the LPS stimulus induced a significant down-regulation of CLDN2 (p-value = 0.0030) and OCLN (p-value = 0.0021). For ZO-1 and ZO-2, a statistically insignificant decrease in the gene expression was reported. Conversely, an increase in CLDN1 mRNA levels was observed (p-value = 0.0003). The treatment with Serobioma induced a dose-dependent increase in the expression of ZO-1 (p-value = 0.0044, 0.0344, 0.0001, and <0.0001 for the 10:1, 1;1, 1:10, and 1:100 cell/CFU ratio treatments, respectively). However, the ZO-2 was up-regulated only by the 1:100 treatment (p-value < 0.0001). Meanwhile, compared with LPS alone, the 100:1 Serobioma pre-treatment increased the ZO-1 and ZO-2 mRNA levels (p-value = 0017 and p-value = 0061). As shown in , there was a significant up-regulation of OCLN at the 1:10 and 1:100 ratios (p-value < 0.0001 and p-value = 0.0075), and of CLDN1 at the 1:1 and 1:10 ratios (p-value = 0.0003 and p-value = 0.0040). The increase in the CLDN1 gene expression was observed even after the treatment with Serobioma at a 1:1 ratio followed by LPS exposure (p-value = 0.0002). Interestingly, a slight reduction in the gene expression of OCLN (p-value = 0.0027) after the 1:1 cell/CFU Serobioma treatment was also observed. The also shows that the pre-treatment with Serobioma at a 10:1 cell/CFU ratio prevented the LPS-induced changes in the CLDN1, CLDN2, and OCLN mRNA. The 1:10 cell/CFU pre-treatment was able to protect Caco-2 cells against the effect of LPS with respect to the CLDN1 mRNA levels. Finally, if compared with the untreated control, the 100:1 Serobioma pre-treatment followed by the LPS exposure increased the CDH1 gene expression (p-value = 0.0001). . mRNA levels for ZO-1, ZO-2, CLDN1, CLDN2, OCLN, and CDH1 genes in the Caco-2 cell monolayer. Caco-2 cells were challenged using four different cell/colony-forming unit (CFU) ratios (i.e., 10:1, 1:1, 1:10, and 1:100) for 24 h, in the absence of or in combination with an inflammatory stimulus (i.e., LPS at 1 µg/mL concentration) introduced 4 h after probiotic exposure. Results are expressed as the fold change (2 −∆∆Ct ) and summarized as the mean ± SEM of three independent experiments. Statistical analysis: one-way ANOVA followed by Dunnett's post hoc. * p-value < 0.05 with respect to the untreated control. # p-value < 0.05 with respect to the LPS-treated sample.
## Protein expression
The changes in the TJP amount induced by Serobioma were evaluated by the Western blotting analysis. The protein levels of OCLN, CADH1, CLDN1, and CLDN2 were detected in the Caco-2 cells. Results are summarized in . Moreover, uncut and unedited whole membranes are provided in the Supplementary [fig_ref] Figure 1: Figure 1 [/fig_ref]. In agreement with the RT-PCR findings, the administration of the probiotic formulation at a 1:1 ratio enhanced the CLDN1 protein levels (p-value = 0.022 and 0.015 in presence of LPS), while the CLDN2 protein expression decreased following 1:10 LPS (p-value = 0.035), 1:100 (p-value = 0.012) and 1:100 (p-value = 0.039) treatments. Quite the opposite, LPS decreased the CLDN1 protein levels and increased the CLDN2 levels, in contrast to what was detected using RT-PCR. The CDH1 levels where enhanced after the 1:1 and 1:10 treatments (p-value = 0.009 and 0.005), even in presence of LPS (p-value = 0.007 and 0.012). Moreover, an increase in the OCLN (p-value = 0.006) was observed after Serobioma exposure at the 1:10 ratio. . Western blot membranes and results for CADH1, OCLN, CLDN1, and CLDN2 proteins in Caco-2 cells challenged using four different cell/colony-forming unit (CFU) ratios (i.e., 10:1, 1:1, 1:10, and 1:100) for 24 h, in the absence of or in combination with an inflammatory stimulus (i.e., LPS at 1 µg/mL concentration) introduced 4 h after probiotic exposure. The western blot signals were normalized using β-tubulin as the loading control. Results are expressed as mean ± SEM of the protein variation relative to the untreated sample. Statistical analysis: one-way ANOVA followed by Dunnett's post hoc. * p-value < 0.05 with respect to the untreated control. # p-value < 0.05 with respect to the LPS-treated sample.
# Discussion
The aim of this study was to analyze the effect of the commercial formulation Serobioma on intestinal permeability in an in vitro model.
The Caco-2 cell monolayer was exposed to Serobioma for 24 h, in the absence of or in combination with the LPS added 4 h after the probiotic exposure. The 4 h time point for the pretreatment with the probiotic formulation was used, as this is the time required for an efficient metabolism that can impact the bioactivity of probiotics [bib_ref] Anti-Inflammatory Effect of Multistrain Probiotic Formulation (L. rhamnosus, B. lactis, and B...., Sichetti [/bib_ref] [bib_ref] Inhibitory Effect of Probiotic Escherichia Coli Strain Nissle 1917 on Adhesion to..., Boudeau [/bib_ref].
In the present study, the epithelial cell monolayer was challenged with the LPS. The LPS inflammatory damage was confirmed by the reduction of the TEER values in LPStreated cells compared to the untreated control. Notably, Serobioma was able to prevent the loss of the monolayer integrity at all the concentrations tested [fig_ref] Figure 2: TEER measurement results [/fig_ref]. The mechanisms underlying this observation could be multiple and simultaneous, such as the production of bioactive metabolites (e.g., butyrate), the bacterial adhesion that prevents LPS binding, and the modulation of TJPs. For example, it was reported that exopolysaccharide (EPS) produced by lactic acid bacteria (such as Lactobacilli and Bifidobacteria) could interact with intraluminal water to produce a protective film [bib_ref] Gut Health Benefit and Application of Postbiotics in Animal Production, Zhong [/bib_ref].
Among the strains included in Serobioma formulation, B. longum BL536 has been recognized as one of the most effective probiotic strains. It acts mainly through microbiota modulation [bib_ref] Beneficial Effects of Bifidobacterium Longum Subsp. Longum BB536 on Human Health: Modulation..., Wong [/bib_ref] , and it is able to modulate the immune response [bib_ref] Clinical Effects of Probiotic Bifidobacterium Longum BB536 on Immune Function and Intestinal..., Akatsu [/bib_ref]. Moreover, this strain is capable of stabilizing TJPs through EPS and the production of butyrate as active metabolites [bib_ref] Gut Health Benefit and Application of Postbiotics in Animal Production, Zhong [/bib_ref]. L. rhamnosus LR32 exhibited immunomodulatory effects [bib_ref] A Key Role of Dendritic Cells in Probiotic Functionality, Foligne [/bib_ref] [bib_ref] Manipulation of Epithelial Integrity and Mucosal Immunity by Host and Microbiota-Derived Metabolites, Kayama [/bib_ref]. Finally, B. lactis BL04 was shown to be useful against dyslipidemia in children.
Due to the close relationship between barrier integrity and TJPs, the authors focused their attention on the expression of these proteins.
The administration of the probiotic formulation containing selected species of both Lactobacillus and Bifidobacterium genera promoted the expressions of ZO-1, CLDN1, and OCLN. These proteins are the most frequently altered during infections and diseases. Indeed, pathogenic microorganisms, such as enteropathogenic and enterohemorrhagic E. coli, can alter the barrier properties by affecting OCLN and ZO-1.
A previous study showed that LPS increases the intestinal permeability by affecting OCLN, CLDN1, and ZO-1 and by inducing mitochondrial dysfunction and mitophagy in piglets [bib_ref] LPS Challenge Increased Intestinal Permeability, Disrupted Mitochondrial Function and Triggered Mitophagy of..., Cao [/bib_ref]. In addition, experimental and clinical studies revealed that the downregulation of ZO-1, OCLN, and CLDN1 was associated with increased intestinal permeability in IBD patients [bib_ref] Increase in the Tight Junction Protein Claudin-1 in Intestinal Inflammation, Poritz [/bib_ref].
In our study, Serobioma treatment increased CLDN1, while it decreased CLDN2 at both the mRNA and protein levels, suggesting that the probiotic complex proves beneficial and has protective effects, against the LPS inflammatory response. Indeed, CLDN1 is known to form continuous sealing filaments. Conversely, CLDN2 is involved in pore architecture and regulates the paracellular permeability of small ions and molecules. In particular, junctional complexes, in which CLDN2 levels are increased, are characterized by the presence of discontinuous filaments [bib_ref] Manner of Interaction of Heterogeneous Claudin Species within and between Tight Junction..., Furuse [/bib_ref]. Likely, in Chron's disease, the increase in CLDN2 accompanied by the decrease in CLDN1 constitutes the molecular basis of discontinuous filaments, which can lead to the conversion of TJs into leaky junctions [bib_ref] Changes in Expression and Distribution of Claudin 2, 5 and 8 Lead..., Zeissig [/bib_ref]. Quite the opposite, LPS decreased the CLDN1 protein levels and increased the CLDN2 levels, in contrast to what was detected using RT-PCR. This apparent discrepancy implies that we observed a reparative response to LPS-associated damage. We also observed a dose-dependent increase in the ZO-1 expression in the Serobioma-treated Caco-2 cells, suggesting that probiotics play a significant role in improving the integrity of the intestinal epithelial barrier [bib_ref] Prebiotics and Epithelial Tight Junctions: A Promising Approach to Modulate Intestinal Barrier..., Rose [/bib_ref].
In the present study, it emerged that CDH1 protein levels were enhanced by Serobioma. However, CDH1 mRNA levels did not change, suggesting that Serobioma has a timedependent effect on gene expression or that it activates intracellular pathways that regulate CDH1 turnover. Various studies have reported that CDH1 coordinates the establishment of the apical-basolateral polarity with the formation of adhesion junctions, desmosomes, and TJs [bib_ref] Adaptation of Core Mechanisms to Generate Cell Polarity, Nelson [/bib_ref] [bib_ref] Tissue Organization by Cadherin Adhesion Molecules: Dynamic Molecular and Cellular Mechanisms of..., Niessen [/bib_ref]. CDH1 plays a crucial role in the regulation of actomyosin-dependent tension, TJ positioning, and barrier formation [bib_ref] E-Cadherin Integrates Mechanotransduction and EGFR Signaling to Control Junctional Tissue Polarization and..., Rübsam [/bib_ref]. In addition, the CDH1-dependent cell adhesion system supports enterocyte differentiation and TJ establishment [bib_ref] Integrin-Mediated Functional Polarization of Caco-2 Cells through E-Cadherin-Actin Complexes, Schreider [/bib_ref] [bib_ref] Depletion of E-Cadherin Disrupts Establishment but Not Maintenance of Cell Junctions in..., Capaldo [/bib_ref]. In light of the preceding observations, it is hypothesized that Serobioma also acts during the cell polarization process and aids in the formation of a more robust barrier.
The results presented in this study indicate how the probiotic formulation was able to modulate the expression of TJPs across the epithelial barrier. Overall, the most significant data were recorded when the probiotic bacteria were administered in the rations of 10:1, 1:1, and 1:10 (cells/CFU).
A previous in vitro screening confirmed the anti-inflammatory properties of the Serobioma formulation [bib_ref] Anti-Inflammatory Effect of Multistrain Probiotic Formulation (L. rhamnosus, B. lactis, and B...., Sichetti [/bib_ref]. Taken together, our results suggest that Serobioma exerts a protective effect against intestinal epithelial barrier dysfunction. In particular, it has a preventive activity against inflammation-associated damage.
We suggest that the molecular mechanisms are attributable to the modulation of proand anti-inflammatory cytokine release, the modulation of TJ gene expression, and the activation of inflammatory intracellular pathways.
Our results are corroborated by numerous studies, in which the effectiveness of Lactobacilli and Bifidobacteria in preserving the intestinal barrier function was evaluated. Beneficial properties of Bifidobacteria have been reported in in vivo models of inflammation. For example, feeding with B. bifidum promoted the ZO-1 expression in a dextran sodium sulphate colitis mouse model [bib_ref] Inhibitory Effect of Bifidobacterium Bifidum ATCC 29521 on Colitis and Its Mechanism, Din [/bib_ref]. B. infantis has been shown to possess barrier-preserving properties [bib_ref] Bifidobacteria Stabilize Claudins at Tight Junctions and Prevent Intestinal Barrier Dysfunction in..., Bergmann [/bib_ref]. B. longum increased the TEER values and decreased the paracellular permeability of Caco-2 cells stimulated by LPS [bib_ref] Longum K5 Alleviates Inflammatory Response and Prevents Intestinal Barrier Injury Induced by..., Zhao [/bib_ref]. B. longum BB 536 was able to induce remission in patients with ulcerative colitis [bib_ref] Efficacy of Probiotic Treatment with Bifidobacterium Longum 536 for Induction of Remission..., Tamaki [/bib_ref]. It has been reported that B. lactis 420 and B. lactis HN019 increased TJ integrity in Caco-2 cells [bib_ref] Effect of Four Probiotic Strains and Escherichia Coli O157:H7 on Tight Junction..., Putaala [/bib_ref]. On the other hand, certain Lactobacillus species prevented barrier disruption through the upregulation of TJPs. In addition, L. acidophilus and L. plantarum increased OCLN expression within in vivo and in vitro models, respectively [bib_ref] Lactobacillus Plantarum MB452 Enhances the Function of the Intestinal Barrier by Increasing..., Anderson [/bib_ref] [bib_ref] Effect of Lactobacillus on the Gut Microflora and Barrier Function of the..., Qin [/bib_ref]. L. rhamnosus CNCM I-3690 partially restored the function of the intestinal barrier and increased the levels of OCLN and CADH1 [bib_ref] Lactobacillus rhamnosus CNCM I-3690 and the Commensal Bacterium Faecalibacterium prausnitzii A2-165 Exhibit..., Laval [/bib_ref]. The probiotics E. coli Nissle 1917, L. rhamnosus GG (LGG), and a mixture of Lactobacilli and Bifidobacteria prevented the increase in intestinal permeability in vivo [bib_ref] Probiotic Escherichia Coli Nissle 1917 Inhibits Leaky Gut by Enhancing Mucosal Integrity, Ukena [/bib_ref] [bib_ref] Probiotic Mixture VSL#3 Protects the Epithelial Barrier by Maintaining Tight Junction Protein..., Mennigen [/bib_ref] [bib_ref] Lactobacillus rhamnosus GG Attenuates Interferon-γ and Tumour Necrosis Factor-Alpha-Induced Barrier Dysfunction and..., Donato [/bib_ref].
The predictability of the in vitro experimental model is a topic of crucial importance. The in vitro model offers the advantage of evaluating the probiotic effects as such, while minimizing the contribution of environmental factors. However, the environment of the gut is much more complex. The pH of the intestinal lumen may be a determining factor. Indeed, studies reported that acidification elicits the instability of the epithelial TJ complexes in both in vivo and in vitro models [bib_ref] Acute Mechanism of Medium Chain Fatty Acid-Induced Enhancement of Airway Epithelial Permeability, Coyne [/bib_ref] [bib_ref] Dissociation and Dispersion of Claudin-3 from the Tight Junction Could Be One..., Oguro [/bib_ref] [bib_ref] The Degradation of Airway Tight Junction Protein under Acidic Conditions Is Probably..., Xu [/bib_ref]. In our experimental model, the 1:100 cell/CFU ratio could have caused the acidification of the culture medium due to the probiotic metabolism, as ascertained by the medium color shift from red to yellow, which may have had a negative impact on the TEER measurements. Strains of Lactobacillus are known to acidify the environment and make it hostile to pathogen proliferation. However, in humans, the mucus layer protects the gut epithelium against excessive acidification. Although it is still difficult to determine whether the in vitro model is predictive of what occurs in vivo, the results of this study provide a rationale for future in vivo and clinical studies.
Finally, there are various forms of commercial probiotic preparations, and their efficacies could differ considerably depending on whether they contain a single strain or multiple strains. Studies have already indicated that the synergistic effects of multi-strain probiotic formulations differ from those of single-strain formulations [bib_ref] Genome-Wide Immune Modulation of TLR3-Mediated Inflammation in Intestinal Epithelial Cells Differs between..., Macpherson [/bib_ref]. In vitro studies have shown that some multi-strain probiotics may show significant better inhibitory effects against enteropathogens and greater benefits when compared to single-strain preparations. Some multi-strain probiotics may reduce the absorption of harmful chemicals due to their ability to absorb heavy metals within their cell walls [bib_ref] A Prophylactic Multi-Strain Probiotic Treatment to Reduce the Absorption of Toxic Elements:..., Astolfi [/bib_ref]. Multi-strain probiotics may be able to create a probiotic niche that improves bacterial colonization overall. In particular, in vivo, strains with an optimal pH range of 6-7, typical of the upper intestinal tract, show rapid growth, leading to a decrease in the optimal pH of the most acidophilic bacterial strains [bib_ref] Probiotics: Determinants of Survival and Growth in the Gut, Bezkorovainy [/bib_ref]. Overall, multi-strain probiotics are more consistent in their actions than single-strain probiotics [bib_ref] Multistrain and Multispecies Probiotics-A Comparison of Functionality and Efficacy, Timmerman [/bib_ref].
For all these reasons, multi-strain probiotic supplementation with Serobioma could be effective method to protect the intestinal barrier function against inflammatory attacks.
# Conclusions
In this study, it was found that the Serobioma formulation, containing a combination of Lactobacilli and Bifidobacteria strains, is capable of preserving the integrity and functioning of the intestinal barrier from damage caused by the LPS inflammatory stimulus, as well as modulating the expression of genes and proteins belonging to TJs in Caco-2 cell monolayers.
The present study provides additional insight into the mechanisms by which selected probiotic strains prevent intestinal epithelial barrier dysfunction and contribute to sustaining gut health. Our data suggest a broad spectrum of positive effects exerted by the Serobioma formulation, demonstrating a high potential for its therapeutic use in modern medicine. Future studies should aim at analyzing and characterizing other molecular mechanisms and intracellular pathways that mediate probiotic efficacy.
[fig] Figure 1: Figure 1. MTT results expressed as the relative cell survival. Caco-2 cells were challenged using four different cell/colony-forming unit (CFU) ratios (i.e., 10:1, 1:1, 1:10, and 1:100) for 24 h, in the absence of or in combination with an inflammatory stimulus (i.e., LPS at a 1 µg/mL concentration) added 4 h after probiotic exposure. Statistical analysis: one-way ANOVA followed by Dunnett's post hoc. * p-value < 0.05 with respect to the untreated control. # p < 0.05 with respect to the LPS-treated sample. [/fig]
[fig] Figure 2: TEER measurement results. Caco-2 cells were challenged using three different cell/colonyforming units (CFU) ratios (i.e., 10:1, 1:1, and 1:10) for 24 h, in the absence of or in combination with an inflammatory stimulus (i.e., LPS at a 1 µg/mL concentration) introduced 4 h after probiotic exposure. Statistical analysis: one-way ANOVA followed by Dunnett's post hoc. Student's t-test was used to compare control cells against the LPS-treated sample. * p-value < 0.05. [/fig]
[fig] Supplementary: Materials: The following supporting information can be downloaded at: https://www. mdpi.com/article/10.3390/cells11162617/s1, Figure S1: uncut and unedited western blot membranes. Author Contributions: Conceptualization, G.T.; methodology, G.T., C.C. and R.d.V.; validation, R.d.V. and C.C.; formal analysis, R.d.V.; investigation, R.d.V. and C.C.; resources, G.T.; data curation, R.d.V. and C.C.; writing-original draft preparation, G.T., R.d.V. and C.C.; writing-review and editing, G.T., C.C. and R.d.V.; visualization, R.d.V.; supervision, G.T.; project administration, G.T.; funding acquisition, G.T. All authors have read and agreed to the published version of the manuscript. [/fig]
[fig] Funding: This research was funded by Fondo Ricerca di Base di Ateneo, University of Perugia, Perugia, Italy, grant numbers 6RICBASE20; 6RICBASE21. Institutional Review Board Statement: Not applicable. Informed Consent Statement: Not applicable. Data Availability Statement: All data are included in the manuscript. [/fig]
[table] Table 1: Primer sequences used for RT-PCR. [/table]
|
Population Pharmacokinetic Analysis of Fluticasone Furoate/Umeclidinium/Vilanterol via a Single Inhaler in Patients with COPD
A population pharmacokinetic analysis was conducted from a subset of samples obtained from the Lung Function and Quality of Life Assessment in Chronic Obstructive Pulmonary Disease with Closed Triple Therapy trial to characterize the pharmacokinetics of fluticasone furoate, umeclidinium, and vilanterol in patients with symptomatic COPD following treatment with fluticason furoate-umeclidinium-vilanterol combined in a single inhaler. This was a randomized, double-blind, double-dummy study comparing 24 weeks of once-daily triple therapy (fluticason furoate-umeclidinium-vilanterol, 100 μg/62.5 μg/25 μg; Ellipta inhaler) with twice-daily dual therapy (budesonide/formoterol 400 μg/12 μg; Turbuhaler). The analyses were conducted in a subset of 74 patients who received fluticason furoate-umeclidinium-vilanterol and provided serial or sparse samples. Monte Carlo simulations and a model-based estimation approach both indicated that systemic drug concentrations of fluticasone furoate, umeclidinium, and vilanterol after administration of fluticason furoate-umeclidinium-vilanterol triple combination therapy from a single inhaler were within the ranges observed following administration of these drugs as monotherapy (fluticasone furoate, umeclidinium, and vilanterol) or as dual-combination therapy (fluticasone furoate/vilanterol or umeclidinium/vilanterol).
Treatment of chronic obstructive pulmonary disease (COPD) commonly includes triple therapy with one or more long-acting bronchodilators (long-acting muscarinic receptor antagonists [LAMA]), long-acting beta 2 -adrenergic receptor agonists and inhaled corticosteroids (ICS) for patients with more advanced disease who have significant symptoms and a high risk for exacerbations.Studies have shown that the use of ICS/LAMA/LABA delivered via multiple inhalers in patients with moderate to severe COPD demonstrates greater improvements in lung function and health-related quality of life compared with ICS/LABA or LAMA therapy alone. [bib_ref] A randomised trial of fluticasone furoate/vilanterol (50/25 μg; 100/25 μg) on lung..., Kerwin [/bib_ref] [bib_ref] Fluticasone furoate/vilanterol (100/25; 200/25 μg) improves lung function in COPD: a randomised..., Martinez [/bib_ref] Recently, a once-daily single-inhaler triple therapy of fluticasone furoate, umeclidinium, and vilanterol (fluticason furoate-umeclidinium-vilanterol) 100 μg/62.5 μg/25 μg was approved in the United States for the treatment of COPD. The Lung Function and Quality of Life Assessment in COPD with Closed Triple Therapy (FULFIL) trial was the first study to compare once-daily single-inhaler triple therapy (ICS/LAMA/LABA) with twice-daily dual therapy (ICS/LABA) in patients with symptomatic COPD who were at risk for exacerbations. [bib_ref] FULFIL trial: once-daily triple therapy for patients with chronic obstructive pulmonary disease, Lipson [/bib_ref] The FULFIL study demonstrated statistically significant and clinically meaningful improvements at 24 weeks of therapy in lung function and health-related quality of life with once-daily single-inhaler triple therapy delivered using the Ellipta inhaler (GlaxoSmithKline, Research Triangle Park, North Carolina), compared with twice-daily budesonide/formoterol (400/12 μg) delivered using the Turbuhaler (AstraZeneca, Wilmington, Delaware). The study also demonstrated a significant reduction in the annual rate of moderate/ severe exacerbations with fluticason furoateumeclidinium-vilanterol compared with budesonide/ formoterol. The safety profile of the fluticason furoateumeclidinium-vilanterol single-inhaler triple therapy was similar to that of the individual components, with no new safety findings.
The pharmacokinetics (PK) of fluticasone furoate, umeclidinium, and vilanterol have been described previously in monotherapy or dual combination therapy (fluticasone furoate/vilanterol or umeclidinium/vilanterol) studies. [bib_ref] Population pharmacokinetics of inhaled umeclidinium and vilanterol in patients with chronic obstructive..., Goyal [/bib_ref] [bib_ref] Population pharmacokinetics of inhaled fluticasone furoate and vilanterol in subjects with chronic..., Siederer [/bib_ref] In healthy volunteer studies, there was no evidence of a pharmacokinetic or pharmacodynamic interaction when the 3 molecules were combined in 1 device compared with dual therapies. [bib_ref] Pharmacokinetics of fluticasone furoate, umeclidinium, and vilanterol as a triple therapy in..., Brealey [/bib_ref] The current report provides the first analysis of the population PK of fluticasone furoate, umeclidinium, and vilanterol when administered as a fixed-dose combination from a single inhaler. The study was carried out according to the Declaration of Helsinki, good clinical practice guidelines of the International Conference on Harmonization, and other applicable regulatory requirements. The study protocol was approved by institutional review boards for human studies associated with the clinical sites and written consent was obtained from each patient or their surrogates prior to study participation. A complete list of all study sites and details of their respective institutional review boards is provided as Supplemental Data.
# Methods
The complete design of the FULFIL study was previously reported. [bib_ref] FULFIL trial: once-daily triple therapy for patients with chronic obstructive pulmonary disease, Lipson [/bib_ref] In brief, eligible patients were aged ࣙ40 years with COPD (forced expiratory volume in 1 second [FEV 1 ] <50% and COPD Assessment Test [CAT TM ; GlaxoSmithKline, Research Triangle Park, North Carolina] score ࣙ10, or FEV 1 ࣙ50-<80% and CAT score ࣙ10, and either ࣙ2 moderate exacerbations in the past year or ࣙ1 severe exacerbation in the past year). Patients had to have received daily maintenance COPD therapy for ࣙ3 months prior to screening. Patients were randomly assigned 1:1 to receive 24 weeks of treatment with once-daily fluticason furoate-umeclidinium-vilanterol or twice-daily budesonide/formoterol; a subset of patients received blinded study treatment for up to 52 weeks.
## Pharmacokinetic assessments
Blood samples were collected from a preplanned subset of approximately 130 subjects for the sparse sampling scheme and approximately 20 subjects for the serial sampling scheme in order to achieve an adequate number of samples from subjects randomly assigned to the fluticason furoate-umeclidinium-vilanterol arm. Blood samples at weeks 12 and 24 were analyzed from 74 patients who had been randomly assigned to the fluticason furoate-umeclidinium-vilanterol arm (fluticason furoate-umeclidinium-vilanterol PK population). The sampling scheme included 2 mutually exclusive subsets of subjects (sparse and serial subsets) that provided PK blood samples. For the sparse sampling, two 6-mL samples (n = 64) were collected at week 12 (prior to dosing and 5-15 minutes after dosing) and week 24 (5-15 minutes and 45-90 minutes after dosing). For the serial sampling, seven 6-mL samples were collected at week 24 (prior to dosing and 5-15 minutes, 45-90 minutes, 2.5-4 hours, 6-8 hours, 10-12 hours, and 23-24 hours after dosing).
Plasma samples were analyzed for fluticasone furoate, umeclidinium, and vilanterol concentrations using validated bioanalytical methods based on solid phase extraction followed by high-pressure liquid chromatography with tandem mass spectrometry for analyte detection. The lower limit of quantification for fluticasone furoate, umeclidinium, and vilanterol was 10 pg/mL; the higher limit of quantification was 1000 pg/mL for fluticasone furoate and vilanterol and 2000 pg/mL for umeclidinium.
# Population pharmacokinetic analysis
Population PK analyses used nonlinear mixed effects modeling with NONMEM program version 7.1.2 (GloboMax, Hanover, Maryland) for modeling and simulations.
Because FULFIL (CTT116853) data are within the prediction limits of the existing models based on extensive clinical pharmacology data (including population PK) from the mono-(umeclidinium) and dual-therapy programs (fluticasone furoate/vilanterol, umeclidinium/vilanterol), a covariate analysis was not planned. The same covariate relationship was assumed and consistent parameter estimates between the current and previous model supported that assumption [fig_ref] Table 1: Final fluticasone furoate Pharmacokinetic Model [/fig_ref]. Data were analyzed using 2 different approaches in this study. In the first approach, Monte Carlo simulations of previously reported population PK models for fluticasone furoate, umeclidinium, and vilanterol in patients with COPD who received fluticasone furoate, umeclidinium, and vilanterol as monotherapy or as dual combination therapy (fluticasone furoate/vilanterol or umeclidinium/vilanterol) were undertaken. [bib_ref] Population pharmacokinetics of inhaled umeclidinium and vilanterol in patients with chronic obstructive..., Goyal [/bib_ref] [bib_ref] Population pharmacokinetics of inhaled fluticasone furoate and vilanterol in subjects with chronic..., Siederer [/bib_ref] Specifically, the final PK model for fluticasone furoate was a 2-compartment model with first-order absorption and first-order elimination rate, with covariate effect of "race" on apparent clearance (CL/F, inhaled). A 3-compartment linear model with zero-order absorption and first-order elimination was The observed time versus plasma concentration data for each analyte from the current study, FULFIL, were overlaid on the 90% prediction intervals from the respective simulations to gauge adequacy of these existing models to describe the observed data. The maximum a posteriori (MAP) Bayesian estimates of the PK parameters were obtained for all 74 patients, which were used to compute steady-state maximum plasma drug concentration (C max ) and area under the plasma drug concentration-time curve (AUC) over the dosing interval.
In the second approach, a model-based estimation was applied. The plasma data for fluticasone furoate, umeclidinium, and vilanterol from FULFIL were combined with the respective historical data used to develop population PK models for fluticasone furoate, umeclidinium, and vilanterol and were analysed using previously reported population PK models for fluticasone furoate, umeclidinium, and vilanterol. [bib_ref] Population pharmacokinetics of inhaled umeclidinium and vilanterol in patients with chronic obstructive..., Goyal [/bib_ref] [bib_ref] Population pharmacokinetics of inhaled fluticasone furoate and vilanterol in subjects with chronic..., Siederer [/bib_ref] The data below the quantifiable limit (BQL) were treated as censored data and analyzed using a joint probability model that maximizes the likelihood of data above quantification limit and treats BQL data as censored with the full likelihood approach. [bib_ref] Likelihood based approaches to handling data below the quantification limit using NONMEM..., Ahn [/bib_ref] The MAP Bayes PK parameter estimates were compared following analysis of the combined dataset (historical data set, including the FULFIL data) versus the original historical data (without the FULFIL data).
# Results
The demographics and baseline characteristics of patients included in the population PK analysis were similar to the intent-to-treat population of the FULFIL study [fig_ref] Table 3: Summary of Patient Demographics [/fig_ref]. As shown in [fig_ref] Figure 1: Observed concentration-time data from FULFIL and historical datasets [/fig_ref] , the observed plasma concentration-time data for fluticasone furoate, umeclidinium, and vilanterol in FULFIL were all within the range observed for historical fluticasone furoate, umeclidinium, and vilanterol data, respectively. In addition, the observed plasma concentration-time data for all 3 analytes (fluticasone furoate, umeclidinium, and vilanterol) in FULFIL were generally contained within the 90% prediction intervals, as indicated by the 5% and 95% percentiles obtained from the Monte Carlo simulations of the population PK models for fluticasone furoate, umeclidinium, and vilanterol from previous models, with fluticasone furoate/vilanterol and umeclidinium/vilanterol suggesting the adequacy of existing population PK models [fig_ref] Figure 2: Comparison of observed versus predicted concentration-time data profiles [/fig_ref]. The model also considered censored data and adequately characterized the proportion of BQL data for each drug in this population [fig_ref] Figure 3: Proportion of values data below the quantifiable limit [/fig_ref]. This was further confirmed using the model-based estimation approach in which there were no marked changes in the PK parameter estimates of fluticasone furoate, umeclidinium, and vilanterol using the combined dataset (FULFIL and historical data) versus the historical data (without data from FULFIL) [fig_ref] Figure 1: Observed concentration-time data from FULFIL and historical datasets [/fig_ref].
Based on the individual estimated MAP Bayes PK parameter estimates for fluticasone furoate, umeclidinium, and vilanterol administered as the triple fixeddose combination (fluticason furoate-umeclidiniumvilanterol), steady-state systemic exposures were computed and are summarized in [fig_ref] Table 4: Summary of Steady-State Systemic C max and AUC Results for fluticasone furoate,... [/fig_ref]. The steady-state AUC over dosing interval for fluticasone furoate, ume-clidinium, and vilanterol estimated from the population PK analysis in the FULFIL study (with fluticason furoate-umeclidinium-vilanterol) was consistent with historical data in patients with COPD given mono-or dual-combination therapies. The C max at steady state for fluticasone furoate following fluticason furoateumeclidinium-vilanterol was also consistent with historical data. Corresponding C max values for umeclidinium and vilanterol were generally lower with fluticason furoate-umeclidinium-vilanterol (likely due to inadequate characterization of time to peak plasma concentration in a small number of subjects) compared with historical data from dual combinations. The lower C max The steady-state estimates were calculated using individual maximum a posteriori (MAP) Bayesian estimates.
for umeclidinium and vilanterol was not expected to be clinically relevant with respect to safety or efficacy following administration of the single-inhaler triple fluticason furoate-umeclidinium-vilanterol combination, as observed in the overall analysis of the FULFIL study. 3
# Discussion
FULFIL is the first study providing data on the PK of fluticasone furoate, umeclidinium, and vilanterol in patients with COPD when administered as fluticason furoate-umeclidinium-vilanterol in combination using a single inhaler. This report describes the population pharmacokinetic PK analysis of data from the subset of patients who participated in the FULFIL study.
The population PK design following fluticason furoate-umeclidinium-vilanterol dosing in this study focused on the initial characterization of the systemic exposure (rate and extent) of the individual analytes fluticasone furoate, umeclidinium, and vilanterol in a cohort of patients with symptomatic COPD. As mentioned above, a descriptive population PK analysis was undertaken in the current study, without any covariate analysis, by leveraging the availability of extensive clinical pharmacology data (including population PK models) for fluticasone furoate, umeclidinium, and vilanterol from previous mono-and dual-therapy programs.
In the FULFIL study, consistency of the predicted PK parameter estimates using the modelbased simulation and estimation approaches suggested no major differences in the patient population characteristics in the FULFIL study in relation to the historical COPD population studied with fluticasone furoate, umeclidinium, and vilanterol as mono-and dual therapies. In addition, simulationbased diagnostics demonstrated that the population PK models adequately predicted the fluticasone furoate, umeclidinium, and vilanterol PK data from the FUL-FIL study. The model also considered censored data and adequately characterized the proportion of BQL data for each drug in this population. The modelbased estimates of C max and AUC values at steadystate from FULFIL were within the range observed in historical studies involving fluticasone furoate, umeclidinium, and vilanterol administered in subjects with COPD. Previous data from healthy volunteers also confirmed the absence of a relevant pharmacokinetic/pharmacodynamic interaction when these 3 molecules were combined. [bib_ref] Pharmacokinetics of fluticasone furoate, umeclidinium, and vilanterol as a triple therapy in..., Brealey [/bib_ref] The extensive clinical pharmacology programs for mono-and dual-therapy combinations did not demonstrate any PK interactions between fluticasone furoate, umeclidinium, and vilanterol. The population PK analysis in the present study provided reasonable estimates of rate and extent of systemic exposure to fluticasone furoate, umeclidinium, and vilanterol in individuals with symptomatic COPD. The FULFIL population PK analyses showed that systemic drug levels of fluticasone furoate, umeclidinium, and vilanterol following fluticason furoate-umeclidinium-vilanterol administration in one inhaler (single-inhaler triple combination) were within the range observed with administration through individual single inhalers (fluticasone furoate, umeclidinium, and vilanterol).
[fig] Figure 1: Observed concentration-time data from FULFIL and historical datasets. FF: Fluticasone furoate; UMEC: umeclidinium; VI: vilanterol. [/fig]
[fig] Figure 2: Comparison of observed versus predicted concentration-time data profiles. FF: Fluticasone furoate; UMEC: umeclidinium; VI: vilanterol. [/fig]
[fig] Figure 3: Proportion of values data below the quantifiable limit (BQL). Solid lines = observed intervals; dashed lines = prediction intervals. FF: Fluticasone furoate; UMEC: umeclidinium; VI: vilanterol. [/fig]
[table] Table 1: Final fluticasone furoate Pharmacokinetic Model: PK Parameter Estimates [/table]
[table] Table 2: Final [/table]
[table] Table 3: Summary of Patient Demographics [/table]
[table] Table 4: Summary of Steady-State Systemic C max and AUC Results for fluticasone furoate, umeclidinium, and vilanterol from FULFIL and Historic Data [/table]
|
Is Corticospinal Tract Degeneration Caused by Sjögren Syndrome?
A 36-year-old woman with a history of dry eyes and recurrent oral ulcers presented with an 8-month history of progressive motor weakness of both legs. She was documented as experiencing mild weakness (4+/5) of both legs with no sensory loss 2 months after onset. Cervical MRI revealed a T2 hyperintense lesion from C2 to C7 without cord swelling or contrast enhancement(Fig. 1A). The rest of the spinal cord and the brain MRI were normal. A cerebrospinal fluid (CSF) examination was unrevealing. She was treated with intravenous and oral corticosteroids based on the presumptive diagnosis of transverse myelitis (TM).Her motor weakness continued to worsen. On admission examination, she showed spastic paraplegia with 3/5 strength with no sensory or bladder/bowel signs or symptoms. Her deep tendon reflexes were all brisk, and the Babinski sign was positive bilaterally. Follow-up cervical MRI revealed focal areas of atrophy along the bilateral lateral columns, with sparing of the posterior columns(Fig. 1B). Workup to exclude metabolic or infectious causes of myelitis were unremarkable. Electromyography findings excluded amyotrophic lateral sclerosis, and no pathologic variants related to hereditary spastic paraplegia were found on wholeexome sequencing. Serologic studies revealed positive anti-Ro/SSA (102.3 U/mL), IgG β2glycoprotein 1 antibody, and anticardiolipin antibody, and negative for the AQP4 antibody.
## Jcn
been responsible for the selective noninflammatory CST degeneration in this case. [bib_ref] Cerebellar degeneration associated with Sjögren's syndrome, Kim [/bib_ref] Additional possibilities include vascular insufficiency or primary lateral sclerosis (PLS). pSS-associated vascular myelopathy is more compatible with an etiology of pure motor paraparesis following selective degeneration of CSTs. CSTs in the spinal cord have greater spinal cord blood flow and metabolic activity, which make them more vulnerable to ischemic injury. [bib_ref] Degeneration of axons in the corticospinal tract secondary to spinal cord ischemia..., Blisard [/bib_ref] This case also technically fulfills the diagnostic criteria for PLS proposed by Pringle et al. [bib_ref] Primary lateral sclerosis. clinical features, neuropathology and diagnostic criteria, Pringle [/bib_ref] However, this case seems less likely to be PLS considering the young age at onset, rapid worsening of paraplegia and CSTs atrophy, lack of progression to other body regions, and the presence of another disease (pSS) that was more like to underlie the myelopathy. [bib_ref] Progression in primary lateral sclerosis: a prospective analysis, Floeter [/bib_ref] The study protocol was approved by the Institutional Review Board at Chonnam National University Hospital, and the subject consented to the publication of her case.
## Conflicts of interest
The authors have no financial conflicts of interest.
[fig] Figure 1: Cervical MRI scans of the patient. A: At 2 months after symptom onset, a T2-weighted sagittal image shows no significant abnormality, but T2weighted axial images show suspicious focal hyperintensities (arrows) in lateral motor tracts bilaterally in the cervical spinal cord. B: MRI images obtained 8 months after the onset show spinal cord atrophy primarily of the lateral motor tracts, with preservation of the posterior columns. C: At 20 months after the onset, a T2-weighted sagittal image shows a linear hyperintensity (asterisk) extending to the upper thoracic levels of the vertebral column (from C2 to T1), and T2-weighted axial images show more-prominent atrophy in both the lateral and ventral columns (arrow heads), which correspond to the area of lateral and anterior corticospinal tracts. A B C [/fig]
|
The management of acute and chronic hyponatraemia
Hyponatraemia is the most common electrolyte abnormality encountered in clinical practice; despite this, the work-up and management of hyponatraemia remain suboptimal and varies among different specialist groups. The majority of data comparing hyponatraemia treatments have been observational, up until recently. The past two years have seen the publication of several randomised control trials investigating hyponatraemia treatments, both for chronic and acute hyponatraemia. In this article, we aim to provide a background to the physiology, cause and impact of hyponatraemia and summarise the most recent data on treatments for acute and chronic hyponatraemia, highlighting their efficacy, tolerability and adverse effects.
# Introduction
Hyponatraemia, defined as a plasma sodium concentration (pNa) < 135 mmol/L, is most common electrolyte abnormality in clinical practice. [bib_ref] The syndrome of inappropriate antidiuretic hormone: current and future management options, Sherlock [/bib_ref] Hyponatraemia poses diagnostic and therapeutic challenges, 2 as it is usually a pathophysiological consequence of an underlying medical condition which has led to disturbed water homeostasis. [bib_ref] Diagnosis and treatment of hyponatremia: compilation of the guidelines, Hoorn [/bib_ref] Hyponatraemia can manifest with a wide range of subtle clinical symptoms. Mild chronic hyponatraemia may be asymptomatic, though nausea, malaise, headache, lethargy, muscle cramps, disorientation and restlessness may all occur. Evidence has emerged in recent years to indicate that even apparent mild hyponatraemia is associated with gait abnormalities and cognitive deficits, which lead to increased frequency of falls and consequent bone fractures. [bib_ref] Hyponatremia is associated with increased osteoporosis and bone fractures in a large..., Usala [/bib_ref] In contrast, more severe hyponatraemia, with pNa < 120 mmol/L, can present with seizures, coma and cardiorespiratory arrest, particularly if the onset of hyponatraemia is rapid; acute hyponatraemia often requires emergency assessment and treatment as a consequence. [bib_ref] Society for Endocrinology Endocrine Emergency Guidance: emergency management of severe symptomatic hyponatraemia..., Ball [/bib_ref] Hyponatraemia has been shown to be associated with increased mortality in almost every published series in the literature spanning a diverse range of medical conditions. [bib_ref] Moderate hyponatremia is associated with increased risk of mortality: evidence from a..., Corona [/bib_ref] Recent data have demonstrated that mortality rates are cause-specific. In a prospective cohort study evaluating mortality in hyponatraemia stratified by volume status, have demonstrated higher mortality rates for hypervolaemic and hypovolaemic hyponatraemia, compared with syndrome of inappropriate antidiuresis (SIAD), although mortality rates in SIAD remained higher than eunatraemic controls. [bib_ref] Mortality rates are lower in SIAD, than in hypervolaemic or hypovolaemic hyponatraemia:..., Cuesta [/bib_ref] Further analysis found that mortality associated with SIAD is agedependent, with increased risk of in-hospital death reported in patients aged < 65 years with SIAD, but not in older patients with SIAD. [bib_ref] Active management of hyponatraemia and mortality in older hospitalised patients compared with..., Thorpe [/bib_ref] A recent population-based cohort study from Switzerland has also demonstrated an increased risk of in-hospital mortality, as well as 30-day hospital admission, in patients with hyponatraemia; subgroup analysis failed to demonstrate increased mortality risk in patients with SIAD, however the groups were not analysed by age. 8
## Physiology of hyponatraemia
In normal physiology, plasma sodium concentration is maintained within such tight parameters that clinical conditions in which hyponatraemia occur represent major abnormalities in the physiology of water balance. This tight homeostasis is achieved through the synergistic actions of The management of acute and chronic hyponatraemia the secretion and antidiuretic effect of vasopressin (AVP) and the activation of the thirst mechanism, resulting in day-to-day variations of plasma osmolality (determined by pNa) of less than 2%. [bib_ref] Neurogenic disorders of osmoregulation, Robertson [/bib_ref] Changes in plasma osmolality are sensed by osmoreceptor cells in the organum vasculosum lamina terminalis (OVLT) and subfornical organ of the anterior hypothalamus. This initiates neural signals to the supraoptic nuclei (SON) and paraventricular nuclei (PVN), which depolarize, stimulating the synthesis and secretion of AVP.AVP is synthesised along with copeptin and neurophysin II, as a preprohormone, which is then cleaved into the three component parts during transport down the axons of the magnocellular neurons to the neurohypophysis, from where AVP is released into the systemic circulation. AVP binds to the V2 receptor in the collecting duct of the kidney, generating new aquaporin-2 (AQ2) and stimulating the insertion of preformed AQ2 into the luminal membrane of the collecting tubules. This action culminates in a reduction of free water clearance and concentration of urine. Increases in plasma osmolality also simultaneously stimulate the thirst centre in the cerebral cortex. The resulting increased fluid intake, combined with actions of AVP, lead to increased plasma water and normalisation of plasma osmolality.
## Cause of hyponatraemia
Hyponatraemia should be initially classified by extracellular volume status into hypovolaemic, euvolaemic and hypervolaemic hyponatraemia 11 [fig_ref] Table 1: Classification of hyponatraemia based on volume status and urinary sodium concentration [/fig_ref]. This is a clinical approach, based on signs which are not completely sensitive or specific, and some acumen is required for accurate classification. A clinical diagnosis of hypovolaemia can be made on the basis of orthostatic hypotension, tachycardia, dry mucous membranes and decreased skin turgor, together with laboratory clues such as raised urea to creatinine ratio, elevated uric acid concentration and urinary sodium concentration < 20 mmol/L. A definitive diagnosis is not always possible at the time of presentation -for instance the distinction between euvolaemia and mild hypovolaemia is sometimes difficult to make on clinical grounds 12 and in this scenario the US guidelines recommend a therapeutic trial of isotonic saline with close monitoring of pNa response. [bib_ref] Diagnosis, evaluation, and treatment of hyponatremia: expert panel recommendations, Verbalis [/bib_ref] Despite it's limitations, categorising patients according to ECF volume status and urine electrolyte excretion does allow the initiation of cause-specific investigations and treatment. This approach to classification also has implications for prognosis, as there is good evidence from recent data that outcomes vary according to volume status, with significantly higher mortality rates in hypervolaemic and hypovolaemic hyponatraemia, compared with SIAD. [bib_ref] Mortality rates are lower in SIAD, than in hypervolaemic or hypovolaemic hyponatraemia:..., Cuesta [/bib_ref] It is important to exclude pseudohyponatraemia which occurs if the blood sample is lipemic or has a high concentration of immunoglobulins such as in patients with myeloma; the problem is largely seen where flame photometry is used for the measurement of electrolytes and is not seen with ion selective electrode measurements. In addition, hyperglycaemia lowers plasma sodium concentration, which needs no action other than lowering of blood glucose concentration. A discrepancy between measured serum osmolality and calculated osmolality raises the possibility of pseudohyponatraemia. 14 SIAD SIAD is a clinical and biochemical syndrome characterised by inappropriate urinary concentration and reduced free water excretion, in the setting of euvolaemic hyponatraemia. [bib_ref] Neurogenic disorders of osmoregulation, Robertson [/bib_ref] [bib_ref] SIAD: practical recommendations for diagnosis and management, Cuesta [/bib_ref] Schwartz & Bartter in 1957 reported the first cases of SIAD in two patients with small-cell lung cancer and proposed the diagnostic criteria for the condition [fig_ref] Table 2: Diagnostic criteria for SIAD [/fig_ref]. Although the vast majority of cases of SIAD are caused by inappropriately increased plasma AVP concentrations, a small minority of cases of SIAD result from a nephrogenic syndrome of inappropriate antidiuresis which occur due to gain of function mutations in the V2 vasopressin receptor located in the distal tubules and collecting ducts of the kidney. [bib_ref] SIAD: practical recommendations for diagnosis and management, Cuesta [/bib_ref] SIAD has been reported in a wide variety of medical conditions, including central nervous system (CNS) disorders, pulmonary disorders and malignancies. Medications account for 8-27% of cases of SIAD in hospital inpatients, [bib_ref] The syndrome of inappropriate antidiuretic hormone secretion: distribution and characterization according to..., Shepshelovich [/bib_ref] and an even greater proportion in elderly cohorts. Commonly implicated drugs include selective serotonin reuptake inhibitors, phenothiazine antipsychotics, carbamazepine, sodium valproate, chemotherapeutic agents particularly vincristine and cyclophosphamide, opioid analgesics and nonsteroidal antiinflammatory agents.
Ascertainment of the key diagnostic parameters for SIAD in routine clinical practice has invariably been poor. The international hyponatraemia registry confirmed only 47% of physician diagnosed SIAD had appropriately confirmed the diagnosis with plasma and urine osmolality and urine sodium measurements, and only 27% had excluded adrenal insufficiency and/or thyroid disease. [bib_ref] Current treatment practice and outcomes. Report of the hyponatremia registry, Greenberg [/bib_ref] The failure to make a correct diagnosis can potentially negatively impact patient outcomes.
A key differential in the diagnosis of SIAD is ACTH deficiency, as it presents with a similar biochemical picture. Mechanisms accounting for the development of hyponatraemia in adrenal insufficiency include elevated vasopressin even in the presence of plasma hypo-osmolality and impaired renal water handling in the absence of circulating free cortisol. [bib_ref] Hyponatremia and glucocorticoid deficiency, Garrahy [/bib_ref] In a recent prospective series, with full ascertainment of the minimum diagnostic criteria for SIAD in 83% of cases, Cuesta et al. showed that 3.8% of patients presenting with euvolaemic hyponatraemia to our institution had adrenal insufficiency. ACTH deficiency should be particularly suspected in patients who develop hyponatraemia after neurosurgical conditions or withdrawal of long-term glucocorticoid treatment. 19
## Prevalence of hyponatraemia
The prevalence of hyponatraemia in the literature varies significantly depending on the patient cohort and the biochemical cut-off used to define hyponatraemia. [bib_ref] Diagnosis, evaluation, and treatment of hyponatremia: expert panel recommendations, Verbalis [/bib_ref] Large population studies have reported rates of 2-8%, [bib_ref] Prevalence of hyponatremia and association with mortality: results from NHANES, Mohan [/bib_ref] higher rates (37%) were reported in a Danish study which used a higher plasma sodium cut-off of < 137 mmol/L. [bib_ref] Hyponatraemia at hospital admission is a predictor of overall mortality, Balling [/bib_ref] Hyponatraemia is commonly seen in oncology patients, particularly those with small cell lung cancer, [bib_ref] Diagnosis and management of hyponatremia in cancer patients, Castillo [/bib_ref] and in patients with pneumonia, [bib_ref] Hyponatraemia in patients with communityacquired pneumonia; prevalence and aetiology, and natural history..., Cuesta [/bib_ref] including COVID-19 infection, [bib_ref] Serum sodium alterations in SARS CoV-2 (COVID-19) infection: impact on patient outcome, Berni [/bib_ref] congestive heart failure and liver failure. [bib_ref] Hyponatremia in cirrhosis: results of a patient population survey, Angeli [/bib_ref] Hyponatraemia occurs in 10-34% of patients in the intensive care unit (ICU), and is particularly common in neurosurgical units. The highest incidence is seen following SAH (20-56%), [bib_ref] Hyponatremia following mild/moderate subarachnoid hemorrhage is due to SIAD and glucocorticoid deficiency..., Hannon [/bib_ref] [bib_ref] Prognostic significance of hypernatremia and hyponatremia among patients with aneurysmal subarachnoid hemorrhage, Qureshi [/bib_ref] [bib_ref] The incidence and pathophysiology of hyponatraemia after subarachnoid haemorrhage, Sherlock [/bib_ref] followed by intracranial tumours (16%), traumatic brain injury (TBI) (10-15%) [bib_ref] Posterior pituitary dysfunction after traumatic brain injury, Agha [/bib_ref] [bib_ref] Acute glucocorticoid deficiency and diabetes insipidus are common after acute traumatic brain..., Hannon [/bib_ref] and pituitary surgery (6-25%). 37,38 SIAD is the most common underlying pathophysiology, but acute ACTH deficiency is increasingly recognised as a cause of euvolaemic hyponatraemia in this patient group. [bib_ref] Glucocorticoid deficiency and syndrome of inappropriate antidiuresis: an underdiagnosed association?, Garrahy [/bib_ref] Cerebral salt wasting (CSW) is another differential diagnosis in neurosurgical patients, although it is exceedingly rare. A prospective study of 100 cases of SAH by Hannon et al. [bib_ref] Hyponatremia following mild/moderate subarachnoid hemorrhage is due to SIAD and glucocorticoid deficiency..., Hannon [/bib_ref] failed to reveal a single case, even after careful clinical and biochemical assessment of volume status, and some experts challenge whether the diagnosis is a real entity. CSW can be distinguished from SIAD by the demonstration of an initial period of natriuresis and hypovolaemia preceding the onset of hyponatraemia. 40
## Impact of treatment of hyponatraemia; why should we treat it?
While acute hyponatraemia is a medical emergency which can lead to death if untreated, chronic hyponatraemia has long been thought of as a benign condition and is often left under-investigated and untreated. However, there is now emerging evidence of the positive impact of treating chronic hyponatraemia on outcomes such as cognition, gait, and quality of life. In a seminal study by Renneboog et al., [bib_ref] Mild chronic hyponatremia is associated with falls, unsteadiness, and attention deficits, Renneboog [/bib_ref] active treatment of chronic hyponatraemia (with urea in the majority of cases) led to improvements in gait parameters and reaction times. In a follow-up study, the same group demonstrated lower baseline scores of tandem gait and attention, and greater gains when hyponatraemia was actively managed, in elderly patients compared with younger subjects with the same degree of hyponatraemia. [bib_ref] Attention and postural balance are much more affected in older than in..., Renneboog [/bib_ref] More recently, Brinkkoetter et al. [bib_ref] Impact of resolution of hyponatremia on neurocognitive and motor performance in geriatric..., Brinkkoetter [/bib_ref] have shown that an increase of pNa of at least 5 mmol/L in patients aged > 70 years leads to improvements in cognitive scores (Mini Mental State Exam, MMSE) and scores of independence (Barthel index of Activities of Independent Living) compared with eunatraemic controls. Largest effect sizes were seen in patients with euvolaemic hyponatraemia and multivariable linear regression analysis confirmed change in pNa to be an independent predictor in MMSE improvement. It has yet to be proven in prospective studies that treatment of hyponatraemia reduces risk of falls and fractures; these studies are awaited.
A 2015 meta-analysis has reported, for the first time, that improvement in plasma sodium in hyponatraemic patients is associated with a reduction in overall mortality (OR 0.57, p = 0.002), which persisted at 12 months; the greatest reduction was seen in older patients and those with lowest plasma sodium at baseline. [bib_ref] Hyponatremia improvement is associated with a reduced risk of mortality: evidence from..., Corona [/bib_ref] While a cause-effect relationship cannot be extrapolated from the data included in this analysis, this report provided the first evidence for a potential mortality benefit from treating chronic hyponatraemia. More recent data from our group has demonstrated that increased specialist input and active management of severe hyponatraemia over a 10-year period, were associated with a fall in mortality rates, to rates comparable with moderate hyponatraemia. 45
## Treatment of hyponatraemia
General approach. There are a number of key issues which influence the implementation of optimum management of hyponatraemia. Accurate assessment of the underlying cause of hyponatraemia is essential to institute cause-specific therapy.
There are causes of hyponatraemia where simply treatment of the underlying cause reverses hyponatraemia without specific treatment for the electrolyte disorder; in community-acquired pneumonia for instance, SIAD reverses to normal with antibiotic therapy and resolution of infection. [bib_ref] Hyponatraemia in patients with communityacquired pneumonia; prevalence and aetiology, and natural history..., Cuesta [/bib_ref] Equally, volume status is essential to quantify, as the treatment of hypovolaemic hyponatraemia, with intravenous saline, is very different to that of hypervolaemic hyponatraemia, for which the therapeutic intervention is diuretic therapy [fig_ref] Figure 1: General approach to treatment of hyponatraemia [/fig_ref].
Awareness of the chronicity of low plasma sodium concentration is the key issue that determines the urgency of treatment and the speed of reversal of hyponatraemia. Patients with acute hyponatraemia require emergency treatment to prevent cerebral herniation and death. In contrast, patients with chronic hyponatraemia are much less likely to develop neurological symptoms as cerebral adaptation protects against the development of raised intracranial pressure. In addition, patients with chronic hyponatraemia are particularly vulnerable to osmotic demyelination (ODS) if there is a rapid rise in extracellular tonicity caused by overaggressive treatment. Therefore, risk-benefit analysis favours a slow rise in pNa in this patient group. First described in 1959, osmotic demyelination syndrome occurs when correction of chronic hyponatraemia exceeds the ability of the brain to reverse compensatory mechanisms. Failure to recapture organic solutes as pNa rises results in an inverse osmotic gradient leading to dehydration of the cells and possible demyelination of white matter, [bib_ref] Central pontine and extrapontine myelinolysis: a systematic review, Singh [/bib_ref] characterised clinically by spastic quadriparesis and pseudobulbar palsy, and in most severe cases 'locked in' syndrome.
Close monitoring of pNa is recommended during the active correction phase. Hourly urine output monitoring is very helpful, as often treatment of the underlying cause of hyponatraemia (e.g. volume replacement in hypovolaemic hyponatraemia) can lead to suppression of AVP, and a rapid aquaresis, resulting in rapid rise in pNa. Both European and American consensus guidelines recommend that plasma sodium rise in chronic hyponatraemia generally should not exceed 10 mmol/l per 24 hours, with the US guidelines specifying that the rise should ideally be restricted to <8 mmol/l. [bib_ref] Clinical practice guideline on diagnosis and treatment of hyponatraemia, Spasovski [/bib_ref] [bib_ref] Diagnosis, evaluation, and treatment of hyponatremia: expert panel recommendations, Verbalis [/bib_ref] The US guidelines also include revised targets for patients deemed to be at greater risk of osmotic demyelination, such as those with liver disease, alcohol misuse, hypokalaemia or malnutrition. In these scenarios, targets are revised downwards to a maximum recommended elevation in plasma sodium of < 6 mmol/l and should not exceed 8 mmol/L in 24 hours [fig_ref] Table 3: Targets for elevation in plasma sodium concentration recommended to avoid osmotic demyelination... [/fig_ref].
In patients with acute hyponatraemia, when the acuity of onset is certain, the rise in pNa need not be restricted. Both guidelines now recommend 3% hypertonic saline bolus in the treatment for acute and symptomatic hyponatraemia, the goal being an initial brisk rise in pNa to reduce cerebral oedema. In chronic symptomatic hyponatraemia, the goal then quickly shifts to maintaining 24-and 48-hour correction within the safe thresholds for chronic hyponatraemia.
## Reversal of overcorrection.
If overcorrection is anticipated based on trend in pNa and urine output, further hyponatraemia-targeted therapy should be held and treatment instigated to prevent overcorrection. Plasma sodium concentration can be kept stable either with the administration of IV hypotonic fluids (enteral water or IV 5% dextrose) to match urine output or the administration of 2-4 mcg parenteral desmopressin to clamp urine output and prevent further aquaresis. If overcorrection does occur, therapeutic relowering of pNa can be considered, by administering 2-4 mcg parenteral desmopressin with bolus of 3 ml/ kg hypotonic fluids, until target pNa is reached. [bib_ref] Diagnosis, evaluation, and treatment of hyponatremia: expert panel recommendations, Verbalis [/bib_ref] The US guidelines stipulate that this is probably only necessary if starting pNa was <120 mmol/L, [bib_ref] Diagnosis, evaluation, and treatment of hyponatremia: expert panel recommendations, Verbalis [/bib_ref] although the European guidelines do not make this distinction. [bib_ref] Prognostic significance of hypernatremia and hyponatremia among patients with aneurysmal subarachnoid hemorrhage, Qureshi [/bib_ref] It should be emphasised that the clinical impact of therapeutic re-lowering of pNa has not been well studied, and both guidelines include the caveat that the therapeutic approaches to overcorrection of hyponatraemia are not validated in prospective randomised controlled trials.
## Treatment of acute hyponatraemia
The administration of hypertonic saline is the recommended treatment of choice for acute hyponatraemia. In a major therapeutic policy change, bolus 3% hypertonic saline is now recommended to treat symptomatic acute hyponatraemia in place of continuous infusion of 3% saline. [bib_ref] Clinical practice guideline on diagnosis and treatment of hyponatraemia, Spasovski [/bib_ref] [bib_ref] Diagnosis, evaluation, and treatment of hyponatremia: expert panel recommendations, Verbalis [/bib_ref] This bolus approach was first applied in the management of exercise induced hyponatraemia. [bib_ref] Diagnosis and treatment of hyponatremia: compilation of the guidelines, Hoorn [/bib_ref] The aim of this approach is to achieve a rapid early rise in pNa of 4-6 mmol/L over the initial four hours of treatment, a target derived from neurosurgical data in which an increment of 5 mmol/L reverses clinical signs of trans-tentorial herniation and reduces intracerebral pressure by nearly 50% within an hour of hypertonic saline administration in eunatraemic patients [bib_ref] Reversal of transtentorial herniation with hypertonic saline, Koenig [/bib_ref] This is a potentially life-saving treatment for cerebral oedema secondary to hyponatraemia, as the increased extracellular sodium concentration immediately removes water from the intracellular space.
The European guidelines recommend the administration of 150 ml of hypertonic saline over a period of 20 minutes, at which point a repeat plasma sodium measurement is taken, and another infusion of 150 ml 3% hypertonic saline commenced; this can be repeated twice or until a rise in pNa of 5 mmol/L is achieved. [bib_ref] Clinical practice guideline on diagnosis and treatment of hyponatraemia, Spasovski [/bib_ref] If hyponatraemia is presenting with moderate symptoms, then a once off bolus of 150 ml hypertonic saline can be administered. [bib_ref] Clinical practice guideline on diagnosis and treatment of hyponatraemia, Spasovski [/bib_ref] The US guidelines differ slightly; in symptomatically severe hyponatraemia they recommend 100 ml bolus hypertonic saline over 10 minutes to be repeated up to twice more. In patients with moderate symptoms of hyponatraemia, the US guidelines recommended a continuous slow infusion of 3% normal saline at a rate of 0.5-2 ml/kg per hour. [bib_ref] Diagnosis, evaluation, and treatment of hyponatremia: expert panel recommendations, Verbalis [/bib_ref] Patients should be in a monitored environment, with close supervision of neurological status, urine output and pNa.
In 2019, Garrahy et al. [bib_ref] Continuous versus bolus infusion of hypertonic saline in the treatment of symptomatic..., Garrahy [/bib_ref] compared prospectively collected clinical and biochemical outcomes in patients with symptomatic hyponatraemia secondary to SIAD treated with 100 ml boluses of 3% saline and compared them to a historical cohort treated with the traditional continuous infusion of 3% saline (20 ml/hr, rate adjusted according to response) [fig_ref] Figure 2: Change in plasma sodium concentration from baseline in patients with SIAD treated... [/fig_ref]. We found that bolus 3% saline resulted in a faster initial rise within the first six hours of treatment (6 mmol/L versus 3 mmol/L, p < 0.0001) with quicker restoration in GCS compared to traditional continuous infusion. Following twenty-four hours of closely observed treatment, plasma sodium concentrations were similar in both groups. Interventions to prevent overcorrection such as intravenous dextrose/dDAVP were more frequently used in the bolus group (5/22 versus 0/28, p = 0.008), particularly in those patients who received three boluses. [bib_ref] Continuous versus bolus infusion of hypertonic saline in the treatment of symptomatic..., Garrahy [/bib_ref] The following year, the SALSA randomised controlled clinical trial found that both weight-based rapid bolus infusion and slow continuous infusion of 3% hypertonic saline were efficacious and safe. The incidence of overcorrection, the primary endpoint of the study, was similar in both groups occurring in 17.2% in the rapid infusion group compared with 24.2% in the continuous infusion group, p = 0.26. However, the incidence of therapeutic re-lowering treatment was significantly lower in the rapid infusion group, 41.4% versus 57.1%, p = 0.04. A greater rise in plasma Na was achieved in the first hour of treatment in the rapid bolus infusion group, supporting the role of bolus hypertonic saline in acute life-threatening hyponatraemia. [bib_ref] Risk of overcorrection in rapid intermittent bolus versus slow continuous infusion therapies..., Baek [/bib_ref] There were no cases of ODS in either study. More recently, in 2021, in an observational study by Chifi et al., 28% of patients with moderate or severely symptomatic hyponatraemia treated with 150 ml boluses of hypertonic saline were given 5% dextrose/or dDAVP to re-lower plasma sodium concentration, which was successful in maintaining rise of pNa within target in less than half of cases. Patients treated with bolus hypertonic saline, however, had less fluctuations in pNa and reached correction thresholds more often than those treated with 'conventional' treatments. [bib_ref] Treatment of symptomatic hyponatremia with hypertonic saline: a real-life observational study, Chifu [/bib_ref] While the above-mentioned studies differ in design, inclusion criteria, control group and bolus dose, in general, they demonstrate that bolus hypertonic saline produces a quicker initial rise in pNa when clinically needed in acute hyponatraemia, but carries a significant risk of overcorrection. Cautious monitoring of urine output and pNa is required, along with early implementation of stalling or re-lowering measures where the duration of hyponatraemia is not known to be acute, to reduce the risk of ODS. An alternative strategy proposed by some experts to mitigate this risk of overcorrection is to administer desmopressin at the onset of treatment. This acts to clamp urinary losses of electrolyte free water, and 3% saline is then titrated according to plasma sodium response. [bib_ref] Hypertonic saline and desmopressin: a simple strategy for safe correction of severe..., Sood [/bib_ref] While a recent retrospective study failed to show any reduction in overcorrection with this 'pro-active' approach, [bib_ref] Evaluation of desmopressin in critically ill patients with hyponatremia requiring 3% hypertonic..., Tran [/bib_ref] it may be of use in those at a particularly high risk of rapid water diuresis with overcorrection of plasma sodium, e.g. hypovolaemic hyponatraemia, thiazide induced hyponatraemia, glucocorticoid deficiency and primary polydipsia, where treatment of the underlying condition results in restoration of the kidneys ability to excrete electrolyte-free water. 11
## Treatment of chronic hyponatraemia
Treatment of chronic hyponatraemia should be directed at reversal or removal of the underlying cause. Patients with chronic hyponatraemia are much less likely to develop neurological symptoms but are at higher risk of osmotic demyelination. Therefore, risk benefit assessment favours a slower rise in plasma sodium concentration in this patient group.
Treatment of SIAD. Treatment of the underlying cause of SIAD is always the first therapeutic step, e.g. removal of offending medication or treatment of pneumonia. Key aspects of currently available treatments for SIAD are summarised in [fig_ref] Table 4: Comparison of treatment options for chronic SIAD [/fig_ref].
Until recently, the majority of data for treatments of SIAD were observational and/or retrospective. The SALT trials published in 2006 were the first randomised controlled trials of a hyponatraemia treatment, and in the past two years there have been a further three randomised controlled trials investigating treatment options for SIAD. The results of these studies are summarised in [fig_ref] Table 5: Summary of recent randomised controlled trials investigating treatments for SIAD [/fig_ref].
Fluid restriction. Fluid restriction is the recommended first-line treatment for SIAD across all international guidelines, [bib_ref] Clinical practice guideline on diagnosis and treatment of hyponatraemia, Spasovski [/bib_ref] [bib_ref] Diagnosis, evaluation, and treatment of hyponatremia: expert panel recommendations, Verbalis [/bib_ref] and the most commonly employed treatment strategy for hyponatraemia in clinical practice. [bib_ref] Current treatment practice and outcomes. Report of the hyponatremia registry, Greenberg [/bib_ref] Despite being the first-line treatment, observational data from the large Hyponatraemia Registry study reported that less than half of patients (44%) with SIAD treated with fluid restriction corrected their pNa by >5 mmol/L and that the majority of patients required additional treatment to reach clinical goals. [bib_ref] Diagnosing and treating the syndrome of inappropriate antidiuretic hormone secretion, Verbalis [/bib_ref] Several clinical and biochemical factors have been proposed as predictors of response to FR, [fig_ref] Table 6: Predictors of failure to respond to fluid restriction [/fig_ref]. [bib_ref] Diagnosis, evaluation, and treatment of hyponatremia: expert panel recommendations, Verbalis [/bib_ref] [bib_ref] Predictors of nonresponse to fluid restriction in hyponatraemia due to the syndrome..., Winzeler [/bib_ref] In a two centre analysis of patients embarking on fluid restriction for SIAD, 40% of patients had one predictor of failure to respond to FR, and 60% had two predictors. [bib_ref] Predictors of failure to respond to fluid restriction in SIAD in clinical..., Cuesta [/bib_ref] It is therefore hardly surprising that FR does not deliver the clinical impact that clinicians require.
The Furst equation ratio, calculated by dividing the sum of urinary sodium and potassium concentrations by plasma sodium, has been utilised to predict the degree of fluid restriction which is required for effective elevation of pNa. [bib_ref] The urine/ plasma electrolyte ratio: a predictive guide to water restriction, Furst [/bib_ref] In patients with a ratio greater than 1 (concentrated urine), a 500 ml fluid restriction per day is advised for correction of hyponatraemia. Clearly, this stringent target is extremely difficult to adhere, as downwards re-setting of the thirst threshold in SIAD dictates that patients continue to experience thirst despite hyponatraemia. [bib_ref] Downward resetting of the osmotic threshold for thirst in patients with SIADH, Smith [/bib_ref] This undoubtedly impacts negatively on long-term compliance with fluid restriction. [bib_ref] New approach to disturbances in the plasma sodium concentration, Rose [/bib_ref] Where compliance is maintained, the rigour of this degree of fluid restriction may predispose to volume depletion. [bib_ref] Efficacy of furosemide, oral sodium chloride, and fluid restriction for treatment of..., Krisanapan [/bib_ref] [bib_ref] Treatment outcomes in syndrome of inappropriate antidiuresis: improvements in hyponatremia may reflect..., Garrahy [/bib_ref] In addition, fluid restriction may not always be possible in the clinical context of adjunctive therapies; for example a patient with SIAD due to underlying neoplastic disease may not be possible to fluid restrict because of the requirements for intravenous fluids as part of chemotherapy.
Although fluid restriction has been recommended as first-line treatment for SIAD, for years there were no data on safety and efficacy. However, a recent prospective randomised controlled trial from our group compared fluid restriction with no hyponatraemia treatment, in ambulatory patients with chronic SIAD who had no reversible or transient cause. The results of this study demonstrated that patients with SIAD, who were fluid restricted to 1 L/day, had an early rise in pNa which was maintained at 30 days and significantly greater than that seen in untreated patients (3 versus 1 mmol/L after 3 days, p = 0.005). Although fluid restriction was statistically better than 'no treatment', the rise in plasma sodium concentration was modest, and less than one in five patients treated with fluid restriction had a rise in pNa of ⩾ 5 mmol/L after 3 days of treatment. Fewer than half of patients achieved this target after 30 days of fluid restriction. The proportion of patients achieving a pNa ⩾ 130 mmol/L was 61% and 71% after 3 and 30 days treatment respectively. [bib_ref] Fluid restriction therapy for chronic SIAD; results of a prospective randomized controlled..., Garrahy [/bib_ref] While the modest improvements in pNa and the favourable safety profile, justify the position of FR as first-line therapy, the data from this study challenge the effectiveness of the treatment in a significant proportion of patients. This highlights the need for effective and affordable second-line treatments for chronic SIAD.
The rise in plasma sodium concentration in two other recently published randomised trials was greater than that which was published in our study. In a comparison of empagliflozin versus FR, and a separate analysis of frusemide versus FR, the rise in pNa in the FR groups was 7 and 5 mmol/L after 3 days, respectively. However, neither study restricted inclusion to patients with chronic SIAD, so the higher rise in pNa in these studies may have reflected a cumulative effect of FR and the treatment and/or resolution of the underlying cause of hyponatraemia. Vasopressin receptor antagonists. The development of vasopressin receptor antagonists (VRA) represented a major advance in SIAD treatment. This class of drug competitively bind to the V2 receptor in the collecting duct, displacing vasopressin, promoting free water clearance and rise in pNa. [bib_ref] A metabolically stable apelin-17 analog decreases AVP-induced antidiuresis and improves hyponatremia, Flahault [/bib_ref] Tolvaptan is an oral selective antagonist of V2 receptor available in Europe and USA, while conivaptan is an intravenous antagonist of V1 and V2 receptors available in the USA.
In the large multicentre, randomised doubleblind placebo controlled trials, SALT1 and SALT2, tolvaptan was shown to produce an effective rise in plasma sodium concentration at day 4 and day 30 in patients with euvolaemic and hypervolaemic hyponatraemia, compared to patients in the placebo group. [bib_ref] Tolvaptan, a selective oral vasopressin V2-receptor antagonist, for hyponatremia, Schrier [/bib_ref] Subsequent subgroup analysis confirmed the efficacy of tolvaptan in the SIAD cohort in the study. [bib_ref] Efficacy and safety of oral tolvaptan therapy in patients with the syndrome..., Verbalis [/bib_ref] Plasma sodium concentration fell following cessation of the drug at 30 days, confirming that the positive effect was related to tolvaptan and not spontaneous recovery of hyponatraemia. The long-term safety of tolvaptan has been confirmed in the SALTWATER study, a multicentre open-label four-year extension of the SALT-1 and SALT-2 trials. [bib_ref] Oral tolvaptan is safe and effective in chronic hyponatremia, Berl [/bib_ref] Plasma sodium concentrations were well maintained over the long-term with a favourable adverse effect profile. The most commonly reported adverse effects related to tolvaptan therapy include urinary frequency and polyuria, both of which are expected due to the mechanism of action of the drug, and good indicators that the drug is functioning. Patients also reported thirst, dry mouth and polydipsia. We have previously reported two cases (and encountered a third) of acquired resistance to tolvaptan. All three cases occurred in patients with SIAD associated with small cell lung cancer, and resistance to tolvaptan effect occurred despite increasing doses of the drug. Resistance to the aquaretic effects of tolvaptan occurred in the setting of progression of malignant disease, with escalating plasma AVP concentrations which we postulate overwhelmed the drug at the V2 receptor. [bib_ref] Secondary resistance to tolvaptan in two patients with SIAD due to small..., Garrahy [/bib_ref] The European guidelines specifically advised against the use of tolvaptan, because of concerns about overcorrection of hyponatraemia. [bib_ref] Clinical practice guideline on diagnosis and treatment of hyponatraemia, Spasovski [/bib_ref] Overcorrection was reported in 1.8% of patients treated with tolvaptan in the SALT studies, [bib_ref] Tolvaptan, a selective oral vasopressin V2-receptor antagonist, for hyponatremia, Schrier [/bib_ref] and no patient displayed symptoms of osmotic demyelination. However, subsequent real world studies have reported much higher rates of overcorrection, 12.1-31%. [bib_ref] Current treatment practice and outcomes. Report of the hyponatremia registry, Greenberg [/bib_ref] [bib_ref] Reallife experience of tolvaptan use in the treatment of severe hyponatraemia due..., Tzoulis [/bib_ref] [bib_ref] Clinical experience of the efficacy and safety of low-dose tolvaptan therapy in..., Chatzimavridou-Grigoriadou [/bib_ref] Overcorrection is a possibility with all effective therapies for hyponatraemia, and the key issues are awareness of the possibility, frequent monitoring of plasma sodium concentration in the 24 hours after initiation of therapy, and willingness to correct early with dextrose or dDAVP. Concurrent use of other hyponatraemia treatments should be avoided.
Overcorrection is more likely where the starting plasma sodium concentration is very low. [bib_ref] Reallife experience of tolvaptan use in the treatment of severe hyponatraemia due..., Tzoulis [/bib_ref] [bib_ref] Rapidity of correction of hyponatremia due to syndrome of inappropriate secretion of..., Morris [/bib_ref] A case series of 61 hospital inpatients by Tzoulis et al. reported a significant negative correlation between baseline plasma sodium concentration and 24 hour sodium rise; 29% of patients with starting pNa <125 mmol/L exceeded the safe rate for pNa correction at any timepoint compared with none of those with pNa ⩾125 mmol/L. [bib_ref] Reallife experience of tolvaptan use in the treatment of severe hyponatraemia due..., Tzoulis [/bib_ref] [bib_ref] Rapidity of correction of hyponatremia due to syndrome of inappropriate secretion of..., Morris [/bib_ref] Lower doses of tolvaptan (7.5 mg or less) are efficacious, [bib_ref] Clinical experience of the efficacy and safety of low-dose tolvaptan therapy in..., Chatzimavridou-Grigoriadou [/bib_ref] [bib_ref] Equivalent efficacy and decreased rate of overcorrection in patients with syndrome of..., Hanna [/bib_ref] [bib_ref] Low-dose tolvaptan for the treatment of hyponatremia in the syndrome of inappropriate..., Harbeck [/bib_ref] and may be safer, though it is important to stress that plasma sodium responses to lower doses of tolvaptan have not been subjected to rigorous prospective study.
Tolvaptan has been utilised in the treatment of SIAD post-pituitary surgery, with a German group reporting that tolvaptan 7.5 mg was more efficacious than fluid restriction in correcting hyponatraemia in this cohort of patients. However, it should be noted with caution that one third of patients treated with tolvaptan had overcorrection of pNa, most likely because the aquaretic effect of tolvaptan are compounded by the transient self-correcting nature of SIAD in this clinical context. [bib_ref] Tolvaptan versus fluid restriction in the treatment of hyponatremia resulting from SIADH..., Kleindienst [/bib_ref] The principal drawback to vaptan therapy perceived by most clinicians is the cost of the drugs, which are not uniformly reimbursable in all countries. Although there have been cost/benefit analyses which have sought to justify the use of vaptans, 75 they remain prohibitively expensive given that the main indication is for chronic SIAD, which may need prolonged therapy, particularly in the absence of data on reduction of hard clinical end points.
Demeclocycline. Demeclocycline is a tetracycline derivative that has been used clinically since the 1970s to treat chronic hyponatraemia secondary to SIAD. [bib_ref] Evidence for the use of demeclocycline in the treatment of hyponatraemia secondary..., Miell [/bib_ref] The mode of action in the treatment of hyponatraemia is not yet fully established, however, it has been observed that it interferes with the action of vasopressin on the renal ducts inducing diabetes insipidus in 60% of individuals with SIAD. [bib_ref] The syndrome of inappropriate antidiuretic hormone: current and future management options, Sherlock [/bib_ref] [bib_ref] Evidence for the use of demeclocycline in the treatment of hyponatraemia secondary..., Miell [/bib_ref] [bib_ref] Demeclocycline in the treatment of the syndrome of inappropriate secretion of antidiuretic..., Perks [/bib_ref] Unfortunately, the onset of action is erratic, usually occurring between 3 and 5 days, but often occurring much later, [bib_ref] Evidence for the use of demeclocycline in the treatment of hyponatraemia secondary..., Miell [/bib_ref] [bib_ref] Demeclocycline in the treatment of the syndrome of inappropriate secretion of antidiuretic..., Perks [/bib_ref] [bib_ref] Renal failure associated with demeclocycline in cirrhosis, Carrilho [/bib_ref] [bib_ref] Clinical management of SIADH, Gross [/bib_ref] and this means that dose adjustment is less predictable. Side effects include nausea and vomiting, renal failure, skin photosensitivity and antibiotic resistance. [bib_ref] Serious hyponatremia in patients with cancer: management with demeclocycline, Trump [/bib_ref] [bib_ref] Irreversible nephrotoxicity from demeclocycline in the treatment of hyponatramia, Curtis [/bib_ref] Renal failure is normally reversible on cessation of the drug. [bib_ref] Renal failure associated with demeclocycline in cirrhosis, Carrilho [/bib_ref] In some patients polyuria can be profound and can lead to hypernatraemia if fluid intake is restricted. [bib_ref] The syndrome of inappropriate antidiuretic hormone: current and future management options, Sherlock [/bib_ref] [bib_ref] Severe hypernatremia following treatment of the syndrome of inappropriate antidiuretic hormone secretion, Soudan [/bib_ref] There are no prospective randomised controlled data to support the use of demeclocycline, and it is no longer universally recommended across consensus guidelines for hyponatraemia management.
Urea. Oral urea has been used to treat hyponatraemia in SIAD since 1980. [bib_ref] Treatment of the syndrome of inappropriate secretion of antidiuretic hormone by urea, Decaux [/bib_ref] It is an osmotically active agent that increases urinary free water excretion and decreases urinary sodium excretion. [bib_ref] Treatment of the syndrome of inappropriate secretion of antidiuretic hormone by urea, Decaux [/bib_ref] In an 1981 study by Decaux & Genette, seven patients with confirmed SIAD with a mean pNa of 115 ± 6 mmol/l and mean urine osmolality of 514 ± 79 mOsm/kg H 2 O were administered between 30 and 60 g of urea once daily. After treatment, a significant rise in serum Na to 136 ± 3.5 mmol/l (p < 0.001) was demonstrated, with a concomitant rise in urine osmolality to 652 ± 95 mOsm/kg H 2 O (p < 0.001). [bib_ref] Urea for long-term treatment of syndrome of inappropriate secretion of antidiuretic hormone, Genette [/bib_ref] Retrospective studies have shown urea to be effective in SIAD induced by SAH and in critical care patients, [bib_ref] Treatment of euvolemic hyponatremia in the intensive care unit by urea, Decaux [/bib_ref] as well as in twelve patients who transitioned to urea therapy following 12 months of vaptan therapy in a clinical trial. [bib_ref] Efficacy and tolerance of urea compared with vaptans for long-term treatment of..., Soupart [/bib_ref] A more recent retrospective non-controlled study from the US (n = 58) compared outcomes in patients treated with an oral commercially available formulation of urea and those treated by other means, over a one year period. Treatment with urea resulted in a statistically significant rise in pNa following 24 hours (2.8 mmol/L versus -0.5 mmol/L in the non-urea group; p < 0.05), [bib_ref] Urea for the treatment of hyponatremia, Rondon-Berrios [/bib_ref] and a median rise in pNa of 6 mmol/L over 4 days, without overcorrection or other adverse events. [bib_ref] Urea for the treatment of hyponatremia, Rondon-Berrios [/bib_ref] Contraindications to the use of urea in the treatment of hyponatraemic include liver cirrhosis, adrenal insufficiency and states of hypovolemia. [bib_ref] The syndrome of impending hepatic coma in patients with cirrhosis of the..., Phillips [/bib_ref] SGLT-2 inhibitors. Sodium glucose co-transport 2 inhibitors (SGLT-2i) are a well-established class of antidiabetic medication. They act by promoting osmotic diuresis via urinary glucose excretion and therefore offer a promising new management option for SIAD. [bib_ref] A randomized trial of empagliflozin to increase plasma sodium levels in patients..., Refardt [/bib_ref] Preliminary data from Refardt et al. [bib_ref] Empagliflozin increases short-term urinary volume output in artificially induced syndrome of inappropriate..., Refardt [/bib_ref] demonstrated significantly increased urinary water excretion in healthy volunteers with artificially induced SIAD within 6 hours of receiving a dose of empagliflozin, an oral SGLT2 inhibitor; this suggests the drug may have a beneficial effect on hyponatraemia acutely. The same group recently published a double-blind randomised placebo-controlled trial, in which hospitalised patients with SIAD were randomised to standard treatment of 1000 ml/day fluid restriction with or without empagliflozin once daily for 4 days. Patients treated with empagliflozin had a greater increase in pNa compared to those treated with standard fluid restriction alone, 10 versus 7 mmol/L, p = 0.04. [bib_ref] A randomized trial of empagliflozin to increase plasma sodium levels in patients..., Refardt [/bib_ref] Empagliflozin may be a favourable treatment option in the treatment of hyponatraemia going forward, however, longterm efficacy or safety data in the outpatient setting are not yet available. There are however extensive data on the use of empagliflozin in patients with type 2 diabetes, renal disease and heart failure; the drug has been shown to have a favourable safety profile, beneficial cardio-protective and renal protective profiles and therefore may be an excellent treatment option for comorbid patients with SIAD. [bib_ref] A randomized trial of empagliflozin to increase plasma sodium levels in patients..., Refardt [/bib_ref] [bib_ref] Empagliflozin and progression of kidney disease in type 2 diabetes, Wanner [/bib_ref] [bib_ref] Empagliflozin, cardiovascular outcomes, and mortality in type 2 diabetes, Zinman [/bib_ref] Apelin. In normal physiological conditions, apelin and AVP are released in proportional concentrations dependent on plasma osmolality. Water resorption occurs once AVP binds to V2-R, increasing aquaporin-2 insertion. Apelin is a neuro-vasoactive peptide that works centrally to inhibit AVP release and at the level of the collecting duct promoting water excretion through its action on apelin-R counteracting the antidiuretic effect of AVP. A recent investigation has proven that a metabolically stable K17F analogue, LIT01-196, restores water homeostasis in SIAD in animal models. LIT01-196 inhibits the antidiuretic effect of AVP, by increasing urine output and reducing urinary osmolality, thereby causing a rise in pNa. This study illustrates metabolically stable apelin analogues have a potential role in the treatment of hyponatraemia in patients with SIAD, pending appropriate clinical trial data. [bib_ref] A metabolically stable apelin-17 analog decreases AVP-induced antidiuresis and improves hyponatremia, Flahault [/bib_ref] Treatment of hypervolemic hyponatraemia A combination of fluid and salt restriction and loop diuretics, plus neurohormonal antagonism is the recommended first-line approach for hyponatraemia in the settings of heart and liver failure. Vaptans offer an attractive alternative or adjunct to loop diuretics, as they are potassium neutral and cause less neuro-hormonal activation from intravascular volume depletion compared with diuretics, and tolvaptan is an FDA-approved treatment for hypervolaemic hyponatraemia. Forty-three percent of patients recruited into the SALT RCTs of tolvaptan mentioned earlier had hypervolemic hyponatraemia, and although subgroup analyses of this group were not published separately, tolvaptan was shown to produce a significantly greater rise in pNa compared with placebo in the overall group. [bib_ref] Tolvaptan, a selective oral vasopressin V2-receptor antagonist, for hyponatremia, Schrier [/bib_ref] Several large RCTs have subsequently confirmed tolvaptan's beneficial effect on plasma sodium, weight and dyspnoea in heart failure patients, [bib_ref] Effects of oral tolvaptan in patients hospitalized for worsening heart failure: the..., Konstam [/bib_ref] [bib_ref] The short-term and long-term effects of tolvaptan in patients with heart failure:..., Xiong [/bib_ref] while in patients with hyponatraemia and ascites, short term tolvaptan use has been shown to normalise plasma sodium in 27-50% of patients. [bib_ref] Therapy of hyponatremia in cirrhosis with a vasopressin receptor antagonist: a randomized..., Gerbes [/bib_ref] The routine use of tolvaptan is not recommended in patients with cirrhosis, based on the liver injury seen in higher doses used in the TEMPO trial. [bib_ref] Tolvaptan in patients with autosomal dominant polycystic kidney disease, Torres [/bib_ref] Treatment of hypovolaemic hyponatraemia Volume replacement with isotonic saline, and specific management of the underlying cause, form the approach to management of hypovolaemic hyponatraemia. Correction of hypovolaemia eliminates the stimulus for AVP secretion resulting in a rapid aquaresis which can potentially lead to overcorrection, and thus careful monitoring is required. Treatment of thiazide diuretic induced hyponatraemia requires particular caution as withdrawal of the drug combined with correction of hypovolaemia may result in a rapid aquaresis
# Conclusion
Acute symptomatic hyponatraemia is a medical emergency, and current practice guidelines have adapted to recommend the use of bolus hypertonic saline in this setting; however recent trial data have emphasised that caution must be taken to prevent overcorrection when the duration of hyponatraemia is unclear. Chronic asymptomatic hyponatraemia is traditionally thought of as clinically benign and is thus often underinvestigated and undertreated. Traditional treatments for SIAD have been limited to date by poor efficacy, side-effects, cost or lack of supportive randomised control trial data. The past two years have seen the publication of several much-needed prospective randomised controlled trials that have demonstrated modest effects of fluid restriction in patients with chronic SIAD, albeit with good tolerability and safety, thereby emphasising the need for second-line therapies. Cost-reduction, more widespread reimbursement and use of a lower starting dose may allow expansion of the use of tolvaptan in treatment of SIAD. Recent studies have also given cause for optimism for potential future treatments such as empagliflozin; an RCT examining use of the drug in chronic euvolemic and hypervolemic hyponatraemia is underway.
Treatment of hyponatraemia must be individualised to the patient, taking into account the acuity and cause of hyponatraemia, the indications for treatment, and treatment goals. For example, it is reasonable to consider an initial trial of fluid restriction in an asymptomatic patient with incidentally noted chronic SIAD. On the other hand, correction of hyponatraemia may be more urgent in patients awaiting chemotherapy for example, and in this scenario early consideration of a low dose of dose of tolvaptan (7.5 mg) is appropriate.
It is becoming increasingly clear that treatment of chronic hyponatraemia is associated with reduction in length of hospital stay, improvements in gait and mentation, and a reduction in mortality. Mortality rates associated with hyponatraemia have been shown to differ according to volume status, and this should be considered when designing and interpreting future outcome studies. Prospective studies, employing effective treatments that will increase plasma sodium concentration by a clinically significant amount, in a large enough cohort to classify patients by volume status, are required to confirm the long-term clinical benefits of chronic hyponatraemia treatment.
[fig] Figure 1: General approach to treatment of hyponatraemia. [/fig]
[fig] Figure 2: Change in plasma sodium concentration from baseline in patients with SIAD treated with bolus 3% saline (red) and slow infusion of 3% saline (blue). Data expressed as median and IQR. Asterisks indicate a p-value of less than 0.05. pNa; plasma sodium concentration, h; hours. With permission, from Garrahy et al. (Garrahy, Dineen and Hannon, 2019).8 journals.sagepub.com/home/tae [/fig]
[table] Table 1: Classification of hyponatraemia based on volume status and urinary sodium concentration. [/table]
[table] Table 2: Diagnostic criteria for SIAD. [/table]
[table] Table 3: Targets for elevation in plasma sodium concentration recommended to avoid osmotic demyelination in chronic hyponatraemia.13 [/table]
[table] Table 4: Comparison of treatment options for chronic SIAD. [/table]
[table] Table 5: Summary of recent randomised controlled trials investigating treatments for SIAD. [/table]
[table] Table 6: Predictors of failure to respond to fluid restriction. [/table]
|
Dopaminergic and Cholinergic Modulation of Large Scale Networks in silico Using Snudda
Neuromodulation is present throughout the nervous system and serves a critical role for circuit function and dynamics. The computational investigations of neuromodulation in large scale networks require supportive software platforms. Snudda is a software for the creation and simulation of large scale networks of detailed microcircuits consisting of multicompartmental neuron models. We have developed an extension to Snudda to incorporate neuromodulation in large scale simulations. The extended Snudda framework implements neuromodulation at the level of single cells incorporated into large-scale microcircuits. We also developed Neuromodcell, a software for optimizing neuromodulation in detailed multicompartmental neuron models. The software adds parameters within the models modulating the conductances of ion channels and ionotropic receptors. Bath application of neuromodulators is simulated and models which reproduce the experimentally measured effects are selected. In Snudda, we developed an extension to accommodate large scale simulations of neuromodulation. The simulator has two modes of simulation -denoted replay and adaptive. In the replay mode, transient levels of neuromodulators can be defined as a time-varying function which modulates the receptors and ion channels within the network in a cell-type specific manner. In the adaptive mode, spiking neuromodulatory neurons are connected via integrative modulating mechanisms to ion channels and receptors. Both modes of simulating neuromodulation allow for simultaneous modulation by several neuromodulators that can interact dynamically with each other. Here, we used the Neuromodcell software to simulate dopaminergic and muscarinic modulation of neurons from the striatum. We also demonstrate how to simulate different neuromodulatory states with dopamine and acetylcholine using Snudda. All software is freely available on Github, including tutorials on Neuromodcell and Snudda-neuromodulation.
# Introduction
The nervous system depends on fast interaction via ionotropic receptors, which are activated by a variety of neurotransmitters including glutamate, GABA, and glycine. Besides these ionotropic receptors, neuromodulation via metabotropic receptors have a profound effect on network dynamics via slower processes. On a single cell level, these neuromodulators act via a variety of receptor subtypes, which modulate the excitability of neurons and influence their synaptic properties. The interactions between neuromodulators and their targets are complex. For example, a single ion channel type can be modulated by several different neuromodulators. While a single neuromodulator can affect multiple ion channel types and signaling pathways within neurons. On a circuit level, neuromodulation of neurons and synapses can massively alter the network activity. A challenge is to bridge the levels between single neurons and the circuit to understand the role of neuromodulators in shaping network behavior.
There are models of both invertebrate and different types of vertebrate which attempt to bridge the gap from cellular and synaptic properties to circuit function, and where the effects of neuromodulation is considered. For neuromodulation, the challenge is that the effect of a single neuromodulator in a single cell setting could be very different once that cell is embedded within a circuit. In the circuit, other components, such as other neurons and synapses, would also be under neuromodulatory control and hence contribute to more complex circuit interactions. Therefore, computational models have to incorporate the ability to bridge these levels. Furthermore, it has been shown that different parts of a neuron (dendrites/soma) can be modulated differently depending on receptor sub-type and target. Such compartmental differences can be implemented in multicompartmental models where reconstructed neuronal morphologies are used and ion channel models are distributed throughout the morphology.
We use the in silico striatal microcircuitas a demonstrative example, which consists of 95% striatal projection neurons (SPNs) and 5% interneurons (fast-spiking (FS), low-threshold spiking and cholinergic interneurons). Here, we simulate neuromodulation of SPNs and FS. The striatum is the input stage of the basal ganglia, a group of subcortical nuclei involved in action-selection, motor control, and habitual learning. Neuromodulation is important in the striatum and especially dopamine is essential for basal ganglia function. The cholinergic interneuron (ChIN) is a spontaneously active interneuron within the striatum, which releases acetycholine (ACh). Hence, in addition to the dopaminergic modulation, cholinergic modulation can modulate several components of the striatal network, which has extensive expression of muscarinic receptors. Furthermore, several studies have demonstrated the complex interactions between DA and ACh within the striatum, ranging from modulating dopamine release via presynaptic nicotinic receptors to direct dopaminergic modulation of ChINs via D2 receptors.
Previously, Snudda, a Python package for creating datadriven networks of neurons, placing synapses using touch detection between axons and dendrites and setting up large scale simulations was developed by. The software included fast synaptic transmission but neuromodulation was limited to dopamine. Hence, the Snudda package required further development to accommodate the generalized implementation of neuromodulation, together with all the other functionalities contained within Snudda. Following the development of Snudda.neuromodulation, it was necessary to create a software for generating and selecting parameter sets which reproduce neuromodulation. Therefore, we decided to develop Neuromodcell, which extracts the modulatory parameters necessary to reproduce neuromodulation on a single cell level for multicompartmental neuron models .
In this study, we applied Neuromodcell and Snudda.neuromodulation to the dopaminergic and cholinergic modulation of the in silico striatal microcircuit to demonstrate the functionalities included in these software packages. In general, Neuromodcell and Snudda can be used to generate and simulate networks of multicompartmental models with a wide range of neuromodulators. Within these simulations, a variety of questions could be addressed. Hence, not merely to simulate neuromodulation, but to predict which components are important for certain features and also redundancies within the network. Furthermore, simulations of multiple neuromodulators could predict how these interact on a network level. Both Neuromodcell and Snudda are open-source and freely available on Github, for further enhancement and expansion.
# Materials and methods
## Software setup
Neuromodcell is freely available to be downloaded from its Github site 1 , written in Python with requirements specified in requirements.txt including NEURON 2 and Jupyter Notebook. To install Neuromodcell use pip install neuromodcell. The software is compatible with Linux operating system and supercomputer clusters (Cray XC40 system). The simulations of large networks use Snudda, a Python package for creating data-driven networks of neurons available from its Github site 3 . To simulate neuromodulation, within the Snudda framework, additional simulation classes and associated neuromodulation subpackages were created, snudda.neuromodulation. To use Snudda, install via pip install snudda and follow the instructions on its wiki page (see text footnote 3).
## Multicompartmental models
Neuromodcell and Snudda require multicompartmental models specified for NEURON simulator described by three files: a morphology file (SWC), a parameter file (JSON) and a mechanism file (JSON) including a directory with the ion channel model files (.mod files, NEURON model description language). The models used in this publication were optimized FIGURE 1 | Neuromodcell structure with classes and methods. Neuromodcell specifies the model and optimization parameters using the DefineModulation class. The associated methods define the protocols, parameters, modulation, and selection criteria for the optimization. Following the simulation, the OptimisationResult class assists in loading and analyzing the results and saving the final modulation (modulation.json).
using BluePyOptin previous publicationswith modification for muscarinic modulation; and include direct and indirect striatal projection neuron (dSPN and iSPN) and fast-spiking interneuron (FS) models from the striatum.
## Parameterization
To modulate the models during the simulation, we introduce additional parameters for the specific neuromodulator within the .mod files. The modulation is implemented in a phenomenological manner, where a modulation parameter is multiplied with the target (for example the conductance of an ion channel). Hence, the conductance (in this case) can be regulated through the simulation. For example, for dopamine, the three parameters are maxModDA, modDA and levelDA (using the key, DA to indicate dopamine), based on previous formalism developed in. In general, the convention is 'maxMod * ' , 'level * ' and 'mod * ' and the * should be replaced with a specific name for the neuromodulator. The maxMod * defines the modulation degree (which can vary, with 0.6 being 40% reduction). The level * defines the transient level of modulation throughout the simulation. The mod * defines the activation of modulation (0 inactive, 1 fully active). Within the examples, dopaminergic and muscarinic modulation was included (with muscarinic modulation using key ACh). The list of modulated ion channels for each cell type is presented in Supplementary
## Neuromodcell
Neuromodcell is a software for optimizing and simulating neuromodulation in detailed multicompartmental neuron models. The software loads and simulates a population of models with user-defined parameter variations of maxMod * , which is how the program induces the changes associated with neuromodulation. The software also incorporates current clamp and voltage clamp parameters to simulate specific biological experiments. Hence, the user can define the simulated experiment in terms of both neuromodulator and the specific protocol to be used. Following the simulation, the parameter sets that reproduce the experimental data are selected based on a user-specified selection criteria. We provide several electrophysiological features by default, but custom features can be included.
Here, Neuromodcell was used to optimize dopaminergic modulation for dSPN, iSPN and FS, as well as muscarinic modulation for dSPN and iSPN. The optimizations for dSPN and iSPN were validated against data fromand Lahiri and Bevan (2020) (see for dopaminergic modulation andfor muscarinic modulation. The dopaminergic modulation for FS was validated against data from.
## Neuromodulation in snudda
Neuromodulation in Snudda is implemented as a separate simulation module called snudda.neuromodulation. Within the module, there are two simulation classes for neuromodulation, SnuddaSimulateNeuromodulation and SnuddaSimulateNeuromodulationSynapse.
These classes represent the two strategies used to simulate neuromodulation, replay and adaptive. The replay mode plays a predetermined transient through the level * -parameters and hence modulates each component of the circuit in a predefined time dependent manner. On the other hand, the adaptive mode originates from the fact that a neuromodulatory neuron integrated within the microcircuit could not modulate via a predetermined transient. Instead, the spiking activity of such a neuron would have to be continuously translated into an instantaneous level of neuromodulation. Therefore, an alternative approach was developed where an intermediate mechanism would integrate the spiking activity. Adaptive simulations couple one or several presynaptic neurons to postsynaptic neurons via an integrating mechanism called conc * , which modifies the neuromodulatory parameters (level * ) in the circuit (for example concDA to levelDA for dopamine). Hence, neuromodulation within the simulation can either follow a fixed recipe ('replay' mode), or dynamically change based on the activity in the network ('adaptive' mode). The mode used within a particular simulation will therefore depend on the network structure. Using either approach, the level of neuromodulation would be updated every time step via the level * -parameter. The replay mode loads an array which has set values for the level * for each time step of the simulation. This would suffice if the neuromodulation is synchronous and the neuromodulatory neurons are not affected by activity of the circuit. On the other hand, if the activity of the microcircuit could for example inhibit the neuromodulatory neurons, like striosomal dSPNs in the striatum, replay mode does not suffice and adaptive mode would be used instead, as the presynaptic neuromodulatory neurons can interact with the microcircuit throughout the simulation.
For replay, the maxMod * , mod * and level * are introduced as RANGE variables. In the adaptive mode, which is based on pointers, the parameters are introduced with the keyword POINTERS 4 and 'ptr' is added to the filename (i.e., na.mod to na_ptr.mod).
Lastly, both types of simulations require a configuration file (JSON) which is created using either Neuromodulation class (snudda.neuromodulation.modulation_network, for replay simulations) or NeuromodulationSynapse (snudda.neuromodulation.modulation_synapse, for adaptive simulations). For examples of how to use these classes (see text footnote 4).
To demonstrate the features of Snudda.neuromodulation, we simulated several networks using both replay and adaptive mode. Firstly, using the Snudda framework, a network of 10 000 neurons was created and simulated with dopamine and acetylcholine transients; and with a cortical activation at 1 s, with cortical and thalamic background activity. The network is based on previous 4 github.com/jofrony/Neuromodulation-software publicationsand the connection probabilities follows the diagram in . The transients consisted of a tonic level of ACh accompanied with a burst or pause of ACh and a DA burst. The simulations were performed on a super-computer cluster (Cray XC40 system at PDC Center for High Performance Computing, KTH). A smaller network with dSPNs was simulated to show the effect of DA and ACh without network interactions from iSPNs and FS. Furthermore, several dopamine transients with different start times were created using the Neuromodulation class and applied to a network of 20 dSPNs with a cortical activation lasting for 500 ms. Lastly, a network of dSPNs was simulated with presynaptic spiking neurons using the adaptive mode. The presynaptic neurons represented dopaminergic neurons and were simulated with and without activation (i.e., bursting). The activity of the neurons within the simulations were measured by the percentage of spiking neurons and the mean firing frequency during the cortical activations using custom Python code and Electrophysiology Analysis Toolkit 5 .
## Tutorials
A tutorial on Neuromodcell and an example of dopaminergic modulation of dSPN are available (see text footnote 1).
Examples of simulations using Snudda.neuromodulation module are available (see text footnote 4).
# Results
## Overview
Neuromodcell performs parameter variation of ion channels and receptor models within multicompartmental models to simulate neuromodulation. The components of the software are presented in with the main class, DefineModulation and its associated class methods, which are used to define the parameters, protocols, selection criteria and modulation for the specific cell and neuromodulator. The user specifies the range of parameter variation for each ion channel and/or receptor as well as the experimental protocol to be replicated in the simulation . The models with parameter variations are simulated using the optimise module, which uses mpi4py for parallelization.
Following the simulation, a selection process is performed, where the models are compared to the experimental data provided in the setup (using DefineModulation). The parameter variations of the models which pass validation are saved in a separate file. The code for defining the selection criteria is freely available (module selection_criteria and selection_function) and can be modified to measure a specific electrophysiological feature of the simulation, for example firing frequency or the number of action potentials.
Below, we describe the process of setting up, optimizing and selecting the models that pass validation. Following the completion of the optimization, the results are transferred to FIGURE 2 | Code for setting up the parameters for neuromodulation using Neuromodcell. The parameter, "path_to_model," defines the path to the model and which specific parameterID, parameter_id. mod_set.define_neuromodulation, the naming defines the neuromodulation key, DA for dopamine. The example has a simulation time of 1000 ms and population of 1000 different models.
FIGURE 3 | Code for setting up the parameter variation for neuromodulation. modset.define_modulation_parameter accepts the ion channel name, neuromodulator, part of the morphology and the intervals of modulation. Here we have used data from previous publications.
Snudda and simulated using snudda.neuromodulation module. A step-by-step tutorial is available (see text footnote 4).
## Defining neuromodulation in silico
We performed optimizations of dopaminergic and muscarinic modulation for dSPN, iSPN, and FS. In the following examples, the dopaminergic optimization of dSPNs is used to demonstrate the features of Neuromodcell. We instantiated the DefineModulation class as shown in . The class requires several arguments to define the optimization, including the FIGURE 4 | Code for setting up the selection criteria. Dopamine level is simulated as a bath application. Two simulated current clamps are added to the set up to replicate the experiment from. The effect of dopamine is measured by the number of action potentials and compared with experimental data. model directory and the number of model variations to simulate (population). The next step is parameter variation. In , the ion channels which are modulated by DA within dSPN (based on previous publicationswere set using the define_modulation_parameter method within the class. The modulation can also be defined to be somatic, dendritic (basal or apical, due to the SWC naming convention) or axonal.
## Experimental setup and selection criteria
The experimental data used in the example optimization was taken from infor dSPNs. The experiment used patch clamp recordings and bath application of dopamine to measure the effect of DA on dSPNs. Following the application of DA, there was an increase in the intrinsic excitability of dSPNs measured by the number of action potentials. Hence, an experimental setup was introduced into modset, with a current clamp protocol and bath application of dopamine . Furthermore, the method define_selection_criteria was used to introduce the measurement used in the validation step (i.e., the change in the number of action potentials). Both the define_modulation_function and define_selection_criteria can select several pre-defined functions but they can also be customized to accommodate specific transients or criteria, respectively.
## Optimization
The optimization uses the OptimiseModulation, from optimise module in Neuromodcell. The optimization for dSPN used a custom Python script called optimise_dspn.py, which loaded the path_to_model and the specific seed for the optimization . The seed is used to randomize the modulation parameters defined using define_modulation_parameter. The modulation parameters sets (i.e., population) are simulated in parallel with the specific model (defined in DefineModulation), as illustrated in .
# Analysis
The result of the optimization was visualized using dSPNanalysis (inheritance from OptimiseAnalysis class in Neuromodcell) . The class loads the output files following the optimization and enables the user to plot and/or analyze the results. In, the result of dopaminergic modulation of dSPN is shown with the model variations (modulated) and control simulation (in black). The dopaminergic modulation increases the intrinsic excitability of dSPNs via D1 receptors. The result of the Neuromodcell optimization produced model variations which reproduce the experimental data from. The control simulation (in black,has four spikes following current injection. The simulated dopamine modulation increases the number of spikes and the change in the number of action potentials are within the range extracted from the experimental data.
## Transfer files to snudda simulation
Following the optimization, the parameter sets which reproduce the neuromodulation are saved as modulation.json and moved to the model directory (i.e., the model is now defined by parameters.json, morphology.swc, mechanisms.json and modulation.json). The modified .mod files should also be transferred to a common Snudda mechanism directory.
## Dopamine
Using the replay mode in Snudda.neuromodulation, we simulated a network of 10 000multicompartmental neuron FIGURE 5 | Code for running the optimization.
FIGURE 6 | Code for analyzing the results from the optimization using dSPNanalysis. dSPNanalysis (inheritance from OptimiseAnalysis class) loads the results from the optimization. The results can be plotted by custom methods and the final modulation is saved. models of dSPNs, iSPNs, and FS to exemplify the effect of the dopamine modulation. The simulation included a cortical stimulation of dSPNs, iSPNs and FS at 1 s, lasting for 0.5 s. In, the effect of dopamine on network activity is shown. Following cortical activation, 6% of dSPNs within the network were spiking. By modulating the ion channels defined in the optimization result of Neuromodcell, dopamine modulation increases the percentage of spiking dSPNs to the double. iSPNs are modulated by D2 receptors, which reduces the intrinsic excitability. Hence, in, the percentage of spiking iSPNs within the network decreases by approximately 3%. The middle panel inshows an example trace of an iSPN, where the dopamine modulation reduces the number of spikes. FS are modulated by D1-like receptors, which depolarizes the membrane potential, as seen in. Following, an optimization using Neuromodcell, the parameter sets for FS dopamine modulation increase the number of spiking neurons within the network. The cortical activation occurs via activating glutamate ionotropic receptor models. These receptors can also be modulated by the same formalism as previously described which leads to changes in the amplitude and release probability .
## Dopamine and acetylcholine
The neuromodulation extension to Snudda can also simulate multiple neuromodulators. To demonstrate this, we simulated five scenarios of dopamine and acetylcholine modulation with the same network presented in. As previously described, dSPNs and FS are modulated by D1 receptors while iSPNs are modulated by D2 receptors. In addition, we used Neuromodcell to optimize for muscarinic modulation for dSPNs and iSPNs.
In, we show the transients used in the simulations, which consisted of an acetylcholine burst or pause and a dopamine burst. In the striatum, acetylcholine is released continuously by cholinergic interneurons due to their spontaneous activity. Hence, the acetylcholine transients had a tonic modulation level of 0.5 with burst and pause response causing an increase or decrease (by 0.5), respectively. These transients were simulated individually but also simultaneously (DA and ACh burst; DA and ACh pause). As shown in, dopamine increased the percentage of spiking dSPNs within the network. The acetylcholine burst also slightly increases the percentage of spiking dSPNs, although in the simultaneous simulation, light purple), the effect is not additive. Instead, the dopamine burst and acetylcholine pause produces a larger response. This is contrary to previous results in, where DA and ACh had additive effects when single neurons were simulated. Hence, we performed a simulation with only dSPNs to investigate the effect of DA and ACh. Without network interactions, the effect of DA and ACh was indeed additive . Hence, the decrease in the percentage of spiking dSPNs could be attributed to the network interactions and the changes seen in iSPNs and FS. As described in, the intrinsic excitability of iSPNs is decreased following dopamine modulation, while ACh modulates iSPNs via M1 receptor, which lead to the opposite effect. Compared to dSPNs, iSPNs have a larger increase in the number of spiking neurons following an ACh burst. During the ACh pause, the percentage of spiking iSPNs is expected to decreased, due to the reduction in ACh levels. In contrast, the percentage of spiking iSPNs was larger than in the control. In, SPNs were recorded in vivo and cholinergic interneurons were inhibited optogenetically. They observed a reduction in the spontaneous activity of SPNs. The control simulation inwas, however, run without any neuromodulation, which could explain the inconsistency. Hence, we simulated a small network of SPNs with a cholinergic pause, but we modified the control simulation to contain a continuous tonic level of ACh. By comparing the mean firing frequency of SPNs with and without the pause, we observed the expected reduction in SPN activity . Therefore, the tonic level of ACh has an influence on the activity of SPNs. For FS, the main effect is seen within the dopamine burst modulation, while for ACh pause and burst there is no visible effect (due to the lack of direct muscarinic modulation of FS).
## The timing of dopamine
The user can construct different transients to investigate the effect of timing on network activity. We created a smaller network of 30 dSPNs to illustrate the effect such an investigation could have on network activity. We defined six transients which were identical, in terms of time constants and modulation level. We then adjusted the start time in relation to cortical stimulation at 1s (of a 3 s simulation). The different delays caused the peak modulation to occur at different times during the cortical stimulation.shows that the transient −100 ms results in the largest increase in mean firing frequency. While −/+500 ms transients are not different from control. This is due to the dopamine level being too low and the peak dopamine response occurring much earlier and later compared to the cortical stimulation. The +100 ms delay produced a lower increase in the mean firing frequency but the percentage of spiking neurons were similar to the −100 ms case. The user could conduct similar investigations into other features of the transients such as the time constants, tonic levels or more complex transients.
## Adaptive modulation
The previous simulations have used a transient(s) to modulate the network. In , we simulate neuromodulation using the adaptive mode in Snudda. In adaptive mode, the spiking activity of the presynaptic neurons are translated to modulation levels as described in . The top panel in shows the spike times of two presynaptic neurons (green and red). The mean frequency of dSPNs following the different DA delays. In control (without dopamine modulation), the dSPNs within the network do not spike following cortical activation. Dopamine modulation starts at 500 ms before cortical input (-500 ms), does not produce enough modulation to induce spiking. While -300 ms, increases the mean firing frequency and the following delays except 500 ms. (C) The percentage of spiking neurons within the network at different dopamine delays. At 300 ms, the percentage of spiking cells was 7%. -100 and 100 ms produced a similar effect on the number of spiking neurons within the network.
FIGURE 11 | Simulating presynaptic spiking neurons and dopaminergic neuromodulation. In (A), a schematic representation of the adaptive mode, concDA (orange) placed on a segment within the multicompartmental model, which receive input from presynaptic cells (red and green). Every spike input triggers a transient inside the concDA which is sent to the ion channel (green) on that particular segment, rightmost panel in (A) hence modulating the current. In (B), a simulation of a network of dSPNs, where neuromodulation depends on presynaptic spontaneously active neurons. Compared to control (without modulation) (black), the simulation with the presynaptic neurons produces a depolarization in dSPNs (red). If the frequency of the presynaptic neuron increases (C), this translates to an increase in the neuromodulation. The dSPN depolarises more than in the spontaneous case and this results in an increase in the percentage of spiking neurons. In (D), the bar chart shows the comparison between control and spontaneous versus bursting presynaptic cell (and control for bursting simulation, where the connections between the cells were disabled). The percentage of spiking cells within the network increases following the dopamine modulation which occurs when the presynaptic cell firing frequency increased. -80 mV marked by yellow line.
These neurons are connected to an integrating mechanisms called concDA (for dopamine, in orange in . The concDA integrates the spiking activity of the two cells into modulation levels (as seen in in left most box). The concDA sends the modulation level from both neurons into the ion channel model (right most box in . The level of current follows the modulation as the conductance of the ion channel is modulated and hence affecting the neuronal activity.
We created a small network of dSPNs to illustrate the effect of changing the firing frequency of the presynaptic neurons. We placed spontaneously active neurons to modulate the dSPNs via D1 receptors. In control, 50% of the dSPNs within the network were spiking and the spontaneous activity of the presynaptic neurons did not change the percentage . Although in , the spontaneous activity has an effect on the membrane potential of the dSPNs. By stimulation the presynaptic spontaneous neurons, we increased the firing frequency . As expected the level of neuromodulation increased as the concDA mechanism integrates the number of spikes into neuromodulation. This leads to a depolarization of the dSPN membrane potential. On a network scale, the increased firing frequency leads to simulated D1 receptor modulation and an increase in the percentage of spiking dSPNs.
# Discussion
In this article, we present a framework for simulating neuromodulation in detailed multicompartmental neuronal models being part of large-scale microcircuits, using Neuromodcell and Snudda. Our priority was to create an extension to Snudda, a software for large scale simulation of networks of multicompartmental models, which also simulates transients of neuromodulator(s). Firstly, we developed Neuromodcell which enables the user to optimize and select parameters for a specific type of neuromodulator. The neuromodulation is defined by introducing parameters within the ion channel and receptors models, hence enabling dynamic control of for example conductance. Secondly, we developed an extension to Snudda, which incorporates the neuromodulation parameters found using Neuromodcell. Thirdly, we simulated a range of different neuromodulatory states with the striatal microcircuit to show the versatility of the tool.
We simulated an example network of dSPNs, iSPNs, and FS. These neurons are modulated by dopamine D1 and D2 receptors. We used Neuromodcell to optimize for parameters sets which reproduced the available experimental data on dopaminergic modulation of these cell types. The simulations of dopaminergic modulation showed that DA can modulate the excitability of the network. Hence, the simulations of dopamine are in line with known effects of DA in the striatum. Currently, the Neuromodcell optimization produces a population of modulated models, which are selected according to predefined parameters. In the future, the Neuromodcell optimization could be improved and utilize more elaborate optimization algorithms, for example the genetic algorithm used within BluePyOpt software .
Within the striatum, cholinergic interneurons release acetylcholine which modulates the network via nicotinic and muscarinic receptors. Both dSPNs and iSPNs are modulated by the M1 receptor, while only the M4 receptor modulates dSPNs. Using Snudda.neuromodulation, we simulated muscarinic modulation throughout the whole network. We defined a burst and pause transient with a tonic background level of muscarinic modulation to replicate the acetylcholine levels reported in the striatum. We could show that this changes the excitability of dSPNs and iSPNs which is consistent with reports. Furthermore, we could replicate the effect of ACh pause on SPNs reported by.
Within the striatum, there are several in silico investigations which can be performed using Snudda.neuromodulation. In Parkinson's disease (PD), the degeneration of dopaminergic neurons results in a change in the balance between dopamine and acetylcholineand anti-cholinergic drugs were used to treat PD before DOPA therapy emerged. Although, the interaction between dopamine and acetylcholine is not fully understood. Therefore, future simulations could investigate how the changes of DA and ACh on the single cell level affect network activity; and potentially dissect the important components using Neuromodcell and Snudda. Furthermore, in, they showed that cholinergic activity and DA levels were coordinated during spontaneous movement. Using Snudda.neuromodulation such transients could be simulated to understand how the coordination between ACh and DA affects the activity of dSPNs and iSPNs.
Recent advances in biosensor technology are enabling research to visualize neuromodulator levels within neuronal networks. These biosensors can provide high spatial and temporal resolution on the action of neuromodulators like dopamine, serotonin, and opioids amongst others. Hence, a simulation platform like Snudda, can incorporate such data to understand how the underlying neural circuit responds to these transients.
Neuromodulation affects many aspects of neurons and neural circuits. Presently, Neuromodcell and Snudda focused on the effect on ion channels and receptors. Hence, we have included a limited part of the effects that neuromodulators have on network activity. Many of the targets of neuromodulators are not incorporated into multicompartmental models, including transcription factors and other subcellular processes. Simulations of such effects would require other types of models and simulators. Although, a possible future development would be to include synaptic plasticity within the large scale simulations. Currently, Snudda includes receptor models of glutamate (NMDA/AMPA) and GABA receptors with short-term plasticity. In several brain areas, neuromodulators can regulate the long-term potentiation (LTP) and long-term depression (LTD). Hence, synaptic plasticity could be coupled to neuromodulation levels during the simulation and simulate the effect of LTP vs. LTD, although currently this is not included in Snudda.
Snudda is a general tool for simulating large scale networks of neurons in any part of the nervous system with NEURON simulator. Our aim was to develop an extension to Snudda, which simulates neuromodulation on a large scale and can include any neuromodulator. We then developed Neuromodcell, which provides a tool for investigating neuromodulation at single cell level, in addition to the integration into the Snudda framework. These tools can provide a link between the single cell experiments and circuit level experiments of neuromodulation, which is currently not possible with the available simulation tools.
# Data availability statement
The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found below: The code for single neuron optimizations available at github.com/jofrony/Neuromodcell.git and Snudda.neuro modulation is available at github.com/Hjorthmedh/Snudda.git.
# Author contributions
JF designed Neuromodcell and generalization of neuromodulation in Snudda and wrote the manuscript with some input from the co-authors. JHj, SG, and JH contributed throughout the study by discussing the results. All authors contributed to the article and approved the submitted version. |
MicroRNA and Male Infertility: A Potential for Diagnosis
MicroRNAs (miRNAs) are small non-coding single stranded RNA molecules that are physiologically produced in eukaryotic cells to regulate or mostly down-regulate genes by pairing with their complementary base-sequence in related mRNA molecules in the cytoplasm. It has been reported that other than its function in many physiological cell processes, dysregulation of miRNAs plays a role in the development of many diseases. In this short review, the association between miRNAs and some male reproductive disorders is surveyed. Male factor Infertility is a devastating problem from which a notable percentage of couples suffer. However, the molecular mechanism of many infertility disorders has not been clearly elucidated. Since miRNAs have an important role in numerous biological cell processes and cellular dysfunctions, it is of interest to review the related literature on the role of miRNAs in the male reproductive organs. Aberrant expression of specific miRNAs is associated with certain male reproductive dysfunctions. For this reason, assessment of expression of such miRNAs may serve as a suitable molecular biomarker for diagnosis of those male infertility disorders. The presence of a single nucleotide polymorphism (SNP) at the miRNAs' binding site in its targeted mRNA has been reported to have an association with idiopathic male infertility. Also, a relation with male infertility has been shown with SNP in the genes of the factors necessary for miRNA biogenesis. Therefore, focusing on the role of miRNAs in male reproductive disorders can further elucidate the molecular mechanisms of male infertility and generate the potential for locating efficient biomarkers and therapeutic agents for these disorders.Citation:
# Introduction
MicroRNAs (miRNAs) are small non-coding single stranded RNA molecules of approximately 22 nucleotides that are physiologically produced in eukaryotic cells (1). After being expressed in the nucleus, miRNA plays a role in regulation of gene expression by pairing with its complementary base-sequence of the targeted mRNA molecule in the cytoplasm. This usually leads to gene silencing through degradation of targeted mRNA or interference with its translation (2, 3). After discovery of miRNAs, as single-stranded non-protein-coding regulatory RNA molecules in C. elegans by in 1993, over 1500 miRNAs have currently been reported to be encoded by the human genome (5-7) which may target around 60% of mammalian genes in various human cell types (8). [fig_ref] Fig 1: Biogenesis of miRNA [/fig_ref] schematically depicts miRNA biogenesis. Aside from the physiological regulation of gene expression in multiple biological processes such as cell cycle control and differentiation (9), cell growth and apoptosis (10) and embryo development (11), the role of miRNAs in development of many human diseases has been studied and reported. Mutation, dysfunction of biogenesis and dysregulation of miRNAs and their targets may lead to various diseases such as cardiovascular diseases (12), cancers (13), schizophrenia (14), renal function disorders (15), psoriasis (16), muscular disorders (17) and diabetes (18). A comprehensive database of whole miRNA-disease association, miRNA vation and miRNA function has been published focus on male infertility with the intent to conduct a literature review of the role of miRNAs in this field.
## Drosha (a class 2 ribonuclease iii) and its rna-binding partner, dgcr8*, cut the stem loop structure to make pre-mirna. exportin 5 transports the pre-mirna from the nucleus to cytoplasm. in the cytoplasm, dicer (a rnase iii endonuclease) and its rna-binding partner, trbp**, cleave pre-mirna to generate a short double-stranded rna molecule named mirna duplex. one strand of this molecule is incorporated into the mirna-induced silencing complex (mirisc) containig protein argonaute-2 (ago2) as a catalytic component
Infertility is a problem in 10-15% of couples worldwide . Approximately, 50% of the infertility cases are attributed to male factors and 60-5% of these cases are idiopathic (20). Of note, the molecular mechanisms of many male infertility disorders are not clear . The aim of this review is to pursue the relation between miRNAs and male infertility.
## Mirna and spermatogenesis
The presence of miRNA in testis has been verified a few years ago (21). Later, researchers focused on finding the function of miRNAs in germinal tract physiology. [fig_ref] Table 1: Discoveries related to microRNAs [/fig_ref] shows a brief description of the chronological order of discoveries related to miRNAs in testis molecular physiology [for additional information please
## Potential of extracellular mirnas as biomarkers for infertility
In addition to the presence of miRNA at the cellular level, miRNAs have also been reported to (34), saliva (35, 36), vaginal secretions, menstrual blood and semen (36). Aberrant expression of extracellular miRNA is attributed to different disorders (37). Altered expression of miRNAs, as with (38-41) has been recently reported in male infertility disorders, hence the possibility of locating potential miRNA biomarkers in male infertility has been proposed.
pression in the testis of patients with non-obstructive azoospermia (NOA) has been reported by . Of note, that NOA which has a testis-origin, presents a diverse range of defects from hypospermatogenesis and sperm maturation arrest to Sertoli-cell-only-syndrome. Hence, assessment of differential expressions of miRNAs in these infertile individuals is a prerequisite.
Wang et al. (43) examined pooled semen samples obtained from infertile men and compared the results with normal fertile individuals as controls. They found alterations in miRNA profiles by Solexa Sequencing (a sequencing method based on reversible dye-terminators technology and engineered polymerases) in both azoospermia and asthenozoospermia. These authors considered a cut of value of 50-fold higher or lower expression as significant, which was further validated by real time RT-qPCR for 7 miRNAs (miR-34c-5p, miR-122, miR-146b-5p, miR-181a, miR-374b, miR-509 -5p, and miR-513a-5p). The level of these 7 miRNAs was significantly lower in azoospermia and higher in asthenozoospermia compared to the control. Another noticeable finding of these authors was the stability of the above miRNAs to different conditions. They attributed this phenomenon to the small size of the miRNAs and their ability to bind with complex organic molecules in cellfree semen samples. Finally, they proposed that these 7 miRNAs might have confirmative molecular diagnostic value for male infertility.
In this regard, Wu et al. (44) evaluated the pattern of expression related to miR-I9b and let-7a in idiopathic infertile individuals with NOA or oligozoospermia by quantitative RT-PCR. They showed that these two miRNAs distinctively expressed at higher levels in infertile cases compared with fertile individuals. Consequently, they concluded that miR-I9b and let-7a were good diagnostic molecular biomarkers for idiopathic infertile cases with NOA or oligozoospermia.
## Single nucleotide polymorphism (snp) and function of mirna in spermatogenesis
Considering the diverse role of miRNA in spermatogenesis, therefore any polymorphism in related genes might lead to infertility. Zhang et al.
sociation between miRNA-binding site single nucleotide polymorphism (SNP) and idiopathic male infertility. They systematically surveyed all SNPs in the 3´ UTR of 140 mammal spermatogenesisrelated genes and observed that some SNPs within miRNA-binding sites might be related to idiopathic male infertility.
Among 140 surveyed genes, 6 SNPs (two in CYP19, Serpina5 and four in CGA, CPEB1, and CPEB2), involved in down regulation of gene expression in spermatogenesis and meiosis and were predicted to have possibility for altering the binding affinity of miRNA by using bioinformatically specialized algorithms (Pictar, miRanda, Targetscan, and RNAhybrid). The effectiveness of these SNPs in male infertility was analyzed in a case-control study using PCR followed by restriction digestion of the amplified fragments for genotyping. They realized that T substituted for A in rs6631 which lead to diminished binding ability of miR-1302 to its binding site in CGA in vitro, with a subsequent overexpression of CGA. Therefore an association between this special SNP and male infertility was concluded.
Recently evaluated the rela-(rs10719, rs2291109, rs17409893 and rs642321) and infertility. Performing genotyping by means of real time PCR, these researchers showed that SNPs in rs10719, rs12323635 and rs642321 were related with male infertility in the examined Han-Chinese population.
# Conclusion
Considering limited number of studies performed in this field, little has been clarified about miRNA-mediated gene regulation in spermatogenesis. In addition, there are few numbers of published studies that focus on the role of miRNAs in male infertility. This review has aimed to elaborate the role of miRNA in the male reproductive system. Any disorder or failure in this system can result in male infertility. Therefore, a dysfunction in miRNA processing as and infertility. Other anomalies such as dysregulation in expression of certain miRNAs or SNP in the miRNA binding site or SNP in the genes involved in biogenesis of miRNA can be related to infertility. Hence, in this regard, scientists suggest that assessment of miRNA may have future diagnostic value and shed more light on the molecular mechanisms of male infertility. Thus, new paths can be opened for future treatments of male infertility or even in the design of new contraceptive drugs.
[fig] Fig 1: Biogenesis of miRNA: RNA polymerase II transcribes miRNA gene to produce Pri-miRNA having a stem loop structure. [/fig]
[table] Table 1: Discoveries related to microRNAs (miRNAs) in the male reproductive system [/table]
|
Endothelial Protein C Receptor (EPCR), Protease Activated Receptor-1 (PAR-1) and Their Interplay in Cancer Growth and Metastatic Dissemination
Endothelial protein C receptor (EPCR) and protease activated receptor 1 (PAR-1) by themselves play important role in cancer growth and dissemination. Moreover, interactions between the two receptors are essential for tumor progression. EPCR is a cell surface transmembrane glycoprotein localized predominantly on endothelial cells (ECs). It is a vital component of the activated protein C (APC)-mediated anticoagulant and cytoprotective signaling cascade. PAR-1, which belongs to a family of G protein-coupled cell surface receptors, is also widely distributed on endothelial and blood cells, where it plays a critical role in hemostasis. Both EPCR and PAR-1, generally considered coagulation-related receptors, are implicated in carcinogenesis and dissemination of diverse tumor types, and their expression correlates with clinical outcome of cancer patients. Existing data explain some mechanisms by which EPCR/PAR-1 affects cancer growth and metastasis; however, the exact molecular basis of cancer invasion associated with the signaling is still obscure. Here, we discuss the role of EPCR and PAR-1 reciprocal interactions in cancer progression as well as potential therapeutic options targeted specifically to interact with EPCR/PAR-1-induced signaling in cancer patients.
# Introduction
Since the landmark publications by Bouillaud and Trousseau, presenting the relationship between coagulation and cancer, numerous studies have been performed to elucidate the exact mechanisms by which the procoagulant phenotype might promote the dissemination of malignant cells. It is now clear that the protease activated cell surface transmembrane receptors, endothelial protein C receptor (EPCR) and protease activated receptor 1 (PAR-1), apart from their roles in hemostasis and inflammation, influence cancer biology [bib_ref] ALEX1, a novel tumor suppressor gene, inhibits gastric cancer metastasis via the..., Pang [/bib_ref] [bib_ref] Endothelial cell protein C receptor-dependent signaling, Pendurthi [/bib_ref] [bib_ref] Endothelial cell protein C receptor: A multiliganded and multifunctional receptor, Rao [/bib_ref]. These receptors have been demonstrated to promote cancer invasion and metastasis in part by activating antiapoptotic pathways as well as to facilitate tumor cell migration, proliferation, angiogenesis and interactions with host vascular cells.
PAR-1, is a thrombin-activated receptor of a specific pathway of activation with proteolytic cleavage of the N-terminus. Interestingly, thrombin-mediated PAR-1 activation in ECs may be modulated by interactions with the EPCR, thereby inducing alternate cascades of intracellular events. That is to say, thrombin-activated protein C (PC) in complex with EPCR and co-factor thrombomodulin (TM), may activate PAR-1 localized on ECs to exert anticoagulant, anti-inflammatory, and crucially for cancer biology, cytoprotective effects. That is why the interactions between EPCR and PAR-1 have become the focus of research in terms of tumor progression. The APC/EPCR/PAR-1 pathway induces motility and proliferation of ECs as well as promotes angiogenesis via vascular-protective signaling and tube formation, thereby facilitating cancer dissemination [bib_ref] Activated protein C enhances cell motility of endothelial cells and MDA-MB-231 breast..., Gramling [/bib_ref]. The interaction axis of EPCR/thrombin/PAR-1 on endothelial cells represents an attractive target for inhibiting metastasis, as our previous studies with α-thrombin demonstrated that the enzyme increases adhesion of W256 carcinoma cells to rat aortic ECs and fibronectin by 50-300% [bib_ref] Thrombin enhances tumor cell adhesive and metastatic properties via increased alpha IIb..., Wojtukiewicz [/bib_ref]. Thrombin mediated EPCR/PAR-1 interactions are also observed in bone marrow where these proteins regulate repopulation, survival and chemoresistance of blood-forming progenitor cells by regulating NO production [bib_ref] PAR1 signaling regulates the retention and recruitment of EPCR-expressing bone marrow hematopoietic..., Gur-Cohen [/bib_ref] [bib_ref] Regulation of long-term repopulating hematopoietic stem cells by EPCR/PAR1 signaling, Gur-Cohen [/bib_ref]. Furthermore, EPCR and PAR-1 are co-expressed in gastric cancer (GC) cells, where EPCR exerts pro-carcinogenic effects by inducing PAR-1-dependent ERK1/2-MAPK pathway signaling to ultimately regulate proliferation and migration of tumor cells. Another hypothesis posits that differences in duration of PAR-1 signaling in endothelial cells may also influence signaling specificity of coagulation proteases [bib_ref] Activated protein C-cleaved protease activated receptor-1 is retained on the endothelial cell..., Schuepbach [/bib_ref].
On the other hand, there are data that APC-mediated EPCR activation triggers cytoprotective signaling cascades resulting in vascular barrier enhancement, reduced cancer cell extravasation and inhibition of metastasis [bib_ref] Endothelial cell protein C receptor opposes mesothelioma growth driven by tissue factor, Keshava [/bib_ref] [bib_ref] Role of activated protein C and its receptor in inhibition of tumor..., Bezuhly [/bib_ref]. The role of the APC/EPCR signaling pathway in limiting cancer cell metastasis may relate to the findings that vitamin K antagonists (VKA) were ineffective in cancer patients in the clinical setting [bib_ref] The effects of vitamin K-antagonists on survival of patients with malignancy: A..., Smorenburg [/bib_ref]. The conflicting data may result from the fact that in tumor cells expressing both receptors the final biological effects depends on reciprocal interaction influenced by a variety of ligands and cofactors like PC and its inhibitor (PCI), plasminogen activator inhibitor (PAI)-1, thrombin, thrombomodulin (TM), sphingosine 1 and 3 phosphate receptor S1P1, S1P3 and procoagulants, e.g., tissue factor (TF) [bib_ref] Protein C and its inhibitor in malignancy, Suzuki [/bib_ref]. Additionally, factor VIIa (FVIIa) modulates thrombin-mediated as well as EPCR-elicited PAR-1 activation in ECs in a likely complex with APC [bib_ref] Factor VIIa induces anti-inflammatory signaling via EPCR and PAR1, Kondreddy [/bib_ref]. There are recognized signaling pathways elicited by the APC/EPCR/PAR-1 axis that are independent of protease cleavage sites, which may reconcile some conflicting data in this area [bib_ref] Activated protein C promotes protease-activated receptor-1 cytoprotective signaling through β-arrestin and dishevelled-2..., Soh [/bib_ref].
The EPCR-dependent effects of APC as well as the role of PAR-1 in the biology of cancer have been separate topics of recent reviews [bib_ref] ALEX1, a novel tumor suppressor gene, inhibits gastric cancer metastasis via the..., Pang [/bib_ref] [bib_ref] Endothelial cell protein C receptor-dependent signaling, Pendurthi [/bib_ref] [bib_ref] Protease-activated receptors (PARs)-biology and role in cancer invasion and metastasis, Wojtukiewicz [/bib_ref]. Here, we present the newest knowledge regarding EPCR and PAR-1 reciprocal interactions in cancer progression as well as potential therapeutic options targeted specifically to inhibit or activate EPCR/PAR-1 -induced signaling in cancer patients.
## Epcr and cancer
EPCR is a type 1 transmembrane protein expressed prominently on the endothelium of the large vessels, but can also be observed on dendritic cells, leukocytes (lymphocytes, monocytes, neutrophils), epithelial cells, keratinocytes, pneumocytes, cardiomyocytes, chondrocytes and osteoblasts [bib_ref] Thrombomodulin-protein C-EPCR system: Integrated to regulate coagulation and inflammation, Van De Wouwer [/bib_ref]. Interestingly EPCR has been described on hematopoietic stem cells and its expression decreases when cells become lineage specific [bib_ref] Autoantibodies against the endothelial receptor of protein C are associated with acute..., Montes [/bib_ref] indicating that EPCR plays a potential role in stem cell proliferation and differentiation [bib_ref] The endothelial cell protein C receptor, Esmon [/bib_ref]. Additionally, EPCR can be cleaved from the cell surface by matrix metalloproteinases (MMPs) and circulate as a soluble form (sEPCR) in the blood where it still may bind to PC and APC, as well as interact with integrins on neutrophils during sepsis. Soluble EPCR may alter the active site of APC making phospholipid interactions impossible, thus blocking APC anticoagulant function.
EPCR expression is regulated by multiple factors, e.g., transcription is suppressed by lipopolysaccharide, IL-1β, TNFα, and thrombin. There are gene polymorphisms of EPCR, and haplotype A3 is reportedly responsible for elevated levels of plasma EPCR in the course of inflammatory disorders like sepsis or lupus erythematosus [bib_ref] Gene expression profile of antithrombotic protein C defines new mechanisms modulating inflammation..., Joyce [/bib_ref] [bib_ref] Plasma levels of endothelial cell protein C receptor are elevated in patients..., Kurosawa [/bib_ref] [bib_ref] The H3 Haplotype of the EPCR Gene Determines High sEPCR Levels in..., Vassiliou [/bib_ref]. Moreover, haplotype A3-mediated elevation of sEPCR has been associated with elevated markers of prothrombin activation (fragment 1-2), venous thrombosis, unexplained fetal death [bib_ref] High plasma levels of endothelial protein C receptor are associated with the..., Lavigne-Lissalde [/bib_ref] and elevated risk of coronary diseases [bib_ref] Elevated sEPCR, prothrombin F1+2, risk for coronary heart disease, and increased sEPCR..., Ireland [/bib_ref] [bib_ref] Haplotypes of the EPCR gene, plasma sEPCR levels and the risk of..., Uitte [/bib_ref] [bib_ref] Functional analysis of two haplotypes of the human endothelial protein C receptor..., Medina [/bib_ref]. Patients with haplotype 1 EPCR had lower levels of sEPCR and reduced risk of venous thromboembolism. Polymorphism of EPCR may have significance in the biology of cancer as oncology patients commonly have an increased thrombotic state. Apart from haplotypes of EPCR, a single nucleotide polymorphism (SNP) of EPCR, rs2069948, was associated with estrogen receptor (ER) and progesterone receptor (PR) positivity in breast cancer specimens [bib_ref] Increased coagulation activity and genetic polymorphisms in the F5, F10 and EPCR..., Tinholt [/bib_ref].
The structure of the EPCR, confirmed via crystallization, is similar to the major histocompatibility complex class 1/CDI family of proteins, which are commonly involved in immunity, which partially explains the contribution of EPCR to inflammation. EPCR is located in the caveolin-positive lipid compartment of the membrane close to thrombomodulin (TM) [bib_ref] A novel transmembrane domain of the EPCR dictates receptor localization of sphingolipid-cholesterol..., Xu [/bib_ref]. The cytoplasmic tail (Arg-Arg-Cys-COOH) of EPCR, which does not play a direct role in cell signaling, is likely essential for anchoring EPCR into lipid rafts near the source of signal mediators [bib_ref] Endothelial cell protein C receptor-dependent signaling, Pendurthi [/bib_ref]. The extracellular domain of EPCR is comprised of two α-helices that expose residues that react with vitamin K-dependent liver plasma glycoprotein PC or APC [bib_ref] The crystal structure of the endothelial protein C receptor and a bound..., Oganesyan [/bib_ref]. EPCR binds PC and APC with high affinity in a Ca 2+ -dependent manner through the Gla domain of these ligands. EPCR is considered a cofactor of protein C as it accelerates TM-thrombin complex-mediated activation of PC to APC by concentrating protein C near the surface of the vessel wall. APC regulates thrombin production. Therefore, after being activated, APC dissociates from the complex with EPCR and takes part in degradation of cofactors Va and VIIIa, thus finally down-regulating further thrombin formation (anticoagulant mechanism) [fig_ref] Figure 1: Protein C [/fig_ref] [bib_ref] Endothelial cell protein C receptor augments protein C activation by the thrombin-thrombomodulin..., Stearns-Kurosawa [/bib_ref].
Cancers 2019, 11, x 3 of 18 activation (fragment 1-2), venous thrombosis, unexplained fetal death [bib_ref] High plasma levels of endothelial protein C receptor are associated with the..., Lavigne-Lissalde [/bib_ref] and elevated risk of coronary diseases [bib_ref] Elevated sEPCR, prothrombin F1+2, risk for coronary heart disease, and increased sEPCR..., Ireland [/bib_ref] [bib_ref] Haplotypes of the EPCR gene, plasma sEPCR levels and the risk of..., Uitte [/bib_ref] [bib_ref] Functional analysis of two haplotypes of the human endothelial protein C receptor..., Medina [/bib_ref]. Patients with haplotype 1 EPCR had lower levels of sEPCR and reduced risk of venous thromboembolism. Polymorphism of EPCR may have significance in the biology of cancer as oncology patients commonly have an increased thrombotic state. Apart from haplotypes of EPCR, a single nucleotide polymorphism (SNP) of EPCR, rs2069948, was associated with estrogen receptor (ER) and progesterone receptor (PR) positivity in breast cancer specimens [bib_ref] Increased coagulation activity and genetic polymorphisms in the F5, F10 and EPCR..., Tinholt [/bib_ref]. The structure of the EPCR, confirmed via crystallization, is similar to the major histocompatibility complex class 1/CDI family of proteins, which are commonly involved in immunity, which partially explains the contribution of EPCR to inflammation. EPCR is located in the caveolin-positive lipid compartment of the membrane close to thrombomodulin (TM) [bib_ref] A novel transmembrane domain of the EPCR dictates receptor localization of sphingolipid-cholesterol..., Xu [/bib_ref]. The cytoplasmic tail (Arg-Arg-Cys-COOH) of EPCR, which does not play a direct role in cell signaling, is likely essential for anchoring EPCR into lipid rafts near the source of signal mediators [bib_ref] Endothelial cell protein C receptor-dependent signaling, Pendurthi [/bib_ref]. The extracellular domain of EPCR is comprised of two α-helices that expose residues that react with vitamin K-dependent liver plasma glycoprotein PC or APC [bib_ref] The crystal structure of the endothelial protein C receptor and a bound..., Oganesyan [/bib_ref]. EPCR binds PC and APC with high affinity in a Ca 2+ -dependent manner through the Gla domain of these ligands. EPCR is considered a cofactor of protein C as it accelerates TM-thrombin complex-mediated activation of PC to APC by concentrating protein C near the surface of the vessel wall. APC regulates thrombin production. Therefore, after being activated, APC dissociates from the complex with EPCR and takes part in degradation of cofactors Va and VIIIa, thus finally down-regulating further thrombin formation (anticoagulant mechanism) [fig_ref] Figure 1: Protein C [/fig_ref] [bib_ref] Endothelial cell protein C receptor augments protein C activation by the thrombin-thrombomodulin..., Stearns-Kurosawa [/bib_ref]. EPCR is localized in small amounts within the recycling compartment of the cytoplasm, where regulation of its activation occurs via circulation of the EPCR-PC/APC complex between the cell surface and recycling compartment of the endothelial cells. It has been suggested that internalization may promote clearance of PC/APC ligands from circulation [bib_ref] Endothelial cell protein C receptor cellular localization and trafficking: Potential functional implications, Nayak [/bib_ref].
Other serine proteases bind to EPCR with similar affinity as PC, and also in a Ca 2+ -dependent mechanism. These include factor VIIa (FVIIa) and factor Xa (FXa) that are involved in hemostasis, tissue repair, inflammatory processes and cancer dissemination [bib_ref] Factor VIIa induces anti-inflammatory signaling via EPCR and PAR1, Kondreddy [/bib_ref] [bib_ref] Is EPCR a multi-ligand receptor? Pros and cons, Montes [/bib_ref] [bib_ref] The endothelial protein C receptor supports tissue factor ternary coagulation initiation complex..., Disse [/bib_ref].
New EPCR ligands have been discovered and include Plasmodium falciparum erythrocyte membrane protein, and a specific variant of the T-cell receptor present on a Vδ2neg γδ T cells [bib_ref] Endothelial cell protein C receptor-dependent signaling, Pendurthi [/bib_ref].
There is substantial evidence from in vivo studies that EPCR plays a notable role in anticoagulation. EPCR inhibition can increase the thrombotic state of animals, causing vessel occlusion in mice. Similarly, the presence of autoantibodies against EPCR was described in patients experiencing episodes of thrombosis [bib_ref] Autoantibodies against endothelial protein C receptor and the risk of a first..., Van Hylckama Vlieg [/bib_ref]. EPCR is localized in small amounts within the recycling compartment of the cytoplasm, where regulation of its activation occurs via circulation of the EPCR-PC/APC complex between the cell surface and recycling compartment of the endothelial cells. It has been suggested that internalization may promote clearance of PC/APC ligands from circulation [bib_ref] Endothelial cell protein C receptor cellular localization and trafficking: Potential functional implications, Nayak [/bib_ref].
Other serine proteases bind to EPCR with similar affinity as PC, and also in a Ca 2+ -dependent mechanism. These include factor VIIa (FVIIa) and factor Xa (FXa) that are involved in hemostasis, tissue repair, inflammatory processes and cancer dissemination [bib_ref] Factor VIIa induces anti-inflammatory signaling via EPCR and PAR1, Kondreddy [/bib_ref] [bib_ref] Is EPCR a multi-ligand receptor? Pros and cons, Montes [/bib_ref] [bib_ref] The endothelial protein C receptor supports tissue factor ternary coagulation initiation complex..., Disse [/bib_ref].
New EPCR ligands have been discovered and include Plasmodium falciparum erythrocyte membrane protein, and a specific variant of the T-cell receptor present on a Vδ2neg γδ T cells [bib_ref] Endothelial cell protein C receptor-dependent signaling, Pendurthi [/bib_ref].
There is substantial evidence from in vivo studies that EPCR plays a notable role in anticoagulation. EPCR inhibition can increase the thrombotic state of animals, causing vessel occlusion in mice. Similarly, the presence of autoantibodies against EPCR was described in patients experiencing episodes of thrombosis [bib_ref] Autoantibodies against endothelial protein C receptor and the risk of a first..., Van Hylckama Vlieg [/bib_ref]. . Ligands leading to EPCR (endothelial protein C receptor) or PAR-1 (protease-activated receptor-1) activation [bib_ref] Endothelial cell protein C receptor-dependent signaling, Pendurthi [/bib_ref] [bib_ref] Protease-activated receptors (PARs)-biology and role in cancer invasion and metastasis, Wojtukiewicz [/bib_ref]. APC (activated protein C), TF (tissue factor), MMPs (matrix metalloproteinases).
## Receptor epcr par-1
Ligand Factor VIIa Factor Xa TF-VIIa-Xa TF-VIIa Plasmodium falciparum erythrocyte membrane protein T-cell receptor present on a subset of Vδ2neg γδ T cells Thrombin Factor Xa TF-VIIa-Xa APC Plasmin Granzyme A Gingipains-R Trypsin MMP-1, -9, -2, -13, -14 APC/EPCR activation results in cytoprotective, anti-inflammatory and anti-apoptotic cellular effects that have been widely described in multiple injury models, as well as inflammation and ischemic stroke models [bib_ref] Gene expression profile of antithrombotic protein C defines new mechanisms modulating inflammation..., Joyce [/bib_ref] [bib_ref] Protective mechanisms of activated protein C in severe inflammatory disorders, Neyrinck [/bib_ref] [bib_ref] The cytoprotective protein C pathway, Mosnier [/bib_ref] [bib_ref] Cytoprotective-selective activated protein C therapy for ischaemic stroke, Mosnier [/bib_ref] [bib_ref] Scientific Sessions Sol Sherry Distinguished Lecturer in Thrombosis: Thrombotic Stroke: Neuroprotective Therapy..., Griffin [/bib_ref]. It has been demonstrated with in vivo studies that the brain is particularly sensitive to APC-induced signaling. There APC exerts anti-apoptotic and neuroprotective effects resulting in decreased brain ischemia [bib_ref] Cytoprotective-selective activated protein C therapy for ischaemic stroke, Mosnier [/bib_ref] [bib_ref] Scientific Sessions Sol Sherry Distinguished Lecturer in Thrombosis: Thrombotic Stroke: Neuroprotective Therapy..., Griffin [/bib_ref]. APC may be generated in human brain as APC levels appear to decrease in stroke patients [bib_ref] Cytoprotective-selective activated protein C therapy for ischaemic stroke, Mosnier [/bib_ref]. Clinical research has focused on the APC variant, 3K3A-APC, that is cytoprotective and independent of the canonical anticoagulant activity induced by activation of PAR-1 [bib_ref] Cytoprotective-selective activated protein C therapy for ischaemic stroke, Mosnier [/bib_ref] [bib_ref] PAR1 biased signaling is required for activated protein C in vivo benefits..., Sinha [/bib_ref]. Neuroprotective therapy with recombinant 3K3A-APC is being evaluated in ongoing National Institutes of Health (NIH)-funded clinical trials for ischemic stroke [bib_ref] Scientific Sessions Sol Sherry Distinguished Lecturer in Thrombosis: Thrombotic Stroke: Neuroprotective Therapy..., Griffin [/bib_ref]. Additionally, the anti-inflammatory and cytoprotective features of APC/EPCR interactions were exploited to treat patients with severe sepsis in the PROWESS clinical trial [bib_ref] The cytoprotective protein C pathway, Mosnier [/bib_ref] [bib_ref] Cytoprotective-selective activated protein C therapy for ischaemic stroke, Mosnier [/bib_ref] [bib_ref] Protective signaling by activated protein C is mechanistically linked to protein C..., Feistritzer [/bib_ref].
The EPCR-mediated role in vascular barrier function and cytoprotective features in endothelium may potentially contribute to cancer development and progression. Various cancer cells express the EPCR (e.g., colorectal cancer, lung cancer, malignant pleural mesothelioma, breast cancer, ovarian cancer, gastric cancer) [bib_ref] Endothelial protein C receptor is overexpressed in colorectal cancer as a result..., Lal [/bib_ref] [bib_ref] Endothelial cell protein C receptor promotes MGC803 gastric cancer cells proliferation and..., Wang [/bib_ref] [bib_ref] Endothelial protein C receptor function in murine and human breast cancer development, Schaffner [/bib_ref] [bib_ref] Intrapleural Adenoviral-mediated Endothelial Cell Protein C Receptor Gene Transfer Suppresses the Progression..., Keshava [/bib_ref] [bib_ref] EPCR promotes breast cancer progression by altering SPOCK1/testican 1-mediated 3D growth, Perurena [/bib_ref] [bib_ref] PO-47-Microparticles derived from ovarian cancer cell line contained genomic and biologically active..., Besbes [/bib_ref]. Upregulation of EPCR in malignant cells results from gene amplification and DNA hypomethylation [bib_ref] Endothelial protein C receptor is overexpressed in colorectal cancer as a result..., Lal [/bib_ref] ; however, its distribution varies. In studies of gynecological cancers, EPCR was absent from microparticles derived from the ovarian adenocarcinoma cell line (OVCAR-3), but was detected on intact OVCAR-3 cells [bib_ref] PO-47-Microparticles derived from ovarian cancer cell line contained genomic and biologically active..., Besbes [/bib_ref]. The expression of EPCR and key proteins associated with EPCR-dependent signaling (protein S, protein C, thrombomodulin, Factor V, VIII and PAR-1 and PAR-2) was investigated in endometrial and ovarian cancers [bib_ref] -14-Tumour expression of coagulation proteases of the aPC pathway-A role in the..., Martin [/bib_ref] and compared with benign tumors. Interestingly, the gynecological cancers expressed EPCR and associated proteins, but to a lesser extent than benign tumors. PAR-1 expression did not differ between benign and malignant tumors.
The clinical significance of the APC/EPCR complex in cancer biology appears to vary depending on the tumor subtype [fig_ref] Table 2: Distinct cellular effects mediated by endothelial protein C receptor [/fig_ref]. Animal model-based studies revealed that increased expression of EPCR in cancer cells inhibits proliferation and migration through the extracellular signal-regulated kinase, ERK/AKT-dependent signaling pathway [bib_ref] Endothelial protein C receptor is overexpressed in colorectal cancer as a result..., Lal [/bib_ref] [bib_ref] Intrapleural Adenoviral-mediated Endothelial Cell Protein C Receptor Gene Transfer Suppresses the Progression..., Keshava [/bib_ref] and promotes apoptosis via BAX and BCL2 factors, thereby limiting tumor growth and dissemination [bib_ref] Role of activated protein C and its receptor in inhibition of tumor..., Bezuhly [/bib_ref] [bib_ref] Intrapleural Adenoviral-mediated Endothelial Cell Protein C Receptor Gene Transfer Suppresses the Progression..., Keshava [/bib_ref]. Transgenic mice overexpressing EPCR (Tie2-EPCR) had a 50-92% decrease in liver and lung metastases compared with wildtype animals. Additionally, treatment of the mouse with recombinant human APC or a signaling-proficient mutant, APC-2Cys (with reduced anticoagulant activity), also led to a reduction of metastatic melanoma foci by inhibiting transendothelial migration of malignant cells [bib_ref] Role of activated protein C and its receptor in inhibition of tumor..., Bezuhly [/bib_ref] [bib_ref] TR47, a PAR1-based peptide, inhibits melanoma cell migration in vitro and metastasis..., De Oliveira [/bib_ref]. The APC/EPCR interaction, e.g., in melanoma cells, likely decreases expression of endothelial adhesion molecules, such as P-selectin that is essential for tumor cell-endothelium interactions during extravasation, ultimately hindering dissemination [bib_ref] Role of activated protein C and its receptor in inhibition of tumor..., Bezuhly [/bib_ref].
Concurrently, and in contrast with the data that APC and EPCR are tumor suppressors, there are multiple studies that demonstrate EPCR expression by cancer cells can actually promote growth and dissemination as a result of antiapoptotic effects [bib_ref] Protein C and its inhibitor in malignancy, Suzuki [/bib_ref] [bib_ref] EPCR promotes breast cancer progression by altering SPOCK1/testican 1-mediated 3D growth, Perurena [/bib_ref] [bib_ref] Activated protein C promotes breast cancer cell migration through interactions with EPCR..., Beaulieu [/bib_ref] ]. This has been described in lung cancer cells in vitro, where the APC/EPCR complex inhibits apoptosis by triggering the AKT/ERK signaling pathways [bib_ref] Receptor of activated protein C promotes metastasis and correlates with clinical outcome..., Antón [/bib_ref]. Furthermore, in vivo experiments where the APC/EPCR interaction was blocked resulted in reduced metastasis. There appears to be an association between high levels of EPCR and poor prognosis in lung cancer patients, especially at early stages of the disease.
Experiments with human and murine xenograft breast cancer models revealed that EPCR silencing decreased primary tumor growth and the establishment of metastatic foci in distant organs, e.g., bones and lungs. Surprisingly, EPCR stimulation by APC under both in vitro and in vivo conditions did not influence these effects. It has been suggested that EPCR-mediated matricellular secretion of proteoglycan SPOCK1/testican-1 was responsible for inducing breast cancer growth. Analysis of the EPCR transcript in 286 breast cancer patients revealed a correlation between high EPCR levels and patient survival [bib_ref] EPCR promotes breast cancer progression by altering SPOCK1/testican 1-mediated 3D growth, Perurena [/bib_ref]. EPCR (PROCR, protein C receptor, CD201) expression was also assessed in invasive ductal carcinoma of the breast from 271 patients with stage II or III disease. Positive expression of EPCR correlated with development of metastases and decreased disease-free and overall survival [bib_ref] Prevalence of protein C receptor (PROCR) is associated with inferior clinical outcome..., Yan [/bib_ref]. [bib_ref] Influence of endothelial cell protein C receptor on breast cancer development, Keshava [/bib_ref] demonstrated that the role of EPCR in tumor growth varied depending on the time breast cancer cells had been injected into the mammary fat pad. Namely, malignant cells positive for EPCR expression exhibited greater tumor growth than EPCR-negative cells, but only during the first 40 days of implantation into the mouse. Interestingly, after that time the growth of those tumors derived from EPCR-positive cells was slower than that from control cells, i.e., EPCR-negative, such that 60 days post injection the tumor volume of EPCR-positive cells was about 30% smaller than tumor originating from control cells. Moreover, animals bearing EPCR-negative tumors presented with more advanced angiogenesis, swollen lymph nodes and extensive inflammatory reactions on the skin, with some animals exhibiting ulceration above the tumor. Necrosis, a marker of more aggressive breast cancers, was also noticed only in EPCR-negative tumors. None of the EPCR-positive tumors demonstrated skin ulceration and angiogenesis was less advanced than in control tumors. EPCR expression was eventually lost in some tumors that were primarily composed of EPCR-positive cells, which confirms the complexity of EPCR-dependent signaling in tumor progression, and also indicates its potentially protective role in preventing cancer growth and dissemination.
Another study that utilized two breast cancer cell lines, MDA-MB-231 and MDA-MB-435, unveiled that APC induced chemotaxis and invasion through the EPCR/PAR-1-dependent signaling pathway [bib_ref] Activated protein C promotes breast cancer cell migration through interactions with EPCR..., Beaulieu [/bib_ref]. The aggressive, triple-negative breast cancer subtype has been shown to express high levels of EPCR, which is correlated with tumorgenicity [bib_ref] Role of the protein C receptor in cancer progression, Ruf [/bib_ref]. The expression of EPCR was also demonstrated on breast cancer stem cells and, interestingly, this population of cells could initiate tumors [bib_ref] Role of the protein C receptor in cancer progression, Ruf [/bib_ref]. Complete inhibition of APC-mediated effects could be achieved with antibodies directed to EPCR and PAR-1. Additionally, the role of APC in angiogenesis was discovered by APC stimulation of MDA-MB-231 breast cancer cells, which resulted in endothelial tube formation [bib_ref] Activated protein C enhances cell motility of endothelial cells and MDA-MB-231 breast..., Gramling [/bib_ref]. The increased proliferation of endothelial cells in vitro and induced angiogenesis in animal models were regulated Cancers 2019, 11, 51 6 of 18 by APC/EPCR/MAPK and partly by the PAR-1 signaling pathway [bib_ref] Activated protein C induces endothelial cell proliferation by mitogen-activated protein kinase activation..., Uchiba [/bib_ref]. An inhibitor of APC, PCI, reduced tumor cell invasion in vitro by blocking protease activity [bib_ref] Protein C and its inhibitor in malignancy, Suzuki [/bib_ref].
There are conflicting data regarding the association between EPCR expression and chemosensitivity in tumors. In lung cancer, expression of EPCR has been proven to increase chemosensitivity, while in colorectal cancer such effects were not observed as the overexpression of EPCR was accompanied by expression of neighboring chemoresistance genes on chromosome 20q [bib_ref] Endothelial protein C receptor is overexpressed in colorectal cancer as a result..., Lal [/bib_ref]. Chemotherapy itself, e.g., with doxorubicin, may cause down-regulation of EPCR in endothelial cells [bib_ref] Effects of the chemotherapeutic agent doxorubicin on the protein C anticoagulant pathway, Woodley-Cook [/bib_ref]. There is a case report that describes a cancer patient receiving chemotherapy who was deficient in APC and who subsequently developed clinically significant thrombosis [bib_ref] Acquired protein C deficiency in a child with acute myelogenous leukemia, splenic,..., Farah [/bib_ref].
## Par-1 and cancer
Protease activated receptor-1 (PAR-1) is one of four isoforms identified to date. It is a G-protein-coupled receptor, and as with EPCR it is located in the lipid raft compartment of the cell membrane. PAR-1 is composed of seven transmembrane α-helices, a cytoplasmic domain for G-protein coupling, and an extracellular N-terminus. A comprehensive review of PARs was published in 2015 [bib_ref] Protease-activated receptors (PARs)-biology and role in cancer invasion and metastasis, Wojtukiewicz [/bib_ref]. The present review focuses on matters pivotal for understanding the role of PAR-1 in APC/EPCR/PAR-1-associated signaling. PAR-1 expression was documented for nearly all cell types in the blood vessel wall (ECs, fibroblasts, myocytes), and blood (platelets, neutrophils, macrophages, leukemic white cells) with exception of erythrocytes. It was also identified on epithelium, and in neurons, astrocytes and immune cells [bib_ref] Protease-activated receptors (PARs)-biology and role in cancer invasion and metastasis, Wojtukiewicz [/bib_ref]. Importantly, PAR-1 expression was also detected in multiple cancer subtypes, including epithelial carcinomas, melanoma, glioblastoma (GBM), giant cell tumors and sarcoma, and is present on both malignant cells as well as on tumor-associated stromal components (reviewed in [bib_ref] Protease-activated receptors (PARs)-biology and role in cancer invasion and metastasis, Wojtukiewicz [/bib_ref] [bib_ref] TGF-β induced PAR-1 expression promotes tumor progression and osteoclast differentiation in giant..., Wang [/bib_ref]. The latest findings indicate that PAR-1 expression occurs on extracellular vesicles derived from cancer cells, e.g., prostate, breast, pancreas [bib_ref] Thrombotic characteristics of extracellular vesicles derived from prostate cancer cells, Saleh [/bib_ref]. Antibodies to PAR-1 have also been discovered in cancer patients [bib_ref] The Role of PAR1 Autoantibodies in Patients with Primary Epithelial Ovarian Cancer, Kreienbring [/bib_ref].
There are multiple PAR-1 ligands . Among these thrombin, APC and haemostatic factors VIIa, Xa and their complex with TF (TF-VIIa-Xa) have been proven to be associated with EPCR and PAR-1-dependent signaling.
Importantly there are canonical (classical) and non-canonical modes of PAR-1 activation. The main activator of PAR-1, thrombin, represents the canonical signaling pathway. After binding to the receptor N-terminal sequence LDPR41-S42 the R41-S42 peptide bond becomes cleaved thereby generating a new sequence referred to as a tethered ligand of PAR-1. The tethered ligand binds residues 42SFLLRN47 in the conserved region of the second loop of the receptor to induce transmembrane signaling [bib_ref] Protease-activated receptors: Contribution to physiology and disease, Ossovskaya [/bib_ref]. Interestingly thrombin-mediated PAR-1 activation does not require the presence of any cofactors due to highly acidic regions (P4-L38 and P2-P40 residues) that increase thrombin affinity and facilitate proper cleavage of PAR-1. However, for APC to cleave the N-terminus of PAR-1 in cancer cells, EPCR is required as a cofactor [bib_ref] Activation of endothelial cell protease activated receptor-1 by the protein C pathway, Riewald [/bib_ref]. PAR-1 activation mediated by thrombin causes receptor coupling to Gα protein (Gαq, Gαi and Gα12/13) and Gβγ. Heterotrimers composed of PAR-1 and Gαq lead to activation of MAP kinase (mitogen activated protein kinase) and increased Ca 2+ concentration, while complexes of PAR-1 with Gα12/13 activate the small G-protein, RhoA (Ras homolog gene family member A) [bib_ref] Paradigm of Biased PAR1 (Protease-Activated Receptor-1) Activation and Inhibition in Endothelial Cells..., Van Den Eshof [/bib_ref]. Activation of PAR-1 by APC in turn leads to activation of Ras-related C3 botulinum toxin substrate 1 (Rac1).
Finally, thrombin-mediated PAR-1 activation drives production of cytokines, chemokines, growth factors and bioactive lipids to trigger inflammation, adhesion, and endothelial barrier disruption, all of which contribute to tumor growth and survival. It also promotes angiogenesis and transendothelial migration resulting in tumor progression and metastasis [bib_ref] TGF-β induced PAR-1 expression promotes tumor progression and osteoclast differentiation in giant..., Wang [/bib_ref]. Studies done in vitro showed that overexpression of PAR-1 in cancer cells is associated with greater invasiveness and ability to disseminate, making PAR-1 expression in some cancers an unfavorable prognostic factor in terms of overall survival or local recurrence.
To complicate matters, a recent study with a pancreatic ductal adenocarcinoma cell line surprisingly demonstrated that PAR-1-associated signaling instead limits tumor growth, probably via induction and maintenance of a mesenchymal-like cell phenotype [bib_ref] PAR1 signaling on tumor cells limits tumor growth by maintaining a mesenchymal..., Tekin [/bib_ref]. This indicates that there are still unexplored molecular events associated with PAR-1 signaling pathways.
Ligands other than thrombin (e.g., MMP-1, APC, FXa) can activate PAR-1 in noncanonical ways resulting in biological effects that differ from those driven by thrombin. Thrombin-mediated PAR-1 activation increases endothelial permeability, while the APC-mediated effects result in endothelial barrier protection and in some cancer models are antimetastatic. MMP-specific signaling patterns exhibited by PAR-1, known as biased agonism, also result in distinct functional outputs compared to thrombin-induced PAR-1 activation. Cleavage of the extracellular portion of the PAR-1 receptor by thrombin occurs at a canonical R41-S42 site, while MMP-1 cleaves PAR-1 at a novel site (D39-P40) resulting in a tethered ligand that is two amino acids longer (PR-SFLLRN) than that generated by thrombin. Subsequent peak Akt signaling occurs after five minutes in thrombin-triggered activation, while MMP-1-triggered Akt activation occurs after one hour and induces different cellular events [bib_ref] PAR1 is a matrix metalloprotease-1 receptor that promotes invasion and tumorigenesis of..., Boire [/bib_ref].
Noncanonical PAR activation has also been documented for FXa, which leads to tunica intima endothelial receptor tyrosine kinase 2 (Tie2) activation in an EPCR-dependent manner, and ultimately endothelial barrier enforcement by upregulation of zona occludens 1 (ZO-1) to stabilize cell-cell junctions.
## Epcr and par-1 interactions
## G-proteins
There is evidence that most cellular effects of EPCR-bound APC require PAR-1 activation [bib_ref] The occupancy of endothelial protein C receptor by its ligand modulates the..., Rezaie [/bib_ref]. APC is much less effective at cleaving and activating PAR-1 compared to thrombin [bib_ref] PAR1 cleavage and signaling in response to activated protein C and thrombin, Ludeman [/bib_ref] and it also evokes different biological effects [bib_ref] Protease-activated receptor-1 signaling by activated protein C in cytokine-perturbed endothelial cells is..., Riewald [/bib_ref]. Classic thrombin-dependent proteolysis of PAR-1 at canonical Arg41 couples G12/13 and Gq signaling that ultimately, through MAPK and RhoA activation, elicits proinflammatory effects. Interestingly, the APC/EPCR complex causes the dissociation of EPCR from caveolin-1 in lipid rafts resulting in PAR-1 cleavage at non-canonical Arg46 and further activation of protein Gi instead of G12/13 and Gq [bib_ref] Is EPCR a multi-ligand receptor? Pros and cons, Montes [/bib_ref]. Finally, β-arrestin-mediated phosphorylation of Akt leads to activation of Ras-related C3 botulinum toxin substrate 1 (Rac1) signaling. APC-cleaved PAR-1 remains present on the cell surface for some time likely blocking thrombin-mediated PAR-1 activation. The slower and longer activation of PAR-1 by APC compared to thrombin may explain the up-regulation of more cytoprotective and anti-inflammatory genes [fig_ref] Figure 2: Distinct mechanisms of PAR-1 [/fig_ref].
For example, there is increased expression of the sphingosine-1-phosphate receptor, which upregulates Tie2, the Angiopoietin-1/Angipoietin-2 ratio and VE-cadherin to ultimately reinforce vascular barrier protection [bib_ref] Role of activated protein C and its receptor in inhibition of tumor..., Bezuhly [/bib_ref] [bib_ref] Endogenous EPCR/aPC-PAR1 signaling prevents inflammation-induced vascular leakage and lethality, Niessen [/bib_ref]. Regulation of cytoskeleton-mediated cell-cell interactions and blocking of actin stress fiber formation is observed as a result of APC-mediated PAR-1 activation [bib_ref] Endothelial barrier protection by activated protein C through PAR1-dependent sphinosine 1-phospate receptor-1..., Feistritzer [/bib_ref] [bib_ref] Activated protein C mediates novel lung endothelial barrier enhancement: Role of sphingosine..., Finigan [/bib_ref]. The cytoprotective properties of the APC/EPCR/PAR-1 signaling cascade and enhancement of endothelial barrier inhibits tumor cell extravasation and dissemination as a result of the PAR-1-dependent mechanism. Absence of endogenous APC as a result of loss of VE-cadherin junctions increases vascular leakage and cancer cell extravasation. The anti-inflammatory effects of EPCR/PAR-1 interactions are associated with suppression of the nuclear factor kappa light-chain enhancer of activated B cells (NF-kB) signaling pathway [bib_ref] Scientific Sessions Sol Sherry Distinguished Lecturer in Thrombosis: Thrombotic Stroke: Neuroprotective Therapy..., Griffin [/bib_ref] [bib_ref] The occupancy of endothelial protein C receptor by its ligand modulates the..., Rezaie [/bib_ref]. Cellular models confirmed that blockade of either one of these receptors (EPCR, PAR-1, S1P1) eliminated the protective effect of APC. The role of APC/EPCR/PAR-1 pathway activation was also described in keratinocyte proliferation [bib_ref] Endothelial protein C receptor and protease-activated receptor-1 mediate induction of a wound-healing..., Xue [/bib_ref]. The co-localization of EPCR and PAR-1 within lipid compartments and caveolae microdomains in endothelial cells plays a role in interactions with PC. Recent experiments demonstrated that administration of PC to septic mice exerted anti-inflammatory and anti-apoptotic effects in the septic lungs and also resulted in reduced caveolin-1 expression in the lungs, thus facilitating the interaction between EPCR and PC that induced PAR1-dependent cytoprotective signaling.
Arg46 and further activation of protein Gi instead of G12/13 and Gq [bib_ref] Is EPCR a multi-ligand receptor? Pros and cons, Montes [/bib_ref]. Finally, β-arrestin-mediated phosphorylation of Akt leads to activation of Ras-related C3 botulinum toxin substrate 1 (Rac1) signaling. APC-cleaved PAR-1 remains present on the cell surface for some time likely blocking thrombin-mediated PAR-1 activation. The slower and longer activation of PAR-1 by APC compared to thrombin may explain the up-regulation of more cytoprotective and anti-inflammatory genes [fig_ref] Figure 2: Distinct mechanisms of PAR-1 [/fig_ref]. Additionally, it has been suggested that APC light chain amino acid residues outside the EPCR-binding site enable activation of APC-mediated PAR-1 cytoprotective functions [bib_ref] A novel protein C-factor VII chimera provides new insights into the structural..., Gleeson [/bib_ref]. There are molecules termed "parmodulins" that may interact with the cytosolic face of PAR-1 and cause similar APC-like, PAR-1-dependent cytoprotective signaling in endothelium [bib_ref] PAR1 agonists stimulate APC-like endothelial cytoprotection and confer resistance to thromboinflammatory injury, De Ceunynck [/bib_ref].
There are contradictory findings that are of paramount importance with respect to the role of the APC/EPCR/PAR-1 axis in cancer progression [bib_ref] Endothelial cell protein C receptor promotes MGC803 gastric cancer cells proliferation and..., Wang [/bib_ref] [bib_ref] Knockdown of EPCR inhibits the proliferation and migration of human gastric cancer..., Wang [/bib_ref] [bib_ref] EPCR promotes MGC803 human gastric cancer cell tumor angiogenesis in vitro through..., Wang [/bib_ref]. Wang et al. [bib_ref] Endothelial cell protein C receptor promotes MGC803 gastric cancer cells proliferation and..., Wang [/bib_ref] demonstrated that gastric cancer tissue expresses elevated levels of EPCR and that EPCR-mediated APC activation induces proliferation and migration of the MGC803 gastric cancer cells. Additional experiments showed that the EPCR/APC/PAR-1 cascade also induces angiogenesis in gastric cancer cells [bib_ref] EPCR promotes MGC803 human gastric cancer cell tumor angiogenesis in vitro through..., Wang [/bib_ref]. The microvessel density (MVD) of 61 surgically removed primary gastric cancer tumors were assessed by using endothelial markers CD31 and CD34. Reportedly the mean value of MVD was higher in gastric cancer tissues expressing EPCR compared to EPCR-negative tissue. Human umbilical vein endothelial cells (HUVECs) were cultured with tumor-conditioned medium derived from EPCR knockdown or PAR-1-inhibited MGC803 gastric cancer cells. Interestingly, knocking down EPCR by small interference RNA and blocking PAR-1 activity by antibody resulted in inhibition of phosphorylation of extracellular signaling-regulated kinases 1 and 2 (ERK1/2) and Akt, thereby leading to suppression of the proliferation, migration and tubule formation potential of HUVECs. These findings indicate that EPCR and PAR-1 reciprocal interactions contribute to cancer-promoting events in malignant cells [bib_ref] Endothelial cell protein C receptor promotes MGC803 gastric cancer cells proliferation and..., Wang [/bib_ref].
## B-arrestin
Although some mechanisms of triggering protein Gi instead of G12/13 and Gq as a result of APC-EPCR/PAR-1 interactions explain the cytoprotective role of PAR-1, it still remains unclear whether Gi is totally responsible for these effects. There are interesting data that explain the phenomenon of PAR-1 activation by two different proteases. The PAR1-induced cytoprotective activity of APC may not be mediated through either one of the G-proteins, but rather via β-arrestin-2-biased signaling [bib_ref] Occupancy of human EPCR by protein C induces β-arrestin-2 biased PAR1 signaling..., Roy [/bib_ref]. It has been demonstrated that APC occupancy by EPCR leads to the recruitment of G-protein-coupled receptor kinase 5 (GRK5) to the plasma membrane and phosphorylation of the PAR-1 cytoplasmic domain. This phosphorylation suppresses PAR-1-dependent activation of G proteins resulting in β-arrestin-2 biased PAR-1 cytoprotective signaling independent of the protease cleavage site and protein Gi coupling. Interestingly, thrombin also exerts PAR-1-dependent cytoprotective effects by β-arrestin-2-induced recruitment of disheveled 2 (Dvl-2) scaffolding protein and Rac1 signaling.
## Factor vii/viia and factor xa (fxa)
Other serine proteases that may occupy EPCR in a Ca 2+ -dependent mechanism and exert protective effects on the endothelial barrier are factor VII/VIIa and factor X/Xa (FXa), which bind EPCR through the 4-carboxyglutamic acid domains [bib_ref] Factor VIIa induces anti-inflammatory signaling via EPCR and PAR1, Kondreddy [/bib_ref] [bib_ref] The endothelial protein C receptor supports tissue factor ternary coagulation initiation complex..., Disse [/bib_ref] [bib_ref] Coagulation factor Xa cleaves protease-activated receptor-1 and mediates signaling dependent on binding..., Schuepbach [/bib_ref]. The cytoprotective properties mediated by FVIIa binding to EPCR depend on PAR-1 and β-arrestin-1 activation [fig_ref] Figure 3: Factor VIIa-mediated activation of endothelial protein C receptor [/fig_ref]. FVIIa-treated ECs in vitro had reduced expression of cellular adhesion molecules and adherence of monocytes to ECs. These effects were dependent on PAR-1-mediated suppression of tumor necrosis factor α (TNF-α). PAR-1 antibodies or small interfering RNA abolished TNF-α-induced activation of ERK1/2, p38 MAPK, JNK, NF-κB, and C-Jun factors as well as lipopolysaccharide (LPS)-mediated in vivo inflammatory responses in lungs of mice overexpressing EPCR. Additionally, FVIIa in the PAR-1/MAPK/Rac1-dependent pathway has been shown to reduce LPS-as well as vascular endothelial growth factor (VEGF)-induced vascular leakage thereby revealing FVIIa barrier-protective features [bib_ref] Factor VIIa binding to endothelial cell protein C receptor protects vascular barrier..., Sundaram [/bib_ref] [bib_ref] Factor VIIa bound to endothelial cell protein C receptor activates protease activated..., Sen [/bib_ref]. EPCR serves as a receptor for FVIIa on endothelial cells and has recently been found to bind FVIIa on human platelets, which have been widely described to play a role in cancer progression [bib_ref] Human platelets express endothelial protein C receptor, which can be utilized to..., Fager [/bib_ref] [bib_ref] Antiplatelet agents for cancer treatment: A real perspective or just an echo..., Wojtukiewicz [/bib_ref]. PAR-1 cleavage by FXa is also highly enhanced in the presence of EPCR, which, similar to APC, dissociates from caveolin-1 after FXa binding [bib_ref] Coagulation factor Xa cleaves protease-activated receptor-1 and mediates signaling dependent on binding..., Schuepbach [/bib_ref]. Unfortunately, FVIIa and FXa/EPCR-mediated PAR-1 activation is not well explored in cancer tissues.
Cancers 2019, 11, x 9 of 18 signaling independent of the protease cleavage site and protein Gi coupling. Interestingly, thrombin also exerts PAR-1-dependent cytoprotective effects by β-arrestin-2-induced recruitment of disheveled 2 (Dvl-2) scaffolding protein and Rac1 signaling.
## Factor vii/viia and factor xa (fxa)
Other serine proteases that may occupy EPCR in a Ca 2+ -dependent mechanism and exert protective effects on the endothelial barrier are factor VII/VIIa and factor X/Xa (FXa), which bind EPCR through the 4-carboxyglutamic acid domains [bib_ref] Factor VIIa induces anti-inflammatory signaling via EPCR and PAR1, Kondreddy [/bib_ref] [bib_ref] The endothelial protein C receptor supports tissue factor ternary coagulation initiation complex..., Disse [/bib_ref] [bib_ref] Coagulation factor Xa cleaves protease-activated receptor-1 and mediates signaling dependent on binding..., Schuepbach [/bib_ref]. The cytoprotective properties mediated by FVIIa binding to EPCR depend on PAR-1 and β-arrestin-1 activation [fig_ref] Figure 3: Factor VIIa-mediated activation of endothelial protein C receptor [/fig_ref]. FVIIa-treated ECs in vitro had reduced expression of cellular adhesion molecules and adherence of monocytes to ECs. These effects were dependent on PAR-1-mediated suppression of tumor necrosis factor α (TNF-α). PAR-1 antibodies or small interfering RNA abolished TNF-α-induced activation of ERK1/2, p38 MAPK, JNK, NF-κB, and C-Jun factors as well as lipopolysaccharide (LPS)-mediated in vivo inflammatory responses in lungs of mice overexpressing EPCR. Additionally, FVIIa in the PAR-1/MAPK/Rac1-dependent pathway has been shown to reduce LPS-as well as vascular endothelial growth factor (VEGF)-induced vascular leakage thereby revealing FVIIa barrier-protective features [bib_ref] Factor VIIa binding to endothelial cell protein C receptor protects vascular barrier..., Sundaram [/bib_ref] [bib_ref] Factor VIIa bound to endothelial cell protein C receptor activates protease activated..., Sen [/bib_ref]. EPCR serves as a receptor for FVIIa on endothelial cells and has recently been found to bind FVIIa on human platelets, which have been widely described to play a role in cancer progression [bib_ref] Human platelets express endothelial protein C receptor, which can be utilized to..., Fager [/bib_ref] [bib_ref] Antiplatelet agents for cancer treatment: A real perspective or just an echo..., Wojtukiewicz [/bib_ref]. PAR-1 cleavage by FXa is also highly enhanced in the presence of EPCR, which, similar to APC, dissociates from caveolin-1 after FXa binding [bib_ref] Coagulation factor Xa cleaves protease-activated receptor-1 and mediates signaling dependent on binding..., Schuepbach [/bib_ref]. Unfortunately, FVIIa and FXa/EPCR-mediated PAR-1 activation is not well explored in cancer tissues.
## Tissue factor
Tissue factor is the main procoagulant protein expressed on cancer cells, and is responsible for thrombin generation in the tumor microenvironment independently of blood coagulation. Thrombin is pivotal for APC activation. Apart from thrombin generation, TF plays a role in EPCR-mediated PAR-1 activation via complex formation with coagulation factors TF/VIIa/Xa, and is regarded as an EPCR cofactor [bib_ref] The endothelial protein C receptor supports tissue factor ternary coagulation initiation complex..., Disse [/bib_ref]. The activation of p44/42 MAPK signaling induced by TF/FVIIa/FXa on activated endothelial cells is EPCR-dependent.
There is evidence that APC/EPCR/PAR-1 cooperate in the up-regulation of TF to increase cancer cell invasiveness. APC also mediates degradation of TFPI, an inhibitor of TF-dependent coagulation, and through this "procoagulant" pathway surprisingly regulates the pro-metastatic potential of thrombin generated due to TF expression on cancer cells [bib_ref] Activated protein C up-regulates procoagulant tissue factor activity on endothelial cells by..., Reto [/bib_ref].
## Tissue factor
Tissue factor is the main procoagulant protein expressed on cancer cells, and is responsible for thrombin generation in the tumor microenvironment independently of blood coagulation. Thrombin is pivotal for APC activation. Apart from thrombin generation, TF plays a role in EPCR-mediated PAR-1 activation via complex formation with coagulation factors TF/VIIa/Xa, and is regarded as an EPCR cofactor [bib_ref] The endothelial protein C receptor supports tissue factor ternary coagulation initiation complex..., Disse [/bib_ref]. The activation of p44/42 MAPK signaling induced by TF/FVIIa/FXa on activated endothelial cells is EPCR-dependent.
There is evidence that APC/EPCR/PAR-1 cooperate in the up-regulation of TF to increase cancer cell invasiveness. APC also mediates degradation of TFPI, an inhibitor of TF-dependent coagulation, and through this "procoagulant" pathway surprisingly regulates the pro-metastatic potential of thrombin generated due to TF expression on cancer cells [bib_ref] Activated protein C up-regulates procoagulant tissue factor activity on endothelial cells by..., Reto [/bib_ref].
The complexity of APC/EPCR/PAR-1 and TF interactions was shown in an animal model of malignant pleural mesothelioma (MPM) [bib_ref] Influence of endothelial cell protein C receptor on breast cancer development, Keshava [/bib_ref]. Intrapleural injection of mesothelioma cells expressing PAR-1 and TF, but that were EPCR negative, led to larger tumor growth, while ectopic expression of EPCR in aggressive mesothelioma cells attenuated their proliferative potential. Interestingly, EPCR inhibition in non-aggressive MPM cells that overexpressed TF increased their tumorgenicity, thus demonstrating the protective role of EPCR in cancer promotion.
## Thrombomodulin, tm
Thrombomodulin is a tumor suppressive protein that when expressed in several types of cancer cells decreased tumor invasiveness and improved patient survival [bib_ref] Thrombomodulin Influences the Survival of Patients with Non-Metastatic Colorectal Cancer through Epithelial-To-Mesenchymal..., Chang [/bib_ref] [bib_ref] Thrombomodulin reduces tumorigenic and metastatic potential of lung cancer cells by up-regulation..., Zheng [/bib_ref]. The level of soluble TM shed from cancer cells to the plasma correlates with disease advancement and is associated with poor prognosis [bib_ref] Thrombomodulin suppresses invasiveness of HT1080 tumor cells by reducing plasminogen activation on..., Higuchi [/bib_ref] [bib_ref] Thrombomodulin (TM) in tumor cell differentiation and periphery blood immune microenvironment in..., Song [/bib_ref]. Expression of TM on cancer cells was associated with better differentiation in patients with squamous cell carcinoma of the oral cavity [bib_ref] Thrombomodulin (TM) in tumor cell differentiation and periphery blood immune microenvironment in..., Song [/bib_ref]. Furthermore, TM decreases PAR-1 and NF-κB activation and significantly suppresses pancreatic cancer tumor growth.
## Hematopoietic stem cells
Gur and Cohen [bib_ref] PAR1 signaling regulates the retention and recruitment of EPCR-expressing bone marrow hematopoietic..., Gur-Cohen [/bib_ref] [bib_ref] Regulation of long-term repopulating hematopoietic stem cells by EPCR/PAR1 signaling, Gur-Cohen [/bib_ref] demonstrated that long-term repopulating hematopoietic stem cells (LT-HSCs) originating from adult murine bone marrow (BM) and that expressed EPCR and PAR-1 had the highest repopulation and self-renewal potential. The EPCR-positive LT-HSCs accumulate in close proximity to BM endothelial subpopulations expressing elevated thrombomodulin, which in the presence of thrombin induces APC-mediated PAR-1 activation. EPCR/PAR-1 signal activation and restriction of nitric oxide (NO) production induces LT-HSC BM repopulation, retention, and chemotherapy resistance to induce anti-apoptotic effects and prolong cell survival.
PAR-1-associated signaling in LT-HSCs also results in their recruitment to the blood circulation and allows them to retain developmental potential. The PAR-1-dependent signaling pathway results in nitric oxide (NO) reduction leading to Cdc42 downregulation and increased adhesion associated with VLA4. The integrin VLA4 binds fibronectin and VCAM-1 and is crucial for CXCL12/CXCR4-mediated BM HSC retention. The balance between retention of LT-HSCs in BM or trafficking to the blood is dependent on PAR-1 activation by APC or thrombin, levels of NO, and surface CXCR4 expression.
Additionally, the thrombin-PAR-1-activated signaling cascade plays a role in BM cell trafficking under conditions of stress or inflammation, when increased levels of thrombin are generated and rapidly enter the BM. This induces PAR1-dependent hematopoietic stem and progenitor cell mobilization, motility and recruitment of LT-HSCs into the bloodstream.
Summing up, along with pro-and anticoagulant signaling, which has also been widely described for malignancies, factors of the coagulation system cooperate within the bones in non-hemostatic pathways to regulate bone-remodeling processes, stem cell interactions, BM cell retention or release to the blood stream during both steady-state and inflammatory conditions. The expression of PAR-1 has been detected in acute myeloid leukemia cells as well as leukemia stem cells. Therefore, the hypothesis that EPCR and PAR-1-dependent signaling is implicated in development of hematopoetic malignancies is justified, especially in light of the fact that myeloid cells such as osteoclasts express TF and prothrombin.
## Microbiome and epcr/pars interactions
Microbial factors and infection-associated inflammatory processes have aroused great interest with respect to carcinogenesis. Proteases are essential for the normal functioning of bacteria and may initiate pathological events such as increased enteric permeability and inflammation by acting through protease receptors [bib_ref] Enteric bacterial proteases in inflammatory bowel disease-pathophysiology and clinical implications, Carroll [/bib_ref]. Streptococcus pyogenes and Staphylococcus aureus on the skin induce inflammatory reactions by producing proteases that activate PARs on keratinocytes [bib_ref] Protease-armed bacteria in the skin, Koziel [/bib_ref]. Epidemiologic studies have reported an association between periodontitis and Porphyromonas gingivalis (Pg) infection with increased risk of oral squamous cell carcinoma (OSCC) and orodigestive cancer death. Interestingly Pg promotes an oral squamous cell carcinoma through the PARs-activated signaling pathway that results in increased MMP9 expression and thus cancer invasion [bib_ref] Arginine-specific protease from Porphyromonas gingivalis activates protease-activated receptors on human oral epithelial..., Lourbakos [/bib_ref]. The newest study reported that Pg induces host molecular changes that promote EMT (epithelial-mesenchymal-transition), a process highly regulated by PAR-1 activity in cancer cells and responsible for dissemination [bib_ref] Human Primary Epithelial Cells Acquire an Epithelial-Mesenchymal-Transition Phenotype during Long-Term Infection by..., Lee [/bib_ref] [bib_ref] Protease-activated receptor-1 (PAR1) promotes epithelial-endothelial transition through Twist1 in hepatocellular carcinoma, Xiao [/bib_ref].
There are reports that EPCR is also involved in EMT among mammary stem cells, a subpopulation of cells determined to have greater invasive potential [bib_ref] Identification of multipotent mammary stem cells by protein C receptor expression, Wang [/bib_ref] [bib_ref] Multiple lineages of human breast cancer stem/progenitor cells identified by profiling with..., Hwang-Verslues [/bib_ref]. Further studies are needed to explore the role of the APC/EPCR/PAR-1 cascade in this part of the metastatic process.
Increased permeability of endothelium or intestine as a result of the action of bacterial proteases is an event similar to that induced by metastasizing cancer cells for dissemination. That is why the findings that the same receptors are involved in both processes are so exciting and require extensive investigation [bib_ref] Enteric bacterial proteases in inflammatory bowel disease-pathophysiology and clinical implications, Carroll [/bib_ref] [bib_ref] Protease-armed bacteria in the skin, Koziel [/bib_ref] [bib_ref] Arginine-specific protease from Porphyromonas gingivalis activates protease-activated receptors on human oral epithelial..., Lourbakos [/bib_ref].
The cyclic dipeptide with documented anti-inflammatory and anti-cancer activity that is naturally produced by bacteria, fungi, marine sponges, gorgonian and red algae has been reported to inhibit tumor necrosis factor (TNF)-α, and interleukin (IL)-1β-induced EPCR shedding in human umbilical vein endothelial cells (HUVECs) [bib_ref] Inhibitory effects of three diketopiperazines from marine-derived bacteria on endothelial protein C..., Lee [/bib_ref].
## Epc/par-1 and neurons
The altered nerve function orchestrated by PAR-1-dependent signaling in the central nervous system and in peripheral nerves has been intensively investigated as an important pathogenic factor of many inflammatory and neurodegenerative diseases [bib_ref] Protease signaling through protease activated receptor 1 mediate nerve activation by mucosal..., Buhner [/bib_ref] [bib_ref] Protease-activated receptor-1 activation by granzyme B causes neurotoxicity that is augmented by..., Lee [/bib_ref]. The influence of inflammation in cancer is undeniable. Thus the molecular connections between neurotoxicity and malignancy are interesting [bib_ref] Vanilloid receptor-1 regulates neurogenic inflammation in colon and protects mice from colon..., Vinuesa [/bib_ref]. Data show that neurons within the inflamed central nervous system and peripheral tissues (lung, intestine) overexpress PAR-1, and in this way are exposed to neurotoxic inflammatory mediators resulting in brain atrophy or enteritis [bib_ref] Inhibitory effects of three diketopiperazines from marine-derived bacteria on endothelial protein C..., Lee [/bib_ref] [bib_ref] Protease signaling through protease activated receptor 1 mediate nerve activation by mucosal..., Buhner [/bib_ref] [bib_ref] Thrombin and trypsin directly activate vagal C-fibres in mouse lung via protease-activated..., Kwong [/bib_ref].
Neurons are another cell population where the activation of PAR-1 either by thrombin or the anticoagulant APC has differential effects [bib_ref] The anticoagulant activated protein C (aPC) promotes metaplasticity in the hippocampus through..., Maggio [/bib_ref].
It has been documented that thrombin/PAR-1/N-methyl-D-aspartate receptors (NMDAR) regulate long-term potentiation (LTP) in the hippocampus thus causing persistent strengthening of synapses and production of long-lasting increases in signal transmission between neurons. Activated protein C may increase synaptic plasticity in a mechanism that requires the activation of S1P1R and intracellular Ca 2+ stores. These APC/EPCR/PAR-1-mediated effects on hippocampus functions may be useful in the therapy of neurodegenerative disease and may also be due to antiapoptotic effects exerted by APC in neurons [bib_ref] Activated protein C prevents neuronal apoptosis via protease activated receptors 1 and..., Guo [/bib_ref]. Moreover, there are findings that APC reduces the nuclear level of NF-κB p65 in hippocampal neurons from glutamate-induced excitotoxicity via binding to EPCR and subsequent PAR-1-activated signaling [bib_ref] Activated protein C prevents glutamate-and thrombin-induced activation of nuclear factor-κB in cultured..., Gorbacheva [/bib_ref]. Although there is a lack of direct evidence of APC/APCR/PAR-1-dependent signaling in brain tumors the findings of their reciprocal interactions in neuronal tissue suggest that such processes may exist.
## Treatment
The fact that the results of in vitro and in vivo cancer models assessing the role of APC/EPCR/PAR-1 in tumor progression are contradictory makes the discoveries of therapeutic drugs very difficult. The basic question of whether there are inhibitors or activators of the APC/EPCR/PAR-1 cascade that inhibit cancer development and dissemination remains unanswered.
The APC/EPCR/PAR-1 axis has been demonstrated to reduce organ damage in models of multiple pathologies like stroke, sepsis and autoimmune disease so that a clinical trial (NCT00533546) with APC treatment in ischemic stroke has been initiated (http://www.clinicaltrials.gov). Moreover, APC has been approved by the Food and Drug Administration (FDA) for the treatment of severe sepsis due to results from the PROWESS clinical trial that drotrecogin alfa (DrotAA), recombinant human APC, reduces mortality in patients with severe sepsis [bib_ref] The cytoprotective protein C pathway, Mosnier [/bib_ref] [bib_ref] Cytoprotective-selective activated protein C therapy for ischaemic stroke, Mosnier [/bib_ref] [bib_ref] Protective signaling by activated protein C is mechanistically linked to protein C..., Feistritzer [/bib_ref] [bib_ref] Efficacy and safety of recombinant human activated protein C for severe sepsis, Bernard [/bib_ref]. Although DrotAA has been withdrawn from ischemic patient therapy, the clinical trials with APC variants (e.g., 3K3A-APC) are conducted based on cytoprotective, not anticoagulant (like with DrotAA) properties of APC [bib_ref] Human recombinant activated protein C for severe sepsis, Martí-Carvajal [/bib_ref].
Based on the cytoprotective effects exerted by APC in the above-mentioned diseases it has been suggested that exogenous APC administration might also serve as a new therapeutic option for cancer patients to protect from or treat metastasis. Unfortunately, the very short half-life of APC, approximately 15-min, requires continuous infusion of the drug, which may be problematic for cancer patients.
Additionally, the bleeding complication frequently observed during APC therapy may limit use of this drug in practice. The in vitro and in vivo studies that demonstrate the potential of APC and PAR1-based peptides (e.g., TR47 given at a dose of 125 µg/animal) to enhance endothelial barrier function, to decrease transendothelial migration of malignant cells, and finally reduce metastasis to distant organs, has stimulated a search for activators of the APC/EPCR/PAR-1 axis that might be a valuable therapeutic approach to prevent metastasis.
It is now accepted that thrombin facilitates the interaction between tumor cells and platelets and endothelial cells, thereby enabling malignant cells to seed and metastasize. Therefore, this is an important target for therapeutic research [bib_ref] The involvement of thrombin in the pathogenesis of glioblastoma, Krenzlin [/bib_ref]. Dabigatran, a thrombin inhibitor, delivered in combination with gemcitabine, inhibited primary tumor growth and prevented tumor cell dissemination from pancreatic cancer in mice [bib_ref] Dabigatran potentiates gemcitabine-induced growth inhibition of pancreatic cancer in mice, Shi [/bib_ref]. Additionally, there are data that inhibition of APC/EPCR or the PAR-1 cascade is effective for killing cancer. Inhibition of the thrombin-induced PAR-1/Rho GTPase signaling pathway activation by ALEX1 (a novel tumor suppressor gene, and member of the armadillo family) limits gastric cancer metastasis [bib_ref] ALEX1, a novel tumor suppressor gene, inhibits gastric cancer metastasis via the..., Pang [/bib_ref]. Importantly, the expression of ALEX1 may be decreased in a variety of solid tumors correlated with clinical features, such as cell differentiation and staging. The anticancer activity of some chemotherapeutic agents has also been shown to be associated with membrane receptors [bib_ref] Doxycycline inhibits breast cancer EMT and metastasis through PAR-1/NF-κB/miR-17/E-cadherin pathway, Zhong [/bib_ref]. Namely, doxycycline inhibits breast cancer migration and metastatic potential through the PAR-1/NF-κB/miR-17/E-cadherin pathway.
The newest study reporting the inhibitory role of PAR-1 in pancreatic ductal cancer growth and all studies presenting an inhibitory role for EPCR in cancer progression suggests that the contribution of APC/EPCR/PAR-1 signaling in cancer biology differs between tumor subtypes. Thus, targeting should be applied with care as this area still remains unexplored.
# Conclusions
The interactions between EPCR and PAR-1 are complex and pleiotropic. Understanding of the mechanisms induced by these receptors improved considerably over the last several years, which creates a chance for future pharmacological interventions that affect receptor system.
[fig] Figure 1: Protein C (PC) activation mediated by thrombin, thrombomodulin (TM) and EPCR (endothelial protein C receptor) and its diverse biological functions (anticoagulant and cellular effects). [/fig]
[fig] Figure 2: Distinct mechanisms of PAR-1 (protease-activated receptor 1) activation by thrombin-and activated protein C (APC)/endothelial protein C receptor (EPCR) complex resulting in different biological effects. MAPK, mitogen activated protein kinase, proteins Rho, Rac, Akt, Ras. [/fig]
[fig] Figure 3: Factor VIIa-mediated activation of endothelial protein C receptor (EPCR) and protease-activated receptor 1 (PAR-1) leading to suppression of tumor necrosis factor alfa (TNF-α)-dependent signaling. ERK-extracellular signal-regulated kinase; MAPK-mitogen activated protein kinase, c-Jun, JNK-N-terminal kinase. [/fig]
[table] Table 2: Distinct cellular effects mediated by endothelial protein C receptor (EPCR) in different cancer settings. [/table]
|
Double intussusception secondary to Meckel’s diverticulum in a seventeen-year-old female: a case report
# Introduction
In 1809, Meckel´s diverticulum (MD) was described as a vestigial embryonic remnant of the omphalomesenteric duct. It is more commonly identified in children, as adults are rarely symptomatic. It is the most common congenital malformation of the gastrointestinal tract and represents 2-4% of the general population. It is, therefore, most often discovered incidentally in adults with several potential management approaches. Herein, we report a case of double ileal intussusception that was successfully treated by laparoscopic-assisted mini-laparotomy after immediate exploratory laparoscopy.
## Patient and observation
A 17-year-old female presented to the emergency department with vomiting and acute peri-umbilical abdominal pain for 12 hours. She had no symptoms of obstipation. Her past medical history was unremarkable. On physical examination: the patient had normal vital signs and an unremarkable abdominal examination with no tenderness. Laboratory workup showed an elevated WBC count of 20.03 10 9 /L, hemoglobin level of 15.0 g/dL, platelet count 354 10 9 /L, and a C-reactive protein of <3.0 mg/L. Ultrasound examination showed a suprapubic intussusception measuring 3.2cm in diameter and extending approximately 8cm in length with significant liquid stasis proximally. Otherwise, no intra-abdominal, pelvic, or retroperitoneal abnormalities were identified. After informed consent was obtained, we took the patient for urgent exploratory laparoscopy. We discovered a double ileal intussusception with the presence of multiple mesenteric lymph nodes without lymphadenopathy. The first was at the terminal ileum followed by a second intussusception, which was 30cm from the terminal ileum. The distal intussusception of 5cm in length waseasily reduced without inadvertent injury. The second intussusception of 15cm in length had signs of ischemia (color change and venous congestion) with a high risk of perforation and its reduction via the laparoscopic route was difficult. Therefore, we decided to reduce it through a minilaparotomy via Pfannenstiel incision. After reduction of the 5cm MD, the bowel perfusion improved with return of normal color and contractility. Resection of MD is indicated upon discovery in any circumstance, especially in the setting of an intussusception with the MD as a pathologic lead point and . The patient was discharged on the second post-operative day following an uncomplicated hospital course. Histological examination revealed the presence of gastric heterotopia with no malignancy and confirmed the diagnosis of Meckel´s diverticulum. One month later, the follow up appointment was unremarkable and the patient was cleared for unrestricted activity.
# Discussion
MD is most often asymptomatic and discovered incidentally intra-operatively, or on an imaging study. However, up to 7% of cases present with complications, which include perforation, obstruction, hemorrhage, neoplasia, or fistulac. There is no gold standard imaging study in the absence of complications in adults. A contrast-enhanced CT scan may miss the diagnosis of MD in the absence of a complication. In the presence of perforation or obstruction, a CT scan is the best imaging modality. Ultrasonography has two advantages over CT scan. It is sensitive in the presence of complications and avoids radiation exposure [7].
Ectopic mucosa tends to develop in MD with the majority being gastric in up to 26% of cases.
Other types of ectopic tissue such as pancreatic, duodenal, colonic, endometrial and hepato-biliary can be seen within the MD. Furthermore, the risk of metaplasia and neoplasia is not negligible as the mean annual incidence rate of MD cancer was 1.44 per 10 million populations between 1973 -2006 (± standard deviation of 1.12). Twothirds of these cases involve carcinoid metaplasia with adenocarcinoma, sarcoma, stromal tumors and lymphoma also described. The removal of the risk of malignant degeneration is a known benefit to resection, which essential in our case.
In the literature, there are multiple surgical options for MD, which are segmental small bowel resection with primary anastomosis, wedge resection, or tangential stapling. It is challenging to perform wedge resection or tangential stapling in cases of perforation or hemorrhagic ulceration. Linear tangential stapling, or wedge resectionare practiced even if there is a small risk of leaving heterotopic tissues. In our case, we performed a mini-laparotomy followed by linear tangential stapling for the proximal intussusception due to the fact that; the reduction was difficult and risked traumatizing the bowel laparoscopically and benefit from palpation to ensure that all thickened tissues were removed.
Finally, various cases of MD are discovered incidentally. Therefore, is expectant non-operative management a reasonable alternative in adult patients? Some studies advocate for abstention for two main arguments: older individuals have fewer complications of MD, and higher risk of postoperative complications . On the other hand, routine resection is advocated, as complications from MD itself are higher than postoperative complications [17] and the association of cancer potential. Moreover, a retrospective study at the Mayo Clinic of 1476 patients investigated the risk factors of complications in MD and they concluded a combination of age less than 50 years old, male gender, length more than 2cm and macroscopic abnormalities increased complications up to 70%. Thus, promoting prophylactic excision. It is debatable to operate cases of MD with incidental findings. Nevertheless, we believe that resection has better potential to decrease the risk of malignant degeneration and the occurrence of complications, especially in the presence of risk factors nowadays, with the advancement of imaging modalities and minimally invasive surgery.
# Conclusion
MD is rarely manifests in the adult population as these are predominantly asymptomatic. In a patient with atypical abdominal symptoms, however, imaging findings may alert general surgeons to the possibility of this diagnosis. It is evident that exploratory laparoscopy is vital in cases presenting with a complication of MD, yet the modality of choice in treatment is dependent on surgeons' experience and presenting complications.
# Authors' contributions
All authors discussed the case presentation and contributed to the final manuscript. Feras Sendy wrote the manuscript with support and input from Thibaut D´escrivan, Anthony Joubert and Nicolae Fetche; Nicolae Fetche helped supervise the project. All the authors have read and agreed to the final manuscript. |
Price, Availability, and Youth Obesity: Evidence From Bridging the Gap
After a decade of analyzing environmental influences on substance use and its consequences among youth in the United States, the Robert Wood Johnson Foundation's Bridging the Gap program has begun studying the effect of environmental factors on youth physical activity, diet, and weight outcomes. Much of this research has focused on access to food, as reflected by availability and price. Program researchers have documented disparities in access to healthy foods and opportunities for physical activity; healthier food outlets and opportunities for physical activity are relatively less available in communities with lower income and larger proportions of racial/ethnic minority populations. They also have found that healthier environments are associated with more fruit and vegetable consumption, more physical activity, lower body mass index, and reduced likelihood of obesity among youth.
# Introduction
The Robert Wood Johnson Foundation's Bridging the Gap program (BTG), created in 1997, has studied the effect of policies, programs, and other environmental influences on adolescent alcohol, tobacco, and illicit drug use and related outcomes in the United States. BTG is an interdisciplinary, multisite initiative that brings together a mix of original and archival data at multiple levels of social organization, including schools, communities, and states. During its first decade, BTG helped to build the evidence base on the effects of prices, control policies, and other environmental influences on youth substance use and abuse among youth; this evidence has been a key factor in the adoption of policies that have reduced youth tobacco and other substance use during this time. BTG focuses on youth because these behaviors largely begin during adolescence and because of the lasting effect of effective interventions to reduce initiation and uptake during adolescence.
As recognition of the obesity epidemic has increased, researchers, policy makers, and public health professionals have focused more attention on the role of environmental factors as determinants of healthy and unhealthy eating, physical activity and inactivity, and weight outcomes. BTG's research focus has turned to applying the approach it successfully used for a decade to looking at the effect of relevant environmental factors on youth physical activity, diet, and weight outcomes. The conceptual framework for BTG's research on adolescent obesity draws on multiple disciplines and emphasizes environmental factors and their interaction with individual and social factors in affecting diet, physical activity, and weight status. Much of BTG's initial work on youth obesity has focused on access to food and physical activity outlets, as reflected by availability and price. We provide a brief overview of BTG's research on environmental factors and related implications for policies to reduce obesity and its consequences.
contribute to overweight and obesity concentrated on documenting differences in the availability of food and physical activity-related outlets across communities. Most of this research used data from the Dun & Bradstreet MarketPlace Database, compiled in 2000 by zip code to analyze differences in availability of these outlets based on community characteristics, including racial/ethnic composition of the population and median household income. Measures of population characteristics for the 28,050 zip codes in which nearly all of the population resided in 2000 were taken from that year's census database.
Measures for the density of food-related outlets based on standard industry classification codes were developed for chain supermarkets, independent supermarkets, convenience stores, grocery stores, fast-food restaurants, and full-service restaurants. Comparable measures for the density of physical activity-related outlets were developed for membership and nonmembership health clubs, spas, and other fitness facilities; membership sports and recreation clubs; and dance studios, schools, and public dance halls. One study used measures of publicly available opportunities for physical activity, including sports areas, parks and green spaces, playgrounds, public pools and beaches, and bike paths collected around nationally representative samples of US secondary schools in 2002 and 2003 as part of BTG's extensive original data collection efforts (5).
This research consistently documented racial/ethnic and income-related disparities in the availability of healthy food and opportunities for physical activity across communities. For example, availability of free opportunities for physical activity was lower in communities with lower income levels and in those where a greater percentage of the local population was African American (5). Similar differences were observed for commercial physical activity-related outlets, which were also less available in communities with larger proportions of Latinos.
Fewer chain supermarkets were in lower-income communities; predominantly African American communities had approximately half as many chain supermarkets as predominantly white communities, and predominantly Latino communities had approximately one-third as many as did largely non-Latino communities. In contrast, smaller groceries and independent supermarkets were more available in communities with relatively larger minority populations. Previous research has shown that larger food stores (eg, supermarkets) are more likely to stock healthful foods than smaller food stores (eg, groceries and convenience stores) (10,11) and to offer foods at a lower cost. Similarly, absolute and relative (as a share of all restaurants) availability of fast-food restaurants is higher in lower-and middle-income communities than in highincome communities. Restaurants of all types were less available in predominantly minority communities, but a significantly higher proportion of restaurants were fastfood restaurants in these communities. Finally, fast-food restaurants and convenience stores are readily available around US secondary schools, particularly high schools; the highest availability is around schools in larger cities and lower-income neighborhoods.
Given the evidence that overweight and obesity are more prevalent among lower-income and minority populations, including adolescents, BTG's research highlighting community-level disparities in access to healthier food outlets and opportunities for physical activity suggests a link between access and weight outcomes. BTG's more recent research has turned to establishing relationships between access (both physical access, as reflected by availability, and economic access, as reflected by prices), healthy eating, physical activity, and obesity among adolescents.
## How availability of food and opportunities for physical activity influence health behaviors among adolescents
BTG's research assessing the effect of availability of food and physical activity-related outlets on adolescents' eating, physical activity, and weight outcomes links the business list data to outcomes from the annual Monitoring the Future (MTF) surveys of US secondary school students. These are nationally representative surveys of approximately 45,000-50,000 8th-, 10th-, and 12th-grade students conducted each spring in approximately 420 schools by the University of Michigan's Institute for Social Research. These surveys are a key source of information on adolescent tobacco, alcohol, and illicit drug use in the United States, and they yield data on related attitudes, beliefs, and perceived access, as well as key socioeconomic and demographic information.
Since 1986, the MTF surveys have also collected information on self-reported height and weight, which can be used to calculate body mass index (BMI) and indicators for youth obesity. These surveys also collect information on dietary habits (eg, frequency of eating fresh fruit and vegetables), physical activity (eg, frequency of exercise, vigorous exercise and sports participation, including participation on a school team), sedentary behavior (eg, frequency of television watching and leisure time computer use), and other behaviors potentially related to weight outcomes (eg, sleep patterns). Consistent with available data from other nationally representative surveys such as the National Health and Nutrition Examination Surveyand the Youth Risk Behavior Surveillance System, data from the MTF surveys show generally increasing BMI and prevalence of obesity among US adolescents, coupled with rising screen time and reductions in physical activity, healthy eating, and sleep.
## Effect on adolescent physical activity
Data from 1997 through 2003 on the availability of commercial physical activity-related outlets (fitness facilities, membership sports and recreation clubs, and dance studios, halls, and schools) have been linked to 8th-, 10th-, and 12th-grade student self-reports of physical activity (frequency of participation in sports/athletics and exercise). Outlet density measures were matched by zip code, based on the zip code of each school participating in the MTF study during these years. Living in a community with greater availability of commercial physical activity outlets was associated with increased reports of frequent participation in sports and exercise, and larger associations were observed for girls and for older students (12th graders). Simulations conducted using these estimates implied, for example, that, for 12th-grade girls, living in a community with a relatively high availability of these outlets (8 facilities) was associated with a 9% higher prevalence of frequent vigorous exercise than in a community with limited availability (1 facility); for 12th-grade boys, the comparable higher prevalence was 6%.
## Effect on adolescent healthy eating and weight
Using 8th-and 10th-grade students' self-reported height and weight to calculate BMI and an indicator for obesity, researchers (24) examined the associations between food store availability and adolescent weight outcomes for the years 1997 through 2003. Outlet density measures for supermarkets (chain and independent), grocery stores, and convenience stores were matched to the MTF survey data based on the zip code of participating schools. Differential effects are expected by store type, given previous research on the relationship between store type and availability and prices for healthier foods. Indeed, the results from the MTF study show that more availability of chain supermarkets was associated with lower BMI and reduced likelihood of obesity among adolescents, whereas more availability of convenience stores was associated with higher BMI and higher probability of being obese. Both associations were stronger for adolescents in households where their mother worked full-time, and the association for chain supermarkets was stronger for African American adolescents.
In contrast, researchers found little association between fast-food and full-service restaurant availability and adolescent diet or weight outcomes. Using similarly calculated measures of restaurant outlet density matched to the MTF survey data based on the zip codes of participating schools, they found no significant associations between either fast-food or full-service restaurant availability and any of the outcomes examined, except for a small positive association between the availability of fullservice restaurants and frequency of fruit and vegetable consumption.
## How food prices influence eating behavior and weight among adolescents
BTG researchers have similarly examined the associations between food prices and adolescent diet and weight outcomes by using data from the American Chamber of Commerce Research Association (ACCRA) quarterly cost-of-living index reports matched to the MTF survey data. The ACCRA reports include prices for goods and services, including 3 fast foods (McDonald's Quarter Pounder with cheese, Pizza Hut or Pizza Inn pizza, and a KFC or Church's Chicken chicken meal) and 7 fruits and vegetables (bananas, canned peaches, canned sweet peas, canned tomatoes, lettuce, potatoes, and frozen corn). BTG researchers used these price data to create 2 price indices, 1 for fast-food prices (a simple average of the 3 prices) and 1 for fruits and vegetables (a weighted average based on household expenditure shares). ACCRA (now the Council for Community and Economic Research) collects prices for about 300 metropolitan statistical areas each quarter; prices were matched to MTF schools based on the prices in the metropolitan statistical area nearest each school.
Researchers used MTF 1997-2003 data on 8th-and 10thgrade students to analyze the effect of prices on frequency of fruit and vegetable consumption, BMI, and an indicator for obesity. As expected given economic theory, they found that lower prices for fruits and vegetables were associated with more fruit and vegetable consumption, while lower prices for fast foods were associated with less fruit and vegetable consumption. Consistent with the evidence that healthier eating results in better weight outcomes, they found that lower fruit and vegetable prices were associated with lower BMI, and lower fast-food prices were associated with higher BMI. Finally, they found that lower fast-food prices were also associated with an increased likelihood of obesity. Their estimates imply that a 10% increase in fast-food prices would raise the probability of frequent fruit and vegetable consumption by 3%, reduce BMI by 0.4%, and lower the likelihood of obesity by 5.9% among youth.
A subsequent BTG studylooked at the differential effect of prices on youth at different BMI levels and explored whether prices had more or less of an effect on youth at increased risk for obesity (those at higher BMIs). Using quantile regression methods applied to the same data, researchers found that associations between fast-food and fruit and vegetable prices were strongest for youth at the upper end of the BMI distribution. For example, they found that the association between fast-food prices and BMI was approximately 4 times greater for youth above the 90th percentile than it was for the overall sample of youth, implying that a 10% increase in fastfood prices would lower BMI by 10% to 11% among youth who weighed above the 90th percentile. Similarly, the association between fruit and vegetable prices and BMI was approximately 5 times greater for youth who weighed above the 95th percentile than for the overall sample of youth. Estimates implied that a 10% drop in fruit and vegetable prices would lower BMI by 5% to 6% in this group.
# Discussion
BTG researchers have documented that healthier food outlets and opportunities for physical activity are less available in communities with lower incomes and larger proportions of racial/ethnic minorities. These findings for youth across the nation are consistent with observations in studies that focused on 1 or a few communities. Likewise, BTG researchers have begun to demonstrate that healthier environments are associated with more fruit and vegetable consumption, more activity, lower BMI, and reduced likelihood of obesity.
These studies, however, have several limitations. For example, some studies have focused on food and physical activity environments only and have not taken into account the variety and range of environments to which a person is exposed. Moreover, because people of different ages have different interactions with their environments, these studies of youth have produced findings that do not translate to all age groups. In addition, these studies are cross-sectional, demonstrating associations and not causation. Also, the business list data and price data used to assess physical and economic availability may be subject to some measurement errors. Finally, these studies used BMI and obesity measures based on self-reported height and weight, which may be inaccurately reported.
The empirical findings presented here (based on US data) and similar study findings (based on data from other countries) have several implications for policies that could be used to curb the rising obesity epidemic. Governments can use a variety of policies to promote healthier environments, from zoning and land use regulations to fiscal policies. One of the fiscal tools that governments can use to promote development of new supermarkets, fitness facilities, and others is to provide tax incentives to businesses that are considering opening facilities in underserved communities; several have begun to do this to attract supermarkets to "food deserts" -areas with little or no ready access to relatively healthy foods. Similarly, governments can directly provide opportunities for physical activity by investing in local recreation centers, parks, and other resources in underserved areas.
Governments also can use taxes and subsidies to alter the relative prices of foods and beverages. Most food subsidy programs to date aim to improve food security by subsidizing food purchases by low-income households (eg, food stamps) and have few restrictions on what products can be purchased. Subsidies that target healthy foods would reduce prices for these products relative to those of less healthy foods and likely result in healthier eating. Similarly, governments could tax less healthy foods and beverages to raise the relative prices of these products and discourage their consumption. Forty states already impose modest sales taxes on at least 1 soft drink, candy, or snack item. Higher taxes that raise the prices of less healthy products would likely reduce the consumption of these products and improved weight outcomes. In addition to promoting healthier eating, such taxes could generate new revenues that could be used to support other efforts to promote healthy eating and increased physical activity, such as subsidizing healthier food products, investing in opportunities for physical activity, and supporting mass-media campaigns to promote healthy eating and increased activity.
As governments adopt policies such as taxes on less healthy products and subsidies for healthier products, investments in food and physical activity outlets, research on the effect of these policies will be needed to support future, evidence-based efforts to curb the obesity epidemic. Particularly useful will be longitudinal studies that assess the effect of these policy changes on people's behavior and on changes in weight outcomes.
The opinions expressed by authors contributing to this journal do not necessarily reflect the opinions of the US Department of Health and Human Services, the Public Health Service, the Centers for Disease Control and Prevention, or the authors' affiliated institutions. Use of trade names is for identification only and does not imply endorsement by any of the groups named above.
AcknowledgmentsWe gratefully acknowledge research support from the Robert Wood Johnson Foundation through Bridging the |
Mosaic Intrachromosomal Triplication of (12)(p11.2p13) in a Patient with Pallister-Killian Syndrome
Pallister-Killian syndrome (PKS) is a rare genetic disorder usually characterized by mosaic tetrasomy of isochromosome 12p detected in cultured fi broblast cells. We describe here a patient with PKS and intrachromosomal triplication of the short arm of chromosome 12. Her karyotype was mos 46,XX,inv trp(12)(p11.2p13)[34]/ 46,XX[16] de novo by conventional cytogenetics and fl uorescent in situ hybridization (FISH) analysis. However, this chromosomal abnormality was not detected from the patient's cultured blood lymphocytes. We report here the third patient with intrachromosomal triplication on the short arm of chromosome 12, presenting a PKS phenotype.
# Introduction
Pallister-Killian syndrome (PKS, OMIM #601803) is a rare sporadic disorder caused by mosaicism for tetrasomy of the short arm of chromosome 12 resulting from an supernumerary isochro-mosome 12p [bib_ref] The pallister mosaic syndrome, Pallister [/bib_ref]. The karyotype in the cultured blood lymphocytes is normal in most cases, but one supernumerary isochromosome 12p is present in high percentage in cultured skin fi broblasts and bone marrow cells of the patients [bib_ref] The use of interphase FISH for prenatal diagnosis of Pallister-Killian syndrome, Mowery-Rushton [/bib_ref]. Supernumerary analphoid inverted duplicated chromosome 12p and a supernumerary ring chromosome consisting of two copies of chromosome 12p have also been reported in the rare cases with PKS [bib_ref] Tetrasomy 12pter-12p13.31 in a girl with partial Pallister-Killian syndrome phenotype, Vermeesch [/bib_ref] [bib_ref] Pallister-Killian syndrome caused by mosaicism for a supernumerary ring chromosome 12p, Yeung [/bib_ref]. Clinical features of this syndrome include; mental retardation, pigmentary skin abnormalities, seizures, prominent forehead with temporal balding, hypertelorism, short nose, short neck, fl at nasal bridge, fl at occiput and macrosomia [bib_ref] Tetrasomy 12p (Pallister-Killian syndrome), Schinzel [/bib_ref] [bib_ref] Pallister-Killian syndrome: rare phenotypic features and variable karyotypes, Liehr [/bib_ref].
Intrachromosomal triplications leading to partial tetrasomies of the certain chromosome 12 regions have been reported in two patients [bib_ref] Intrachromosomal triplication 12p11.22-p12.3 and gonadal mosaicism of partial tetrasomy 12p, Eckel [/bib_ref] [bib_ref] Mosaic tetrasomy 12p with triplication of 12p detected by array-based comparative genomic..., Powis [/bib_ref]. In this report, we present the third patient who has mosaic intrachromosomal triplication on the short arm of chromosome 12.
## Case report
A 9-year-old girl was referred to us because of developmental delay and mental retardation. She was the third child of non consanguineous parents. Both parents were 43 years old at the time of the patient's birth. The family history was remarkable because the child of the patient's paternal uncle had severe mental retardation and a supernumerary inverted duplicated marker chromosome derived from chromosome 15. The patient's siblings were clini-INTRACHROMASOMAL TRIPLICATION 12p cally normal. She was born at term by Cesarian section (birth weight was 6200 gr).
On examination, she could not speak or walk, her weight was 20.5 kg (3-10 percentile) and height was 110 cm (below 3 percentile). She had a prominent forehead, high frontal hair line, low-set ears, sparse eyebrows, hypertelorism, full cheeks, long philtrum, high arched palate, macroglossia, gingival hypertrophy, mandibular prognathism and muscular hypotonia [fig_ref] Figure 1: Frontal view of the patient's head at 9 months of age [/fig_ref]. She also had hypermobile joints and pes equinovarus of the left foot. Cranial computed tomography showed bilateral frontal subdural hygroma [fig_ref] Table 1: Comparison of the patient's clinical fi ndings with a previously reported patient... [/fig_ref]. Her other system examinations were normal. A short-term phytohemagglutinin-stimulated peripheral blood lymphocyte culture was performed according to a standard procedure [bib_ref] Chromosome staining and banding techniques, Benn [/bib_ref]. Analysis of 100 of the patient's lymphocyte metaphases by Gbands by trypsin using Giemsa banding showed the karyotype to be 46,XX [fig_ref] Figure 3: The patient's partial karyotype and an ideogram showing chromosome 12 pair [/fig_ref]. Skin fi broblasts from the patient were cultured in two different fl asks including 2 mL of AmnioGrow (Cytogen GmbH, Bienenweg, Germany) and Chang Medium-D (Irvine Scientifi c, Santa Ana, CA, USA) at 37°C in a CO 2 incubator and harvested by standard methods [bib_ref] Chromosome staining and banding techniques, Benn [/bib_ref]. Intrachromosomal triplication of the p11.2p13 region on one chromosome 12 was found in 68% of the metaphases of the skin fi broblasts. C-band by barium hydroxide using Giemsa analysis showed the triplicated chromosome to be monocentric. The parents' lymphocytes showed a normal karyotype.
Fluorescent in situ hybridization (FISH) analysiswas performed on metaphase plates from the patient using To TelVysion subtelomeric probe set (Vysis Inc., Downers Grove, IL, USA) according to the manufacturer's instructions. Images were recorded using a Zeiss Axioplan epifl uorescence microscope equipped with a CCD camera (Photometrics Sensys, Tucson, AS, USA) and analyzed using MacProbe v4.3 software. The FISH results on 15 metaphases showed two signals for the 12p subtelomeric region on the abnormal chromosome 12 and one on the normal chromosome 12. This analysis confi rmed that the tetrasomy 12p arose by an inverted duplication mechanism. The patient's karyotype was designated as mos 46,XX,inv trp(12)(p11.2p13)[34]/46,XX [bib_ref] Intrachromosomal triplication of 2q11.2-q21 in a severely malformed infant: case report and..., Wang [/bib_ref] according to International System for Human Cytogenetic Nomenclature (ISCN 2009).
# Discussion
Intrachromosomal triplications are rare events and the middle segment usually being inverted in orientation, as in our case [bib_ref] Intrachromosomal triplication 12p11.22-p12.3 and gonadal mosaicism of partial tetrasomy 12p, Eckel [/bib_ref]. In the majority of cases, triplications were interstitial, whereas there is only one published report of triplication of a whole chromosome arm [bib_ref] Mosaic tetrasomy 12p with triplication of 12p detected by array-based comparative genomic..., Powis [/bib_ref]. In most cases, triplications had originated from maternal chromosomes [bib_ref] Mechanism of intrachromosomal triplications 15q11-q13: a new clinical report, Vialard [/bib_ref]. Our case is the second patient with intrachromosomal triplication of a whole chromosome arm.
To the best of our knowledge, only two patients with intrachromosomal triplication of the short arm of chromosome 12 have been presented in the literature. In the article by Eckel et al. [bib_ref] Intrachromosomal triplication 12p11.22-p12.3 and gonadal mosaicism of partial tetrasomy 12p, Eckel [/bib_ref] , the 12p11.22-p12.3 region was triplicated and this region does not cover the critical PKS region, which defi ned the chromosomal region 12pter-p12.3. Therefore, the patient's clinical phenotype was similar to trisomy 12p syndrome rather than the PKS. Unexpectedly, the intrachromosomal triplication was found in all peripheral blood lymphocytes. Also, this intrachromosomal triplication was found in 12% of the mother's peripheral blood lymphocytes. The clinical fi ndings were compatible with the PKS that were reported by Powis et al. [bib_ref] Mosaic tetrasomy 12p with triplication of 12p detected by array-based comparative genomic..., Powis [/bib_ref]. During conventional cytogenetic analyses, triplicated chromosome 12 was observed in 30% of the cultured fi broblasts but not in peripheral blood lymphocytes. However, array-based comparative genomic hybridization analysis showed a triplication of the 12p in the peripheral blood lymphocytes in a low level mosaicism. A parental transmission of the triplication could not be excluded since parental blood samples were not available [bib_ref] Mosaic tetrasomy 12p with triplication of 12p detected by array-based comparative genomic..., Powis [/bib_ref]. However, this has been excluded in our patient because her parents' karyotypes were normal.
Mechanisms for the formation of the intrachromosomal triplication include; i) fusion of the inverted duplicated supernumerary marker chromosome with the normal homologue; ii) unequal crossover or interhomologue translocation followed by the inverted insertion at the former breakpoint junction; iii) two U-type exchanges among three chromatids [bib_ref] Intrachromosomal triplication of 15q11-q13, Schinzel [/bib_ref] [bib_ref] Intrachromosomal triplication of distal 7p, Rivera [/bib_ref] [bib_ref] Intrachromosomal triplication of 2q11.2-q21 in a severely malformed infant: case report and..., Wang [/bib_ref]. Powis et al. [bib_ref] Mosaic tetrasomy 12p with triplication of 12p detected by array-based comparative genomic..., Powis [/bib_ref] speculated that triplication of chromosome 12p could have arisen from telomere to telomere fusion of supernumerary analphoid isochromosome 12p with a normal chromosome 12. Indeed, there are three reports about PKS patients with analphoid inverted duplicated supernumerary marker chromosomes consisting of chromosome 12p [bib_ref] Tetrasomy 12pter-12p13.31 in a girl with partial Pallister-Killian syndrome phenotype, Vermeesch [/bib_ref] [bib_ref] Partial tetrasomy 12pter-12p12.3 in a girl with Pallister-Killian syndrome: extraordinary fi nding..., Dufke [/bib_ref] [bib_ref] Pallister-Killian syndrome: tetrasomy of 12pter®12p11.22 in a boy with an analphoid, inverted..., Huang [/bib_ref].
The clinical fi ndings of the patient reported by Powis et al. [bib_ref] Mosaic tetrasomy 12p with triplication of 12p detected by array-based comparative genomic..., Powis [/bib_ref] included hypotonia, brachycephaly with upslanting palpebral fi ssures, thin upper lip, low nasal bridge, high arched palate, a single transverse palmar crease on each hand and abnormal hair pattern. The clinical differences between this case and our case could result from differences in the degree of mosaicism and in the distribution of abnormal cells in different tissues.
As a result, we report here the third patient with intrachromosomal triplication of the short arm of chromosome 12. We conclude that intrachromosomal triplication might be a new mechanism of formation for PKS.
## Aims and scope of the journal
Balkan Journal of Medical Genetics (BJMG) was founded in 1998, and encompasses all aspects of human genetics: cytogenetics, molecular genetics, clinical genetics, immunogenetics, oncogenetics, pharmacogenetics, population genetics, genetic screening and diagnosis of monogenic and polygenic diseases, prenatal and preimplantation genetic diagnosis, genetic counseling, advances in treatment and prevention.
The journal is supported by an international Editorial Board.
The BJMG welcomes the submission of manuscripts that meet the general criteria of significance and scientific excellence, and will publish original articles, case studies, reviews, surveys, opinions, commentaries and essays.
# Identification statement
BJMG is published on the web as articles become available. The journal is published twice a year in June and November by the Macedonian Academy of Sciences and Arts.
[fig] Figure 1: Frontal view of the patient's head at 9 months of age. [/fig]
[fig] Figure 3: The patient's partial karyotype and an ideogram showing chromosome 12 pair. [/fig]
[fig] Figure 2: a) The patient's GTG-banded karyotype, b) The patient's CBG-banded metaphase chromosomes. [/fig]
[table] Table 1: Comparison of the patient's clinical fi ndings with a previously reported patient with an intrachromosomal triplication.Table is adapted from Huang et al. [19]. BALKAN JOURNAL OF MEDICAL GENETICS Yakut S, Mıhcı E, Altiok Clark O, Cetin Z, Keser I, Berker S, Luleci G,* [/table]
|
The Reduction of Peripheral Blood CD4+ T Cell Indicates Persistent Organ Failure in Acute Pancreatitis
ObjectiveFew data are available on the potential role of inflammatory mediators and T lymphocytes in persistent organ failure (POF) in acute pancreatitis (AP). We conducted a retrospective study to characterize their role in the progression of POF in AP.MethodsA total of 69 AP patients presented within 24 hours from symptom onset developing organ failure (OF) on admission were included in our study. There were 39 patients suffering from POF and 30 from transient OF (TOF). On the 1st, 3rd and 7th days after admission, blood samples were collected for biochemical concentration monitoring including serum IL-1β, IL-6, TNF-α and high-sensitivity C-reactive protein (hs-CRP). The proportions of peripheral CD4 + and CD8 + T lymphocytes were assessed based on flow cytometry simultaneously.ResultsPatients with POF showed a significantly higher value of IL-1β and hs-CRP on day 7 compared with the group of TOF (P < 0.05). Proportions of CD4 + T cells on days 1, 3, 7 and CD4 + / CD8 + ratio on day 1 were statistically lower in the group of POF patients (P < 0.05). A CD4 + T cell proportion of 30.34% on day 1 predicted POF with an area under the curve (AUC) of 0.798, a sensitivity with 61.54% and specificity with 90.00%, respectively.ConclusionsThe reduction of peripheral blood CD4 + T lymphocytes is associated with POF in AP, and may act as a potential predictor.
# Introduction
Acute pancreatitis (AP) is a disease characterized by local and systemic immunological response, which may result in systemic inflammatory response syndrome (SIRS), multiple organ failure and mortality [bib_ref] Acute pancreatitis with organ dysfunction associates with abnormal blood lymphocyte signaling: controlled..., Oiva [/bib_ref]. With the publication of the 2012 revised Atlanta classification [bib_ref] Classification of acute pancreatitis-2012: revision of the Atlanta classification and definitions by..., Banks [/bib_ref] , severe AP (SAP) has been redefined as AP with POF, i.e. if OF persists more than 48h. OF that resolves within 48h is regarded as transient OF (TOF) and belongs to the group of moderately severe AP (MSAP). A recent study by Nawaz et al. [bib_ref] Revised Atlanta and determinant-based classification: application in a prospective cohort of acute..., Nawaz [/bib_ref] found a mortality of 15.4% among SAP patients, while no death was observed among MSAP and mild AP (MAP). Also, some other studies showed a mortality of 36-67% in patients who developed POF in the early phase of AP [bib_ref] Classification of acute pancreatitis-2012: revision of the Atlanta classification and definitions by..., Banks [/bib_ref] [bib_ref] Persistent organ failure during the first week as a marker of fatal..., Johnson [/bib_ref] [bib_ref] Association between early systemic inflammatory response, severity of multiorgan dysfunction and death..., Mofidi [/bib_ref] [bib_ref] Persistent early organ failure: defining the high-risk group of patients with severe..., Lytras [/bib_ref].
At present, researchers suggest that the cytokine cascade from the innate immune system (mainly monocytes, macrophages and neutrophils) and the activated adaptive immune system (including CD4 + , CD8 + T and CD19 + B lymphocytes) are central to the developments of SIRS, compensatory anti-inflammatory response syndrome and OF in AP [bib_ref] Acute pancreatitis with organ dysfunction associates with abnormal blood lymphocyte signaling: controlled..., Oiva [/bib_ref] [bib_ref] Serum antigen(s) drive the proinflammatory T cell response in acute pancreatitis, Sweeney [/bib_ref] [bib_ref] Inflammation and immunosuppression in severe acute pancreatitis, Kylanpaa [/bib_ref] [bib_ref] Acquired immunityplays an important role in the development of murine experimental pancreatitis..., Nakayama [/bib_ref]. The innate immune system, and more specifically the inflammatory mediator, has been extensively explored over the past two decades. However, the activated adaptive immune system still remains to be systemically studied. T lymphocytes play a pivotal role in the regulation of the adaptive immune system, and possess a particular influence on macrophage and neutrophil activations. In a mice model of acute experimental pancreatitis, T cells, predominantly CD4 + subset invaded the pancreas and infiltrated border acini [bib_ref] CD4(+) T cells play an important role in acute experimental pancreatitis in..., Demols [/bib_ref]. Pietruczuk et al. [bib_ref] Alteration of peripheral blood lymphocyte subsets in acute pancreatitis, Pietruczuk [/bib_ref] found a significant depletion of circulating CD4 + T cell population in the early stage of AP, while CD8 + cells were in the normal range.
Alterations of the immune systems in AP patients with OF should be deeply explored given the fact that OF has become the leading cause of morbidity and mortality in these patients [bib_ref] Persistent organ failure during the first week as a marker of fatal..., Johnson [/bib_ref] [bib_ref] Multiple organ dysfunction associated with severe acute pancreatitis, Halonen [/bib_ref]. In this study, we, therefore, aimed to evaluate the role of peripheral inflammatory mediator and T lymphocyte in the progression of POF in patients with AP.
# Materials and methods
## Patients
Between November 1 st 2012 and September 1 st 2013, a total of 69 AP patients with OF on admission presented to the Pancreatic Disease Institute of Wuhan Union Hospital were reviewed. In all patients, the time between the abdominal pain onset and admission to the hospital was within 24 hours. The diagnostic criteria were based on the presence of two or more of the following three features: 1) abdominal pain consistent with AP; 2) serum amylase and/or lipase elevation 3 times the upper limit of normal; and / or 3) computed tomography (CT) findings characteristic of AP [bib_ref] Classification of acute pancreatitis-2012: revision of the Atlanta classification and definitions by..., Banks [/bib_ref]. The exclusion criteria included any of the following: age smaller than 18 years or greater than 80 years, a diagnosis of recurrent or chronic pancreatitis, a history of chronic pulmonary, renal or cardiovascular disease, pancreatitis induced by trauma or pregnant, and any surgical intervention taken in the first 3 days after admission. The study was conducted according to the principles of the Declaration of Helsinki. The need for informed consent was waived by the Medical Ethics Committee because the study was an observational one using a database from which patients' identification information had been removed. The ethics review board of Wuhan Union Hospital approved this study.
There were 39 patients suffering from POF and 30 from TOF, respectively. OF was diagnosed according to a modified Marshall scoring system when the following cutoffs were exceeded: 1) cardiovascular failure if systolic blood pressure was < 90 mmHg with no response to fluid resuscitation; 2) respiratory failure if the ratio of PaO 2 /FiO 2 was < 300 mmHg; and 3) renal failure if serum creatinine was 1.9 mg/dl (S1 . The acute physiology and chronic health evaluation II (APACHE II) scores [bib_ref] APACHE II: a severity of disease classification system, Knaus [/bib_ref] were recorded on admission and on day 1.
## Treatment
All patients received standard treatment based on the United Kingdom and Chinese Medical Association guidelines for the management of acute pancreatitis . Mechanical ventilation (noninvasive or invasive) and dialysis were instituted when needed. Antibiotic prophylaxis was administered for no more than 10 days, unless clinical symptoms of persistent sepsis were shown. Fluid supplementation followed international guidelines. Both crystalloids and colloids were used based on the assessment of residents. Parenteral nutrition was given, and after the function of gastrointestinal tract recovered, patients received both enteral feeding and parenteral nutrition. If a patient was confirmed of an obstruction in the common bile duct or in the ampulla of Vater by CT, an endoscopic retrograde cholangiopancreatography (ERCP) was given. CT scan was performed within 72 hours of admission before any surgical intervention. A contrastenhanced CT was repeated after 7 to 10 days to determine the morphology of acute pancreatitis. All CT scans were reviewed by a radiologist who was blinded to the study. The presence of infected pancreatic necrosis was presumed when there was extraluminal gas in the pancreatic and / or peripancreatic tissues on contrast-enhanced CT or when percutaneous, image-guided fine needle aspiration was positive for bacteria and / or fungi on Gram stain and culture [bib_ref] Classification of acute pancreatitis-2012: revision of the Atlanta classification and definitions by..., Banks [/bib_ref]. CD4 + and CD8 + T lymphocytes. Surface monocyte antigens (cluster designation, CD) of heparinized peripheral blood lymphocytes were assessed using double-stained (fluorescein isothiocyanate (FITC) / phycoerythrin (PE)) monoclonal antibodies (BD Bioscience, San Jose, CA, USA) and a flow cytometer (FACSCanto II, BD Bioscience, San Jose, CA, USA).
## Sample collection and measurements
# Statistical analysis
Statistical analysis was undertaken using SPSS 16.0 (SPSS Inc, Chicago, IL). Data are shown as mean ± standard deviation (SD). Group comparisons used Student's t test for normally distributed variables and non-parametric Mann-Whitney U test for non-normally distributed variables. Categorical variables were compared with a chi-square test. Data between days 1, 3 and 7 were compared using Friedman repeated-measures ANOVA on ranks. Differences were considered to be significant at P < 0.05. Receiver-operating curves (ROC) were generated to determine the cutoff values for optimal sensitivity, specificity, positive and negative predictive values.
# Results
## Patient characteristics
Baseline characteristics of the two groups are summarized in [fig_ref] Table 1: Baseline characteristics of the AP patients [/fig_ref] with transient multiple OF. Twenty-three patients in the group of TOF developed acute peripancreatic fluid collection, five suffered from acute necrotic collection and two from interstitial oedematous pancreatitis. Three patients developed pancreatic pseudocyst, and one developed walled-off necrosis. There were 22 patients in the group of POF suffering from acute peripancreatic fluid collection, and 17 from acute necrotic collection. Pancreatic pseudocyst was found in 7 and walled-off necrosis in 5 of the POF ones. Infection of pancreatic necrosis was found in 3 TOF patients and in 11 POF ones, which was confirmed by positive fine needle aspiration of necrotic tissue. Infection was observed from 6 to 21 days after hospitalization, with a median time of 12 days. Twelve patients underwent pancreatic debridement, with a mean operation time of 24 days. The remaining one was treated conservatively with percutaneous drainage of organized pancreatic necrosis. No laparoscopic debridement was given. Eight patients (5 men and 3 women) died among the 39 POF patients due to infected necrosis during hospitalization, with a mortality rate of 20.5%. No death was observed in the 30 TOF ones [fig_ref] Table 1: Baseline characteristics of the AP patients [/fig_ref].
## Inflammatory mediator
Serum peak value of IL-6 was observed on day 1, and then decreased rapidly on days 3 and 7 in both POF patients (63.08 ± 48.02 pg/ml vs. 41.10 ± 27.34 pg/ml and 25.24 ± 30.28 pg/ml, P < 0.01) and TOF ones (55.14 ± 48.62 pg/ml vs. 31.84 ± 28.49 pg/ml and 19.94 ± 15.90 pg/ml, lower in the TOF group on day 7 compared with the POF group, while values of IL-6 and TNF-α showed no significant difference between the two groups on all three days [fig_ref] Fig 1: Sequential changes in the values of IL-1β [/fig_ref].
## Cd4 + and cd8 + t lymphocytes
Proportions of CD4 + T lymphocytes on days 1, 3 and 7 were significantly decreased in POF patients compared with TOF patients (27.73 ± 8.25% vs. 37.26 ± 7.04%, 31.81 ± 7.87% vs. 40.79 ± 7.07% and 33.55 ± 9.31% vs. 42.61 ± 7.81%, P < 0.0001). Proportions of CD8 + T cells were lower in the group of POF than in the group of TOF on days 1, 3 and 7, but the differences were not statistically significant. A significant reduction of the CD4 + / CD8 + ratio on day 1 was noticed in POF patients compared with TOF ones (1.54 ± 0.54 vs. 2.00 ± 0.85, P < 0.05), but not on day 3 and day 7 [fig_ref] Fig 2: Sequential changes in the proportion of CD4 + T lymphocytes [/fig_ref].
To determine early predictors of POF, we optimized the sensitivity and specificity, and calculated the optimal cutoff values for the proportion of CD4 + T lymphocytes and for the CD4 + / CD8 + ratio on day 1 using ROC curves. CD4 + T lymphocyte proportion on day 1 presented an area under the curve (AUC) of 0.798, with a sensitivity of 61.54% and specificity of 90.00%. The positive predictive value (PPV) was 88.89%, and the negative predictive value (NPV) was 64.29%. The optimal threshold was 30.34%. However, the CD4 + / CD8 + ratio on day 1 showed an AUC of 0.670, with an optimal threshold of 2.10 [fig_ref] Table 2: ROC analysis of CD4 + T cell proportion, CD4 + / CD8... [/fig_ref] ; [fig_ref] Fig 3: ROC curve of CD4 + T cell proportion versus CD4 + /CD8... [/fig_ref].
## Apache ii score
The APACHE II score on admission was 11.56 ± 4.15 in POF patients, significantly higher than those in TOF patients (8.50 ± 1.15, P < 0.01). A significant difference was also observed on day 1 (POF: 10.95 ± 3.85 vs. TOF: 8.80 ± 1.36, P < 0.05) as well. ROC analysis was performed for the APACHE II scores on admission and on day 1 [fig_ref] Fig 2: Sequential changes in the proportion of CD4 + T lymphocytes [/fig_ref].
# Discussion
Studies show that 37-76% of AP patients develop OF within 24h after symptom onset [bib_ref] Double blind, randomised, placebo controlled study of a platelet activating factor antagonist,..., Johnson [/bib_ref]. POF leads to almost all deaths in the early phase of AP. It is well accepted that patients suffering from SIRS have a high risk of developing OF [bib_ref] Acute pancreatitis with organ dysfunction associates with abnormal blood lymphocyte signaling: controlled..., Oiva [/bib_ref]. As a part of the host response, systemic inflammation is activated by both innate and adaptive immune systems. The former is directed by monocytes, macrophages and neutrophils, while the later is mainly made up of CD4 + T, CD8 + T and CD19 + B lymphocytes. Pancreatic inflammation triggers the innate immune system, and induces a cytokine cascade. However, the adaptive immune system is activated by antigens and cytokines. Exaggerated and uncontrollable host response leads to SIRS, and eventually progresses into OF [bib_ref] Revised Atlanta and determinant-based classification: application in a prospective cohort of acute..., Nawaz [/bib_ref] [bib_ref] Persistent organ failure during the first week as a marker of fatal..., Johnson [/bib_ref] [bib_ref] Inflammatory mediators in human acute pancreatitis: clinical and pathophysiological implications, Mayer [/bib_ref]. The reductions of peripheral blood T cell subpopulations in the course of AP have been reported previously. Curley et al. [bib_ref] Reduction in circulating levels of CD4-positive lymphocytes in acute pancreatitis: relationship to..., Curley [/bib_ref] noticed a significant decrease in the proportion of CD4 + T cells in AP with local complications (necrosis, abscess or pseudocyst) compared with the mild form. Pezzilli and colleagues [bib_ref] Circulating lymphocyte subsets in human acute pancreatitis, Pezzilli [/bib_ref] observed a significant decrease of both CD4 + and CD8 + T cells in severe cases, whose severity was assessed by the clinical outcome and the number of complications developed during hospitalization. Uehara et al. [bib_ref] Immune function in patients with acute pancreatitis, Uehara [/bib_ref] reported that CD4 + T cells were further decreased in the severe form compared with the mild form, while the CD8 + cells and the CD4 + / CD8 + ratio were not different. In that study, the severity was determined by the Ranson's score, the APACHE II score and the Japanese 1990 criteria. Pietruczuk et al. [bib_ref] Alteration of peripheral blood lymphocyte subsets in acute pancreatitis, Pietruczuk [/bib_ref] observed a significant decrease of the CD4 + T lymphocyte population in AP patients, as diagnosed according to the Ranson's score, the Balthazar's criteria and serum CRP levels, while CD8 + T cells remained in the normal range.
In this study, we focused on the relationship between T lymphocyte subgroups and POF, since POF is the main cause of early mortality in AP. Significant reduction of peripheral CD4 + cells in AP patients with POF was observed in a week-long period after admission compared with TOF patients, indicating that sustained lower levels of CD4 + T cells may be related to the progress of POF and to poor outcomes.
At present, the reasons for the reduction of peripheral CD4 + T lymphocytes in AP remain unclear. One of the possible reasons is the migration of activated CD4 + cells to the inflammatory sites, including the pancreas and other tissues like the lung and kidney [bib_ref] Pathophysiology of acute pancreatitis, Bhatia [/bib_ref]. In an experimental model, CD4 + T cell-depleted mice and nude mice (T cell-deficient) developed less severe pancreatitis than immunocompetent mice; after adaptive CD4 + lymphocytes transfer, this effect was partially reversed [bib_ref] CD4(+) T cells play an important role in acute experimental pancreatitis in..., Demols [/bib_ref]. The decrease in peripheral blood CD4 + T lymphocytes may indicate activated CD4 + cells being recruited to vital organs. Another possible explanation is an excessive elimination of lymphocytes by apoptosis [bib_ref] Alteration of peripheral blood lymphocyte subsets in acute pancreatitis, Pietruczuk [/bib_ref] [bib_ref] Peripheral lymphocyte reduction in severe acute pancreatitis is caused by apoptotic cell..., Takeyama [/bib_ref] , but the precise mechanism needs further research.
Early prediction of POF contributes to the appropriate management of AP patients, thus decreases morbidity and mortality. To the best of our knowledge, it is the first time that the peripheral CD4 + T lymphocyte proportion and the CD4 + / CD8 + ratio could be used as early predictive factors for POF in AP patients. At a cutoff value of 30.34% of CD4 + T cells, POF was predicted with a specificity of 90.00%.
TNF-α and IL-1β are pro-inflammatory cytokines. They are mainly released by macrophages, leading to an enhanced capillary permeability by activating neutrophils and vascular endothelia, and inducing the release of other cytokines. Studies revealed that these two mediators play a fundamental role in the progression of OF as well as pancreatic necrosis [bib_ref] Acute pancreatitis as a model of sepsis, Wilson [/bib_ref] [bib_ref] Risk and markers of severe acute pancreatitis, Papachristou [/bib_ref] [bib_ref] Identification of a key pathway required for the sterile inflammatory response triggered..., Chen [/bib_ref] [bib_ref] Tumour necrosis factor alpha secretion induces protease activation and acinar cell necrosis..., Sendler [/bib_ref]. IL-6 is produced by monocytes and fibroblasts, and acts as the main inducer of CRP synthesis in the liver [bib_ref] The role of proinflammatory cytokines in inflammatory and metabolic responses, Tompkins [/bib_ref]. Except the immune system, IL-6 and IL-1β can be produced by multiple organs, indicating their major roles in the pathophysiology of OF [bib_ref] Diagnostic accuracy of interleukin-6 and interleukin-8 in predicting severe acute pancreatitis: a..., Aoun [/bib_ref]. There is now concordant evidence that plasma level of IL-6 correlates with the severity of AP according to the 1992 criteria [bib_ref] Severe acute pancreatitis: pathogenetic aspects and prognostic factors, Al [/bib_ref].
In the present study, no significant differences in respect to serum values of IL-6 and TNF-α on days 1, 3 and 7 were observed between POF and TOF, suggesting that they may not play an important role in the progression of POF. On day 7, IL-1β and hs-CRP levels were significantly higher in the POF group, indicating that persistent high levels of IL-1β and hs-CRP are associated with POF. However, neither of them can be used as an early predictor.
Previously, the APACHE II score system was used to estimate the need of intensive care support for critically ill patients and to distinguish SAP from the mild form. According to the 1992 Atlanta criteria, this system may be used to determine the severity at any time during the course of AP, and the severe category was defined as a score 8 [bib_ref] Classification of acute pancreatitis-2012: revision of the Atlanta classification and definitions by..., Banks [/bib_ref]. However, another study [bib_ref] Dynamic nature of early organ dysfunction determines outcome in acute pancreatitis, Buter [/bib_ref] classified patients as SAP if they scored > 6. Johnson et al. [bib_ref] Persistent organ failure during the first week as a marker of fatal..., Johnson [/bib_ref] observed that a fatal outcome was associated with an APACHE II score > 8. In the present study, the APACHE II scores were significantly different between the POF and TOF groups at admission and on day 1, but they showed weaker predictive abilities compared with the proportion of CD4 + T lymphocytes on day 1. Therefore, the proportion of CD4 + T lymphocytes on day 1 is comparable to the APACHE II score in the prediction of POF.
Our study had several limitations. First, as this was a derivation study the true value of CD4 + T lymphocytes in predicting POF remained to be investigated further in a prospective study. Second, race and ethnicity might have important bias on T cell subsets. As a result, researches including both Asian Chinese and other populations were needed. Moreover, the number of patients in our study was limited (n = 69), and we were not able to include MSAP patients without OF, as well as MAP and healthy controls. The retrospective nature did not allow us to evaluate peripheral blood lymphocyte count and anti-inflammatory cytokine IL-10 in our patients. Besides, parameters after clinical and laboratory remission were not assessed.
# Conclusions
The present study revealed that the reduction of CD4 + T lymphocytes in the early phase of AP is related to the progression of POF. Decreased CD4 + T cells may act as a potential predictor of POF in AP. The mechanism underlying needs to be further clarified in a large prospective multicenter study.
Supporting Information S1
[fig] Fig 1: Sequential changes in the values of IL-1β (A), IL-6 (B), TNF-α (C) and hs-CRP (D) in the POF and TOF groups. doi:10.1371/journal.pone.0125529.g001 CD4 + T Cell Reduction Indicates Persistent Organ Failure in AP PLOS ONE | DOI:10.1371/journal.pone.0125529 May 4, 2015 [/fig]
[fig] Fig 2: Sequential changes in the proportion of CD4 + T lymphocytes (A), the proportion of CD8 + T [/fig]
[fig] Fig 3: ROC curve of CD4 + T cell proportion versus CD4 + /CD8 + ratio in predicting POF.doi:10.1371/journal.pone.0125529.g003 CD4 + T Cell Reduction Indicates Persistent Organ Failure in AP PLOS ONE | DOI:10.1371/journal.pone.0125529 May 4, 2015 [/fig]
[fig] Fig 4: ROC curve of APACHE II score on admission versus APACHE II score on day 1 in predicting POF. doi:10.1371/journal.pone.0125529.g004 [/fig]
[table] Table 1: Baseline characteristics of the AP patients.Data are presented in either means and standard deviations or frequencies and percentages. [/table]
[table] Table 2: ROC analysis of CD4 + T cell proportion, CD4 + / CD8 + ratio and APACHE II scores in diagnosing POF. [/table]
[table] Table: Modified Marshall scoring system for organ failure. (DOC) S1 Dataset. Relevant data underlying the findings described in manuscript. (XLS) [/table]
|
Nurse-led, telephone-based follow-up after acute coronary syndrome yields improved risk factors after 36 months: the randomized controlled NAILED-ACS trial
We investigated whether a nurse-led, telephone-based follow-up including medical titration was superior to usual care in improving blood pressure (BP) and low-density lipoprotein cholesterol (LDL-C) values 36 months after acute coronary syndrome (ACS). We screened all patients admitted with ACS at Östersund hospital, Sweden, between January 1, 2010, and December 31, 2014, for inclusion based on ability to participate in a telephone-based follow-up. Participants were randomly allocated to usual care or an intervention group that received counselling and medical titration to target BP < 140/< 90 mmHg and LDL-C < 2.5/< 1.8 mmol/L. The primary outcome was LDL-C at 36 months. Of 962 patients, 797 (83%) were available for analysis after 36 months. Compared to controls, the intervention group had a mean systolic BP (SBP) 4.1 mmHg lower (95% confidence interval [CI] 1.9-6.5), mean diastolic BP (DBP) 2.9 mmHg lower (95% CI 1.5-4.5), and mean LDL-C 0.28 mmol/L lower (95% CI 0.135-0.42). All P < 0.001. A significantly greater proportion of patients reached treatment targets with the intervention. After 36 months of follow-up, compared to usual care, the nurse-led, telephone-based intervention led to significantly lower SBP, DBP, and LDL-C and to a larger proportion of patients meeting target values.Trial registration: ISRCTN registry. Trial number ISRCTN96595458. Retrospectively registered.
Participants. The county of Jämtland, Sweden, has only one hospital (Östersund), which has a catchment area of around 130,000 people. During the inclusion period, which ran from January 1, 2010, until December 31, 2014, all patients admitted with ACS were eligible for inclusion and screened. The definition of ACS was either acute myocardial infarction type 1 (AMI: either ST-elevation myocardial infarction or non-STEMI [NSTEMI]) or unstable angina (UA) [bib_ref] Universal definition of myocardial infarction, Thygesen [/bib_ref]. Exclusion criteria were based on the patient's inability to participate in the telephone-based follow-up, and patients were excluded if they had hearing loss, aphasia, severe dementia, an inability to communicate in Swedish or English, or could not use a telephone. Participation in another ongoing trial was also considered a reason for exclusion. A previous study explored reasons for non-participation Randomisation. Randomisation occurred via computer allocation in blocks of four, with stratification based on type of ACS (AMI, UA) and sex. Patients were randomised into two parallel groups, a control group and an intervention group.
Data collection and follow-up. During the initial hospitalisation, baseline data, demographic information, comorbidities, the use of medication and health status were collected via interview and medical records. At 1, 12, 24, and 36 months after hospital discharge, blood pressure (BP) readings and blood lipid measurements were collected via the closest health care provider. Shortly thereafter, a study nurse telephoned patients in both study groups.
Patients in both the intervention and control group were interviewed in regard to general status, level of physical activity, smoking and medication intake. Results were recorded on pre-printed standardised forms and stored in binders organised by patient number. The data from standardised BP and LDL-C measurements were registered in both the electronic journal system and also on standardised paper forms. LDL-C values were calculated using Friedewald's formula based on fasting values of cholesterol and triglycerides. Blood pressure readings were made with the patient in a seated position after 5 min of rest. [fig_ref] Figure 1: Study flow chart [/fig_ref] show the study flow chart.
Usual care. Regardless of whether randomised into the intervention or control group, all patients received the standard follow-up via the cardiology clinic. This standard follow-up include a visit to a cardiology nurse approximately one month after discharge, and after 2-3 months a visit to a cardiologist. This follow-up is the same for trial patients and non-trial patients. During the standard follow up via the clinic, patients receive general secondary preventive guidance in regard to lifestyle factors and an overview of medication and secondary preventive targets. In most cases, patients would then be referred to their GP for continued secondary preventive treatment. For both groups, secondary preventive medication was initiated in-hospital in accordance to national guidelines.
Intervention group. During the telephone based follow-up patients randomised to the intervention were counselled on the importance of medical adherence, physical activity, and exercise and in applicable cases, smoking cessation. The study nurses were educated in regard to motivational interviewing. Patients in the intervention group were given advice regarding diet and exercise in accordance with recommendations by the National Food Administration and the National Board of Health and Welfare. Current smokers were given advice about smoking cessation and were recommended available resources. At follow-up, if an intervention patient had BP or LDL-C values above target levels, a physician would be contacted and the patient would have their medication titrated to achieve target levels. Titration was individualized and choice of medication and dosage was up to the treating study physician. Following every titration, a follow-up via telephone was scheduled approximately one month after titration, and if needed further medical adjustments was made.
Control group. For the control group, there was no intervention or titration made by the study physicians and the patients received no counselling or medical advice from the study nurses. Patients simply provided BP and LDL-C measurements and were interviewed. The blood lipid and blood pressure measurements from the control group were available to both the study nurses and the patient's primary care provider. The patients received what we refer to as usual care as previously described.
Target levels. Target levels for BP and LDL-C were based on current local guidelines at the time of the trial.
Target levels for BP were < 140/90 mmHg. In regard to LDL-C the initial target was < 2.5 mmol/L. In March 2013 there was a local guideline change for diabetic patients in which a lower target of < 1.8 mmol/L was set. The [bib_ref] Implementation of a new guideline in cardiovascular secondary preventive care: Subanalysis of..., Jakobsson [/bib_ref]. In 2017, local guidelines were updated again and a target LDL-C value of < 1.8 was set for all patients with a history of ACS.
# Statistical analysis.
We performed analyses in accordance to an intention-to-treat principle, in which data were analysed regardless of patient adherence to treatment. We did not use imputation for missing data. Results are presented as means for continuous variables and as percentages for categorical variables. Baseline characteristics between the intervention and control groups were compared using t-tests for continuous variables and chi-square tests for categorical variables. Adjusted means were compared between the intervention and control groups via general linear regression, adjusted for our randomisation variables, sex, and type of ACS. Paired samples t-tests were used for comparisons within groups. A P < 0.05 was considered significant. All analyses were carried out using IBM SPSS v24.
Power calculation. The power calculation was based on being able to detect a mean difference of 5 mmHg in SBP with a standard deviation (SD) of 19, and a mean difference of 0.5 mmol/L in LDL-C (SD 1.0). Based on a two-tailed alpha of 0.05 and 80% power, we calculated that a minimum of 200 participants in each study group www.nature.com/scientificreports/ would be needed. We included substantially more patients to maintain statistical power for analyses after longterm follow-up and in subgroups.
Trial registration. The NAILED-ACS trial was registered in the International Standard Randomized Controlled Trial Number (ISRCTN) registry on August 24, 2011 (trial number: ISRCTN96595458). Unfortunately, before the strict requirement of prospective registration came to our attention, recruitment had already begun. Thus, this study is classified as retrospectively registered. We confirm that all related and on-going trials are now registered.
Ethics. The Regional Ethics Committee, Umeå, approved the study on October 28, 2009. The study was conducted in accordance with relevant guidelines and regulations regarding scientific research. All participants signed an informed, written consent document (D-nr 09-142M).
# Results
Participants. In total, 962 patients were randomised. Of 486 patients in the intervention group, 406 (83.5%) completed the 36-month follow-up. In the control group, there were 476 patients enrolled, and 391 (82.1%) completed the 36-month follow-up. In the intervention group, 42 patients died between randomisation and the 36-month follow-up, 14 chose to withdraw participation, 12 were unable to continue or were unreachable, 4 moved, and 8 participated in drug trials. In the control group, 53 patients died, 10 withdrew, 11 were unable to continue or were unreachable, 10 moved, and 1 patient had missing data [fig_ref] Figure 1: Study flow chart [/fig_ref]. There were no significant differences in baseline variables between the two groups .
Blood pressure at the 36-month follow-up. Use of medicine at the 36-month follow-up. [fig_ref] Table 2: Medication at 36 months [/fig_ref] shows the use of medicine at the 36-month follow-up. In the intervention group, the use of calcium channel blockers was more common, and the use of statins, angiotensin-receptor blockers, and anticoagulants as well as diuretics were numerically higher, although nonsignificantly compared to the control group.
Trends over time. As seen in [fig_ref] Figure 3: Adjusted mean values for SBP [/fig_ref] and below, the adjusted means for SBP, DBP, and LDL-C showed a clear difference between groups. For the intervention group, the titrations were associated with a distinct decrease in SBP, DBP, and LDL-C, although at the subsequent scheduled annual assessment, the effect of the intervention had decreased.
# Discussion
In this randomised controlled trial, a nurse-led telephone-based follow-up that included medical titration was superior to usual care. After 36 months of follow-up, the intervention group had significantly lower values for SBP, DBP, and LDL-C. The proportions of patients reaching target values were also significantly higher in the intervention group. There was a trend towards rising values in the control group, and lower values in the intervention group. The differences were numerical, and not statistically significant with the exception of lower SBP in the intervention group and higher LDL-C in the control group. Overall, this result suggests that the intervention promoted achievement of lower levels and helped patients avoid the rise, particularly in LDL-C, seen in the control group.
## Effect of the intervention on bp and ldl-c and clinical relevance.
At the end of the study followup at 36 months, comparing the intervention group to controls, the mean SBP was 4.1 mmHg lower, mean DBP was 2.9 mmHg lower, and mean LDL-C was 0.28 mmol/L lower. A lowering of BP and LDL-C in secondary prevention reduce cardiovascular events as shown in several prior studies.
The clinical relevance of the present numerically small differences between the intervention and control group can be placed in the context of previously published analyses, such as results of a meta-analysis showing that a reduced SBP by 10 mmHg translated into a 22% reduction in coronary heart disease [bib_ref] Use of blood pressure lowering drugs in the prevention of cardiovascular disease:..., Law [/bib_ref]. Regarding LDL-C, other studies have shown that a reduction in LDL-C by 1 mmol/L resulted in a decreased relative risk of cardiovascular death by approximately 20% [bib_ref] Association between lowering LDL-C and cardiovascular risk reduction among different therapeutic interventions:..., Silverman [/bib_ref]. The combined effect of lowering blood pressure and LDL-C and the effect of the marked reduction in risk factor levels after titration is difficult to estimate.
Effects on mortality and morbidity were not the focus of this study. A numerically higher number of participants had died by the 36-month follow-up in the control group as compared to the intervention group (53 vs 42) corresponding to percentages of 11% and 8.6%, respectively, from randomisation to the 36-month followup. A separate study regarding adjudicated clinical endpoints in the entire NAILED trial population is recently completed and will be reported in accordance with the study plan (ISRCTN30433343; Scientific Reports 2021, accepted for publication).
Target level achievement. Secondary preventive measures following acute coronary syndrome have room for improvement. The proportion of patients reaching target values for blood pressure and LDL-C tends to be inadequate, as was evident in the large Euroaspire IV-survey 5 , in which roughly two-fifths of patients reached treatment targets for blood pressure (140/90) and less than two-thirds attained treatment targets for LDL-C (< 2.5 mmol/L). In the present study, a comparatively larger proportion of patients in the intervention group reached treatment target levels (77.6% for SBP, 90.9% for DBP, and 65.5% for LDL-C). Compared to the results in the Euroaspire IV survey, our control group exhibited a higher proportion of patients who achieved treatment target levels at 36 months of follow-up for blood pressure (62.9% for SBP and 80.8% for DBP). The proportion of patients reaching target levels for LDL-C were in line with the results of the survey, with 53.1% reaching target levels. The higher proportion of patients in the control group who achieved treatment target levels compared to patients in the Euroaspire IV-study needs to be considered when evaluating the effect of the intervention. The relative improvement in the intervention group was achieved despite a well-treated control group.
## Intervention and titration.
Our results demonstrate that although the intervention did not result in a continuous reduction of BP and LDL-C a reduced risk factor burden was maintained throughout the study period. Shortly after titration, the difference was more pronounced, but the effect abated, as can be seen in [fig_ref] Figure 3: Adjusted mean values for SBP [/fig_ref]. . Adjusted means over time. www.nature.com/scientificreports/ Even though the effect was not lasting, the reduction in overall risk factor burden in the intervention group, measured as a reduced area under the curve, may be speculated to add to the benefit related to the lower point estimates of blood pressure and LDL-C. Of note, the titration could result in long lasting lowering of BP and LDL-C on an individual basis, but on a group level the effect diminished over time. We can only speculate as to why the effect of the titration declined over time for the whole group. Some patients that were previously within target levels could experience a worsening of BP and LDL-C over time which could be due to a number of factors, such as poor medical adherence, changes in lifestyle factors or for other medical reasons.
Follow-up and medication. At the yearly follow-up, patients were interviewed and asked about their current medication. Study nurses also evaluated both prescription and laboratory data to identify discrepancies such as reporting statin adherence but having increased LDL-C. At the 36-month follow-up, there were no significant differences regarding use of medication apart from a higher proportion of intervention patients receiving calcium channel blockers, although there was a nonsignificant trend towards higher use of statins, diuretics, and angiotensin receptor blockers. However we do not have data about the dose of the different blood pressure lowering medications but it is possible that the intervention group in general were prescribed higher doses. This is supported by our previous findings that the intervention led to increased adherence to statins and a greater use of high-intensity lipid lowering therapy [bib_ref] Statin treatment after acute coronary syndrome: Adherence and reasons for non-adherence in..., Daniel [/bib_ref].
Overall, most patients were treated with medications in accordance with European Society of Cardiology guideline recommendations [bib_ref] ESC Guidelines for the management of acute myocardial infarction in patients presenting..., Ibanez [/bib_ref] [bib_ref] ESC Guidelines for the management of acute coronary syndromes in patients presenting..., Roffi [/bib_ref].
This indicates that a reported high proportion of patients on treatment does not automatically translate to a high proportion reaching treatment target levels. This could be due to a number of reasons, such as inadequate doses or poor adherence.
Poor adherence to secondary preventive medications has been previously described [bib_ref] Impediments to adherence to post myocardial infarction medications, Desai [/bib_ref] [bib_ref] Long-term medication adherence after myocardial infarction: Experience of a community, Shah [/bib_ref]. It is possible that the yearly instructions on blood pressure measurement and blood lipid testing led to a higher proportion of patients with adequate treatment in both groups, as compared to numbers seen in the previous referenced studies. We earlier examined adherence to statin treatment specifically in NAILED ACS study participants with a mean of 3.9 years of follow-up. In the intervention group, 89% were adherent compared to 85% in the control group, so figures were high in both groups [bib_ref] Statin treatment after acute coronary syndrome: Adherence and reasons for non-adherence in..., Daniel [/bib_ref].
A Norwegian study on adherence to secondary preventive drugs after myocardial infarction with up to 2 years of follow-up showed higher adherence compared to older studies, but still slightly lower than in the present study [bib_ref] Initiation of and long-term adherence to secondary preventive drugs after acute myocardial..., Halvorsen [/bib_ref].
Another explanation for inadequate secondary prevention is therapeutic or clinical inertia which has long been acknowledged and is defined as not acting in accordance with or adhering to guidelines in the treatment of various symptoms and diseases [bib_ref] Clinical inertia in the pharmacological management of hypertension: A systematic review and..., Milman [/bib_ref] [bib_ref] The concept and definition of therapeutic inertia in hypertension in primary care:..., Lebeau [/bib_ref]. This inertia is likely part but not solely the explanation for the higher mean BP and increased LDL-C levels in the control group. The goal-oriented medical titration in the intervention group may have helped to lower this therapeutic inertia and aid in achieving treatment targets. Titration was not always possible, either because of non-adherence by the patient or decision of the study physician because of an already maximum dosage, co-morbidity, or adverse effects.
## Relevance of nailed acs.
Comparing our results to other trials are difficult because of a lack of longterm perspective 9,10,28 . Other nurse-led or nurse-coordinated secondary prevention studies have also shown promising results but direct comparisons with the NAILED ACS trial are difficult to make due to heterogeneity in trial participants, design and time frame [bib_ref] Effect of a nurse-coordinated prevention programme on cardiovascular risk after an acute..., Jorstad [/bib_ref] [bib_ref] Global secondary prevention strategies to limit event recurrence after myocardial infarction: Results..., Giannuzzi [/bib_ref].
Secondary prevention should be seen as a lifelong engagement, and this study shows that a nurse-led telephone-based intervention can improve control of relevant risk factors in the long-term. The overall design of the NAILED-ACS trial could be integrated into clinical practice, at least for developed nations, in a similar setting to diabetic follow-up in primary care. If so, further research with a focus on clinical outcomes is warranted.
# Strengths and limitations.
The patients included in this study consist of a representative and clinically relevant population. The Jämtland county in middle Sweden consist of both urban and rural settings. There is only one hospital with one cardiology clinic, which enabled us to conduct the study in a controlled manner; however, the single centre design might limit external validity. The trial population in NAILED ACS consist of relatively unselected ACS patients and represents patients typically encountered in a clinical setting. The population and setting is comparable to other western countries. To our knowledge, no other population-based longterm secondary preventive intervention studies have focused on telephone-based follow-up. One limitation of this study is that the control patients received instruction each year to measure their blood pressure and blood lipid levels and had a short interview over the telephone with the study nurses. This interaction could remind the patient of the importance of blood pressure management and medication adherence. An underestimate of the effect of the intervention is therefore possible. An analysis of the importance of lifestyle factors was beyond the scope of the present trial.
# Conclusion
After 36 months of follow-up, the nurse-led, telephone-based intervention led to significantly lower SBP, DBP, and LDL-C values and increased the proportion of patients reaching their targets. Our data imply that a secondary prevention strategy must be sustained beyond the first year to maintain risk factor reduction.
[fig] Figure 1: Study flow chart. Scientific Reports | (2021) 11:17693 | https://doi.org/10.1038/s41598-021-97239-x [/fig]
[fig] Figure 2: depicts the proportion of patients reaching treatment targets at 36 months of follow-up. The proportion of patients within target levels for SBP was 62.9% in the control group, compared to 77.6% in the intervention group (P < 0.001). This proportion for DBP was 80.8% of patients in the control group reaching the target vs 90.9% in the intervention group (P < 0.001). For LDL-C targets, 53.1% in the control group met targets vs 65.6% in the intervention group (P < 0.001). [/fig]
[fig] Figure 3: Adjusted mean values for SBP (a), DBP (b), and LDL-C (c) at assessments at 1, 12, 24, and 36 months and after titration. SBP systolic blood pressure, DBP diastolic blood pressure, LDL-C low-density lipoprotein cholesterol. Means adjusted for type of ACS and sex. [/fig]
[table] Table 2: Medication at 36 months. a Other antiplatelet drug, e.g., clopidogrel, ticagrelor. [/table]
|
A seed resource for screening functionally redundant genes and isolation of new mutants impaired in CO2 and ABA responses
The identification of homologous genes with functional overlap in forward genetic screens is severely limited. Here, we report the generation of over 14 000 artificial microRNA (amiRNA)-expressing plants that enable screens of the functionally redundant gene space in Arabidopsis. A protocol was developed for isolating robust and reproducible amiRNA mutants. Examples of validation approaches and essential controls are presented for two new amiRNA mutants that exhibit genetically redundant phenotypes and circumvent double mutant lethality. In a forward genetic screen for abscisic acid (ABA)-mediated inhibition of seed germination, amiRNAs that target combinations of known redundant ABA receptor and SnRK2 kinase genes were rapidly isolated, providing a strong proof of principle for this approach. A new ABA-insensitive amiRNA line that targets three avirulence-induced gene 2(-like) genes was isolated . A thermal imaging screen for plants with impaired stomatal opening in response to low CO 2 exposure led to the isolation of a new amiRNA targeting two essential proteasomal subunits, PAB1 and PAB2. The seed library of 11 000 T2 amiRNA lines (with 3000 lines in progress) generated here provides a new platform for forward genetic screens and has been made available to the Arabidopsis Biological Resource Center (ABRC). Optimized procedures for amiRNA screening and controls are described.
# Introduction
The presence of large gene families in plants, including Arabidopsis (Arabidopsis Genome [bib_ref] Analysis of the genome sequence of the flowering plant Arabidopsis thaliana, Arabidopsis Genome Initiative [/bib_ref] , leads to functional genetic redundancies or partial functional overlap among closely related genes. Functional overlap and partial or complete redundancy between different family members has been proposed to provide a buffer for loss or gain of function mutation events and mechanistic robustness of cellular networks [bib_ref] Distributed robustness versus redundancy as causes of mutational robustness, Wagner [/bib_ref]. This is considered to be a main reason for the lack of observable phenotypes in single-gene deletion mutants and increasing severity of phenotypes in higher order mutants of homologous genes [bib_ref] Regulators of PP2C phosphatase activity function as abscisic acid sensors, Ma [/bib_ref] [bib_ref] Abscisic acid inhibits type 2C protein phosphatases via the PYR/PYL family of..., Park [/bib_ref]. Identification and characterization of functionally overlapping genes in genetic screens is limited, as is evident from the relatively low number (591 of all Arabidopsis genes) of genes not associated with a single mutant phenotype [bib_ref] A comprehensive dataset of genes with a loss-of-function mutant phenotype in Arabidopsis, Lloyd [/bib_ref]. Analysis of genome-wide gene family definitions showed that the Arabidopsis genome includes over 22 000 genes belonging to gene families [bib_ref] A genomic-scale artificial microRNA library as a tool to investigate the functionally..., Hauser [/bib_ref]. Strategies and tools have been developed to enable screens of the functionally redundant gene space. Recently, an artificial microRNA (amiRNA)-based computational design approach was introduced [bib_ref] A genomic-scale artificial microRNA library as a tool to investigate the functionally..., Hauser [/bib_ref]. AmiRNAs designed to specifically target diverse combinations of gene family members or combinations of subfamily members enable the screening of partial overlapping homologous gene functions at a genome-wide scale. The presented platform also provides an approach for the capture of homologous gene silencing phenotypes, for which higher order loss of function mutants would lead to lethality, as illustrated by a mutant identified here.
Here, we report the generation of over 11 000 T2 amiRNA lines and 3000 additional amiRNA lines by transformation of Arabidopsis Col-0 with a previously published amiRNA library [bib_ref] A genomic-scale artificial microRNA library as a tool to investigate the functionally..., Hauser [/bib_ref] and screening of T2 amiRNA lines for abscisic acid (ABA)-insensitive seed germination phenotypes or plants with low-CO 2 -insensitive high-leaf-temperature phenotypes. Methods are described to identify robust amiRNA mutants and how to avoid pitfalls of this approach. The screen rapidly identified two amiRNAs that target three PYR/RCAR ABA receptor- [bib_ref] Regulators of PP2C phosphatase activity function as abscisic acid sensors, Ma [/bib_ref] [bib_ref] Abscisic acid inhibits type 2C protein phosphatases via the PYR/PYL family of..., Park [/bib_ref] or six SNF1-related kinase-(SnRK2; [bib_ref] Arabidopsis OST1 protein kinase mediates the regulation of stomatal aperture by abscisic..., Mustilli [/bib_ref] [bib_ref] ABA-activated SnRK2 protein kinase is required for dehydration stress signaling in Arabidopsis, Yoshida [/bib_ref] [bib_ref] Arabidopsis mutant deficient in 3 abscisic acidactivated protein kinases reveals critical roles..., Fujii [/bib_ref] encoding genes known to be involved in ABA-mediated control of seed germination. One candidate line that shows an ABA-insensitive seed germination phenotype contains an amiRNA that targets three genes of unknown function, which are annotated as Avirulenceinduced gene 2 (AIG2A; AT3G28930), Avirulence-induced gene 2-like protein A (AIG2LA; AT5G39720), and Avirulence-induced gene 2-like protein B (AIG2LB; AT5G39730). One amiRNA that causes a low-CO 2 -insensitive high-leaf-temperature phenotype targets two genes encoding proteasomal α2-subunits, annotated as PAB1 (AT1G16470) and PAB2 (AT1G79210), for which double mutation causes lethality. New amiRNA lines that target the gene for proteasomal α7-subunit, annotated as PAG1 (AT2G27027), were constructed resulting in a similar stomatal phenotype. Together these observations indicate a rate-limiting role of the intact proteasome for stomatal opening responses.
# Materials and methods
## Plant material, growth conditions and transformation
Arabidopsis accession Columbia-0 was used as the background for all amiRNA transformations of the library. Surface-sterilized seeds (15 min 70% ethanol, 0.1% sodium dodecyl sulfate; three to four washes with ~100% ethanol; alternatively 10 min 50% bleach, 0.05% Tween-20; four to six washes with water; [bib_ref] Standardized method for high-throughput sterilization of Arabidopsis seeds, Lindsey [/bib_ref] of Arabidopsis were coldtreated for 2-5 d at 4 °C and germinated on half-strength Murashige and Skoog basal medium supplemented with Gamborg's vitamins (Sigma-Aldrich [bib_ref] A revised medium for rapid growth and bio assays with tobacco tissue..., Murashige [/bib_ref] [bib_ref] Nutrient requirements of suspension cultures of soybean root cells, Gamborg [/bib_ref] , 0.8% Phytoagar (Difco, Franklin Lakes, NJ, USA) and pH adjusted (pH 5.8; 2.6 mM MES titrated with potassium hydroxide). After 5-7 d, plants were transferred to plastic pots containing sterilized premixed soil (Sunshine Professional Blend LC1 RS; Sunshine; supplemented with an appropriate amount of insecticide (Marathon, Gnatrol)) and propagated under the following conditions: long day (16 h light/8 h dark); 23-27 °C; 20-70% humidity, 60-100 mmol m −2 s −1 light.
Plant transformation by floral dip was performed as described elsewhere [bib_ref] Floral dip: a simplified method for Agrobacterium-mediated transformation of Arabidopsis thaliana, Clough [/bib_ref] with the following modifications. Agrobacterium tumefaciens GV3101::pMP90 [bib_ref] The promoter of TL-DNA gene 5 controls the tissue-specific expression of chimaeric..., Koncz [/bib_ref] was grown under selection of all markers, i.e. genomic (rifampicin), Ti-plasmid (gentamicin), pSOUP (tetracycline) and T-DNA plasmid (spectinomycin). The infiltration medium for resuspension of the bacteria and floral dip contained 5% sucrose (w/v) and 0.02% (v/v) Silwet L-77 [bib_ref] Floral dip: a simplified method for Agrobacterium-mediated transformation of Arabidopsis thaliana, Clough [/bib_ref].
Large scale transformation with the amiRNA library pools [bib_ref] A genomic-scale artificial microRNA library as a tool to investigate the functionally..., Hauser [/bib_ref] was performed as described elsewhere [bib_ref] Random GFP∷cDNA fusions enable visualization of subcellular structures in cells of Arabidopsis..., Cutler [/bib_ref] with the following modifications. One microgram of DNA of each amiRNA sublibrary [bib_ref] A genomic-scale artificial microRNA library as a tool to investigate the functionally..., Hauser [/bib_ref] was electroporated into a total of 500 µl electrocompetent A. tumefaciens cells. The 20 bp and 21 bp amiRNA sublibrary variants for each pool were individually electroporated and combined at this stage. After 2 h at 30 °C in non-selective Luria-Bertani-Miller medium (LB, Teknova), the cells were spread on 20 LB plates (1.5% agar; 150 mm diameter) containing all the appropriate antibiotics (rifampicin, gentamycin, tetracycline, spectinomycin) and grown for 3 d at 30 °C. The bacteria were scraped from the plates, resuspended in 5 ml LB and concentrated by centrifugation for 20 min at 5855 g. Plants were transformed by spraying the flowers with this suspension of the bacteria in infiltration medium (adjusted to an optical density at 600 nm of 0.5) twice with 1 week between the treatments. T1 plants were selected on plates supplemented with 75 µM phosphinotricin or directly on soil by spraying diluted herbicide (1000× dilution, Finale ® ; Bayer, Research Triangle Park, NC, USA) four times with 2-7 d between the treatments. Herbicide-resistant plants were transferred to soil and grown to full maturity and T2 seeds collected from individual plants. When appropriate, media for growth of bacteria or plant selection contained the following concentrations of antibiotics (mg ml −1 ): carbenicillin 100, gentamycin 25, kanamycin 30, rifampicin 50, spectinomycin 100, tetracycline 10, and phosphinotricin 15.
Screen for abscisic acid-insensitive cotyledon emergence phenotype T2 plants were screened individually for insensitivity of seed germination to ABA in 96-well plates (100 µl 0.1% agarose supplemented with 2 µM (±)-ABA, Sigma-Aldrich). Approximately 10-20 seeds were used from each T2 plant. For the pooled screening, approximately 10-50 seeds of 90 individual T2 plants were mixed, surface sterilized and sprinkled onto agar plates (3 μM (±)-ABA; Sigma-Aldrich). As control for ABA insensitivity, abi4-1 (ABRC, CS8104) or abi5-1 (ABRC, CS8105) was used and Col-0 was used as a wildtype control. A putative ABA-insensitive phenotype was scored in a binary manner for similarity to the abi4-1 or abi5-1 phenotype and difference from wild-type after 5-8 d using green cotyledons as indicator [bib_ref] The protein phosphatase AtPP2CA negatively regulates abscisic acid signal transduction in Arabidopsis,..., Kuhn [/bib_ref]. For lines that showed a putative ABA insensitivity, the seed germination assay was repeated by propagating individual T2 seedlings to the next generation (T3) and using seeds of the T3 generation for ABA sensitivity assays. This time, seeds were placed on plates with and without ABA (2 µM (±)-ABA; Sigma-Aldrich) and images were taken daily for 7 d and emergence of radicles and cotyledons was counted manually using Fiji [bib_ref] Fiji: an opensource platform for biological-image analysis, Schindelin [/bib_ref]. For candidates of the individual screen the T2 seeds were used for the repetition of the germination assay.
For candidates of the pooled screen ABA-insensitive seedlings were transferred to plates containing 75 µM phosphinotricin, and after 7-10 d resistant seedlings were transferred to soil, grown up to full maturity, and the T3 seeds used for the validation of the ABA-insensitive germination phenotype.
## Screen for co 2 -insensitive leaf temperature phenotype
Seeds of T2 plants were germinated in 96-pot-flats (254 mm×508 mm; East Jordan Plastics, East Jordan, MI, USA) on soil with each pot containing seeds from one plant. After 7 d, seedlings were sprayed with a 1000× dilution of Finale ® (Bayer), and 2-3 d later pale seedlings were removed and only one healthy dark green seedling was left per pot. After 19 d under standard growth conditions, the plants were exposed to 150 ppm CO 2 for 2 h in a Percival growth chamber. A first set of thermal images was taken with a FLIR A320 thermal imaging camera (FLIR, Wilsonville, OR, USA). Subsequently the plants were exposed to ≥ 800 ppm CO 2 and after 2 h a second set of thermal images was taken. Control plants included in the experiments were ht1-2 , ost1-3 and wild-type Col-0. Thermal images were converted into Flexible Image Transport System format (fits) using the ExaminIR software (FLIR). For the screen using the 96-pot-flat format, the temperature of plant leaves and the surrounding soil were measured using Fiji [bib_ref] Fiji: an opensource platform for biological-image analysis, Schindelin [/bib_ref]. The soil temperature served as a location-specific reference to compensate for temperature variation depending on the position in the 96-pot flat. Either the temperature difference between plant leaves and surrounding soil or the average temperature of plant leaves was used as a quantitative measure. Plants with more than 1 °C difference from soil were considered as primary candidates and subject to further testing. The high-temperature leaf phenotype of ht1-2 was used as a reference for CO 2 insensitivity. To test the reproducibility of the CO 2dependent leaf temperature phenotype of putative candidates, T2 plants were grown in triplicate and assayed again alongside with ht1-2 and wildtype control plants.
## Identification of amirna sequences and testing reproducibility
Genomic DNA from candidates with a robust and reproducible phenotype was isolated as described elsewhere [bib_ref] A simple and rapid method for the preparation of plant genomic DNA..., Edwards [/bib_ref] and the sequence of the amiRNA present was determined by sequencing of the PCR product (primers pha2804f and pha3479r; see [fig_ref] Table 1: Overview of the 10 amiRNA libraries as described by [/fig_ref] at JXB online). Using the Target Search function available on the WMD3 website [bib_ref] Gene silencing in plants using artificial microRNAs and other small RNAs, Ossowski [/bib_ref] , putative amiRNA target genes were predicted. For independent confirmation of the phenotype, independent lines were generated by cloning the identified amiRNA into pFH0032 [fig_ref] Table 2: AmiRNA sequences and predicted target genes found in candidate T3 plants showing... [/fig_ref] ; [bib_ref] A genomic-scale artificial microRNA library as a tool to investigate the functionally..., Hauser [/bib_ref] and transforming it into Arabidopsis Col-0. Confirmed phenotypes were further analysed by using single knock-out mutants, higher order mutants generated by crossing and/or generating amiRNAs that target subsets of the initial target genes (see .
## Gas exchange analyses
Stomatal conductance of H 2 O (g s ) was measured in leaves of 5-to 6-week-old plants using portable gas exchange systems (LI-6400 and LI-6800, LI-COR, Lincoln, NE, USA), starting 2 h after growth chamber light onset. For intact single leaf ABA treatments, a LED light source set at 150 μmol m −2 s −1 (10% blue) and a chamber temperature of 21 °C was used. Leaves were equilibrated for 1 h at a relative humidity of 70-72%, airflow of 200 rpm and CO 2 concentration of 400 ppm. After 1 h, steadystate stomatal conductance was recorded 10 min before the addition of ABA to the petioles submerged in water at the indicated concentration. For light-response measurements, plants were kept in the dark for 18 h prior to experiments. Stomatal conductance of a single intact leaf in the dark was recorded for 10 min, followed by red light treatment of 600 μmol m −2 s −1 . After 20 min of red light treatment, additional blue light was applied at 10 μmol m −2 s −1 . The incoming air humidity was kept at 62-65% and air flow at 200 rpm. For stomatal conductance measurements of single intact leaf CO 2 responses, incoming relative air humidity was kept at 62-65% and the imposed changes in CO 2 concentration were applied as indicated. Leaves were attached to intact plants and were equilibrated for 1 h before the measurements. The data presented represent n≥3 leaves with each leaf from independent plants per genotype per treatment.
# Qrt-pcr analysis
Total RNA (500 ng) was reverse transcribed using the first-strand cDNA synthesis kit (GE Healthcare). qRT-PCR analyses were performed using 3-fold-diluted cDNA (Maxima SYBR Green Rox/qPCR Master Mix; Thermo Fisher Scientific) on a CFX Connect PCR cycler (BioRad). The housekeeping PDF2 gene was used as an internal control. The threshold cycle (C T ) was determined by the instrument (CFX Manager Software, BioRad), and the ΔΔC T method was used to calculate the fold change in each gene [bib_ref] Analysis of relative gene expression data using real-time quantitative PCR and the..., Livak [/bib_ref]. For RAB18 gene expression measurements, total RNA was extracted from 2-week-old seedlings that were treated with ABA for 9 h and final concentration of 20 µM.
# Results and discussion
## Generation of amirna library plants
We have previously described the generation of an amiRNA library consisting of 10 sublibraries that represent 22 000 individual amiRNA designs [bib_ref] A genomic-scale artificial microRNA library as a tool to investigate the functionally..., Hauser [/bib_ref]. Deep sequencing of these 10 sublibraries showed that ≥95% of the designed amiRNAs were present in these sublibraries [bib_ref] A genomic-scale artificial microRNA library as a tool to investigate the functionally..., Hauser [/bib_ref]. The amiRNA library was transformed first into Agrobacterium tumefaciens and then into Arabidopsis Col-0. Over a period of over 4 years, the amiRNA library consisting of 10 sublibraries was transformed and T1 seeds harvested. Using plate-or soil-based selection methods, herbicide-resistant T1 plants were grown and T2 seeds from over 14 000 individual plants were harvested [fig_ref] Table 1: Overview of the 10 amiRNA libraries as described by [/fig_ref]. The transformation rate varied over a range from 0.08% to 0.76% with an average of 0.25%. During the course of this research, approximately 3000 additional T2 lines [bib_ref] A transportome-scale amiRNA-based screen identifies redundant roles of Arabidopsis ABCB6 and ABCB20..., Zhang [/bib_ref] were generated expressing amiRNAs that target homologous transporter-encoding gene family members. These 3000 lines will also be made available to the ABRC, such that in the end over 14 000 total T2 lines will be submitted for use by the community.
## Screen for aba-insensitive seed germination phenotype
In total over 2500 T2 amiRNA lines were screened individually and over 5000 T2 amiRNA lines were screened in pools for ABA-insensitive germination phenotypes [fig_ref] Figure 1: Overview of primary screen performed with over 7500 T2 amiRNA lines [/fig_ref]. In the primary screen using individual plants in a 96-well plate format, 59 putative candidates were identified. In the primary screen using pools of 90 plants with 25-80 seeds per line, approximately 340 putative candidates representing an unknown number of lines were identified [fig_ref] Figure 1: Overview of primary screen performed with over 7500 T2 amiRNA lines [/fig_ref].
These candidates were subjected to further analysis in a secondary screen [fig_ref] Figure 2: Overview of secondary ABA insensitivity screen performed with T3 candidate amiRNA lines... [/fig_ref]. The cotyledon emergence phenotype of 24 T3 seedlings from a total of 76 retested plants showed a more reduced ABA sensitivity that was clearly different from wildtype and less severe than the abi4-1 and abi5-1 controls [fig_ref] Figure 2: Overview of secondary ABA insensitivity screen performed with T3 candidate amiRNA lines... [/fig_ref]. From the 59 putative candidates identified using the individual screening approach, the amiRNA line p8l1257 showed a reproducible partial insensitivity to ABA in the T3 generation [fig_ref] Figure 3: Avirulence-induced genes [/fig_ref]. Only the amiRNA in candidates with the most robust phenotypes was determined by sequencing. Two of the amiRNAtargeted gene sets identified in 24 seedlings with reproducible phenotypes are known core components of the ABA signal transduction cascade [fig_ref] Figure 2: Overview of secondary ABA insensitivity screen performed with T3 candidate amiRNA lines... [/fig_ref]. These include amiRNA lines that target the three ABA receptors PYR1 (RCAR11), PYL4 (RCAR10), and PYL6 (RCAR9) [fig_ref] Figure 2: Overview of secondary ABA insensitivity screen performed with T3 candidate amiRNA lines... [/fig_ref] , C; [fig_ref] Table 2: AmiRNA sequences and predicted target genes found in candidate T3 plants showing... [/fig_ref]. Furthermore, amiRNA-expressing plants that target six members of the SnRK2 protein kinase family [bib_ref] ABA-activated SnRK2 protein kinase is required for dehydration stress signaling in Arabidopsis, Yoshida [/bib_ref] [bib_ref] Arabidopsis mutant deficient in 3 abscisic acidactivated protein kinases reveals critical roles..., Fujii [/bib_ref] were isolated in this screen, including the three SnRK2 protein kinases, SnRK2.2, SnRK2.3, and SnRK2.6 (OST1), that are known to be required for abscisic acid signaling [fig_ref] Figure 2: Overview of secondary ABA insensitivity screen performed with T3 candidate amiRNA lines... [/fig_ref] , C; [fig_ref] Table 2: AmiRNA sequences and predicted target genes found in candidate T3 plants showing... [/fig_ref] ; [bib_ref] Arabidopsis OST1 protein kinase mediates the regulation of stomatal aperture by abscisic..., Mustilli [/bib_ref] [bib_ref] ABA-activated SnRK2 protein kinase is required for dehydration stress signaling in Arabidopsis, Yoshida [/bib_ref] [bib_ref] Arabidopsis mutant deficient in 3 abscisic acidactivated protein kinases reveals critical roles..., Fujii [/bib_ref].
Notably, [fig_ref] Figure 2: Overview of secondary ABA insensitivity screen performed with T3 candidate amiRNA lines... [/fig_ref] shows a strong variation in the cotyledon emergence phenotype among plants expressing the same amiRNA that targets six SnRK2 kinase transcripts. This variation might be responsible for the high number of variable candidates that did not show a robust phenotype following the primary screen. Additional amiRNA lines were isolated as putative mutants and the amiRNA sequence was determined . Although some of the predicted targets might be expected to affect abscisic acid responses, rescreening of these putative mutants did not show consistently robust reproducible phenotypes. Thus, amiRNAs appear to produce phenotypes that may be variable even within the same line. These findings [bib_ref] A genomic-scale artificial microRNA library as a tool to investigate the functionally..., Hauser [/bib_ref] , the number of amiRNAs designed for each library, and the number of individual T2 amiRNA transformants that were generated here led us to develop a protocol in which: (i) only putative mutants that showed a consistent phenotype when screening seeds from the next generation of plants were selected, and (ii) Only lines that showed similar phenotypes upon re-transformation with new amiRNAs that are predicted to target the same genes were selected. Furthermore, based on the variation observed here in the secondary screen it is advisable to investigate over 10 independent transformed lines [bib_ref] Highly specific gene silencing by artificial microRNAs in Arabidopsis, Schwab [/bib_ref] [bib_ref] A genomic-scale artificial microRNA library as a tool to investigate the functionally..., Hauser [/bib_ref] in the future to determine which amiR-NAs produce phenotypes that can be carried forward. The isolation of amiRNA lines targeting functionally overlapping PYR/RCAR ABA receptor and SnRK2 protein kinase genes, which could not be isolated in traditional forward genetic mutant screens, provides a proof of principle that functionally redundant genes can be isolated in forward genetic screening using this new amiRNA resource. The inclusion of control lines and the validation steps described above should enable screening for diverse phenotypes using the lines generated here that are being provided to ABRC.
AmiRNA lines targeting three avirulence-induced genes show partial insensitivity to ABA inhibition of seed germination but not to ABA-induced stomatal closure
The amiRNA in line p8l1257 isolated in the present screen targets a new set of three genes [fig_ref] Figure 3: Avirulence-induced genes [/fig_ref]. Previous research [fig_ref] Table 2: AmiRNA sequences and predicted target genes found in candidate T3 plants showing... [/fig_ref]. Wild-type control (Col-0) and abi4-1 and abi5-1 as reference for insensitivity are shown in the other two rows. (C) Time course of cotyledon emergence in the presence of ABA for wild-type control (wt), abi4-1 as a reference for ABA insensitivity and two representative amiRNA lines that target a set of six SnRK2 protein kinase genes (amiRNA-SnRK2) or a set of three PYR/RCAR ABA receptor genes (amiRNA-PYR/ RCAR). Per genotype 74 ± 46 seeds were phenotyped. (D) Bar plot of variation of cotyledon emergence phenotype (day 6; 2 µM ABA) in the T3 generation of plants isolated in the primary screen that were selected as candidates based on their seed germination phenotype in the T2 generation. Sequencing of the amiRNA confirmed that all 18 of these amiRNA-expressing plants contain an amiRNA that targets a set of six SnRK2 kinase genes (amiRNA-SnRK2; see [fig_ref] Table 2: AmiRNA sequences and predicted target genes found in candidate T3 plants showing... [/fig_ref]. Wild-type control (wt, Col-0) and abi4-1 as reference for ABA insensitivity are shown.
annotated these genes based on their mRNA upregulation in a transcriptomic study after infection with Pseudomonas syringae pv maculicola carrying avrRpt2 (avrRpt2-induced gene, AIG2) [bib_ref] Isolation of Arabidopsis genes that differentiate between resistance responses mediated by the..., Reuber [/bib_ref]. However, these genes have not been previously described to be involved in ABA-mediated control of seed germination or other phenotypic responses in plants.
The line p8l1257 was named amiRNA-AIG here and was further tested by analysing seed germination with additional T2 generation seeds from the original p8l1257 stock [fig_ref] Figure 3: Avirulence-induced genes [/fig_ref]. Germination properties were compared with a control amiRNA line targeting the human myosin 2 (amiRNA-HsMYO), which has no targeted genes in Arabidopsis [bib_ref] A genomic-scale artificial microRNA library as a tool to investigate the functionally..., Hauser [/bib_ref]. After 12 d on plates containing 2 µM ABA, the amiRNA-AIG line showed cotyledon greening in contrast to the control amiRNA-HsMYO line [fig_ref] Figure 3: Avirulence-induced genes [/fig_ref]. The effect of the amiRNA-AIG on the expression of a known ABA-induced gene, RAB18, was analysed by qRT-PCR (see [fig_ref] Figure 1: Overview of primary screen performed with over 7500 T2 amiRNA lines [/fig_ref]. The ABA-mediated induction of RAB18 expression was substantially reduced in the amiRNA-AIG line indicating a role of the targeted avirulence-induced genes (AIG2s) in ABA signal transduction.
[formula] 18 TGGATATGCTCCAACCGGCAT AT1G10940 SNRK2.4 AT1G60940 SNRK2.10 AT2G23030 SNRK2.9 AT3G50500 SNRK2.2 AT4G33950 SNRK2.6 (OST1) AT5G66880 SNRK2.3 5 TATCAACGACGTAAGACTCGT AT2G38310 PYL4 (RCAR10) AT2G40330 PYL6 (RCAR9) AT4G17870 PYR1 (RCAR11) 1 TTAATACATGGATGCACACGT AT3G28930 AIG2 AT5G39720 AIG2LA AT5G39730 AIG2LB [/formula]
Since two out of the three genes are tandemly repeated, generation of double mutants using T-DNA insertion knockouts would be limited. Therefore, five new amiRNA lines were generated that target subsets of genes targeted by the original amiRNA-AIG to verify the relevance of the predicted AIG2 target genes. AmiRNA1, 2 and 3 targeted each a single AIG2 [fig_ref] Figure 3: Avirulence-induced genes [/fig_ref]. AmiRNA4 targeted two tandem-repeat AIG2 genes and amiRNA5 targeted all three AIG genes targeted in the original amiRNA-AIG line, but with a different amiRNA sequence [fig_ref] Figure 3: Avirulence-induced genes [/fig_ref] ; for amiRNA sequences). When the T2 seeds expressing these five new amiRNAs were tested in a seed germination assay with 0.5 µM ABA, only the amiRNA4 and amiRNA5expressing lines showed less sensitivity to ABA compared with the control amiRNA-HsMYO line in cotyledon greening [fig_ref] Figure 3: Avirulence-induced genes [/fig_ref]. The expression of all three putative target genes (AT5G39720, AT5G39730, AT3G28930) was analysed using qRT-PCR in the originally isolated amiRNA-AIG line and in all the amiRNA lines 1-5 (see [fig_ref] Figure 2: Overview of secondary ABA insensitivity screen performed with T3 candidate amiRNA lines... [/fig_ref]. The amiRNA efficiency of transcriptional inhibition varies between the lines, target transcript(s) and amiRNA sequence. Note that microRNA silencing in plants occurs via two mechanisms, (i) the degradation of transcripts and (ii) inhibition of translation [bib_ref] Widespread translational inhibition by plant miRNAs and siRNAs, Brodersen [/bib_ref]. Thus, quantification of targeted transcripts may not fully show the degree of silencing of target genes. Combined, these data provide evidence that the original amiRNA-AIG phenotype is attributable to silencing of more than one AIG2 gene, suggesting overlapping homologous gene functions.
The original amiRNA-AIG line was also investigated to determine if it affects ABA-induced stomatal closure using an intact leaf gas exchange analysis approach. When ABA was applied to the transpiration stream of intact leaves at a final concentration of 2 µM, both the control amiRNA-HsMYO line and the amiRNA-AIG line showed an ABA-induced decrease in stomatal conductance to H 2 O (g s , [fig_ref] Figure 4: The isolated amiRNA line targeting three AIG2 genes [/fig_ref]. Normalization of the stomatal conductance data showed no dramatic difference in ABA-induced stomatal closure between amiRNA-HsMYO and amiRNA-AIG [fig_ref] Figure 4: The isolated amiRNA line targeting three AIG2 genes [/fig_ref]. Together, the present data show that the isolated amiRNA-AIG line is less sensitive to ABA inhibition of seed germination.
The AIG2 genes are functionally annotated as putative γ-glutamyl cyclotransferases (GGCTs, EC:4.3.2.9) based on their similarity to the human orthologue (HsGGCT; O75223). AIG2LA and AIG2LB share only 16% and 17% identity, respectively, to the human orthologue. GGCTs have been described to cleave γ-glutamyl-amino acid dipeptides to release the free amino acid and 5-oxoproline [bib_ref] The identification and structural characterization of C7orf24 as γ-glutamyl cyclotransferase. An essential..., Oakley [/bib_ref]. Further research will be required to determine the mechanism by which AIG2s affect ABA inhibition of seed germination.
## Screen for co 2 -insensitive leaf temperature phenotype
In total, over 2500 T2 amiRNA lines were screened individually for an altered leaf temperature response to a low CO 2 concentration (150 ppm) by infra-red thermal imaging . Leaf temperature depends on various parameters including radiation absorption, air temperature, and humidity . Low ambient CO 2 concentration leads to stomatal opening in Arabidopsis, causing an increased transpiration rate and thus a decrease in leaf temperature compared with the surrounding air. Mutants impaired in CO 2 -induced stomatal opening appear warmer compared with wild-type plants. In the screen, we used the soil temperature as reference to compensate for the local temperature differences due to various factors including humidity of the soil. Wild-type plants and plants of the HIGH LEAF TEMPERATURE1-2 (ht1-2) mutant were included in all trays as controls. Based on visual inspection of the thermal images, plants with relatively higher leaf temperature under low [CO 2 ] compared with the other plants in the same image were selected and the difference between the average leaf temperature and the surrounding soil was determined. The difference between leaf temperature and soil temperature was determined as reference for overall temperature and to compensate for local temperature differences. A set of 106 plants with more than one degree difference between the leaf temperature and the surrounding soil was defined as initial putative candidates for further testing (see Methods for details). For the rescreening of putative mutants, we set a high threshold for temperature differences in the selection of mutants compared with the wild-type strain of 1 °C. The constitutive CO 2 response mutant ht1-2, when exposed to low [CO 2 ], shows a delta temperature above 1 °C between leaf and soil. Rescreening of these candidates in the T2 generation revealed an amiRNA line (p9l22) with a robust and reproducible impaired response to low CO 2 .
After exposure to low [CO 2 ], the leaf temperature of the p9l22 line was compared with wild-type (Col-0) and with the constitutive high-CO 2 -response mutant ht1-2 [fig_ref] Figure 6: The amiRNA line p9l22 is defective in light-and low-CO 2 -induced stomatal... [/fig_ref] ; . The leaves of the p9l22 line had a higher temperature than wild-type leaves and a similar temperature to ht1-2 leaves [fig_ref] Figure 6: The amiRNA line p9l22 is defective in light-and low-CO 2 -induced stomatal... [/fig_ref]. Stomatal index (SI) and density (SD) were calculated for wild-type, the control amiRNA-HsMYO, and p9l22 lines. No noteworthy differences were found between the genotypes (Supplementary [fig_ref] Figure 3: Avirulence-induced genes [/fig_ref] ; amiRNA-HsMYO versus p9l22 line, one-way ANOVA, P>0.05 for SI and SD).
To measure [CO 2 ] responses in a time-resolved fashion, we measured stomatal conductance (g s ) using a gas exchange analyser. In the amiRNA-HsMYO control line, the shift from ambient (400 ppm) to low (150 ppm) [CO 2 ] led to a rapid increase in stomatal conductance [fig_ref] Figure 6: The amiRNA line p9l22 is defective in light-and low-CO 2 -induced stomatal... [/fig_ref]. AmiRNA line p9l22 responded to the same treatment with a lower magnitude of stomatal opening [fig_ref] Figure 6: The amiRNA line p9l22 is defective in light-and low-CO 2 -induced stomatal... [/fig_ref]. Both lines showed stomatal closure in response to high (800 ppm) [CO 2 ] exposure at similar rates [fig_ref] Figure 6: The amiRNA line p9l22 is defective in light-and low-CO 2 -induced stomatal... [/fig_ref]. To test whether line p9l22 is defective in response to other stimuli that cause stomatal opening, light-induced g s responses were investigated [fig_ref] Figure 6: The amiRNA line p9l22 is defective in light-and low-CO 2 -induced stomatal... [/fig_ref]. The control amiRNA-HsMYO and p9l22 lines were kept in the dark for 18 h prior to the experiments and steady-state g s was measured. When red light (at 600 µmol m −2 s −1 ) was applied, the p9l22 line showed a reduced rate of g s increase when compared with the control line. The same was observed when blue light (at 10 µmol m −2 s −1 ) was superimposed on the red light background [fig_ref] Figure 6: The amiRNA line p9l22 is defective in light-and low-CO 2 -induced stomatal... [/fig_ref]. Thus, the amiRNA line p9l22 causes reduced responses to low CO 2 concentration, red light, and blue light.
The amiRNA in the p9l22 line was sequenced and is predicted to target two homologous proteasomal subunit genes (PAB1, AT1G16470; and PAB2, AT1G79210). PAB1 and PAB2 are the sole genes that encode the 20S proteasome α2 subunit [bib_ref] The proteasome: paradigm of a self-compartmentalizing protease, Baumeister [/bib_ref]. First, we attempted to isolate a double mutant (pab1 pab2) using T-DNA insertion lines (SALK_099950 and SALK_144987; [bib_ref] Genome-wide insertional mutagenesis of Arabidopsis thaliana, Alonso [/bib_ref]. After genotyping over 100 plants in the F2 generation, no homozygous double mutant was recovered. We concluded that the double mutant is very likely lethal.
Alternatively, a new amiRNA sequence targeting solely the PAB1 gene was cloned and transformed into the pab2-1 single mutant (SALK_144987). This new amiRNA line, pab2-1mut pab1amiRNA, was investigated in stomatal conductance analyses of [CO 2 ] responses [fig_ref] Figure 7: New amiRNA lines targeting the two PAB genes [/fig_ref]. Leaves were first exposed to high (900 ppm) [CO 2 ] for 1 h and steady-state g s was recorded. Shifts to low (150 ppm) [CO 2 ] led to an increase in g s in both the pab2-1mut pab1amiRNA line and the control amiRNA-HsMYO line [fig_ref] Figure 7: New amiRNA lines targeting the two PAB genes [/fig_ref]. The normalized stomatal conductance data show that the pab2-1 amiRNA-PAB1 line responds to low [CO 2 ] with a reduced magnitude compared with the control line [fig_ref] Figure 7: New amiRNA lines targeting the two PAB genes [/fig_ref].
Initial experiments were pursued to determine if modifications in the α-ring of the 20S proteasome might be linked to the above phenotypes, or whether this mutation is specific to only α2 subunit mutations of the proteasome. The α-ring of the 20S proteasome is composed of seven α-subunits [bib_ref] Structure, function and regulation of plant proteasomes, Kurepa [/bib_ref]. The p9l22 amiRNA targets the only two genes that encode the α2 subunit of the proteasome [fig_ref] Figure 7: New amiRNA lines targeting the two PAB genes [/fig_ref] , inset highlighted in red). To determine whether other α-subunits also affect the response to low [CO 2 ], a second amiRNA line was generated, which targets the PAG1 gene (α7 subunit, inset in [fig_ref] Figure 7: New amiRNA lines targeting the two PAB genes [/fig_ref] , . Thermal imaging screen for mutants with impaired response to low CO 2 identified the amiRNA line p9l22. (A) Thermal images of amiRNA lines in the primary screen after exposure to low CO 2 (150 ppm). Plants were grown in 96-pot flats under ambient CO 2 and after 26 d exposed to low (150 ppm) CO 2 for 2 h and then thermal images of the entire flat were taken in eight separate images per flat. An image is shown from the primary screen in which the plant p9l22 (white box) was flagged as a candidate with a putative altered response to CO 2 based on the leaf temperature. The average leaf temperature was computationally calculated across all rosette leaves for flagging putative mutants. (B) Differences of leaf temperature (T plant ) and surrounding soil (T soil ) for putative mutants exposed to low [CO 2 ] (150 ppm) for 2 h. Average leaf temperature was computed by image analysis of the most fully expanded rosette leaves. Bars show average ±SD (n=3-5 leaves). WT (Col-0) (orange) and ht1-2 (green) were used as control.
named amiRNA-PAG1. The α7 subunit is encoded by a single gene in Arabidopsis [bib_ref] Structure, function and regulation of plant proteasomes, Kurepa [/bib_ref]. When a amiRNA-PAG1 line was tested in g s responses to [CO 2 ] shifts, it showed a lower rate of stomatal opening when compared with the control amiRNA-HsMYO line [fig_ref] Figure 7: New amiRNA lines targeting the two PAB genes [/fig_ref] one-way ANOVA P>0.05). The expression levels of PAB1, PAB2, and PAG1 were analysed in the p9l22 amiRNA line and also in pab2-1mut pab1amiRNA and amiRNA-PAG1 lines using qRT-PCR (see [fig_ref] Figure 4: The isolated amiRNA line targeting three AIG2 genes [/fig_ref]. With the exception of the severely reduced PAB2 expression in the pab2-1mut pab1amiRNA line when compared with control lines, no clear evidence for knock down at the transcriptional level could be detected in the amiRNA lines, which may point to amiRNAmediated inhibition of translation [bib_ref] Widespread translational inhibition by plant miRNAs and siRNAs, Brodersen [/bib_ref].
The present findings show that the p9l22 amiRNA line is also partially impaired in red light-induced stomatal opening. Red light mediates stomatal opening in part via activation of photosynthesis and the resulting drop in internal concentration of CO 2 (C i ) [bib_ref] CO 2 provides an intermediate link in the red light response of..., Roelfsema [/bib_ref] [bib_ref] The HT1 protein kinase is essential for red light-induced stomatal opening and..., Matrosova [/bib_ref]. Inserts show a representation of the 26S proteasome, with the α2 subunits highlighted in red and α7 subunits highlighted in green. The initial slope of g s response for amiRNA-HsMYO and amiRNA-PAG1 was calculated and one-way ANOVA was used to compare the values (P>0.05).
In addition, the p9l22 line is also partially impaired in blue light-induced stomatal opening. This suggests that a general regulator of stomatal opening is impaired in this line. As the proteasome mediates the degradation of proteins and reduced functions of α-ring subunits are predicted to increase protein levels, it is tempting to speculate that the phenotype observed might be correlated with an increased abundance of a negative regulator of stomatal opening. Further research will be required to test this or other hypotheses. In other studies, the 26S α2 subunit, when overexpressed, enhanced thermotolerance and adaptation in rice and Arabidopsis, suggesting that proteasomal subunits can have rate-limiting roles in regulating plant physiological responses [bib_ref] Natural alleles of a proteasome α2 subunit gene contribute to thermotolerance and..., Li [/bib_ref].
## Summary and future use
A library of over 11 000 plus 3000 additional T2 generation amiRNA lines was created as a new resource to screen the redundant gene space in Arabidopsis. These amiRNAexpressing lines are being provided as individual lines to the Arabidopsis Biological Resource Center (ABRC). Given the observations and findings in the present study, lines will be available for high-throughput screening in pools of 90 lines per pool with approximately 25-50 seeds per individual amiRNA line in each pool. In each pool, the pooled seeds for screening will originate from one of the 10 sublibraries that target gene family members with defined functional annotations [fig_ref] Table 1: Overview of the 10 amiRNA libraries as described by [/fig_ref] [bib_ref] A genomic-scale artificial microRNA library as a tool to investigate the functionally..., Hauser [/bib_ref]. This approach will increase the probability of identifying interesting putative mutants in future screens despite the biological variability in amiRNA silencing lines found here [fig_ref] Figure 2: Overview of secondary ABA insensitivity screen performed with T3 candidate amiRNA lines... [/fig_ref].
The screen for ABA-insensitive seed germination phenotypes identified two amiRNAs targeting PYR/RCAR ABA receptor genes and SnRK2 genes, which are both known groups of redundant key genes and proteins required for ABA signal transduction [bib_ref] Regulators of PP2C phosphatase activity function as abscisic acid sensors, Ma [/bib_ref] [bib_ref] Abscisic acid inhibits type 2C protein phosphatases via the PYR/PYL family of..., Park [/bib_ref]. Isolating amiRNA lines in these known components serves as proof of principle for our approach. Moreover, screening this amiRNA population enables the identification of mutants that require co-silencing of homologous gene family members, which are less likely to be found in forward genetic screens of ethyl methanesulfonate or T-DNA mutagenized seed populations.
Overall the presented amiRNA screen shows that amiRNA lines are prone to showing a high rate of variable candidates with weak or non-robust phenotypes. Nevertheless, as shown here new mutants can be isolated. Furthermore, during the course of this research, this amiRNA library resource has been used to isolate long-sought functionally redundant auxin transporter genes (e.g. ABCB6, ABCB20; [bib_ref] A transportome-scale amiRNA-based screen identifies redundant roles of Arabidopsis ABCB6 and ABCB20..., Zhang [/bib_ref]. Approaches to circumvent the inherent limitations of forward genetic screening with amiRNAs were developed in the present study. As a first step, it is recommended to rescreen the next generation to identify robust and reproducible phenotypes in individually isolated putative mutant lines. As a second step, the amiRNA sequence of confirmed mutant lines needs to be determined (see Methods). AmiRNA sequences linked to the observed phenotypes are retransformed, and testing over 10 independent lines for the phenotype is recommended. Alternatively, amiRNA on one line without break that target a subset of the initially predicted targets can be used to narrow down the causative genes (e.g. [fig_ref] Figure 3: Avirulence-induced genes [/fig_ref]. For cases where only two to three genes are targeted, T-DNA lines or CRISPR/Cas9 mutants can be used to narrow down the genes relevant for the phenotype.
Over 95% of the amiRNAs in this library were designed to target only two to five genes [bib_ref] A genomic-scale artificial microRNA library as a tool to investigate the functionally..., Hauser [/bib_ref] , meaning that identification of causal genes is facilitated. Using the above approach, we report on two newly identified mutants: (i) amiRNA lines targeting three genes encoding avirulence induced gene (2-like) proteins show an ABA-insensitive seed germination phenotype; and (ii) amiRNA lines targeting two proteasomal subunits show insensitivity to low-CO 2 -induced stomatal opening. Further analyses of the two targeted genes in this amiRNA line suggest that stronger T-DNA alleles result in lethality. This indicates the usefulness of the generated amiRNA lines for forward genetic isolation of higher order mutants that would be lethal upon knock-out. Our data suggest that the wild-type expression levels of two α-subunits of the 20S proteasome, α2 and α7, are required for fully functional stomatal opening mediated by physiological stimulation. This indicates that the proteasomal subunits are likely controlling an unknown general negative regulator of stomatal opening. The amiRNA seed resource generated here provides a new and potent tool to identify redundant genes and also lethality causing higher order mutants in many biological processes in Arabidopsis.
In conclusion, the amiRNA library resource is well suited for screening of phenotypes that can be easily verified in subsequent generations. This population may be best suited for screens that permit high throughput or medium throughput screening for phenotypes with a large dynamic range. Many such powerful screens have been performed in classical Arabidopsis mutant screens that were previously not designed to identify functionally redundant genes.
## Supplementary data
Supplementary data are available at JXB online. [fig_ref] Figure 1: Overview of primary screen performed with over 7500 T2 amiRNA lines [/fig_ref]. The induction of RAB18 gene expression by ABA is lower in the amiRNA-AIG lines. [fig_ref] Figure 2: Overview of secondary ABA insensitivity screen performed with T3 candidate amiRNA lines... [/fig_ref]. The expression of AIG2 genes in amiRNA-AIGs lines. [fig_ref] Figure 3: Avirulence-induced genes [/fig_ref]. The p9l22 amiRNA line has normal stomatal indices and density. [fig_ref] Figure 4: The isolated amiRNA line targeting three AIG2 genes [/fig_ref]. The expression of PAB1, PAB2, and PAG1 genes in amiRNA lines. [fig_ref] Table 1: Overview of the 10 amiRNA libraries as described by [/fig_ref]. List of relevant primers used in this study. [fig_ref] Table 2: AmiRNA sequences and predicted target genes found in candidate T3 plants showing... [/fig_ref]. List of relevant plasmids used in this study. . List of new amiRNAs designed and cloned in this study. . AmiRNA sequences and predicted target genes found in candidate plants.
[fig] Figure 1: Overview of primary screen performed with over 7500 T2 amiRNA lines. (A) Approximately 2500 T2 amiRNA lines were screened individually by adding seeds to 96-well plates each well containing 100 µl 0.1% agarose supplemented with 2 µM ABA and approximately 10-20 seeds from one plant. Wild-type (Col-0), abi4-1, and abi5-1 were used as controls. Based on visual inspection of cotyledon greening, around 59 lines were considered as candidates for further testing in the T3 generation. (B) Approximately 5000 T2 amiRNA lines were screened in pools. Each pool contained 10-50 seeds from 90 individually stored amiRNA lines (see main text). Approximately 340 lines were selected for further testing in the T3 generation. [/fig]
[fig] Figure 2: Overview of secondary ABA insensitivity screen performed with T3 candidate amiRNA lines identified in the primary screen that targets known genes involved in ABA signal transduction. (A) Heat map representation of cotyledon emergence time course of the T3 generation obtained from candidates with putative ABA-insensitive seed germination or cotyledon emergence phenotypes. Each row represents the percentage of cotyledon emergence of one individual line. The rows are ordered by hierarchical clustering. Wild-type control (WT, Col-0) and abi4-1 and abi5-1 as reference for ABA insensitivity are shown. (B) Images of seeds germinating on plates in secondary screen containing 2 µM ABA at the indicated time points. Shown are representative plants of amiRNA plants targeting a set of six SnRK2 kinase genes (amiRNA-SnRK2) or a set of three PYR/RCAR ABA receptor genes (amiRNA-PYR/RCAR; see [/fig]
[fig] Figure 3: Avirulence-induced genes (AIG2s) targeted by an amiRNA cause a reduced ABA sensitivity in cotyledon emergence assays. (A) Seedlings of the control amiRNA line, which has no target gene in Arabidopsis (amiRNA-HsMYO control), and an amiRNA line that targets three AIG2 genes (amiRNA-AIG) germinated in the presence of 0 or 2 µM ABA. Photographs were taken after 12 d of exposure. (B) Five new amiRNAs were designed to target the three AIG2(-like) genes. Three of these amiRNAs, amiRNA1, 2 and 3, target a single gene each. AmiRNA4 targets two tandem-repeat AIG2(-like) genes and amiRNA5 targets all three genes at non-identical nucleotides compared with the original amiRNA isolated in the primary screen (amiRNA-AIG, seeSupplementary Table S2for amiRNA sequences). (C) The new T2 amiRNA lines were tested in cotyledon emergence assays. Seedlings were germinated in the presence of 0 or 0.5 µM ABA. Photographs were taken after 4 d of incubation. [/fig]
[fig] Figure 4: The isolated amiRNA line targeting three AIG2 genes (amiRNA-AIG) responds to ABA in whole leaf gas exchange analyses. Time-resolved stomatal conductance to H 2 O (g s ) in response to application of 2 µM ABA to the transpiration stream (red arrows) is shown in the amiRNA control line (amiRNA-HsMYO) and in the amiRNA line targeting three AIG2 genes (amiRNA-AIG). Stomatal conductance was analysed using whole leaf gas exchange. (A) Stomatal conductance in mol m −2 s −1 . (B) Data from (A) were normalized to stomatal conductance at the beginning of the experiment. Data are the mean of n=3 leaves per genotype ±SEM. [/fig]
[fig] Figure 6: The amiRNA line p9l22 is defective in light-and low-CO 2 -induced stomatal opening. (A) The p9l22 line shows an elevated leaf temperature phenotype at low ambient [CO 2 ] treatment (150 ppm). Top, thermal imaging; bottom, image of the same plants. The calibration bar shows the pseudocolored scale for temperature. (B) Time-resolved stomatal conductance responses at the imposed [CO 2 ] shifts (bottom in ppm) in the control amiRNA-HsMYO line and in the p9l22 line were analysed using intact whole-leaf gas exchange. The plots represent average ±SEM (n=4 leaves from four plants per genotype). (C) Time-resolved stomatal conductance responses from darkness to the imposed light intensity and light quality shifts (bottom moon shape: darkness; red dashed line: red light at 600 μmol m −2 s −1 ; and blue dashed line: blue light at 10 μmol m −2 s −1 ) in the control amiRNA-HsMYO line and in the p9l22 line were analysed using intact whole-leaf gas exchange. The plots represent average ±SEM (n=3 leaves from three plants per genotype). [/fig]
[fig] Figure 7: New amiRNA lines targeting the two PAB genes (α2 subunit) and PAG1 gene (α7 subunit) of the 26S proteasome show partial impairment in low-CO 2 -induced stomatal opening. (A, C) Time-resolved stomatal conductance responses at the imposed [CO 2 ] shifts (bottom in ppm) in the control line amiRNA-HsMYO and (A) in the pab2-1mut pab1amiRNA line (pab2 gene T-DNA knockout and pab1 gene amiRNA knockdown, α2 subunit) and (C) amiRNA-PAG1 line (pag1 gene amiRNA knockdown, α7 subunit) were analysed using intact whole-leaf gas exchange. The plots represent average ±SEM (n=3-4 leaves from different plants per genotype). (B, D) Data from (A, C) were normalized to the average stomatal conductance of the first 10 min. [/fig]
[table] Table 1: Overview of the 10 amiRNA libraries as described by [/table]
[table] Table 2: AmiRNA sequences and predicted target genes found in candidate T3 plants showing robust ABA-insensitive seed germination phenotypes in the T2 screen and subsequently in the T3 generation [/table]
|
Goal-oriented cognitive rehabilitation in early-stage dementia: study protocol for a multi-centre single-blind randomised controlled trial (GREAT)
Background: Preliminary evidence suggests that goal-oriented cognitive rehabilitation (CR) may be a clinically effective intervention for people with early-stage Alzheimer's disease, vascular or mixed dementia and their carers. This study aims to establish whether CR is a clinically effective and cost-effective intervention for people with early-stage dementia and their carers. Methods/design: In this multi-centre, single-blind randomised controlled trial, 480 people with early-stage dementia, each with a carer, will be randomised to receive either treatment as usual or cognitive rehabilitation (10 therapy sessions over 3 months, followed by 4 maintenance sessions over 6 months). We will compare the effectiveness of cognitive rehabilitation with that of treatment as usual with regard to improving self-reported and carer-rated goal performance in areas identified as causing concern by people with early-stage dementia; improving quality of life, self-efficacy, mood and cognition of people with early-stage dementia; and reducing stress levels and ameliorating quality of life for carers of participants with early-stage dementia. The incremental cost-effectiveness of goal-oriented cognitive rehabilitation compared to treatment as usual will also be examined.Discussion: If the study confirms the benefits and cost-effectiveness of cognitive rehabilitation, it will be important to examine how the goal-oriented cognitive rehabilitation approach can most effectively be integrated into routine health-care provision. Our aim is to provide training and develop materials to support the implementation of this approach following trial completion. Trial registration: Current Controlled Trials ISRCTN21027481
# Background
There is a greater need than ever before to identify effective and beneficial interventions for people with early-stage dementia. There are thought to be over 750,000 people with dementia in the UK, a figure that is expected to have doubled by 2040. Current policy targets include ensuring early diagnosis and good quality early intervention for all and supporting people with dementia in living as full and active a life as possible.
Early diagnosis of dementia creates an opportunity to equip patients and carers to manage the disease effectively and to live well with dementia. In this context early intervention offers the possibility of reducing or delaying the progression of functional disability, depression or behavioural difficulties, helping to maintain independence, supporting management of co-morbidity and hence avoiding or reducing hospitalisation, maintaining quality of life and ultimately delaying institutionalisation. At present, however, the chances of accessing early psychosocial intervention following a diagnosis of dementia are low, and research evidence regarding the efficacy of early psychosocial interventions remains limited. Research priorities set out by the Ministerial Advisory Group on Dementia Research (MAGDR) in 2011 indicate a need to evaluate the effect of psychosocial interventions for people with dementia living in the community, including those based on 're-ablement' , and to identify ways of improving quality of life for people with dementia and their carers.
Relatively little attention has been given to developing psychosocial early intervention approaches aimed at helping people to live well with dementia. Traditionally, efforts have focussed instead on attempting to address the underlying impairments in memory and other cognitive functions, which are a defining feature of early-stage dementia. A number of research studies have examined the potential of cognitive training to benefit people with dementia. Cognitive training involves repeated, structured practice on tasks targeting specific cognitive domains, such as working memory or attention. Evidence is mixed, with some studies reporting modest benefits and others reporting no benefits, but a Cochrane systematic review [bib_ref] Cognitive rehabilitation and cognitive training for early-stage Alzheimer's disease and vascular dementia, Clare [/bib_ref] found no evidence for significant benefits. Even where improvements on cognitive tasks assessing trained domains are reported, there is no evidence that these generalise to other areas, have any impact in the real-life context or offer any benefits as regards engagement in everyday activities [bib_ref] Cognitive training in Alzheimer's disease: a meta-analysis of the literature, Sitzer [/bib_ref]. That is to say, these approaches, which target underlying impairment, albeit with limited success, fail to reduce functional disability. Yet there is evidence for preservation of some degree of cognitive and neural plasticity in early-stage dementia [bib_ref] Cognitive plasticity in healthy, mild cognitive impairment (MCI) subjects and Alzheimer's Disease..., Fernández-Ballesteros [/bib_ref] , and it should be possible to harness this potential to deliver beneficial intervention effects. According to neuropsychological models of memory [bib_ref] Memory, hippocampus, and brain systems, Squire [/bib_ref] , while some cognitive functions (e.g. long-term episodic recall) are significantly impaired in early-stage Alzheimer's disease, others are relatively spared (e.g. procedural memory for skills, routines and actions, semantic knowledge and implicit memory) [bib_ref] Assessment and intervention in dementia of Alzheimer type, Clare [/bib_ref] , and people with early-stage dementia are capable of some new verbal and behavioural learning [bib_ref] The retention of new information in senile dementia, Little [/bib_ref] , although they are likely to require extra support to achieve it [bib_ref] Memory training and memory improvement in Alzheimer's disease: rules and exceptions, Backman [/bib_ref]. Consequently, there are possibilities for behaviour change to occur.
Conceptualising dementia within the framework of a disability modelhighlights the distinction between the underlying impairment, resulting from pathological changes, and the resulting limitations on engaging in activity (disability) and restrictions on social participation (handicap). Activity limitation and participation restriction are not solely determined by the degree of impairment, but are subject to a range of personal, social and environmental influences. Negative influences can contribute to the development and maintenance of 'excess' disability, where the extent of functional disablement is greater than would be predicted by the degree of impairment; an example would be where an individual loses confidence, gives up previously enjoyed activities, and becomes socially withdrawn and depressed in reaction to receiving the diagnosis of dementia, with consequent effects on cognitive and functional ability. This is similar to Kitwood's description of the way in which a negative, unsupportive social environment can undermine well-being for people with dementia. Equally, positive influences can support optimal functioning and overcome some of the potential impact of impairment, enabling people to live well with dementia. A focus on addressing barriers to activity and participation, and encouraging adaptive behaviours, can therefore be expected to produce benefits for people with dementia and their family members.
Interventions that aim to reduce functional disability by targeting activity and participation, drawing on retained strengths to support adaptive behaviour, are typically described as forms of rehabilitation. Rehabilitation interventions aim to 'enable people who are disabled by injury or disease to achieve their optimum physical, psychological social well-being' [bib_ref] Functional recovery and the principles of disability medicine, Mclellan [/bib_ref]. The rehabilitation of people who have cognitive, as opposed to purely physical, impairments is termed 'cognitive rehabilitation' (CR). Although rehabilitation is most often associated with nonprogressive conditions such as brain injury, it is equally applicable to people with chronic and progressive conditions. There is considerable evidence for the efficacy of cognitive rehabilitation with a range of clinical groups. Rehabilitation interventions are generally highly individualised, as clients have a diverse range of impairments, needs, circumstances and preferences. Central to the practice of rehabilitation is the identification of realistic and personally meaningful individual rehabilitation goals for each client and the development of tailored interventions to address these. Goal-based approaches have been applied in numerous conditions, including brain injury [bib_ref] Use of Goal Attainment Scaling in measuring clinically important change in cognitive..., Rockwood [/bib_ref] [bib_ref] Occupational therapy and achievement of self-identified goals by adults with acquired brain..., Trombly [/bib_ref] , stroke [bib_ref] Reliability and validity of the Canadian Occupational Performance Measure in stroke patients, Cup [/bib_ref] , neurological illness [bib_ref] Use of Goal Attainment Scaling in inpatient rehabilitation for persons with multiple..., Khan [/bib_ref] , physical disability [bib_ref] Goal Attainment Scaling in the rehabilitation of patients with lower-extremity amputations: a..., Rushton [/bib_ref] and chronic pain [bib_ref] Awareness of activity limitations and prediction of performance impairments in patients with..., Fischer [/bib_ref] [bib_ref] Assessing clinically meaningful change following a programme for managing chronic pain, Fisher [/bib_ref] , as well as for frail older people [bib_ref] Responsiveness of goal attainment scaling in a randomized controlled trial of comprehensive..., Rockwood [/bib_ref]. Goals are, wherever possible, negotiated collaboratively between client and therapist. Such interventions may be regarded as inherently person-centred.
It has been suggested that rehabilitation provides a useful overarching conceptual framework for the care and support of people with dementia and for the design of interventions to meet their needs. A few early examples of interventions that addressed meaningful individual goals relating to self-care or activity participation supported the possible utility of this approach [bib_ref] Use of a digital alarm chronograph as a memory aid in early..., Kurlychek [/bib_ref] [bib_ref] Supporting everyday activities in dementia: an intervention study, Josephsson [/bib_ref]. Hence feasibility studies were undertaken to explore the application of cognitive rehabilitation to help people with early-stage dementia and their families manage the impact of the condition.
A series of studies using single-case experimental designs or small-group pre/post comparisons demonstrated that it was possible to identify meaningful personal goals and use evidence-based restorative or compensatory rehabilitation methods [bib_ref] Fundamentals of cognitive rehabilitation, Mateer [/bib_ref] to bring about behaviour change in these areas for people with early-stage dementia [bib_ref] Errorless learning of face-name associations in early Alzheimer's disease, Clare [/bib_ref] [bib_ref] Intervening with everyday memory problems in early Alzheimer's disease: an errorless learning..., Clare [/bib_ref] [bib_ref] Cognitive rehabilitation as a component of early intervention in dementia: a single..., Clare [/bib_ref]. Restorative approaches build on retained abilities and use a range of instructional or prompting techniques to promote new learning or relearning, whether of information, habits or strategies; examples include the application of the spaced retrieval method to support retention of information [bib_ref] Facilitation of new learning in Alzheimer's disease, Camp [/bib_ref]. Compensatory methods use a range of aids and adaptations to support functioning and overcome limitations resulting from cognitive impairments; examples include the use of memory books to support engagement in conversation [bib_ref] Enhancing conversation skills in patients with Alzheimer's disease using a prosthetic memory..., Bourgeois [/bib_ref]. Rehabilitation interventions for people with dementia need to offer practical benefits in daily life. In the context of cognitive impairment it is particularly challenging to ensure that learning and behavioural change generalises from one setting to another; to circumvent this obstacle, the interventions in our smallscale studies were carried out in the person's everyday setting rather than in the clinic. The behavioural changes observed, although focused on specific targeted goals, led to wider benefits in everyday life; for example, learning names of other participants in a social club helped to maintain attendance and participation and reduced the risk of social isolation [bib_ref] Errorless learning of face-name associations in early Alzheimer's disease, Clare [/bib_ref] , and using a memory aid to reduce repetitive questioning reduced carer frustration and tensions between the participant and carer [bib_ref] Intervening with everyday memory problems in early Alzheimer's disease: an errorless learning..., Clare [/bib_ref]. There was also some evidence for generalisation of the problemsolving approach to other everyday situations and challenges [bib_ref] Intervening with everyday memory problems in early Alzheimer's disease: an errorless learning..., Clare [/bib_ref] [bib_ref] Cognitive rehabilitation as a component of early intervention in dementia: a single..., Clare [/bib_ref]. Gains were maintained for several months post-intervention, and one longer-term study demonstrated maintenance of gains up to 3 years postintervention [bib_ref] Long-term maintenance of treatment gains following a cognitive rehabilitation intervention in early..., Clare [/bib_ref]. Further studies investigated the efficacy and applicability of specific memory rehabilitation techniques, such as errorless learning and spaced retrieval methods [bib_ref] Relearning of face-name associations in early-stage Alzheimer's disease, Clare [/bib_ref] [bib_ref] Memory rehabilitation for people with early-stage dementia: a single case comparison of..., Clare [/bib_ref] [bib_ref] Learning face-name associations in early-stage dementia: Comparing the effects of errorless learning..., Dunn [/bib_ref]. These findings were augmented by reports from other research groups [bib_ref] Behavioural difficulties and cued recall of adaptive behaviour in dementia: experimental and..., Bird [/bib_ref] [bib_ref] Training early Alzheimer patients to use a mobile phone, Lekeu [/bib_ref]. A Cochrane systematic review found no randomised controlled trials of cognitive rehabilitation [bib_ref] Cognitive rehabilitation and cognitive training for early-stage Alzheimer's disease and vascular dementia, Clare [/bib_ref]. The findings from the feasibility studies, therefore, formed the basis for developing an intervention protocol that could be tested in a pilot randomised controlled trial.
The design of trials to evaluate the efficacy of rehabilitation interventions must take into account the fact that rehabilitation typically focuses on the attainment of highly individual goals that are functionally, socially and contextually relevant [bib_ref] Goal attainment scaling in rehabilitation, Malec [/bib_ref]. When evaluating service or programme outcomes in rehabilitation settings, goal attainment scaling has been used to identify goals and rate progress on a standardised scale [bib_ref] Use of Goal Attainment Scaling in measuring clinically important change in cognitive..., Rockwood [/bib_ref] [bib_ref] Responsiveness of goal attainment scaling in a randomized controlled trial of comprehensive..., Rockwood [/bib_ref] [bib_ref] Goal attainment scaling in rehabilitation, Malec [/bib_ref] [bib_ref] Goal Attainment Scaling: a general method for evaluating comprehensive community mental health..., Kiresuk [/bib_ref]. However, where the focus is on treatment outcomes for the individual client, as opposed to overall efficacy of a multidisciplinary or multi-component programme, goal setting and goal achievement are more readily evaluated by means of client-centred approaches in which the client plays a central role in a collaborative goal-setting process, and the client's perceptions of change serve as the primary outcome measure. The most widely used example of this approach is provided by the Canadian Occupational Performance Measure (COPM), which offers a structured format for eliciting individual goals and a standardised means of rating goal performance and satisfaction with performance. There is evidence for the reliability, construct validity, sensitivity and responsiveness of this measure as well as for its clinical utility [bib_ref] Reliability and validity of the Canadian Occupational Performance Measure in stroke patients, Cup [/bib_ref] [bib_ref] The Canadian Occupational Performance Measure: a research and clinical literature review, Carswell [/bib_ref] [bib_ref] A comparison of goal attainment scaling and the Canadian Occupational Performance Measure..., Cusick [/bib_ref] [bib_ref] Individualized outcome measures: a review of the literature, Donnelly [/bib_ref] [bib_ref] The reproducibility of the Canadian Occupational Performance Measure, Eyssen [/bib_ref] [bib_ref] Utility of the Canadian Occupational Performance Measure in community-based brain injury rehabilitation, Jenkinson [/bib_ref] [bib_ref] Clinical utility of the Canadian Occupational Performance Measure -Swedish version, Wressle [/bib_ref]. When using this measure in research, it is possible to elicit goals and performance ratings at baseline and to have participants in both treatment and control groups re-rate goal performance at follow-up. Where clients have cognitive impairments, it is helpful to supplement self-ratings with independent ratings made by professionals or caregivers for comparison purposes [bib_ref] Reliability and validity of the Canadian Occupational Performance Measure in stroke patients, Cup [/bib_ref] [bib_ref] Utility of the Canadian Occupational Performance Measure in community-based brain injury rehabilitation, Jenkinson [/bib_ref]. The goal-oriented approach accords with person-centred values in dementia care, allowing the person with dementia to engage in an intervention that is specifically tailored to his/her own needs and preferences, while also providing for a standardised group-level comparison. Therefore, more recently, the Bangor Goal-Setting Interview (BGSI) has been developed for research purposes [bib_ref] The AgeWell study of behaviour change to promote health and wellbeing in..., Clare [/bib_ref]. This interview has similar aims to those of the COPM, but has been developed primarily as a research tool and incorporates a number of different features. The BGSI is based on the social cognitive theory of behaviour change [bib_ref] Health promotion by social cognitive means, Bandura [/bib_ref] and on the concept of motivational interviewing [bib_ref] Mash B: Motivational interviewing, Rollnick [/bib_ref] , and the relevant domains of functioning in relation to which goals will be considered are selected according to the requirements of the given study.
The CR intervention is focussed on the identification and attainment of individual goals. The recently developed BGSI will be used in the GREAT trial. However, for the pilot trial, perceived goal performance, rated using the COPM, was selected as the primary outcome. A pilot trial of individual, goal-oriented CR, funded by the Alzheimer's Society and published in the American Journal of Geriatric Psychiatry, was conducted in North Wales from 2005 -2009 [bib_ref] Goal-oriented cognitive rehabilitation for people with early-stage Alzheimer disease: a single-blind randomized..., Clare [/bib_ref]. This was a single-site, single-blind randomised controlled trial comparing an eight-session CR intervention to relaxation therapy (RT), which involved equivalent therapist time and attention, and was expected to be pleasurable for participants without addressing the areas targeted in CR, and to treatment as usual (TAU). All participants received acetylcholinesterase-inhibiting (AChEI) medication and routine outpatient monitoring, and had access to the range of voluntary sector services available at the time in the area. The primary outcome was goal performance and satisfaction with performance. Sixty-nine participants were randomised by computer algorithm, independently operated by the North Wales Organisation for Randomised Trials in Health, Clinical Trials Unit (NWORTH CTU), to one of the three conditions (CR, n = 23; RT, n = 24; TAU, n = 22).
Following intervention, goal performance and satisfaction ratings improved for the CR group and showed no change in the other two groups (see [fig_ref] Figure 1: Effects of intervention on goal performance and satisfaction [/fig_ref]. Analysis of covariance indicated a significant effect of CR on performance (F(2,58) = 7.880, P <0.001) and satisfaction (F(2,58) = 8.270, P < 0.001). For both measures, CR differed significantly to both RT and TAU (performance: 1.459 ± 0.936 for RT and 1.128 ± 0.989 for TAU; satisfaction: 1.686 ± 1.041 for RT and 1.193 ± 1.090 for TAU). For the CR group, achievement of therapy goals was corroborated in three ways through within-group analyses [bib_ref] Goal-setting in cognitive rehabilitation for people with early-stage Alzheimer's disease, Clare [/bib_ref]. First, participants rated performance and satisfaction with performance for each goal targeted, recording significant increases (performance: t(25) = −3.742, P < 0.001; satisfaction: t(25) = −4.877, P < 0.001). Second, a separate therapist rating of goal performance was made at the start and end of therapy; this reflected significant improvements (t(25) = −8.027, P < 0.001). Third, a simplified goal attainment scaling procedure was used, whereby for each therapy goal behavioural indicators of full and partial attainment were established by the research team at the start of therapy and each goal was rated accordingly at the end of therapy. This classified 12 (46%) goals as fully implemented, 13 (50%) as partially implemented and 1 (4%) as not implemented. It was noted that many of the partially implemented goals would likely have been fully achieved given a little more time.
Secondary outcomes were evaluated in terms of effect sizes (Cohen's d) for the CR group compared to the pooled control (RT and TAU) groups, as no differences were observed between the two control groups on any measure. Outcomes examined for the person with dementia were quality of life, mood and cognition (effect sizes are shown in [fig_ref] Table 1: Effect sizes [/fig_ref]. CR produced benefits in all three areas, and quality of life continued to improve at 6-month follow-up. It should be noted that for the most part mood was within the normal range at baseline and hence scope for improvement was limited. For carers, CR reduced stress and improved psychological wellbeing and quality of life (effect sizes are shown in [fig_ref] Table 1: Effect sizes [/fig_ref] , and in some cases these were maintained or continued to improve at follow-up.
The CR intervention was acceptable to, and well received by, participants and carers. Across all three groups, the attrition rate between randomisation and post-intervention assessment was 7%; five individuals discontinued because of physical illness (1), death (1), incorrect diagnosis (1) and self-withdrawal. Attrition between post-intervention assessment and 6-month follow-up was 12%; eight individuals were lost to follow up because of death (2), moving out of the area (3) and self-withdrawal (3). Thus, the overall rate of elective self-withdrawal for the trial was only 7% (2 each from CR and RT, and 1 from TAU).
This pilot trial demonstrated that participants with early-stage dementia can identify personally meaningful goals relating to managing everyday activities, and, with a modest amount of support from a therapist, make significant progress towards implementing these. Goal performance constituted a sensitive and specific measure of change. As performance and satisfaction ratings were closely associated, performance ratings should suffice in future work. The addition of carer ratings of performance would be informative. The trial provided valuable experience in collaborative identification of specific, measurable, achievable and realistic goals [bib_ref] Goal-setting in cognitive rehabilitation for people with early-stage Alzheimer's disease, Clare [/bib_ref]. Results suggested that a slightly longer intervention might be advisable in order to fully establish and consolidate gains. The trial showed that CR can bring benefits with regard to cognition, well-being and quality of life for the person with dementia, as well as the well-being and quality of life of the carer. The lack of observed differences between the two control groups (RT and TAU) suggested that in a definitive trial TAU could be adopted as an appropriate comparison condition for CR. Findings from the pilot provided information about intervention parameters, outcomes and effect sizes that has informed the design of the planned multi-site trial presented below.
# Methods/design
This is a multi-centre single-blind randomised controlled trial comparing cognitive rehabilitation (CR) to treatment as usual (TAU) for people with early-stage Alzheimer's, vascular or mixed dementia, with outcomes assessed at 3 and 9 months post randomisation. Participants will be recruited from memory clinics, old age mental health services and general practitioner (GP) practices. CR will be delivered in participants' homes, with a carer involved where possible. The study will be run from the co-ordinating centre at Bangor University with six recruiting centres around the UK: North-West England, South-West England, West Midlands, London, South Wales and North Wales. At each recruiting centre, a part-time therapist (with an appropriate professional background, e.g. occupational therapy or psychology) will conduct the interventions, and a research assistant, blind to group allocation, will carry out assessments at baseline and at 3 and 9 months post-randomisation. Participants will be recruited to the trial between 1 April 2013 and 30 September 2015. A CONSORT-style flow chart for the trial is shown in [fig_ref] Figure 2: GREAT trial CONSORT-style flow chart [/fig_ref].
## The cognitive rehabilitation intervention
CR is an individualised approach for people with dementia (PwD) aimed at managing or reducing functional disability and maximising engagement and social participation. PwD and their carers work together with a health professional over a number of sessions to identify personally relevant goals and devise and implement strategies for achieving these. CR will be delivered in ten individual sessions over 3 months, followed by four maintenance sessions over 6 months. Carers will be involved in part of each session where possible. Involvement of a carer helps to ensure that skills are maintained and applied to novel situations and facilitates communication about how current or possible future difficulties might be managed. Over the course of the 10 weekly sessions, participants with dementia will work in collaboration with the therapist to address personal rehabilitation goals. Drawing on the goals identified at baseline assessment, up to three behavioural goals will be operationalised, and strategies for addressing these will be devised and implemented. Goals will be introduced one at a time, in a flexible manner depending on rate of progress. Following introduction and modelling of strategies and skills during the therapy sessions, the participant and carer will work on the selected goal between sessions following an agreed schedule of activities. Progress with each goal will be reviewed and the strategies adopted will be adjusted as necessary on a weekly basis. Performance for each goal will be independently rated at the outset and in week 10 by the participant, carer and therapist, and the therapist will rate the extent of goal attainment following the sessions in week 10. Work on goals will be supplemented by the following components, designed to support the ability to make progress with the selected goals and address barriers to progress, which will be systematically introduced across the ten sessions:
1. Introduction of, and practice in applying, a solutionfocused problem-solving approach by following a short sequence of steps to specify and test possible solutions. 2. Introduction of anxiety management strategies, building on participants' existing strengths and preferences in this area, and practice in strategy use and application. 3. Monitoring of activity levels, leading to plans for increasing engagement in meaningful and enjoyable activity, and implementation of these plans. 4. Practice in strategies for maintaining or improving attention and concentration. 5. A review of compensatory strategy use (e.g.
calendars, diaries, reminder systems) and development and implementation of plans for improving strategy use, which might include increasing the efficiency of existing strategies and introduction of new strategies. 6. A review of current use of restorative strategies for retaining new information or improving recall, and practice in key strategies (mnemonics, semantic association, spaced retrieval), enabling participants to identify a preferred strategy and apply this in everyday situations.
In addition, to ensure sensitivity to the wider context, where appropriate a discussion of carer well-being and strategies for managing stress will be initiated, and information will be provided about additional sources of support and help, with encouragement to access these. The four maintenance sessions will focus on supporting maintenance of gains and encouraging continued goal performance and strategy use. The therapist will re-rate the extent of goal attainment following the session in week 14.
The effects of the CR intervention will be compared to treatment as usual (TAU). In the pilot, CR was compared with TAU and with an attention placebo condition (relaxation therapy). There was no evidence of a difference between the two control groups, which suggests that TAU can serve as an appropriate comparator. For the CR group, the CR will be provided in addition to TAU, while the control group will receive only TAU and will have no contact with the research team between assessments. TAU will consist of acetylcholinesterase-inhibiting medication where prescribed, and any other services normally provided apart from specific programmes of cognitive rehabilitation or other cognition-focused intervention. TAU may include, for example, routine monitoring by the Memory Clinic, information provision, attendance at drop-in groups or support groups, or carer participation in support groups. Service receipt during the intervention period, including dementia-specific services, monitoring and interventions provided by Memory Clinics, will be documented for all participants. All participants will be free to access services such as those offered by the Alzheimer's Society, and the extent of this will be recorded.
## Participant selection
Participants will be recruited from memory clinics, old age mental health services and GP practices, and will have been diagnosed with early-stage Alzheimer's disease (AD), vascular dementia, or mixed AD and vascular dementia. For each participant, a carer (a family member or close friend who is either co-resident or in regular contact) will also be involved.
Inclusion criteria:
1. The participant must have been assigned an ICD-10 diagnosis of Alzheimer's disease (AD), vascular dementia, or mixed AD and vascular dementia. AD accounts for 62% of dementia diagnoses, and vascular dementia for 17%, with mixed AD and vascular dementia accounting for a further 10%. These categories together capture 89% of those diagnosed with dementia. There is no reason to assume that people with rarer subtypes of dementia, including dementia with Lewy bodies (4%), frontotemporal dementia (2%) and Parkinson's dementia (2%), could not benefit from CR, but these forms of dementia have specific features that would require a distinct approach. For this reason, and because numbers are likely to be too small to allow for subgroup analysis, we are not proposing to include these groups in the current trial. 2. The participant must be in the early stages of dementia, as indicated by an MMSE score of 18 or above. This is to ensure that participants recruited to the trial have a level of cognitive functioning that is sufficient to allow them to complete the selected outcome measures without undue difficulty for the duration of their participation in the study. Use of a cutoff point, while inevitably somewhat arbitrary, provides protection for people who may be overly burdened by the assessments. The selected cutoff of a score of 18 on the MMSE is frequently used in research studies and worked well in the pilot trial.
We have not placed any upper limit on the MMSE score, since it can be expected that a small proportion of people meeting diagnostic criteria for dementia will have high MMSE scores [bib_ref] Can you have dementia with an MMSE score of 30?, Shiroky [/bib_ref]. 3. If taking acetylcholinesterase inhibitors, the participant must have been receiving a stable dose for 1 month prior to trial entry, and there should be no intention to change the dose over the period of participation in the study unless clinically indicated. This is to ensure that intervention effects are not confounded by changes in medication status. 4. The participant must have a carer who is willing to participate. While having a carer is not an essential prerequisite for receiving a CR intervention, it is important for the purposes of research to obtain an informant perspective on the effects of the intervention, and in this trial carers will be asked to provide an independent rating of goal performance. It is also important to determine the effects of the intervention on carer well-being; positive effects on the carer are likely to bring added benefits in the longer term for the person with dementia. 5. The participant must be able to give informed consent. People in the early stages of dementia are normally expected to have capacity to consent to participation. When recruiting participants, the research team will use a checklist to ensure that all relevant information is considered and that the participant is able to give informed consent. While CR principles may be applied at any stage of dementia, the intervention to be tested here involves engaging the person with dementia in a collaborative process of identifying and addressing personally meaningful goals, and therefore the participant needs to be able to understand this and to make a positive choice to take part.
Exclusion criteria:
1. Participants will be excluded if they have a prior history of stroke, brain injury or other significant neurological condition. Such conditions would be expected to affect cognitive, behavioural and emotional functioning, and people who have one of these conditions prior to developing dementia would have additional rehabilitation needs. While such individuals might benefit from CR, their inclusion would represent a potential confounding factor. 2. Participants will be excluded if they are unable to speak English. This criterion is applied for practical reasons, because the standardised outcome measures we plan to use are only available in English. No official data are available to indicate what proportion of the UK population cannot speak English; while it is estimated that about 3% of the population use a language other than English at home, with over 100 different languages represented (source: The National Centre for Languages), many of the individuals concerned also speak English. The time and costs involved in translating standardised measures and providing interpreters for assessment and therapy sessions would be substantial. However, we predict that very few individuals would be excluded from participation because of inability to communicate in English.
# Ethical considerations
The study has been reviewed by the North Wales Research Ethics Committee-West, which issued a favourable opinion on 25 June 2012 (Reference 12/WA/ 0185) and has been approved by the Bangor University School of Psychology Research Ethics Committee. Based on previous findings, participants randomised to receive CR may be expected to derive some benefits in terms of managing everyday activities and general wellbeing. Their caregivers may also be expected to show reduced stress and improved well-being. Availability of evidence from a definitive trial may be expected to have a positive influence on the future provision of interventions to support people with early-stage dementia and their carers. Previous findings also suggest that participants randomised to TAU are expected to show little or no change; thus, they will not be harmed by this allocation. As the trial will provide the first evidence from a large-scale trial regarding the benefits of CR, it cannot at this stage be considered unethical to withhold this treatment from the control group, and the control group will still have access to the care typically provided by memory clinics and GPs, and to voluntary sector services.
There are no known risks or side effects associated with CR. It is possible that some participants may find it challenging to confront their difficulties, but the therapist will provide support as they engage in this process, and the intervention protocol incorporates attention to managing emotional reactions. Neither the feasibility studies nor the pilot trial have suggested that this represents a significant risk to participants. The research team will be trained to be alert to any concerns about participants' well-being. If there are serious concerns about a participant, these will be referred, wherever possible with the permission of the individual concerned and the carer, to the clinician responsible for the participant's care.
Participants with early-stage dementia, and carers, will be fully informed prior to entry into the trial about the intervention and about the current state of knowledge regarding possible benefits and risks, and this information will be updated if additional evidence becomes available during the course of their participation.
Informed consent will be obtained from all participants and carers. People with early-stage dementia are expected to have the capacity to consent to participation. While consent provided at the outset provides an initial mandate for entry to the trial and commencement of trial procedures, consent is an ongoing process, and this is crucial for psychosocial interventions where participants' active engagement is required. Therefore, research assistants and therapists will be trained to monitor ongoing consent and to respond appropriately to any indication of a possible withdrawal of consent. As participants will be in the early stages of dementia, loss of capacity to consent during the course of participation is expected to be infrequent. However, on entry to the trial participants will be asked whether, should they lose capacity to consent, they are willing to continue to be included in the trial and to have their data used.
Personally identifiable information will be retained only until publication of the trial report unless the participant has consented to retention of details for potential further follow-up, while anonymised data will be retained for 5 years after publication unless a longer period is required by the Research Ethics Committee or other regulatory authorities. Consent forms will be retained for 25 years following trial closure.
The governance and management of the study will be undertaken within the Department of Health Research Governance Framework for Health and Social Care (2nd edition, 2005). This will ensure the highest standards of clinical research, covering scientific quality, ethical standards and all related management issues. The trial will adhere to the Standard Operating Procedures (SOPs) of NWORTH CTUfor all trial and data management, statistical and regulatory matters. This is not a clinical trial of an investigational medical product (CTIMP) and therefore it does not come under the provisions of the Medicines for Human Use (Clinical Trials) Regulations (2004).
All research staff and therapists will undergo training in Good Clinical Practice with regard to the conduct of clinical trials. Trial-specific training requirements will be addressed throughout the study period and regularly reviewed. Orientation and project-specific training will be provided for CLRN, MHRN and NISCHR CRC staff.
## Sample size
Power calculations and attrition rates are based on findings from the pilot trial. We wish to confirm the finding that the primary effectiveness outcome of goal performance was improved in the treatment group. The difference observed in the pilot was large (standardised effect >1 at post-intervention assessment). However, we also wish to detect any effect sizes in the order of 0.3 for important secondary outcomes, as we judge that this will give confirmation of effects that are large enough to have substantive clinical benefits. For the present study, intervention length has been increased and now includes a maintenance phase in order to further strengthen demonstrable effect sizes in secondary outcomes. We have elected to be conservative in all aspects of our estimate of power, and we have made a larger estimation of potential attrition than the <20% observed in that study, based on the 27% rate observed in the recent REMCARE trial [bib_ref] Reminiscence groups for people with dementia and their family carers: pragmatic eight-centre..., Hounsome [/bib_ref]. To achieve 80% power to detect a medium effect size of 0.3, with alpha 0.05, in primary and secondary outcomes, 175 PwD, with their carers, need to complete the trial in each arm. Adjusting for potential attrition, we aim to randomise 480 PwD, each with a carer. There will be one practitioner involved at each centre, and we will test for differences between sites in the analysis. If there is a change of practitioner at one or more centres during the course of the trial we will also test for practitioner effects.
## Outcome measures (1) primary outcome measure
Bangor Goal-Setting Interview (BGSI) The BGSI [bib_ref] The AgeWell study of behaviour change to promote health and wellbeing in..., Clare [/bib_ref] provides a structured format for the goal-setting process. The interview proceeds in three stages. First, relevant domains of functioning are discussed in turn with a view to eliciting issues that might form the basis for behavioural goal-setting. For each area, the participant rates perceived importance of making changes in this area, and readiness to make changes in this area, on a 1 -10 scale (where 1 is not at all important/not at all ready and 10 is extremely important/completely ready). Once all areas have been discussed, the second stage involves revisiting each area in turn and negotiating specific behavioural goals that conform to 'SMART' principles (specific, measurable, achievable, realistic and timedelineated). Additionally, goal attainment indicators are specified, providing clear descriptors of what would constitute 25%, 50% and 75% goal attainment. In the final stage of the interview, the participant is asked to rate, for each of the goals that have been identified, current performance and satisfaction with performance on a 1 -10 scale (where 1 is unable to perform/extremely dissatisfied and 10 is able to perform perfectly/extremely satisfied). Mean scores for performance and satisfaction with performance across goals are calculated by dividing in each case the sum of the scores for all goals identified by the number of goals set. For the present study, details of the goals identified in the baseline BGSI assessment will be provided to the therapist for each participant allocated to the CR condition and will form the starting point for the therapist's intervention. At follow-up assessment, the participant re-rates current performance and satisfaction with performance for each goal so that changes in ratings can be examined. The interviewer elicits information about current performance and uses the previously specified goal attainment indicators to rate the extent of progress towards achieving the goal. There is scope within the interview format for informant ratings to be obtained; in this study, carer ratings of performance and descriptions of goal attainment will be elicited at baseline and follow-up for comparison purposes.
(2) Secondary outcome measures for participants with dementia DEMQOL DEMQOL [bib_ref] Development of a new measure of health-related quality of life for people..., Smith [/bib_ref] has been designed to assess health-related quality of life of people with dementia across the full range of severity and subtypes, and shows high internal consistency (Cronbach's alpha 0.87) and good test-retest reliability (ICC 0.76) in people with mild to moderate dementia. It consists of a 28-item interviewer-administered questionnaire for the person with dementia and a 31-item interviewer-administered questionnaire on which the caregiver provides proxy ratings. These may be used together or separately. In this study, only self-ratings by the person with dementia will be taken. An algorithm has been developed to generate quality-adjusted life year (QALY) scores from DEMQOL scores for use in economic evaluation [bib_ref] Development of DEMQOL-U and DEMQOL-Proxy-U: Generation of preference-based indices from DEMQOL and..., Mulhern [/bib_ref].
Generalised Self-Efficacy Scale (GSES) The ten-item Generalised Self-Efficacy Scale (GSES)was created to assess a general sense of perceived self-efficacy, the potential to influence one's situation through one's own actions. Responses are made on a 4-point scale. Responses to all ten items are summed to yield the final composite score with a range from 10 to 40. Cronbach's alphas range from 0.76 to 0.90 [bib_ref] General self-efficacy in various domains of human functioning: evidence from five countries, Luszczynska [/bib_ref].
## Hospital anxiety and depression scale (hads)
The HADScontains 14 items forming two subscales: anxiety and depression. Each item is rated on a 4-point scale, giving maximum scores of 21 for anxiety and for depression. Scores of 11 or more on either subscale are considered to be a significant 'case' of psychological morbidity, with scores of 8-10 classified as 'borderline' and 0-7 'normal'. The HADS has been employed and validated in studies of people with dementia and carers [bib_ref] Awareness of carer distress in people with dementia, Ablitt [/bib_ref] [bib_ref] Coping strategies and anxiety in caregivers of people with Alzheimer's disease: The..., Cooper [/bib_ref].
Brief cognitive assessment battery This will consist of brief tests of memory, attention and executive function, suitable for people with early-stage dementia, each taking less than 5 min to administer: (1) Memory: Rivermead Behavioural Memory Test (RBMT), story recall subtest. The RBMT is a well-established, ecologically valid test of everyday memory. In the story recall task, the researcher reads out a short story, similar to a brief report of a newsworthy event in a daily newspaper, and the participant is asked for immediate and delayed (after 20 min) recall of the content. Recall is scored following a standard protocol (inter-rater reliability > 0.9) with a maximum possible score of 21 for the immediate and for the delayed component. Four equivalent versions are available to permit reassessment without the risk of practice effects; practice effects are not anticipated with test-retest intervals of 3 and 6 months, but as a precaution a different version will be used at each time point. Raw scores will be used in the analysis as they provide a greater range than the condensed standardised profile score that is used in calculation of the overall RBMT score. (2) Attention: Test of Everyday Attention (TEA), elevator counting and elevator counting with distraction subtests. The TEA is a well-established, ecologically valid test of everyday attention, with subtests assessing different components of attention. The elevator counting subtest assesses sustained attention. Participants are required to count a short string of monotonous tones and give the total number. Seven strings are presented, and the total score is the number of strings correctly counted. The elevator counting with distraction subtest assesses auditory selective attention. Further strings of tones are presented, this time also including distractor (high-pitched) tones that are not to be counted. The total score is the number of strings correctly counted. Three equivalent versions of each subtest are available to permit reassessment without the risk of practice effects; as above, practice effects are not anticipated but as a precaution a different version will be used at each time-point. (3) Executive function: Letter fluency subtest of the Delis-Kaplan Executive Function System (D-KEFS). D-KEFS consists of a set of standardised tests of executive function. The verbal letter fluency task evaluates the executive subdomains of initiation, response generation and inhibition [bib_ref] The elusive nature of executive functions: a review of our current understanding, Jurado [/bib_ref] and draws on semantic memory and language ability. In this task, the participant is asked to list as many words as possible beginning with a specific letter of the alphabet in a 1-min period, excluding proper nouns and repetitions. Three letters, F, A and S, are used. The total number of correct responses to the three letters is used in analysis. This task has been extensively examined in people with early-stage dementia [bib_ref] Verbal fluency performance in dementia of the Alzheimer's type: a meta-analysis, Henry [/bib_ref]. Evidence suggests that even in healthy participants there are no practice effects for most components of this task even at testretest intervals of less than 2 weeks; there are minimal practice effects for the switching component with testretest intervals of less than 2 weeks, but not with longer intervals.
(3) Secondary outcome measures for the carer Relatives' Stress Scale (RSS) The RSS [bib_ref] Measuring behavioural disturbance of elderly demented patients in the community and its..., Greene [/bib_ref] is a 15-item dementia-specific measure of caregiver stress with items rated on a 5-point Likert scale and summed. A higher overall score indicates higher levels of caregiving-specific stress.
## Euroqol (eq5d)
The EQ-5Dis a standardised measure of health status and health outcome, applicable to a wide range of health conditions. In the first section, the respondent is asked to select one of three options for each of five dimensions: mobility, self-care, usual activities, pain/discomfort and anxiety/depression. For each dimension, the three response options are coded on a 3-point scale from 1 (no problem) to 3 (unable to perform/extreme problem). This yields a descriptive profile (e.g. 11232) across the five dimensions. The second part of the measure is a visual analogue scale for selfrating of health-related quality of life ('your health state today'). This measure is included because the EQ5D score will be used to generate QALY scores using societal weights.
## Who quality of life -bref(whoqol-bref)
The WHOQOL-BREF [bib_ref] The World Health Organization's WHOQOL-BREF quality of life assessment: psychometric properties and..., Skevington [/bib_ref] is a 26-item scale assessing perceived quality of life, giving scores in four domains: environment, social relationships, psychological and physical health.
## (4) service utilisation
Client Services Receipt Inventory (CSRI) The CSRIprovides a template that can be adapted to the needs of each specific study. Respondents are asked about their use of health care services for a period preceding baseline assessment and during the study period. The questions cover contact with a range of health and social care professionals, prescription of medications, hospital appointments and stays, participation in local authority funded activities such as day centres, participation in activities run by voluntary organisations and the contribution of informal carers. Questions to examine the nature and extent of any dementia-specific treatment received from the Memory Clinic will be included.
(5) Demographic and background information for the person with dementia and carer Details such as gender, age, relationship between person with dementia and carer and whether they live together, age of onset of dementia, educational level, social class and co-morbidities will be collected. This will allow us to examine effects of demographic and social variables on treatment efficacy.
## (6) process measures for the cr group
For the cognitive rehabilitation group, process measures will be taken to provide convergent evidence about change in goal performance. In-session parallel ratings of goal performance by participant, carer and therapist will be made when each goal is introduced and in session 10. A simplified goal attainment scaling procedure [bib_ref] Goal attainment scaling in rehabilitation, Malec [/bib_ref] will be applied, as described for the pilot trial; clearly specified behavioural indicators of full and partial goal achievement will be defined when each goal is introduced, and progress according to these criteria will be rated by the therapist following session 10 and again following session 14.
(7) Therapist adherence to protocol Therapist adherence to the treatment protocol will be monitored through therapy logs and structured supervision sessions. Therapists will receive monthly telephone supervision and face-to-face supervision meetings will be held every 3 months. Therapy logs reporting session content (with participant details anonymised) will be submitted to the supervisor for scrutiny prior to supervision sessions.
## (8) treatment compliance
Treatment compliance will be indexed by the number of sessions completed for each participant.
## Procedure
Initial identification of participants will be made by National Institute for Health Research (NIHR) Comprehensive Local Research Network (CLRN) and Mental Health Research Network (MHRN) staff in England and National Institute of Social Care and Health Research Clinical Research Collaboration (NISCHR CRC) staff in Wales. Participants will be contacted by or on behalf of the clinician responsible for their care and invited to respond directly to the research team to express an interest in finding out more about the study. Interested participants and carers will then be contacted by telephone by the local research assistant, who will provide additional information and send out written details. This will be followed by a further telephone call; for those interested in finding out more, a meeting will be arranged at which the research assistant will explain the study, answer any questions they may have, re-check eligibility and ensure that the person with dementia has the capacity to consent. Consent from the participant and the carer will be taken at, or following, this visit.
At each centre, once participants have given informed consent, they will be visited by the research assistant who will conduct the baseline assessment. Following this assessment, the research assistant will trigger randomisation. Results of the randomisation will be sent to the therapist, who will telephone the participant and the carer to explain the next steps. Participants allocated to CR will receive 10 weekly visits from the therapist over a 3-month period. The therapist will trigger the postintervention assessment for all participants. The research assistant will visit each participant to conduct the assessment. Following the post-intervention assessment, participants in the CR group will receive four maintenance sessions with the therapist over a 6-month period. The research assistant will visit all participants 6 months after the post-intervention assessment to carry out the final 6-month follow-up assessment. All primary and secondary outcome measures, and service utilisation measures, will be administered at each assessment point.
After consent and baseline assessment, participants will be individually randomised. Randomisation to GREAT will be achieved by secure web access to the remote randomisation centre, NWORTH CTU, at Bangor University. This system will be set up, maintained and monitored independently of the trial statistician or other trial staff. The randomisation will be performed by dynamic allocation [bib_ref] Generalized method for adaptive randomization in clinical trials, Russell [/bib_ref] to protect against subversion while ensuring that the trial maintains good balance to the allocation ratio of 1:1 both within each stratification variable and across the trial. Participants will be stratified by centre, gender, age (under 75 vs. 75 and above) and MMSE score (under 24 vs. 24 and above). For validation purposes, additional information will be taken including the participant's trial number, initials, and date of birth, and details of the person requesting the randomisation. This is a single-blind trial. The researchers taking the measures will be blind to allocation, as will the data analysts. The importance of maintaining blinding will be emphasised in the training for both research assistants and therapists. As the participants are not blind to their treatment, at post-intervention and follow-up assessments participants will be specifically asked not to comment on the nature of their involvement in the study and not to reveal to the researcher whether or not they were visited by the therapist. Following each assessment, the blinded researcher will note to which condition s/he thought the participant had been allocated and how certain s/he was of the allocation. Sensitivity analyses will be performed to determine whether this knowledge affected participant scores. If there is evidence to suggest that consequential bias is present, the analysis will be adjusted to counteract this effect.
Other protection from bias will include the method of allocation to groups. The randomisation will be performed independently of the data analysis team by the CTU using a dynamic, stratified, web-based system designed to protect from bias by the unpredictability of the algorithm and the security of the web-based programme. Blinding will be maintained by automatic generation of randomisation codes and distribution via e-mail directly to the therapists responsible for implementing the treatment. Further bias protection will come from a "treatment as allocated" analysis, which will be the principal analysis performed on both primary and secondary outcomes. Treatment compliance measures will be restricted to inclusion in secondary analyses.
We will collect basic anonymised demographic data and reasons for not progressing to trial participation for all those people identified as warranting screening and invitation to the trial but declining to be screened or to participate. These data will be reported on a CONSORT diagram, together with information on the amount and nature of missing data, to enable readers to assess bias arising from recruitment or acceptability issues within the trial.
# Statistical analysis
Demographic and baseline data will be fully described and all outcome data will be analysed and reported. Significance will be assumed to be 5% throughout, and 95% confidence intervals will be quoted. All data will be anonymised and coded so that data collection and statistical analysis are blind to treatment allocation. The code will be broken only after the primary analysis has been completed. A fully pre-specified analysis plan will be prepared prior to the data being released to analysts. The analysis will be performed on a "Treatment as Allocated" principle to ensure protection against unintended bias. The data will be fully imputed in line with the predefined statistical analysis plan to minimise data loss due to missing values or time points. Sensitivity analyses (best case/worst case) will be performed to assess the influence of differing imputation assumptions. All trial reporting will be CONSORT-compliant [bib_ref] Statement: updated guidelines for reporting parallel group randomised trials, Schulz [/bib_ref].
For each outcome measure, at both post-intervention and follow-up, three analyses will be presented, the first two being unadjusted and adjusted treatment-as -allocated analyses and the third a treatment received analysis:
1. An unadjusted two-sample t-test by allocation group.
2. An analysis of covariance with baseline score and stratification variables as the covariates and allocated group as the condition factor. Between-group effect sizes with confidence intervals will be calculated using Cohen's d. Centre will be added as a fixed factor to test and quantify any site-specific effects, and if the number of practitioners is greater than the number of centres, practitioner will be added as a random factor. 3. A repeat of analysis 2 with treatment compliance factored in.
If CR is shown to be effective, additional forward stepping regression modelling will be undertaken to identify factors important in maximising the observed effects. Factors that will be investigated will include diagnostic category, medication status, educational level, social class, caregiver relationship to the person with dementia (spouse, adult child, other), and whether or not the carer is residing with the person with dementia.
For the cost-effectiveness analysis (CEA), service utilisation and carer input data will be collected using the CSRI. Unit costs will be attached to service use measures (from national reference costs, the PSSRU compendiumor calculated anew), CR costed in liaison with providers, and carer inputs valued using opportunity and replacement cost options. The CEA will look at changes over 9 months from each of two perspectives (health and social care; societal) in four analyses: cost of achieving an incremental change in BGSI; cost of achieving incremental changes in self-efficacy for participants with dementia; cost of achieving incremental QALY gains for participants with dementia; cost of achieving incremental QALY gains for carers. Incremental cost-effectiveness ratios will be computed as required and acceptability curves plotted for a range of willingness-to-pay values. Net-benefit regressions will make it possible to control for site, baseline outcome measures (where appropriate) and baseline costs, as well as gender, age and MMSE score. Sensitivity analyses will be conducted to test for different assumptions in the attachment of costs. We will also estimate the investment costs and net costs to the NHS and the social care system of making CR available nationally.
## Research governance
The research will be sponsored by Bangor University. The sponsor will ensure that appropriate indemnities are in place. The research will be overseen by a Trial Steering Committee and a Data Monitoring and Ethics Committee. Safety data will be routinely reported to the trial Data Monitoring and Ethics Committee (DMEC), and any suspected unexplained adverse reactions (SUSARs) noted will be reported to both the DMEC and the trial sponsor within established timeframes. Given the nature of the intervention, no interim analyses are planned, but such analyses may be requested at any time by the DMEC.
The issue of potentially competing studies will be monitored carefully. At the time of developing the protocol, only one study that could be perceived as potentially competing had been identified by local networks; that study was recruiting people with lower MMSE scores, and recruitment was due to be completed by May 2013. GREAT is a pragmatic trial and therefore participation would not preclude involvement in other clinical trials per se, unless those trials involved cognition-focused intervention. However, participant burden and inclusion/exclusion criteria in other, fastidious trials may preclude such dual participation. Each local PI will review any situation where there may be a conflict of recruitment pathways between trials to ensure that all potential participants are offered the most suitable option based on closest fit to eligibility criteria and participant preference.
## Service user involvement
Service users have been involved in the feasibility and pilot stages of the research leading to the development of this trial. The pilot trial benefitted greatly from the involvement of Alzheimer's Society Research Network Volunteers. During the development of the present study we again sought, and took into account, the views of Alzheimer's Society Research Network Volunteers and of service users contributing to the Dementia and Neurodegenerative Diseases Research Network (DeNDRoN) Patient and Public Involvement (PPI) programme. PPI for the GREAT trial will be provided through a partnership with the Alzheimer's Society; this will ensure that service users are fully involved with the design, delivery and dissemination of the research. Service users will be consulted at each stage of the trial to ensure optimal tailoring of study protocols and procedures. To ensure that PPI is integrated throughout the study, two service user representatives will sit on the Trial Steering Committee. The Alzheimer's Society Research Network will also contribute to dissemination activities towards the end of the study, ensuring that outcomes are communicated to lay audiences and policy-makers.
## Implementation within routine health-care provision
This will be the first multi-centre trial of individualised, goal-oriented cognitive rehabilitation for people with early-stage dementia. The CR approach offers a practical means of engaging people with dementia and carers in an early intervention process that aims to reduce functional disability and maximise engagement and participation, contributing to the possibility of living well with dementia. This approach can readily be offered by memory clinics in the period following diagnosis. Several UK memory services have already expressed interest in implementing CR. CR is also becoming acknowledged internationally; for example, it has recently been authorised for insurance reimbursement in Belgium, and has been conducted with the aid of trained volunteers in Canada. People with dementia have themselves begun to advocate for a rehabilitation approach. The trial will provide the necessary evidence base to extend these developments, if the findings demonstrate that CR is indeed both clinically beneficial and costeffective. Towards the end of the trial, we will build on the experience gained at each site to demonstrate how the CR approach can most effectively be integrated into routine health-care provision. Once recruitment for the trial comes to an end, we will develop materials and offer training for clinical teams and therapists, as well as preparing information for people with dementia, carers and the general public. Information about good practice in this area will be disseminated through a range of routes.
## Trial status
The GREAT trial started on 1 October 2012 and will be recruiting participants from 1 April 2013 to 30 September 2015. The end date for the trial is 31 December 2016.
[fig] Figure 1: Effects of intervention on goal performance and satisfaction (COPM ratings) for participants in each condition in the pilot trial: significant improvements for CR and no change for RT or TAU. [/fig]
[fig] Figure 2: GREAT trial CONSORT-style flow chart. [/fig]
[table] Table 1: Effect sizes (Cohen's d) on secondary outcome measures obtained in the pilot trial for the CR group compared to the pooled RT and TAU groups HADS Hospital Anxiety and Depression Scale, RBMT Rivermead Behavioural Memory Test, TEA Test of Everyday Attention, FAS letter fluency for letters F, A, and S, TEA ECD, TEA elevator counting with distraction, ILS Independent Living Scales, RSS Relatives' Stress Scale, GHQ-12, General Health Questionnaire. [/table]
|
Expanding the spectrum of “mesenchymal” tumors of the central nervous system
This is an open access journal distributed in accordance with the CC-BY-NC-ND (Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International) license: the work can be used by mentioning the author and the license, but only for non-commercial purposes and only in the original version. For further information: https://creativecommons. org/licenses/by-nc-nd/4.0/deed.enSummaryIn this review, we summarize the clinical, histopathological, and molecular features of central nervous system (CNS) tumors with BCOR internal tandem duplication, intracranial mesenchymal tumor with FET/CREB fusion, CNS CIC-rearranged sarcomas and primary intracranial sarcoma DICER1-mutant, now included in the 2021 WHO classification of CNS tumors. Possible relationships between tumors occurring in the CNS and their systemic counterparts are discussed.
# Introduction
In recent years, large molecular studies have permitted a better classification of undifferentiated, poorly differentiated round cell tumors, mainly under the definition of CNS-PNETs, or spindle cell neoplasms occurring in the CNS 1 . One important result of these studies was the widening of the spectrum of "mesenchymal" neoplasms potentially occurring in the CNS; this group includes now mesenchymal tumor with FET/CREB fusion, CIC-rearranged sarcomas and DICER1-mutant intracranial sarcoma. Furthermore, these analyses led to the identification of neuroepithelial tumors which share a common genetic background with pediatric sarcomas, like CNS tumors with BCOR internal tandem duplication (BCOR ITD). Even though many aspects of biology of these rare tumors have been better defined, some intriguing issues are still to be definitively addressed, and especially their exact relationships with their systemic counterparts.
## Incidence, distribution and localisation
Although a large number of cases have been progressively documented in recent years, these neoplasms are rare and precise information regarding their incidence and prevalence in general population is still not available. Based on available literature, the two most frequently reported entities seem to be mesenchymal tumor with FET/CREB fusion and the CNS tumor BCOR ITD 2 . Intracranial mesenchymal tumor with FET/CREB fusion occurs more frequently in adult population: in a large series, the median patient age at presentation was 17 years with a predominance in the female population [bib_ref] Intracranial mesenchymal tumor with FET-CREB fusion -A unifying diagnosis for the spectrum..., Sloan [/bib_ref]. They are mainly extra-axial or intraventricular tumors, localized, in particular, in hemispheric meninges, the falx and the tentorium In the CNS, tumor in the BCOR ITD and CIC-rearranged sarcoma affect mainly the pediatric population: infants for CNS tumor BCOR ITD (mean age 3.5 years) [bib_ref] High-grade neuroepithelial tumor with BCOR exon 15 internal tandem duplication -a comprehensive..., Ferris [/bib_ref] , while adolescents and young adult (usually ≤ 21 years) for CIC-rearranged sarcoma. CNS tumors BCOR ITD showed predominantly a cerebellar localization, especially in younger patients (< 5 years old), although they can also occur in cerebral hemispheres 2,6,7 ; CNS CIC-rearranged sarcomas may show variable localization including cases at spinal level 1,8-10 . The mean age of patients with primary DICER1-mutant intracranial sarcoma is 6 years with a wide range of age and predominant intracranial localization [bib_ref] Primary intracranial spindle cell sarcoma with rhabdomyosarcoma-like features share a highly distinct..., Koelsche [/bib_ref] [bib_ref] DICER1-associated central nervous system sarcoma in children: comprehensive clinicopathologic and genetic analysis..., Kamihara [/bib_ref] [bib_ref] Expanding the spectrum of dicer1-associated sarcomas, Warren [/bib_ref].
## Integrated diagnosis
A combined histopathological-and molecular-based approach is often necessary to reach the final diagnosis. A variable histopathological spectrum and an unspecific immunohistochemical profile are common, hindering an easy recognition in routine diagnostic neuropathology. Intracranial mesenchymal tumor with FET/CREB fusion is a mesenchymal neoplasm characterized by the presence of fusion of a FET RNA-binding protein family gene (usually EWSR1, rarely FUS) with a member of the CREB transcription factors family (CREB1, ATF1, or CREM). Such cases have been reported in the past as angiomatoid fibrous histiocytoma of the meninges (AFH) or intracranial myxoid mesenchymal tumors (IMMT) [bib_ref] Intracranial mesenchymal tumor with FET-CREB fusion -A unifying diagnosis for the spectrum..., Sloan [/bib_ref] [bib_ref] Primary Intracerebral Angiomatoid Fibrous Histiocytoma, Dunham [/bib_ref] [bib_ref] Primary intracranial angiomatoid fibrous histiocytoma presenting with anaemia and migrainelike headaches and..., Hansen [/bib_ref] [bib_ref] Primary myxoid mesenchymal tumour with intracranial location: report of a case with..., Sciot [/bib_ref] [bib_ref] ESWR1-CREM Fusion in an Intracranial Myxoid Angiomatoid Fibrous Histiocytoma-Like Tumor: A Case..., Gareton [/bib_ref] [bib_ref] Intracranial Myxoid Mesenchymal Tumor With EWSR1-ATF1 Fusion, Ballester [/bib_ref] [bib_ref] Intracranial angiomatoid fibrous histiocytoma with rhabdoid features: a mimic of rhabdoid meningioma, Vizcaino [/bib_ref] [bib_ref] Intracranial myxoid mesenchymal tumors with EWSR1 -CREB family gene fusions: myxoid variant..., Bale [/bib_ref]. They may display extremely variable histopathological features, including presence of myxoid stroma, desmoplastic areas, epithelioid and spindle cell cytology 3,15-22 [fig_ref] Figure 1: Histopathological features of intracranial mesenchymal tumors with FET/CREB fusion [/fig_ref]. Angiomatous areas and intralesional inflammatory infiltrates are common, as seen in AFH of soft tissues [bib_ref] Intracranial mesenchymal tumor with FET-CREB fusion -A unifying diagnosis for the spectrum..., Sloan [/bib_ref] [bib_ref] An integrative histopathological and epigenetic characterization of primary intracranial mesenchymal tumors, FET:CREB-fused..., Tauziède-Espariat [/bib_ref] [bib_ref] Primary intracranial angiomatoid fibrous histiocytoma presenting with anaemia and migrainelike headaches and..., Hansen [/bib_ref]. Meningioma-like areas and amianthoid fibers can be detected. The immunophenotype of tumor cells is variable. The cells may often express vimentin, EMA and CD99 and could be CD68 and CD163 positive [bib_ref] Intracranial mesenchymal tumor with FET-CREB fusion -A unifying diagnosis for the spectrum..., Sloan [/bib_ref] [bib_ref] Intracranial Myxoid Mesenchymal Tumor With EWSR1-ATF1 Fusion, Ballester [/bib_ref] [bib_ref] Intracranial angiomatoid fibrous histiocytoma with rhabdoid features: a mimic of rhabdoid meningioma, Vizcaino [/bib_ref] [bib_ref] Intracranial myxoid mesenchymal tumors with EWSR1 -CREB family gene fusions: myxoid variant..., Bale [/bib_ref]. A common and unique feature is a focal immunoreactivity for desmin [bib_ref] Intracranial mesenchymal tumor with FET-CREB fusion -A unifying diagnosis for the spectrum..., Sloan [/bib_ref] [bib_ref] An integrative histopathological and epigenetic characterization of primary intracranial mesenchymal tumors, FET:CREB-fused..., Tauziède-Espariat [/bib_ref] [bib_ref] Intracranial myxoid mesenchymal tumor with EWSR1-CREB1 gene fusion: a case report and..., Komatsu [/bib_ref] [bib_ref] ESWR1-CREM Fusion in an Intracranial Myxoid Angiomatoid Fibrous Histiocytoma-Like Tumor: A Case..., Gareton [/bib_ref] [bib_ref] Intracranial Myxoid Mesenchymal Tumor With EWSR1-ATF1 Fusion, Ballester [/bib_ref] [bib_ref] Intracranial angiomatoid fibrous histiocytoma with rhabdoid features: a mimic of rhabdoid meningioma, Vizcaino [/bib_ref] [bib_ref] Intracranial myxoid mesenchymal tumors with EWSR1 -CREB family gene fusions: myxoid variant..., Bale [/bib_ref]. Cytokeratins, glial markers, melanocytic markers are usually negative [bib_ref] Intracranial mesenchymal tumor with FET-CREB fusion -A unifying diagnosis for the spectrum..., Sloan [/bib_ref]. These tumors show often a low proliferative activity 20 , but cases with increased proliferative and mitotic activity have been reported [bib_ref] An integrative histopathological and epigenetic characterization of primary intracranial mesenchymal tumors, FET:CREB-fused..., Tauziède-Espariat [/bib_ref] [bib_ref] Intracranial Myxoid Mesenchymal Tumor With EWSR1-ATF1 Fusion, Ballester [/bib_ref] [bib_ref] Intracranial angiomatoid fibrous histiocytoma with rhabdoid features: a mimic of rhabdoid meningioma, Vizcaino [/bib_ref]. The final diagnosis relies on the identification of translocation involving FET family genes with CREB transcription factors genes. Notably, epithelioid features seem to be associated to EWSR1-ATF1 fusions [bib_ref] Intracranial mesenchymal tumor with FET-CREB fusion -A unifying diagnosis for the spectrum..., Sloan [/bib_ref] [bib_ref] Intracranial angiomatoid fibrous histiocytoma with rhabdoid features: a mimic of rhabdoid meningioma, Vizcaino [/bib_ref]. Because molecular alterations involving EWSR1 are not specific and can be identified in other CNS tumors, extreme caution must be used in the interpretation of results obtained with FISH-based methods: a further confirmation of the presence of a specific FET-CREB fusion with other methods should be recommended. Intracranial mesenchymal tumor with FET/CREB fusion may present additional mutations in several genes, including BRAF 22 . CNS tumor BCOR ITD is as a malignant neoplasm characterized by a predominantly solid or microcystic growth pattern, glial-like cytology, a dense capillary network, formation of pseudo-rosettes, and by the presence of an ITD in exon 15 of the BCOR gene. Differently from other entities described herein, this tumor is considered to derive from a neuroepithelial cell of origin (see below) and therefore included in the CNS embryonal tumors along to medulloblastoma, ETMR and ATRT 2 . This neoplasm can present variable neuropathological features and immunohistochemical profile. CNS tumor BCOR ITD often resembles a glial neoplasm and can show solid, microcystic or perivascular architecture, often with a prominent, dense vascular stroma [fig_ref] Figure 2: Histopathological features of CNS tumor BCOR ITD [/fig_ref]. The cells, usually round or oval, often appear embedded in a glioma-like fibrillary stroma; Homer-Wright rosettes can be found; the tumor may show in part an infiltrative growth pattern; usually these neoplasms display high proliferative and mitotic activity 6,23-25 . The differential diagnosis, given the common occurrence in the posterior fossa in children, should include medulloblastoma, ependymomas and ET-MR. From an immunohistochemical point of view, the neoplasm, usually CD56 and vimentin positive, may present a variable expression of glial markers such as GFAP, OLIG2 and S100 but may also show positivity for NeuN; the expression of other neuronal markers is uncommon 6,23-25 . SATB2, BCL2 and TLE1 as well as pan-NTRK are usually positive [bib_ref] Intra-and extra-cranial BCOR-ITD tumours are separate entities within the BCOR-rearranged family, Bouchoucha [/bib_ref]. The widespread positivity for BCOR, EGFR and Cyclin D1 is helpful but BCOR over-expression is unspecific and can be encountered in numerous other CNS neoplasms, including high-grade gliomas [bib_ref] BCOR involvement in cancer, Astolfi [/bib_ref]. The definitive diagnosis relies on the identification of the presence of duplication in the 15 exon of BCOR gene [fig_ref] Figure 3: BCOR Internal Tandem Duplication [/fig_ref]. Notably, several BCOR alterations have been also described in other CNS neoplasms, including mutations and fusions [bib_ref] BCOR involvement in cancer, Astolfi [/bib_ref]. Mutations have been identified in retinoblastomas, in various glial tumors, particularly those with high-grade histol-ogy, including high-grade astroblastomas. Moreover, BCOR genetic alterations were also found in about 5% of medulloblastomas being apparently more common in infantile SHH-medulloblastoma subgroup [bib_ref] BCOR involvement in cancer, Astolfi [/bib_ref]. CIC-rearranged sarcoma of the CNS is a high-grade, . Angiomatous vessels were present at the periphery (C). Another case, showed a more solid, fibro-histiocytic histology (D) with abundant inflammatory infiltrates at its periphery (E). The tumor was partially positive for desmin (F). Another tumor, occurred in a 55 years old patient and localized in the frontal meninges, showed delicate myxoid features (H). Prominent vascularity and inflammatory infiltrates were present (arrow)(I). All tumors presented EWSR1 rearrangement in FISH analysis. showing a more solid growth pattern; necrosis is common; epithelioid or spindle cell cytology as long as desmoplastic or myxoid stroma have been also described [bib_ref] Primary Spinal Epidural CIC-DUX4 Undifferentiated Sarcoma in a Child, Donahue [/bib_ref] [fig_ref] Figure 4: Histopathological features of CIC-rearranged sarcoma [/fig_ref]. In most cases, the differential diagnosis includes other round cell neoplasms, like rhabdomyosarcoma and Ewing's sarcoma. WT1 and ETV4 positivity, a weak CD99 expression and negativity for NKX2-2 and FLI1 stainings can be very helpful to orientate the diagnosis [bib_ref] Primary Spinal Epidural CIC-DUX4 Undifferentiated Sarcoma in a Child, Donahue [/bib_ref] [bib_ref] Sarcomas With CICrearrangements Are a Distinct Pathologic Entity With Aggressive Outcome: A..., Antonescu [/bib_ref]. Heterogeneous ERG/CD31 co-expression in a subset of CIC-rearranged sarcoma may be a potential pitfall in differential diagnosis with vascular tumors [bib_ref] Co-expression of ERG and CD31 in a subset of CIC-rearranged sarcoma: a..., Kojima [/bib_ref]. Expression of pan-cytokeratin, smooth muscle actin, and neurofilament protein has been reported [bib_ref] A case report of CIC-rearranged undifferentiated small round cell sarcoma in the..., Ito [/bib_ref] [bib_ref] Primary Spinal Epidural CIC-DUX4 Undifferentiated Sarcoma in a Child, Donahue [/bib_ref]. The final diagnosis relies on demonstrating CIC-rearrangement using FISH, NGS-or RT-based methods. Primary intracranial sarcoma DICER1-mutant composed of spindle or pleomorphic tumor cells, with evidence of myogenic and occasionally of chondroid differentiation. The tumors show often spindle cell cytology with anaplastic features, focal rhabdomyoblastic differentiation, foci of primitive embryonaltype tissue and in some cases chondroid differentiation 11,31-34 [fig_ref] Figure 5: Histopathological features of DICER1-mutant sarcoma [/fig_ref]. The presence of eosinophilic globules is also a typical feature 11 : they are PAS+ and are stained with alpha-1-antitrypsin. The tumor shows patchy expression of muscular markers (like desmin and myogenin) highlighting the rhabdomyosarcomatous component. Nuclear positivity for TLE1 expression [bib_ref] Loss of histone H3 trimethylation on lysine 27 and nuclear expression of..., Alexandrescu [/bib_ref] and loss of H3K27me3 were observed in primary intracranial sarcoma DICER1-mutant [bib_ref] DICER1-associated central nervous system sarcoma in children: comprehensive clinicopathologic and genetic analysis..., Kamihara [/bib_ref] [bib_ref] Loss of histone H3 trimethylation on lysine 27 and nuclear expression of..., Alexandrescu [/bib_ref] : therefore, these tumors may be included in the differential diagnosis of high-grade cellular malignant spindle cell neoplasms with loss of H3K27me3 (in particular MPNST) [bib_ref] DICER1-associated central nervous system sarcoma in children: comprehensive clinicopathologic and genetic analysis..., Kamihara [/bib_ref]. The histology of primary intracranial sarcoma DICER1-mutant overlaps considerably with other DICER1-associated tumors, notably to type II/III pleuropulmonary blastoma (PPB): therefore, distinguishing primary intracranial sarcoma, DICER1-mutant from metastatic PPB is pivotal [bib_ref] DICER1 tumor predisposition syndrome: an evolving story initiated with the pleuropulmonary blastoma, González [/bib_ref]. Information, if any, on a possible presence of a DIC-ER1 syndrome or presence of other DICER1-related neoplasms in clinical history of the patient significantly may further facilitate the neuropathological approach to diagnosis. Confirmation of presence of DICER1 mutation in sporadic cases is mandatory for the final diagnosis.
## Taxonomy and oncogenesis
The most intriguing issue for these CNS neoplasms in particular for CNS BCOR ITD, CIC-rearranged sarcoma and for intracranial mesenchymal tumor with FET/ CREB fusion is the relationship with their systemic counterparts. Tumors harboring BCOR-internal tandem duplication represent a histologically heterogeneous group of neoplasms, comprising CNS tumors and sarcomas; the latter group includes clear cell sarcomas of the kidney (CCSK), high-grade endometrial stromal sarcomas (HG-ESS), myxoid mesenchymal tumor of infancy (PMMTI) and undifferentiated round cell sarcoma (URCS) in bone and soft tissues [bib_ref] Consistent in-frame internal tandem duplications of BCOR characterize clear cell sarcoma of..., Ueno-Yokohata [/bib_ref]. BCL-6 transcriptional corepressor (BCOR) (located at Xp11.4) encodes a protein which functions as a corepressor bounding to BCL-6 and forming part of the Polycomb Repressive Complex 1 which through ubiquitination leads epigenetic silencing [bib_ref] BCOR involvement in cancer, Astolfi [/bib_ref]. Besides ITD, various alterations affecting BCOR gene have been described: they include fusions (i.e., BCOR-CCNB3, BCOR-MAML3 and ZC3H7B-BCOR), mutations and internal tandem duplications (ITD) [bib_ref] BCOR involvement in cancer, Astolfi [/bib_ref]. The main similarities between CNS and non-CNS tumors BCOR ITD neoplasms are clinical and include the median age of the patients at the time of diagnosis, the local presentation and the poor prognosis. The histopathology, however, varies greatly and depends on the tissue and site of occurrence [bib_ref] Intra-and extra-cranial BCOR-ITD tumours are separate entities within the BCOR-rearranged family, Bouchoucha [/bib_ref] [bib_ref] CNS high-grade neuroepithelial tumor with BCOR internal tandem duplication: a comparison with..., Yoshida [/bib_ref]. Moreover, CNS tumors BCOR ITD and BCOR ITD sarcomas present close but distinct transcriptomic signature and DNA methylation profile, suggesting the possibility of a common, acquired oncogenic pattern in distinct cell types within specific tissues. Notably, expression analysis revealed that the CNS BCOR ITD group seems to be enriched in genes expressed in neuro-glial cells whereas BCOR ITD sarcomas predominantly expressed embryonal/developmental genes: this correlate well at histopathological level, with the evidence that CNS tumors BCOR ITD, but not BCOR ITD sarcomas express neuro-glial markers (like GFAP, Olig2, Neu-N). The difference in molecular signatures is probably dependent on their different cell of origin. As previously mentioned, the putative origin of CNS BCOR ITD from neuroepithelial cells led to the inclusion of this tumor among CNS embryonal tumors, rather than mesenchymal neoplasms in WHO classification of CNS tumors. The relationships between intracranial mesenchymal tumors with FET/CREB fusion and their systemic counterpart is more difficult be addressed. Soft tissue AFH, pulmonary CREB sarcoma and intracranial mesenchymal tumors with FET/CREB fusion share many features but also present some differences at histopathological level (i.e. desmin expression) [bib_ref] An integrative histopathological and epigenetic characterization of primary intracranial mesenchymal tumors, FET:CREB-fused..., Tauziède-Espariat [/bib_ref]. Methylation profiling analysis added interesting information on taxonomy of those tumors occurring in the CNS. CNS tumors seem to be different from their soft tissue counterparts, suggesting the existence of two epigenetic subgroups [bib_ref] An integrative histopathological and epigenetic characterization of primary intracranial mesenchymal tumors, FET:CREB-fused..., Tauziède-Espariat [/bib_ref] [bib_ref] Intracranial mesenchymal tumors with FET-CREB fusion are composed of at least two..., Sloan [/bib_ref]. The first subgroup includes tumors that have epigenomic similarities with the systemic AFH and SFT and have mostly EWSR1-ATF1 and ESWR1-CREB1 fusions; these tumors occur mostly in adolescents and young adults, show a spindle cell cytology and a hemangioma-like vascularity. The second subgroup includes tumors with epigenomic similarities with the clear cell sarcoma of soft tissue (CCS) enriched in cases mostly with ESWR1-CREM and FUS-CREM fusions; they occur in early childhood and show a round or epithelioid/rhabdoid morphology and lack of hemangioma-like features [bib_ref] Intracranial mesenchymal tumors with FET-CREB fusion are composed of at least two..., Sloan [/bib_ref]. There is no sufficient evidence to distinguish CIC-rearranged sarcoma in the CNS from histologically and genetically similar tumors in other extra CNS tissues. Despite a similar histopathology, the majority of intracranial sarcomas with CIC-rearrangement studied to date showed NUTM1 and LEUTX as the fusion partner, whereas those in extracranial bone and soft tissue had DUX4, FOXO4 and NUTM2A as the fusion partner [bib_ref] Clinicopathologic Features of CIC-NUTM1 Sarcomas, a New Molecular Variant of the Family..., Loarer [/bib_ref] [bib_ref] cIMPACT-NOW update 6: new entity and diagnostic principle recommendations of the cIMPACT-Utrecht..., Louis [/bib_ref] [bib_ref] Report: A Unique Case of Pediatric Central Nervous System Embryonal Tumor Harboring..., Hu [/bib_ref]. CIC gene product encodes a member of the high mobility group (HMG)-box superfamily of transcriptional repressors. Gene truncation (besides fusions) may be also sufficient to enable an oncogenic de-repression of transcription. Primary intracranial sarcoma, DICER1-mutant, as previously indicated, overlaps considerably in terms of histopathology with other DICER1-associated tumors particular with PPB. These tumors occur in a specific genetic setting defined by mutations in the DICER1 gene (either somatic or germline as part of the DICER1 syndrome). DICER1 syndrome is a rare autosomal dominant familial tumor predisposition disorder with a heterozygous germline mutation of DICER1 gene (chromosome 14, region q32.13) that increases the risk of development of different types of malignant and benign tumors [bib_ref] Extending the phenotypes associated with DICER1 mutations, Foulkes [/bib_ref]. Patients with DIC-ER1 syndrome commonly develop pleuropulmonary blastoma (PPB), multinodular goiter, ovarian Sertoli-Leydig cell tumors, and rarely CNS tumors including ETMR, pituitary blastoma and pineoblastoma [bib_ref] Extending the phenotypes associated with DICER1 mutations, Foulkes [/bib_ref]. The DICER1 gene encodes an RNase III endoribonuclease that facilitates the activation of the RNA-induced silencing complex essential for double stranded-RNA and mi-RNA processing. Disruption of this pathway results in alterations in protein expression and in cell proliferation as well as derangement of cell differentiation and DNA repair [bib_ref] DICER1 tumor predisposition syndrome: an evolving story initiated with the pleuropulmonary blastoma, González [/bib_ref]. DICER1 behaves as either a tumor suppressor gene due to loss-of-function mutations or an oncogene, due to gain-of-function mutations. It retained function as a haploinsufficient tumorsuppressor gene with the loss of one allele leading to tumor progression, but loss of both alleles having an inhibitory effect for tumor development; therefore, one intact allele is needed for cell survival [bib_ref] DICER1 tumor predisposition syndrome: an evolving story initiated with the pleuropulmonary blastoma, González [/bib_ref]. Recent investigations revealed that some genetic alterations seem to be more common in intracranial DICER1-mutant sarcoma compared with other DIC-ER1-associated tumors; they include TP53 inactivation and activating alterations of genes in the RAS pathway (KRAS and NF1) [bib_ref] Primary intracranial spindle cell sarcoma with rhabdomyosarcoma-like features share a highly distinct..., Koelsche [/bib_ref]. Moreover, it seems that primary intracranial sarcoma, DICER1-mutant, may have a significantly higher tumor mutational burden in comparison to other DICER1-related tumors 12 .
## Prognosis and outcome
Given their rarity, it is difficult to draw any assumptions in term of outcome and prognosis. CIC-rearranged sarcoma and CNS tumor BCOR ITD show generally aggressive clinical behavior 8 ; the initial therapeutic strategy is mainly multimodal including most often the first-line surgery, radiotherapy and adjuvant chemotherapy [bib_ref] High-grade neuroepithelial tumor with BCOR exon 15 internal tandem duplication -a comprehensive..., Ferris [/bib_ref]. However, cases with prolonged survival have been described [bib_ref] High-grade neuroepithelial tumor with BCOR exon 15 internal tandem duplication -a comprehensive..., Ferris [/bib_ref] [bib_ref] A single supratentorial highgrade neuroepithelial tumor with two distinct BCOR mutations, exceptionally..., Bremer [/bib_ref]. For CIC-rearranged sarcoma, it remains uncertain whether the specific fusion partner may influence the biology and therefore the prognosis of the patients with these tumors. The clinical course of intracranial mesenchymal tumor with FET/CREB fusion is unpredictable, ranging from cases with a relatively indolent behavior, to tumors prone to rapid recurrence 3 . The prognosis for patients with DICER1-mutant primary intracranial sarcoma remains unknown, because only limited data clinical for patient with longterm follow-up are available. Moreover, the possible prognostic relevance of tumors arising in the settings of germline DICER1 mutations has not been defined There is no evidence that the presence of a germline or somatic mutation may influence the prognosis of a patient with DICER1-mutant intracranial sarcoma.
# Conclusions
In conclusion, the WHO classification of CNS tumors 2021 now includes new entities among the groups of mesenchymal non-meningothelial tumors and embryonal tumors. These tumors share specific molecular and histopathological features. Given their rarity in the CNS, particular awareness is needed for practicing neuropathologists to suspect the presence of such tumors in routine neuropathology: a combined histopathological and molecular-based approach is therefore necessary to pinpoint the final diagnosis.
## Aknowledgments
Special thanks to Prof. Caterina Giannini for providing histopathological images of the CIC-rearranged sarcoma case and to Dr. Isabella Giovagnoni e Dr. Sabina Barresi for the BCOR ITD exon 15 graphics.
[fig] Figure 1: Histopathological features of intracranial mesenchymal tumors with FET/CREB fusion. A case of intracranial mesenchymal tumors with FET/CREB fusion, in the posterior fossa of a 60 years old patient, showed dense cellularity and spindle cell cytology, alternating with myxoid areas (A-B) [/fig]
[fig] Figure 2: Histopathological features of CNS tumor BCOR ITD. A CNS Tumor BCOR ITD affecting the cerebellum in a twoyear-old boy. The tumor presented with solid or microcystic growth pattern (A, C). The cells were embedded in a fibrillary stroma and appeared arranged around vascular structures (B, D). The tumor was partially OLIG2 positive (E) and diffusely BCOR positive (F). [/fig]
[fig] Figure 3: BCOR Internal Tandem Duplication (BCOR ITD).Schematic plot of internal tandem duplication in BCOR exon 15 (BCOR ITD) at chromosome X (in A). In OGM Circle Plot (Access Software, Bionano Technologies) chromosomal rearrangement are shown as green line (fusion) involving parts of BCOR gene (in B.). The in-frame fusion (depicted in light blue, arrow) affects the BCOR PUFD domain leading to an altered PCGF1-binding affinity, dysregulating transcription (in C.). [/fig]
[fig] Figure 4: Histopathological features of CIC-rearranged sarcoma. A CIC-rearranged sarcoma (harboring a CIC-LEUTX fusion) occurred in the spinal cord of a 4 years old boy. The tumor was extramedullary. The tumor was composed of large epithelioid cells with abundant cytoplasm (A). Necrotic changes were present (B). [/fig]
[fig] Figure 5: Histopathological features of DICER1-mutant sarcoma. A DICER1-mutant intracranial sarcoma in a 18 years old female patient localized in the temporal lobe. The tumor showed high cellularity, partial spindle cell cytology and myxoid stroma(A, B). Pleomorphic cells and hyaline globules were seen (C) along as highly cellular round cell undifferentiated areas (D): here focal myogenin positive staining was present (E). Partial loss of H3K27me3 was found (F). The tumor harbored a DICER1 p.E1813A mutation. [/fig]
|
The challenge to verify ceramide's role of apoptosis induction in human cardiomyocytes - a pilot study
Background: Cardioplegia and reperfusion of the myocardium may be associated with cardiomyocyte apoptosis and subsequent myocardial injury. In order to establish a pharmacological strategy for the prevention of these events, this study aimed to verify the reliability of our human cardiac model and to evaluate the pro-apoptotic properties of the sphingolipid second messenger ceramide and the anti-apoptotic properties of the acid sphingomyelinase inhibitor amitryptiline during simulated cardioplegia and reperfusion ex vivo. Methods: Cardiac biopsies were retrieved from the right auricle of patients undergoing elective CABG before induction of cardiopulmonary bypass. Biopsies were exposed to ex vivo conditions of varying periods of cp/rep (30/10, 60/20, 120/40 min). Groups: I (untreated control, n = 10), II (treated control cp/rep, n = 10), III (cp/rep + ceramide, n = 10), IV (cp/rep + amitryptiline, n = 10) and V (cp/rep + ceramide + amitryptiline, n = 10). For detection of apoptosis anti-activated-caspase-3 and PARP-1 cleavage immunostaining were employed. Results: In group I the percentage of apoptotic cardiomyocytes was significantly (p < 0.05) low if compared to group II revealing a time-dependent increase. In group III ceramid increased and in group IV amitryptiline inhibited apoptosis significantly (p < 0.05). In contrast in group V, under the influence of ceramide and amitryptiline the induction of apoptosis was partially suppressed. Conclusion: Ceramid induces and amitryptiline suppresses apoptosis significantly in our ex vivo setting. This finding warrants further studies aiming to evaluate potential beneficial effects of selective inhibition of apoptosis inducing mediators on the suppression of ischemia/reperfusion injury in clinical settings.
# Introduction
Cardioplegia and reperfusion of the myocardium are essential techniques employed in many cardiac surgical procedures when a temporarily arrested myocardium is required. However, as a consequence of exposure to cardioplegia and reperfusion apoptosis of cardiomyocytes may occur [bib_ref] Role of nitric oxide in Ca2+ sensitivity of the slowly activating delayed..., Bai [/bib_ref]. Apoptosis is the ultimate result of multiple convergent signalling pathways, which are triggered by events such as nutrient and oxygen deprivation, intracellular calcium overload and excessive reactive oxygen species production [bib_ref] Role of nitric oxide in Ca2+ sensitivity of the slowly activating delayed..., Bai [/bib_ref]. In the setting of cardiac surgery these events can finally result in contractile dysfunction of the myocardium [bib_ref] Protein kinase Cdelta activation induces apoptosis in response to cardiac ischemia and..., Murriel [/bib_ref] and atrial fibrillation [bib_ref] Determination of histopathologic risk factors for postoperative atrial fibrillation in cardiac surgery, Ak [/bib_ref]. Apoptosis of cardiac non-myocytes also contributes to maladaptive remodelling and the transition to decompensated congestive heart failure [bib_ref] A synopsis of research in cardiac apoptosis and its application to congestive..., Khoynezhad [/bib_ref]. Regarding this potentially impact of apoptosis on clinical outcomes, there is a demand for pharmacological strategies. Pharmacological blockade has been shown to reduce apoptosis during extracorporeal circulation in an animal model [bib_ref] Hasichaonu : Carvedilol attenuates CPB-induced apoptosis in dog heart: regulationof Fas/FasL and..., Zhang [/bib_ref]. In contrast to that we have successfully established a human cardiac model, which we have presented recently [bib_ref] The nonselective beta-blocker carvedilol suppresses apoptosis in human cardiac tissue: a pilot..., Usta [/bib_ref] [bib_ref] Suppressing apoptosis with milrinone simulating extracorporeal circulation: a pilot study, Usta [/bib_ref] [bib_ref] Human cardiac tissue in a microperfusion chamber simulating extracorporeal circulationischemia and apoptosis..., Usta [/bib_ref].
Our present pilot study was performed just as a sequel to our recent work [bib_ref] The nonselective beta-blocker carvedilol suppresses apoptosis in human cardiac tissue: a pilot..., Usta [/bib_ref] [bib_ref] Suppressing apoptosis with milrinone simulating extracorporeal circulation: a pilot study, Usta [/bib_ref] [bib_ref] Human cardiac tissue in a microperfusion chamber simulating extracorporeal circulationischemia and apoptosis..., Usta [/bib_ref] to further evaluate our presented human cardiac model during simulated cardioplegia and reperfusion ex vivo respectively the end-points feasibility and reliability. We conducted this study to clarify if another pathway of apoptosis induction in cardiomyocytes exists. Our aim was to evaluate during ex vivo simulated cardioplegia and reperfusion the effect of the sphingolipid second messenger ceramide and the anti-apoptotic properties of the sphingomyelinase inhibitor amitryptiline respectively the end-point apoptosis induction and reduction in cardiomyocytes which to our knowledge has not been described in such an experimental setting yet. The results should clarify if any clinical potential utilization could be favoured.
# Materials and methods
# Ethics declaration
The investigation conforms with the principles outlined in the Declaration of Helsinki. In addition, approval was granted by the Ethics Committee of the Faculty of Medicine of the Eberhard-Karls-University, Tübingen, Germany (approval reference number 40/2007 V).
## Patient characteristics
The study protocol was approved by the ethics committee of the Faculty of Medicine of the Eberhard-Karls-University Tübingen. 20 patients undergoing elective CABG surgery were included in this study and gave informed consent for study participation. Mean patient age was 65 years (range 45-70). Mean body mass index 28 kg/m 2 (range [bib_ref] Death receptors: signaling and modulation, Ashkenazi [/bib_ref] [bib_ref] Hypoxia and acidosis activate cardiac myocyte death through the Bcl-2 family protein..., Kubasiak [/bib_ref] [bib_ref] Cell death: critical control points, Danial [/bib_ref] [bib_ref] Intracellular transport and metabolism of sphingomyelin, Koval [/bib_ref] [bib_ref] Sphingolipids: second messengers, mediators and raft constituents in signaling, Prieschl [/bib_ref] [bib_ref] De novo ceramide regulates the alternative splicing of caspase 9 and Bcl-x..., Chalfant [/bib_ref] [bib_ref] Caspases and caspase inhibitors, Villa [/bib_ref] [bib_ref] ICE-LAP6, a novel member of the ICE/Ced-3 gene family, is activated by..., Duan [/bib_ref]. Mean left ventricular ejection fraction 63% (range 55-75). Mean number of diseased coronary vessels 3 (range 2-3). Mean number of infarctions 1 (range 1-3) in patients history. The basic medication of all patients consisted of β-blockers (Beloc Zok™ 47.5 mg twice per die, angiotensin converting enzyme inhibitors, statins and diuretics. All patients had a sinus rhythm.
# Material
Human tissue was retrieved from the auricle of the right atrium of patients before cardiopulmonary-bypass (CPB) and was processed immediately. Each biopsy was transmuraly divided in thirteen pieces with [0.5 to 1 cm 2 ] size, which were placed separately in microperfusion chambers with continuous perfusion. Cardiac specimens were outside the body before being mounted and tested in the chamber system for a maximum of 30 min, but during this period the oxygen supply was maintained continuously by bubble-oxygenating the Krebs-Henseleit buffer in the petri dish (Greiner Bio-One, Frickenhausen Germany).
## Chemicals and buffer solutions
## Cardioplegic solution
Cardioplegic solution was prepared on the basis of Cafree KH consisting of 115 mM NaCl, 4.5 mM KCl, 1.18 mM MgCl 2 , 0.5 mM EGTA, 1.23 mM NaH 2 PO 4 , 1.19 mM Na 2 SO 4 , 80 mM Glucose, and 10 mM HEPES, pH adjusted to 7.4 at 37°C with NaOH. Furthermore, a solution containing 20 mM Tris hydroxymethyl-aminomethane, 60 mmol K + and anionic polypeptides to the isoionic point was added in a 1:4 proportion to Ca-free KH buffer. This solution served as cardioplegic solution and was administered at 4°C, in analogy to our clinical regimen. The resulting K + concentration in this mixture was 16.5 mM.
## Ceramide
Sphingolipids are constituents of cellular membranes and of lipoproteins. The common backbone is the long chain amino base sphingosine (trans-4-sphingenine), and the ceramides refer to the N-acyl derivatives of sphingosine. For a decade now, ceramides have been widely studied as regulators of major cellular functions, i.e., apoptosis, proliferation, or senescence [bib_ref] Functions of ceramide in coordinating cellular responses to stress, Hannun [/bib_ref] [bib_ref] Regulation of ceramide production and apoptosis, Kolesnick [/bib_ref] [bib_ref] Ceramide in the antiapoptotic effect of ischemic preconditioning, Argaud [/bib_ref]. Apoptosis induction with short chain ceramide (20-50 μM) supports the view that ceramides are able to trigger apoptosis [bib_ref] Inhibition of Akt kinase by cell-permeable ceramide and its implications for ceramide-induced..., Zhou [/bib_ref]. The concentration of ceramide employed in this study was 50 μM, similar to previous experimental settings [bib_ref] Inhibition of Akt kinase by cell-permeable ceramide and its implications for ceramide-induced..., Zhou [/bib_ref].
## Amitryptiline
Amitryptiline (systematic taxonomy: 3-(10,11-dihydro-5H-dibenzo[[a, d]]cycloheptene-5-ylidene)-N, Ndimethyl-1-propanamine) is a tricyclic antidepressant. Besides its known clinical use it has been identified as an acid sphingomyelinase inhibitor with lowering ceramide levels and thus carrying out anti-apoptotic properties [bib_ref] Ceramide accumulation mediates inflammation, cell death and infection susceptibility in cystic fibrosis, Teichgraber [/bib_ref] [bib_ref] Fas/CD95/Apo-I activates the acidic sphingomyelinase via caspases, Brenner [/bib_ref].
## Cell viability
The viability of cardiomyocytes in tissue samples was assessed by trypan blue exclusion before each experiment. Only samples consisting of ≥ 99% viable cardiomyocytes were further processed in the experiments of this study.
## Microperfusion chamber
Our self developed, previously described [bib_ref] The nonselective beta-blocker carvedilol suppresses apoptosis in human cardiac tissue: a pilot..., Usta [/bib_ref] [bib_ref] Suppressing apoptosis with milrinone simulating extracorporeal circulation: a pilot study, Usta [/bib_ref] [bib_ref] Human cardiac tissue in a microperfusion chamber simulating extracorporeal circulationischemia and apoptosis..., Usta [/bib_ref] microperfusion chamber was modified to investigate larger specimens. It consisted of two components [fig_ref] Figure 1: Microperfusion chamber [/fig_ref]. The first component a temperature-controlled plexiglas block contained a rectangular cavity forming the chamber with following dimensions (length × width × height, 5.5 × 1.5 × 1.25 cm). The second component was mounted over the first, and consisted of another plexiglas block forming the ceiling of the chamber. In this chamber nylon net with a pore size of 400 μm was mounted diagonally. To enable perfusion of the chamber, a thin pipe was introduced at one end of the plexiglas component, entered the chamber and exited at the other end. A thin rubber layer between each component sealed the microperfusion chamber. The biopsy was fixed physically at the nylon net by the laminar flow (perfusion velocity of 5 ml/min) of the hydrostatic perfusion system through the chamber.
## Experimental groups
The protocol was designed to simulate clinical routine procedures administering cardioplegic solution with the same K + concentration (16.5 mM) and temperature (4°C). Five different groups (I -V) were arranged as follows: I (untreated control, n = 10), II (treated control cp/rep, n = 10), III (cp/rep + ceramide, n = 10), IV (cp/ rep + amitryptiline, n = 10) and V (cp/rep + ceramide + amitryptiline, n = 10). In group III cardiomyocytes were continuously treated with 50 μM ceramid. In In group IV cardiomyocytes were continuously treated with 100 μM amitryptiline. In contrast to that in group V cardiomyocytes were continuously treated with both drugs ceramid [50 μM] and amitryptiline [100 μM]. In general, each assay was carried out with the specimens of one patient, i.e. specimens of patients were analysed separately.
## Ischemia/reperfusion assay
The cardiac specimens in the microperfusion chambers were initially equilibrated with KH for 5 min (32°C and continuously bubble-oxygenated with carbogen (95% O 2 and 5% CO 2 ) to attain a PO 2 of 25-30 kPa and pH 7.4. After that the cardioplegic solution (4°C) was administered for 5 min. To induce ischemic injury during the cardioplegia period the perfusion of the microperfusion chamber was stopped and the oxygen supply was discontinued. The cardiac specimens were subjected to various periods of cardioplegia (30, 60 or 120 min) followed by 1/3 of the chosen cardioplegia time as reperfusion (10, 20 or 40 min), as in our surgical routine. For reperfusion 35°C KH was used. Finally, the cardiac specimens were snap-frozen in liquid nitrogen.
## Immunohistochemical apoptosis detection
The slides with the cryosections of the samples (10 μm) were processed prior to the staining according to the manufacturer's recommendation (Epitomics, Inc., Burlingame, CA, USA). The described chemicals were purchased from Biochrom, Berlin Germany. In brief, the cryosections were immersed into the staining dish containing the antigen retrieval solution: 9 ml of stock solution A (0.1 M citric acid solution) and 41 ml of stock solution B (0.1 M sodium citrate solution) were added to 450 ml of destillated H 2 O and adjusted to pH 6.0. After warming for 30 min in a rice cooker and cooling down the slides were washed with TBST (Tris-Buffered Saline and 0.1% Tween 20) for 5 min on a shaker. For the inactivation of endogenous peroxidases the slides were covered with 3% hydrogen peroxide for 10 min and later washed with TBST. After that the slides were immersed into the blocking solution (PBS (Dulbecco's Phosphate Buffered Salts) and 10% bovine serum albumin) for 1 hour.
Later the cryosections were incubated overnight in a humidified chamber (4°C) with antibodies against PARP-1 (Anti-Poly-(ADP-Ribose)-Polymerase)-cleavage (Epitomics, Inc.). PARP is a zinc-dependent DNA binding protein that recognizes DNA strand breaks and is presumed to play a role in DNA repair. PARP is cleaved in vivo by caspase-3 [bib_ref] Yama/CPP32 beta, a mammalian homolog of CED-3, is a CrmAinhibitable protease that..., Tewari [/bib_ref]. The antibody only recognizes p25 cleaved-form of PARP-1.
On the other hand cryosections were stained with antibodies against activated Caspase-3 (Epitomics, Inc.), also. Caspases are a family of cytosolic aspartate-specific cysteine proteases involved in the initiation and execution of apoptosis. Caspase-3 (apopain, SCA-1, Yama and CPP32) is a member of the apoptosis execution functional group of caspases, and is either partially or totally responsible for the proteolytic cleavage of many key proteins during apoptosis. Caspase-3 is a cytosolic protein found in cells as an inactive 35 kDa proenzyme. It is Later for detection to each section secondary HRPconjugated anti-rabbit antibody (Epitomics, Inc.) diluted in the blocking solution per manufacturer's recommendation was applied and incubated for 1 hour at room temperature.
## Fluorescence microscopy
The number of cells on the cryosections was determined by counting the nuclei of cardiomyocytes after staining with DAPI (4',6-Diamidino-2-phenylindole 2 HCl), a dye known to form fluorescent complexes with natural double-stranded DNA, under a fluorescence microscope (Zeiss, Jena, Germany). In each analysis three different areas of the cryosections were counted using 40-fold magnification. Apoptotic cells were identified by condensation and fragmentation of the nuclei and fluorescent conglomerates in the cytoplasm. They were quantified by counting a total of 200 nuclei from each cryosection and calculating the percentage of apoptotic nuclei. After DAPI counterstaining the greater nuclei of cardiomyocytes allow their distinction from fibroblasts with smaller nuclei. In anti-activated caspase-3 positive, apoptotic cardiomyocytes the cytoplasm reveales an intensive granular fluorescence . In contrast to that PARP-1 cleavage positive, apoptotic cardiomyocytes nuclei feature an intensive granular fluorescence intensity with granular staining of the nucleus.
Fluorescence images (blue) of DAPI loaded cardiac specimens were obtained at an excitation wavelength of 360 nm, with an emission wavelength of 460 nm. DAPI was purchased from Sigma-Aldrich, Germany.
# Statistical analysis
Analysis of calcium recordings and graphics were obtained using Sigma Plot software (version 9.0, SPSS Inc., Chicago, IL). Data are expressed as the mean± standard error of deviation (SD) and statistical analysis was performed using GraphPad Prism (version 5.0, GraphPad Software, Inc., CA, USA). Comparison of groups was performed using repeated measures one-way ANOVA followed by Tukey's HSD post hoc test. A p value of less than 0.05 was considered to indicate a statistically significant difference.
# Results
## Immunohistochemical apoptosis detection
Anti-activated-caspase-3
Cardiomyocytes in the untreated group I revealed a significant (p < 0.05) low percentage of apoptotic cells (12 ± 5%) in comparison to the treated control group II [fig_ref] Figure 3: Demonstrating the effect of ceramide and amitryptline on apoptosis in human cardiomyocytes [/fig_ref]. There was a significant (p < 0.05) lower percentage of apoptotic cells in the amitryptiline treatment group IV if compared to group III with ceramide [fig_ref] Figure 3: Demonstrating the effect of ceramide and amitryptline on apoptosis in human cardiomyocytes [/fig_ref].
## Parp-1 cleavage
Cardiomyocytes in the untreated group I featured a significant (p < 0.05) low percentage of apoptotic cells (12 ± 4%) in comparison to the treated control group II [fig_ref] Figure 3: Demonstrating the effect of ceramide and amitryptline on apoptosis in human cardiomyocytes [/fig_ref]. There was a significant (p < 0.05) lower percentage of apoptotic cells in the amitryptiline treatment group IV if compared to group III with ceramide [fig_ref] Figure 3: Demonstrating the effect of ceramide and amitryptline on apoptosis in human cardiomyocytes [/fig_ref].
# Discussion
In the present study our first goal was to apply ceramide to evaluate the proapoptotic potential during cardioplegia and reperfusion [bib_ref] Functions of ceramide in coordinating cellular responses to stress, Hannun [/bib_ref] [bib_ref] Ceramide is involved in triggering of cardiomyocyte apoptosis induced by ischemia and..., Bielawska [/bib_ref] in an ex vivo setting with human cardiomyocytes which to our current knowledge has not been reported yet. Our second goal was to investigate if the proapoptotic effect of ceramide could be inhibited by amitryptiline [bib_ref] Physiology of apoptosis, Gulbins [/bib_ref]. Our third goal was just in accordance to our clinical routine to administer cardioplegia and reperfusion to simulate the extracorporeal circulation in our experimental model and evaluate if the induction or inhibition of apoptosis could be influenced.
In our experimental model human cardiomyocytes were kept in their natural environment as intact cardiac tissue. Otherwise human papillary muscle could be employed but obtaining it before cardioplegic arrest is not an imaginable and feasible option during Representative fluorescent image of cardiomyocytes treated with ceramide during cardioplegia (60 min) and reperfusion (20 min) (group III). After DAPI counterstaining the greater nuclei of cardiomyocytes allow their distinction from fibroblasts with smaller nuclei. In anti-activated caspase-3 positive, apoptotic cardiomyocytes the cytoplasm reveales an intensive granular fluorescence (marked with stars). The exemplary images represent a single experiment. During the cryosection procedure artifacts presenting as nuclei conglomerates could not be avoided; these were excluded from analyses. clinical routine. The simulation of ischemia in isolated cardiomyocyte models can provide important insights into the pathophysiology of myocardial ischemic injury and its underlying molecular mechanisms as was the subject in previous studies in isolated mammalian cardiomyocytes [bib_ref] Differential effects of amrinone and milrinone upon myocardial inflammatory signaling, Chanani [/bib_ref] , isolated papillary muscle preparations [bib_ref] Effects of epinephrine and phosphodiesterase III inhibitors on bupivacaineinduced myocardial depression in..., Azuma [/bib_ref] or animal heart models [bib_ref] Effects of pimobendan and EGIS 9377, cardiotonic agents, and OG-VI, a nucleoside-nucleotide..., Fukutomi [/bib_ref]. The distinctive difference of our experimental assay was utilizing human atrial cardiac tissue as a model for apoptosis studies inducing apoptotis just in accordance to our clinical routine with cardioplegia and reperfusion without induction of ischemia with N 2 perfusion like in previous studies [bib_ref] Cardiotrophin-1 protects the human myocardium from ischemic injury. Comparison with the first..., Ghosh [/bib_ref] [bib_ref] Reperfusion, not simulated ischemia, initiates intrinsic apoptosis injury in chick cardiomyocytes, Vanden Hoek [/bib_ref]. Like presented above in our experimental assay the cardioplegia and reperfusion stimulus proved to be an adequate stimulus for apoptosis induction and is comparable with those in the literature [bib_ref] The nonselective beta-blocker carvedilol suppresses apoptosis in human cardiac tissue: a pilot..., Usta [/bib_ref] [bib_ref] Suppressing apoptosis with milrinone simulating extracorporeal circulation: a pilot study, Usta [/bib_ref] [bib_ref] Human cardiac tissue in a microperfusion chamber simulating extracorporeal circulationischemia and apoptosis..., Usta [/bib_ref] [bib_ref] Ca2+ dysregulation induces mitochondrial depolarization and apoptosis: role of Na+/Ca2+ exchanger and..., Miyamoto [/bib_ref].
Further we wanted to enlighten the major mediators of apoptosis occurring during postischemic reperfusion. Apoptosis is an important mechanism of active cellular death that is distinct from necrosis and has been implicated in the pathogenesis of a variety of degenerative and ischemic human diseases [bib_ref] Functional consequences of caspase activation in cardiac myocytes, Communal [/bib_ref]. The family of caspases is key mediator of apoptosis. An extrinsic pathway involving cell surface death receptors [bib_ref] Death receptors: signaling and modulation, Ashkenazi [/bib_ref] and an intrinsic pathway with intracellular and extracellular death signals which are transmitted to the mitochondria through members of the Bcl-2 family [bib_ref] Hypoxia and acidosis activate cardiac myocyte death through the Bcl-2 family protein..., Kubasiak [/bib_ref] exist. Several intracellular stimuli, including oxidative stress, translocate Bax and/or Bak to the mitochondria, leading to dysfunction of this organelle, the release of proapoptotic proteins, and the activation of caspase-9 [bib_ref] Cell death: critical control points, Danial [/bib_ref]. Another important stimulus for apoptosis derive from sphingolipids like ceramides which have been described as second messengers for several events like differentiation, senescence, proliferation and cell death in different cell lines [bib_ref] Functions of ceramide in coordinating cellular responses to stress, Hannun [/bib_ref]. Sphingolipids are found in most subcellular membranes. In the plasma membrane they are predominantly found in the outer leaflet [bib_ref] Intracellular transport and metabolism of sphingomyelin, Koval [/bib_ref]. The metabolism of sphingolipids has been proved to be a dynamic process and their metabolites (such as ceramide, sphingosine, and sphingosine 1-phosphate (S1P)) are now recognized as messengers playing essential roles in cell growth, survival, as well as cell death [bib_ref] Functions of ceramide in coordinating cellular responses to stress, Hannun [/bib_ref] [bib_ref] Sphingolipids: second messengers, mediators and raft constituents in signaling, Prieschl [/bib_ref]. Sphingomyelin (SM) is a ubiquitous component of animal cell membranes, where it is by far the most abundant sphingolipid. Ceramide can be formed through sphingomyelinases (SMase)-dependent catabolism of SM and by de novo synthesis. SMases are specialized enzymes with phospholipase C activity that can hydrolyze the phosphodiester bond of SM. It is well known that ceramide can modulate many different cellular processes. Ceramide directly regulates protein phosphatase 1 (PP1), inducing dephosphorylation of SR proteins and splicing of caspase-9 and Bcl-x genes [bib_ref] De novo ceramide regulates the alternative splicing of caspase 9 and Bcl-x..., Chalfant [/bib_ref]. Interaction of ceramide with protein kinase-c can inhibit translocation of the kinase to the plasma membrane and therefore inhibits its catalytic activity. Finally the intrinsic and extrinsic pathways of apoptosis induction converge and lead to the activation of caspases which have been characterized as major executioners of apoptosis [bib_ref] Caspases and caspase inhibitors, Villa [/bib_ref]. During oxidative stress reactive oxygen species trigger the release of cytochrome c from mitochondria and, subsequently, caspase activation. Active caspases promote cellular demolition by activating other destructive enzymes, such as DNAses, and by directly targeting key structural proteins, such as lamin and actin, and regulatory proteins, thus leading to chromatin margination, DNA fragmentation, nuclear condensation and collapse [bib_ref] Caspases and caspase inhibitors, Villa [/bib_ref] , which we could demonstrate in our immunohistochemical assays.
In our experiments, we found that caspase-3 was already activated at the end of the ischemia, thus suggesting that the mitochondrial pathway of apoptosis is a very early event in myocardial injury. Caspase-3 has been shown to cleave the 112 kDa nuclear protein PARP into an 85 kDa apoptotic fragment [bib_ref] ICE-LAP6, a novel member of the ICE/Ced-3 gene family, is activated by..., Duan [/bib_ref] , and this cleavage by caspase-3 has been shown to be necessary for apoptosis [bib_ref] Yama/CPP32 beta, a mammalian homolog of CED-3, is a CrmAinhibitable protease that..., Tewari [/bib_ref]. In this regard, the nuclear presence of proteolytic fragments of PARP has been considered a hallmark of an apoptotic cell. However, the role of PARP-1 in apoptosis remains to be determined because conflicting data have been reported. Some investigators have shown that neurons or hepatocytes from PARPdeficient mice do not exhibit any altered sensitivity to apoptotic stimuli, whereas others have demonstrated that pharmacological or genetic inhibition may increase apoptosis in cells subjected to alkylating agents [bib_ref] Poly (ADP-ribose) polymerase, nitric oxide and cell death, Pieper [/bib_ref] [bib_ref] Importance of poly(ADP-ribose) polymerase and its cleavage in apoptosis. Lesson from an..., Oliver [/bib_ref]. The family of Bcl-2-related proteins constitutes the most relevant class of apoptotic regulators and, more specifically, the ratio of anti-or pro-apoptotic proteins determines whether the cell will survive or die [bib_ref] Cell metabolism in the regulation of programmed cell death, Plas [/bib_ref] [bib_ref] The proto-oncogene Bcl-2 and its role in regulating apoptosis, Kroemer [/bib_ref]. On the other hand, expression of Bcl-2 protein prevents the induction of apoptosis caused by a variety of oxidative stresses, and it can influence the level of caspase activation [bib_ref] Cell metabolism in the regulation of programmed cell death, Plas [/bib_ref] In accordance to this referred data in our presented study we could demonstrate that apoptosis can be suppressed effectively in our experimental setup. Considering our immunohistochemical apoptosis detection there is a significant reduction of apoptosis in cardiomyocytes treated with amitryptiline in contrast to the treatment with ceramide after cardioplegia and succeeding reperfusion. The high apoptosis rate in the treated control group especially after 120 min cardioplegia and 40 min reperfusion should not be extrapolated into the in vivo situation without any caution as atrial and ventricular myocardium possess specific characteristics that may influence the susceptibility to ischaemia/reperfusion injury. One explanation is the reported difference in the distribution of potassium channels [bib_ref] Ravens U: Differences between outward currents of human atrial and subepicardial ventricular..., Amos [/bib_ref] , which contributes to the characteristic differences between atrial and ventricular action potentials and may determine a different response to cardioplegia/reperfusion.
Our presented data provide evidence that one of the key signaling pathways controlling apoptosis could mediate, at least in part, ischemia-reperfusion induced injury. Furthermore, the results of our study suggest that, although proapoptotic signalling plays an important role in the development of reperfusion-induced damage, acid sphingomyelinase inhibition by amitryptiline aside from dose-dependency may not afford alone a complete protection against postischemic damage. This characteristic has been described in previous studies [bib_ref] Fas/CD95/Apo-I activates the acidic sphingomyelinase via caspases, Brenner [/bib_ref] and could be an explanation for the partial inhibition of apoptosis due to the treatment with amitryptiline like presented in this study.
# Limitations
The present study has few potential limitations. First, clinical ischemia might be quite different from the simulated ischemia we use. Unfortunately, there is currently no accepted standard that constitutes a clinically relevant "simulated ischemic exposure" for cells. Simulating the ischemic environment of the extracellular fluid that bathes the cells is quite complex due to the fact that there are alterations in many factors, simulating all of these events is not currently possible. So, whereas the use of simulated ischemia is not perfect, we believe it recreates a number of the important components of clinical ischemia. Further in this study only a single ceramid and amitryptiline concentration was employed, but analogous to previous studies in a pharmacological relevant concentration [bib_ref] Acute exposure of ceramide enhances cardiac contractile function in isolated ventricular myocytes, Relling [/bib_ref]. Therefore, detailed doseresponse relationships of neither ceramide nor amitryptiline on apoptotic events were not investigated. Nevertheless, with the concentration employed in this study, apoptotic events could be triggered or inhibited considerably. Furthermore the primary purpose of this study was to test its effect on apoptotic events in cardiomyocytes in this new experimental setting rather than to study doseresponse relationships. Our next step would be to verify our current findings in an animal model. However our results indicate a definite beneficial effect of amitryptiline on apoptotic events.
# Conclusions
In human cardiomyocytes there is a remarkable induction of apoptosis due to the pro-apoptotic second messenger ceramide.
The treatment of human cardiomyocytes in an ex vivo experimental setting with simulated cardioplegia and reperfusion can result in considerable reduction of apoptotic events by adding amitryptiline. These findings warrant further studies in order to evaluate potentially beneficial effects of acid sphingomyelinase inhibition by amitryptiline in the in vivo setting of cardioplegia as employed in cardiac surgery.
[fig] Figure 1: Microperfusion chamber. The perfusate enters the chamber, constructed from plexiglas (2), through the pipe (1) and fills the rectangular shaped chamber(3). Once laminar flow is constituted the cardiac tissue is physically fixed before the nylon net (not featured), which spans in a 135°angle. The fluid exits on the opposite side(4). Between the bottom and the upper part of the chamber a rubber layer was placed for sealing and fastened with 4 screws. [/fig]
[fig] Figure 3: Demonstrating the effect of ceramide and amitryptline on apoptosis in human cardiomyocytes. Percentage of anti-activated caspase-3 (3A) and anti-PARP-1 cleavage (3B) positive cardiomyocytes. In the treated control group the time-dependent increase of apoptotic cardiomyoctes is significant (p < 0.05) if compared to the untreated control group. Ceramide had a higher impact on apoptosis if compared to the treated control group. Amitryptiline applied together with ceramide suppressed the proapoptotic effect of ceramide significantly (p < 0.05) (*). Results shown represent mean±SD of combined results from n = 10 independent assays. [/fig]
|
A pretraining domain decomposition method using artificial neural networks to solve elliptic PDE boundary value problems
A pretraining domain decomposition method using artificial neural networks to solve elliptic PDE boundary value problemsJeong-Kweon SeoDeveloping methods of domain decomposition (DDM) has been widely studied in the field of numerical computation to estimate solutions of partial differential equations (PDEs). Several case studies have also reported that it is feasible to use the domain decomposition approach for the application of artificial neural networks (ANNs) to solve PDEs. In this study, we devised a pretraining scheme called smoothing with a basis reconstruction process on the structure of ANNs and then implemented the classic concept of DDM. The pretraining process that is engaged at the beginning of the training epochs can make the approximation basis become well-posed on the domain so that the quality of the estimated solution is enhanced. We report that such a well-organized pretraining scheme may affect any NN-based PDE solvers as we can speed up the approximation, improve the solution's smoothness, and so on. Numerical experiments were performed to verify the effectiveness of the proposed DDM method on ANN for estimating solutions of PDEs. Results revealed that this method could be used as a tool for tasks in general machine learning.Partial differential equations (PDEs) contain principled physical laws and govern the object's physical or engineering systems. For instance, PDEs are mathematically analyzed to understand scientific and physical phenomena such as heat, electricity, fluid dynamics, and so on. However, due to the analytic complexity of a given PDE, it is usually impossible to have their explicit solutions. In correspondence, in the science of numerical solutions of PDEs, many studies have been conducted to efficiently solve problems by applying various discretization methods such as the finite difference method 1 , finite element method 2 , and other high-order methods such as the spectral method 3 . Eventually, as the era of the artificial intelligence approaches 4 with their applications, for example, to cognitive science 5 , biomedical science 6,7 , ecology 8,9 , manufacturing 10 , and more, many numerical methodologies of computational physics based on machine learning techniques have also been actively studied, including radial basis function networks 11 , Gaussian Kernels 12 , least squares support vector machine 13 , dimension reduction methods 14-16 , deep learning with Monte Carlo method 17 , autoencoder 18 , and convolutional neural networks 19 .On the other hand, Li et al.20have briefly summarized the use of machine learning into three categories: (1) implementing physics into machine learning by generating a dataset of physical laws 21 , (2) designing a physicsguided machine learning method with physics-guided model as a black box 22 , and (3) imposing physics into the training process, such as the method of "physics-informed neural networks (PINN)" 23 . Among these various types of machine learning technologies, this paper will focus on artificial neural networks (ANN) of the third category[24][25][26]. Unlike most of other state-of-the-art machine learning techniques that lack robustness or fail to guarantee convergence with small data regimes, the method of PINN encodes physical information into the training process and well generalizes to the right solution even when only a few training examples are available 23 .Regarding the development of ANN-based PDE solvers, Lagaris et al.27,28have made early contributions to the application of ANN towards solving systems of PDEs. Using analogues of their model, subsequent research has been performed on empirical methodologies to use ANNs for the calculus of PDEs, including irregular boundary problems, arbitrary boundary problems 29,30 , initial-boundary value problems 31 , time-constrained optimal
control problems [bib_ref] Fixed-final-time-constrained optimal control of nonlinear systems using neural network HJB approach, Cheng [/bib_ref] , the methods of unsupervised learning [bib_ref] Multilayer perceptron neural networks with novel unsupervised training method for numerical solution..., Shirvany [/bib_ref] , immersed boundary problem [bib_ref] An immersed boundary neural network for solving elliptic equations with singular forces..., Balam [/bib_ref] , and a constrained backpropagation approach [bib_ref] A constrained backpropagation approach for the adaptive solution of partial differential equations, Rudd [/bib_ref].
In 2020, devising a domain decomposition methodology (DDM), Jagtap et al. [bib_ref] Conservative physics-informed neural networks on discrete domains for conservation laws: Applications to..., Jagtap [/bib_ref] proposed conservative physics-informed neural networks to solve Burgers' equations, incompressible Navier-Stokes equations, and compressible Euler equations. Among many variations of ANN models, in Refs. [bib_ref] Conservative physics-informed neural networks on discrete domains for conservation laws: Applications to..., Jagtap [/bib_ref] , the authors implemented PINN-based domain decomposition methodologies and showed their approximations' effectiveness by hprefinement. They divided their computational domain into several subdomains and constructed the main loss function by employing weight parameters on terms of boundary/initial conditions, the residual from governing equations and interfaces, and applied their method to forward and inverse problems of several challenging nonlinear PDEs. Since hp-refinement technologies have been one of the main concerns for research performing numerical analysis of PDEs [bib_ref] Domain decomposition algorithms for indefinite elliptic problems, Cai [/bib_ref] , their breakthroughs employing the ANN-based decomposition of the domain could not be regarded as anything other than impressive.
Solving PDE by ANN models, many studies have developed efficient technologies as listed in above. However, there is no model that employs a concept of pretraining technique. As the main contribution of our proposed method in our work, we proposed a pretraining scheme to improve the performance of a given model of ANN to solve elliptic PDEs.
In this study, as another approach for an ANN-based DDM, we developed a pretraining scheme DDM. The pretraining process that is engaged in at the beginning of the training epochs helps the approximation basis get well-posed on the domain so that the accuracy of the estimated solution is enhanced, i.e., the speed of approximation is improved rather than non-pretrained schemes. We report that such a well-organized pretraining scheme is effective as an NN-based PDE solver. In our work, we did not focus on handling disadvantages of existing methods. Instead, we focused on developing an ad-hoc technology by proposing a pretraining scheme so that it could be applied to any given existing ANN models to enhance their numerical performances.
The proposed pretraining scheme consists of two parts. The first part is to train along with governing equations without any engagement of the boundary or subdomain's interfacial conditions. The second part is to train for neighboring subdomains to be flattened, i.e., a subdomain's function of the estimation is compelled to be zero at data points located on neighboring subdomains. In this article, we call the first part the smoothing process and the second part the basis reconstruction (BR) process. With numerical experiments, we show that as the proposed pretraining process continues, the estimated solution's output of the fine-tuning improves. Moreover, we report that the BR process helps the numerical basis of the ANN become well-posed to approximate sorts of high degrees of polynomials that are in solutions as the BR process makes the basis of ANN of a subdomain closely gather on its region such that the initial state of the ANN, at the starting point of the fine-tuning, has high smoothness.
Other than existing methodologies of the NN-based domain decomposition method that employ hyperparameters within the loss function, our proposed method consists of a scheme of pretraining and fine-tuning approach that allows us to use the scheme in various architectures of deep-learning algorithms. As the main contribution of our proposed method, the pretraining process that is engaged at the beginning of the training epochs helps the approximation basis being well-posed on the domain so that the quality of the estimated solution is enhanced. We report that such well-organized pretraining scheme may give effectiveness to any NN-based PDE solver as we get a speed-up in approximation and improvement in the solution's smoothness and so on.
In "Methods" section, we introduce our domain decomposition method for ANNs. In "Numerical experiments" section, we present numerical results of using our method to solve elliptic boundary value PDEs and the XOR problem as a demonstration of the method's performance for application to the general machine learning task. We then discuss properties of our proposed method. In "Conclusion" section, we present some conclusions drawn from our study.
# Methods
The feedforward multi-layered neural networks and the system of PDEs. Let the architecture of the multi-layered ANN (abbreviated as ANN thereafter) be given as follows:
where v ∈ R H and w ∈ R H×(d+1) denote learnable weights, p represents them as general parameters, the input variable is denoted by x , an extension of x such that x = (x, b) T with a constant b for the computation of the bias weight in w , and " σ : R H → R H " is a component-wise activation function. In our study, we employed the sigmoidal activation function. The d-dimensional problem structured by N x, p is illustrated in [fig_ref] Figure 1: d-Dimensional ANN structure [/fig_ref]. www.nature.com/scientificreports/ Following 28 , we considered the following system of PDEs: where A(x) satisfies the boundary condition such that �(x) = A(x), ∀x ∈ ∂� , and N x, p is the single-hiddenlayered ANN. The scalar function B(x) is constructed to satisfy B(x)N x, p = 0, ∀x ∈ ∂� D , where ∂� D is the portion of ∂� where the essential boundary condition is imposed. For details of techniques on how to impose constraints, please refer to Refs. [bib_ref] Artificial neural networks for solving ordinary and partial differential equations, Lagaris [/bib_ref] [bib_ref] Neural-network methods for boundary value problems with irregular boundaries, Lagaris [/bib_ref].
[formula] (1) N x, p = v T σ (wx),where �(x) , with x = (x(1), · · · , x(d)) ∈ R d , [/formula]
To estimate the trial solution t x, p solving the system of Eq. (2), we defined discretization of such that � = ξ j ∈ �; j = 1, . . . , m . Then Eq. (2) is in the form of the collocation method 3 as follows:
To train weight parameters of the minimization problem in Eq. (4), we applied the backpropagation method used in gradient descent [bib_ref] Artificial neural network methods in quantum mechanics, Lagaris [/bib_ref] [bib_ref] Artificial neural networks for solving ordinary and partial differential equations, Lagaris [/bib_ref] [bib_ref] A constrained backpropagation approach for the adaptive solution of partial differential equations, Rudd [/bib_ref].
## Implementation of the gradients and training.
To compute the gradient of � t ξ , p , ξ ∈ � , we discretized the 1st-and 2nd-order derivatives by the backward and midpoint Euler scheme, respectively, such that where for a positive real number δ > 0 , ξ i,−δ and ξ i,+δ were defined as varying the i th component of ξ by ±δ , i.e., ξ i,−δ = (ξ(1), ξ(2), . . . , ξ(i) − δ, . . . , ξ(d)) , ξ i,+δ = (ξ (1), ξ (2), . . . , ξ (i) + δ, . . . , ξ(d)) . Note that, for the implementation of numerical experiments in this paper, we set δ = 10 −6 . Using the numerical approximation theory of derivatives, one may set it as sufficiently small.
As for our example problem of Eq. (2), the Poisson equation is given as follows:
where F(x) is the source function.
To estimate the solution of Eq. (6), we discretized the minimization problem of Eq. (6) as follows:
Now, by employing the gradient descent method with a suitable learning rate, we obtained the iterative solution of Eq. (7). For weight parameters w and v of p , using Eqs. (5) and (7), their n th epoch's updates of gradients �w (n) and �v (n) are given by the chain rule of the derivatives of Eq. (7) for w (n) and v (n) , respectively, as follows:
where σ ′ k is the derivative of the k th component of σ.
The pretraining domain decomposition method for ANNs. In our study, the given computational domain was divided into several non-overlapping smaller computational regions over which the governing PDEs were inherited. We call these small-sized computational domains as subdomains of our domain decomposition method. Now, the core of the DDM scheme we proposed began with the flux-continuity on the interfaces of the subdomains. After that, the construction of the basis related to the solutions on the subdomains was considered. To illustrate the scheme, we defined a pair of consecutive neighborhood subdomains r , s ⊂ where r and
[formula] (2) G x, �(x), ∇�(x), ∇ 2 �(x) = 0, x ∈ �, (3) � t x, p = A(x) + B(x)N x, p , (4) min p 1 2 m j=1 G ξ j , � t ξ j , p , ∇� t ξ j , p , ∇ 2 � t ξ j , p 2 , ξ j ∈ �. (5) ∂� t( ξ ,p) ∂x(i) = � t( ξ ,p)−� t � ξ i,−δ ,p � δ ∂ 2 � t( ξ ,p) ∂x(i) 2 = � t � ξ i,+δ ,p � −2� t( ξ ,p)+� t � ξ i,−δ ,p � δ 2 , (6) −ν∇ 2 �(x) − F(x) = 0, x ∈ � �(x) = 0, [/formula]
x ∈ ∂� , www.nature.com/scientificreports/ s were disjoint except for their interfaces. Trial functions t,r ·, p r and t,s ·, p s were given for r and s , respectively.
[formula] (7) min p 1 2 m j=1 −ν d i=1 ∂ 2 � t ξ j , p ∂x(i) 2 − F(ξ j ) 2 , ξ j ∈ �. (8) �w (n) k,l = − ν δ 2 d i=1 −ν d i=1 ∂ 2 � t ξ j , p (n) ∂x(i) 2 − F ξ j × B ξ i,+δ j v (n) k σ ′ k w (n) ξ i,+δ j ξ i,+δ j (l) − 2B ξ j v (n) k σ ′ k w (n) ξ j ξ j (l) + B ξ i,−δ j v (n) k σ ′ k w (n) ξ i,−δ j ξ i,−δ j (l) , �v (n) k = − ν δ 2 d i=1 −ν d i=1 ∂ 2 � t ξ j , p (n) ∂x(i) 2 − F ξ j × B ξ i,+δ j σ k w (n) ξ i,+δ j − 2B ξ j σ k wξ j + B ξ i,−δ j σ k w (n) ξ i,−δ j , [/formula]
Flux-continuity on interfaces. The training for the interfacial flux-continuity was divided into two parts: the nodal continuity and the continuity of the normal directional derivatives.
(1) Nodal continuity:
(2) Continuity of the outward normal directional derivatives:
For normal directional derivatives in Eq. (10), we applied direct differentiation on � t,r ζ , p r and � t,s ζ , p s for respect to normal directions on their domains. Neglecting the subscript of r or s , we have Hence,
where − → n is the outward directional normal vector on the interface.
## Construction of the basis on subdomains.
As the key subject of our study, a pre-processing approach was taken for the model of multi-subdomains by proposing a pretraining scheme that could initialize training weight parameters well. We aimed to enhance the convergence power. This pretraining can be divided into two parts:
(1) Pretraining for the loss function on each subdomain (smoothing process) In this step, we strengthened the convergence power interior of each subdomain except for terms of the flux-continuity. From this, we performed an early estimation of the smoothness interior before the interfacial flux-continuity exchanges. The pretraining scheme implemented to optimize the problem is given as follows:
where the subscript q denotes the arbitrary index of subdomains. (2) Pretraining to impel the basis into each subdomain (basis reconstruction (BR) process) In this step, we aim to train the network to nullify the basis arising in the neighboring subdomains. We hypothesize that if the function-shaping basis exists outside of the interface, it may weaken the power of the polynomial basis that shapes the interior estimations. Hence, to prevent the basis from being non-zero outside of the interfaces, we performed nullification pretraining such that where the ǫ q was the ǫ-expansion of q for a given positive number ǫ > 0 ∈ R defined by
The definition of ǫ q is illustrated in [fig_ref] Figure 2: A 2D rectangular domain and ǫ q [/fig_ref]. Assuming that a two-dimensional (2D) rectangular domain is given, the gray-shaded region depicts q and the region enclosed by a red curve represents ǫ q .
Note that the above two parts of the pretraining occurred subsequently in each epoch. The main purpose of the pretraining scheme is to enhance computational smoothness of basis functions made of activation functions. As we nullified solutions outside of a given subdomain, we assumed that the smooth region in the basis function moved near the region of the subdomain so that the well-posed basis into the domain might prevent leakage of the computational basis' smoothness off the domain.
The proposed ANN version of DDM is illustrated in [fig_ref] Figure 3: The structure of the pretraining version of ANN-based DDM [/fig_ref].
[formula] (9) � t,r ζ , p r − � t,s ζ , p s = 0, ζ ∈ � r ∩ � s . (10) ∂� t,r ζ , p r ∂ − → n r + ∂� t,s ζ , p s ∂ − → n s = 0, ζ ∈ � r ∩ � s. (11) ∂� t ζ , p ∂x(l) = ∂A(ζ ) ∂x(l) + ∂B(ζ ) ∂x(l) v T σ wζ + B(x) N h k=1 v T k w k,l σ ′ wζ ,(12)∂� t ζ , p ∂ − → n = ∇� t ζ , p · − → n ,(13)min p q 1 2 m q j=1 −ν d i=1 ∂ 2 � t ξ j , p q ∂x(i) 2 − F(ξ j ) 2 , ξ j ∈ � q ,(14)� t,q η, p q = 0, η ∈ �\ � ǫ q ,(15)� ǫ q = η ∈ �\ � q : |η − ξ | ≥ ǫfor allξ ∈ � q . [/formula]
## Numerical experiments
The first numerical experiment we performed was a preliminary study to investigate the pretraining's effect on one-dimensional (1D) problems. We applied the smoothing process as pretraining without basis reconstruction or nullification for the Poisson problem with the Dirichlet boundary condition. Note that in our work, the benchmark algorithm, which is conceptually the same as the one with the algorithm introduced in Ref. [bib_ref] Conservative physics-informed neural networks on discrete domains for conservation laws: Applications to..., Jagtap [/bib_ref] , i.e., a DDM model without pretraining process, was one of the recently proposed algorithms to apply the traditional DDM method calculated by preserving the flux continuity between subdomains to ANN as it is (refer to refs [bib_ref] Conservative physics-informed neural networks on discrete domains for conservation laws: Applications to..., Jagtap [/bib_ref].
## 1d problem: effect of smoothing. the 1d poisson equation with dirichlet boundary condition is:
where the solution u satisfies equations over the domain = [0,2] and the boundary ∂� of is ∂� = �\(0,2) . As a test solution, we let u = u(x) , for x ∈ , such as www.nature.com/scientificreports/ [fig_ref] Figure 4: Solutions with 150, 200, and 300 hidden units from the 1D domain... [/fig_ref] shows plots of the estimated solution made from the ordinary single domain method. They were compared to the exact solution. To compare, we verified the use of hidden units which we set to be 150, 200, and 300. To organize the set of training data, we discretized the domain equi-spatially into 3 × 2 4 nodes. Among estimations, only the case of using 300 hidden units gave a good approximation to the solution. Note that when updating neural weights in the training, we conducted the smallest batch model that updated weights for each input data, like the stochastic batch mode. Now, to assess the computing power effectiveness of the DDM model, we conducted training with a threesubdomain method without any pretraining process. [fig_ref] Figure 5: The solution without pretraining process on the 1D domain is decomposed into... [/fig_ref] shows a plot of these results. Here, hidden units are set to be 20, 30, and 100 on each subdomain. Hence, the total number of the hidden units is 150. As show in [fig_ref] Figure 5: The solution without pretraining process on the 1D domain is decomposed into... [/fig_ref] , preset hidden units of the DDM failed to estimate the solution. Thus, the ordinary scheme of the DDM with ANN did not improve the results further in this case.
[formula] (16) −ν�u = f in � u = 0 on ∂� , (17) u(x) = e x 2 sin(2πx). [/formula]
Next, to improve DDMs by using smoothing for pretraining, we applied it to the same structure of subdomains with the same sets of hidden units. The pretraining used 10,000 epochs. For fine-tuning, the ratio of the training epochs at the boundary, subdomain's interface, and the governing equation was set to be 20:2:1, i.e., the training took place at boundary nodes for 20 epochs, then for 2 epochs at the interface, and then 1 epoch for the governing equation on interior nodes of each subdomain. Here, the number of training epochs at the boundary was set relatively high since we noted that the structure of our ANN was not equipped with an ansatz employed www.nature.com/scientificreports/ to hold the boundary condition. We propose that further training at the boundary will allow estimations to be stable and maintain their analytic uniqueness. The resulting configuration of the estimated solution and the plots of the progression of the loss versus epoch are shown in Figs. 6 and 7, respectively. It was revealed that the model of smoothing well approximated the exact solution in Eq.. However, the single domain method and the ordinary approach of the DDM failed to make an accurate estimate [fig_ref] Figure 6: The solution with the smoothing pretraining process on the 1D domain is... [/fig_ref]. This demonstrates that such highly-oscillatory solutions in Eq., which the ordinary single domain method hardly solves with lower amounts of hidden units, are more effectively solved by the DDM model with its effective alignment of the subdomains' numerical basis delivered from the DDM's inherent domain separation scheme property. Note that, for our numerical results, the learning rate is set to be 10 −6 . From our preliminary tests, the convergence rate deteriorated for some scales of the learning rates much greater than 10 −6 . Even for DDM solutions, in the subregion where oscillation or size change of the solution was large, the convergence rate was significantly lower than that in other regions.
In the 1D problem, we have defined the area of the domain to reach a relatively longer region, i.e., we have defined the domain as . In this case of domain region, we found that the ordinary ANN with some small numbers of hidden units failed to converge to the right solution. For this reason, as a special case of the solution to reveal the effectiveness of DDM applied for the solution type made of ANN structure, ANN solution made of some hundreds of hidden units was used for comparison to show the effectiveness of DDM. In the 1D problem, we found that, even for a hardly learnable solution due to some touch conditions regarding the domain, DDM could technically provide guarantees to reach the right convergence. [fig_ref] Table 1: Settings and relevant parameters of the ANN on each subdomain [/fig_ref]. Note that the initial weight parameters of the ANN were given randomly since we intended to verify prominent and consistent effects of methods even in random situations. The ratio of the training epochs at the boundary, subdomain's interface, and the governing equation was set to be 40:2:2, i.e., the training took place at boundary nodes for 40 epochs, then for 2 epochs at the interface, and then 2 epochs for the governing equation on the interior nodes of each subdomain.
The basis reconstruction scheme on the 2 × 2 square subdomains of our test problem is illustrated in . In [fig_ref] Figure 9: Results of using six hidden units [/fig_ref] , the estimated solutions from the case of 6 hidden units are depicted in the subplots (on the left-hand side) with a decaying process of the loss function (on the right-hand side). The solution estimated on the single domain model with 6 hidden units is denoted as '1-domain' . The ordinary DDM model without pretraining is 'No-pretraining' . The pretraining model of DDM with only smoothing is 'Smoothing' . DDMs of BR methods with ǫ of 0.05, 0.1, 0.2 are indicated as 'BR-0.05' , 'BR-0.1' , and 'BR-0.2' , respectively. For the case of '1-domain' , the solution estimation was incomplete and the loss remained high, suggesting that shortages of the numerical basis in the ANN caused the estimation to fail. For other methods using DDM structures, the loss declined. These methods relaxed the shortages in the approximation basis of the single domain model. Hence, from a computational effectiveness perspective, when one applies parallel processors, DDM models might improve the convergence rate as independent calculators (processors) are distributed to subdomains, which may save training costs per processor as each processor's allocated portion of the training data is reduced as the domain is divided (i.e., the overall order of the basis is strengthened whereas the training load on each processor is relieved).
To demonstrate further, in Figs. 10 and 11, configurations of estimated solutions and the loss along epochs are given for cases with 8 and 10 hidden units, respectively. It was found that the single domain approximation method still struggled to reduce the loss. DDM methods were comparatively more computationally powerful, although the dominance among them was indiscriminate when the hidden layer's size was increased.
As shown in Figs. 9, 10 and 11, the DDM-based lower-order ANN model had a more convergent solution than a higher-order single domain model. When training is performed using a given input data, computational costs for both the DDM-based model and the single domain model would be the same if both models' numbers of hidden nodes are the same. In our work, we focused on the computational efficiency of our proposed DDM technique since single-domain models might have difficulty to obtain the right solutions for some problems while our proposed methods were validated to handle these problems.
Results of comparing different DDM models are presented in [fig_ref] Figure 1: d-Dimensional ANN structure [/fig_ref] , which shows resulting values of the loss at the end of each training with increasing number of hidden units. As shown in [fig_ref] Figure 1: d-Dimensional ANN structure [/fig_ref] , when the size of hidden layers is increased, results are improved. However, the problem is that weight parameters that generate [bib_ref] Learning and meta-learning of stochastic advection-diffusion-reaction system from sparse measurements, Chen [/bib_ref] u(x) = 1 e π − e −π sin(πx 1 ) e πx 2 − e −πx 2 . In the next subsection, we will present that, by pretraining with the BR method, the smoothness of the basis function generated by the hidden layer is improved. Hence, we can interpret that the effect of ǫ can cause the BR method to enrich the order of polynomials which shape the basis.
Comparing polynomial degrees using the H 2 semi-norm. We confirmed effects of using the H 2 semi-norm to compare degrees of polynomials in the basis as a measure of smoothness according to pretraining. Although we did not investigate all of the basis-degree with the measure in the H 2 semi-norm reflecting indirect comparison of the strength of the basis, our numerical results gave us confidence that the proposed pretraining method might help the initial basis reform its polynomials to higher degrees. www.nature.com/scientificreports/ From the structure of the ANN, we know that the solution is a linear combination of the outputs of activated hidden units. Hence, the H 2 semi-norm of functions shaped by the activation of the hidden units, i.e., the measure of the degree of the polynomials of each hidden unit's activation function, can be considered as a measure of the degree of the polynomials of the numerical basis of the ANN. Precisely, we measured the H 2 semi-norm of one hidden unit's output values activated in each subdomain. The sum of the values of the hidden units on their subdomains was then added to make a total measure (see [fig_ref] Table 2: H 2 semi-norm comparison along with the number of hidden units for... [/fig_ref]. In [fig_ref] Table 2: H 2 semi-norm comparison along with the number of hidden units for... [/fig_ref] , the mean of the H 2 semi-norm is [fig_ref] Table 2: H 2 semi-norm comparison along with the number of hidden units for... [/fig_ref] are plotted. [fig_ref] Figure 1: d-Dimensional ANN structure [/fig_ref] (where the x-axis is H) clearly shows that the values of the H 2 semi-norm for the BR method are almost more than those for the smoothing method. In [fig_ref] Figure 1: d-Dimensional ANN structure [/fig_ref] , the x-axis denotes the method of pretraining. Values tended to increase as ǫ decreased. This means that the basis of BR with low values of ǫ has improved its polynomial degrees as a result of the pretraining. Results shown in [fig_ref] Figure 1: d-Dimensional ANN structure [/fig_ref] (where the BR method with ǫ = 0.3 and 0.4 generated a relatively low-quality decay of the loss) could be interpreted that the BR method with a relatively low ǫ could give a benefit to fitting the solution's estimation in higher orders due to enrichment of the basis. Note that the computation of the H 2 semi-norm on the domain is conducted as follows. Let us define the output of the activation of the hidden units as k such that where w k ∈ R H×(d+1) denotes learnable weights defined in Eq. (1) and the subscript k denotes the index of subdomains. The discrete form of the second-order derivative is straightforward as shown in Eq. [bib_ref] A physics-guided neural network framework for elastic plates: Comparison of governing equationsbased..., Li [/bib_ref] : , ξ i,+δ = (ξ (1), ξ (2), · · · , ξ (i) + δ, . . . , ξ(d)). The measure of the H 2 semi-norm of k , denoted | k | 2 , is defined by the following:
[formula] (19) k = σ (w k x),(20)∂ 2 k (ξ ) ∂x(i) 2 = k ξ i,+δ − 2 k (ξ ) + k ξ i,−δ δ 2 , [/formula]
We summed all individual | k | 2 so that we could have k | k | 2 for values listed in [fig_ref] Table 2: H 2 semi-norm comparison along with the number of hidden units for... [/fig_ref].
## Verification of the learning efficiency for general ann tasks by simulation. if the br method
can be effectively applied to the estimation of solutions of PDEs, it stands to reason that it may also be effectively applied to a general task of machine learning with ANNs. To generally prove that the BR method can enhance the performance of ANNs as a tool for machine learning, we tested the BR method on the problem of the differentiation of the exclusive OR (XOR) logic gate, a well-known low-dimensional problem of machine learning on which to apply an ANN. To arrange the problem of XOR in a way that could fit the concept of the BR method, we reorganized the dataset in the computational domain ⊂ R H , where = [0, 1] 2 , as follows (see [fig_ref] Figure 1: d-Dimensional ANN structure [/fig_ref]. The XOR dataset is organized as (0,0), (1/2, 0), (0, 1/2), (1/2, 1/2), with labels of 1, 0, 0, 1, respectively. We defined the domain q as � q = [0, 1/2] 2 which included the dataset (see [fig_ref] Figure 1: d-Dimensional ANN structure [/fig_ref] and denoted q as q = [(0,0), (1/2, 0), (0, 1/2), (1/2, 1/2)] with the set of labels Y = . Note that the definition of ǫ q is given in Eq. [bib_ref] Numerical solutions of Burgers equation by reduced-order modeling based on pseudo-spectral collocation..., Seo [/bib_ref]. Now, we can employ the BR scheme which is illustrated in [fig_ref] Figure 1: d-Dimensional ANN structure [/fig_ref]. Since BR is a pretraining scheme, we determined the efficiency as a function of the number of epochs in the pretraining process. In our experiments, we set the number of pretraining epochs to be 500, 1000, 2000, and 3000. Results are compared with the ordinary training scheme which does not apply any pretraining method. We discretized each subdomain of except q into 21 equi-spatial nodes along each axis so that there were about 400 nodes in total.
In [fig_ref] Figure 1: d-Dimensional ANN structure [/fig_ref] , results of the experiments for BR with varying ǫ are shown for the case with 4 hidden units. The values of ǫ used were 0.05, 0.1, 0.2, 0.3, and 0.4. In the plots, data for 500, 1000, 2000, and 3000 epochs of pretraining are indicated as PT500, PT1000, PT2000, and PT3000, respectively. The ordinary training scheme which had no pretraining process was denoted as PT0. Note that initial values of neural weights were set to be the same for each case since we aimed to verify the progression as we increased epochs of the pretraining of each method. Our results revealed that as the number of epochs of pretraining increased, the convergence of the loss to 0 was accelerated. As the size of ǫ increased, the speed of convergence was further accelerated. However, for the case of PT0, the convergence progress appeared to stop in the middle and the training failed to optimize the loss.
To make a comparison in the case of a larger hidden layer, we repeated the experiments with 200 hidden units and plotted the results for ϵ = 0.05, 0.1, and 0.2 as shown in [fig_ref] Figure 1: d-Dimensional ANN structure [/fig_ref]. Similar to previous results with 4-hidden units, as the number of pretraining epochs increased, the convergence of the loss accelerated, although the difference in the speed of convergence was insignificant [fig_ref] Figure 1: d-Dimensional ANN structure [/fig_ref]. Experimental results shown above suggest that the BR technique applied in the pretraining seems to help the ANN model estimate the solution to the XOR problem. Furthermore, the finding that the convergence speed increased with the number of epochs in the pretraining showed us that the training results might not be caused by a random phenomenon with some vague factors and that the BR method strongly contributed to the effective utilization of the optimization power of the hidden units.
[formula] (21) | k | 2 = H � j=1 δ � x∈� k � d � i=1 ∂ 2 i ∂x(i) 2 � 2 1/2 [/formula]
. [fig_ref] Figure 1: d-Dimensional ANN structure [/fig_ref]. The domain of the XOR problem, data points with labels in red letters (left), and the domain structure to apply the BR method with ǫ (right). www.nature.com/scientificreports/ On the other hand, to demonstrate that the basis reconstruction method improved the approximation power of ANNs, the basis function of each hidden unit was developed into a higher state of the degree of polynomials so that the estimation power of the hidden layer could be improved. The approximated function representing the output function of the XOR logic gate given in [fig_ref] Figure 1: d-Dimensional ANN structure [/fig_ref] shows a wrong approximation of XOR by PT0 on the left and a higher-quality approximation on the right. The solution of the XOR function might have a high order of curvature's smoothness. Thus, the approximating numerical basis functions of the ANN also need to be in the form of some high-degree polynomials. With prerequisite knowledge, we observed shapes of the hidden www.nature.com/scientificreports/
# Conclusion
The estimation's accuracy using NN depends on the solution's analytic complexity. If the solution consists of high order polynomials, it needs to employ high order hidden units in NN because the number of hidden units determines the estimation's order of polynomials. In this study as basic research, we applied the proposed method for a single hidden layered architecture based on the universal approximation theory of the single hidden layered NN [bib_ref] Multilayer feedforward networks are universal approximators, Hornik [/bib_ref]. In our study of the implementation of an ANN-PDE solver, we proposed a pretraining version of ANNbased DDM which used a basis reconstruction technique where weight parameters in the hidden layer were rearranged to enrich the quality of the numerical basis and estimated solutions. The BR method was implemented by setting the output of the ANN to be zero on the neighboring subdomains. The main idea is that, as the ANN is nullified on other subdomains, eventually the core degrees of polynomials constituting the hidden units' numerical basis might be rearranged to center on their main subdomains. Experiments were conducted with one-and two-dimensional Poisson's equations for the XOR problem. When the BR method was applied to PDE problems, it was observed that the BR method enhanced the computing power to accelerate the rate at which poor estimations of an ANN on single domains were improved in the DDM model by reducing the computational costs relative to single-domain models. On the other hand, to numerically demonstrate the basis reconstruction property and benefits for the weight parameter's initialization, we compared the after-pretraining status of the hidden units' activation in the XOR problem. Results showed that degrees of polynomials of activations by the BR method, understood by measuring the curvature, were enriched. This gives a superior approximation compared with ordinary training procedures.
In this work of developing DDM in the structure of ANNs, the basis reconstructing treatment was effective for estimating solutions to PDEs. It could also be applied well to general machine learning tasks. Thus, it may be applicable as a tool for subjects such as regularization, weight initialization, and speeding computations in training.
Nowadays, lots of research have been conducted applying more deeper architectures. For instance, researchers have used ResNet 42 which provides more promising performances than any basic deep neural network (DNN)based architecture due to the skip-connection's property to find optimal number of hidden layers without too laborious tests made for some normal DNN architectures. However, for a deeper architecture of neural network, we have to apply the proposed method and test it to know how effective it will be in future works.
Received: 2 November 2021; Accepted: 9 August 2022
[fig] Figure 1: d-Dimensional ANN structure. Scientific Reports | (2022) 12:13939 | https://doi.org/10.1038/s41598-022-18315-4 [/fig]
[fig] Figure 2: A 2D rectangular domain and ǫ q . [/fig]
[fig] Figure 3: The structure of the pretraining version of ANN-based DDM. Scientific Reports | (2022) 12:13939 | https://doi.org/10.1038/s41598-022-18315-4 [/fig]
[fig] Figure 4: Solutions with 150, 200, and 300 hidden units from the 1D domain with 3 × 2 4 nodes as estimated by the single domain method. [/fig]
[fig] Figure 5: The solution without pretraining process on the 1D domain is decomposed into three subdomains with 2 4 nodes in each subdomain, where the total number of hidden units is 150.Scientific Reports | (2022) 12:13939 | https://doi.org/10.1038/s41598-022-18315-4 [/fig]
[fig] Figure 6: The solution with the smoothing pretraining process on the 1D domain is decomposed into three subdomains with 2 4 nodes for each subdomain, where the total number of hidden units is 150. [/fig]
[fig] Figure 7: A comparison of the decay of the cost function's loss for the training of the solution on the 1D domain decomposed into three subdomains with 2 4 nodes for each subdomain, where the total number of hidden units is 150. : effects of basis reconstruction. In this section, numerical results for 2D elliptic PDEs are presented.As a test problem, we solved the type of Poisson equations given in Eq.(16) with Dirichlet boundary condition for the domain = [0,1] 2 . As our test solution, we defined u = u(x) , for x = (x 1 , x 2 ) ∈ � , as follows:For the structure of the ANN on each subdomain, we set the number of hidden units to be 6, 8, or 10 and applied drop-out 41 of 1, 3, or 5, respectively. As in the case of the 1D problem, we separated the domain and employed the same amount of hidden units and drop-out on subdomains for each test. The parameter ǫ of the BR's nullifying range was set to be 0.05, 0.1, 0.2, 0.3, and 0.4. Settings of the ANN on each subdomain for the experiments are listed in [/fig]
[fig] Figure 9: Results of using six hidden units (left) and a comparison of the decay of the loss (right) for both the 1-domain and 4-subdomain methods (no-pretraining, smoothing, BR-0.05, BR-0.1, BR-0.2). [/fig]
[table] Table 1: Settings and relevant parameters of the ANN on each subdomain.Figure 8. The basis reconstruction scheme with parameter ǫ on a 2 × 2 square subdomain structure. hidden units' values might be poorly posed with only a few hidden layers. This causes the training process to have shortages of numerical bases. Especially for the BR method with ǫ of 0.3 and 0.4, values of the loss were highly unstable. This suggests that the value of ǫ may affect the ANN's ability to estimate accurately. Of course, more reliable evidence of ǫ 's effect needs to be established by performing relevant well-developed experiments. [/table]
[table] Table 2: H 2 semi-norm comparison along with the number of hidden units for several pretraining models. [/table]
|
Rapidly progressing subperiosteal orbital abscess: an unexpected complication of a group-A streptococcal pharyngitis in a healthy young patient
Introduction: Complications associated to group-A streptococcal pharyingitis include non-suppurative complications such as acute rheumatic fever and glomerulonephritis and suppurative complications such as peritonsillar or retropharyngeal abscess, sinusitis, mastoiditis, otitis media, meningitis, brain abscess, or thrombosis of the intracranial venous sinuses. Case presentation: We described a case of a 15-year-old patient with a history of acute pharyngodinia early followed by improvise fever and a progressive formation of a diffuse orbital edema, corneal hyperaemia, diplopia and severe decrease of visual acuity. The patient was surgically treated with functional endoscopic sinus surgery (FESS) after the response of a maxillofacial computed tomography scans that showed a pansinusitis complicated by a left orbital cellulites. Numerous colonies of Streptococcus pyogenes were found in the samples of pus and an antibiotic therapy with meropenem was initiated on the basis of the sensitivity test to antibiotics. The patient was finally discharged with diagnosis of left orbital cellulites with periorbital abscess, endophtalmitis and acute pansinusitis as a consequence of streptococcal pharyngitis. Conclusion: The case highlights the possible unusual complication of a group-A streptococcal pharyingitis in a immunocompetent child and the needing of a prompt surgical and medical approach toward the maxillofacial complications associated to the infection.
# Introduction
Group A beta-haemolitic streptococcus (GAS) being the most common etiology of sore throats caused by bacteria. It has been estimated that GAS is responsible for around 15-30% of cases of acute pharyngitis in children [bib_ref] Group A streptococcal infections in children, Steer [/bib_ref]. Streptococcal pharyngitis is most common in children 5 to 12 and presents with a predominant sore throat and a temperature higher than 38.5°C. Symptoms include fever, chills, myalgias, headaches and nausea. Physical findings may include pharyngeal and tonsillar erithema and exudates and cervical adenopathy [bib_ref] Pharyngitis: a survey of the microbiologic etiology, Van Cauwenberge [/bib_ref]. Sequelae associated to the GAS infection include non-suppurative (or post-streptococcal) complications as rheumatic fever and glomerulonephritis and suppurative complications as cervical lymphadenitis, peritonsillar or retropharyngeal abscess, sinusitis, mastoiditis, otitis media, meningitis, thrombosis of the intracranial venous sinuses, endocarditis, pneumonia, sepsis. In rare cases, necrotizing fascitis, myositis and streptococcal toxic shock syndrome have been described [bib_ref] Management of Group A beta-hemolytic streptococcal pharyngitis, Hayes [/bib_ref].
We report the case of a previously healthy 15-year-old girl affected by pansinusitis and subperiosteal orbital abscess following an acute GAS pharyngitis and successfully treated with functional endoscopic sinus surgery (FESS).
## Report of a case
A 15-year-old girl was admitted to the Department of Ophtalmology (University of Trieste) for high-grade fever (body temperature 40.2°C), severe decrease of visual acuity, left eyelid edema and hyperaemia. Five days before, she had started to complain pharyngodinia, odynophagia, and fever treated with antipyretics by the general physician. Symptoms were followed by progressive pain at the medial left orbital canthus and headache. Her past medical history was unremarkable, and no previous visual compromise was reported. At admission, clinical ophthalmological examination demonstrated a severe edema and erythema of the eyelids with a slight proptosis associated with a conjunctival chemosis and hyperaemia. An ophthalmoplegia was present and the patient referred diplopia and pain on eye movements. Visual acuity was 20\20 in right eye and 20\50 in left eye. In addition, a pharyngeal diffuse erythema with a bilateral anterior cervical lymphadenomegaly was present. A pharmacological therapy with acetaminophen 500 mg per os (when fever was higher than 38°C) and cefaclor 500 mg per os every six hours was administered. Laboratory findings showed normal value of the white cell blood count (9.01 × 10 3 /mm 3 ) with moderate neutrophilia (77%). S-C-reactive protein (CRP) and erythrosedimentation rate were increased with values of 142.8 mg/L (normal value < 5 mg/L) and 92 mm/h (normal value < 10 mm/h), respectively.
CT scans showed an increased thickness of skin and soft tissue overlying the medio-lateral area of the left orbit with a decreased transparency of the fat tissue. At the anteromedial ethmoidal cell area, a circumscribed liquid collection with a convex border directed toward the internal orbit was appreciable. The liquid mass caused a slight deviation of the medial rectum muscle that appeared thickened but regular in morphology. A complete obliteration of all the paranasal cavities with little hydro-air levels in both maxillary sinuses and thinning of the ethmoidal bone plates were observed with a possible bilateral interruption of the papyracea lamina [fig_ref] Figure 1 A: -B [/fig_ref]. After three hours from the admission, the patient was transferred to the Department of Head and Neck Surgical Sciences (University of Trieste) to undergo functional endoscopic sinus surgery (FESS) under general anaesthesia. An extensive osteoplastic procedure of the natural ostium of the left maxillary sinus was performed. During the anteroposterior ethmoidectomy and revision of the left anterior ethmoidal cells, near the nasolacrimal duct, the papyracea lamina was surgically interrupted for a short trait and a large purulent collection was evacuated from the orbita. The latter finding was consisting with a subperiosteal abscess. Samples of purulent fluid from the orbital area and maxillary sinus were collected for microbiological culture, and samples of mucosal tissues were sent for the hystopathologic examination. The patient was given an antibiotic therapy with meropenem 1g three times daily intravenously as indicated by the specialist in infectious diseases. The same day of the surgery, the temperature dropped at 36.6°C. At day 1 post-surgery the patient was afebrile, the ophthalmologic clinical status was improving, and CRP decreased to 112.6 mg/L while erythrosedimentation rate remained unvaried (93 mm/h).
A cerebral angio-magnetic resonance imaging (A-MRI) did not show any intracranial lesion including the left eyeball and the ophthalmic nerve. A diffuse signal alteration was present in the left retro-bulbar adipose tissue associated to a moderate edema and hyperaemia of the extrinsic musculature [fig_ref] Figure 2: Axial contrast-enhanced fat suppressed T1-weighted MR [/fig_ref]. A dental examination was performed to evaluate the possible odontogenic origin of the infection. The patient was not affected by dental or periodontal disease although the vertical percussion of posterior maxillary teeth and the bilateral palpation of the vestibular maxillary region exacerbate a localized pain (VAS= 4). An orthopantomography showed a complete bilateral opacity of the maxillary sinuses but no periapical lesions or any dental involvement [fig_ref] Figure 3: Orthopantomography [/fig_ref].
In the second day after surgery the patient had no fever (36.5°C) and the pharingodinia decreased. Histology of the mucosal samples showed connective tissue characterized by diffuse and conspicuous acute and chronic inflammatory infiltrate. Culture of the orbital and ethmoidal essudates showed numerous colonies of group-A Streptococcus pyogenes which was sensitive to penicillin and cephalosporin derivates, to macrolides, tetracycline, chinupristine/ dalfopristine, vancomycine and clindamycin. An intermediate sensibility was observed for levofloxacine. At day 7 post-surgery, the patient remained afebrile, her left orbital swelling, proptosis and diplopia were disappearing and her visual acuity in her left eye rose to 20\20. The patient was discharged on antibiotic therapy with amoxicillin plus clavulanic acid 1g three times daily, orally for 5 day. After 6 months post-hospital admission, the clinical follow-up of our patient was completely favourable.
# Discussion
The orbital septum divides the preseptal space (soft tissues of the eyelid) from the orbital space (postseptal space) so that periorbital or preseptal cellulitis involve only the lids and not the orbit, whereas orbital or postseptal cellulitis is much more uncommon and involves the soft tissues of the bony orbit. An opthalmological examination is mandatory in assessing proptosis, chemosis, opthalmoplegia or decreased visual acuity as these findings highlight the presence of postseptal orbital cellulitis. However, the distinction beteween preseptal cellulitis and orbital involvement cannot be made with clinical examination alone and delay in treatment can result in blindness in up to 10% of patients. Orbital cellulitis is a serious infection in children and can result in significant complications as blindness, cavernous sinus thrombosis, meninigitis, subdural empyema and brain abscess [bib_ref] Orbital cellulitis in children, Nageswaran [/bib_ref]. In the preantibiotic era, 20% of patients with orbital cellulitis had permanent loss of vision and 17% died for central nervous system complications, today these percentages decreased but have not still been eliminated (15 to 30% of patients develop visual sequelae) [bib_ref] Management of orbital subperiosteal abscess in children, Rahbar [/bib_ref].
Orbital complication accounts for 74 -85% of complications arising from acute sinusitis and usually this is secondary to acute ethmoidal sinusitis since the ethmoid sinus is separate from the orbit only by the papyracea lamina [bib_ref] A ten year retrospective review of orbital complications secondary to acute sinusitis..., Suhaili [/bib_ref]. In a paediatric series, Nageswaran et al. found that 98% and 71% of their patients with orbital cellulitis were affected by ethmoid or maxillary sinusitis respectively [bib_ref] Orbital cellulitis in children, Nageswaran [/bib_ref]. Furthermore bilateral pansinusistis is the most common presentation [bib_ref] Management of orbital subperiosteal abscess in children, Rahbar [/bib_ref]. As in this case, an abscess may be present in the subperiostium of the lateral wall of the lamina papyracea [bib_ref] Periorbital versus orbital cellulitis, Givner [/bib_ref]. It was estimated that the incidence of a subperiosteal abscess in orbital infections is about 15% in children [bib_ref] Management of orbital subperiosteal abscess in children, Rahbar [/bib_ref]. The etiology of orbital cellulitis is usually unknown because blood cultures are often negative. Sinus cultures reveal typical acute sinusitis pathogens including Streptococcus pneumoniae, Haemophylus influenzae, Moraxella catharralis, Streptococcus pyogens, Staphylococcus aureus, αand nonhemolytic streptococci and anaerobic bacteria of the upper respiratory tract [bib_ref] Orbital cellulitis in children, Nageswaran [/bib_ref]. Subperiosteal abscess cultures showed S. pneumoniae, group A streptococci, H. influenzae, as the major pathogens in a previous paediatric series [bib_ref] Subperiosteal orbital abscess in children: diagnosis, microbiology, and management, Skedros [/bib_ref]. Polymicrobial infections are also common and may be more frequent in older versus younger children [bib_ref] Pediatric medial subperiosteal orbital abscess: medical management where possible, Brown [/bib_ref].
The management of an "acute orbit" depends on the cause and severity of the infection. All patients affected by orbital cellulitis should be treated with intravenous antibiotics whereas they should undergo surgical drainage of abscesses and involved sinuses only in presence of large abscess, complete ophtalmoplegia or significant visual impairment, as in the our patient [bib_ref] Periorbital versus orbital cellulitis, Givner [/bib_ref]. Maxillofacial CT scan is indicated to evaluate the extension of the infection and to identify children who are most likely to benefit from surgical intervention [bib_ref] Subperiosteal orbital abscess in children: diagnosis, microbiology, and management, Skedros [/bib_ref]. Because of the aggressive nature of a subperiosteal orbital abscess, we agree with Rahbar et al. in obtaining a CT scan even if the only presentation is preseptal cellulitis [bib_ref] Management of orbital subperiosteal abscess in children, Rahbar [/bib_ref]. In our case, CT images showed a massive involvement of the paranasal cavities including the ethmoidal cells and the subperiosteal abscess of left orbit. Impairment of vision, periorbital erythema and hyperemia, proptosis, together with radiological findings, indicated an immediate surgical approach to avoid the potential loss of vision and the devastating morbidity associated to the subperiosteal orbital abscess [bib_ref] Management of orbital subperiosteal abscess in children, Rahbar [/bib_ref]. The surgical drainage of the abscess was afforded with FESS. This transnasal endoscopic technique provided a quick and safe drainage of the paranasal sinuses, orbita, anterior skull base avoiding facial scars as well as hastening the post-operative recovery period [bib_ref] Functional endoscopic sinus surgery of orbital subperiosteal abscess in children, Deutsch [/bib_ref] [bib_ref] Long-term follow-up for children treated with surgical intervention for chronic rhinosinusitis, Lusk [/bib_ref]. Rahbar et al. found that orbital subperiosteal abscess in children can be successfully and safely managed by a transnasal endoscopic approach in selected patients [bib_ref] Management of orbital subperiosteal abscess in children, Rahbar [/bib_ref]. The choice of method of surgical drainage should be based on the location of the abscess and the experience of the surgeon. Medial and inferior orbital abscesses can be treated with an endoscopic approach while superior localization generally requests an external drainage.
After the emergency surgical treatment, a polispecialistic examination was immediately requested to choose the best antibiotic option, to examinate the dental status, and to exclude intracranial spreading of the infection. Considering this last complication, intravenous therapy with meropenem, a broad-spectrum antibacterial agent of the carbapenem family, was chosen. Meropenem is indicated as empirical therapy prior to the identification of causative organisms, or for disease caused by single or multiple susceptible bacteria in both adults and children with a broad range of serious infections [bib_ref] Meropenem: a review of its use in the treatment of serious bacterial..., Baldwin [/bib_ref].
The dental status was evaluated since the odontogenic nature of the sinusal and orbital infection has to be excluded, especially when the localization is monolateral. Finally, the intracranial spreading of the infection was excluded by A-MRI.
The result of the microbiological analysis (positive for GAS) led to consider a strong correlation between the pharyngitis and the sinusal/orbital involvement and to discharged the patient with the final diagnosis of left orbital cellulitis with periorbital abscess, endophtalmitis and acute pansinusitis as a consequence of GAS pharyngitis.
Group A streptococcal pharyngitis is usually a selflimited disease, and therapy can generally be safely postponed for up to 9 days after the onset of symptoms to prevent the occurrence of major nonsuppurative sequelae. However, according to guidelines of the Infectious Diseases Society of America, early initiation of antibiotic therapy results in faster resolution of signs and symptoms. Furthermore a Cochrane review of randomized, placebo-controlled trials showed that antibiotic therapy significantly reduces the risk of acute otitis media (relative risk 0.30; 95% CI, 0,15-0.58) and peritonsillar abscess (relative risk 0.15; 95% CI, 0.05-0.47) [bib_ref] Antibiotics for sore throat, Mar [/bib_ref]. Thus, antimicrobial therapy is recommended for subjects with symptomatic pharyngitis in the presence of GAS in the throat confirmed by culture or RADT [bib_ref] Practice Guidelines for the diagnosis and management of group A streptococcal pharyngitis, Bisno [/bib_ref]. Clinical scoring systems have been developed to predict the likelihood of streptococcal infection among children and adults with sore throat [bib_ref] Clinical practice. Streptococcal pharyngitis, Wessels [/bib_ref]. Presence of fever (temperature > 38°C), absence of cough, tonsillar swelling or essudate, tender and enlarged anterior cervical lymphnodes are correlated to approximately 30 to 50% probability of positive results of a throat culture or RADT. In this case, despite the clinical presentation, the streptococcal infection was not previously diagnosed. However, the serious complications seen in the patients are usually unexpected in a 15-years immunocompetent girl although it has been recently observed an increase in the incidence of head and neck infections especially associated with acute sinusitis due to group A streptococcal infections in children. This trend may reflect an increase of virulence related to the evolving biology of streptococcal organism and can justify the onset of aggressive and rapidly progressive infections also in previously health children. Much of the increase in invasive dissemination of Streptococcus Pyogenes (noted in last 15 years) has been associated with M protein types M1 and M3 that prevent phagocytosis of the bacteria by inhibiting the interaction with complement.
# Conclusion
In conclusion, our case focused the attention on:
the possible spreading of a streptococcal pharyngeal infection towards the orbital involvement that may require a multispecialty emergency approach; the need of a prompt diagnosis and therapy for streptococcal pharyngitis that might prevent such potentially sight-and life-threatening complication; the surgical management of subperiosteal orbital abscess that can by safely performed with FESS in young patients affected by of a subperiosteal abscess as a consequence of a streptococcal pharyngeal infection.
## Consent
Written informed consent was obtained from the parents of the patient for publication of this case report and any accompanying images. A copy of the written consent is available for reviewer by the Editor-in-Chief of this journal.
[fig] Figure 1 A: -B. Axial (A) and coronal (B) CT unenhanced scan through the middle orbit. The images show: increased thickness of skin and soft tissue overlying the medio-lateral area of the left orbit, decreased transparency of the ipsilateral intraconal and extraconal fat tissue and a subperiosteal abscess along the medial wall of the left orbit (white arrows), adjacent to the opacified ethmoid air cells with resultant lateral displacement of the medial rectus muscle (b) A partial obliteration of sphenoid (A) and maxillary sinuses (B) is also visualized. [/fig]
[fig] Figure 2: Axial contrast-enhanced fat suppressed T1-weighted MR. The image demonstrates ethmoid and sphenoid sinusitis and an heterogeneous enhancement of left eyelid tissues and of orbital fat. The left orbital extraocular muscles show a moderate enhancement due to edema and hyperaemia. Left medial rectus muscle is also moderately thickened. [/fig]
[fig] Figure 3: Orthopantomography. Image shows a bilateral opacity of the maxillary sinuses. [/fig]
|
Cleavage of Dicer Protein by I7 Protease during Vaccinia Virus Infection
Dicer is the key component in the miRNA pathway. Degradation of Dicer protein is facilitated during vaccinia virus (VV) infection. A C-terminal cleaved product of Dicer protein was detected in the presence of MG132 during VV infection. Thus, it is possible that Dicer protein is cleaved by a viral protease followed by proteasome degradation of the cleaved product. There is a potential I7 protease cleavage site in the C-terminus of Dicer protein. Indeed, reduction of Dicer protein was detected when Dicer was co-expressed with I7 protease but not with an I7 protease mutant protein lack of the protease activity. Mutation of the potential I7 cleavage site in the C-terminus of Dicer protein resisted its degradation during VV infection. Furthermore, Dicer protein was reduced dramatically by recombinant VV vI7Li after the induction of I7 protease. If VV could facilitate the degradation of Dicer protein, the process of miRNA should be affected by VV infection. Indeed, accumulation of precursor miR122 was detected after VV infection or I7 protease expression. Reduction of miR122 would result in the suppression of HCV sub-genomic RNA replication, and, in turn, the amount of viral proteins. As expected, significant reduction of HCVNS5A protein was detected after VV infection and I7 protease expression. Therefore, our results suggest that VV could cleave Dicer protein through I7 protease to facilitate Dicer degradation, and in turn, suppress the processing of miRNAs. Effect of Dicer protein on VV replication was also studied. Exogenous expression of Dicer protein suppresses VV replication slightly while knockdown of Dicer protein does not affect VV replication significantly.
# Introduction
MicroRNAs (miRNAs) are a group of small noncoding RNAs about 22 nt in length that downregulate gene expression by either of the two posttranscriptional mechanisms: mRNA cleavage or translational repression [bib_ref] Mechanisms of post-transcriptional regulation by microRNAs: are the answers in sight?, Filipowicz [/bib_ref]. miRNAs have been implicated in a vast array of cellular processes including cell differentiation, proliferation and apoptosis [bib_ref] MicroRNAs: genomics, biogenesis, mechanism, and function, Bartel [/bib_ref]. Links between miRNAs and To clone the plasmid expressing pre-miR122, cDNA library derived from HuH7 cells was used as template and forward and reverse PCR primers (5'GGAATTCGGAGTGTGA CAATGGTGTTTG 3' and 5'GCTCTAGATTTAGTGTGATAATGGCGTTTG 3') were used to amplify the gene fragment. After PCR, the DNA fragment was digested by restriction enzymes (EcoRI/XbaI) and cloned into the expression vector pcDNA3 (linearized by EcoRI/XbaI).
All of the expression plasmids were verified by sequencing.
## Virus infection
Vaccinia virus WR strain [bib_ref] Vaccinia virus 4c (A26L) protein on intracellular mature virus binds to the..., Chiu [/bib_ref] was used to infect HeLa and HuH 7 cells in this study, following previously published procedures for virus amplification and plaque assay [bib_ref] Increased ATP generation in the host cell is required for efficient vaccinia..., Chang [/bib_ref] [bib_ref] Comparison of methods for detection of vaccinia virus in patient specimens, Fedorko [/bib_ref]. Cytosine arabosinide (ara C), where used, was added to the cells at a concentration of 40 μg/ml. Recomninant virus vI7Li, a gift kindly provided by Dr. Bernard Moss, was also used to infect HeLa cells following previously published procedures [bib_ref] Role of the I7 protein in proteolytic processing of vaccinia virus membrane..., Ansarah-Sobrinho [/bib_ref]. Influenza A virus WSN33 was used to infect MDCK cells following previously published procedures for virus amplification and plaque assay [bib_ref] Rescue of influenza A virus from recombinant DNA, Fodor [/bib_ref].
## Western blotting (wb) analysis
Our previous procedures were followed for WB analysis [bib_ref] Characterization of the cleavage of signal peptide at the C-terminus of hepatitis..., Ma [/bib_ref] [bib_ref] Enterovirus type 71 2A protease functions as a transcriptional activator in yeast, Yang [/bib_ref] [bib_ref] SARS-CoV nucleocapsid protein interacts with cellular pyruvate kinase protein and inhibits its..., Wei [/bib_ref]. The primary antibodies used for the analyses in this study were mouse monoclonal antibodies against the C-terminus (a.a. 1813-1912) of Dicer (M01, clone 2F12, Abnova, Taiwan), against the a.a. 378-385 of Dicer (clone 5D12.2, Millipore, USA), against T7 tag (Novagen, USA), against HCV NS5A protein (Biodesign, USA), against V5 tag (Serotec, USA) and rabbit polyclonal antibodies against ERK-2 protein (Santa Cruz Biotechnology, USA), against NPT protein (Upstate, USA), against β -actin (GeneTex, CA, USA) and against I7 protein (a gift kindly provided by Dr. Bernard Moss). Usually, 200 ug protein is enough for the WB analysis. However, due to the insensitivity of the antibodies against the C-terminus of Dicer, 800 ug protein is needed to detect the endogenous Dicer protein. When this antibody was used to detect exogenous T7-Dicer protein from expressing plasmid, the endogenous Dicer protein could not be detected.
Real-time reverse transcriptase-polymerase chain reaction (real-time RT-PCR) Total RNAs extracted from HeLa or HuH7 cells (mock or vaccinia virus infection) were converted into cDNAs using oligo-dT as the primer. Our previous procedures were followed for real-time RT-PCR [bib_ref] Increased ATP generation in the host cell is required for efficient vaccinia..., Chang [/bib_ref] [bib_ref] Visualization of the structures of the hepatitis C virus replication complex, Chan [/bib_ref]. Specific primers identical to those used in the Takamizawa's report were used to detect precursor miRNA122 and U6 RNA. Primers (5' TGGAGACGCCAA TAGCAATA 3' and 5' TGCTGCTGCAGTGAATTCTT 3') were used to detect Dicer mRNA while primers (5'CGCTGGTCAGTTCGTGATTA 3' and 5'AACTCAGGCCCATTTCCTTT 3') were used to detect TFRC mRNA as a control.
# Statistical analysis
Experiments were performed three times. Data were analyzed using student t test. P<0.05 was considered statistically significant (p<0.05, Ã ; p<0.01, ÃÃ ; p<0.001, ÃÃÃ ).
# Results
## Reduction of dicer protein during vaccinia virus infection in different cell types
The amount of endogenous Dicer protein was reduced after VV infection in HeLa cells as detected by antibodies against its extreme C-terminus [fig_ref] Fig 1: Reduction of Dicer protein during vaccinia virus infection [/fig_ref]. This reduction could be recovered after AraC, an inhibitor of VV DNA replication, was added after virus infection (lane 3, [fig_ref] Fig 1: Reduction of Dicer protein during vaccinia virus infection [/fig_ref]. Unlike the TFRC mRNA decreased after VV infection in a dose dependent manner, the Dicer mRNA level, though decreased, was not specifically reduced after VV infection in S1 File). Exogenously expressed Dicer protein with a T7 tag at its N-terminus was also reduced after VV but not influenza A virus infection in HeLa cells as detected by antibodies against its extreme C-terminus [fig_ref] Fig 1: Reduction of Dicer protein during vaccinia virus infection [/fig_ref]. A smaller protein with tiny amount was detected using antibodies against T7 tag at the N-terminus of this fusion protein (lane 3, left panel of [fig_ref] Fig 1: Reduction of Dicer protein during vaccinia virus infection [/fig_ref]. This smaller protein was also detected after addition of MG132, an inhibitor of proteasome degradation, after virus infection (lane 7, right panel of [fig_ref] Fig 1: Reduction of Dicer protein during vaccinia virus infection [/fig_ref].
The amount of endogenous Dicer protein was also reduced after VV infection in HuH7 cells [fig_ref] Fig 1: Reduction of Dicer protein during vaccinia virus infection [/fig_ref]. Again, unlike the TFRC mRNA, the mRNA level of Dicer, though reduced, was not specifically reduced after VV infection in these cells in S1 File). Similar results were found in HCV replicon cells (data not shown, also see below), which are derived from HuH7 cells with HCV subgenomic RNAs [bib_ref] Reactive oxygen species suppress hepatitis C virus RNA replication in human hepatoma..., Choi [/bib_ref].
## Cleavage of dicer protein by i7 protease facilitates dicer degradation
It is possible that during VV infection Dicer protein is cleaved by a viral protease at the Cterminus and the cleaved products are then degraded by proteasome degradation pathway. VV G1 protein, a predicted metalloprotease, is essential for the morphogenesis of infectious virions but not for the cleavage of major core proteins [bib_ref] Vaccinia virus G1 protein, a predicted metalloprotease, is essential for morphogenesis of..., Ansarah-Sobrinho [/bib_ref]. In addition, I7 protease is responsible for processing most or all viral core and membrane proteins in the late stage of VV life cycle [bib_ref] Role of the I7 protein in proteolytic processing of vaccinia virus membrane..., Ansarah-Sobrinho [/bib_ref] [bib_ref] The vaccinia virus I7L gene product is the core protein proteinase, Byrd [/bib_ref] [bib_ref] Molecular dissection of the vaccinia virus I7L core protein proteinase, Byrd [/bib_ref]. These proteolytic events are involved in the transformation of immature virions into mature virions. There are five predicted cleavage sites for I7 protease (AG/X) in Dicer protein sequence. Therefore, I7 protease may involve in the cleavage of Dicer protein. To address this issue, I7 protease gene was cloned and expressed (Figure SII in S1 File). When co-expressed with I7 protease, exogenously expressed Dicer protein was reduced as detected using antibodies against the extreme C-terminus of Dicer protein [fig_ref] Fig 2: Cleavage of Dicer protein by I7 protease facilitates Dicer degradation [/fig_ref]. When MG132 was added, full-length Dicer protein was increased in the absence of I7 protease [fig_ref] Fig 2: Cleavage of Dicer protein by I7 protease facilitates Dicer degradation [/fig_ref] , lanes 2 and 4) while it was further reduced in the presence of I7 protease [fig_ref] Fig 2: Cleavage of Dicer protein by I7 protease facilitates Dicer degradation [/fig_ref] , lanes 3-5) as detected using antibodies against the extreme C-terminus of Dicer protein. The same samples were analyzed using antibodies recognizing the N-terminal T7 tag of Dicer protein. A band with smaller size was detected in the presence of MG132 when Dicer and I7 proteins were coexpressed [fig_ref] Fig 2: Cleavage of Dicer protein by I7 protease facilitates Dicer degradation [/fig_ref]. Expression of I7 protease could only be detected by Western blotting analysis in the presence of MG132 to stabilize this labile protein (lane 5, [fig_ref] Fig 2: Cleavage of Dicer protein by I7 protease facilitates Dicer degradation [/fig_ref]. The plasmid expressing a mutant I7 protease lack of the protease activity was constructed by replacing amino acid residue 328 from Cys to Ala [bib_ref] Role of the I7 protein in proteolytic processing of vaccinia virus membrane..., Ansarah-Sobrinho [/bib_ref]. Reduction of Dicer protein was not detected when it was co-expressed with this mutant I7 protein [fig_ref] Fig 2: Cleavage of Dicer protein by I7 protease facilitates Dicer degradation [/fig_ref].
Five potential viral I7 protease cleavage sites in Dicer protein are after a.a. 13, 323, 465, 1079 and 1817. It is possible that Dicer protein was cleaved by I7 protease after a.a. 1817, and the cleaved protein (about 199 kDa) was further degraded through proteasome ubiquitin degradation pathway [fig_ref] Fig 2: Cleavage of Dicer protein by I7 protease facilitates Dicer degradation [/fig_ref]. To this end, a plasmid expressing the Dicer protein with mutation in a.a. 1816 and 1817 was constructed. Compared with the wild-type Dicer protein, the protein amount of mutated Dicer protein was no longer reduced after VV infection [fig_ref] Fig 2: Cleavage of Dicer protein by I7 protease facilitates Dicer degradation [/fig_ref].
To further demonstrate the reduction of Dicer protein by VV infection was through I7 protease, a recombinant VV vI7Li expressing I7 protease protein under IPTG regulation was used [bib_ref] Role of the I7 protein in proteolytic processing of vaccinia virus membrane..., Ansarah-Sobrinho [/bib_ref]. Indeed, Dicer protein was reduced dramatically by recombinant VV vI7Li after but not before the induction of I7 protease [fig_ref] Fig 3: Dicer protein was reduced after the induction of I7 protease expression during... [/fig_ref].
## Inhibition of mir122 processing either by vaccinia virus infection or by i7 protease expression
Pre-miRNAs transited by exportin-5 to the cytoplasm were cleaved by Dicer protein to generate miRNAs. Thus, the degradation of Dicer protein should block the formation of miRNAs and results in the accumulation of pre-miRNAs during VV infection. miRNA repertoires are highly cell type specific and change markedly during development or upon cell activation [bib_ref] Origins and Mechanisms of miRNAs and siRNAs, Carthew [/bib_ref]. miR122 is abundant in HuH7 cells [bib_ref] Modulation of hepatitis C virus RNA abundance by a liver-specific MicroRNA, Jopling [/bib_ref]. To determine the effect of VV infection on the miRNA processing, the endogenous pre-miR122 level was analyzed in HuH7 cells after VV infection. As expected, the endogenous pre-miR122 level was increased after VV infection in a dose dependent manner [fig_ref] Fig 4: Inhibition of miR122 processing either by VV infection or by I7 protease... [/fig_ref]. Similar results were found in HCV replicon cells after VV infection (data not shown, also see below). miR122 is scarce in HeLa cells. The plasmid expressing pre-miR122 was constructed and transfected into HeLa cells. The exogenous pre-miR122 level was analyzed in HeLa cells after VV infection. Again, the exogenous pre-miR122 level was increased after VV infection in a dose dependent manner [fig_ref] Fig 4: Inhibition of miR122 processing either by VV infection or by I7 protease... [/fig_ref] while the Dicer mRNA level was not affected significantly in S1 File). The exogenous pre-miR122 level was also analyzed in HeLa cells cotransfected with the plasmids expressing pre-miR122 and I7 protease. The exogenous pre-miR122 level was also increased in the presence of I7 protease in a dose dependent manner [fig_ref] Fig 4: Inhibition of miR122 processing either by VV infection or by I7 protease... [/fig_ref] while the Dicer mRNA level was not affected by I7 protease (Figure SIC in S1 File).
## Inhibition of mir122 function either by vaccinia virus infection or by i7 protease expression
It has been previously reported that miRNA122 could facilitate HCV replication [bib_ref] Modulation of hepatitis C virus RNA abundance by a liver-specific MicroRNA, Jopling [/bib_ref] [bib_ref] Liver-specific microRNA miR-122 enhances the replication of hepatitis C virus in nonhepatic..., Chang [/bib_ref]. Degradation of Dicer protein would reduce the production of miRNA122 and, in turn, should repress the HCV RNA replication. Thus, the replication of HCV subgenomic RNA, and in turn the amount of proteins encoded from this RNA, should be reduced in HCV replicon cells after VV infection [bib_ref] Visualization of the structures of the hepatitis C virus replication complex, Chan [/bib_ref]. Indeed, the HCV NS5A and core-NPT protein levels were suppressed, accompanying with the reduction of Dicer protein in HCV replicon cells after VV infection in a dose dependent manner [fig_ref] Fig 5: Inhibition of miR122 function either by VV infection or by I7 protease... [/fig_ref]. The plasmid expressing I7 protease was also transfected into the HCV replicon cells. As expected, accompanying with the reduction of Dicer protein, the HCV NS5A and core-NPT protein levels were suppressed in a dose dependent manner [fig_ref] Fig 5: Inhibition of miR122 function either by VV infection or by I7 protease... [/fig_ref]. Again, expression of I7 protease could only be detected by Western blotting analysis in the presence of MG132 to stabilize this labile protein (lane 6, [fig_ref] Fig 5: Inhibition of miR122 function either by VV infection or by I7 protease... [/fig_ref].
## Suppression of vv replication slightly by exogenous expression of dicer protein
To determine the effect of Dicer protein on VV replication, the gain-of-function (overexpression of Dicer) and loss-of-function (knockdown of Dicer) assays were performed. Exogenously expressed Dicer protein suppressed viral protein amount (e.g., A type inclusion protein) inside the cells slightly [fig_ref] Fig 6: Suppression of vaccinia virus replication slightly by exogenously expressed Dicer protein [/fig_ref] and also reduced the number of the secreted viral particles in the medium to around 60% [fig_ref] Fig 6: Suppression of vaccinia virus replication slightly by exogenously expressed Dicer protein [/fig_ref]. The number of secreted viral particles was further reduced when the Dicer protein with mutations in a.a. 1816 and 1817 was exogenously expressed [fig_ref] Fig 1: Reduction of Dicer protein during vaccinia virus infection [/fig_ref]. Twenty-four hrs after infection, mRNAs were extracted and converted into cDNA. Then, real-time PCR assay was performed to detect the amount of miR122 using U6 mRNA as the internal control for normalization. (C) HeLa cells were mock-transfected or cotransfected with the plasmids expressing pre-miR122 and I7 protease with the indicated amount. Forty-eight hrs after transfection, mRNAs were extracted and converted into cDNA. Then, real-time PCR assay was performed to detect the amount of miR122 using U6 mRNA as the internal control for normalization. (p<0.05, *; p<0.01, **; p<0.001, ***).
# Discussion
Our results in this study showed that Dicer protein was reduced in VV-infected cells [fig_ref] Fig 1: Reduction of Dicer protein during vaccinia virus infection [/fig_ref] , and in turn, the processing and the function of miR122 were blocked [fig_ref] Fig 4: Inhibition of miR122 processing either by VV infection or by I7 protease... [/fig_ref]. Reduction of Dicer protein should affect the processing of universal miRNAs. Indeed, a recent report indicated that Dicer protein was suppressed in VV-infected cells that was associated with a universal reduction of host miRNAs expression. miRNAs have been implicated in a vast array of cellular processes including cell proliferation and apoptosis [bib_ref] MicroRNAs: genomics, biogenesis, mechanism, and function, Bartel [/bib_ref]. Thus, reduction of Dicer protein during VV infection is probably one of the many factors responsible for the viral pathogenesis.
Reduction of Dicer protein in VV-infected cells [fig_ref] Fig 1: Reduction of Dicer protein during vaccinia virus infection [/fig_ref] may be caused by several different mechanisms. One possible mechanism is due to mRNA reduction (Figures SIA and SIB in S1 File), which is also reported previously. VV is known to enhance the degradation of host mRNAs by two decapping enzymes encoded by the virus, D9 and D10 [bib_ref] Characterization of a second vaccinia virus mRNA-decapping enzyme conserved in poxviruses, Parrish [/bib_ref] [bib_ref] Vaccinia virus D10 protein has mRNA decapping activity, providing a mechanism for..., Parrish [/bib_ref]. There are still other possible mechanisms. The reduction of Dicer protein during VV infection could be recovered by adding araC to inhibit the VV DNA replication [fig_ref] Fig 1: Reduction of Dicer protein during vaccinia virus infection [/fig_ref]. Therefore, in addition to the reduction of Dicer mRNA, a viral protease expressed after VV DNA synthesis is also probably responsible for the reduction of Dicer protein. I7 protease is required for AG/X-specific cleavages of viral membrane and core proteins during VV assembly [bib_ref] Role of the I7 protein in proteolytic processing of vaccinia virus membrane..., Ansarah-Sobrinho [/bib_ref]. Our results further demonstrated that Dicer protein was first cleaved by I7 protease after a.a. 1817 and the cleaved product was then degraded by the proteasome-ubiquitin degradation pathway during VV infection [fig_ref] Fig 1: Reduction of Dicer protein during vaccinia virus infection [/fig_ref]. Comparing with the I7 protease encoded by VV infection [fig_ref] Fig 3: Dicer protein was reduced after the induction of I7 protease expression during... [/fig_ref] , I7 protease derived from expressing plasmid is much more labile because it could only be detected by Western blotting analysis in the presence of MG132 [fig_ref] Fig 2: Cleavage of Dicer protein by I7 protease facilitates Dicer degradation [/fig_ref]. This also indicated that tiny amount of I7 protease should be sufficient for the cleavage of Dicer protein (Figs. 2A and 5B). However, Dicer protein was reduced after the induction of I7 protease protein in the presence of more than 50 uM IPTG but not in less than 50 uM IPTG [fig_ref] Fig 3: Dicer protein was reduced after the induction of I7 protease expression during... [/fig_ref] , indicating that not only the tiny amount of I7 protease but also other factors (e.g., whether I7 protease interacts with Dicer protein or not) are important to cleave Dicer protein and facilitate its degradation.
There are many different ways of interactions between viruses and miRNAs. Firstly, host miRNAs could suppress or facilitate viral replication [bib_ref] RNAi and cellular miRNAs in infections by mammalian viruses, Haasnoot [/bib_ref] [bib_ref] Dicing with viruses: microRNAs as antiviral factors, Muller [/bib_ref] [bib_ref] Modulation of hepatitis C virus RNA abundance by a liver-specific MicroRNA, Jopling [/bib_ref]. Secondly, many DNA viruses, including herpesviruses, adenovirus, polyomaviruses and papillomavirus, have evolved to encode viral miRNAs to potentially control various phases of the viral life cycle, such as latency, reactivation, replication, etc. [bib_ref] Kaposi's sarcoma-associated herpesvirus expresses an array of viral microRNAs in latently infected..., Cai [/bib_ref] [bib_ref] Epstein-Barr virus microRNAs are evolutionarily conserved and differentially expressed, Cai [/bib_ref] [bib_ref] The functions of herpesvirus-encoded microRNAs, Grey [/bib_ref] [bib_ref] Identification of virus-encoded micro-RNAs, Pfeffer [/bib_ref] [bib_ref] Identification and function of human cytomegalovirus microRNAs, Grey [/bib_ref]. We do not expect VV would encode its viral miRNAs, similar to those typical cellular miRNAs, because this virus is replicated in the cytoplasm. Thirdly, both our present study and Dr. Grinberg's reportshowed that a suppression of host miRNA expression was followed by the reduction of Dicer protein during VV infection. Other DNA viruses may not suppress Dicer protein since this would affect the processing of their own viral miRNAs. Therefore, it is not surprising that VV, different from the other DNA viruses, uses a novel way to interact with miRNAs.
RNAi could serve as an innate antiviral mechanism in plants, fungi and animals. Human viruses, like plant viruses, encode suppressor proteins or RNAs that block or modulate the RNAi pathway [bib_ref] RNAi suppressors encoded by pathogenic human viruses, De Vries [/bib_ref]. Several mammalian viruses contain viral proteins with RSS activity, that usually involved two mechanisms: Dicer binding and siRNA binding [bib_ref] Viral RNA silencing suppressors (RSS): novel strategy of viruses to ablate the..., Bivalkar-Mehla [/bib_ref]. The results of this study indicated that I7 protease of VV could cleave Dicer protein to facilitate Dicer degradation. Thus, VV I7 protease possesses RSS activity with a novel mechanism.
Suppression of VV replication slightly was demonstrated by exogenous expression of Dicer protein [fig_ref] Fig 6: Suppression of vaccinia virus replication slightly by exogenously expressed Dicer protein [/fig_ref]. However, knockdown of Dicer protein did not facilitate VV replication significantly [fig_ref] Fig 6: Suppression of vaccinia virus replication slightly by exogenously expressed Dicer protein [/fig_ref]. This may be simply due to the efficient cleavage of Dicer protein during VV infection [fig_ref] Fig 1: Reduction of Dicer protein during vaccinia virus infection [/fig_ref].
In conclusion, results in this study indicate that, during VV infection, the cleavage of Dicer protein by I7 protease facilitates Dicer degradation, and in turn, suppresses the processing of miRNAs.
## Supporting information
[fig] Fig 1: Reduction of Dicer protein during vaccinia virus infection. (A) HeLa cells were either mock infected (lane 1) or infected with VV (M.O.I. = 10) in the absence (lane 2) or in the presence (lane 3) of AraC. Sixteen hrs after infection, cell lysates were analyzed by SDS-PAGE and Western blotting with antibodies against the extreme C-terminus of Dicer protein (upper panel), A type inclusion protein (middle panel) or β -actin protein (lower panel) as the loading control. (B) HeLa cells were either mock transfected (lane 1) or transfected with the plasmid expressing Dicer protein with a T7 tag at its N-terminus (T7-Dicer, lanes 2-6). Forty-eight hrs after transfection, these cells were either mock infected (lane 2) or infected with VV (M.O.I. = 3, lanes 3-5). After the time indicated, cell lysates were analyzed by SDS-PAGE and Western blotting with antibodies against the extreme C-terminus of Dicer protein (upper panel), NPT protein as the transfection control, A type inclusion protein or Erk2 protein (bottom panel) as the loading control. (C) HeLa cells were either mock transfected (lanes 1 and 5) or transfected with the plasmid expressing T7-Dicer (lanes 2-4 and 6-8). Forty-eight hrs after transfection, these cells were either mock infected, infected with VV (M.O.I. = 3, lanes 3, 7 and 8) or with influenza A virus (M.O.I. = 3, lane 4). MG132 (lane 7) or araC (lane 8) was also added in the culture. Twenty-four hrs after infection, cell lysates were analyzed by SDS-PAGE and Western blotting with antibodies against the T7 tag at the N-terminus of Dicer protein (upper panel), against NPT protein (middle panel) as the transfection control, or against Erk2 protein (bottom panel) as the loading control. (D) HuH7 cells were either mock infected (lane 1) or infected with VV in M.O.I = 1 (lane 2), 5 (lane 3) or 10 (lanes 4 and5). araC was also added in the culture (lane 5). Twenty-four hrs after infection, cell lysates were analyzed by SDS-PAGE and Western blotting with antibodies against the Dicer protein (upper panel), A type inclusion protein (middle panel), β-actin protein or Erk2 protein (lower panel). Both β-actin and Erk2 proteins were served as the loading control. [/fig]
[fig] Fig 2: Cleavage of Dicer protein by I7 protease facilitates Dicer degradation. (A) HeLa cells were mock-transfected (lane 1), co-transfected with the plasmids expressing T7-Dicer and empty vector (lanes 2 and 4) or the plasmids expressing T7-Dicer and I7 protease with V5 tag (lanes 3 and 5). Thirty-two hrs after transfection, 10 uM MG132 was also added (lanes 4 and 5). Sixteen hrs later, cell lysates were analyzed by SDS-PAGE and Western blotting with antibodies against the extreme C-terminus of Dicer protein (upper panel), V5-tag to detect the expression of I7 protease, NPT protein as the transfection control, or Erk2 protein as the loading control (bottom panel). The arrow marks the position of I7 in lane 5. (B) HeLa cells were mock-transfected (lane 4), cotransfected with the plasmids expressing T7-Dicer and empty vector (lane 1) or the plasmids expressing T7-Dicer and I7 protease (lanes 2 and 3). Thirty-two hrs after transfection, 10 uM MG132 was also added (lane 3). Sixteen hrs later, cell lysates were analyzed by SDS-PAGE and Western blotting with antibodies against the T7 tag at the N-terminus of Dicer protein (upper panel) or NPT protein as the transfection control. (C) HeLa cells were transfected with empty vector (Vec), plasmids expressing I7 protease (I7) or I7 protease containing C328A mutation (I7m). Twenty-four hrs after transfection, DMSO (left panels) or 20 uM MG132 (right panels) was also added. Twenty-four hrs after treatment, cell lysates were analyzed by SDS-PAGE and Western blotting with antibodies against Dicer (upper panel), NPT protein as transfection control (middle panel) or β-actin for the loading control (bottom panel). (D) Different functional domains in the Dicer protein. Five potential viral I7 protease cleavage sites (a.a. 13, 323, 465, 1079 and 1817) in Dicer protein are marked by arrows. (E) HeLa cells were mock-transfected (lane 1), transfected with the plasmid expressing T7-Dicer (lanes 2-4) or transfected with the plasmid expressing T7-Dicer with mutations in a.a. 1816 and 1817 (mT7-Dicer, lanes 5-7). Twenty-four hrs after transfection, these cells were either mock infected (lanes 2 and 5), or infected with VV (M.O.I. = 5) in the presence (lanes 4 and 7) or absence (lanes 3 and 6) of 100 ug araC. Twenty-four hrs later, cell lysates were analyzed by SDS-PAGE and Western blotting with antibodies against the N-terminal T7 tag of Dicer protein (upper panel), NPT protein as the transfection control or Erk2 protein as the loading control. doi:10.1371/journal.pone.0120390.g002 [/fig]
[fig] Fig 3: Dicer protein was reduced after the induction of I7 protease expression during recombinant virus vI7Li infection. HeLa cells were mock-infected, infected with wild-type vaccinia virus (WT) or recombinant virus vI7Li (M.O.I. = 5). IPTG was also added as the indicated amount. Twenty-four hrs after infection, cell lysates were analyzed by SDS-PAGE and Western blotting with antibodies against the Dicer, A type inclusion protein, I7, β-actin for the loading control or GFP which is constitutively expressed during vI7Li infection. [/fig]
[fig] Fig 4: Inhibition of miR122 processing either by VV infection or by I7 protease expression. (A) HuH7 cells were either mock-infected or infected with VV in different M.O.I. (1, 5, and 10). Twenty-four hrs after infection, mRNAs were extracted and converted into cDNA. Then, real-time PCR assay was performed to detect the amount of miR122 using U6 mRNA as the internal control for normalization. (B) HeLa cells were mock-transfected, transfected with empty vector (2 ug) or with the plasmid expressing pre-miR122 (2 ug). Twenty-four hrs after transfection, the cells with pre-miR122 were mock-infected or infected with VV in different M.O.I. (1, [/fig]
[fig] Fig 5: Inhibition of miR122 function either by VV infection or by I7 protease expression. (A) HCV replicon cells were mock-infected (lane 2) or infected with VV in M.O.I. = 1 (lane 3), 5 (lane 4), 10 (lanes 5 and 6). 100 ug araC was also added (lane 6). Twenty-four hrs later, cell lysates were analyzed by SDS-PAGE and Western blotting with antibodies against Dicer protein (upper panel), NS5A and NPT proteins for the expression of HCV subgenomic RNAs, or Erk2 protein as the loading control. Cell lysates from mock-infected HuH7 cells (lane 1) were served as a negative control for the detection of NS5A and NPT. (B) HCV replicon cells were mock-transfected or transfected with the indicated amount of the plasmid expressing I7 protease (lanes 1-6). Thirty-two hrs after transfection, 10 uM MG132 was also added (lanes 5 and 6). Sixteen hrs later, cell lysates were analyzed by SDS-PAGE and Western blotting with antibodies against Dicer protein, NS5A protein for the expression of HCV subgenomic RNAs, NPT protein (the thin arrow marks the position of NPT protein expressed from the transfected plasmids while the thick arrow marks the position of core-NPT fusion protein encoding from HCV subgenomic RNAs.), V5 tag to detect the expression of I7 (The arrow marks the position of I7 in lane 6.) or Erk2 protein as the loading control. doi:10.1371/journal.pone.0120390.g005 [/fig]
[fig] Fig 6: Suppression of vaccinia virus replication slightly by exogenously expressed Dicer protein. (A, B) HeLa cells were transfected with empty vector (2 ug, lanes 1 and 2), with the plasmid expressing T7-Dicer protein (2 ug, lanes 3 and 4) or with the plasmid expressing T7-Dicer with mutations in a.a. 1816 and 1817 (2 ug, lanes 5 and 6). Twenty-four hrs after transfection, the cells were mock-infected (lanes 1, 3 and 5) or infected with VV (M.O.I. = 10) (lanes 2, 4 and 6). Twenty-four hrs after infection, cell lysates were analyzed by SDS-PAGE and Western blotting with antibodies against the Dicer, A type inclusion protein and β-actin (A) while the secreted virus particles from the supernatants (lanes 2, 4 and 6) were analyzed by the plaque assay (B). (p<0.01, **). (C) HeLa cells were transfected with scramble siRNA as a control or siRNA against Dicer. Twenty-four hrs after transfection, cell lysates were analyzed by SDS-PAGE and Western blotting with antibodies against the Dicer (upper panel) and β-actin proteins (bottom panel) as the loading control. (D) HeLa cells were transfected with scramble or Dicer siRNA. Twenty-four hrs after transfection, the cells were infected with VV (M.O.I. = 10). Twenty-four hrs after infection, the secreted viral particles from the supernatants were analyzed by plaque assay. NS: not significant. doi:10.1371/journal.pone.0120390.g006 [/fig]
[fig] S1: File. This file contains Figures SI and SII. (DOCX) [/fig]
|
Analysis of causes for poor persistence of CAR-T cell therapy in vivo
# Introduction
In recent years, great progress has been made in chimeric antigen receptor T cell (CAR-T cell) therapy for hematological and solid tumors, but patients still experience recurrence after treatment. Poor persistence of CAR-T cells in vivo is an important cause of relapse [bib_ref] Recent advances in car-T cell engineering, Huang [/bib_ref]. This review summarizes three factors limiting why CAR-T cells persistence in vivo: the structure of CARs, the proportion of memory CAR-T cells and the TME. The CARs structure of CAR-T cells has undergone four generations of evolution. Second-generation CARs include additional costimulatory domains that are lacking in firstgeneration CARs, third-generation CARs have two costimulatory domains, and fourth-generation CARs have both costimulatory domains and other domains that can regulate cytokines or other molecules. With the structural additions of each generation of CARs, the survival and persistence of CAR-T cells in vivo has improved [bib_ref] Chimeric fv-zeta or fv-epsilon receptors are not sufficient to induce activation or..., Brocker [/bib_ref] [bib_ref] Enhancing car T cell persistence through icos and 4-1bb costimulation, Guedan [/bib_ref] [bib_ref] Activation of resting human primary T cells with chimeric receptors: Costimulation from..., Finney [/bib_ref] [bib_ref] Cd27 costimulation augments the survival and antitumor activity of redirected human T..., Song [/bib_ref] [bib_ref] Chimeric receptors containing Cd137 signal transduction domains mediate enhanced survival of T..., Milone [/bib_ref] [bib_ref] Car T cells releasing il-18 convert to T-Bet(High) Foxo1 (Low) effectors that..., Chmielewski [/bib_ref] [bib_ref] Inducible secretion of il-21 augments anti-tumor activity of piggybac-manufactured chimeric antigen receptor..., Stach [/bib_ref] [bib_ref] Cd40 ligand-modified chimeric antigen receptor T cells enhance antitumor function by eliciting..., Kuhn [/bib_ref]. Studies have shown that the higher the percentage of memory T cells, the more durable the T cells are in the vivo [bib_ref] Chimeric antigen receptor T cell persistence and memory cell formation, Mclellan [/bib_ref]. Therefore, it is essential to improve the proportion of memory CAR-T cells, which can be increased in four ways: preventing T cell differentiation [bib_ref] Determinants of response and resistance to Cd19 chimeric antigen receptor (Car) T..., Fraietta [/bib_ref] [bib_ref] Immunotherapy of non-hodgkin's lymphoma with a defined ratio of Cd8+ and Cd4+..., Kim-Schulze [/bib_ref] ; reprogramming terminally differentiated T cells ; shortening the culture time of CAR-T cells [bib_ref] Enhancement of the in vivo persistence and antitumor efficacy of Cd19 chimeric..., Ghassemi [/bib_ref] and delaying the senescence of CAR-T cells (21). In addition, immune checkpoint molecules and certain cytokines in the TME can also shorten the lifespan and reduce the function of T cells [bib_ref] Pd-1 blockade with nivolumab in relapsed or refractory hodgkin's lymphoma, Ansell [/bib_ref] [bib_ref] End of phase 1 results from zuma-6: Axicabtagene ciloleucel (Axi-cel) in combination..., Jacobson [/bib_ref] [bib_ref] Targeting tumors with il-21 reshapes the tumor microenvironment by proliferating pd-1inttim-3-Cd8+ T..., Deng [/bib_ref]. [bib_ref] T-Cell activation by recombinant immunoreceptors: Impact of the intracellular signalling domain on..., Heuser [/bib_ref] The structure of CARs can influence CAR-T cell proliferation and persistence CAR-T cell therapy consists of editing T cells to express specific CARs and reinjecting the cells expanded in vitro into the patient to eradicate tumors [bib_ref] Enhancing car T cell persistence through icos and 4-1bb costimulation, Guedan [/bib_ref]. A CAR is a receptor consisting of three main portions: an extracellular antigen recognition domain, a transmembrane domain, and an intracellular signaling domain (25).
## First-generation cars of car-t cells
First-generation CARs contain a single-chain variable fragment (scFv) that is linked to the intracellular signaling domain of CD3z, providing the signal that is necessary to activate T cells; however, these CARs only provide the T cell priming signal and have no costimulatory molecules. The lack of costimulatory molecules results in low activation and antitumor effects of T cells in vivo (2, 3)..
## Second-generation cars of car-t cells
Developed based on first-generation CARs, second-generation CARs include a costimulatory domain (CD28, 4-1BB, OX-40, ICOS and CD134) that provides a second signal for T-cell activation [bib_ref] Activation of resting human primary T cells with chimeric receptors: Costimulation from..., Finney [/bib_ref]. The second signal prevents the exhaustion of CAR-T cells and promotes their continued proliferation and cytokine secretion, resulting in increased potency in the killing of target cells in vivo [bib_ref] Chimeric receptors containing Cd137 signal transduction domains mediate enhanced survival of T..., Milone [/bib_ref]. This review focuses on the CD28 and 4-1BB costimulatory domains .
CD28 is a transmembrane protein and a member of the immunoglobulin superfamily (26). The CD28 costimulatory receptor is expressed on both CD8+ T and CD4+ T cells. The extracellular portion of CD28 binds to the B7 family of ligands [CD80, CD86 (B7-2), and CD275 (B7-H2) (27-29)] to activate Evolution of chimeric antigen receptors (CARs). Each CAR generation contains an extracellular single-stranded variable region (scFv), an intracellular CD3 domain (CD3z) and a T-cell receptor transmembrane domain. The first generation of CARs has only the CD3 domain in the cell. The second generation of CARs includes the CD3 domain and the costimulatory domain (CM). Developed on the basis of secondgeneration CARs, the third generation of CARs has two different costimulatory domains in the cell. Developed based on the structure of second-generation of CARs, fourth-generation of CARs have an additional intracellular domain that regulates the expression of cytokines or other costimulatory molecules.
intracellular signaling cascades, such as the PI3K, NF-kB,AP-1 and NFAT signaling pathways, which can regulate T-cell proliferation and survival and promote IL-2 production (26, 30, 31). 4-1BB (also known as CD137 and TNFRSF9) is also a transmembrane protein and a member of the tumor necrosis factor receptor superfamily (TNFRSF). 4-1BB is found on antigen-activated CD8+ and CD4+ T cells [bib_ref] Cdna sequences of two inducible T-cell genes, Kwon [/bib_ref]. 4-1BB binds to its sole ligand (4-1BBL) to activate downstream signaling by recruiting TRAF proteins [bib_ref] 4-1bb enhancement of car T function requires nf-Kb and trafs, Li [/bib_ref].
The CD28 signaling domain represents an "early" costimulatory signal that promotes the proliferation of T cells, accelerates the differentiation of effector memory T (TEM) cells and CD8+ T cells, and leads to increased more IL-2 production. Encoding a fully functional CD28 signaling domain on the CAR polypeptide chain may produce overstimulation, increase the incidence of cytokine release syndrome (CRS), promote T-cell exhaustion, and reduce persistence [bib_ref] 4-1bb enhancement of car T function requires nf-Kb and trafs, Li [/bib_ref] [bib_ref] Distinct signaling of coreceptors regulates specific metabolism pathways and impacts memory development..., Zhao [/bib_ref]..
The 4-1BB signaling domain is responsible for the "late" costimulatory signal that promotes T cell differentiation into central memory T(TCM) cells, increases the oxidative metabolism and glycolysis of CAR-T cells, enhances T cell persistence and improves the antitumor ability of T cells [bib_ref] Distinct signaling of coreceptors regulates specific metabolism pathways and impacts memory development..., Zhao [/bib_ref] [bib_ref] Chimeric antigen receptors combining 4-1bb and Cd28 signaling domains augment Pi3kinase/Akt/Bcl-xl activation..., Zhong [/bib_ref].
Studies showed that the 4-1BB domain in CAR activated ncNF-kB signaling in human T cells, thereby increasing the expansion and the survival of 4-1BBz CAR-T cells in vitro (37).
In preclinical studies of CAR-T cells in pre-B cell acute lymphoblastic leukemia (pre-BALL) mice, compared with CD28 costimulation or CD3z signaling alone, 4-1BB costimulation of CAR-T cells improved the survival of tumor-bearing mice .
In relapsed or refractory B-cell acute lymphoblastic leukemia (r/r B-ALL), 4-1BB-based CAR-T cells have been shown to show higher antitumor efficacy, longer persistence, and fewer serious adverse events than CD28 CAR-T cells [bib_ref] Efficacy and safety of Cd28-or 4-1bb-Based Cd19 car-T cells in b cell..., Zhao [/bib_ref].
It has been shown that CAR-T cells can live approximately 30 days in patients who are treated with CD28z CAR-T cells, while CAR-T cells can persist for more than 4 years in those who are treated with 4-1BBz CAR-T cells [bib_ref] Cd19-targeted T cells rapidly induce molecular remissions in adults with chemotherapy-refractory acute..., Brentjens [/bib_ref] [bib_ref] T Cells expressing Cd19 chimeric antigen receptors for acute lymphoblastic leukaemia in..., Lee [/bib_ref] [bib_ref] Chimeric antigen receptor T cells persist and induce sustained remissions in relapsed..., Porter [/bib_ref].
These results indicate that compared to CD28zCAR-T cells, 4-1BBzCAR-T cells are more persistent but have weaker killing effects in vivo. To improve the shortcomings of both CAR-T cells, studies have been conducted to integrate the two costimulatory molecules into the structure of CARs, leading to the development of thirdgeneration CARs of CAR-T cells.
## Third-generation cars of car-t cells
To increase T cell survival, cytokine production, and antitumor potential, researchers have modified the second-generation CARs by adding another costimulatory domain to the CARs. These modifications include the combination of CD28 and OX40 costimulatory molecules in CD30 CAR-T cells, CD28 and 4-1BB costimulatory molecules in CD19 CAR-T cells and 4-1BB and DAP10 costimulatory molecules in NKG2D(z) CAR-T cells. All of these factors enhance the expansion and persistence of CAR-T cells in vivo [bib_ref] Combination of 4-1bb and Dap10 promotes proliferation and persistence of Nkg2d(Bbz) car-T..., Wei [/bib_ref].
Nevertheless, third-generation CARs have risks of off-target effects and excess cytokine production. In addition, there is a lack of clinical data for third-generation CARs. There was clinical efficacy, but the treatment did not cause serious side effects in patients who were not cured [bib_ref] Cd28 costimulation improves expansion and persistence of chimeric antigen receptor-modified T cells..., Ramos [/bib_ref] [bib_ref] Evaluation of intracellular signaling downstream chimeric antigen receptors, Karlsson [/bib_ref]. Hence the fourth-generation of CARs have been generated.
## Fourth-generation cars of car-t cells
Fourth-generation CARs which were developed on the basis of second-generation CARs, include not only costimulatory domains but also domains that regulate the expression of cytokines or other costimulatory molecules.
IL-12 is a proinflammatory cytokine with strong tumor inhibitory activity, that not only stimulates T cells to secrete IFN-g to enhance the cytotoxicity of CAR-T cells, but also limits the activation of regulatory T (Treg) cells, reshaping the TME in an IFN-g-dependent manner [bib_ref] Reversing tumor immune suppression with intratumoral il-12: Activation of tumor-associated T Effector/Memory..., Kilinc [/bib_ref] IL-23 is a two-subunit cytokine composed of the IL-23ap19 and IL12b p40 subunits, both of which are expressed by activated macrophages and dendritic cells [bib_ref] Interleukin-23: A key cytokine in inflammatory diseases, Duvallet [/bib_ref]. In chronic lymphocytic leukemia, IL-23-activated STAT3 contributes to the enhanced function of CAR-T cells [bib_ref] Determinants of response and resistance to Cd19 chimeric antigen receptor (Car) T..., Fraietta [/bib_ref]. IL-23-producing CAR-T cells (P40-TD CAR-T) have been constructed and show higher antitumor ability, increased granzyme B expression and decreased PD-1 expression [bib_ref] Interleukin-23 engineering improves car T cell function in solid tumors, Ma [/bib_ref].
The above studies have added domains that can regulate cytokine expression to the structure of second-generation CARs, and the expression of these cytokines increases the persistence of CAR-T cells in vivo by improving the TME or promoting the generation of memory T cells. More different cytokine domains can be added to the structure of CARs in the future to increase the persistence of CAR-T cells in vivo.
## Others
In addition, targeting transcription factors can reduce the depletion of CAR-T cells and increase their persistence. CAR-T cells lacking the NR4A1, NR4A2 and NR4A3 transcription factors have downregulated expression of PD-1 and TIM3 and stronger antitumor effects [bib_ref] Nr4a transcription factors limit car T cell function in solid tumours, Chen [/bib_ref]. CAR-T cells overexpressing c-Jun have increased secretion of IL-2 and IFN-g an increased the proportion of memory T cells, and prolonged the survival of tumor-bearing mice [bib_ref] C-jun overexpression in car T cells induces exhaustion resistance, Lynn [/bib_ref].
In the report, the persistence of CAR-T cells was positively correlated with the number of memory T cells. Researchers have found that weak TCR signaling favors memory T cell differentiation, while strong TCR signaling promotes differentiation into effector T cell subsets. The amount and type of immunoreceptor tyrosine-based activation motif (ITAM) domains in CD3 and the TCR complex are related to TCR signal intensity (58). Feucht et al. found that knockout of CD3 ITAMs could induce a memory T-cell phenotype, which demonstrating better antitumor ability. The study demonstrated that the CAR-T cell therapy achieved a lasting and complete tumor response [bib_ref] Adoptive transfer of effector Cd8+ T cells derived from central memory cells..., Berger [/bib_ref] [bib_ref] Calibration of car activation potential directs alternative T cell fates and therapeutic..., Feucht [/bib_ref].
In addition, the persistence of CAR-T cells can be improved by changing the length and composition of extracellular spacers. Modifying long IgG4/IgG2-derived spacers in CAR constructs reduced antigen-independent tetanic signaling in anti-GPRC5D CAR-T cell models, thereby increasing specific antigen-dependent CAR activation [bib_ref] Gprc5d is a target for the immunotherapy of multiple myeloma with rationally..., Smith [/bib_ref].
In addition, researchers have designed CD22 CAR and CD133 CAR with either a short or long scFv linker, and found that the short scFv CAR (CART22-short or CD133 CAR-short) had excellent cytotoxicity, secreted more IFN-g, IL2 and TNF-a, resulted in lower expression of exhaustion-associated surface proteins, exhibited significant anti-leukemia activity and improved animal survival. Thus CD22 CAR and CD133 CAR improved the persistence of CAR-T cells in vivo [bib_ref] Antigenindependent activation enhances the efficacy of 4-1bb-Costimulated Cd22 car T cells, Singh [/bib_ref].
In addition, the persistence of CAR-T cells can be improved by using humanized CAR-T cells instead of mouse CAR-T cells due to poor immune responses related to the antigen binding domain.
These studies suggest that T cell persistence can also be increased in vivo by adding domains that regulate transcription factors to the CARs structure or by attenuating TCR signaling in the CARs structure.
In conclusion, first-generation to the fourth-generation CAR structures have mainly been developed with costimulatory signaling molecules and cytokine secretion in mind, while there have been few studies of the signal strength of the TCR itself. In the future, the relationship between TCR signal strength and persistence of CAR-T cells in vivo should be further explored.
## Increasing the proportion of memory t cells in car-t cells to improve car-t cell persistence
There are many subsets of T cells, each with different proliferative potential. Naive T (TN, CD45RA + CCR7 + CD62L + ) cells, stem memory T (TSCM, CD45RA + CCR7 -CD62L + ) cells and central memory T (TCM, CD45RA -CCR7 + CD62L + ) cells have a higher proliferation potential than effector T cells, and higher proportions of these cells increase the persistence of CAR-T cells in vivo [bib_ref] Chimeric antigen receptor T cell persistence and memory cell formation, Mclellan [/bib_ref]. Therefore, to increase the persistence of CAR-T cells, it is necessary to increase the proportion of memory T cells in CAR-T cells, which can be achieved by the following four methods: (1) preventing T cell differentiation, [bib_ref] T-Cell activation by recombinant immunoreceptors: Impact of the intracellular signalling domain on..., Heuser [/bib_ref] reprogramming T cells into terminally differentiated cells, (3) shortening the culture time of CAR-T cells, (4) delaying CAR-T cell senescence.
## Preventing t cell differentiation
## Cytokines
In a study of CD19 CAR-T cell therapy for chronic lymphocytic leukemia (CLL), Fraietta.et. al. found that CAR-T cells in CLL patients with complete remission were rich in the markers IL-6/ STAT3, which increased the proportion of memory T cells (CD27 + CD45RO -CD8 + ) [bib_ref] Determinants of response and resistance to Cd19 chimeric antigen receptor (Car) T..., Fraietta [/bib_ref].
It is well known that T cells can differentiate into effector T cells under the action of self-secreted IL-2 (64), but IL-2 secretion is not good for the persistence of T cells in vivo. Therefore, Fei Mo et al. designed a IL-2 partial agonist (H9T) that can promote the stemness of CD8+ T cells through the STAT5 signaling pathway (65). IL-7 promotes the expansion of primitive T cells and maintains the TCM T cell pool by increasing expression of the anti-apoptotic molecules Bcl-2. Studies have shown that CD19 CAR-T cells expressing IL-7R have a better ability to expand in vivo and resist apoptosis, thus improving the persistence of CAR-T cells in vivo (65).
IL-15 can regulate the homeostatic proliferation of memory CD8 + T cells, and can improve the lytic ability of memory CD8 + T cells by increasing the expression of perforin, granzyme B and IFN-g. The survival of memory CD8 + T cells can also be promoted by increasing the expression of the anti-apoptotic molecules Bcl-2 and Bcl-X L as well as that of the costimulatory molecule 4-1BB (17). The addition of a IL-15 domain into the CAR improved the survival rate of GD2 CAR-T cells and prolong antitumor activity [bib_ref] Eradication of neuroblastoma by T cells redirected with an optimized Gd2-specific chimeric..., Chen [/bib_ref].
Recently, it was found that culturing naive CD19 CAR-T cells with IL-7 and IL-15 promoted their differentiation into CAR-TSCM cells and showed antitumor activity in vitro and in mouse models [bib_ref] Efficient derivation of chimeric-antigen receptor-modified T cells, Kranz [/bib_ref].
IL-21 can not only direct CD8+ T cells to express L-selectin during secondary stimulation to promote the secondary proliferation of CAR-T cells, but also inhibit Treg cells [bib_ref] Immunotherapy of non-hodgkin's lymphoma with a defined ratio of Cd8+ and Cd4+..., Kim-Schulze [/bib_ref].
The combined expression of IL-15 and IL-21 maintained the expression of T cytokine 1 (TCF-1), a transcription factor essential for T cell development and survival. Research shows that Gpc3 CAR-T cells coexpressing IL-15 and IL-21 enriched poorly differentiated T cells pools, exhibited the strongest peak amplification and persistence in vivo, and mediated increased tumor control and survival in hepatocellular carcinoma (HCC) tumor-bearing mice compared to cytokines alone or controls (69).The above studies show that the expression of IL-7, IL-15 and IL-21 can promote the expansion of memory CD8+ cells and improve their antitumor activity.
## Medicine
Many small molecule drugs can enhance the effect of T cell therapy by regulating TCR, cytokines, costimulatory ions and growth factor receptors to alter T cell differentiation. For example, the mTOR inhibitor rapamycin can promote the formation of memory CD8+ T cells by regulating the expression of the transcription factors T-bet and eomesodermin, thereby enhancing antitumor function [bib_ref] The mtor kinase determines effector versus memory Cd8+ T cell fate by..., Rao [/bib_ref]. In addition, metformin, a drug forn anti-type 2 diabetes drug, can restore CD8+ memory cells and enhance survival in TRAF6-deficient mice by regulating AMPK-activation and mitochondrial fatty acid oxidation (FAO) [bib_ref] Enhancing Cd8 T-cell memory by modulating fatty acid metabolism, Pearce [/bib_ref]. Besides, immunomodulatory drugs (IMiDs; lenalidomide and pomalieradomide), which bind to the bispecific Tcell engager molecule AMG 701, can not only induce T-celldependent cytotoxicity (TDCC) in multiple myeloma (MM) cells, but also improve the proportion of TSCM cells in vitro and have longlasting antitumor effects in MM mice [bib_ref] The immunomodulatory drugs lenalidomide and pomalidomide enhance the potency of amg 701..., Cho [/bib_ref]. Furthermore, GSK3b inhibitors can promote the self-renewal of memory T cells through upregulation of the Wnt/b-catenin pathway (60). A small molecule inhibitor of lactate dehydrogenase (LDH) can inhibit aerobic glycolysis and maintain a metabolically quiescent state, and can inhibit CD8+ T cell depletion by regulating the mRNA expression of members of the NR4A family of nuclear receptors, as well as Prdm1 and Xbp1, and can promote TSCM cell production and produce powerful antitumor effects in collaboration with IL-21.
These studies suggested that small-molecule drugs can regulate the proliferation and differentiation of memory T cells through different mechanisms. Therefore, these small-molecule drugs can be used in combination with CAR-T cell therapy to prolong the persistence of CAR-T cells in vivo.
## Proportion of cd4+ t cells and cd8+ t cells in injected car-t cells
The ratio of CD4+ and CD8+ T cells is highly variable in patients with B-ALL and NHL, and studies have shown that the injection a 1:1 ratio of CD4 to CD8 CAR-T cells can ameliorate this phenomenon, resulting in predictable polyclonal proliferation of CD4 + and CD8 + T cells in patients, thereby increasing the persistence of CAR-T cells (15, 16).
## Reprogramming terminally differentiated t cells
In response to the constant stimulation of tumor cells, CAR-T cells injected in vivo continued to differentiate into tumor-infiltrating lymphocytes (TILs), going down the terminal differentiation a path of terminal differentiation and no longer providing sustained and effective antitumor effects. Therefore, the persistence of CAR-T cells could be improved by modifying the terminal differentiation status of CAR-T cells using gene editing. Studies have shown that TILs can be forced into induced pluripotent stem (IPS) cells capable of maintaining the variables (V), diversity (D), and junction (J) rearrangement region of the TCR chain via expression of the SOX2, OCT4, MYC and KLF4 transcription factors. However, this approach is inefficient, and thus a method to reverse terminal differentiation by forcing late differentiated terminal effector T (TEF) cells to express TN and TSCM cell related transcription factors through direct reprogramming has also been developed. Studies have used MEK1/2 inhibition (MEKi) to reprogram CD8+ T cells into TSCM cells, so that they can self-renew, longer existence in vivo, and stronger antitumor effect [bib_ref] Mek inhibition reprograms Cd8 T lymphocytes into memory stem cells with potent..., Verma [/bib_ref]. It has also been reported that the NOTCH-FOXM1 axis can be reprogrammed by mitochondrial metabolism to transform traditional human CAR-T cells into TSCM-like CAR-T cells, thus increasing longevity and achieving greater antitumor potential in vivo (75).
## Shortening the culture time of car-t cells
Currently, most T cell engineering regimens typically amplify T cells in vitro for 9 to 14 days. CD19 CAR-T cells show less differentiation and improved effector function when harvested from cultures at earlier time points (day 3 or day 5) than at later time points (day 9) in vitro. The results of studies in mouse models have also indicated that CAR-T cells that are cultured for a shorter time have better therapeutic potential [bib_ref] Enhancement of the in vivo persistence and antitumor efficacy of Cd19 chimeric..., Ghassemi [/bib_ref].
## Delaying car-t cell senescence by preventing telomere loss
Transient addition of a modified mRNA encoding telomerase reverse transcriptase to the CAR structure of CD19 CAR-T cells increases telomerase activity in these cells. Compared to conventional CD19 CAR-T cells, these cells have increased proliferation, enhanced persistence and improved antitumor ability (21).
Therefore, there are many ways to promote the maintenance of CAR-T cells: adding specific cytokines such as IL-5, IL-7, and IL-21 to CAR-T cell therapy or adding small-molecule drugs; reprogramming terminally differentiated CAR-T cells; shortening the culture time of CAR-T cells in vitro; and preventing the loss of telomerase in CAR-T cells to slow aging. In the future, we can continue to explore ways to help CAR-T cells differentiate into TSCM cells with these four strategies.
4 Improving the TME to increase CAR-T cell persistence CAR-T cell therapy has shown remarkable therapeutic efficacy against leukemia and lymphoma. However, over time, T cell exhaustion appears in the TME. This is related to the overexpression of inhibitory molecules such as PD-1, TIM-3, CTLA-4 and LAG-3 by dysfunctional T cells, as well as cytokines in the TME.
## Overcoming checkpoint inhibition to increase car-t cell persistence
PD-1, LAG3 and CTLA-4 are common immune checkpoint molecules in the TME. There are some common ways to inhibit these molecules. The first way is to use checkpoint inhibitors (nivolumab, atezolizumab and pembrolizumab) combined with CAR-T cells [bib_ref] End of phase 1 results from zuma-6: Axicabtagene ciloleucel (Axi-cel) in combination..., Jacobson [/bib_ref] [bib_ref] Disruption of ctla-4 expression on peripheral blood Cd8 + T cell enhances..., Zhang [/bib_ref] [bib_ref] Efficacy of programmed cell death 1 inhibitor maintenance therapy after combined treatment..., Mu [/bib_ref]. The second way is to construct CAR-T cells that can directly express immune checkpoint inhibitors (PD-1/ CD28 CD19 CAR-T cells) [bib_ref] Cd19 car-T expressing pd-1/Cd28 chimeric switch receptor as a salvage therapy for..., Blaeschke [/bib_ref]. The third way is to utilize gene editing technology (CRISPR-Cas9, TALEN) to directly knock out immune checkpoint related genes [bib_ref] Disruption of ctla-4 expression on peripheral blood Cd8 + T cell enhances..., Zhang [/bib_ref] [bib_ref] A versatile system for rapid multiplex genome-edited car T cell generation, Ren [/bib_ref] [bib_ref] Nucleofection with plasmid DNA for Crispr/Cas9-mediated inactivation of programmed cell death protein..., Hu [/bib_ref]. All of these approaches have demonstrated better T cell proliferation, cytokine production, killing capabilities and persistence in vitro than CAR-T cells alone.
In addition to the use of immune checkpoint inhibitors and the construction of novel CAR-T cells by knockout of immune checkpoint genes using gene editing methods, the expression of immune checkpoint molecules can also be regulated by transcription factors (BATF, NFAT, AP-1, and downstream molecules).
In the TME, CAR-T cells overexpressing BATF have increased expansion capacity and cytotoxicity, produce more cytokines and granzymes, prevent T cell depletion and reduce the secretion of the depletion-related transcription factor thymocyte selection-associated HMG BOX (TOX) [bib_ref] Batf and Irf4 cooperate to counter exhaustion in tumor-infiltrating car T cells, Seo [/bib_ref]. The high-mobility group (HMG)-box transcription factors (TOX and TOX2) and the NR4A family of orphan nuclear receptors are downstream targets of the transcription factor NFAT. Inhibiting the expression of these cytokines increases the expression of cytokines in CAR-TILs while decreasing the expression of inhibitory receptors. It also has the advantages of inhibiting tumor growth and prolonging the survival of tumor-bearing mice [bib_ref] Eradication of neuroblastoma by T cells redirected with an optimized Gd2-specific chimeric..., Chen [/bib_ref] [bib_ref] Tox and Tox2 transcription factors cooperate with Nr4a transcription factors to impose..., Seo [/bib_ref]. Deficiency of c-Jun transcription factors of the AP-1 family is associated with T cell depletion. Therefore, CAR-T cells overexpressing c-Jun show enhanced expansion potential, reduced terminal differentiation, and improved antitumor efficacy (57).
## Utilizing cytokine signaling to increase car-t cell persistence
IL-1 is a major proinflammatory cytokine in the TME and is involved in the development and invasion of several tumors (102). In chronic myeloid leukemia (CML), recombinant antibodies targeting IL-1 receptor antagonist (IL-1RA) and IL-1 receptor accessory protein (IL1RAP) can block IL-1 signaling in CML leukemia stem cells (LSCs) and inhibit their growth (85, 86). The combination of IL-1 signaling blockade with a tyrosine kinase inhibitor (TKI) significantly inhibited LSCs growth compared to the TKI alone. In acute myeloid leukemia (AML) and CML, blocking IL-1 signaling (for example, with anti-IL-1RA or anti-IL1RAP antibodies, IRAK1/4 inhibitors, IL-11/4 inhibitors, AP antibodies) eliminates LSCs in the TME, thereby preventing recurrence in patients (85-88).
IL-4 can induce apoptosis of AML cells in vitro and in vivo through Caspase-3 activation and STAT6 phosphorylation, and can also mediate P53-dependent apoptosis via IL-4 on LICs through the endogenous CyPG-PPARg axis, which is a negative regulator of normal AML cells [bib_ref] Interleukin 4 induces apoptosis of acute myeloid leukemia cells in a Stat6-dependent..., Pena-Martinez [/bib_ref] [bib_ref] Bisphenol a triggers the malignancy of acute myeloid leukemia cells Via regulation..., Zhang [/bib_ref].
IL-6 family cytokines are defined as those that use the common signaling receptor subunit glycoprotein 130 kDa (gp130), which is associated BCR/ABL activity in the TME (92). Disrupting IL-6 paracrine signaling of can inhibit the activity of BCR/ABL, and IL-6 inhibitors combined with PD-1 inhibitors can have an antitumor effect, thus contributing to the clearance of cancer cells in the CML TME [bib_ref] Il-6 controls leukemic multipotent progenitor cell fate and contributes to chronic myelogenous..., Reynaud [/bib_ref]. In CAR-T therapy, IL-6/STAT3 blocked reduced CAR-T cell proliferation [bib_ref] Interleukin-23 engineering improves car T cell function in solid tumors, Ma [/bib_ref].
IL-10 is a V-shaped homodimer that can be generated by multiple cell types, and IL-10 has important effects on the TME (103). In the CLL TME, IL-10 is highly expressed, which disrupts the synergy between NFAT and AP-1 through IL-10R-STAT3 signaling, thereby increasing the expression of PD-1 on CD8+ T cells and preventing T cells from playing an antitumor role [bib_ref] Il-10 rescues cll survival through repolarization of inflammatory nurse-like cells, Domagala [/bib_ref] [bib_ref] Interleukin-10 receptor signaling promotes the maintenance of a pd-1 tcf-1 Cd8 T..., Hanna [/bib_ref]. CAR-T cells targeting the interleukin-10 receptor (IL-10R) also have strong tumor cytotoxicity in AML treatment [bib_ref] The opposite effects of il-15 and il-21 on cll b cells correlate..., Chen [/bib_ref].
IL-12 is a proinflammatory cytokine with strong tumor inhibitory activity that not only stimulates T cells to secrete IFN-g to enhance the cytotoxicity of CAR-T cells, but also generates resistance to Treg cells effects, reshaping the TME in an IFN-g-dependent manner [bib_ref] Reversing tumor immune suppression with intratumoral il-12: Activation of tumor-associated T Effector/Memory..., Kilinc [/bib_ref]. Autocrine IL-12 with CAR-T cells in the TME significantly improved the killing effect of CAR-T cells and achieved long-lasting antitumor effects (49-51).
In the TME, studies have shown that the secretion of IL-21 can increase memory and affect CD8+ T cells, inhibit Treg cells, and can cooperate with immune checkpoint inhibitors to improve the depletion of T cells [bib_ref] Immunotherapy of non-hodgkin's lymphoma with a defined ratio of Cd8+ and Cd4+..., Kim-Schulze [/bib_ref] [bib_ref] Targeting tumors with il-21 reshapes the tumor microenvironment by proliferating pd-1inttim-3-Cd8+ T..., Deng [/bib_ref].
TGF-b is a common cytokine in the TME and is secreted by malignant cells and their stroma. It promotes tumor growth and metastasis, and potently inhibits APCs and the growth and effector function of T cells and NK cells. TGF-b interacts with TGF-b receptor I (TGF-bRI) or TGF-bII to form type I and type II receptor dimers, resulting in phosphorylation of TGF-b RI and endowing it with the ability to phosphorylate Smad2 and Smad3, which then enter the nucleus to interact with transcription factors. On the one hand, TGFb can regulate cell growth and inhibit tumor-specific cellular immunity, on the other hand, TGF-bcan inhibit the effects of perforin, granzyme A, granzyme B, Fas ligand and IFN-g, and inhibit the cytotoxicity of CTLs (104). In Hodgkin's lymphoma, depletion of endogenous TGF-negative EBV-positive Hodgkin's CAR-T cells using the CRISPR/Cas9 technique reduced induced Treg transformation and prevented CAR-T cell depletion [bib_ref] Tgf-B inhibition Via crispr promotes the long-term efficacy of car T cells..., Tang [/bib_ref]. The specific kinase inhibitor SD-208 can also be administered to block TGF-b-receptor signaling, synergizing the antitumor effects in the TME with CAR-T cells in the TME (100) [fig_ref] TABLE 1: The role of cytokines in tumor microenvironment [/fig_ref].
In summary, there are many factors in the TME that inhibit the function of CAR-T cells, such as immunosuppressive molecules, inhibitory cytokines, and Treg cells. Ways to optimize the TME for via these factors and improve the persistence of CAR-T cells in vivo.
## Concluding remarks
This review focuses on ways to increase the persistence of CAR-T cells in vivo by modifying the structure of CAR-T cells, increasing the proportion of memory CAR-T cells, and improving the TME.
Although memory T cells are required in CAR-T cell therapy to increase their persistence in vivo, terminally differentiated effector T cells are still required to kill tumor cells. Therefore, how to increase the proportion of stem cells while ensuring the proportion of effector T cells needed to kill tumor is a problem worth thinking about.
In the TME, the expression of inhibitory immune checkpoint proteins on the surface of CAR-T cells was increased, thereby affecting the function of CAR-T cells. The current research focuses on the suppression of immune checkpoint proteins, including the use of immune checkpoint inhibitors, the construction of suppressantproducing CAR-T cells and the construction of CAR-T cells that knock out the genes related to the immune checkpoint proteins. The first two methods are to directly inhibit immune checkpoint proteins, while the latter method is to directly block the expression of immune checkpoint molecules at the gene level. The knockout of related genes using gene editing technology (CRISPR-Cas9, TALEN) can fundamentally reduce the expression of immunosuppressive molecules, which will be a popular method in the future immunotherapy.
Gene editing technology can not only knock out the genes associated with immune checkpoints, but also the genes associated with inhibitory cytokine secretion. It can also insert genes regulated to memory T cells. Therefore, the structure of CARs can be modified by this technology in many aspects to enhance the persistence of CAR-T structure. However, the stability and security of these domains should also be considered.
In addition, the persistence of CAR-T cells in vivo is not always better. For example, CD19 CAR-T cells can target both malignant cells and non-normal B cells. If the persistence of CAR-T cells is increased, it will have adverse effects on normal B cells in the patient. While attention should be paid to the persistence of CAR-T cells in vivo, treatment side effects (such as CRS and neurotoxicity) should also be paid. In the future, there is an urgent need to observe the side effects and safety of novel CAR-T cell therapies in clinical trials to determine whether they can be safely applied to clinical treatment.
# Author contributions
## Conflict of interest
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
## Publisher's note
All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.
## Cytokines
Application in tumor microenvironment IL-1 Blocking IL-1 signaling in CML and AML can inhibit tumor cell proliferation through ADCC-mediated kill sensitivity IL-4 Inducing apoptosis of AML cells in vitro and in vivo through caspase-3 activation, STAT6 phosphorylation and endogenous CyPGs-PPARg axis IL-6 Playing an antitumor role by destroying the paracrine signaling and combining with PD-1 inhibitors IL-10 Blocking IL-10R-STAT3 signaling can regulate the synergy between NFAT and AP-1 IL-12 Secreting IFN-g to enhance the cytotoxicity of CAR-T cells, but also generates resistance to Treg cells
## Il-21
Promoting the production of memory and effector CD8 + T cells, inhibiting the growth of Treg cells, and cooperating with immune checkpoint inhibitors to improve the depletion phenomenon in tumor microenvironment TGF-ß Blocking TGF-b-receptor II signaling can induced Treg transformation and preventing CAR-T cell depletion generation car T cells. J Immunother Cancer (2021) 9(10):e003354. doi: 10.1136/jitc-2021-003354
[fig] Funding: This work was supported by the National Natural Science Foundation of China (82270183, 81873440 and 81873444), Key R & D plan of Hubei Province (No.2020BCB021 & 2020BCB043), The excellent young science foundation project of Tongji Hospital (No.2020YQ0012). [/fig]
[table] TABLE 1: The role of cytokines in tumor microenvironment. [/table]
|
Evolution of global development cooperation: An analysis of aid flows with hierarchical stochastic block models
Despite considerable scholarly attention on the institutional and normative aspects of development cooperation, its longitudinal dynamics unfolding at the global level have rarely been investigated. Focusing on aid, we examine the evolving global structure of development cooperation induced by aid flows in its entirety. Representing annual aid flows between donors and recipients from 1970 to 2013 as a series of networks, we apply hierarchical stochastic block models to extensive aid-flow data that cover not only the aid behavior of the major OECD donors but also that of other emerging donors, including China. Despite a considerable degree of external expansion and internal diversification of aid relations over the years, the analysis has uncovered a temporally persistent structure of aid networks. The latter comprises, on the one hand, a limited number of major donors with far-reaching resources and, on the other hand, a large number of mostly poor but globally well-connected recipients. The results cast doubt on the efficacy of recurrent efforts for "aid reform" in substantially changing the global aid flow pattern.conducts such an investigation to fill this research gap. Focusing on one of the most traditional policy instruments for development cooperation, that is, aid, the study aims to elucidate the evolving global structure of development cooperation induced by aid flows among various international actors (e.g., states, international organizations, and major public and private funds) over the past decades. For this purpose, we employ cutting-edge methods, including stochastic block models (SBMs), which have been developed in complex network science [6], and we apply these analytical tools to extensive aid-flow data that cover not only the aid behavior of Organisation for Economic Co-operation and Development (OECD) countries but also that of other emerging donors, including China.Network science itself has made significant inroads into IR in recent years [7,8]. Different aspects of IR have been represented as networks of co-affiliation to intergovernmental-organization [9-11], democratic peace [10], trade [10], voting behavior in the United Nations[12], and bilateral cooperation agreements [12], among others[13,14]. Unfortunately, the structure of global development cooperation, including aid, has received only scant attention in this context. There are indeed a few notable studies that employ network analysis to study aid[15][16][17][18][19]. Many of these studies calculate node-level local measures (e.g., degree centrality and eigenvector centrality) from given network representations, and then input these quantities into regression equations that largely model state-level behavior and/or state-to-state dyadic relationships. Although such a methodological orientation is fully consistent with the dominant literature on aid relations and aid effectiveness in IR[20][21][22][23][24], the aid networks represented therein contain far more information than these local quantities can capture. By fully leveraging this information, it is possible to rigorously measure and analyze the entire structure of aid relations unfolding at the systemic level. By performing such measurement and analysis, we aim to demonstrate the still-unexplored potential of network science for studying global development cooperation.
# Introduction
Over the past 60 years, international society has sustained an increasingly institutionalized pattern of development cooperation. A countless number of programs, funds, and institutions have been launched for promoting economic and social development in different countries and regions; various landmark policy documents (e.g., Berg Report, Comprehensive Development Framework, Paris Declaration) have been adopted for more efficient and effective development efforts; and an ever-expanding array of goals and targets, including the Millennium Development Goals and Sustainable Development Goals, have been globally embraced for channeling these efforts to all countries and all people in need. Not surprisingly, these conspicuous aspects of development cooperation have attracted the attention of international relations (IR) researchers, especially those working in the liberal (e.g., international regime) and constructivist (e.g., aid norm diffusion) traditions [bib_ref] The development aid regime at fifty: policy challenges inside and out, Hook [/bib_ref].
However, the actual dynamics of cooperation unfolding at the global level have rarely been investigated. These dynamics have emerged and evolved over years under successive policy and normative initiatives, such as those mentioned above. Our study
# Materials and methods
## Foreign aid networks
We begin by building foreign aid networks using open datasets provided by the OECDand AidData [bib_ref] More dollars than sense: Refining our knowledge of development finance using, Tierney [/bib_ref] [bib_ref] Aid, China, and growth: Evidence from a new global development finance dataset, Dreher [/bib_ref]. Both datasets track foreign aid provided by both bilateral and multilateral donors.
The OECD dataset on foreign aid is based on reports provided by donor countries, which include the names of donors and recipients, years, and the amount of aid. We focus on official development assistance (ODA) commitments and do not include other financial flows in our analysis. ODA needs to satisfy several criteria for public development aid, which are set by the Development Assistance Committee (DAC) of the OECD.
The OECD data do not cover several emerging donors (e.g., China), but AidData collects a similar set of information (donors and recipients, years, and amount of aid) for both emerging and traditional donor countries [bib_ref] More dollars than sense: Refining our knowledge of development finance using, Tierney [/bib_ref]. Therefore, we also use AidData for the donors that are not included in the OECD data but appear in AidData. Data from AidData are based on not only the official reports of donors to OECD but also their original data collection. In particular, the information about foreign aid provided by China is separately collected based on various information sources, including media reports [bib_ref] Aid, China, and growth: Evidence from a new global development finance dataset, Dreher [/bib_ref]. Although China hardly follows the DAC criteria, we include the financial flow that have a similar nature to ODA, which are labeled as "ODA-like" by AidData. Chinese "ODA-like" assistance as a donor is included only after 2000. In addition, while data for DAC donors are reported and recorded from an earlier period, it is likely that the quality of data for non-DAC donors other than China has also improved. This suggests that data quality improves over time, which may affect the evolution of the aid networks. Therefore, while we must accept the possibility that our analysis may be biased due to data limitations, OECD.stat and AidData are still the sources that many studies rely on, and the results of our analysis maintain some validity.
We construct a weighted directed network by aggregating the amount of aid in the dataset for each year in 1970-2013. The networks, which we call foreign aid networks (FANs), represent the yearly financial flow between the actors, including governments, intergovernmental organizations (e.g., the World Bank), and private foundations (e.g., the Bill & Melinda Gates Foundation). Specifically, we define the weight of an edge from node i to j as the normalized quantity f ij /s out i s in j , where f ij is the amount of aid i gave to j, s out i is the total amount of aid i gave, and s in j is the total amount of aid j received in the year. This normalization is introduced because an analysis of the network without the normalization yields results that reflect only the financial flows between a few major players and largely ignore many others, as the amount of aid flow f ij is distributed over a wide range of magnitude. The weight of an edge from a donor to a recipient is equal to one, the maximum, when the donor provided aid only to the recipient and the recipient obtained aid only from the donor in the year. Even if the amount of aid from a donor to a recipient was large, the edge weight would be small when the donor provided (the recipient obtained) a large amount of aid to other recipients (from other donors) in the year.
## Stochastic block models
In addition to the local network attributes, such as average degree, we identify larger-scale structures in networks by dividing nodes into communities or blocks. While there are diverse ways to identify communities in networks [bib_ref] Community detection in graphs, Fortunato [/bib_ref] , one of the most principled approaches to performing this task is inferring the parameters for SBMs; this is arguably one of the simplest generative models incorporating the idea of groups of nodes. The SBM is a simple generative model for random networks, where nodes are partitioned into B blocks b = {b i } and the edges between the blocks are generated according to the parameter θ, where θ rs is the probability of forming edges between block r and s. Given these constraints, the edges are then placed randomly. Hence, nodes that belong to the same block possess the same probability of being connected with other nodes of the network.
It is possible to calculate the probability of generating a network A by the SBM with b and θ, P (A|θ, b), in an analytic form. Therefore, if we observe an empirical network A, the posterior probability that it was generated by a given partition b is obtained by
[formula] P (b|A) = θ P (A|θ, b)P (θ, b) P (A) ,(1) [/formula]
where P (θ, b) is the prior probability of the model parameters and
[formula] P (A) ≡ θ,b P (A|θ, b)P (θ, b) [/formula]
is the total probability of observing the empirical network A. By finding a network partition that maximizes Eq. (1), we can infer the most likely partitions of empirical networks or sample the partitions from the posterior distribution. Note that the block structure does not always correspond to communities. On the one hand, communities are sets of nodes that are more densely connected inside of communities compared with outside. On the other hand, the SBM covers a broader class of structures, such as core-periphery structures, in which nodes in core blocks are likely to connect with nodes in both core and periphery blocks but nodes in periphery blocks tend to connect only to nodes in core blocks. As FANs are weighted networks, we use the extended version of the SBM, which treats edge weights as covariates sampled from some distribution conditioned on the node partition [bib_ref] Learning latent block structure in weighted networks, Aicher [/bib_ref] [bib_ref] Nonparametric weighted stochastic block models, Peixoto [/bib_ref]. In other words, the weight of all the edges follows the same probability distribution if the edges connect the same directed pair of blocks (e.g., from block r to block s) We try a couple of probability distributions, including exponential and normal distributions, and find that the log-normal distribution best fits our dataset. This is because blocks in FANs inferred with normal and exponential distribution tend to include only a single or a few nodes; thus we could not find any meaningful division of FANs. Therefore, the results for the log-normal model are shown below.
In addition, we use a hierarchical version of the model, which replaces the prior by a nested sequence of priors and hyperpriors, as described in [bib_ref] Nonparametric weighted stochastic block models, Peixoto [/bib_ref]. In other words, the nodes are partitioned into blocks, and each of these blocks is partitioned into higher-order blocks, and so on. Without allowing the hierarchical structure, the SBM could fail to find small blocks compared with the size of networks.
Partition and other parameters, including the number of blocks and layers in a hierarchy, are determined by optimizing minimum description length (MDL), which measures the amount of information required to describe the data. A smaller MDL means that we need less additional information to describe the network and indicates that the assumption of block structure is suitable to explain the network. The inferences of the block structures are conducted by graph-tool, which is one of the standard libraries for network analysis. The source code for the analysis is available online to enable reproducibility of the results.
# Results
## Local network characteristics
Before moving to more global characteristics (i.e., block structure), we first examine local network characteristics, such as the number of nodes and average degree. [fig_ref] Fig 1: Time evolution of the local network quantities of FANs [/fig_ref] shows the growing trend of the FANs during the observation period. As shown in [fig_ref] Fig 1: Time evolution of the local network quantities of FANs [/fig_ref] , the numbers of donors and recipients steadily increase from the 1970s to the 1990s, where donors and recipients are defined as the nodes with outgoing and incoming edges, respectively. The number of recipients temporarily increased after the end of the Cold War and then quickly declined. The increase was due to the birth of new states created by the dissolution of the Soviet Union and Yugoslavia with the end of the Cold War, and the subsequent decline came with the conversion of countries such as Korea and the oil-producing countries in the Middle East from recipients to donors. Thereafter, the number of recipients more or less stabilizes, while the number of donors continues to increase. Note that most nodes are either donors or recipients. Thus, the FANs are close to bipartite networks. However, there is a small fraction of nodes with both incoming and outgoing edges, which are represented as donor-cum-recipients in [fig_ref] Fig 1: Time evolution of the local network quantities of FANs [/fig_ref]. [fig_ref] Fig 1: Time evolution of the local network quantities of FANs [/fig_ref] shows the average out-degree of the donors and the average in-degree of the recipients, which again shows the increasing trend over the entire observation period. Moreover, the average out-degree of the donors is always much larger than the average in-degree of the recipients, which indicates a small number of donors provide aid to a large number of recipients (see also S1 [fig_ref] Fig 7: The number of LDCs in each stable block in [/fig_ref].
## Time evolution of block structures
Next, we examine the block structures of the FANs using hierarchical SBMs in order to study larger-scale network characteristics. [fig_ref] Fig 2: Block structure of FANs [/fig_ref] shows samples of obtained block structures of foreign aid networks in 1970, 1990, and 2010. For each FAN, we take 100 samples from the posterior probability distribution and keep only stable blocks that comprise the nodes that belong to the block in at least 95% of the samples. Nodes in each stable block are listed in S1, S2, and S3 Tables . The donors and recipients are split into different blocks, indicating that the inference of the block structure works as expected. We also find finer structures within the donor blocks and recipient blocks. Note that the division between donor and recipient blocks shows that the SBM inference correctly captures the fact that FANs are close to bipartite networks. This is not the case for the inference of communities that are sets of nodes more densely connected internally as compared to externally.
The division between the donor blocks and recipient blocks is also observed in and 2010 have six and nine blocks, respectively. While the recipients are mostly divided into two blocks for these networks, the number of donor blocks increases from two to seven, indicating increasing complexity of the FANs, largely due to the divisions of the donor blocks. We also confirm the increase in the number of blocks in [fig_ref] Fig 4: The number of blocks [/fig_ref] which shows the number of blocks averaged over 100 samples taken from the posterior probability distribution.
We also find that the time evolution of the donor block structures is different from those of the recipients. [fig_ref] Fig 5: The alluvial diagram of stable blocks [/fig_ref] shows the change in the membership of blocks between different time points (alluvial diagram generated by. In most of the time points, the two bottom blocks are recipients while the top blocks are donors. Although the two or three recipient blocks occasionally change their membership (e.g., in 1985 and 1995), recipient blocks are more stable than the donor blocks (see also S2 . Meanwhile, the donor blocks frequently split and change their members, which largely contributes to the increase in the number of blocks in the observation period. Despite these unsettling dynamics, we note that the most resourceful block in terms of total aid commitment has almost always had a stable set of powerful donors at its core, including the U.S., Japan, and Germany (see [fig_ref] Fig 3: Aid flow networks of blocks [/fig_ref].
## Characterization of blocks
We also analyze the characteristics of each block in the FANs. [fig_ref] Fig 6: Regions or types in FAN stable blocks in [/fig_ref] shows the regions or types of actors in the blocks in 1970, 1990, and 2010, indicating that each block consists of similar nodes in terms of the region or type. For example, most sub-Saharan African countries and Oceanic countries belong to the same block (i.e., Blocks 1 and 2, respectively). This means that the nodes (especially recipients) in the same region tend to have similar connection patterns. Moreover, multilateral donors tend to form their own blocks: Block 4 in 1990 and Blocks 3-5 in 2010. This finding means that multilateral donors have different patterns of connections to those of the bilateral donors and there is a variety of such patterns, even among multilateral donors.
Next, we examine the relationship between the poverty of the nodes and the block structure of the FANs. [fig_ref] Fig 7: The number of LDCs in each stable block in [/fig_ref] shows the number of least developed countries (LDCs) in each block in 1990 and 2010 (LDC data are not available for 1970). As shown in the figure, most LDCs are included in the largest block (Block 0), indicating that LDCs have received a similar pattern of the development aid.
# Discussion and conclusions
We constructed FANs for 1970-2013 and studied their local and global properties to examine the evolving structure of global development cooperation among numerous international actors. The results of this largely exploratory investigation corroborate several important observations and insights. First, the global structure of aid flows, having been continuously incorporating new participants, especially on the donor side, has steadily expanded and become more complex over time (see [fig_ref] Fig 4: The number of blocks [/fig_ref]. This largely confirms the prevailing view of aid relations as an ever-expanding domain of IR [bib_ref] The development aid regime at fifty: policy challenges inside and out, Hook [/bib_ref]. Second, and despite the former conclusion, the global aid-flow structure, at its core, has largely stable components (see [fig_ref] Fig 3: Aid flow networks of blocks [/fig_ref] : namely, a limited number of lead donors, such as the U.S., Japan, and the EU, which are financially connected, to a varying degree, to an entire array of recipient groups (and even some donor groups); and a large number of poor recipients, including most sub-Saharan African countries, which are well connected to a broad array of donor groups, especially the lead donor group. This observation may be somewhat surprising given recurrent efforts for "aid reform" during the past decades, which have been aimed at bringing meaningful changes to global aid flow patterns. Of course, changes have occurred, most notably, the expanding roster of multilateral and bilateral donors with accompanying diversification in aid-flow channels. In comparison with the long-established lead donors, however, the roles of these other donors remain complementary, if not peripheral.
Third, for a large part of the study period, especially since the 1980s, we observed the recurrent formation of small but distinct donor-cum-recipient blocks comprising emerging countries, such as Kuwait, Korea, and Turkey. Since the 2000s, unlike the previous donor-cum-recipient blocks, the block tends to comprise only China (although it sometimes comprises other emergent donors, e.g., Turkey). Furthermore, China's block, while being deeply embedded in the global aid structure, is highly distinct and persistent. In comparison with other leading donors, however, its presence remains limited. This is probably because the AidData dataset covers the aid given by China There are numerous limitations of our study, which we will continue to work on in the immediate future. First, it is necessary to iterate the present line of analysis in different modeling specifications to establish the robustness of our findings. Data availability, as just mentioned, is another obvious challenge. Fortunately, in this regard, AidData has recently released a new version of the China dataset, which extends its temporal coverage beyond 2013 to 2017. Incorporating a more recent part of China's aid behavior into the analysis might substantially change its standing in the derived block configurations from the modest contribution that we have seen so far.
Finally, the present analysis is exploratory as well as descriptive. We believe that such an endeavor is an essential step to follow, but we also recognize the need to go beyond it. A useful next step would be to establish some theoretical baseline for understanding and evaluating the obtained results, since currently, there are not obvious conclusions about the extent to which a given global structure of aid relations is effective in attaining coordination among numerous donors, aligning development efforts between donors and recipients, or delivering needed resources to intended beneficiaries, etc. We aim to expand our inquiry into these more challenging domains by again employing various modeling devices provided by network science (e.g., exponential random graph models).
[fig] Fig 1: Time evolution of the local network quantities of FANs. (a) The numbers of donors, recipients, and nodes with both incoming and outgoing edges (donor-cum-recipients). (b) The average out-degree of donors, and the average in-degree of recipients. [/fig]
[fig] Fig 2: Block structure of FANs. The block structure of FANs in (a) 1970, (b) 1990, and (c) 2010. Circles represent nodes whose colors indicate their blocks. The arrows between nodes represent edges whose colors indicate the edge weights. Nodes in each block are listed in [/fig]
[fig] Fig 3: Aid flow networks of blocks. A coarse-grained representation of the FANs in (a) 1970, (b) 1990, and (c) 2010, where the nodes in this figure represent the blocks. The edges show the aid flow between the nodes in each block. The aid flow from block b 1 to block b 2 is the total aid given by actors in b 1 to actors in b 2 . The size of the nodes depends on the total inward and outward aid flow summed over all members of the blocks. The labels show the names of the actors with the three largest inward and outward aid flows in each block. The brightness of the blocks indicates the fractions of outward aid flow, in other words, bright blocks mainly contain donors. The edge width shows the volume of aid flow. [/fig]
[fig] Fig 4: The number of blocks. Averaged over 100 samples in posterior probability distribution sampling. Error bars show the standard deviation. [/fig]
[fig] Fig 5: The alluvial diagram of stable blocks. The diagram shows the change of memberships of the stable blocks from 1970 to 2010 every 5 years (from left to right). The flow from block b 1 in year t to block b 2 in year t + 5 represents a situation in which actors who had belonged to b 1 in year t belonged to b 2 in year t + 5. The height of each block indicates the number of nodes (actors) in the block. Colors are used to make blocks distinguishable and do not represent any block characteristics. The blocks surrounded by black lines are the recipient blocks. [/fig]
[fig] Fig 6: Regions or types in FAN stable blocks in (a) 1970, (b) 1990, and (c) 2010. The number of actors of each region or type in each block. Governments are labeled by the regions of their countries, while IGOs and NGOs are labeled as multilateral and private, respectively. The labels are available from OECD data. The stable blocks are in descending order of size. [/fig]
[fig] Fig 7: The number of LDCs in each stable block in (a) 1990 and (b) 2010. The stable blocks are in descending order of size. only for the period from 2000 to 2013, which is largely before the launch of China's massive Belt and Road Initiative. [/fig]
[fig] Fig S2: Time evolution of the sizes of the largest and second largest stable blocks. [/fig]
[table] Table S1: List of actors in stable blocks in 1970. block ID actors 0 Afghanistan, Algeria, Botswana, Brazil, Burkina Faso, Burundi, Cambodia, Cameroon, Central African Republic, Congo, Democratic Republic of the Congo, Ethiopia, Gabon, Ghana, India, Indonesia, Kenya, Korea, Lao People's Democratic Republic, Madagascar, Mali, Morocco, Niger, Nigeria, Pakistan, Peru, Philippines, Rwanda, Senegal, Sri Lanka, Tanzania, Thailand, Tunisia, Turkey, Uganda, Viet Nam, Zambia [/table]
|
Associations of Toenail Arsenic, Cadmium, Mercury, Manganese, and Lead with Blood Pressure in the Normative Aging Study
Background: Arsenic, cadmium, mercury, and lead are associated with cardiovascular disease in epidemiologic research. These associations may be mediated by direct effects of the metals on blood pressure (BP) elevation. Manganese is associated with cardiovascular dysfunction and hypotension in occupational cohorts. oBjectives: We hypothesized that chronic arsenic, cadmium, mercury, and lead exposures elevate BP and that manganese lowers BP. Methods: We conducted a cross-sectional analysis of associations between toenail metals and BP among older men from the Normative Aging Study (n = 639), using linear regression and adjusting for potential confounders. results: An interquartile range increase in toenail arsenic was associated with higher systolic BP [0.93 mmHg; 95% confidence interval (CI): 0.25, 1.62] and pulse pressure (0.76 mmHg; 95% CI: 0.22, 1.30). Positive associations between arsenic and BP and negative associations between manganese and BP were strengthened in models adjusted for other toenail metals. conclusions: Our findings suggest associations between BP and arsenic and manganese. This may be of public health importance because of prevalence of both metal exposure and cardiovascular disease. Results should be interpreted cautiously given potential limitations of toenails as biomarkers of metal exposure. key words: arsenic, blood pressure, cadmium, epidemiology, lead, manganese, mercury, metals. Environ Health Perspect 120:98-104 (2012). http://dx.
volume 120 | number 1 | January 2012 - Environmental Health Perspectives Research Arsenic, cadmium, mercury, and lead are associated with cardiovascular disease in epidemiologic research [bib_ref] Vascular effects of chronic arsenic exposure: a review, Engel [/bib_ref] [bib_ref] The role of mercury and cadmium heavy metals in vascular disease, hypertension,..., Houston [/bib_ref] [bib_ref] Lead exposure and cardiovascular disease-a systematic review, Navas-Acien [/bib_ref]. These associations may be mediated by direct effects on blood pressure (BP) elevation. In animals, arsenic, cadmium, mercury, and lead induce hypertension [bib_ref] Lead exposure and cardiovascular disease-a systematic review, Navas-Acien [/bib_ref] [bib_ref] Kidney dysfunction and hypertension: role for cadmium, p450 and heme oxygenases?, Satarug [/bib_ref] [bib_ref] Hypertension induced by methyl mercury in rats, Wakita [/bib_ref] [bib_ref] Lifelong inorganic arsenic compounds consumption affected blood pressure in rats, Yang [/bib_ref] , and manganese causes hypotension [bib_ref] Cardiovascular toxicities upon manganese exposure, Jiang [/bib_ref]. All five metals are plausibly linked with BP, based on mechanistic and experimental data [bib_ref] The role of mercury and cadmium heavy metals in vascular disease, hypertension,..., Houston [/bib_ref] [bib_ref] Cardiovascular toxicities upon manganese exposure, Jiang [/bib_ref] [bib_ref] Lead exposure and cardiovascular disease-a systematic review, Navas-Acien [/bib_ref] [bib_ref] Metals, toxicity and oxidative stress, Valko [/bib_ref].
Epidemiologic studies show consistent associations between high arsenic exposure and BP [bib_ref] Increased prevalence of hypertension and long-term arsenic exposure, Chen [/bib_ref] [bib_ref] Hypertension and arsenic exposure in Bangladesh, Rahman [/bib_ref]. Few studies have assessed lower-level exposure. Research on cadmium and BP is inconsistent [bib_ref] Exposure to cadmium and conventional and ambulatory blood pressures in a prospective..., Staessen [/bib_ref] [bib_ref] Cadmium exposure and hypertension in the 1999-2004 National Health and Nutrition Examination..., Tellez-Plaza [/bib_ref]. Lead is a known risk factor for BP elevation [bib_ref] Lead exposure and cardiovascular disease-a systematic review, Navas-Acien [/bib_ref] , but no studies have evaluated toenail lead. High manganese exposure is associated with cardiovascular dysfunction and hypotension in occupational cohorts [bib_ref] Cardiovascular toxicities upon manganese exposure, Jiang [/bib_ref]. Research at lower exposures is inconsistent (Gonzalez-Muñoz et al. 2010; [bib_ref] Mineral factors controlling essential hypertension -a study in the Chandigarh, India population, Taneja [/bib_ref]. Studies examining mercury and BP report mixed results [bib_ref] Relationship between mercury concentration in blood, cognitive performance, and blood pressure, in..., Johansson [/bib_ref] [bib_ref] Relationship between mercury in blood and 24-h ambulatory blood pressure in Greenlanders..., Pedersen [/bib_ref] [bib_ref] Environmental mercury exposure and blood pressure among Nunavik Inuit adults, Valera [/bib_ref]. Data on low-level mercury are sparse.
We hypothesized that chronic arsenic, cadmium, mercury, and lead exposures elevate BP and that manganese lowers BP. We evaluated these hypotheses in a cohort of elderly men from the Greater Boston area.
# Materials and methods
Study population. Subjects are from the on going, longitudinal Veterans Administration Normative Aging Study (NAS) [bib_ref] The Normative Aging Study: an interdisciplinary and longitudinal study of health and..., Bell [/bib_ref]. Participants are males with no known chronic medical conditions at recruitment. They are evaluated at study visits every 3-5 years. This study has been approved by all relevant institutional review boards. All participants gave their written informed consent.
We asked NAS participants to bring toenail clippings to their study visit between January 1999 and January 2009 (n = 818 eligible NAS participants). For our analysis, we excluded NAS participants who did not bring toenail clippings (n = 165) or were missing information on BP (n = 3), race/ethnicity (n = 12), education (n = 22), alcohol intake (n = 11), age (n = 1), body mass index (BMI; n = 1), smoking history (n = 1), packyears (n = 1), or season of clinical visit (n = 1). Analyses included 639 men with toenail samples and complete information on BP and relevant covariates on at least one visit.
For each metal, we used data from first visits of the participants with complete information on the specific metal of interest, BP, and relevant covariates in regressions against BP. In contrast, when reporting results for the study population as a whole (i.e., participant characteristics, toenail metal concentrations, and correlations between toenail metals), we used data from the first study visit of the participants with donated toenails (even if some metal concentrations were missing) and complete information on BP and covariates. Many NAS participants had more than one study visit between 1999 and 2009. Sometimes, data were missing on only certain metals at a given visit. Thus, the same participant may contribute data from different visits for, for example, arsenic-BP versus cadmium-BP regressions. Hence, sample sizes sometimes differ across metal-BP regressions even after adjustment for all other toenail metals.
Physical parameters and medical history. We asked participants to fast and abstain from smoking overnight prior to their morning visit. Height and weight were measured during physical examinations. Information on smoking and medications was obtained from questionnaires and confirmed by an onsite physician. Alcohol and seafood intakes were determined using a standardized semiquantitative food frequency questionnaire [bib_ref] Reproducibility and validity of a semiquantitative food frequency questionnaire, Willett [/bib_ref].
Toenail samples. Toenails were collected from all toes, sonicated for 15 min in ~ 10 mL 1% Triton X-100 solution, rinsed with distilled deionized water, dried at 60°C in a drying oven for 24 hr, and weighed. Mercury concentrations were measured by a different assay than the other metals. A sufficient quantity of toenails was not always available for both assays. If weight was > 0.1 g, a portion of the sample was used for mercury analysis, and the rest was acid-digested and analyzed for multi-elements. Samples weighing < 0.1 g were not analyzed for mercury.
Among men otherwise eligible for inclusion in metal-BP regressions, some were excluded because of cadmium concentrations below the detection limit (DL; n = 41) and/or samples of insufficient weight to assess metal concentrations (7 for cadmium, 72 for mercury). One man had two otherwise eligible study visits during the period of toenail collection but was excluded from cadmium-BP regressions because of cadmium concentration below the DL in one visit and insufficient sample weight in the other visit. No eligible participants were excluded because of metal concentrations below the DL for arsenic, mercury, manganese, or lead, or because of samples of insufficient weight for arsenic, manganese, or lead. We found no meaningful differences in characteristics between those with sufficient and insufficient sample weight for mercury assays (data not shown).
Arsenic, cadmium, manganese, lead. Samples were dissolved in HNO 3 acid for 24 hr, diluted to 5 mL with deionized water, and analyzed by an inductively coupled plasma-mass spectrometer (Elan 6100; PerkinElmer, Norwalk, CT, USA). Analyses were conducted through external calibration with thalium, indium, and tellurium as internal standards for lead, cadmium, and manganese, and arsenic, respectively.
Quality control (QC) measures included analysis of the initial calibration verification standard (National Institute of Standard and Technology Standard Reference Material 1643e, trace elements in water; Gaithersburg, MD, USA), a 1-ng/mL mixed-element standard solution, continuous calibration standards, and a procedural blank. Certified reference material GBW 07601 was used as the QC sample; we used a large preparation (2 g/L) to monitor daily variation. Results were given as the average of five replicate measurements. Recovery of the analysis of the QC standard by this procedure was 90-110% with 95% precision.
The between-assay coefficient of variation was 3.6% for lead, 3.0% for cadmium, 3.3% for manganese, and 10% for arsenic. The DL for the analytical solution for all four metals was 0.2 × 10 -3 µg/mL. The DL for the sample depends on sample weight. Sample weight varied from 0.002 g to 0.9 g, and the DL varied from 0.001 µg/g to 0.42 µg/g (mean = 0.02 µg/g).
Mercury. We conducted assays with the Direct Mercury Analyzer 80 (Milestone Inc., Monroe, CT, USA). Samples were analyzed using a matrix-matched calibration curve created with different weights of certified reference material GBW 07601 (human hair; Institute of Geophysical and Geochemical Exploration, Langfang, Hebei Province, People's Republic of China).
QC steps included daily calibration verification of the high and low ends of the calibration curve, a procedural blank, and analysis of National Institute for Environmental Studies certified reference material 13 (human hair; Ibaraki, Japan). Mercury recovery was 90-110%, with > 90% precision.
The DL for the mercury analysis was 0.5 × 10 -3 µg. The DL for the sample varied according to sample weight. Sample weight varied from 0.01 g to 0.15 g, and the DL varied from 0.003 µg/g to 0.05 µg/g (mean = 0.018 µg/g). BP measurements. Systolic BP (SBP) and fifth-phase diastolic BP (DBP) were measured by a physician at the same visit as toenail collection. BP readings were taken in each arm (to the nearest 2 mmHg) with standard mercury sphygmomanometers with 14-cm cuffs. The means of the arm measurements were used as the SBP and DBP of the participants, and pulse pressure was calculated as SBP -DBP. BP was taken seated, immediately after a seated patient history.
Statistical methods. We examined crosssectional relations between toenail metals and BP from the same study visit using multivariable linear regression.
Antihypertensive medication use is a potential source of bias when examining associations with BP. People taking antihypertensives have controlled and likely artificially low BP. We address this by adding a constant (10 mmHg) to the SBP and DBP of participants using antihypertensives [bib_ref] Antihypertensive treatments obscure familial contributions to blood pressure variation, Cui [/bib_ref] [bib_ref] Adjusting for treatment effects in studies of quantitative traits: antihypertensive therapy and..., Tobin [/bib_ref]. To test the robustness of our results, we ran a sensitivity analysis examining associations between toenail metals and BP without adding a constant to the SBP and DBP of antihypertensive users. Results were similar when not correcting for medication use (data not shown).
We examined interactions with statins for metals associated with BP in any regression by adding interaction terms between metals and statin use (yes/no) to main effects models and by modeling stratified regressions.
The following potential confounders were identified from the literature and included in all models: age, cigarette smoking (never, current, or former), pack-years, BMI (kilograms per square meter), alcohol intake (two or more vs. fewer than two drinks/day), race/ethnicity, education (years), and season and year of clinical visit. For mercury, we adjusted for darkmeat fish, shellfish, canned tuna, and other fish intake. To evaluate potential confounding by other toenail metals, we included all five metals together in the same regressions, retaining all covariates.
To examine dose-response relationships between arsenic and BP, we refit regression models using the generalized additive model in R (R Project for Statistical Computing, Vienna, Austria). We fit a penalized spline and chose the optimum penalty using generalized cross validation.
We evaluated associations between an interquartile range (IQR) increase in toenail metals and hypertension (defined as antihypertensive medication use or SBP ≥ 140 mmHg or DBP ≥ 90 mmHg at study visit) using logistic regression and adjusting for the aforementioned confounders.
We conducted a sensitivity analysis using inverse probability weighting to adjust for potential selection bias [bib_ref] Instrumental variables and inverse probability weighting for causal inference from longitudinal observational..., Hogan [/bib_ref]. We ran a logistic regression assessing relations between covariates and participation status, retaining variables that were significant at the p < 0.10 level (age, BMI, smoking). We obtained predicted probability of participation for each subject, and weighted regression models by inverses of these probabilities. Results were not appreciably altered by inverse probability weighting (data not shown).
Because a number of otherwise eligible participants were excluded from cadmium-BP regressions due to toenail concentrations below the DL, we also conducted a sensitivity analysis in which we assigned participants with cadmium concentrations below the DL a concentration value of the mean sample DL/2 (0.01 µg/g).
Using Student's t-test and chi-square analy sis, we compared participants with nonparticipants presenting during the same time period. For this comparison, we used data from the first NAS visit of the nonparticipants between January 1999 and January 2009 (the period of toenail collection). We also calculated medians and IQRs for toenail metal levels after stratifying by potential confounders and effect modifiers. Correlations between toenail metals were evaluated using Spearman correlation. Regression results were similar, although somewhat weaker, when using log-transformed metal concentrations (data not shown).
# Results
Study population. At their first study visit with donated toenails and complete information on BP and covariates, participants were mostly volume 120 | number 1 | January 2012 - Environmental Health Perspectives former smokers. Mean BMI was 28 kg/m 2 , and 20%, 53%, and 27% of participants had a BMI of < 25, 25-30, and ≥ 30 kg/m 2 , respectively. Participants were younger than nonparticipants (mean ages 72 and 74, respectively, p = 0.03) [fig_ref] Table 1: Descriptive statistics by participation status [/fig_ref]. Mean uncorrected SBP and DBP were 133 mmHg and 78 mmHg, respectively, with 71% of participants classified as hypertensive based on medication use or BP at study visit. Sixty percent used antihypertensives; 37% used statins.
Detailed information on metal concentrations by participant characteristics is provided in Supplemental Material, [fig_ref] Table 1: Descriptive statistics by participation status [/fig_ref] (http://dx.doi.org/10.1289/ehp.1002805). For example, arsenic was higher among men consuming two or more versus fewer than two alcoholic drinks per day (0.09 vs. 0.07 µg/g); among non-Hispanic blacks compared with non-Hispanic whites (0.09 vs. 0.08 µg/g), cadmium was higher among men consuming two or more versus fewer than two alcoholic drinks per day (0.02 vs. 0.01 µg/g) and among men with fewer years of education (< 15 years vs. ≥ 15 years, 0.02 vs. 0.01 µg/g); mercury was higher among men with greater dark-meat fish (0.31 vs. 0.15 µg/g), other fish (0.26 vs. 0.15 µg/g), shellfish (0.27 vs. 0.18 µg/g), and tuna intake (0.25 vs. 0.12 µg/g) (one or more servings per month vs. less than one serving per month for each seafood category). Manganese was lower among statin users (0.24 vs. 0.31 µg/g) and was higher among non-Hispanic whites compared with non-Hispanic blacks (0.28 vs. 0.18 µg/g); lead was higher among men consuming two or more versus fewer than two alcoholic drinks per day (0.38 vs. 0.30 µg/g) and among men with fewer years of education (< 15 years vs. ≥ 15 years, 0.35 vs. 0.26 µg/g). Toenail cadmium levels were similar across categories defined by smoking history or pack-years [Supplemental Material, [fig_ref] Table 1: Descriptive statistics by participation status [/fig_ref] (http://dx.doi.org/10.1289/ehp.1002805)], even when participants with cadmium concentrations below the DL were assigned a concentration value of 0.01 µg/g (mean sample DL/2; data not shown). Correlations between toenail metals are reported in Supplemental Material, Table 2 (http://dx.doi.org/10.1289/ ehp.1002805). Briefly, we found statistically significant correlations between arsenic and cadmium (r = 0.39), mercury (r = 0.09), manganese (r = 0.56), and lead (r = 0.45), between cadmium and manganese (r = 0.49) and lead (r = 0.58), and between manganese and lead (r = 0.50) (Supplemental Material, [fig_ref] Table 2: Linear regression models estimating the change in BP parameters associated with an... [/fig_ref].
Arsenic. An IQR (0.06 µg/g) increase in arsenic was associated with higher SBP [0.93 mmHg; 95% confidence interval (CI): 0.25, 1.62] and pulse pressure (0.76 mmHg; 95% CI: 0.22, 1.30) and showed a weak positive association with DBP [fig_ref] Table 2: Linear regression models estimating the change in BP parameters associated with an... [/fig_ref]. Adjustment for other metals strengthened associations with SBP (1.43 mmHg, 95% CI: 0.34, 2.51) and DBP (0.63 mmHg, 95% CI: 0.05, 1.21). The association with pulse pressure was essentially unchanged but was less precise [fig_ref] Table 2: Linear regression models estimating the change in BP parameters associated with an... [/fig_ref].
We evaluated the dose-response relationship between arsenic and SBP, using penalized splines [Supplemental Material, Data are given as mean ± SD or n (%). Participant data are from first study visit with donated toenails and complete information on BP and covariates. For comparison with eligible NAS subjects who were not included in the metal-BP analyses (i.e., nonparticipants), we used data from first NAS visit of nonparticipants between January 1999 and January 2009 (the period of toenail collection). slope of 1, indicating a linear association. Thus, there is no evidence of deviation from linearity in the arsenic-SBP model.
Cadmium. An IQR (0.02 µg/g) increase in cadmium was not significantly associated with BP [fig_ref] Table 2: Linear regression models estimating the change in BP parameters associated with an... [/fig_ref]. Adjustment for other metals did not meaningfully alter these results [fig_ref] Table 2: Linear regression models estimating the change in BP parameters associated with an... [/fig_ref]. Assigning participants with cadmium levels below the DL a concentration value of DL/2 did not appreciably alter results (data not shown).
Mercury. An IQR (0.31 µg/g) increase in mercury was negatively but not significantly associated with BP (SBP: -0.56, 95% CI: -2.27, 1.16; DBP: -0.21 mmHg, 95% CI: -1.16, 0.75; pulse pressure: -0.35 mmHg, 95% CI: -1.70, 1.00) [fig_ref] Table 2: Linear regression models estimating the change in BP parameters associated with an... [/fig_ref]. Adjustment for other metals did not meaningfully affect these results [fig_ref] Table 2: Linear regression models estimating the change in BP parameters associated with an... [/fig_ref]. Mercury regressions have a considerably smaller sample size than regressions for other metals [fig_ref] Table 2: Linear regression models estimating the change in BP parameters associated with an... [/fig_ref] because of toenail samples of insufficient weight to assess mercury concentrations as well as adjustment for fish and shellfish intake.
Manganese. Manganese was negatively but not significantly associated with BP (SBP: -0.45 mmHg, 95% CI: -1.18, 0.28; DBP: -0.12 mmHg, 95% CI: -0.51, 0.28; pulse pressure: -0.33 mmHg, 95% CI: -0.91, 0.25) [fig_ref] Table 2: Linear regression models estimating the change in BP parameters associated with an... [/fig_ref]. After adjusting for other metals, an IQR (0.40 µg/g) increase in manganese was more strongly associated with decreased SBP (-1.09 mmHg; 95% CI: -2.08, -0.10) and DBP (-0.62 mmHg; 95% CI: -1.15, -0.09).
Lead. An IQR (0.52 µg/g) increase in lead was not associated with BP (SBP: -0.07 mmHg, 95% CI: -0.67, 0.53; DBP: 0.24 mmHg, 95% CI: -0.08, 0.57; pulse pressure: -0.31 mmHg, 95% CI: -0.78, 0.16) [fig_ref] Table 2: Linear regression models estimating the change in BP parameters associated with an... [/fig_ref]. Adjustment for other metals did not meaningfully alter these results [fig_ref] Table 2: Linear regression models estimating the change in BP parameters associated with an... [/fig_ref].
Hypertension. Estimated associations of hypertension with an IQR increase in toenail metals were consistent with linear regression results [arsenic, odds ratio (OR) = 1.07, 95% CI: 0.97, 1.20; cadmium, OR = 1.01, 95% CI: 0.95, 1.06; mercury, OR = 0.85, 95% CI: 0.68, 1.05; manganese, OR = 0.96, 95% CI: 0.87, 1.05; lead, OR = 1.00, 95% CI: 0.92, 1.08]. Upon adjusting for other metals, manganese was negatively associated with hypertension (OR = 0.88; 95% CI: 0.77, 0.999). Similar adjustment did not meaningfully alter the other results (data not shown).
Effect modification. We found a significant interaction between arsenic and statins for DBP but not for SBP or pulse pressure [fig_ref] Table 3: Effect modification of the association between an IQR increase in toenail arsenic... [/fig_ref]. Among men not taking statins (n = 404), an IQR increase in arsenic was associated with an increase in DBP (0.76 mmHg; 95% CI: 0.20, 1.33). Arsenic showed a non significant negative association with DBP among statin users (n = 235; -0.30 mmHg; 95% CI: -0.80, 0.21). Within the sample included in arsenic-BP analyses, statin and antihypertensive use were strongly associated with each other based on chi-square analysis and age-adjusted logistic regression (data not shown).
We found no significant interactions between manganese and statins (SBP: p = 0.76, DBP: p = 0.39; pulse pressure: p = 0.84). Results were similar after adjustment for other metals (data not shown).
# Discussion
In a cohort of elderly men with low toenail arsenic concentrations, we found positive associations between toenail arsenic and BP levels. Our results also tentatively suggest negative associations between toenail manganese and BP levels. We found little evidence of associations with toenail cadmium, mercury, or lead. If confirmed, these results may be important because of the high prevalence of both cardiovascular disease and low to moderate metal exposures in the United States.
No investigations have assessed relations between arsenic, cadmium, mercury, manganese, or lead and BP using toenails as biomarkers. Toenail samples are convenient to collect and store, grow more slowly than hair, are more protected from external contaminants, and represent longer-term exposures than blood or urine [bib_ref] Validity of human nails as a biomarker of arsenic and selenium exposure:..., Slotnick [/bib_ref]. Toenails likely reflect metal exposures from the preceding 12-18 months, with slower growth among the elderly [bib_ref] Validity of human nails as a biomarker of arsenic and selenium exposure:..., Slotnick [/bib_ref].
Arsenic. Arsenic may elevate BP through pathways related to oxidative stress [bib_ref] Metals, toxicity and oxidative stress, Valko [/bib_ref] , inflammation [bib_ref] Gene expression of inflammatory molecules in circulating lymphocytes from arsenic-exposed human subjects, Wu [/bib_ref] , and endothelial dysfunction/nitric oxide inhibition [bib_ref] Decreased serum concentrations of nitric oxide metabolites among Chinese in an endemic..., Pi [/bib_ref].
Epidemiologic studies have mostly evaluated inorganic arsenic, which is more toxic than organic arsenic. Our toenail assay does not distinguish between forms. However, seafood is the main source of organic arsenic [bib_ref] Validity of human nails as a biomarker of arsenic and selenium exposure:..., Slotnick [/bib_ref]. Toenail arsenic is very weakly or not at all correlated with fish (dark-meat fish: r = 0.06; other fish: r = 0.09; tuna: r = -0.004) and shellfish (r = 0.06) intake among our participants.
Arsenic sources include soil, dust, air, and food (Benbrahim-Tallaa and Waalkes 2008; [bib_ref] Validity of human nails as a biomarker of arsenic and selenium exposure:..., Slotnick [/bib_ref]. Water is a potentially important inorganic arsenic source [bib_ref] Validity of human nails as a biomarker of arsenic and selenium exposure:..., Slotnick [/bib_ref]. The Massachusetts Water Resources Authority (MWRA), in which arsenic is consistently undetectable (< 1.0 µg/L) (MWRA 2011), supplies most of the Greater Boston area.
Arsenic is found in toenails because of its high affinity to sulfhydryl groups [bib_ref] Measurement of low levels of arsenic exposure: a comparison of water and..., Karagas [/bib_ref]. Toenails are validated biomarkers of arsenic exposure [bib_ref] Toenails as a biomarker of inorganic arsenic intake from drinking water and..., Slotnick [/bib_ref] [bib_ref] Validity of human nails as a biomarker of arsenic and selenium exposure:..., Slotnick [/bib_ref]. Toenail concentrations in our study appear similar to those reported in most U.S. general population studies [bib_ref] Toenail trace element levels as biomarkers: reproducibility over a 6-year period, Garland [/bib_ref] [bib_ref] Skin cancer risk in relation to toenail arsenic concentrations in a US..., Karagas [/bib_ref] [bib_ref] Arsenic exposure predicts bladder cancer survival in a US population, Kwong [/bib_ref] [bib_ref] Profiles of trace elements in toenails of Arab-Americans in the Detroit area, Slotnick [/bib_ref] [bib_ref] Toenails as a biomarker of inorganic arsenic intake from drinking water and..., Slotnick [/bib_ref] [bib_ref] Evaluation of exposure to arsenic in residential soil, Tsuji [/bib_ref]. Other U.S. studies report somewhat higher [bib_ref] Total arsenic concentrations in toenails quantified by two techniques provide a useful..., Adair [/bib_ref] [bib_ref] Evaluation of a food frequency questionnaire-food composition approach for estimating dietary intake..., Macintosh [/bib_ref] or lower [bib_ref] Toenail arsenic content and cutaneous melanoma in Iowa, Beane Freeman [/bib_ref] arsenic levels.
Several investigations have assessed relations between high-level arsenic exposure and BP. These include two population-based analyses that found dose-response relationships between well-water arsenic and hypertension [bib_ref] Increased prevalence of hypertension and long-term arsenic exposure, Chen [/bib_ref] [bib_ref] Hypertension and arsenic exposure in Bangladesh, Rahman [/bib_ref]. Exposure categories used in these studies (10-700, 710-900, > 900 µg/L, and < 500, 500-1,000, > 1,000 µg/L, respectively) are much greater than the maximum contaminant level for arsenic in the United States (10 µg/L), although in practice this target value is often exceeded in U.S. drinking water sources [Agency for Toxic Substances and Disease Registry (ATSDR) 2007]. A recent Iranian study also found positive associations between high-level arsenic exposure and BP [bib_ref] Arsenic exposure, dermatological lesions, hypertension, and chromosomal abnormalities among people in a..., Dastgiri [/bib_ref].
Two population-based studies in Bangladesh and Inner Mongolia (with exposure categories of < 8, 8.1-40.8, 40.9-91.0, 91.1-176.0, 176.1-864.0 µg/L, and < 20, 21-50, 51-100, > 100 µg/L drinking water, respectively) reported positive associations between more moderate arsenic levels and BP-related outcomes [bib_ref] Drinking water arsenic exposure and blood pressure in healthy women of reproductive..., Kwok [/bib_ref]. A U.S. study found an association between well-water arsenic levels of ≥ 10 µg/L versus < 2 µg/L and self-reported high BP [bib_ref] Prevalence of chronic diseases in adults exposed to arsenic-contaminated drinking water, Zierold [/bib_ref].
We found positive associations between arsenic and BP levels, consistent with the current literature. Our study likely evaluated considerably lower arsenic exposure levels than previous investigations, because arsenic is consistently undetectable in the main water supply for the Greater Boston area (MWRA 2011). Our study was the first to use a biomarker of arsenic dose when assessing relations with BP. Cadmium. Cadmium may act on BP through mechanisms related to oxidative stress [bib_ref] Metals, toxicity and oxidative stress, Valko [/bib_ref] , inflammation [bib_ref] Cadmiuminduced renal damage and proinflammatory cytokines: possible role of IL-6 in tubular..., Kayama [/bib_ref] , endothelial dysfunction, partial agonism of calcium channels, increased vasoconstriction, and activation of the sympathetic nervous system [bib_ref] • Environmental Health Perspectives as an environmental factor of hypertension in animals:..., Varoni [/bib_ref]. It may also act through renal tubular damage, sodium retention, and volume overload [bib_ref] Kidney dysfunction and hypertension: role for cadmium, p450 and heme oxygenases?, Satarug [/bib_ref].
Food from cadmium-contaminated soils is a major exposure source [bib_ref] Adverse effects of chronic exposure to low-level cadmium in foodstuffs and cigarette..., Satarug [/bib_ref]. Other sources include tobacco smoke, soil, dust, water, and air pollution [bib_ref] Endocrine disruption by cadmium, a common environmental toxicant with paradoxical effects on..., Henson [/bib_ref] [bib_ref] Adverse effects of chronic exposure to low-level cadmium in foodstuffs and cigarette..., Satarug [/bib_ref].
Toenail cadmium concentrations were lower in our study than in other U.S. general population studies [bib_ref] Prediagnostic toenail cadmium and zinc and subsequent prostate cancer risk, Platz [/bib_ref] [bib_ref] Profiles of trace elements in toenails of Arab-Americans in the Detroit area, Slotnick [/bib_ref] [bib_ref] Mercury and risk of coronary heart disease in men, Yoshizawa [/bib_ref]. Data on validity of toenails as cadmium biomarkers are sparse. Most relevant studies found associations between cadmium exposure and concentrations in toenails or nails [bib_ref] Arsenic, cadmium, and lead levels in hair and toenail samples in Pakistan, Anwar [/bib_ref] [bib_ref] Reference intervals of cadmium, lead, and mercury in blood, urine, hair, and..., Mortada [/bib_ref] [bib_ref] Case-control study of toenail cadmium and prostate cancer risk in Italy, Vinceti [/bib_ref]. However, toenail levels did not vary with smoking (a major cadmium source) among our participants. This may reflect shortcomings of toenails as cadmium biomarkers or other factors, such as the potential influence of non tobacco cadmium sources. In addition, few participants were active smokers, and number of pack-years likely reflects an exposure window too distant to influence current toenail concentrations.
The literature regarding cadmium and BP is inconsistent. Investigations report positive [bib_ref] Cadmium-induced nephropathy in the development of high blood pressure, Satarug [/bib_ref] [bib_ref] Cadmium exposure and hypertension in the 1999-2004 National Health and Nutrition Examination..., Tellez-Plaza [/bib_ref] and null [bib_ref] Blood pressure, the prevalence of cardiovascular diseases, and exposure to cadmium: a..., Staessen [/bib_ref] [bib_ref] Exposure to cadmium and conventional and ambulatory blood pressures in a prospective..., Staessen [/bib_ref] associations with BP. One ecologic analysis reported a negative relation [bib_ref] Case-control study on cardiovascular function in females with a history of heavy..., Kagamimori [/bib_ref]. We found little evidence of associations between cadmium and BP. This should be interpreted cautiously in light of an unvalidated biomarker of exposure.
Mercury. Mercury may act on BP through mechanisms related to oxidative stress, vascular inflammation, endothelial dysfunction/ nitric oxide inhibition, renal tubular dysfunction, and proteinuria [bib_ref] The role of mercury and cadmium heavy metals in vascular disease, hypertension,..., Houston [/bib_ref].
Mercury exposure occurs primarily through seafood and dental amalgams [bib_ref] Lead and mercury exposures: interpretation and action, Brodkin [/bib_ref]. Toenail mercury is a reliable, wellvalidated biologic marker of long-term exposure to the metal [bib_ref] Toenail trace element levels as biomarkers: reproducibility over a 6-year period, Garland [/bib_ref] [bib_ref] The relationship between amalgam restorations and mercury levels in male dentists and..., Joshi [/bib_ref] [bib_ref] Evaluation of a food frequency questionnaire-food composition approach for estimating dietary intake..., Macintosh [/bib_ref] [bib_ref] Toenail mercury and dietary fish consumption, Rees [/bib_ref]. Toenail mercury concentration in our analysis was similar to other general population studies [bib_ref] Toenail trace element levels as biomarkers: reproducibility over a 6-year period, Garland [/bib_ref] [bib_ref] Toenail mercury and dietary fish consumption, Rees [/bib_ref] but lower than in the U.S. Health Professionals Follow-Up Study [bib_ref] The relationship between amalgam restorations and mercury levels in male dentists and..., Joshi [/bib_ref].
The overall evidence on mercury and BP is inconsistent, with studies reporting positive and null associations [bib_ref] Relationship between mercury concentration in blood, cognitive performance, and blood pressure, in..., Johansson [/bib_ref] [bib_ref] Relationship between mercury in blood and 24-h ambulatory blood pressure in Greenlanders..., Pedersen [/bib_ref] [bib_ref] Environmental mercury exposure and blood pressure among Nunavik Inuit adults, Valera [/bib_ref]. Data on associations between low-level mercury and BP are sparse. We found little evidence of an association between mercury and BP.
Manganese. Manganese may act on BP by lowering vascular sensitivity to alpha-adrenergic receptor activation, decreasing dopamine levels, and inducing oxidative stress. It may also act through antagonism of the alpha-adrenergic receptor in blood vessels, calcium channel antagonism, or effects on the autonomic nervous system [bib_ref] Cardiovascular toxicities upon manganese exposure, Jiang [/bib_ref].
Manganese exposure occurs through food, water, air, and soil [bib_ref] Manganese exposure, essentiality and toxicity, Santamaria [/bib_ref]. Manganese is associated with hypotension and decreased BP in animal [bib_ref] Cardiovascular toxicities upon manganese exposure, Jiang [/bib_ref] and occupational research [bib_ref] Effects of manganese exposure on cardiovascular functions of male workers, Jiang [/bib_ref] [bib_ref] Investigation on health in manganese ferroalloy workers, Su [/bib_ref]. Postural hypotension was observed in manganese dipyridoxyldiphosphate overdosed patients [bib_ref] A toxicologic risk for using manganese complexes? A literature survey of existing..., Misselwitz [/bib_ref]. Toenail manganese concentration in our investigation is similar, although slightly lower, relative to another U.S. general population study [bib_ref] Profiles of trace elements in toenails of Arab-Americans in the Detroit area, Slotnick [/bib_ref].
Environmental manganese is generally considered to lower hypertension risk [bib_ref] Trace elements and cardiovascular diseases, Houtman [/bib_ref]. However, non occupational studies have reported positive [bib_ref] Mineral factors controlling essential hypertension -a study in the Chandigarh, India population, Taneja [/bib_ref] , null [bib_ref] The content of elements in rainwater and its relation to the frequency..., Tubek [/bib_ref] , and negative [bib_ref] Differences in metal and metalloid content in the hair of normo-and hypertensive..., Gonzalez-Muñoz [/bib_ref] associations with BP-related outcomes. Results may differ across investigations because of variations in study design, confounder control, metal measurement, definition or measurement of the BP outcome, manganese exposure level, and participant characteristics.
We found negative relations between toenail manganese and BP in models adjusted for other metals. This is consistent with results observed in occupational investigations. However, these tentative results should be interpreted extremely cautiously, because toenails are not validated biomarkers of environmental manganese exposure.
Lead. Proposed mechanisms linking lead to BP elevation include effects on kidney function, oxidative stress, and nitric oxide and guanylate cyclase levels, and changes in the renin-angiotensin system [bib_ref] Lead exposure and cardiovascular disease-a systematic review, Navas-Acien [/bib_ref].
Lead is associated with hypertension and BP elevation in occupational and experimental animal research and in many environmental epidemiologic studies [bib_ref] Lead exposure and cardiovascular disease-a systematic review, Navas-Acien [/bib_ref]. A recent systematic review concluded that this association is causal [bib_ref] Lead exposure and cardiovascular disease-a systematic review, Navas-Acien [/bib_ref]. Bone lead, which represents cumulative lead exposure [bib_ref] The epidemiology of lead toxicity in adults: measuring dose and consideration of..., Hu [/bib_ref] , is consistently associated with BP outcomes in the NAS [bib_ref] Bone lead and blood lead levels in relation to baseline blood pressure..., Cheng [/bib_ref] [bib_ref] The relationship of bone and blood lead to hypertension. The Normative Aging..., Hu [/bib_ref] [bib_ref] Cumulative community-level lead exposure and pulse pressure: the Normative Aging Study, Perlstein [/bib_ref]. Toenail lead levels in our analysis were similar to those of adults in another U.S. general population study [bib_ref] Profiles of trace elements in toenails of Arab-Americans in the Detroit area, Slotnick [/bib_ref].
Exposure occurs primarily through inhalation of lead dust and ingestion of contaminated food and water [bib_ref] The long-term consequences of exposure to lead, Rosin [/bib_ref]. Toenails are not validated biomarkers of lead exposure. In fact, some research suggests that nails may be inappropriate biomarkers for lead [bib_ref] A critical review of biomarkers used for monitoring human exposure to lead:..., Barbosa [/bib_ref]. We found no association between toenail lead and BP, perhaps because of a number of factors. For example, toenail lead represents a shorter averaging time than bone lead. Alternatively, toenails may be a poor lead biomarker. Among men included in lead-BP analyses, toenail lead was weakly correlated with blood lead (r = 0.28, p < 0.0001), but not tibia (r = 0.04, p = 0.55) or patella (r = 0.03, p = 0.64) lead (blood and bone lead were measured as described by [bib_ref] The relationship of bone and blood lead to hypertension. The Normative Aging..., Hu [/bib_ref].
Effect modification. Statins reduce inflammation and oxidative stress, including in the vascular endothelium, and increase nitric oxide levels [bib_ref] Effects of atorvastatin on the different phases of atherogenesis, Rubba [/bib_ref].
We found statistically significant interaction between arsenic and statins for DBP, with associations limited to statin non users. Thus, it is possible that statins blunt effects of arsenic on DBP, although this interaction needs to be evaluated in other studies. We found no evidence of interactions between manganese and statin use.
## Study limitations and future research directions
The cross-sectional design of our investigation precludes us from ascertaining causality. We had limited power to assess interactions. In addition, our study population consists almost entirely of white, elderly men.
Another limitation is using toenails as biomarkers. Although toenails are validated biomarkers for mercury and arsenic, their validity is unclear for cadmium and manganese and has been called into question for lead. It is very important to interpret results with caution. BP readings are based on only one study visit, although BP is variable and affected by many stimuli. This variability may attenuate results.
We did not correct for multiple comparisons. We have examined five metals. However, they all have pathway-based support for associations with BP, and we believe that our study should not be penalized relative to investigations examining only individual metals under these circumstances.
More research is needed to address relations between metals and BP among larger, more diverse prospective cohorts. Potential interactions between arsenic and statins should be evaluated in future analyses.
# Conclusions
Our findings suggest that environmental arsenic may be associated with BP. Our analyses also tentatively suggest that toenail manganese may be associated with decreased BP. We found no consistent associations with IQR increases in toenail mercury, cadmium, or lead. These findings should be interpreted in light of the difficulty of ascertaining fish intake [the beneficial cardiovascular effects of this important mercury source [bib_ref] Does supplementation of diet with 'fish oil' reduce blood pressure? A meta-analysis..., Appel [/bib_ref] may confound associations between toenail mercury and BP] and potential shortcomings of toenails as biomarkers for metal exposure.
Although changes in BP reported here are not clinically relevant on an individual level, shifting the population distribution of BP has public health significance [bib_ref] Reducing the population burden of cardiovascular disease by reducing sodium intake: a..., Dickinson [/bib_ref]. Findings of associations for arsenic at these levels, if confirmed, suggests that current efforts to regulate arsenic exposure in the general population may be inadequate.
[table] Table 1: Descriptive statistics by participation status. [/table]
[table] Table 2: Linear regression models estimating the change in BP parameters associated with an IQR increase in metal levels. a,b For each metal, the same IQR value is examined in relation to BP for both metal-adjusted and non-metal-adjusted models. b All regression models are adjusted for age, cigarette smoking, pack-years of smoking, season of clinical visit, year of clinical visit, BMI, years of education, race/ethnicity, and alcohol intake. Mercury-BP models are adjusted for fish intake. c In addition to the covariates in footnote b, this regression model is also adjusted for the four other toenail metals. *p < 0.05. **p < 0.01. [/table]
[table] Table 3: Effect modification of the association between an IQR increase in toenail arsenic level a and BP by statin use (n = 639). b [/table]
|
Characterization of an iPSC line NCHi006-A from a patient with hypoplastic left heart syndrome (HLHS)
Hypoplastic left heart syndrome (HLHS) is a severe congenital heart defect characterized by underdeveloped structures on the left side of the heart, including hypoplasia of the left ventricle and stenosis or atresia of the aortic and mitral valves. Here, we generated an iPSC line from the peripheral blood mononuclear cells of a male patient with HLHS through Sendai virus-mediated transfection of 4 Yamanaka factors. This iPSC line exhibited normal morphology, expressed pluripotency markers, had a normal karyotype, and could differentiate into cells of three germ layers. This iPSC line can be used for studying cellular and developmental etiologies of HLHS.
## Resource
## Resource utility
Derived from a patient with hypoplastic left heart syndrome (HLHS), this iPSC line can be differentiated into cardiac cell lineages for modeling congenital heart defects. It may serve as a useful in vitro biological system to study underlying mechanisms of HLHS, screen for candidate therapeutics, and increase our understanding of human cardiac development.
## Resource details
Hypoplastic left heart syndrome is a severe congenital heart defect where the left ventricle of the heart is underdeveloped, causing aberrant hemodynamics [bib_ref] Hypoplastic left heart syndrome: A new paradigm for an old disease?, Grossfeld [/bib_ref]. The cause of HLHS is still poorly understood due to the lack of experimental models. Here, we established and characterized an iPSC line derived from a male infant with HLHS to provide a biological system that retains the genetic information of the proband and can be utilized to model stages of human development through differentiation into the three germ layers. We envision this iPSC line to be used as a patient-specific biological model to study human cardiac development, especially to interrogate the mechanisms that govern the pathogenesis of HLHS [bib_ref] Probing single ventricle heart defects with patient-derived induced pluripotent stem cells and..., Hall [/bib_ref] [bib_ref] Decoding genetics of congenital heart disease using patient-derived induced pluripotent stem cells..., Lin [/bib_ref].
To generate iPSC line NCHi006-A, blood was drawn from a patient clinically diagnosed with HLHS and mitral/aortic stenosis, with no other observed heart defects (see .
The peripheral blood mononuclear cells (PBMCs) were isolated and transfected with 4 Yamanaka factors to produce an iPSC line that exhibited normal morphology and colony formation [fig_ref] Figure 1: Fig. 1. [/fig_ref]. The majority of cells displayed pluripotency markers TRA-1-60, SOX2, NANOG, and OCT3/4, as detected by immunofluorescence staining [fig_ref] Figure 1: Fig. 1. [/fig_ref]. Genetically, the iPSCs displayed a normal male karyotype (46, XY) as confirmed by a whole-genome array [fig_ref] Figure 1: Fig. 1. [/fig_ref] , and their identity was confirmed using STR analysis to prove their origin from the patient's PBMCs (in archive with the journal). This iPSC line had the ability to differentiate into cells of all three germ layers, as established by positive immunofluorescence staining of germ layer-specific markers. Ectodermal-like cells showed expression of PAX6 and OTX2, mesodermal-like cells displayed TBX6 and Brachyury, while endodermal-like cells expressed FOXA2 and SOX17 [fig_ref] Figure 1: Fig. 1. [/fig_ref]. The iPSCs were also tested negative for mycoplasma contamination [fig_ref] Figure 1: Fig. 1. [/fig_ref].
# Materials and methods
## Reprogramming
Patient PBMCs were isolated and incubated for one week in StemPro-34 SFM medium (Thermo Fisher Scientific) supplemented with 100 ng/mL SCF (PeproTech), 100 ng/mL FLT3 (Thermo Fisher Scientific), 20 ng/mL IL3 (PeproTech), 20 ng/mL IL6 (Gibco), 20 ng/mL EPO (Thermo Fisher Scientific), and 1× GlutaMAX (Thermo Fisher Scientific). PBMCs were then transfected using the CytoTune™-iPS 2.0 Sendai Reprogramming Kit (Thermo Fisher Scientific) according to the manufacturer's instructions. Transfected cells were resuspended in supplemented StemPro-34 SFM medium and transferred into a Matrigel-coated plate for one week. Cells were then switched to complete E8 medium (Thermo Fisher Scientific). After two weeks, emerging iPSC clones were picked, expanded over multiple passages, and stored in liquid N 2 .
## Ipsc maintenance and passaging
Cells were maintained in complete E8 media at 37 °C with 5% CO 2 . Upon reaching 90% confluency, cells were washed with DPBS then dissociated with 0.5 mM EDTA for 5-8 min. EDTA was then removed, and iPSCs were manually dislodged with complete E8 media plus ROCK inhibitor (Y-27632, Selleck Chemicals). To split, the cell suspension was replated at a 1:6-1:10 ratio.
## Immunofluorescent staining
The pluripotency of iPSCs (passages 12-13) was assessed by immunofluorescence staining and manual counting. Cells were fixed with 4% paraformaldehyde solution (Electron Microscopy Sciences) for 15 min, then permeabilized with 0.1% Triton X-100 solution (Sigma) for 20 min at room temperature. Cells were then blocked with 0.2% BSA (Sigma) in DPBS, and incubated with primary antibodies (dilution: 1:200) overnight at 4 °C. The next day, secondary antibodies (dilution: 1:2000) in 0.2% BSA were added at room temperature for 1 h, then counterstained with DAPI (dilution: 1:2000) in DPBS at room temperature for 10 min (see . Stained coverslips were mounted onto glass slides using SlowFade Gold Antifade (Thermo Fisher Scientific) and imaged with a fluorescence microscope (Keyence).
## Karyotyping
To detect chromosomal abnormalities using whole genome array, 2 × 10 6 iPSCs (passages 12-13) from were harvested and analyzed using the KaryoStat Assay (Thermo Fisher Scientific).
## Short tandem repeat (str) analysis
Genomic DNA was extracted from iPSCs (passages 14-15) and PBMCs using the Quick-DNA Miniprep Plus Kit (Zymo Research). The PowerPlex 16 System (Promega) was then utilized to amplify genomic materials according to the manufacturer's instructions. Samples were sent for capillary sequencing using an ABI 3730xl Genetic Analyzer (Thermo Fisher Scientific). GeneMapper 5.0 (Thermo Fisher Scientific) software was used to analyze the sequencing data for allele callings for 16 loci per sample. Only strong allele calling signals were considered for analysis.
## Germ layer differentiation
Pluripotency was confirmed by differentiating iPSCs (passages 11-12) into endoderm and ectoderm cells using the Human Pluripotent Stem Cell Functional Identification Kit (R&D Systems) according to the manufacturer's instructions. Mesoderm differentiation was induced through two-day application of 6 μM CHIR99021 (Selleck Chemicals) in RPMI 1640 media (Thermo Fisher Scientific) with B27 minus insulin supplement (Thermo Fisher Scientific). Samples were fixed and stained with respective germ layer-specific markers.
## Mycoplasma detection
Mycoplasma contamination was checked using the MycoAlert™ Detection Kit (Lonza) on iPSC passage 12 supernatant following the manufacturer's protocol.
# Supplementary material
Refer to Web version on PubMed Central for supplementary material. Characterization of an iPSC line derived from an HLHS patient.
[fig] Figure 1: Fig. 1. [/fig]
|
The cerebellum and its network: Disrupted static and dynamic functional connectivity patterns and cognitive impairment in multiple sclerosis
Background: The impact of cerebellar damage and (dys)function on cognition remains understudied in multiple sclerosis. Objective: To assess the cognitive relevance of cerebellar structural damage and functional connectivity (FC) in relapsing-remitting multiple sclerosis (RRMS) and secondary progressive multiple sclerosis (SPMS). Methods: This study included 149 patients with early RRMS, 81 late RRMS, 48 SPMS and 82 controls. Cerebellar cortical imaging included fractional anisotropy, grey matter volume and resting-state functional magnetic resonance imaging (MRI). Cerebellar FC was assessed with literature-based resting-state networks, using static connectivity (that is, conventional correlations), and dynamic connectivity (that is, fluctuations in FC strength). Measures were compared between groups and related to disability and cognition. Results: Cognitive impairment (CI) and cerebellar damage were worst in SPMS. Only SPMS showed cerebellar connectivity changes, compared to early RRMS and controls. Lower static FC was seen in fronto-parietal and default-mode networks. Higher dynamic FC was seen in dorsal and ventral attention, default-mode and deep grey matter networks. Cerebellar atrophy and higher dynamic FC together explained 32% of disability and 24% of cognitive variance. Higher dynamic FC was related to working and verbal memory and to information processing speed. Conclusion: Cerebellar damage and cerebellar connectivity changes were most prominent in SPMS and related to worse CI.
# Introduction
MS is a common neuroinflammatory and neurodegenerative disease of the central nervous system (CNS), affecting the grey matter (GM) and white matter (WM). A brain region commonly excluded in MS research is the cerebellum, 1 especially cerebellar cortex. Previous studies showed that cerebellar cortex is commonly demyelinated 2 and atrophic,with ongoing discussion on specific stagings of damage.In addition, recent studies on the healthy brain discovered strong cerebellar connections with specific functional networks like the fronto-parietal network (FPN)and relations with cognition.Nonetheless, in MS research, how cerebellar pathology influences cognition remains unclear.The field of network neuroscience 9 recently evolved with the discovery of dynamic 10 (or time-varying) functional connectivity (FC), which is strongly related to cognition.Conceptually, static connectivity could represent the amount of information transferred (the total correlation between two signals), while dynamic connectivity could assess the variability of the level of information transfer. As such, these two measures together could provide unique information on network functioning crucial for cognition.Unfortunately, a few studies specifically investigated the cognitive role of cerebellar FC, and even fewer in MS.As such, we investigated at which disease stage cerebellar cortical damage (i.e. diffusion changes and atrophy) and FC alterations (static and dynamic) become apparent in patients with relapse-onset MS and how these relate to cognition. We expected that the cerebellum would show strong disconnection in progressive MS (based on static FC), combined with highly variable cognitive connections (based on dynamic FC). This hypothesis was addressed in a large MS cohort divided into patients with a relatively short disease duration ('early'), those with longer disease durations ('late') and progressive MS.
# Methods
## Participants
Retrospective data from participants of the Amsterdam MS cohortwith sufficient cerebellar coverage (see functional MRI processing) and relapse-onset MS were included. Based on disease duration, the relapsing-remitting multiple sclerosis (RRMS) group was subdivided into 'early' (<15 years) and 'late' (>15 years) RRMS. The shortest disease duration was 4.6 years. The final sample (seeincluded 278 MS patients (74% women, age 47 ± 11 years, 149 early and 81 late RRMS, and 48 SPMS) and 82 matched healthy controls (HCs, 63% women, age 46 ± 11 years). Patients were diagnosed with the revised McDonald criteria.The Expanded Disability Status Scale (EDSS) 20 was used to measure overall disability. Patients were relapse-free and without steroid treatment for at least 2 months prior to participating in the study. Approval was obtained from the local institutional ethics review board, and the subjects gave written informed consent prior to participation.
## Neuropsychological evaluation
Subjects underwent neuropsychological evaluation on the day of magnetic resonance (MR) scanning using an expanded Brief Repeatable Battery of Neuropsychological (BRB-N) tests. Executive functioning (EF, concept shifting test), verbal memory (VM, selective reminding test), verbal fluency (VF, word list generation), information processing speed (IPS, symbol-digit modalities test), visuospatial memory (VSM, spatial recall test), attention (Stroop colour-word test) and working memory (WM, memory comparison test) domains were included, as described in Eijlers et al.Raw cognitive scores were corrected for normal effects of sex, age and education.Z-scores were calculated based on the mean values and standard deviations of the HC group for each subject. For descriptive purposes, all Z-scores were averaged to form an 'averaged cognition' score.
## Magnetic resonance imaging
Subjects underwent 3T MRI (GE Signa HDxt), including three-dimensional (3D) T1-weighted fast spoiled gradient-echo (TR = 7.8 ms, TE = 3.0 ms, TI = 450 ms, FA = 12°, 0.9 mm × 0.9 mm × 1 mm voxel size) and a 3D fluid-attenuated inversion-recovery sequences (FLAIR, TR = 8000 ms, TE = 125 ms, TI = 2350 ms, 1.2 mm sagittal slices and 0.98 mm × 0.98 mm in-plane resolution). Diffusion tensor imaging based on echo planar imaging (EPI) covered the entire brain, using five volumes without directional weighting (i.e. b0) and 30 volumes with non-collinear diffusion gradients (i.e. 30 directions, b = 1000 s/mm 2 , TR = 13,000 ms, TE = 91 ms, FA = 90°, 53 contiguous axial slices of 2.4 mm and in-plane resolution 2 mm × 2 mm). Resting-state (i.e. eyes closed and no task) functional magnetic resonance imaging (MRI) covered the entire brain, using 202 volumes, of which the first two were discarded (EPI, TR = 2200 ms, TE = 35 ms, FA = 80°, 3 mm contiguous axial slices and in-plane resolution 3.3 mm × 3.3 mm).shows an overview of the processing pipeline.
Structural MRI processing: cerebellar GM volume As described in Eijlers et al.,WM lesions were segmented on FLAIR using the k-nearest neighbour classification with tissue-type priors (kNNTTP) yielding lesion maps for lesion volumes and for lesion filling on 3D-T1, to minimize their impact on processing steps. Normalized brain volumes (NBVs) were analysed using SIENAX (part of FSL5). The cerebellar region of interest (ROI) from the Harvard-Oxford atlas (part of FSL) was non-linearly co-registered to each subject's 3D-T1 using inverted FNIRT registration parameters. To specifically investigate cerebellar GM, the binary cerebellar mask was multiplied with SIENAX's GM partial volume estimation (PVE) image. This cerebellar GM PVE map was averaged to calculate the mean quantity of cerebellar GM, multiplied by the number of cerebellar GM voxels and normalized for head size with the V-scaling factor of SIENAX to obtain normalized cerebellar grey matter volume (GMV). The unthresholded cerebellar PVE image was binarized to form the cerebellar cortical ROI for diffusion and FC measurements (see below).
Structural MRI processing: cerebellar cortical fractional anisotropy Diffusion images were pre-processed using eddy current and motion correction with FSL, providing fractional anisotropy (FA) maps. Boundary-based registration (BBR) was used to calculate registration parameters between b0 images and 3D-T1 images. These were inverted and applied on the cerebellar ROI, using nearest-neighbour interpolation, to calculate mean FA within the cerebellar cortex.
## Functional mri processing and atlas
Functional pre-processing used FSL, including basic motion correction and smoothing. Advanced motion correction was subsequently performed using ICA-AROMA, as well as WM and cerebrospinal fluid (CSF) regression and high-pass filtering (100 seconds cut-off); resting-state data were kept in subject space. Cortical regions were defined using the standard space Brainnetome atlas. Similar to the cerebellar mask, the atlas was registered to subject space using inverted non-linear registration parameters and masked with binarized SIENAX-derived GM PVE maps, before adding deep grey matter (DGM) regions derived from FIRST and the cerebellar mask. The complete GM atlas was then co-registered to the subject's functional scan, using an inverted BBR matrix. All registration steps used nearest-neighbour interpolation. As described in Meijer et al.,for each subject, effects of EPI distortion were assessed by calculating the number of voxels within each ROI that contained reliable signal, excluding those with <30% coverage, resulting in the exclusion of bilateral orbitofrontal and nucleus accumbens areas. All subjects had at least 60% cerebellar coverage on functional magnetic resonance imaging (fMRI); mean cerebellar coverage was not significantly different between groups and was >75% in each group. For both static and dynamic connectivity measurements (see below), connections were averaged into seven well-known resting-state networks according to maximum overlap with the previous literature.These variables therefore represent static and dynamic cerebellar connectivity with the default-mode (DMN), fronto-parietal (FPN), ventral and dorsal attention (VAN and DAN), sensorimotor (SMN), visual (VN) and DGM networks. In addition, these variables were averaged to represent one global measure of static and dynamic cerebellar connectivity.
## Functional mri processing: static and dynamic fc
For each of the remaining regions of interest in the atlas, the average signal intensity was calculated for each volume, creating 197 averaged time-series. Static FC was calculated by correlating cerebellar time-series with each of the 196 cerebral atlas regions. Negative correlations were made absolute. Whereas static FC represents the strength of the functional connection that is measured across the entire scan, dynamic FC is a measurement of variability of this functional connection strength over time.Such a measure can be interpreted as a measure of stability (i.e. a low variability) or flexibility (i.e. a high variability) of functional communication, although it should be noted that it remains unclear whether a high or low time-varying FC is to be considered 'optimal', and that this variability in fact seems to be network and state-dependent.To calculate dynamic FC, an in-house MATLAB script was used, 11 which uses a sliding window approach in order to calculate FC values in a range of partially overlapping windows for each of the investigated cerebellar functional connections. Subsequently, the coefficient of variation of these connectivity values was calculated across time windows, by dividing the standard deviation across windows by the average FC, and used as a normalized measure of dynamics. Similar to previous studies, a window length of 60 seconds and a shift of 9 seconds were used. 11 FC variables were compared with null-model data to assess whether the observed dynamics were statistically different from random noise. These models were created using phase-randomization of our data,and FC was averaged over 50 randomization runs. Randomized variables were compared to the empirical data using paired t-tests.
## Statistical analyses
All statistical analyses were performed in SPSS 22. Normality was assessed using the Kolmogorov-Smirnov testing and histogram inspection. Since linear models ideally incorporate variables with normal distributions, some variables needed mathematical transformation to achieve normality (static FC with log 10 (x), due to a right-tailed distribution, and dynamic FC with x 2 , due to a left-tailed distribution, see Supplementary . Multivariate general linear models (GLMs) compared imaging measures between groups, including age, sex and level of education as covariates. Significant cerebellar measures were subsequently related to EDSS and cognition using two multivariate linear regression models with backward selection. Individual cognitive domains were related to FC variables that were significant in the cognition model only, using Pearson's correlations. All reported p-values are Bonferroni-corrected for multiple comparisons. Disease duration was not related to FC, but did relate to average cognition (r = −0.24, p < 0.001), EDSS (r = 0.44, p < 0.001), as well as cerebellar cortical FA (r = −0.25, p < 0.001) and volume (r = −0.40, p < 0.001). Cerebellar lesion volume was not related to cognition, EDSS or FC, but did relate to cerebellar atrophy (rho = −0.13, p = 0.031) and a trend for cerebellar cortical FA (rho = −0.12, p = 0.053).
# Results
## Demographics, cognition and disability
## Null models
Randomized data were generated for all FC variables and compared to the empirical data. While all static FC measures remained unaffected, all dynamic FC measures were significantly different from their randomized counterparts (all p < 0.001).
# Discussion
In this study, we investigated at which disease stage cerebellar alterations, that is, group differences in FA, GM volume and FC, become apparent in relapse-onset multiple sclerosis and how these explain cognitive impairment (CI) and disability. Cerebellar damage was mild in early RRMS, while late RRMS showed signs of cerebellar atrophy, which further worsened in SPMS. Static FC (i.e. strength of connectivity) was only lower in SPMS, while dynamic FC (i.e. variability of connectivity) was only higher in SPMS, compared to early RRMS and controls. Cerebellar dynamic FC (but not static FC) and cerebellar atrophy together explained 32% of disability (together with age and education) and 24% of cognitive variance (together with education). Dynamic connectivity correlated with disability, WM and VM and IPS.
Lower static FC of the cerebellar cortex was only seen in SPMS and only with the DMN and FPN. Both networks are known to be structurally connected with the cerebellum.Interestingly, no early connectivity changes were found. This could contradict the hypothesized construct of functional reorganization,where functional activation and connectivity are thought to increase to compensate for structural damage. Instead, these findings support the hypothesis of a network collapse in progressive MS.The few studies that assessed cerebellar function mostly showed severe cerebellar connectivity alterations in SPMS only.In this study, in late RRMS, cerebellar FC was still normal, while cerebellar cortical volume and FA were already lower compared to controls. These structural differences were possibly driven by demyelination and a loss of tissue organization due to loss of Purkinje cells.While some previous studies also showed cerebellar atrophy 2 and WM integrity loss 26 in RRMS, effects were shown to be especially evident in SPMS.Together, these findings indicate that structural cerebellar damage becomes apparent before functional cerebellar network alterations, which seem to develop in the transitional period towards SPMS. This finding could support the notion of a 'tipping' point in the severity of structural cerebellar damage in the later stages of RRMS, after which the network will destabilize. This sudden 'network collapse' could hereby explain the sudden cognitive worsening in progressive MS.Future work is required, however, to further study these effects over time, in order to pinpoint how brain function is able to remain normal in earlier stages of the disease and the specific underpinnings of this altered FC.
Dynamic FC was markedly altered in many more networks compared to effects of static FC, and again almost exclusively in SPMS. Strongest relations with cognition were found for dynamic cerebellum-DMN FC. Previous work showed that in SPMS, atrophy is prominent in regions that are known DMN hub areas, which is much worse than in RRMSand might explain this DMN specificity. We only observed higher dynamic FC, which might indicate that the connection between the cerebellum and DMN may not be sufficiently stable to properly process information, a process similar to that hypothesized to occur in brain damaged patients with a lowered level of consciousness.This increase could be an attempt to preserve normal functioning, that is, that the cerebellum and DMN continuously attempt to reconnect to preserve normal processing. However, it could also merely be the result of some form of disinhibition and not any form of active reorganization. However, our measure of dynamic FC was limited to the coefficient of variation over time of a specific connection. As such, more in-depth dynamic FC analyses could provide additional information, for instance, investigating specific cerebellar connectivity patterns and how these are organized in time. These so-called functional 'metastates' have previously been indicated to be important for cognition in MS, but have not been explored in the cerebellum.The specificity of altered static cerebellar connectivity with the DMN and FPN could also be explained by results of previous tract-tracing studies, showing strong structural connections between the cerebellum and these brain areas.Correlations with WM specifically, as seen in this study, seem valid, given that the FPN is directly involved in WM,as is the cerebellum itself.Interestingly, cerebellar connectivity was also related to reduced IPS in this study as well as other recent MS works,although the cerebellum is not traditionally implicated in this cognitive domain in studies on the healthy brain.This could represent a 'bleed-through' effect of WM on the symbol-digit modalities test, and/or the other way around, as well as a possible role for the cerebellum in fine-tuning cognition by affecting aspects of IPS.
Some limitations should be acknowledged. First, very early MS could not be investigated, where data remain rare. In addition, due to sample size, we could not group controls into age bins, although we included age as a covariate. Furthermore, we used one cerebellar mask, while segmenting cerebellar lobules could provide additional information. For instance, it is known that cortico-cerebellar connections are extensively mediated by the thalamus,and a previous study found reduced thalamic connectivity with specific parts of the cerebellum in MS.Specifically investigating the integrity of the cortico-thalamo-cerebellar circuit in MS might also increase statistical contrast. Finally, longitudinal assessments could investigate individual trajectories of progression, including assessments of abnormalities in specific network states.In conclusion, these novel findings indicate that cortico-cerebellar FC is especially affected in SPMS, focused on the default-mode and DGM networks. Lower static FC is accompanied by higher variability in FC strength, the latter of which especially relates to the cognitive and physical decline in this phase of the disease. These findings indicate the importance of including the cerebellum in studies investigating cognitive dysfunction, while future longitudinal studies are now required to further investigate the prognostic value and cause of these findings. |
Modest Increase in Fertility Consultations in Female Adolescents and Young Adults with Lymphoma: A Population-Based Study
# Introduction
T he adolescent and young adult (AYA) age group, 15-39 years of age, presents unique challenges in oncology. Hematological malignancies (HMs) make up 21%-34% of cancer diagnoses among this age group. Since the 5-year survival rate for HMs diagnosed before the age of 39 ranges from 50% to 95%, research foci are shifting to improve quality of life (QOL) after cancer.An influencer of QOL is fertility, which can be affected by oncologic diagnoses and treatments. Greaves et al. demonstrated that HM survivors were more likely to remain childless than the general population.In Hodgkin's lymphoma (HL), the most common HM in the AYA population, premature ovarian failure occurs in up to 37% of patients.To address these concerns, the discipline of oncofertility emerged with the hopes to accelerate the inclusion of fertility in oncology care.The American Society of Clinical Oncology (ASCO) has released guidelines necessitating fertility discussions in oncology care for AYA patients.However, oncology patients continue to report dissatisfaction.Specifically, in AYA patients with HM, rates of infertility discussions vary from 17% to 83%.Factors that affect these rates include patient demographics, physician characteristics, and complex societal factors.While the need for fertility consultations led to the ASCO guideline updates, the impact of these guidelines is unknown. Similarly, patient surveys cannot capture the scope of care in a population. Our group published a population-based analysis of referral rates for AYA, breast cancer patients showing a modest improvement after the 2013 guideline.This article investigates a similar trend in lymphoma patients by capturing the province of Ontario. We determined patient, physician, and socioeconomic factors associated with fertility consultation rates, highlighting areas for improvement.
# Methods
This population-based cohort study of AYAs diagnosed with lymphoma from January 2000 to March 2018 identified AYAs residing in Ontario, Canada's largest province by population (13.2 million) using the Ontario Cancer Registry (>98% of cancer incidents).Patients with a history of infertility, sterilization procedures, or previous cancer and patients ineligible for health insurance were excluded.
Datasets were linked by encoded identifiers and analyzed at ICES (www.ices.on.ca). The primary outcome, gynecology consultation about fertility, between the diagnosis of HM and commencement of chemotherapy, was identified as a gynecology consult billed as an International Classification of Diseases, Ninth Revision (ICD-9) code 628 (infertility diagnosis).
Patient and physician demographics were retrieved from the Registered Persons Database and ICES Physician Database, respectively. Parity was defined as a previous live birth (MOMBABY dataset). Chemotherapy billing codes identified treatment commencement. Community deprivation and income quartile scores measured socioeconomic factors.
Individual characteristics were analyzed using chi-square analyses for categorical variables, one-way analysis of variance (ANOVA) for continuous variables, and Kruskal-Wallis tests for medians. Logistic regression examined how factors influenced fertility consults, adjusting for confounding variables through a backward selection modeling. All statistical tests were two sided with p < 0.05 for significance. Data were analyzed using SAS version 9.4 (Cary, NC). This study was approved by the Queen's University Health Sciences and Affiliated Teaching Hospitals Research Ethics Board.
# Results and discussion
This population-based study assessed fertility consultation rates in female AYAs with lymphoma in Ontario, Canada.
A total of 2088 female AYAs diagnosed with lymphoma were identified.presents sociodemographic characteristics with the adjusted model in. HL represented 1323 (63%) of the study population, whereas, 765 (37%) had non-Hodgkin's lymphoma (NHL). The overall referral rate (3.4%) was low, ranging from below 1% before 2006 to 7.9% in 2014-2018, but similar to our previous study of breast cancer.There was a significant increase in fertility consultation rates across the study period ( p for trend <0.001).
A survey-based study investigating HM patients between 1957 and 2006 in Britain showed a slightly higher rate of 12% with improvement after 2000.However, fertility consultation in that study was publicly funded. In contrast, universal funding for fertility treatment began in Ontario in 2015. Referral rate in a private, American system for HMs, breast and gastrointestinal malignancy combined, was 5% between 1993 and 2007, similar to our study.Notably, we used fertility billing codes, which preclude referrals that were suggested, but not completed. While this code, ICD-9 code 628, is the only one for infertility and required for billing insured services in Ontario, its use as a substitute for fertility consultations has not been validated. Furthermore, it does not represent discussions had by other health care professionals, including oncologists, patient flow coordinators, and nursing staff. Together, these limitations may contribute to the low rates observed.
Fertility consults increased after the 2013 ASCO update ( p < 0.001). This modest improvement is consistent with the literature that demonstrates knowledge gaps even after guidelines are released.Campbell et al. identified that 35% of pediatric oncologists surveyed read the 2013 guideline. 14 In addition to guidelines, other mechanisms to promote oncofertility include the following: referral pathways, multidisciplinary teams, and provider education.One Canadian initiative is the establishment of the Canadian National Task Force on Adolescents and Young adults with Cancer, in 2008, which identified improving fertility outcomes in cancer survivors as a priority issue.Patient and socioeconomic factors may affect care. In this study, fertility consultations were not altered by patient age or cancer type. Several studies demonstrate that pediatric and older-AYA (30-39 year old) patients have reduced odds of referral than patients in their 20s.In both this study and the literature, patients with a history of childbirth were less likely to have a fertility consult than nulliparous women (odds ratio [OR] 0.34, 95% confidence interval [CI] 0.16-0.73).Since economic factors alter referral patterns, in Ontario, future research into the impacts of public funding is warranted.In Ontario, public funding of one cycle of in vitro fertilization for a female patient is covered under the Ontario Fertility Program, introduced in December 2015. Prior to this program, the consultation was publicly funded, but treatment was not. Thus, some patients may not have followed through with a consultation if they knew they could not afford fertility preservation.
Likewise, income quintile did not alter fertility consultation odds; however, more complex socioeconomic factors (e.g., education and employment) captured by a higher deprivation score decreased odds of consultation (OR 0.55; 95% CI 0.31-0.96). Being educated was similarly associated with increased referral trends in a review by Loren and Senapati, 9 as well as heterosexuality and being Caucasian. This suggests social determinants of health influence oncofertility care.Fertility consultation was more likely if the time between diagnosis and chemotherapy was >6 weeks (OR 2.03, 95% CI 1.25-3.1). However, this result does not mean that fertility referrals delay treatment. The median time to chemotherapy was 5 weeks (IQR 3-8) and 6 weeks (IQR 4-9) for patients with and without a fertility consultation, respectively, thus demonstrating a modest delay (p = 0.02). Notably, we did not measure cancer stage, and therefore, cannot comment if urgency affects odds of referral. One possibility is that a less severe stage allows for more time for a referral. The literature and guidelines suggest early referrals and referral pathways reduce delays in treatment.Furthermore, more recent ovarian stimulation protocols allow for less delay.Several studies demonstrate that physician characteristics, including sex, attitudes, age, and specialty, affect referrals.Of the physicians who did refer, 60% were female, a higher proportion than female hematologists in Canada (<1/3 in the early 2000s increasing to 48% in 2018).Of the physicians who made a referral, 55% were older than 45. The most common specialty to refer was hematology (48%), which aligns with the target audience of the guidelines, followed by family physicians (20%) and medical oncology (11%).
Fertility consultation improved after the ASCO guidelines for female AYAs with lymphoma. By identifying referral trends, we can tailor future interventions toward demographics in need. Despite the modest improvement demonstrated in this study, oncofertility research demonstrates vast potential for improving care.
# Author disclosure statement
No competing financial interests exist.
# Funding information
The study received funding support from the Faculty of Health Sciences, Queen's University. This study was supported by ICES, which is funded by an annual grant from the Ontario Ministry of Health and Long-Term Care (MOHLTC). Parts of this material are based on data and information compiled and provided by MOHLTC, Ontario Health (Cancer Care Ontario), the Canadian Institute of Hospital Information, and the Ontario Health Insurance Program. The analyses, conclusions, opinions, and statements expressed herein are solely those of the authors and do not reflect those of the funding or data sources; no endorsement is intended nor should be inferred. |
Setting boundaries for genome-wide heterochromatic DNA deletions through flanking inverted repeats in Tetrahymena thermophila
Eukaryotic cells pack their genomic DNA into euchromatin and heterochromatin. Boundaries between these domains have been shown to be set by boundary elements. In Tetrahymena, heterochromatin domains are targeted for deletion from the somatic nuclei through a sophisticated programmed DNA rearrangement mechanism, resulting in the elimination of 34% of the germline genome in ∼10,000 dispersed segments. Here we showed that most of these deletions occur consistently with very limited variations in their boundaries among inbred lines. We identified several potential flanking regulatory sequences, each associated with a subset of deletions, using a genome-wide motif finding approach. These flanking sequences are inverted repeats with the copies located at nearly identical distances from the opposite ends of the deleted regions, suggesting potential roles in boundary determination. By removing and testing two such inverted repeats in vivo, we found that the ability for boundary maintenance of the associated deletion were lost. Furthermore, we analyzed the deletion boundaries in mutants of a known boundary-determining protein, Lia3p and found that the subset of deletions that are affected by LIA3 knockout contained common features of flanking regulatory sequences. This study suggests a common mechanism for setting deletion boundaries by flanking inverted repeats in Tetrahymena thermophila.
# Introduction
Chromatin structures regulate gene expression, maintenance and transmissions in eukaryotes and are often organized in domains [bib_ref] Heterochromatin structure and function, Dillon [/bib_ref]. Heterochromatic domains are condensed and silent in transcription with distinctive molecular components. The DNA packaged in these domains can be defined by specific boundary elements, the loss of which leads to spreading of the heterochromatic state into the neighboring region [bib_ref] Insulators and boundaries: versatile regulatory elements in the eukaryotic genome, Bell [/bib_ref]. Understanding the molecular nature of domain boundary control is critical to the study of gene activities in chromosomes. In ciliated protozoa, one major form of heterochromatin is believed to govern programmed deletion of thousands of specific DNA segments, thus offering a special setting in which to understand the regulation of chromatin boundaries.
Several cis-acting boundary elements have been described in a diverse array of eukaryotes. So-called insulators have been shown to block the propagation of heterochromatin and regulate gene expression [bib_ref] Braking the silence: How heterochromatic gene repression is stopped in its tracks, Donze [/bib_ref]. The propagation of heterochromatin is restricted between the E and I silencers at the silent mating type loci (HML and HMR) in Saccharomyces cerevisiae. In fission yeast, there are two inverted repeats flanking the silent region of the mating type loci. Deletions of these elements caused the methylation of histone H3 on lysine 9 (H3K9) to spread into adjacent sequences [bib_ref] Transitions in distinct histone H3 methylation patterns at the heterochromatin domain boundaries, Noma [/bib_ref]. In the Drosophila 87A7 heat-shock locus, flanking sequences, scs and scs' (specialized chromatin sequences) contain the binding sites for proteins Zw-5 and BEAF-32, which are responsible for the insulator function [bib_ref] Visualization of chromosomal domains with boundary element-associated factor BEAF-32, Zhao [/bib_ref]. The highly-conserved protein CTCF (CCCTC-binding factor) has been showed to bind to insulators and block enhancer activities in vertebrates [bib_ref] CTCF is a uniquely versatile transcription regulator linked to epigenetics and disease, Ohlsson [/bib_ref] [bib_ref] Structural and functional conservation at the boundaries of the chicken -globin domain, Saitoh [/bib_ref]. These results suggest that the interaction between the cis-acting boundary elements and the specific targeting DNA-binding proteins are important in limiting heterochromatin propagation.
Tetrahymena thermophila carries out massive DNA deletions that are regulated by chromatin structures. The organism displays nuclear dualism, with a somatic (macro-) and a germline (micro-) nucleus present in the same cell. The macronucleus contains the necessary genetic information for vegetative cell growth and division, and the micronucleus contains all of the inherited genetic materials. During the growth phase, the macronucleus undergoes amitotic division and is actively transcribed while the micronucleus divides by typical mitosis and is transcriptionally silent. During conjugation, the micronucleus goes through mitosis, meiosis, and cross fertilization to generate zygotic nuclei, which further divide and develop into new macro-and micronuclei [bib_ref] RNA-Guided DNA deletion in tetrahymena: an RNAi-based mechanism for programmed genome rearrangements, Yao [/bib_ref] [bib_ref] Programmed genome rearrangements in tetrahymena, Yao [/bib_ref]. The developing new macronucleus undergoes a series of dramatic programmed DNA rearrangements, including the elimination of ∼34% of the genome (from 157 to 104 Mb) and the fragmentation of the five micronuclear chromosomes into about 225 minichromosomes that are retained in the macronucleus [bib_ref] Programmed genome rearrangements in tetrahymena, Yao [/bib_ref].
Tetrahymena programmed DNA rearrangement was first revealed through comparative genomic DNA hybridization studies [bib_ref] Comparison of the sequences of macro-and micronuclear DNA of Tetrahymena pyriformis, Yao [/bib_ref]. Large amounts of sequences were selectively eliminated from the developing new macronucleus, implicating an intricate mechanism of regulation. Two globally occurring processes were later found: IES (internal eliminated sequence) deletion and chromosome breakage, with IES deletion responsible for eliminating the bulk of the germ-line specific sequences. Several lines of evidence have revealed an RNA-guided DNA deletion mechanism that uses small RNAs to guide chromatin modifications to the DNA segments to be targeted for removal [bib_ref] RNA-Guided DNA deletion in tetrahymena: an RNAi-based mechanism for programmed genome rearrangements, Yao [/bib_ref] [bib_ref] Programmed genome rearrangements in tetrahymena, Yao [/bib_ref]. During conjugation, bidirectional transcripts are generated from selected regions of the micronuclear genome and processed into small RNAs [bib_ref] Nongenic, bidirectional transcription precedes and may promote developmental DNA deletion in Tetrahymena..., Chalker [/bib_ref] [bib_ref] Small RNAs as guardians of the genome, Malone [/bib_ref] [bib_ref] Analysis of a piwi-related gene implicates small RNAs in genome rearrangement in..., Mochizuki [/bib_ref]. These small RNAs target the homologous sequences in the developing macronucleus to trigger histone H3K27 and H3K9 methylation [bib_ref] Communication between parental and developing genomes during tetrahymena nuclear differentiation is likely..., Chalker [/bib_ref] [bib_ref] RNAi-dependent H3K27 methylation is required for heterochromatin formation and DNA elimination in..., Liu [/bib_ref] and recruit other proteins including Pdd1p, a HP1-like chromodomain protein [bib_ref] RNAi-dependent H3K27 methylation is required for heterochromatin formation and DNA elimination in..., Liu [/bib_ref] [bib_ref] Pdd1p, a novel chromodomain-containing protein, links heterochromatin assembly and DNA elimination in..., Madireddi [/bib_ref] [bib_ref] Methylation of histone h3 at lysine 9 targets programmed DNA elimination in..., Taverna [/bib_ref]. The Pdd1p-containing complex in turn recruits Tpb2p, a domesticated piggyBac transposase [bib_ref] A domesticated piggyBac transposase plays key roles in heterochromatin dynamics and DNA..., Cheng [/bib_ref] to execute IES excision [bib_ref] Sequence microheterogeneity is generated at junctions of programmed DNA deletions in Tetrahymena..., Austerberry [/bib_ref]. The broken ends are rejoined through a nonhomologous end-joining (NHEJ) pathway [bib_ref] An essential role for the DNA breakage-repair protein Ku80 in programmed DNA..., Lin [/bib_ref] and result in deletion junctions with certain degrees of sequence microheterogeneity, probably generated from the cutting or the rejoining process. Recent studies have discovered a minor pathway that utilize two other domesticated piggyBac transposases, TPB1 and TPB6, to eliminate a small subset of IESs that target terminal sequences instead of heterochromatin to carry out precise deletion [bib_ref] The piggyBac transposon-derived genes TPB1 and TPB6 mediate essential transposon-like excision during..., Cheng [/bib_ref] [bib_ref] A germline-limited piggyBac transposase gene is required for precise excision in Tetrahymena..., Feng [/bib_ref].
Since most IES deletions are controlled by heterochromatin, there are probably special domain boundaries to limit the extents of deletions. The nature of this boundary determination mechanism remains largely unknown. It is interesting to note that DNA deletions in Tetrahymena can be induced to occur at random locations by dsRNA injection. However, these deletions lack defined boundaries (with variations up to several kbs), and implied the existence of boundary regulatory sequences in natural deletions [bib_ref] Programmed DNA deletion as an RNA-guided system of genome defense, Yao [/bib_ref]. Indeed, previous studies have reported the existence of flanking regulatory sequences (FRSs) that help determine the boundaries of several IESs. The well-characterized M element has two alternative left boundaries and one shared right boundary [bib_ref] Sequence structures of two developmentally regulated, alternative DNA deletion junctions in Tetrahymena..., Austerberry [/bib_ref] [bib_ref] DNA elimination in tetrahymena: a developmental process involving extensive breakage and rejoining..., Yao [/bib_ref]. All three boundaries contain a 10bp polypurine sequence (5 -AAAAAGGGGG or A 5 G 5 ) in their flanking regions a short distance (∼45 bp) away and arranged in opposite orientations, thus appearing as a pair of inverted repeats (IR) for each deletion. Removal of this sequence resulted in the formation of highly variable deletion boundaries, and shifting its location caused the boundary to move with it. These results indicate that the polypurine IR serve as the FRSs of the M element [bib_ref] A programmed site-specific DNA rearrangement in Tetrahymena thermophila requires flanking polypurine tracts, Godiska [/bib_ref] [bib_ref] A distant 10-bp sequence specifies the boundaries of a programmed DNA deletion..., Godiska [/bib_ref]. Furthermore, recent studies have identified a protein, Lia3p, that recognizes A 5 G 5 sequences and affects the boundaries of the M-element and 4 other elements that also contained A 5 G 5 flanking sequences [bib_ref] A parallel G quadruplex-binding protein regulates the boundaries of DNA elimination events..., Carle [/bib_ref]. The depletion of LIA3 reduced progeny production after conjugation to 15%, revealing the functional importance of this G-rich sequence binding protein in IES deletions. Detailed analysis has also identified FRSs for the R-element, although their sequence identities have been more complex [bib_ref] Flanking regulatory sequences of the Tetrahymena R deletion element determine the boundaries..., Chalker [/bib_ref]. Moreover, additional analysis has suggested the presence of other FRSs in mse2.9 and Tlr1, which may also involve inverted repeats [bib_ref] Cis-acting requirements in flanking DNA for the programmed elimination of mse2.9: a..., Fillingham [/bib_ref] [bib_ref] A developmentally regulated deletion element with long terminal repeats has cis-acting sequences..., Patil [/bib_ref] [bib_ref] Alternate junctions and microheterogeneity of Tlr1, a developmentally regulated DNA rearrangement in..., Patil [/bib_ref] [bib_ref] A small family of elements with long inverted repeats is located near..., Wells [/bib_ref].
These cases suggest a possible general mechanism for IES boundary determination in Tetrahymena based on cisacting flanking sequences. Using the macronuclear and the micronuclear genome sequence information [bib_ref] Macronuclear genome sequence of the ciliate Tetrahymena thermophila, a model eukaryote, Eisen [/bib_ref] , it should be possible to test this idea at the genomic level. Here, we investigated the presence of FRSs for IES deletion using genomic sequences from different inbred strains. We found that the occurrences of deletion were mostly, though not always, conserved among strains and that their boundaries show different degrees of variations. We found specialized sequence structures near IES boundaries that could be linked to boundary determination, and experimentally determined the importance of the most prominent ones. Furthermore, we analyzed the macronuclear genomes of LIA3 mutants and found a large number of IESs that are affected by the mutation, and they appeared to share similar cisacting flanking IRs. This study suggests a general rule for IES elimination in Tetrahymena and reveals sequence structures that may mark chromatin domain boundaries. and then replicated to drops with specific drugs to identify progeny cells. Viable progeny cells were transfer to 96 well plates.
# Materials and methods
## Cell and cell culture
## Genomic dna sequencing and alignment
Genomic DNA was prepared using methods previously described [bib_ref] Nucleotide sequence structure and consistency of a developmentally regulated DNA deletion in..., Austerberry [/bib_ref]. We sequenced the genomes of inbred and Lia3 progeny strains to a depth of 49-60 million read-pairs with 2 × 100 bp using Illumina HiSeq 2000 paired-end sequencing (Illumina Inc., San Diego, CA, USA) at the BRC NGS Core Facility in Academia Sinica (Taiwan). Sequencing quality was measured using FastQC software (version 0.11.2; http: //www.bioinformatics.babraham.ac.uk/projects/fastqc). Quality scores across all bases were confirmed to be more than 30. Error corrections for reads were using Musket (version 1.0.6) (40). Sequence alignment was mapped into the MIC genome assembly data (Tetrahymena Comparative Sequencing Project BIoHaM, https://www.ncbi.nlm.nih.gov/ bioproject/?term=Tetrahymena%20broad%20institute) as the reference genome using BWA (version 0.7.15-r1140) [bib_ref] Fast and accurate short read alignment with Burrows-Wheeler transform, Li [/bib_ref] , and SAM/BAM file handling was done by SAMtools (version 1.3) [bib_ref] The Sequence Alignment/Map format and SAMtools, Li [/bib_ref]. The mapped reads were visualized using the Integrative Genomics Viewer (IGV) [bib_ref] Integrative Genomics Viewer (IGV): high-performance genomics data visualization and exploration, Thorvaldsdottir [/bib_ref] and analyzed using home-made Perl scripts. The raw sequence data sets have been deposited at NCBI BioProject (http://www.ncbi.nlm.nih.gov/bioproject) as PRJNA326452 and PRJNA416874.
## Ies identification
The deletions were first predicted by BreakDancer [bib_ref] BreakDancer: Identification of genomic structural variation from paired-end read mapping: BreakDancer: Identification..., Fan [/bib_ref]. The distribution of split reads that were extracted from the files of each strain was compared with the predicted deletions. Note that the hard clipping and the soft clipping were both considered, while the average of the clipping counts per position served as the threshold to remove false positives. The position that was near the predicted IES end (within 200-bp window) and had the highest split reads was considered as the reference IES end. Next, the deletions that were less than 100 bp and that contained the unknown nucleotides Ns at the IES ends were removed. The terminal direct repeats, which produce microhomology at each end after cleavage, were moved to the 'A-end' of each IES according to their orientation in the MIC genome sequences. The A-end and B-end of an IES refer to the ends that appear in the 5 and 3 side of the IES as they appear in the MIC genome sequences.
Two IESs within or among strains that share at least 1bp overlap were defined as two different forms of the same IES. The boundary variations among IES forms were determined by the sum of the difference at both ends between these two forms.
## Maximum boundary variation
To measure and categorize the variation among different forms of the same IES within and between cell strains, we summed up the length difference at both ends between any two forms. The maximum of these values between any pair of forms for an IES is defined as the maximum boundary variation for this IES. Hence, for an IES, let di f f (S i , S j ) be the length difference at both ends between two forms S i and S j . The maximum boundary variation of the IES is defined as
[formula] max S i ,S j ∈ f orms di f f S i , S j . [/formula]
## Flanking regulatory sequence identification
The 100-bp upstream and downstream of IES flanking regions were extracted and the reverse complement of the downstream sequences were used for searching IRs with identical sequences. IRs that were located on both ends and with similar distances to the reference IES ends of CU427 (less than 10-bp difference) were selected, and the occurrences at each position were calculated. The concentricity was defined by IQR (the interquartile range); IQR is represented by the range including the middle 50% of the population, i.e. the difference between the third quartile (75 percentile) and the first quartile (25 percentile). A lower IQR indicated that these IRs were more concentrated in IES flanking regions. The threshold of concentrated pentamer IRs was IQR ≤10 and count ≥3. [fig_ref] Table 1: Number of IESs exhibiting ≥100-bp boundary variations between the three inbred and... [/fig_ref]. DNA fragment of the normal IES with the same length of flanking sequences from supercontig 2.504 was copied from CU428 genomic DNA by PCR reaction. Supercontig 2.89 and supercontig 2.89 without TACCNT were cloned into the NotI site of the pD5H8 rDNA vector [bib_ref] A programmed site-specific DNA rearrangement in Tetrahymena thermophila requires flanking polypurine tracts, Godiska [/bib_ref]. These two insertions were at the opposite direction within the vector. Supercontig 2.504 with or without C-rich IRs were cloned between the PmeI and ApaI site of the pD5H8 rDNA vector.
Biolistic transformation is carried out as previous description [bib_ref] A programmed site-specific DNA rearrangement in Tetrahymena thermophila requires flanking polypurine tracts, Godiska [/bib_ref]. Briefly, DNA was coated on 0.6 m gold particle and delivered to mating cells CU427 and CU428 at 10 hours after mating was initiated using a Biolistic gun (BioRad PSD-1000/He). The transformants were selected by their resistance to paromomycin, and random clones were grown and either pooled or directly examined for their boundary variations using PCR and nucleotide sequencing.
# Results
## Iess are consistently deleted in different inbred strains
In order to understand IES boundary determination, we need to first compare the deletion of IESs among different Tetrahymena strains to determine their variations. The MAC genomes of three inbred B strains, CU427, CU428 and B2086 II (BII), were sequenced using Illumina paired-end sequencing. To locate IESs that were deleted, we mapped reads onto the MIC reference genome and used BreakDancer (44), a tool for predicting genomic structure variation, to detect deletions from the MIC genome in each MAC genome [bib_ref] Programmed minichromosome elimination as a mechanism for somatic genome reduction in Tetrahymena..., Lin [/bib_ref]. There were 10,127, 10,176 and 10,138 deletions detected in CU427, CU428 and BII, respectively. We observed that the deletion boundaries predicted by BreakDancer did not offer sufficient precision, hence, we improved the resolution by extracting split reads located at each junction and used them to identify the exact nucleotide position of the breakage point. Many deletions contained unknown nucleotides at the junction due to incomplete micronuclear genome sequences and were removed. After these refinements, 6913, 7031 and 7088 deleted segments were identified with high confidence in CU427, CU428 and BII, respectively [fig_ref] Figure 1: IES elimination among three Tetrahymena inbred strains [/fig_ref]. During this process, we observed that some deleted segments shared significant overlaps and should be considered alternative forms of the same IES, indicating that a small population of IESs have intra-strain variation. They were further verified by identifying the mapped reads across the junctions. Hence, the number of non-overlapping IESs identified were 6879, 6073, 7006 in CU427, CU428 and BII, respectively [fig_ref] Figure 1: IES elimination among three Tetrahymena inbred strains [/fig_ref] , including some well-defined IESs that are TPB1-dependent (Supplemental [fig_ref] Table 1: Number of IESs exhibiting ≥100-bp boundary variations between the three inbred and... [/fig_ref].
Next, we compared the occurrence of deletions among these strains, and found that the deletion of >95% of IESs (averaging 6733 IESs) are shared by any pair of strains (Supplemental [fig_ref] Figure 1: IES elimination among three Tetrahymena inbred strains [/fig_ref] and that more than 94% of IESs (6599 IESs) are deleted in all strains [fig_ref] Figure 1: IES elimination among three Tetrahymena inbred strains [/fig_ref]. This result indicates that the vast majority, but not all, of IESs are consistently deleted in independently developed macronuclei. We further analyzed the publicly available genome data of another B strain, SB210, and found a similar result (Supplemental [fig_ref] Figure 2: Nucleotide distribution near IES boundaries [/fig_ref].
Interestingly, there were 85 IESs that had, within a single strain, more than one form of deletion, and for one IES up to nine forms were found (Supplemental [fig_ref] Figure 3: IR of the motif 'TACCNT' at similar distance to both ends of... [/fig_ref] and B). Furthermore, for four of these IESs more than one form was found in all three strains, indicating the persistence of multiple rearranged forms at these loci. Note that the developing macronucleus has endoduplicated to a level of about 4-8C when IES elimination occurs, allowing up to eight different deletion forms to be generated at each IES location. Presumably different forms (like different alleles) should be sorted out through amitosis during macronuclear division. These inbred strains have been propagated asexually for many decades, and thus have ample opportunities for assortment. The retention of multiple forms including some that overlapped with expressed genes, especially in all strains, raised the possibility of functional roles for these boundary variations (Supplemental [fig_ref] Figure 3: IR of the motif 'TACCNT' at similar distance to both ends of... [/fig_ref].
## The majority of deletion boundaries show inter-strain microheterogeneity
Since the great majority of IESs were deleted in all inbred strains tested, we next examined their junction sequences for possible inter-strain variations. For each IES, the combined difference in length at both ends between any two strains was calculated and the maximum value was used to indicate the extent of its boundary variation [fig_ref] Figure 1: IES elimination among three Tetrahymena inbred strains [/fig_ref]. For instance, if there were three forms for an IES and the junction difference were 30, 40 and 50 bp between each pair of forms, this IES was put into the group with 41-to-50-bp variation. The results show that the junctions of deletion varied from 0 to 56,391 bp, with 14.71% of IESs showing no boundary variation [fig_ref] Figure 1: IES elimination among three Tetrahymena inbred strains [/fig_ref] and E), 38.87% exhibiting variations of 1-to-20-bp, and 27.79% differing by more than 100-bp [fig_ref] Figure 1: IES elimination among three Tetrahymena inbred strains [/fig_ref]. This result indicates that the majority (53.58%) of IESs showed very limited boundary variations during deletion (20 bp or less).
## Abrupt change in nucleotide distributions near ies boundaries suggests potential cis-regulatory sequences
To explore the possibility that cis-regulatory sequences are commonly used to determine IES boundary, we searched for nucleotide sequence patterns near mapped junctions. We first aligned all 6599 IESs according to their deletion boundaries and examined the nucleotide distribution at each position within 500-bp on each side of the reference end of CU427 (within the IES and in the flanking region). These regions contained a slightly lower GC content than the MIC genome average (25%GC), presumably due to the largely non-coding nature of IESs and their immediate flanking regions. Strikingly, abrupt and significant changes were observed for a short (∼50 bp) interval within the first 100-bp of the flanking regions. This interval includes the locations in which the flanking polypurine sequences of the M-element were located. This result strongly suggests that boundaries of a significant proportion of IESs are marked by special flanking sequences [fig_ref] Figure 2: Nucleotide distribution near IES boundaries [/fig_ref] , which have the potential to play a regulatory role.
## Inverted repeats near ies boundaries as potential regulatory sequences
To identify potential 'flanking regulatory sequences' that may help set the boundary, we searched for shared sequences with particular features, using motif finding tools, eTFBS and MEME [bib_ref] Discovering gapped binding sites of yeast transcription factors, Chen [/bib_ref] [bib_ref] MEME: discovering and analyzing DNA and protein sequence motifs, Bailey [/bib_ref]. For eTFBS, flanking regions within 100-bp from IES ends in CU427 were scanned to find 10 overrepresented motifs that contained the longest conserved sequences (Supplemental [fig_ref] Figure 4: IES elimination among Lia3 strains [/fig_ref]. The IES flanking regions between 100 and 200 bp away from the junctions were used as the background dataset. Most of the motifs identified had high AT patterns, supporting the higher AT content of the 100-bp flanking regions to the background. However, most of them did not display other common features, except two (motifs 2 and 7) that displayed a consistent distance to reference IES ends when occur as IR but not as direct repeats (DR). These two motifs share the same core sequence 'TACCNT' [fig_ref] Figure 3: IR of the motif 'TACCNT' at similar distance to both ends of... [/fig_ref]. Coincidently, the 'TACCNT' motif (Top 7) was also predicted as a significant motif by MEME using the sequences within 100-bp flanking regions of the IESs in CU427 (Supplemental [fig_ref] Figure 4: IES elimination among Lia3 strains [/fig_ref]. There were a total of 1881 copies of these motifs in the 100-bp flanking regions of all IESs, of which 198 occurred at both sides of an IES as IR and 57 as DR [fig_ref] Figure 3: IR of the motif 'TACCNT' at similar distance to both ends of... [/fig_ref] and G and Supplemental [fig_ref] Table 2: Numbers of LIA3-affected IESs in all three B strains [/fig_ref]. Significantly, these IRs occurred at similar distances (∼62 bp) to reference IES ends, with an 11 bp variation on average between the two sides of the same IES [fig_ref] Figure 3: IR of the motif 'TACCNT' at similar distance to both ends of... [/fig_ref]. This common pattern was not found for the DRs [fig_ref] Figure 3: IR of the motif 'TACCNT' at similar distance to both ends of... [/fig_ref].
To directly test whether the TACCNT motif acts as a FRS in vivo, we adopted an assay routinely used to examine the cis-requirement for IES excision and inserted an IES flanked by the TACCGT IR (referred to as T-domain for the following TACCNT IR) into an artificial rDNA mini chromosome transformation vector [bib_ref] A programmed site-specific DNA rearrangement in Tetrahymena thermophila requires flanking polypurine tracts, Godiska [/bib_ref]. After introduction of these vectors into Tetrahymena cells during conjugation, any deletion that occurred in this construct can be detected in the transformed progeny. IES constructs with or without the flanking T-domain were generated and tested and their deletion boundaries determined using PCR and nucleotide sequencing. As expected, the normal IES with the T-domain showed highly regulated boundaries in the clones analyzed. Consistent with the hypothesis that this sequence controls the accuracy of excision, the mutated IES lacking the T-domain lost it defined boundary as excision became highly variable [fig_ref] Figure 3: IR of the motif 'TACCNT' at similar distance to both ends of... [/fig_ref] , I and Supplemental . This result indicates that the TACCNT motif is an essential FRS that controls the boundary of this and likely other IESs with a similar flanking sequence motif.
We noticed interesting common features between the Tdomain and the polypurine sequences of the M element: they are both IRs at similar distances to respective reference IES ends. We thus repeated the search by focusing on IRs that were located at similar distances (less than 10-bp difference) from the two ends of an IES in the CU427 genome dataset. We arbitrarily defined an IR as a pair of pentamer sequences with no mismatches between the copies flanking each IES. We clustered these IRs and determined the distributions of their left copies relative to their proximal IES ends. Since the locations of two copies were similar to the reference IES ends, we assumed that the location distribution of the copy on the right-hand side was similar with the left-hand side at this step. We identified 472 pentamer sequences that occurred as IRs at the flanking regions. The pentamer 'ATTTT' IR occurred at the highest frequency; however, it was widely dispersed with no apparent pattern (Supplemental [fig_ref] Figure 5: G-rich IR and C-rich IR are positioned as the flanking regulatory elements... [/fig_ref]. On the other hand, we found 136 pentameric IRs with their distributions concentrated within a small range (Supplemental .
Interestingly, when some groups with high concentrated distributions that shared the same core sequence were combined, their concentricity was still maintained. There were 2700 IESs that contain pentamer IRs with the core sequence 'TATA', which was the most frequent group with high concentrated distribution (Supplemental and Supplemental . These pentamer IRs had a tight distribution that were about 65-bp away from the reference IES ends (Supplemental . Other cases also showed the same property of having the IR at similar distances from both ends of the IESs (Supplemental , implying a relevant relationship between the location of the IRs and the IES boundaries. In addition, the above identified Tdomain were also grouped as high concentrated IRs where both the 'TAC' and the 'TACC' groups include the 'TAC-CNT' IRs (Supplemental , indicating that this method can sufficiently identify the consensus of IRs that showed distinct patterns near the IES flanking regions. Moreover, pentamer IRs composed of G or C were also highly represented in the high concentrated groups (Supplemental . The common feature of the concentricity of the distance of IRs to the IES boundaries within the same groups suggests that these IRs may represent a type of FRSs for IES boundary determination.
Altogether, 3794 IESs were included in the 6 major IR groups mentioned above, which covered 57.49% of all IESs shared among the three inbred strains. This result implied that IRs could be the major type of regulatory sequences for IES boundary determination.
## Lia3p regulates a distinct subset of iess
Lia3p was recently shown to control the position of boundaries of the M element by binding to its G-rich FRS [bib_ref] A parallel G quadruplex-binding protein regulates the boundaries of DNA elimination events..., Carle [/bib_ref]. LIA3-deficient cells also exhibited imprecise deletion boundaries for five other IESs that had similar G-rich sequences as the M element. We suspected that Lia3p may control many more IESs, many of which could include the IESs we found with G-rich IRs (Supplemental . To reveal the spectrum of IESs with boundaries controlled by Lia3, we generated three progeny lines (3-1, 4-1 and 27-2) from the LIA3 strains and sequenced their macronuclear genomes to identify defects in IES boundaries [fig_ref] Figure 4: IES elimination among Lia3 strains [/fig_ref]. Since these LIA3 strains were also derived from the B inbred lines, we used the three B inbred lines described earlier for comparison (Supplemental . We noticed that the total number of IESs with >100-bp boundary variations was 11.08% higher in these mutant strains (38.87%) than in the inbred strains (27.79%) [fig_ref] Figure 4: IES elimination among Lia3 strains [/fig_ref] , suggesting that Lia3p may regulate a large number of IESs.
## G and c are enriched near the boundary of lia3-affected iess
To identify the subset of IESs with boundaries controlled by Lia3p, we compared IES variations among inbred strains and LIA3 progeny lines, which revealed 519 IESs that showed higher (by at least 100-bp) boundary variations in these LIA3 progeny lines [fig_ref] Table 1: Number of IESs exhibiting ≥100-bp boundary variations between the three inbred and... [/fig_ref]. They are referred to as 'LIA3-affected IESs' thereafter. To look for possible common motifs, we extracted the 100-bp flanking regions from both sides of IESs in this group and calculated the nucleotide ratios in each position. To reduce potential noises caused by IESs that were highly variable even in the inbred strains, we only considered a subset (387 of the 519 IESs) that showed at most 100-bp variation among the inbred strains (Supplemental . Remarkably, we found a distinct enrichment of G at positions 40-60 bp away from the reference IES ends of CU427 [fig_ref] Figure 4: IES elimination among Lia3 strains [/fig_ref]. It agrees very well with the characteristics of FRS of the M-element, and further supports their potential role in the regulation of deletion boundaries. Unexpectedly, we also observed a small peak of Cs at positions 25-40 bp from the reference IES ends [fig_ref] Figure 4: IES elimination among Lia3 strains [/fig_ref] , suggesting the possible existence of some C-rich motifs under Lia3p regulation.
## G-rich and c-rich inverted repeats at the flanking regions of lia3-affected iess
We then searched for common motifs within these 100bp flanking regions using MEME [bib_ref] MEME: discovering and analyzing DNA and protein sequence motifs, Bailey [/bib_ref] , and identified conserved G-rich sequences [fig_ref] Figure 5: G-rich IR and C-rich IR are positioned as the flanking regulatory elements... [/fig_ref]. To minimize background noise we only used the subset of 308 LIA3-affected IESs that showed very low variation (at most 20 bp) among the inbred strains. Note that the conserved sequences predicted from MEME combined both orientations of the motif (both G-rich and C-rich sequences). We then determined the enrichment of these G-rich or C-rich motifs in the 387 IESs that were affected by LIA3. As summarized in [fig_ref] Figure 5: G-rich IR and C-rich IR are positioned as the flanking regulatory elements... [/fig_ref] , there appears to be a strong correlation between LIA3 effects and the presences of G-rich or C-rich IRs with high concentricity, strengthening the possibility that Lia3p acts through these IRs.
To refine the IR sequences associated with LIA3 effects, we tested different similarities of PWM (position weight matrix) value from the consensus we built in [fig_ref] Figure 5: G-rich IR and C-rich IR are positioned as the flanking regulatory elements... [/fig_ref] and B in CU427 and found that in the G-or C-rich IRs with 75% similarity, >60% of Lia3-affected IESs contained one of the two IRs, but only ∼14% of the background IESs contained them (Supplemental . This is a very robust correlation. We thus set the PWM threshold at 75% similarity for subsequent experiments. It should be noted that if we lowered the threshold to 60% similarity, 91.21% of Lia3affected IESs were found to contain G-rich or C-rich IR (Supplemental . However, this would also increase the background to 73%, reducing the distinction between these two groups.
Next, we scanned the flanking regions of these 387 LIA3affected IESs for the two IRs. We analyzed all three B strains, and the results were quite similar. Significantly, 59.95% of Lia3-affected IESs had G-or C-rich IRs among all inbred strains, but only 11.29% of the background IESs had these IRs [fig_ref] Table 1: Number of IESs exhibiting ≥100-bp boundary variations between the three inbred and... [/fig_ref] , indicating a strong correlation between these IRs and LIA3 effects. It is noted that if the G-and C-rich IRs appeared in the same IES, the one with lesser distance differences between both ends of the IES was assigned as the FRS of the IES. However, the overlaps were rare. In CU427, only seven LIA3-affected IESs appeared to have the two IRs at both ends. Moreover, the distances between these IRs and the reference IES ends were very similar among IESs [fig_ref] Figure 5: G-rich IR and C-rich IR are positioned as the flanking regulatory elements... [/fig_ref] , and especially between the two ends of the same IESs (P-value<10 −5 on average). This consistency was absent from those IESs unaffected by LIA3 (but have G-or C-rich IRs) [fig_ref] Figure 5: G-rich IR and C-rich IR are positioned as the flanking regulatory elements... [/fig_ref]. a Difference between IESs among three inbred strains (WT) is less than or equal to the indicated number of base pairs (bp), and the IES differences between the three WT and three Lia3 strains are ≥100 bp. b Similarity of PWM score of the consensus indicated in [fig_ref] Figure 5: G-rich IR and C-rich IR are positioned as the flanking regulatory elements... [/fig_ref] We also considered the G-rich or C-rich motifs as DRs. We identified 41 G-rich DRs and 45 C-rich DRs in the group of LIA3-affected IESs. However, their distances to the reference IES ends were less consistent (53.32 bp ± 23.24 in G-rich DRs and 50.12 bp ± 22.54 in C-rich DRs). In addition, the two copies flanking the two ends of an individual IES showed higher distance variation for DRs (∼25 bp in G-rich DRs and about 17 bp in C-rich DRs) than for IRs, making DRs less likely to serve as boundary regulatory elements. Our results show that LIA3-affected IESs are likely regulated by the IRs of G-rich or C-rich sequences.
To determine if the predicted C-rich motif is indeed a FRS, we generated constructs of an IES with the C-rich IR and its mutant without the IR, and examined their deletion boundary maintenance in vivo. The results show that the boundaries become highly variable in the mutant lacking the C-rich IRs [fig_ref] Figure 5: G-rich IR and C-rich IR are positioned as the flanking regulatory elements... [/fig_ref] and J). In conclusion, we found that not only the G-rich IR, but also the C-rich IRs function as FRSs in LIA3-affected IESs.
## Multiple flanking regulatory sequences exhibited in lia3affected iess with g-rich and c-rich irs
Some IESs that are likely controlled by Lia3p actually show large boundary variations even in normal strains. This could indicate that some IESs have relaxed boundary control or, alternatively, some IESs may exhibit precisely controlled alternative boundaries. This scenario has been described for the M element, which can undergo two equally likely deletion outcomes, removing either 0.6-and 0.9-kb [bib_ref] Specific DNA rearrangements in synchronously developing nuclei of Tetrahymena, Austerberry [/bib_ref]. The two forms have the same right boundary but different left boundaries that are 0.3-kb apart, and each boundary contain the 5 -A 5 G 5 motif positioned ∼45 bp away [bib_ref] A distant 10-bp sequence specifies the boundaries of a programmed DNA deletion..., Godiska [/bib_ref]. Consistent with the possibility that junction variability in wild-type cells represents control of alternative junctions, closer examination revealed that variable junctions each had copies of the same putative FRSs. We observed that a potential FRS could usually be found at a consistent distance to an IES boundary even if the boundary variation was high, suggesting that the same tight distance control was maintained. Supplemental showed another example of a LIA3-affected IES that contained one copy of the G-rich IRs near the right junction and two copies near the left junctions. The two forms of deletion in CU427 and BII used the same outer pair of IR and generated deletions with only 1-bp variation at the right junction, while the single form in CU428 used the inner pair of the IR and showed 61-bp difference at the left junction and 1-bp difference at the right junction from the other two strains. A simple survey revealed 40 and 8 IESs with multiple FRSs in the LIA3-affected IESs with G-and C-rich IRs, respectively, representing 72.73% and 88.9% of the respective group with more than 20-bp variation (Supplemental . Moreover, a T-domain containing IES with the highest level of variation [fig_ref] Table 2: Numbers of LIA3-affected IESs in all three B strains [/fig_ref] , CU427.Supercontig2.222.9221) was actually deleted as two segments that were 908-bp apart in CU428 and BII but as one continuous form in CU427 (Supplemental [fig_ref] Figure 3: IR of the motif 'TACCNT' at similar distance to both ends of... [/fig_ref] and . T-domains or its degenerated sequences were found in most of the flanking regions of these 3 IES forms, suggesting that alternative deletions can occur when several combinations of FRSs are available in the same region.
We described in an earlier section that nearly half of the IESs showed >20-bp variation. The alternative deletion just described may offer a potential explanation. Looking at the genome-wide situation, we were surprised to find that 5573 of these 6599 IESs (84.45%) showed little or no variations (≤20-bp variation) in at least one end. It is likely that there are limited numbers of defined potential boundaries for most IESs. This result implies that the majority of IES boundaries are well regulated and those that do vary may partly be due to the alternative use of multiple FRSs that are present in these IESs. Among the 380 LIA3-affected IESs, about 73% contained the alternative deletions in at least one end of the new boundary when LIA3 was mutated (Supplemental , raising the possibility that secondary FRSs and their interacting proteins are used to set boundaries when Lia3p is depleted. Altogether, our finding supports the mechanism that IES boundaries are determined by flanking regulatory sequences.
# Discussion
In this study we investigated the global regulatory mechanism of IES deletion boundary determination. We observed that the occurrence of deletions was highly, though not completely, conserved among different Tetrahymena strains. For those conserved IESs, the majority of deletion bound-aries exhibited microheterogeneity of 20 bp or fewer at each end. In searching for potential regulatory sequences we discovered that each of several IRs is present outside a subset of IESs, with each copy of the two repeats located at nearly equal distances from each end of an IES. These two copies likely work as a pair. Earlier studies that manipulated these sequences of the M-element have also suggested this possibility [bib_ref] A programmed site-specific DNA rearrangement in Tetrahymena thermophila requires flanking polypurine tracts, Godiska [/bib_ref] [bib_ref] A distant 10-bp sequence specifies the boundaries of a programmed DNA deletion..., Godiska [/bib_ref]. This finding suggests that the boundaries of these IESs can be determined by a mechanism with these IRs serving as flanking regulatory sequences. Thus, the majority of Tetrahymena IESs, which are specified by heterochromatin, could have their boundaries determined by flanking regulatory sequences to limit their variations.
Although the vast majority of IESs are deleted in all strains analyzed, there are 693 IESs that are deleted only in one or two strains. This interesting variation could be caused by at least two possibilities. Firstly, the execution of deletion could be inefficient and only some of the copies in the polyploid MAC are deleted. Random assortment of these copies during cell growth and amitosis could generate clones with or without the deletion. Secondly, the interesting epigenetic effects exerted by the parental MAC could inhibit deletion in some strains [bib_ref] Communication between parental and developing genomes during tetrahymena nuclear differentiation is likely..., Chalker [/bib_ref]. There is also a technical issue to consider that is related to the detection of IESs by BreakDancer. It could be less consistent in particular regions of the genome and contributed to this variation. We have randomly selected ∼10% of IESs from this group for analysis by individual inspection and can unambiguously verify ∼30% of them were true positives. It will be interesting to find out how these events are generated and whether this somatic diversification affects cellular fitness.
The nearly identical distances of the two copies of an IR to the IES ends raised the possibility that these two copies could cooperate with each other. We speculate that IRs interact with their binding proteins to set chromatin domain boundaries, which then recruit Tpb2p to cut at these ends. After Tpb2p directed excision, this structure could further protect the macronuclear-destined regions from nuclease digestion and maintain these two double-stranded ends in close proximity before they are joined by NHEJ . Since new boundaries that are formed after the removal of the T-domain, the C-Rich IR, or the depletion of Lia3p are not at entirely random locations but tend to be clustered, we favor the possibility that secondary or alternative FRSs are used once the predominant FRS becomes non-functional, so as to reduce the risk of spreading deletions to nearby coding regions.
Lia3p was the first known example of regulatory proteins to interact with FRSs, and provides an excellent stage from which to further analyze this process. Our genomic analysis of LIA3 mutants revealed a large group of potential IES targets. Surprisingly, we observed that there are two FRSs in LIA3-affected IESs, namely G-rich and C-rich IRs. More than 90% of the LIA3-affected IESs contained at least one of the FRSs under our threshold of 60% similarity of the PWM score, whereas about 60% of them contained one of the FRSs under a threshold of 75% similarity. This finding indicates that almost all of the 387 LIA3-affected IESs contained G-rich or C-rich IRs, though some exhibited lower similarity. Interestingly, the distances between the IRs and the IES boundaries differed between G-rich and C-rich IRs . A speculation on boundary regulation of IES deletion. We hypothesize that once the heterochromatin is formed through the small RNA mediated process, its boundaries are set by the interaction between the FRS binding protein such as Lia3p (light blue ovals) and the FRSs (arrow). Together they recruit other proteins (pink, purple and orange circle) including Tpb2p that carries out DNA cutting and IES excision. After IES deletion, the FRS binding protein may also protect the macronuclear-destined region and maintain the two broken ends in close proximity to facilitate the NHEJ process. Without FRSs, the heterochromatin boundaries may spread out, and Tpb2p will cut at variable point to excise IESs [bib_ref] Programmed DNA deletion as an RNA-guided system of genome defense, Yao [/bib_ref]. Furthermore, the broken ends are not well protected and are eroded before rejoining, causing additional boundary variation.
(51 and 38 bp, respectively). A recent study showed that Lia3p preferentially binds to single-stranded sequences with five guanine residues, which forms a parallel G-quadruplex in vitro [bib_ref] A parallel G quadruplex-binding protein regulates the boundaries of DNA elimination events..., Carle [/bib_ref] , but its ability to bind C-rich sequences is very poor, suggesting that Lia3p bind to the G strand in both G-rich and C-rich IRs. Interestingly, the represented motif in our study was 'GAGGG', which has been shown to have the most unstable form for maintaining the G-quadruplex structure [bib_ref] Guanines are a quartet's best friend: impact of base substitutions on the..., Gros [/bib_ref] , suggesting potential structural differences from the conventional G-quadruplex. We suspect that the different orientation of the G strand toward the IESs between these two IR types may affect Lia3p dimerization and alter the distances from the TPB2 cutting site.
Our results suggest that Tetrahymena has evolved a special way to harness transposases for IES eliminations. The domesticated piggyBac transposases TPB1/6 are responsible for the excision of 12 special IESs with features of transposons (such as terminal inverted repeats [TIR] and the TTAA cutting site) [bib_ref] The piggyBac transposon-derived genes TPB1 and TPB6 mediate essential transposon-like excision during..., Cheng [/bib_ref]. Tpb2p, on the other hand, has lost its ability to recognize the TIR and has evolved to broaden its target sites by recognizing heterochromatin to cut at its boundaries. Here we revealed many IRs in the genome that could serve as potential FRSs to define the excision boundaries. We speculate that FRSs and their binding proteins may have evolved from existing regulatory components of DNA activities, such as transcription factors and their binding sequences or other transposable elements. They are adopted by the IES elimination machinery once a new heterochromatic region arises from genetic agents that has invaded the genome.
To support the idea derived from sequence analysis, we directly tested two of the newly identified FRSs, TACCNT and C-rich motif, and clearly demonstrated their functions to set boundaries in vivo. Interestingly, when the flank-ing 'TACCNT' IR was removed and the boundary became highly variable, we noticed that there were FRS-like IRs adjacent to the newly formed boundaries (Supplemental , suggesting that a similar mechanism is involved in setting a new boundary. Our search discovered that about 60% of IESs contained one of the six main groups of IRs in their flanking regions. By using this system, we argue that TPB2dependent IES elimination could regulate >6000 IESs by targeting heterochromatin whilst also maintain high degrees of boundary precision.
For the IESs that do not belong to the six groups, we assumed that different kinds of sequence structures might be present in the flanking regions of IESs but were hard to detect through sequence analysis (e.g. the R element). In addition, in this first genome-wide search, we choose to use a stringent method for the motif discovery process, which did not cover all IESs but only the highly confident ones. Our study provides a comprehensive understanding of the IR-regulated IES boundary determination. Further studies will hopefully reveal its detailed mechanism.
This study further supports a remarkable similarity between the mechanism of programmed DNA rearrangements of Tetrahymena and that of the V(D)J recombination of the vertebrate adaptive immune system [bib_ref] Recombination centres and the orchestration of V(D)J recombination, Schatz [/bib_ref]. They both target inverted repeats and use domesticated transposases to perform excisions, and repair the break by the NHEJ pathway. The variable combination among the VDJ region resembles the alternative deletion of the IES region. Importantly, as a consequence every Tetrahymena cell has a different genome sequence with potentials to adapt to environmental changes, much like the adaptive immunity generated by B and T cells. This convergent evolution is interesting and could advance our understanding of vertebrate immune system and Tetrahymena biology.
## Data availability
The raw sequence data sets have been deposited at NCBI BioProject (http://www.ncbi.nlm.nih.gov/bioproject) as PRJNA326452 and PRJNA416874.
[fig] Figure 1: IES elimination among three Tetrahymena inbred strains. (A) Numbers of IESs and IES forms in the B inbred strains CU427, CU428 and BII. (B) The Venn diagram shows that the majority of IESs are shared among the three inbred strains. (C) Schematic illustration of the calculations used for IES variations at one location. S: form; X and Y: the boundary variation on each side. (D) Boundary variation classes of IESs. Note that intra-strain variations at the same location are included. (E) The histogram shows the distribution of IESs with boundary variation within 20 bp. [/fig]
[fig] Figure 2: Nucleotide distribution near IES boundaries. The plot shows the nucleotide distribution of the first 500-bp sequences of all IESs next to an end and the adjacent 500-bp flanking sequences in CU427. Sequences surrounding both ends of all IESs were used in the compilation. Zero indicates the boundary of IESs. The upper and lower dashed lines indicate the average genomic contents of A or T and G or C, respectively. [/fig]
[fig] Figure 3: IR of the motif 'TACCNT' at similar distance to both ends of IESs. (A) Conserved sequence of 'TACCNT'. (B) A cartoon shows the arrangement of IR that flanks an IES. (C) Tight distance distribution of the motifs as IRs near IESs in the CU427 genome. (D) Statistic information of the 'TACCNT' IRs in CU427. (E) A cartoon shows the arrangement of DR. (F) Distance distribution of the motifs as DRs near IESs in CU427. (G) Statistic information of the 'TACCNT' DRs in CU427. A-distance: distance of motif to one end of the IES; B-distance: distance of motif to the other end of the IES; distance difference: difference of the distances of the motif to either end of the IES; s.d.: standard deviation. (H) PCR of genomic DNA isolated from clones of IESs with or without the flanking T-domain. Dark arrow: expect arranged form; gray arrow: unspecific band. (I) Diagram of IES regions based on the sequencing result. Blue arrow: position of the primer set. Tw: single clone of WT IES with T-domain; Tm: single clone of mutated IES without T-domain. Arrow: primer. [/fig]
[fig] Figure 4: IES elimination among Lia3 strains. (A) Numbers of IESs and IES forms in Lia3 strains 3-1, 4-1 and 27-2. (B) Venn diagram showing that the majority of IESs are shared among the three Lia3 strains. (C) Boundary variation classes of IESs in the LIA3 strains. Note that intrastain variations at the same location are included. (D) IES Boundary variation classes within 20 bp in LIA3 strains. (E) A plot shows the nucleotide distribution of flanking sequences near both ends of 387 Lia3-affected IESs in CU427. To generate this figure, we used a stringent definition of Lia3affected IESs, i.e. those having ≤100-bp variation among the three inbred strains and increase by >100-bp variation among the three LIA3 strains. The upper and lower dashed lines indicate the average content of A or T and G or C, respectively. [/fig]
[fig] Figure 5: G-rich IR and C-rich IR are positioned as the flanking regulatory elements in LIA3-affected IESs. (A and B) Shared sequence motifs (analyzed by MEME) in 308 LIA3-affected IESs that show very limited variation (20-bp or lower) among the inbred strains. LIA3 strains increased the variations by at least 100-bp. (B) represents the reverse complement of the motif in (A). (C and D) Cartoons show the arrangement of G-rich or C-rich IR, respectively. (E and F) Motifs of the G-rich IR and C-rich IR in LIA3-affected IESs that increased the variations by at least 100-bp of IESs that show lower degrees of variation among inbred strains. We defined 75% similarity as the minimum score of the PWM indicated in (A) for the threshold. (G and H) Motifs of the G-rich IR and C-rich IR from the entire IES dataset, respectively. The numbers of LIA3-affected IESs indicate in (E) and (F) have been removed. A-distance: distance of the motif to one end of the IES; B-distance: distance of the motif to the other end of the IES; s.d.: standard deviation. (I) PCR of genomic DNA isolated from clones of IESs with or without the flanking C-rich IRs. Dark arrow: the expected rearranged form for normal deletion. (J) Diagram of IES regions based on the sequencing result. Blue arrow: position of the primer set. Cw: single clone of WT IES with C-rich IRs; Cm: single clone of mutated IES without C-rich IRs. Arrow: primer. Noted that the proximal (blue) and the distal (green) reversed primers in the right flanking region were individually paired with the forward primer in the left flanking region in separate PCR tests. [/fig]
[table] Table 1: Number of IESs exhibiting ≥100-bp boundary variations between the three inbred and three Lia3 strains [/table]
[table] Table 2: Numbers of LIA3-affected IESs in all three B strains [/table]
|
Shexiang Baoxin Pills as an Adjuvant Treatment for Chronic Heart Failure: A System Review and Meta-Analysis
Background. Shexiang Baoxin pills (SXBXP), as a Traditional Chinese Medicine, are widely used for chronic heart failure in China. It is essential to systematically assess the efficacy and safety of SXBXP as an adjuvant treatment for chronic heart failure. Methods. Seven English and Chinese electronic databases (PubMed, Embase, Cochrane Library, CBM, Wanfang, VMIS, and CNKI) were searched from inception to July 2017. The Cochrane Risk of Bias tool was used to evaluate the methodological quality of eligible studies. Meta-analysis was performed by Review Manager 5.3. Results. A total of 27 RCTs with 2637 participants were included in this review. Compared to conventional treatment, SXBXP combined with conventional treatment showed potent efficacy when it came to the total efficacy rate (OR, 3.88; 95% CI, 2.87, 5.26; < 0.00001), B-type natriuretic peptide (BNP) (MD = −66.95; 95% CI, −108.57, −25.34; = 0.002), N-terminal pro-brain natriuretic peptide (NT-ProBNP) (MD = −0.15; 95% CI, −0.21, −0.09; < 0.00001), six-minute walking distance (6-MWD) (MD = 38.57; 95% CI, 28.47, 48.67; < 0.00001), cardiac output (CO) (MD = 0.84; 95% CI, 0.68, 0.99; < 0.00001), and Stroke Volume (SV) (MD = 7.43; 95% CI, 4.42, 10.44, < 0.00001).The pooled subgroup analysis indicated that there was a significant difference between SXBXP plus conventional treatment and conventional treatment alone in short term course (OR = 3.51; 95% CI, 2.28, 5.40; < 0.00001), in middle period of treatment (OR = 5.01; 95% CI, 2.61, 9.60; < 0.00001), and in long-term course (OR = 3.77; 95% CI, 2.13, 6.67; < 0.00001). No serious adverse events or reactions were mentioned in these RCTs. Conclusions. As an adjuvant drug, this study suggested that SXBXP provide an obvious efficacy for the treatment of CHF. However, due to small samples and generally low quality studies being applied in this study, more rigorous and well-designed RCTs are needed to confirm these findings.
# Introduction
In spite of a tremendous advance in pharmacology and therapies, chronic heart failure (CHF) remains the most serious cardiovascular disorder all over the world [bib_ref] ESC guidelines for the diagnosis and treatment of acute and chronic heart..., Mcmurray [/bib_ref] [bib_ref] 2013 ACCF/AHA guideline for the management of heart failure: executive summary: a..., Yancy [/bib_ref]. There is a significant high mortality in patients with CHF [bib_ref] ESC guidelines for the diagnosis and treatment of acute and chronic heart..., Ponikowski [/bib_ref] , probably an estimated 50% mortality in 5 years [bib_ref] Heart Disease and Stroke Statistics, Benjamin [/bib_ref]. Moreover, with the aging of the population becoming a more and more serious issue, patients with CHF constitute a high proportion of the aging population, and so patients with CHF increase year after year, which highlights the urgent need for effective treatment strategies [bib_ref] Forecasting the impact of heart failure in the United States: a policy..., Heidenreich [/bib_ref].
CHF is a complex clinical syndrome that results from structural change or functional abnormalities, which may lead to a series of cardiac dysfunctions, such as decrease of cardiac output, increase of intracardiac pressure function, ventricular filling, or impaired ejection at both rest and load conditions [bib_ref] Meta-analysis of the relation of body mass index to all-cause and cardiovascular..., Sharma [/bib_ref]. Cardinal manifestations are various, such as dyspnea and fatigue, which may lead to fluid retention due to limited exercise tolerance and which may bring about pulmonary and peripheral edema [bib_ref] Time to redefine PD? Introductory statement of the MDS Task Force on..., Berg [/bib_ref]. CHF is the terminal stage of various heart diseases, often accompanied by high morbidity and high mortality, especially in the elderly [bib_ref] Galectin-3 is an independent marker for ventricular remodeling and mortality in patients..., Lok [/bib_ref]. Patients with CHF died within 5 years as well as the survival rate being less than 50% within 1 year [bib_ref] Chronic heart failure and aging-effects of exercise training on endothelial function and..., Sandri [/bib_ref]. CHF poses a serious challenge to the health of the people of the world, giving the family and society a heavy burden.
Current Western medicine treatment formed with "angiotensin-converting enzyme inhibitor (ACEI) or Angiotensin receptor antagonist (ARB), beta blockers, or 2 Evidence-Based Complementary and Alternative Medicine Aldosterone receptor antagonist" is the basis of the Golden Triangle treatment program [bib_ref] 2013 ACCF/AHA guideline for the management of heart failure: executive summary: a..., Yancy [/bib_ref]. However, satisfactory results are still difficult to obtain in some patients. A lot of researches show that, in CHF patients with Yang deficiency blood stasis, SXBXP have beneficial Qi Tong Yang Huayu efficacy and played a very effective role in the treatment of CHF. These Western medicine treatments are not only conventional treatments, but also the dominating treatment. However, the existing treatment is not perfect enough [bib_ref] Natriuretic peptide-guided therapy in chronic heart failure: a meta-analysis of 2,686 patients..., Savarese [/bib_ref]. It is well known that long-term use of Western medicine may cause side effects and resistance [bib_ref] Tetramethylpyrazine-mediated regulation of CXCR4 in retinoblastoma is sensitive to cell density, Wu [/bib_ref].
SXBXP, as a traditional and complementary medicine, derives from traditional decoction named Suhexiang pills, which is recorded in the Ministry of Health and the benefits of the party side in Song Dynasty. The traditional decoction has been used for more than a thousand years and it includes Moschus, toad, Panax ginseng, Bos taurus domesticus Gmelin, Cinnamomum cassia Presl, and Borneolum [bib_ref] Pharmacokinetics and tissue distribution of five bufadienolides from the Shexiang Baoxin pill..., Huang [/bib_ref]. From 1981, since the clinical application, SXBXP was widely used in coronary heart disease, angina, myocardial infarction, and other heart diseases. Most clinical trials showed that SXBXP benefited patients with CHF [bib_ref] Shexiang Baoxin Pills for coronary heart disease in animal models: Preclinical evidence..., Zhang [/bib_ref]. Now, a large number of references reported the efficacy of SXBXP on CHF. All the trails did not report obvious side effects and adverse reactions. However, evidence was very limited on the efficacy of SXBXP for CHF [bib_ref] Simultaneous determination of four volatile compounds in rat plasma after oral administration..., Chang [/bib_ref]. In fact, the previous systematic review did not assess the effects of SXBXP as an adjuvant treatment for CHF, too. Therefore, it is necessary for us to assess the efficacy and safety of SXBXP, which act as an adjuvant treatment with conventional treatment.
# Materials and methods
## Search strategy.
Comprehensive searches were conducted in both English and Chinese databases to identify all published RCTs from their inception to July 2017. All relevant RCTs were searched from the following 7 databases including PubMed, Embase, Cochrane Library, CBM, Wanfang, VMIS, and CNKI. The following search terms were used: "Shexiang Baoxin pills" [Title/Abstract] AND "Chronic heart failure" [Title/Abstract] OR "Chronic heart disease" [Title/Abstract]. The literature searches were independently examined by two investigators (Taiwei Dong and Rong Ma) and disagreements were resolved by consensus as well as discussion. The bibliographies of included trials were searched for through references. However, the trials without English abstract would be translated by the investigator Taiwei Dong and checked by the investigator Rong Ma.
## Inclusion criteria. two authors (taiwei dong and rong
Ma) read the titles and abstracts of trials in all searched databases independently to assess the rationality for inclusion. The full text was further read to evaluate for the inclusion criteria. The inclusion criteria were as follows. (1) Investigative object and intervention: all the randomized controlled trails (RCTs) which combined SXBXP with conventional medical treatment (experimental group) compared with conventional medical treatment (control group) alone in CHF were included. The method of intervention was oral administration. (2) Characteristics of patients: patients diagnosed with CHF with New York Heart Association (NYHA) (The Criteria Committee of the New York Heart Association 1994) classified from II to IV were included.
(3) Outcome measures: total efficacy rate. The secondary outcome measures included left ventricular ejection fraction (LVEF), cardiac output (CO), Stroke Volume (SV), B-type natriuretic peptide (BNP), N-terminal pro-brain natriuretic peptide (NT-ProBNP), and six-minute walking distance (6-MWD). RCTs with one or more outcomes were included.
## Exclusion criteria.
The trials conforming to the following conditions were excluded: (1) reduplicative publications reporting the same trials; (2) nonrandomized controlled trials; (3) nonclinical experiments, reviews, literature research, mechanism research, or animal experiment; (4) controlled interventions combined with any other Chinese herbal medicine or acupuncture in control group or experimental group; (5) unavailable or incorrect data for meta-analysis; (6) patients with unclear functional classification; (7) trials with unclear evaluation indicators or basic data for statistic research.
## Data extraction. two investigators (taiwei dong and
Rong Ma) independently extracted the basic information such as title, published year, total cases, cases of experiment group and control group, interventions, outcome measures, NYHA classification, course of disease, safety evaluation, and ADEs or ADRs to conclusive tables. Relevant disagreements were resolved through discussion with investigator (Jian Wang). Symptom improvement was evaluated according to the Guidance for Clinical Research on New Drugs of TCM, Framingham criteria, American College of Cardiology/American Heart Association (ACC/AHA), or textbook criteria as long as the criteria met the international-used diagnostic criteria [bib_ref] The Criteria Committee of the New York Heart Association, Levin [/bib_ref].
## Risk of bias assessment and quality assessment.
The methodological quality of included RCTs was assessed by Review Manager 5.3 according to the Cochrane Risk of Bias tool. The methodological quality of each trial was evaluated by seven domains including random sequence generation (selection bias), allocation concealment (selection bias), blinding of participants and personnel (performance bias), blinding of outcome assessment (detection bias), incomplete outcome data (attrition bias), selective reporting (reporting bias), and other biases. The quality of each trial was classified as "high risk," "unclear risk," or "low risk." The trials that had insufficient information available to make a judgment were classified as unclear risk of bias. The trials with low risk of bias represented a good methodological quality and the trials with high risk of bias represented a low methodological quality. Any disagreement was settled through discussion with investigator (Jian Wang). Full-text articles for further eligibility (Cochrane Collaboration, Oxford, UK). For outcome measures, dichotomous variables were presented as odds ratio (OR) with 95% confidence intervals (CI), while continuous outcomes were expressed as mean difference (MD) with 95% CI. As a quantitative measure of inconsistency, the -square ( 2 ) statistic was used to assess heterogeneity. Fixed effect model was performed with minor heterogeneity when 2 was less than 50%. Random-effect model was applied when 2 was over 50%. A funnel plot was used for assessing the potential publication bias. Furthermore, subgroup analysis was performed due to course of treatment of SXBXP.
# Results
## Basic information
## Description of studies.
A total of 199 records were identified for preliminary screening after searching English and Chinese databases. All the included trials were conducted in China and published in Chinese. As shown in [fig_ref] Figure 1: Flow diagram for searching and selecting study [/fig_ref] , 147 records were reserved for further screening after removing 52 duplicated publications. For the preserved records, 79 obvious irrelevant literatures were excluded by reading the title and abstract. 68 full-text articles were used for further assessment. After reading the full text, 41 more literatures were excluded for the following reasons: participants not meeting the inclusion criteria ( = 16), improper grouping, outcomes, or pharmacy ( = 12), nonrandomized controlled trials ( = 5), and no data available for extraction ( = 8). Finally, 27 RCTs of SXBXP for CHF were included in this review. [fig_ref] Table 1: Principal characteristics of the studies included in the meta-analysis [/fig_ref] , a total of 27 RCTs with 2637 participants were included in this review. The control group consisted of 1313 patients, while the treatment group consisted of 1324 patients. All trials' sample sizes ranged from 56 to 181; sample size is large enough. The ages of the subjects were over 50 years. Moreover, all the trials included NYHA classification among II∼IV. As for the characteristics of intervention, the course of treatment varied from 24 days to 6 months. Only one trial did not mention the course of treatment [bib_ref] To observe the therapeutic effect of Shexiang Baoxin Pills on adjuvant treatment..., Wang [/bib_ref]. The baseline of patients in both groups was balanced.
## Study characteristics. as shown in
The treatment group used SXBXP combined with the same conventional treatment as control group; only one trial used Placebo combined with conventional treatment [bib_ref] Effect of Shexiangbaoxin pills on heart function in patients with chronic ischemia..., Ding [/bib_ref]. Two groups of all trails used the dose of 202.5 mg/d, although most doses are 135 mg/d; only one trial used 67.5 mg/d. SXBXP was given through oral administration three times daily in all included trials. The control group used conventional medical treatment alone, including ACEI, ARB, cardiac glycosides, diuretics, -receptor blockers, antialdosterone drugs, calcium channel blockers, or vasodilators.
None of the included trials reported death. All trails reported NYHA classification. Twenty-two trials reported LVEF. Eighteen trials reported ER and 6 MWT. Four articles reported NT-proBNP. Four articles reported BNP. Eight trails reported CO. Seven trails reported SV.
# Methodological quality.
The methodological quality of the included trials was generally poor. Five trials reported that random sequence was generated by a random number table [bib_ref] Effects of Shexiang Baoxin pills on cardiac function and vascular endothelialfunction in..., Sang [/bib_ref] and only one trail generated it by the envelope method [bib_ref] Study on randomized parallel control of shexiang baoxin pills combined with western..., Wang [/bib_ref] ; the remaining 21 trials only mentioned random allocation without any description of the method of randomization. There was no allocation concealment mentioned in all the articles. Only one trial mentioned blinding of participants and blinding of outcome assessment [bib_ref] Effect of Shexiangbaoxin pills on heart function in patients with chronic ischemia..., Ding [/bib_ref]. Two articles reported incomplete outcome data and selective reporting [bib_ref] Clinical observation on treatment of chronic heart failure with coronary heart disease..., Wu [/bib_ref] [bib_ref] Effects of Shexiang Baoxin pills on cardiac function and heart rate variability..., Ran [/bib_ref]. About 20% trials showed high risk of bias in incomplete outcome data. Other potential sources of bias were unclear. Therefore, the quality of all the included trails was graded as high risk of bias. The details of the methodological quality were presented in and [fig_ref] Table 2: Risk of bias assessment of included studies [/fig_ref].
## Publication bias.
Publication bias was assessed using a funnel plot based on the total efficacy rate reported in 18 trials. The funnel plot was asymmetrical, which indicated that the potential publication bias might influence the results of this review. The bias might result from these reasons: small sample size, poor quality, and a high proportion of positive results . [bib_ref] Effects of Shexiang Baoxin pills on cardiac function and vascular endothelialfunction in..., Sang [/bib_ref] [bib_ref] Study on randomized parallel control of shexiang baoxin pills combined with western..., Wang [/bib_ref] [bib_ref] Clinical observation on treatment of chronic heart failure with coronary heart disease..., Wu [/bib_ref] [bib_ref] Clinical Study on the Treatment of Chronic Systolic Heart Failure with Shexiang..., Xu [/bib_ref] [bib_ref] Effects of Shexiang Baoxin Pills on Chronic Heart Failure in Patients withOld..., Bao [/bib_ref] [bib_ref] Clinical observation on the effects of shexiang baoxin pills on coronary heart..., Han [/bib_ref] [bib_ref] Treatment of chronic heart failure with integrated traditional Chinese and western medicine..., Guo [/bib_ref] [bib_ref] Effects of Shexiang Baoxin pills on chronic heart failure, Zhu [/bib_ref] [bib_ref] Effects of Shexiang Baoxin pills on cardiac function and plasma nt-probnp in..., Zhong [/bib_ref] [bib_ref] Effects of Shexiang Baoxin pills on diastolic heart failure, Dong [/bib_ref] [bib_ref] Effects of Shexiang Baoxin pills on chronic heart failure, Wang [/bib_ref] [bib_ref] Effects of Shexiang Baoxin Pills On chronic congestive heart failure, Sun [/bib_ref] [bib_ref] The effects of combined traditional Chinese and western medicine on the treatment..., Quan [/bib_ref] [bib_ref] Effects of Shexiang Baoxin pills on chronic congestive heart failure, Li [/bib_ref] [bib_ref] Clinical study on the treatment of chronic heart failure with shexiang baoxin..., Gao [/bib_ref] [bib_ref] Clinical Observation on the treatment of chronic heart failure in patients with..., Li [/bib_ref] [bib_ref] Effects of Shexiang Baoxin pills on left ventricular function and serum, Yang [/bib_ref]. There were 831 patients in experiment group and 904 in control group. The result showed that there was no heterogeneity ( = 0.83, 2 = 0%) and the fixed effect model was adopted for analysis. As shown in forest plot, there was a statistically significant difference between SXBXP plus conventional treatment and conventional treatment alone in the total efficacy rate (OR, 3.88; 95% CI, 2.87, 5.26; < 0.00001) .
## Effects of interventions
## Secondary outcome measures
(1) 1B-Type Natriuretic Peptide (BNP) and N-Terminal Pro-Brain Natriuretic Peptide (NT-Pro BNP). Four trails with 390 patients assessed the therapeutic efficacy of NT-Pro BNP [bib_ref] Effect of Shexiangbaoxin pills on heart function in patients with chronic ischemia..., Ding [/bib_ref] [bib_ref] The effects of combined traditional Chinese and western medicine on the treatment..., Quan [/bib_ref] [bib_ref] Effects of Shexiang Baoxin Pills on chronic heart failure, Fang [/bib_ref] [bib_ref] Effects of Shexiang Baoxin pills on B-type brain natriuretic peptide and left..., Deng [/bib_ref]. Four trails with 388 participants assessed the effect of SXBXP plus conventional treatment in decreasing BNP in patients with chronic heart failure [bib_ref] Effects of Shexiang Baoxin pills on cardiac function and plasma nt-probnp in..., Zhong [/bib_ref] [bib_ref] Clinical study on the treatment of chronic heart failure with shexiang baoxin..., Gao [/bib_ref] [bib_ref] Clinical Observation on the treatment of chronic heart failure in patients with..., Li [/bib_ref] [bib_ref] Effects of Shexiang Baoxin pills on left ventricular function and serum, Yang [/bib_ref]. It has considerably high heterogeneity in BNP ( = 0.02, 2 = 71%), and there is no significant difference between SXBXP plus conventional treatment and conventional treatment alone on NT-Pro BNP ( = 1.00, 2 = 0%). The trials reported that SXBXP plus conventional treatment was superior to conventional treatment alone to reduce NT-Pro BNP (MD = −0.15; 95% CI, −0.21, −0.09; < 0.00001) [fig_ref] Figure 5: Forest plot of SXBXP plus conventional treatment versus conventional treatment in the... [/fig_ref] ) and BNP (MD = −66.95; 95% CI, −108.57, −25.34; = 0.002) [fig_ref] Figure 5: Forest plot of SXBXP plus conventional treatment versus conventional treatment in the... [/fig_ref].
## (2) the comparison of 6-minute walking distance (6-mwd).
A total of 15 studies with 1439 subjects reported the level of 6-MWD . There was considerable heterogeneity ( < 0.00001, 2 = 79%) and random-effect model was conducted for analysis. The result showed that SXBXP could substantially increase the level of 6-MWD compared with conventional treatment (MD = 40.15; 95% CI, 30.40, 49.91; < 0.00001) [fig_ref] 2. 6: Strategy for Data Synthesis and Analysis [/fig_ref]. It indicated a significant improvement of SXBXP for CHF in exercise endurance.
(3) Cardiac Output (CO) and Stroke Volume (SV). Seven trials with 704 participants [bib_ref] To observe the therapeutic effect of Shexiang Baoxin Pills on adjuvant treatment..., Wang [/bib_ref] [bib_ref] Effect of Shexiangbaoxin pills on heart function in patients with chronic ischemia..., Ding [/bib_ref] [bib_ref] Effects of Shexiang Baoxin pills on chronic congestive heart failure, Li [/bib_ref] [bib_ref] Clinical study on the treatment of chronic heart failure with shexiang baoxin..., Gao [/bib_ref] [bib_ref] Clinical Observation on the treatment of chronic heart failure in patients with..., Li [/bib_ref] [bib_ref] Effects of Shexiang Baoxin pills on left ventricular function and serum, Yang [/bib_ref] [bib_ref] Effects of Shexiang Baoxin pill on 40 cases of chronic cardiac insufficiency, Lv [/bib_ref] and seven trials with 535 individuals [bib_ref] Clinical observation on treatment of chronic heart failure with coronary heart disease..., Wu [/bib_ref] [bib_ref] Effects of Shexiang Baoxin Pills on Chronic Heart Failure in Patients withOld..., Bao [/bib_ref] [bib_ref] Effects of Shexiang Baoxin pills on chronic congestive heart failure, Li [/bib_ref] [bib_ref] Clinical Observation on the treatment of chronic heart failure in patients with..., Li [/bib_ref] [bib_ref] Effects of Shexiang Baoxin pills on left ventricular function and serum, Yang [/bib_ref] [bib_ref] Effects of Shexiang Baoxin pill on 40 cases of chronic cardiac insufficiency, Lv [/bib_ref] [bib_ref] Effects of Shexiang Baoxin pills on senile chronic congestive heart failure, Niu [/bib_ref] assessed the therapy on cardiac function of SXBXP plus conventional treatment in CO and SV, respectively. There was considerable heterogeneity in CO ( = 0.03, 2 = 56%) and SV ( < 0.00001, 2 = 85%) in trials. Meta-analysis with a randomeffect model showed that, compared with conventional treatment, SXBXP plus conventional treatment significantly enhanced the cardiac function. The pooled analysis indicated that there was a statistically significant difference between SXBXP plus conventional treatment and conventional treatment alone on CO (MD = 0.84; 95% CI, 0.68, 0.99; < 0.00001) [fig_ref] Figure 7: Forest plot of SXBXP plus conventional treatment versus conventional treatment in increasing... [/fig_ref] and SV (MD = 7.43; 95% CI, 4.42, 10.44, < 0.00001) [fig_ref] Figure 7: Forest plot of SXBXP plus conventional treatment versus conventional treatment in increasing... [/fig_ref].
## (4) left ventricular ejection fraction (lvef).
A total of 21 trials evaluated LVEF and were pooled with a random model . The heterogeneity of the LVEF study was considerably high ( < 0.00001, 2 = 89%). Pooled comparisons demonstrated that SXBXP plus conventional treatment had a statistically significant beneficial effect compared to conventional treatment alone in terms of LVEF (MD = 3.89; 95% CI, 2.70, 5.07, < 0.00001) . As an adjuvant drug, SXBXP improve the cardiac function of patients with CHF.
# Subgroup analysis
(1) The Total Efficacy Rate with Different Courses. Subgroup analysis was conducted to assess the efficacy of SXBXP plus conventional treatment on total efficacy rate according to different course of treatment. 8 trails with 681 participants were treated within 2 months [bib_ref] Effects of Shexiang Baoxin pills on cardiac function and vascular endothelialfunction in..., Sang [/bib_ref] [bib_ref] Study on randomized parallel control of shexiang baoxin pills combined with western..., Wang [/bib_ref] [bib_ref] Effects of Shexiang Baoxin pills on chronic heart failure, Zhu [/bib_ref] [bib_ref] Effects of Shexiang Baoxin pills on diastolic heart failure, Dong [/bib_ref] [bib_ref] Effects of Shexiang Baoxin pills on chronic heart failure, Wang [/bib_ref] [bib_ref] Clinical study on the treatment of chronic heart failure with shexiang baoxin..., Gao [/bib_ref] [bib_ref] Clinical Observation on the treatment of chronic heart failure in patients with..., Li [/bib_ref] [bib_ref] Effects of Shexiang Baoxin pills on left ventricular function and serum, Yang [/bib_ref] , which was regarded as short term course. The middle period of treatment included 4 trails with 285 patients; these participants were treated within 2-4 months [bib_ref] Clinical observation on treatment of chronic heart failure with coronary heart disease..., Wu [/bib_ref] [bib_ref] Clinical Study on the Treatment of Chronic Systolic Heart Failure with Shexiang..., Xu [/bib_ref] [bib_ref] Effects of Shexiang Baoxin Pills on Chronic Heart Failure in Patients withOld..., Bao [/bib_ref] [bib_ref] Effects of Shexiang Baoxin Pills On chronic congestive heart failure, Sun [/bib_ref]. Moreover, the last subgroup contained 5 trails with 392 patients ; these participants were treated for over 4 months but less than 6 months. The 2 statistic of short term course showed that there was minor heterogeneity among these trials ( = 0.76, 2 = 0%). It also demonstrated similar result of the middle period of treatment ( = 0.20, 2 = 36%) and the long-term course ( = 0.91, 2 = 0%). All trails used fixed effect model to pool the results. The pool analysis presented that there was a significant difference between two therapy methods in short term course (OR = 3.51; 95% CI, 2.28, 5.40; < 0.00001), in middle period of Heterogeneity: 2 = 220.27; 2 = 66.37, df = 14 (P < 0.00001); I 2 = 79% [fig_ref] 2. 6: Strategy for Data Synthesis and Analysis [/fig_ref] : Forest plot of SXBXP plus conventional treatment versus conventional treatment in increasing 6-MWD. [bib_ref] 2013 ACCF/AHA guideline for the management of heart failure: executive summary: a..., Yancy [/bib_ref] and are the criteria for the heterogeneity test; X is pooled mean difference; -◼-is mean difference and 95% CI.
## Experimental control
Lv .
# Discussion
The basic pathogenesis of CHF is myocardial pathology "reconstruction." CHF is the final stage of most heart diseases, but generally there are two processes that ultimately lead to CHF. One is the occurrence of myocardial death, such as myocardial damage caused by myocardial injury and severe myocarditis; the other is the neuroendocrine system caused by overactivation of systemic reactions [bib_ref] Disease management in the treatment of patients with chronic heart failure who..., Kalter-Leibovici [/bib_ref] [bib_ref] Shexiang Baoxin Pills for coronary heart disease in animal models: Preclinical evidence..., Zhang [/bib_ref]. SXBXP can improve the heart pump function of CHF patients, reverse ventricular remodeling, and improve activity tolerance. The activation of the renin-angiotensinaldosterone system (RAAS) and sympathetic nervous system leads to cardiac hypertrophy, which will result in ventricular remodeling and ultimately decompensation [bib_ref] New strategies for heart failure with preserved ejection fraction: The importance of..., Senni [/bib_ref]. Therefore, delaying this process is particularly important for improving the quality of life and prolonging the survival time of patients with CHF. The indexes such as CO, SV, 6 MWT, LVEF, BNP, and NT-Pro BNP reflect the cardiac function and cardiac output. These indexes can be used to evaluate the cardiac function of patients with CHF and judge the CHF patients' clinical efficacy [bib_ref] Cardiac output response to norepinephrine in postoperative cardiac surgery patients: Interpretation with..., Maas [/bib_ref]. This study found that the combination of SXBXP and conventional treatment appeared to be more effective and safer for the treatment of CHF compared with conventional treatment. It indicated that SXBXP was worthy of clinical application and promotion. SXBXP is a Chinese patent medicine. It is based on the work of Professor Ruihong Dai of Huashan Hospital, who belongs to Fudan University [bib_ref] Pharmacokinetic study of five ginsenosides using a sensitive and rapid liquid chromatographytandem..., Peng [/bib_ref]. A group of Western medicine experts used the Western standard research and development of pure Chinese medicine preparations; it contains Moschus 6%, ginseng extract 27%, bezoar 4%, cinnamon 24%, storsin 16%, toad 4%, and Dryobalanops 19%. Basic Research Certificate showed that storsin and Dryobalanops have slowed the heart rate and had the role of the lifting coronary artery spasm; Moschus extract can dilate blood vessels and enhance cardiac function effectively [bib_ref] The application of metabolomics in traditional Chinese medicine opens up a dialogue..., Cao [/bib_ref]. Ginseng saponins have antioxidant properties and positive muscle force and reduce the role of lipid [bib_ref] Chemical diversity of ginseng saponins from Panax ginseng, Shin [/bib_ref]. Toad has a strong effect with heart [bib_ref] Rapid acclimation to cold allows the cane toad to invade montane areas..., Mccann [/bib_ref]. This is the first systematic review to assess the effects of SXBXP as an adjuvant treatment for CHF. In this metaanalysis review, only two trails reported mild adverse reactions, and none of the included trails reported death. Thus, evidence is limited to make a summary of death and adverse reactions. So on the whole long-term use of SXBXP can be considered safe and effective. This result also proves several expert consensuses that the addition of Chinese medicines can improve the clinical symptoms and quality of life in patients with CHF, simultaneously maintain the Test for overall effect: Z = 4.57 (P < 0.00001)
Heterogeneity: 2 = 1.03, ＞＠ = 4 (P = 0.91); I 2 = 0% Study or subgroup : Forest plot of SXBXP plus conventional treatment versus conventional treatment according to different courses in total efficacy rate. [bib_ref] 2013 ACCF/AHA guideline for the management of heart failure: executive summary: a..., Yancy [/bib_ref] and are the criteria for the heterogeneity test; X is pooled mean difference; -◼-is mean difference and 95% CI.
cardiac function and reduce the rehospitalization rate, and depress mortality of patients with unique comment Western medicine treatment [bib_ref] Prescription pattern of Chinese herbal products for heart failure in Taiwan: A..., Tsai [/bib_ref]. In this study, SXBXP displayed positive impact on clinical efficacy via increasing levels of LVEF and 6-MWT and reducing levels of NT-pro BNP and BNP. It seemed that SXBXP could ameliorate the cardiac function in patients with CHF according to primary outcomes, secondary outcomes, and subgroup analysis according to different course of treatment in experiment group.
The study implemented strict inclusion and exclusion criteria. However, there were still statistical heterogeneity between some of the outcome indicators of the included trials, due to the main consideration and the limited sample size and the variation in the length of treatment. Five issues still remain in all RCTs from the results: (1) the amount of included trials is small, in addition to the lack of high-quality and large sample study; (2) quality is generally low, random application is less, and blind implementation is unknown; (3) long-term follow-up is lacking; most studies did not mention mortality and readmission rate; (4) most of the studies did not report adverse reactions. Therefore, more high-quality with long-term follow-up RCTs were required to elucidate the effectiveness and security of SXBXP for CHF in the future.
# Conclusions
In summary, as an adjuvant treatment for chronic heart failure, this study suggested that SXBXP have an obvious efficacy for the treatment of CHF. In the future, more multicenter, large sample size, randomized double-blind, longterm evaluation RCTs are needed to confirm the efficacy and mechanism of SXBXP for CHF.
[fig] 2. 6: Strategy for Data Synthesis and Analysis. The metaanalysis was performed by Review Manager 5. [/fig]
[fig] Figure 1: Flow diagram for searching and selecting study. [/fig]
[fig] Figure 2, Figure 3: Risk of bias assessment of included studies. Funnel plot of SXBXP plus conventional treatment versus conventional treatment on total efficacy rate in patients with CHF. [/fig]
[fig] Figure 5: Forest plot of SXBXP plus conventional treatment versus conventional treatment in the decrease of BNP and NT-Pro BNP.2 and are the criteria for the heterogeneity test; X is pooled mean difference; -◼-is mean difference and 95% CI. (a) is forest plot of NT-Pro BNP; (b) is forest plot of BNP. [/fig]
[fig] Figure 7: Forest plot of SXBXP plus conventional treatment versus conventional treatment in increasing CO and SV.2 and are the criteria for the heterogeneity test; X is pooled mean difference; -◼-is mean difference and 95% CI. (a) is forest plot of CO; (b) is forest plot of SV. [/fig]
[table] Table 1: Principal characteristics of the studies included in the meta-analysis. Wang et al. 2010 [/table]
[table] Table 2: Risk of bias assessment of included studies. [/table]
|
LDL-cholesterol lowering effect of a new dietary supplement: an open label, controlled, randomized, cross-over clinical trial in patients with mild-to-moderate hypercholesterolemia
Background: Hypercholesterolemia is a major risk factor for cardiovascular disorders and requires specific intervention through an adequate lifestyle (diet and physical exercise) and, if necessary, an appropriate drug treatment. Lipid-lowering drugs, although generally efficacious, may sometimes cause adverse events. A growing attention has been devoted to the correction of dyslipidemias through the use of dietary supplements. The aim of this study was to assess the lipid-lowering activity and safety of a dietary supplement containing monacolin K, L-arginine, coenzyme Q10 and ascorbic acid, named Argicolina (A), compared to a commercially available product containing monacolin K and coenzyme Q10, Normolip 5 (N). Methods: This was a single center, controlled, randomized, open-label, cross-over clinical study enrolling 20 Caucasian outpatients aged 18-75 years with serum LDL-C between 130 and 180 mg/dL. Patients assumed two different dietary supplements (A and N) both containing monacolin K 10 mg for 8 weeks each, separated by a 4-week wash-out period. Evaluated parameters were: Total cholesterol (Tot-C), low density lipoprotein cholesterol (LDL-C), high density lipoprotein cholesterol (HDL-C), triglycerides (TG), fasting blood glucose, aspartate aminotransferase, alanine aminotransferase, creatinekinase, gamma-glutamyl-transpeptidase, brachial arterial pressure and heart rate, measured at the start and at the end of each treatment period. Safety was monitored through the study.Results: LDL-C decreased by 23.3% during treatment with N (p < 0.0001) and by 25.6% during treatment with A (p < 0.0001); the LDL-C mean reduction was 36.4 (95% CI: 45,6-27,1) mg/dL during N treatment and 40.1 (95% CI: 49.2-30,9) mg/dL during A treatment. Tot-C decreased significantly (p < 0.0001) within each treatment period. HDL-C increase was negligible during A whereas it was significant during N. TG diminished markedly during A and not significantly during N. The difference between treatments was not statistically significant for all variables. No serious or severe adverse events occurred during the study.Conclusions:Our results confirm the clinically meaningful LDL-C lowering properties of monacolin K. At variance with a supplement already in the market (N), the novel association (A) of monacolin K with L-arginine, coenzime Q10 and ascorbic acid also produces a significant reduction of triglycerides without significant effects on HDL.Trial registration: ClinicalTrials.gov ID: NCT03425630.
# Background
Cardiovascular diseases (CVDs) rank first as the current leading cause of death. Identification, prevention and management of the risk factors for CVDs are therefore a priority in public healthcare programs [bib_ref] European guidelines on cardiovascular disease prevention in clinical practice: the sixth joint..., Piepoli [/bib_ref]. Among CVDs risk factors, one of the most important is dyslipidemia, primarily hypercholesterolemia, which may be corrected through an adequate lifestyle (diet and physical exercise) and, if necessary, an appropriate drug treatment [bib_ref] ACC/AHA guideline on the treatment of blood cholesterol to reduce atherosclerotic cardiovascular..., Stone [/bib_ref]. The results achieved with lipid-lowering drugs, however, are not always satisfactory, and untoward effects (mainly myalgia and myopathies) may sometimes emerge that cause patients to discontinue the treatment [bib_ref] Statin myotoxicity: a review of genetic susceptibility factors, Needham [/bib_ref] [bib_ref] Clinical response to statins: mechanism (s) of variable activity and adverse effects, Sirtori [/bib_ref]. Over the past few years a growing attention has been devoted to the correction of dyslipidemias through the use of dietary supplements, either because some patients may have milder forms of hypercholesterolemia or as an alternative to statins in patients who may have experienced side effects, although the potential adverse effects caused by supplements have not been fully investigated. There is evidence of a relationship between some food constituents and a reduction in CVD incidence [bib_ref] Food and vessels: the importance of a healthy diet to prevent cardiovascular..., Engelfriet [/bib_ref] [bib_ref] Resveratrol: French paradox revisited, Catalgol [/bib_ref]. Monacolin K is a substance obtained during rice fermentation by the fungus Monascus purpureus, used for thousands of years in China to produce rice wine. Fermented rice owes its red color (hence the name "red yeast rice") to various pigments produced by the fungus, including monascorubramin and rubropunctamin. In 1979 Akira Endo of the Tokyo University isolated a metabolite produced by the fungus with a strong inhibitory activity toward the enzyme HMG-CoA reductase (3-hydroxy-3-methyl-glutaryl coenzyme A reductase), which mediates the step from hydroxymethylglutaryl coenzyme A to mevalonic acid early in the pathway to the endogenous synthesis of cholesterol [bib_ref] A gift from nature: the birth of the statins, Endo [/bib_ref]. This metabolite was later characterised as a member of the monacolin group and identified as monacolin K, a structural analogue of lovastatin. Several clinical studies have demonstrated a lipid-lowering effect of monacolin K, alone [bib_ref] The effect of red yeast rice (Monascus Purpureus) in dyslipidemia and others..., Yang [/bib_ref] [bib_ref] Red yeast rice improves lipid pattern, high-sensitivity C-reactive protein and vascular remodelling..., Cicero [/bib_ref] or in association with other compounds [bib_ref] Effects of poly-byoactive compounds on lipid profile and body weight in a..., Solà [/bib_ref]. Studies have shown a reduction in serum total cholesterol (Tot-C) between 12 and 30% after the administration of products containing from 3 to 10 mg of monacolin K, for treatments ranging from 4 weeks up to 1 year [bib_ref] Red yeast rice improves lipid pattern, high-sensitivity C-reactive protein and vascular remodelling..., Cicero [/bib_ref] [bib_ref] Nutraceutical approach to moderate cardiometabolic risk: results of a randomized, double-blind and..., Ruscica [/bib_ref] [bib_ref] Long-term effects of nutraceuticals (berberine, red yeast rice, policosanol) in elderly hypercholesterolemic..., Marazzi [/bib_ref] [bib_ref] Efficacy and tolerability of a nutraceutical combination (red yeast Rice, policosanols, and..., Gonnelli [/bib_ref]. Some of these trials have reported a lower incidence of myopathies compared to statins [bib_ref] Red yeast rice for dyslipidemia in statin-intolerant patients: a randomized trial, Becker [/bib_ref]. A recent metanalysis on products containing monacolin K have shown that the expected mean reduction, compared to placebo, was 37.5 mg/dL (95%CI, 30.9-43.7; p < 0.00001) for Tot-C and 33.6 mg/dL (95%CI, 27.5-39.8; p < 0.00001) for LDL-cholesterol (LDL-C) [bib_ref] A meta-analysis of red yeast rice: an effective and relatively safe alternative..., Li [/bib_ref]. A novel dietary supplement named Argicolina (A) containing monacolin K, L-arginine, coenzyme Q10 and ascorbic acid has been developed by Damor Pharmaceuticals (Naples, Italy). L-arginine is an essential aminoacid, a substrate in the synthesis of nitric oxide (NO) catalysed by the NO-synthetase expressed in the vessel endothelium, therefore acting as a vasodilator and an inhibitor of platelet aggregation [bib_ref] Nitric oxide synthesised from L-arginine mediates endothelium dependent dilatation in human veins..., Vallance [/bib_ref] [bib_ref] L-arginine improves endothelium-dependent vasodilatation in hypercholesterolemic humans, Creager [/bib_ref] [bib_ref] The L-arginine-nitric oxide pathway: role in the atherosclerosis and therapeutic implications, Böger [/bib_ref] [bib_ref] Increase in fasting vascular endothelial function after short-term oral L-arginine is effective..., Bai [/bib_ref] [bib_ref] Mechanisms of disease: L-arginine in coronary atherosclerosis-a clinical perspective, Tousoulis [/bib_ref] [bib_ref] Arginine: a new therapy for atherosclerosis?, Cooke [/bib_ref]. Experimental work showed that the association of L-arginine with a statin increased NO production by endothelial cells compared to statin alone [bib_ref] Arginine or citrulline associated with a statin stimulates nitric oxide production in..., Berthe [/bib_ref]. Coenzyme Q has antioxidant activity and is widely used in dietary supplements for subjects with raised lipid levels or cardiovascular risk [bib_ref] Co-enzyme Q10 supplementation for the primary prevention of cardiovascular disease, Flowers [/bib_ref]. Ascorbic acid has antioxidant and vasoprotective activity [bib_ref] Review: when is an antioxidant not a antioxidant? A review of novel..., Duarte [/bib_ref] [bib_ref] Long-term orange juice consumption is associated with low LDL-cholesterol and apolipoprotein B..., Aptekmann [/bib_ref]. The diverse and complementary properties of the components of A suggest that this product may be useful in treating subjects with hyperlipidemia or cardiovascular risk. The aim of this study was to assess the lipid-lowering activity and safety of the proposed formulation in patients with mild-moderate hypercholesterolemia, compared to a commercially available dietary supplement containing monacolin K and coenzyme Q10 (Normolip 5: N) (ESI -Albissola Marina, Savona, Italy). The study was conducted according to a randomized cross-over design.
# Methods
## Patients selection
Between July 2016 and April 2017 eligible patients were recruited among the outpatients attending the Obesity Center of the Endocrinology Unit 1, Cisanello Hospital, Pisa, Italy. Patients aged 18-75 years with serum LDL-C between 130 and 180 mg/dL, not significantly modified by an appropriate dietetic regimen were considered eligible for the study. Exclusion criteria were: pregnancy or breast-feeding; known liver, renal or muscle diseases; serum triglycerides (TG) greater than 350 mg/dL; previous cardiovascular events; concomitant neoplastic or immunodepressive diseases; use of lipid-lowering drugs or dietary supplements within the last 3 weeks; concurrent use of thiazide diuretics, oral contraceptives containing estrogen or progestogen, systemic corticosteroids; use of psycho-active substances, drug or alcohol abuse; neurological or psychiatric diseases that could affect consent validity or impair the patient's adherence to the study protocol. Thirty patients, all Caucasian, were screened. Ten were excluded during the screening process because they did not fulfill all the inclusion criteria (screening failure). Twenty patients were thus randomized, 10 to the A > N sequence and 10 to the N > A sequence.
## Study design
The study was conducted in a single center according to a controlled, randomized, open-label, cross-over design. Each patient had to assume, in a randomized sequence, both treatments (A, 1 sachet/day; N, 1 capsule/day) for 8 weeks each separated by a 4-week wash-out period. The study plan included the initial screening visit (V1), an entry visit at start of the first treatment period (V2), a visit at the end of the first treatment period (V3, 56 ± 5 days after V2), a wash-out period of 4 weeks (±5 days), a Data are expressed as mean ± SD (min-max range)
visit at start of the second (crossed over) treatment period (V4), and a visit at the end of the second treatment period (V5, 56 ± 5 days after V4) [fig_ref] Figure 1: Study flow-chart [/fig_ref]. Tot-C, LDL-C, HDL-cholesterol (HDL-C), TG, fasting blood glucose, aspartate aminotransferase (AST), alanine aminotransferase (ALT), creatinekinase (CK), gamma-glutamyl-transpeptidase (GGT), brachial arterial pressure and heart rate were measured at V1, V3, V4 and V5. Blood analyses were centrally performed in the laboratory of the Endocrinology Unit using standard enzymatic techniques; LDL-C was directly measured. Clinical safety was monitored throughout the study. If required, the patient could be re-evaluated at any time during the study, aside of the visits scheduled.
# Statistical methods
The minimum level of statistical significance was set to p < 0.05 two-sided, therefore 95% confidence limits (95%CIs) were calculated. All reported p-values and CIs are two-sided. The primary efficacy variable was the LDL-C change between the start and the end of each treatment period, expressed as a percentage of the initial value. Therefore, mean and 95%CIs of changes within treatment periods (from V2 to V3 and from V4 to V5) for the experimental and the control treatment, irrespective of sequence, were calculated. The main analysis was the determination of the two-sided 95%CI of the between-treatment mean difference in the primary variable, computationally analogous to a paired t test. Setting 0.10 (i.e. 10% of the initial value) as the minimum clinically relevant difference, the two treatments were considered equivalent if the two-sided 95%CI of the difference in their LDL-C change from baseline was entirely between − 0.10 and + 0.10. Parallel calculations were carried out on absolute, rather than relative to baseline, LDL-C changes. Tot-C and HDL-C were analyzed as described above for LDL-C; for TG levels (which were approximately log-normally distributed) analogous calculations were performed on logarithmic transformations and changes were expressed as ratios. Between-treatment comparisons were expressed as A − N differences for cholesterol values and as A/N ratios for TG values. The effects on LDL-C were additionally tested in sensitivity multivariate analyses: variance for cross-over studies on final-baseline changes adjusting for period effects, and analysis of covariance on the difference between the final values adjusted for sequence and for the difference between the
[formula] Δ Final-Initial −40.1 (−49.2 −30.9) −36.4 (−45.6 −27.1) −3.7 (−13.0 +5.6) % Δ(Final-Initial) / Initial −25.6 (−31.1 -20.1) −23.3 (−28.7 −17.9) −2.3 (−7.4 +2.8) [/formula]
Data are expressed as mean ± SD, (95%CI), and analyzed by Student's t test for paired data (end vs. start of treatment), * p < 0.0001
## Sample size
The sample size was calculated for the main efficacy analysis described above, i.e. the determination of the two-sided 95%CI of the between-treatment mean difference in the LDL-C change from baseline. Assuming a standard deviation (SD) of the difference no greater than 0.12, based on a previous cross-over study for the difference between monacolin K and placebo [bib_ref] Red yeast rice improves lipid pattern, high-sensitivity C-reactive protein and vascular remodelling..., Cicero [/bib_ref] , it was estimated that 18 patients were required to prove the equivalence with a power of 0.80. This figure was rounded to 20 enrolled patients allowing for possible exclusions from the analysis.
# Results
## Patients characteristics and compliance
A patient was lost to follow-up after V2 (study entry) and was therefore excluded for analysis. No major violations of the study protocol occurred in the other 19 patients. Therefore the intention-to-treat, per-protocol and safety populations were constituted by the same patients. The demographic characteristics of the 19 patients included in the analysis are shown in [fig_ref] Table 1: Characteristics of the patients at entry [/fig_ref]. The baseline characteristics of the patients assigned to the two different sequences of treatment (A > N or N > A) were similar, only slight differences being observed regarding the parameters studied. Ten patients were affected by one or more metabolic or endocrine diseases: obesity (n = 9), hypothyroidism treated with L-thyroxine (n = 5), and type-2 diabetes mellitus treated with metformin (n = 2). Six patients were on antihypertensive therapy. The ratio between the number of doses presumably assumed (products delivered minus returned) and the treatment duration in days (equivalent to the number of doses to be taken as per the scheduled one-daily dose regimen), a measure of treatment compliance, was greater than 90% in all 19 patients who completed the study.
Efficacy LDL-C values and their changes during the study are reported in [fig_ref] Table 2: LDL-C [/fig_ref]. LDL-C decreased by 23.3% during treatment with N and by 25.6% during treatment with A. The A − N difference, i.e. the mean difference between the within-period changes observed with A and N, was − 2.3% (the minus sign meaning a greater reduction during treatment with A compared to N). The 95%CI of this difference was between − 7.4 and + 2.8%, entirely within the interval of − 10 to + 10% defined as clinical equivalence and used in sample size calculation. In parallel, the absolute (i.e., not referred to the initial value) LDL-C mean reduction was 36.4 mg/dL during N treatment and 40.1 mg/dL during A treatment, the mean A − N difference being − 3.7 mg/dL (95%CI, − 13.0 to + 5.6). LDL-C reductions within each treatment, either as absolute values or expressed as ratio of the initial value, were highly significant (p < 0.0001). The results of the analysis of variance adjusting for period effects on final −baseline LDL-C changes were very similar to those of the main analysis above. The mean A − N difference relative to baseline was − 2.3% (95%CI, − 7.4 to + 2.9; p = 0.36), with p = 0.70 for the period effect. The mean absolute A − N difference was − 3.8 mg/dL (95%CI, − 13.4 to + 5.7; p = 0.41), with p = 0.52 for the period effect. The analysis of Data are expressed as mean ± SD, (95%CI), and analyzed by Student's t test for paired data (end vs. start of treatment), * p < 0.0001 Data are expressed as mean ± SD, (95%CI), and analyzed by Student's t test for paired data (end vs. start of treatment), * p < 0.01 covariance on the difference between the final values adjusted for sequence and for the difference between the baseline values yielded a mean A − N absolute difference of − 3.6 mg/dL, very close to the other estimates reported above, but with a narrower 95%CI (− 9.7 to + 2.5; p = 0.23), with p = 0.49 for the sequence effect and p = 0.83 for the difference between the baseline values. No carry-over effect was observed, based on the fact that plasma LDL-C levels at the end of the wash-out were similar to pre-study values [fig_ref] Figure 2: LDL-C values [/fig_ref]. Individual values at start and end of each treatment showed a decrease of LDL-C in all patients, except for one who changed from 146 to 150 mg/dL during the first-period treatment with N . Irrespective of treatment period, decrease from abnormal (≥ 130 mg/dL) to normal LDL-C values was obtained in 17 of 19 patients during A treatment (89%) and in 14 of 17 during N treatment (82%; 2 patients were still in the normal range after the wash-out and remained so at the end of treatment). Tot-C, HDL-C and TG values measured before and after each treatment and their differences are reported in [fig_ref] Table 3: Total-C [/fig_ref] Like LDL-C, Tot-C decrease within each treatment period was always highly significant (p < 0.0001), while the between-treatment comparison showed a slightly, non-significantly greater reduction with A, both as absolute value (− 8.2 mg/dL; p = 0.16) and as ratio of baseline (− 3.8%; p = 0.086).
HDL-C increase was negligible during A and significant during N [fig_ref] Table 4: HDL-C [/fig_ref] , whereas TG diminished markedly during A and not significantly during N [fig_ref] Table 5: Triglycerides [/fig_ref] ; the difference between treatments, however, was not statistically significant for both variables (p > 0.30 and p = 0.12, respectively). The percent changes of the efficacy variables from start to end of each treatment are summarized in .
## Safety
No serious or severe adverse events occurred during the study; no event required treatment interruption or remedial therapy. Moderate and transient gastrointestinal disturbances occurred in three patients, two receiving N (constipation and flatulence) and one receiving A (diarrhea). Other moderate and transient adverse events referred by one patient each were eczema and headache during treatment with N and myalgia during treatment with A. Serum CK levels above the upper limit of normal (190 U/L) were observed in three patients: one after treatment with A (268 U/L) -previously not treated with N, one after both A and N (276 and 246 U/L, respectively), and one after the wash-out period between A and N (320 U/L). No clinically significant change was observed in the serum levels of fasting blood glucose, AST, ALT and GGT, as well as in blood pressure and heart rate measures.
# Discussion
High cholesterol levels are a major risk factor for cardiovascular disorders and may require specific intervention with the prescription of an appropriate drug treatment [bib_ref] ACC/AHA guideline on the treatment of blood cholesterol to reduce atherosclerotic cardiovascular..., Stone [/bib_ref] which has been shown to be able to reduce morbility and mortality in large cohorts of patients [bib_ref] Efficacy and safety of cholesterol-lowering treatment: prospective meta-analysis of data from 90,056..., Baigent [/bib_ref] and principally based on statins. Red yeast rice, obtained through rice fermentation by the fungus Monascuspurpureus, contains monacolin K a natural product chemically similar to lovastatin [bib_ref] Red yeast rice, Nguyen [/bib_ref]. Monacolin K, similarly to synthetic statins, inhibits the activity of HMG-CoA In this work we have investigated the lipid lowering capacities of a novel association containing monacolin K, L-arginine, coenzyme Q10 and ascorbic acid. L-arginine is an essential aminoacid with NO-mediated vasodilating properties [bib_ref] Nitric oxide synthesised from L-arginine mediates endothelium dependent dilatation in human veins..., Vallance [/bib_ref] [bib_ref] L-arginine improves endothelium-dependent vasodilatation in hypercholesterolemic humans, Creager [/bib_ref] [bib_ref] The L-arginine-nitric oxide pathway: role in the atherosclerosis and therapeutic implications, Böger [/bib_ref] [bib_ref] Increase in fasting vascular endothelial function after short-term oral L-arginine is effective..., Bai [/bib_ref] , coenzyme Q has antioxidant activity [bib_ref] Co-enzyme Q10 supplementation for the primary prevention of cardiovascular disease, Flowers [/bib_ref] while ascorbic acid has antioxidant and vasoprotective activity [bib_ref] Review: when is an antioxidant not a antioxidant? A review of novel..., Duarte [/bib_ref] [bib_ref] Long-term orange juice consumption is associated with low LDL-cholesterol and apolipoprotein B..., Aptekmann [/bib_ref]. The effects of this association were compared with those of a commercially available dietary supplement containing monacolin K and coenzyme Q10. The study was conducted according to a rigorous randomized cross-over design. Our results were in agreement with previous research on monacolin K based dietary supplements and showed a statistically significant reduction of LDL-C levels. We were able to document a marked lowering effect of LDL-C (A: -40.1 mg/dL, − 25,6%; N: -36,4 mg/dL, − 23,3%), in line with previous reports [bib_ref] Red yeast rice improves lipid pattern, high-sensitivity C-reactive protein and vascular remodelling..., Cicero [/bib_ref] [bib_ref] Nutraceutical approach to moderate cardiometabolic risk: results of a randomized, double-blind and..., Ruscica [/bib_ref] [bib_ref] Efficacy and tolerability of a nutraceutical combination (red yeast Rice, policosanols, and..., Gonnelli [/bib_ref] [bib_ref] Testing the short-term efficacy of a lipid-lowering nutraceutical in the setting of..., Cicero [/bib_ref] [bib_ref] Effect of Xuezhikang, an extract from red yeast chinese rice, on coronary..., Lu [/bib_ref] [bib_ref] Efficacy and safety of Monascus purpureus went rice in subjects with hyperlipidemia, Lin [/bib_ref] , with no difference in efficacy between the two formulations studied. In parallel with the reduction of LDL-C a decrease in Tot-C was documented for both formulations, and HDL-C was increased although reaching significant levels only with N. A significant reduction of triglycerides levels was found exclusively for the novel formulation tested (A). A clinically relevant effect of monacolin K-based supplements on triglycerides has been previously but not consistently reported [bib_ref] The effect of red yeast rice (Monascus Purpureus) in dyslipidemia and others..., Yang [/bib_ref] , since this reduction was not documented for N we speculate that other compounds present in A, probably L-arginine, may have contributed to this effect [bib_ref] L-arginine enhances the triglyceride-lowering effect of simvastatin in patients with elevated plasma..., Schulze [/bib_ref]. Regarding side effects (adverse events) in our study 3/19 patients reported mild and transient gastrointestinal discomfort, and in 3/19 subjects elevation of CK occurred, with no associated myalgia. These figures are similar to what expected and already reported in the literature [bib_ref] A nutraceutical approach (Armolipid plus) to reduce total and LDL cholesterol in..., Barrios [/bib_ref]. Among the limitations of this study we have to mention the open (not blinded) design of the study which was chosen because of the different nature of the pharmaceutical preparations of the compounds tested. The relatively low number of patients enrolled is an additional limitation, the impact of which was attenuated by the crossover design.
# Conclusions
Our results confirm the clinically meaningful lipid lowering properties of monacolin K and specifically the efficacy of a novel association of it with L-arginine, coenzyme Q10 and ascorbic acid. Furthermore, our results show a good safety profile of this association.
In addition to the very effective LDL-C lowering capacity it is worth noticing the significant reduction of triglycerides level achieved by A, which might be due to the presence of L-arginine in the formulation.
# Funding
The study was supported by an unrestricted grant to the Endocrinology Unit (Cisanello Hospital, Pisa, Italy) from Damor Pharmaceuticals. ISPharm srl and Studio Associato Airoldi Cicogna Ghirri were granted from Damor for planning and monitoring the study and to perform the statistical analysis, respectively.
## Availability of data and materials
The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.
## Dietary supplements composition
Argicolina: red yeast rice extract mg 200 (corresponding to monacoline K mg 10), citrinine and aflatoxin-free; L-arginine mg 3500; Coenzyme Q10 mg 50; ascorbate mg 100. Normolip 5: red yeast rice extract mg 200 (corresponding to monacoline K mg 10); gamma orizanol mg 90; Coenzyme Q10 mg 10; polycosanols mg 5; chrome picolinate mg 0,2.
Authors' contributions SM performed the study and was involved in writing the manuscript; GC was involved in writing the manuscript; CP had full access to all the data in the study and took responsibility for the integrity and accuracy; RJ, JV, PF and GS collaborated in recruitment of participants; GA performed the statistical analysis; MM designed the study and wrote the basic manuscript; GS designed the study and was involved in writing the manuscript; FS contributed to the study design, supervised the study and revised the manuscript. All authors read and approved the paper.
## Ethics approval and consent to participate
The study protocol, the patient information sheet, the informed consent form, the letter to general practitioner and the "privacy statement" sheet were examined and approved by the reference Ethic Committee of the investigational study site (Italian National Health Institute. Comitato Etico Area Vasta Nord Ovest [CEAVNO] Toscana, Pisa, Italy) prior to any studyrelated procedure was started. The study initiation was notified to the Italian Ministry of Health. The study was conducted according to the principles defined in the Declaration of Helsinki and amendments, and to the procedures of Good Clinical Practice (whenever applicable), expressed in the guideline set out by the International Conference on Harmonization. The decision on study participation was freely and in written taken by the patient, and it was clarified that the consent could have been withdrawn at any time, without penalty or loss of patient' rights of benefits.
[fig] Figure 1: Study flow-chart [/fig]
[fig] Figure 2: LDL-C values (mean and SD) through the study, split by treatment sequenceFig. 3 Individual values of LDL cholesterol (LDL-C, mg/dL) before and after Argicolina (A) and Normolip (N) treatment baseline values. Efficacy analyses had to be performed in the intention-to-treat population, i.e. all patients with at least one post-baseline control. A sensitivity analysis of the primary variable was also planned in the per-protocol population, i.e. all patients without major protocol violations. Safety results had to be reported in all patients who had assumed at least one dose of one study drug. Statistical analyses were performed by the Studio Associato Airoldi Cicogna and Ghirri, Milan, using the SAS Software version 9.4 (SAS Inc., Cary, NC). [/fig]
[table] Table 1: Characteristics of the patients at entry [/table]
[table] Table 2: LDL-C (mg/dL) in the 19 patients who completed the study [/table]
[table] Table 3: Total-C (mg/dL) in the 19 patients who completed the study [/table]
[table] Table 4: HDL-C (mg/dL) in the 19 patients who completed the study [/table]
[table] Table 5: Triglycerides (mg/dL) in the 19 patients who completed the study [/table]
|
Rad51 paralogs Rad55-Rad57 balance the anti-recombinase Srs2 in Rad51 filament formation
13 0.21 µM RPA, ± 0.4 µM Srs2), and the means (ø) ± 1 sd and distributions of filament length classes are shown. Scale bars: 100 nm. White arrows indicate RPA-ssDNA complexes and red arrows Rad51 filaments.Figure 4. Rad55-Rad57 interact with Srs2 and inhibit Srs2 helicase. a, Pull-down with 4 nM (1 pmol) Rad55-Rad57 and 2.7, 8, or 16 nM Srs2. b, Pull-down with 4 nM Rad55-Rad57 and 8 or 16 nM Rad51. c, Quantitation of results in (a) and (b) and additional experiments. d, Pulldown with 4 nM Rad55-Rad57 and 8 nM Srs2 ± 40 nM Rad51. GST was used as control. S: supernatant, W: wash, E: eluate. e, Helicase assay. f, 28 nM Rad51 ± 25 nM Rad55-Rad57 were incubated with 1.5 nM 3'-tailed substrate before addition of 120 nM Srs2 protein. Product yields at 20 min. were quantified as shown in g. HEAT DEN.: heat denatured substrate, shown are means ± 1sd, n=3.MethodsProtein purification: Yeast Rad51, RPA, and Srs2 proteins were purified as described 1,2 . The purification of Rad55-Rad57 was adapted from a previously published protocol 3 . Yeast cells overexpressing GST-Rad55-His6-Rad57 were grown and harvested as described 3 . Cells were disrupted in Buffer B containing 10 mM Na 2 HPO 4 , 1.8 mM KH 2 PO 4 , 2.7 mM KCl, 1M NaCl, 10 % (v/v) Glycerol, 10 mM -mercaptoethanol, and protease inhibitor cocktail (1 mM PMSF, 2 M leupeptin, 1 M pepstatin A, and 1 mM benzamidin) using glass beads (0.5 mm glass beads;BioSpec Products, Inc.). The cell lysate was centrifuged at 40,000 rpm for 45 min using a Ti50.2 rotor. The supernatant was collected and loaded onto a pre-equilibrated Glutathione Sepharose
## 14
4B column (GE Healthcare, NJ). After washing with buffer A (20 mM Tris HCl pH 7.5, 1 mM EDTA, 1 M NaCl, 5 mM -mercaptoethanol, and 10 % glycerol), the GST-tagged proteins were eluted with Buffer A containing 20 mM reduced glutathione plus protease inhibitor cocktail.
Fractions containing the GST-Rad55-His6-Rad57 heterodimer, as determined by 10 % SDS-PAGE, were pooled and dialyzed against Buffer C (50 mM NaH 2 PO 4 pH 8.0, 1 M NaCl, and 10% glycerol) containing the protease inhibitor cocktail. Then the pool was loaded onto a preequilibrated Ni-NTA agarose column and washed with Buffer C plus protease inhibitor cocktail.
The bound complexes were eluted with Buffer C containing 0.5 M NaCl, 0.1 mM PMSF, and 250 mM imidazole, and analyzed by 10 % SDS-PAGE. Fractions containing stoichiometric Rad55-Rad57 heterodimer were pooled, concentrated, dialyzed into the storage buffer containing 20 mM Tris HCl pH 7.5, 0.5 M NaCl, 0.1 mM EDTA, 1 mM DTT, and 10 % glycerol, and then stored in aliquots at -80°C. The absence of contaminating enzymatic activities and DNA in protein preparations was verified as described 4 .
Purification of 600 nt ssDNA: As published previously 5 , 600 bp dsDNA fragments biotinylated at a one 5' end were generated by PCR from PstI-linearized X174 DNA using primers WDHY427 5´-TTATCGAAGCGCGCATAAAT-3´and 5´biotinylated WDHY431 5´-GTCTTCATTTCCATGCGGTG-3´. The biotinylated dsDNA was loaded onto a HiTrap Streptavidin HP column (Amersham Biosciences), and non-biotinylated single stranded DNA was eluted with 60 mM NaOH.
Rad51-ssDNA filament assembly assay: In , Rad51 (0.67 M) was incubated with 4 M (nt) ssDNA, in the presence or absence of 0.11 M Rad55-Rad57, in buffer R containing 20 mM triethanolamine pH 7.5, 4 mM Mg(OAc) 2 , 2.5 mM ATP, 25 g/mL BSA, 1 mM DTT, 90 mM NaCl, and 5 % glycerol for 10 min. Then 0.25 % glutaraldehyde was used to crosslink the protein-DNA complexes for 15 min. The complexes were separated on a 0.5 % agarose gel, stained with SYBR Gold, transferred to nitrocellulose membrane, and blotted with rabbit polyclonal anti-Rad51 or anti-Rad55 antibodies. Complexes were separated on a 0.5 % agarose gel and stained with SYBR Gold. Proteins were transferred to nitrocellulose membrane and blotted with anti-Rad51 antibodies. All DNA concentrations refer to nucleotides (ssDNA) or base pairs (dsDNA).
Protein binding to ssDNA immobilized on magnetic beads: In , a 5´-biotinylated oligonucleotide was immobilized onto magnetic streptavidin beads as previously described 6 .
The oligo sequence is 5´-CCCCCCCCCCCCCAAGATAATTTTTCGACTCATCAGAAATATCCGA AAGTGTTAACTTCTGC GTCATGGAAGCGATAAAACTC-3´. In experiments containing Srs2, 10 µL slurry of beads containing 3 M (nt concentration) ssDNA were incubated with 1 M Rad51 in the presence and absence of 0.1 M Rad55-Rad57 in buffer containing 20 mM trienthanolamine, 5 mM Mg(OAc) 2 , 4 mM ATP, 25 g/mL BSA, 1 mM DTT, 5 % glycerol, and 25 mM NaCl for 10 min at 22°C. Then 0.1 or 0.33 M Srs2 protein was added and further incubated for 10 min. The beads were washed, and bound proteins were eluted and quantified as described 7 . Background protein binding was typically less than 3 %. Protein interaction assays: GST-Rad55-His6-Rad57 (4 nM) or 30 nM GST (GE Healthcare) were incubated with increasing amounts of either Srs2 or Rad51 in buffer P containing 25 mM Tris-HCl (pH 7.5), 10 mM Mg(OAc) 2 , 50 mM NaCl, 1 mM DTT, 10 % glycerol, and 0.05 % NP-40 for an hour at room temperature . Equilibrated and BSA-treated Glutathione-Sepharose 4B beads were added to the mixture and incubated for another hour. The beads and supernatant were separated through centrifugation and the beads were washed twice with binding buffer P. The pulled-down protein complexes were eluted by boiling at 95°C for 3 min in 10 L SDS-PAGE loading buffer, separated through a 10 % SDS-PAGE gel, and the protein bands were visualized through immunoblots and quantitfied by ImageQuant. In -b, 1/16th of the supernatant and wash were loaded. In , 1/7th of the supernatant and wash were loaded.
## Topology-based assay for rad51 dissociation:
For the competition protein binding assay , the two proteins were incubated for 30 min before the addition of an increasing amount of the third challenging protein, as specified in the diagrams. After another 30-min incubation, equilibrated and BSA-treated Glutathione-Sepharose 4B beads were added to the mixture and incubated for another hour.
Analysis and quantitation was performed as described above. The anti-Rad51, Rad55, Rad57 antibodies were generated in rabbits, the anti-Srs2 antibody was purchased from Santa Cruz Biotechnology, Inc.
Helicase assay: The assay followed a published protocol and the substrates were prepared exactly as described before 9 . In , 28 nM Rad51 with 0 or 25 nM Rad55-Rad57 were incubated with 1.5 nM oligo substrate with 3´ tail in buffer containing 20 mM triethanolamine pH 7.5, 4 mM Mg(OAc) 2 , 25 g/mL BSA, an ATP-regenerating system consisting of 2.5 mM ATP, 20 units/ml creatine kinase, and 20 mM creatine phosphate, as well as either 1 mM DTT 18 and 40 mM NaCl or 5 mM DTT and 10 mM NaCl for 10 min at 30 C. Then 120 nM Srs2 protein was added to initiate the helicase reaction. After 20 min incubation, the reactions were stopped by adding 4.5 L stop buffer containing 150 mM EDTA, 2 % SDS, 163 nM unlabeled oligo, and 4.3 mg/mL protease K into 9 L reaction sample. The DNA species were separated through electrophoresis on a 10 % TBE-PAGE gel, which was dried and analyzed by a Storm phosphoimager. The bands were quantified by densitometry using ImageQuant. , was adapted from a published protocol 10 . In brief, Rad51-Rad55-Rad57-ssDNA complexes were assembled as described above and crosslinked with 0.25 % glutaraldehyde for 20 min, before deposition on grids. Grids were blocked in 50 g/mL BSA in TBST for 30 min, and then incubated with goat anti-GST antibody (GE Healthcare) for 30 min. After three 5 min washes with 50 g/mL BSA in TBST, the grids were incubated in TBST plus a 1:5 dilution of gold particles dressed with rabbit anti-goat antibody (BioAssay Works). After two five min washes in 50 g/mL BSA in TBST and one five min wash in 5 mM Mg(OAc) 2 , grids were stained with 2 % (w/v) uranyl acetate before imaging. 20 Recombination assay: Spontaneous recombination rates between direct repeats were determined following a published fluctuation analysis protocol using the method of the median . The direct-repeat recombination substrate has two different ade2 alleles separated by plasmid sequences and the URA3 gene 13 . Yeast strains were grown on YPD plates for 2 days at 30°C for single colonies. For each strain, nine independent single colonies were randomly chosen and the entire colony was used to inoculate 4 mL YPD liquid culture. Liquid cultures were grown for 2-3 days at 30°C to reach stationary phase. Cells were collected, washed with sterile H 2 O, and suspended into 1 mL sterile H 2 O. 100 L of appropriate dilutions of each culture were spread on two plates each of SD-ADE-URA. Cells were incubated for 2 days at 30°C. For each culture, the number of colonies on YPD were counted and totaled to determine the total cell number. The number of colonies on SD-ADE-URA were counted to determine the median number of recombinants. For each strain, recombination rates were measured independently three times and the mean values with standard deviations are shown.
## Saccharomyces cerevisiae strains are listed in
## Mms sensitivity assay:
Yeast strains were grown overnight in liquid YPD to mid-log phase at 30°C, and then diluted to OD 600 = 1. Serial dilutions of these cell cultures were made with sterile H 2 O and spotted onto YPD plates with or without methyl methanesulfonate (MMS). Plates were incubated for three days at 30°C or five days at 22°C before photographing using a FluorChem8900 imaging system (Alpha Innotech). |
Measurement of collateral perfusion in acute stroke: a vessel-encoded arterial spin labeling study
Collateral perfusion is important for sustaining tissue viability in acute ischemic stroke. Conventional techniques for its visualization are invasive, require contrast agents and demonstrate collateral vessels, rather than measuring perfusion directly. In this study we utilize a non-invasive, non-contrast magnetic resonance imaging (MRI)-based method to directly quantify collateral perfusion in acute stroke patients. Vessel-encoded multi-postlabeling delay arterial spin labeling (AsL) was used to separately quantify the blood flow and blood arrival time from four arteries supplying the brain in patients presenting within 18 hours of stroke onset. Twenty-nine acute ischemic stroke patients were scanned with a median time of onset to first MRI of 3 hours. Collateral perfusion at presentation was associated with tissue fate at 1-week. It sustained tissue prior to reperfusion, but was less effective than direct blood flow at maintaining tissue viability in patients who did not reperfuse. Delay in the blood arrival around the ischemic region was found at presentation and reduced over time but was not consistently associated with collateral perfusion. Vessel-encoded multi-postlabeling delay AsL provides a noninvasive tool for direct measurement of collateral perfusion and delayed blood arrival in acute stroke patients.Collateral tissue perfusion is an important determinant of tissue outcome in acute stroke 1 , sustaining tissue viability prior to reperfusion, and maintaining blood flow in the longer term 2 . Patients with extensive collateral vessels have better clinical outcomes 2-4 , and collateral vessel status may be used to select those patients who are likely to benefit from recanalization therapies 5-11 . However, measuring collateral perfusion directly to the tissues is challenging and current approaches infer collateral perfusion from the presence of collateral blood vessels or delayed blood arrival 2 .Arterial spin labeling (ASL) is a non-invasive magnetic resonance imaging (MRI) technique that does not require an exogenous contrast agent. ASL MRI labels the arterial blood in the feeding arteries in the neck using radiofrequency magnetic pulses, and can serially measure absolute cerebral blood flow (CBF) in patients with acute stroke 12-15 . Vessel-encoded pseudocontinuous ASL (VEPCASL) 16 is capable of mapping perfusion within territories of individual arteries, providing flow information that agrees well with digital subtraction angiography 17 , and does not compromise signal-to-noise ratio or CBF quantification 18 .The time taken for the labeled blood to reach the tissue, the arterial transit time (ATT), can be calculated from ASL images when data at multiple postlabeling delays (multi-PLD) have been acquired. Delays in blood arrival have been proposed to identify collateral perfusion 2,19-21 , but delayed ATT can also be observed in other settings including vascular disease, ischemia, and in those with microvascular changes[22][23][24][25].In this study, we present the use of multi-PLD VEPCASL acquired serially in a cohort of patients with acute ischemic stroke. We demonstrate that VEPCASL can concurrently identify collateral perfusion patterns and delayed blood arrival serially in these patients, and assess whether collateral perfusion measured at presentation is associated with tissue fate at follow-up.
www.nature.com/scientificreports www.nature.com/scientificreports/ Methods patients. Patients presenting with acute ischemic stroke within 18 hours of symptom onset were recruited and consented under research protocols agreed by the UK National Research Ethics Service Committee South Central -Oxford C (refs: 12/SC/0292 and 13/SC/0362). Inclusion criteria for this analysis were: presenting scan within 18 hours of symptom onset; Diffusion-weighted imaging (DWI) lesion within the middle cerebral artery (MCA) territory; patient or representative able to give a clear medical history and participate in the consent process; age over 18. Patients with a contraindication to MRI, lacunar stroke defined on DWI, or severely impaired conscious level (score greater than 1 on question 1a of the National Institute for Health Stroke Scale) were not enrolled. Serial imaging was performed at presentation, two hours, 24-hours, 1-week, and 1-month, whenever possible. Where thrombolysis was indicated patients underwent the initial MRI scan during the infusion of alteplase if required. No endovascular treatment options were available at the time of the study. All experiments were performed in accordance with the relevant guidelines and regulations and informed consent or agreement from a consultee was obtained from all individual participants included in the study.
Imaging. All scans were performed on a 3T Verio (Siemens Healthcare, Erlangen, Germany) using a 32-channel head coil. Preliminary scans were as follows: 1) a rapid 3D time-of-flight (TOF) angiogram of the neck (voxel size 0.8 × 0.8 × 1.3 mm, acquisition time 47 s) to position the VEPCASL labeling plane and identify the location of the feeding arteries within this plane: the right and left internal carotid arteries (ICAs), and the right and left vertebral arteries (VAs), as described previously; 2) Diffusion-weighted images (DWI, voxel size 1.8 × 1.8 × 2.0 mm, b = 0 and 1000 s/mm 2 , acquisition time 3 min) to define the ischemic core; and 3) a T1-weighted structural image (voxel size 1.8 × 1.8 × 1.0 mm, acquisition time 4 min) to aid registration.
These were followed by ASL perfusion imaging, using a previously described protocolwhich builds on the minimum standards outlined in a recent consensus paper. The protocol included a 1.4 s duration VEPCASL pulse train which cycled through eight different vessel-encodings: two non-selective (label and control), two left-right, two anterior-posterior and two diagonal. Two repetitions of these encodings were acquired for each of six nominal PLDs (0.25, 0.5, 0.75, 1.0, 1.25 and 1.5 s) with a repetition time of 4.1 s, giving a total of 96 volumes in 6.5 minutes. Images were acquired with a 2D multi-slice echo-planar imaging readout (voxel size 3.4 × 3.4 × 5 mm, matrix size 64 × 64, 6/8 ths partial Fourier, echo time 14 ms, 24 slices acquired in ascending order). The time to acquire all slices was 1085 ms, meaning the average effective PLDs across the brain were 0.79, 1.04, 1.29, 1.54, 1.79 and 2.04 s. Calibration scans were acquired with both head and body coils for signal reception to allow for correction of coil non-uniformity and quantification of absolute CBF.
Finally, a T2-weighted turbo spin echo fluid attenuated inversion recovery (FLAIR) acquisition (voxel size 1.9 × 1.9 × 2.0 mm, echo time 96 ms, acquisition time 2 min) was performed at the 1-week timepoint to define final infarct Image processing. Images were processed using the FMRIB software libraryand Matlab (MathWorks, Natick, MA, USA), as described previously. Pre-processing included motion correction of the VEPCASL raw data 28 , brain extraction 29 and segmentation of the T1-weighted image, and correction of the VEPCASL data for receive coil non-uniformity. Within-timepoint registration across imaging modalities was achieved using linear registration, but across timepoints non-linear registration 31 was used to account for tissue distortion.
Separation of the signals arising from each brain-feeding artery in the vessel-encoded data was achieved using a Bayesian maximum a posteriori solution 32 to the general framework for vessel-encoded analysis, which can account for some patient movement between the TOF and VEPCASL acquisitions. Image calibration was achieved by non-linearly registering a ventricle mask from standard space via the patient's T1-weighted image on to the VEPCASL calibration image to allow estimation of the equilibrium magnetization of cerebrospinal fluid, which is then converted into the equilibrium magnetization of blood. The general ASL kinetic model 20 was fitted to each arterial component separately using a variational Bayes algorithm 34 to yield CBF and ATT estimates from each feeding artery within each voxel. To simplify further analysis, weighted ATT maps were calculated by multiplying the CBF and ATT maps, summing across all feeding arteries and then dividing by the total CBF in each voxel. In voxels supplied by a single artery, the weighted ATT is therefore equal to the ATT of this dominant arterial component, but in voxels supplied by multiple arteries it represents the weighted average ATT across these arteries.
## Definitions and regions of interest.
For the patients in this study with strokes in the MCA territories, Direct CBF was defined as the blood flow to a voxel from the ipsilateral ICA, and Indirect CBF was the sum of the CBF from all arteries other than the ipsilateral ICA. Therefore, for tissue within the MCA territories, which is normally supplied by the ipsilateral ICA, a change in collateral perfusion originating from the contralateral ICA or the VAs should therefore be reflected by a change in the Indirect CBF.
Each patient was scored on the Modified Thrombolysis in Cerebral Infarction (mTICI) scale using the 24-hour CBF maps when available 7 , with patients categorized as reperfusers (mTICI = 2b or 3) or non-reperfusers (mTICI = 0, 1 or 2a).
The ischemic core at presentation was defined using semi-automated delineation of the apparent diffusion coefficient (ADC) mapbelow an externally validated thresholdof 620 × 10 −6 mm 2 /s. Final infarction was defined preferentially on the 1-week FLAIR, or on the 24-hour trace DWI if the 1-week timepoint was not available. These masks enabled two specific ROIs to be generated which were not derived from a perfusion-based definition of tissue at risk:
1. Surviving tissue: The co-registered final infarct mask was dilated using an empirically defined 10 mm radius spherical kernel before subtracting the original mask. This ROI was generated to investigate Indirect CBF and ATT in tissue that was close to the infarct but survived.
www.nature.com/scientificreports www.nature.com/scientificreports/ 2. Peri-core: The ischemic core mask was dilated using a spherical kernel of radius 20 mm before subtracting the original mask. This ROI was generated to assess the relationship between Direct and Indirect CBF at presentation and risk of subsequent infarction. A larger radius than the surviving tissue ROI was chosen to increase the number of voxels selected, particularly surviving voxels.
Equivalent contralateral ROIs were generated in the same fashion from mirrored ischemic core and final infarct masks. All ROIs were restricted to gray matter and manually checked to ensure no voxels from the opposite hemisphere were included. The gray matter mask was derived from the presenting segmented T1-weighted structural image, registered into the space of the ASL data and thresholded at a partial volume of greater than 0.5 Analysis. All data sets were included in the analysis unless affected by significant motion artefacts (assessed in a blinded manner by a clinician, GH), no follow-up data were available to define the final infarct or the ROIs did not contain any voxels after transformation to ASL space and gray matter masking.
The ability of VEPCASL to identify the presence and prevalence of collateral perfusion in tissue that survived was quantified by measuring Indirect CBF in the Surviving Tissue ROI as a proportion of the total CBF. Both the proportion of Indirect CBF and the ATT were compared to the contralateral ROI at presentation and 1-month for those patients where both scans were available. Patient level data were also presented using imaging from all patients at all available timepoints, to reduce any bias arising from exclusion of those lost to follow up. Two-way analysis of variance (ANOVA) of both measures were used to compare Surviving Tissue to the contralateral ROI across the timepoints. If these were significant, post-hoc t-tests were also performed to compare the results within each timepoint.
The effect of collateral perfusion, as measured by VEPCASL, on tissue viability was assessed by measuring the relationship between the Indirect CBF for each voxel in the Peri-Core ROI and whether or not it survived. Voxels that received more than 25 ml/100 g/ of Direct CBF were excluded, leaving only those with a significant risk of infarction if collateral perfusion were not present. Tissue survival was defined as those voxels that fell outside the final infarct mask. The proportion of surviving voxels was calculated across 10 ml/100 g/min Indirect CBF ranges. Statistical significance was assessed by comparing the proportion of surviving tissue in voxels with Indirect CBF above and below 25 ml/100 g/min using a binomial proportion test. In order to investigate the effect of reperfusion on this relationship, the analysis was repeated after splitting the patients into reperfusion and non-reperfusion subgroups.
To assess whether there was a difference between the effect of Direct and Indirect CBF on tissue fate, and whether this depended on reperfusion status, the proportion of all voxels within the Peri-Core mask that survived was separately quantified across a range of Direct and Indirect CBF values.
# Results
29 patients were included in this study. Patient demographics are listed in . The number of multi-PLD VEPCASL scans completed at presentation, two hours, 24-hours, 1-week and 1-month was 24, 17, 20, 17 and 18, respectively. Of these 5, 5, 7, 6 and 0 data sets were excluded at each of the respective timepoints due to the presence of significant motion artefacts. A further 4, 6, 2, 2 and 3 data sets were excluded due to a lack of follow-up infarct mask availability or no voxels within the ROIs, leaving 15, 6, 11, 9 and 15 data sets at each respective time point for analysis. Representative multi-PLD VEPCASL data, clearly showing the phenomena of collateral perfusion and delayed blood arrival at the level of the individual, along with DWI and FLAIR images, can be seen in surviving tissue. For patients with ASL data available at both presentation and 1-month, there was a significantly greater proportion of Indirect CBF in the Surviving Tissue ROI than the contralateral ROI (ANOVA p = 0.016,, which was most marked at presentation (0.37 vs 0.19), indicating the presence of collateral perfusion. There was no significant effect of timepoint on the proportion of Indirect CBF in Surviving or www.nature.com/scientificreports www.nature.com/scientificreports/ . Example data from a patient with no apparent collateral perfusion. CBF and ATT maps from the presenting and 24-hour scans, with the one week T2-weighted FLAIR and presenting DWI, registered to the presenting T1-weighted structural image. The CBF maps are color-coded according to the arterial origin of the blood signal, as shown in the legend. The CBF deficit at presentation in the right MCA territory partially reperfuses by 24-hours, but nevertheless results in infarction (white arrowheads). No collateral CBF to the ischemic area is apparent. In areas around the ischemic region the ATT appears delayed compared to the contralateral hemisphere (blue arrowheads). Times are from symptom onset. ATT values are displayed for voxels with CBF of greater than 25 ml/100 g/min. . Example data from a patient with inter-hemispheric collateral flow. Images are arranged and labeled as per . The presenting CBF deficit in the right MCA territory reperfuses by 27 hours, but tissue in this region has already infarcted (white arrowheads). Other regions receive collateral perfusion originating from the LICA, probably through the anterior cerebral arteries (orange arrowheads). Some collateral supply appears to flow through pial collaterals (yellow arrowhead). Following reperfusion, the vascular territories revert to a standard configuration, and regions that received collateral perfusion survive. Extended ATT in regions receiving collateral perfusion are shown, but also in regions distal to the ischemic region, which persist even after reperfusion (blue arrowheads). RICA occlusion has resulted in the LICA providing blood to both the right anterior cerebral artery and right MCA territories. The small perfusion deficit at presentation leads to infarction despite reperfusion at 27 hours (white arrows). Collateral perfusion from the posterior circulation, which regresses after reperfusion, spares tissue from infarction (orange arrows). Despite the complete collateral supply to the RICA territory there is no apparent delay in ATT. . Patient-level collateral perfusion analysis. The fraction of Indirect CBF (blood that does not arise from the ipsilateral ICA), averaged across patients, is much higher within the Surviving Tissue mask than within the contralateral mask, especially at presentation: (a) patients with imaging both at presentation and 1-month; (b) any available imaging from all patients at each timepoint. The number of patients contributing to the data at each timepoint is quoted below each bar. Error bars represent the standard error. In both cases, the ipsilateralcontralateral difference is significant (ANOVA, p < 0.05). Post-hoc t-test significance (p < 0.05) is marked with an asterisk (*). (2019) 9:8181 | https://doi.org/10.1038/s41598-019-44417-7 www.nature.com/scientificreports www.nature.com/scientificreports/ Delayed blood arrival within the Surviving Tissue ROI was observed relative to the contralateral mask (ANOVA p = 0.047,, but the delay was only marked at presentation, with the mean ATT values being 1.35 s in Surviving Tissue versus 1.22 s on the contralateral side (t-test p = 0.005, n = 12). There was no significant effect of timepoint on the ATT values (p = 0.6). Including data from all patients at all times yields results showing a similar trend for reduction in the mean ATT difference between Surviving and contralateral ROIs over time, although only the effect of ROI was significant (ANOVA p = 0.047, t-test at presentation p = 0.001,.
At the level of the individualit was observed that collateral perfusion and delayed blood arrival do not always coincide, with persistent ATT increases still visible after reperfusion in some cases. In , no apparent collateral perfusion is present but the asymmetrical ATT demonstrates that in addition to incomplete reperfusion there are persistent arrival time delays, even in the absence of Indirect CBF. In , inter-hemispherical collateral perfusion is apparent at presentation, but despite the seemingly complete reperfusion and resolution of the collateral flow, ATT delays also persist at 24-hours in the region of the presenting perfusion deficit. In contrast, delayed blood arrival is not evident ineither before or after reperfusion despite blood flow to the right ICA territory originating entirely from the contralateral ICA. peri-core. Within Peri-core voxels with less than 25 ml/100 g/min Direct CBF at presentation, which were at risk of infarction in the absence of collateral flow, there was a clear increase in tissue survival fraction in voxels with higher levels of Indirect CBF . This indicates that collateral perfusion, as measured by VEPCASL, has a meaningful impact on tissue fate. The tissue survival fraction in voxels receiving less than 25 ml/100 g/min of Indirect CBF was significantly lower than in those above this threshold (60.3% versus 75.2%, p < 0.0001). The relative increase in tissue survival fraction between low (0-10 ml/100 g/min) and high (70-80 ml/100 g/min) levels of Indirect CBF was 45%.
A similar relationship between Indirect CBF and tissue survival was seen in both the reperfusion and non-reperfusion subgroups separately. This association was significant in both cases (p < 0.0001), although as expected, the fraction of tissue surviving was always greater in voxels of patients who reperfused (p < 0.05).
Across all voxels within the Peri-Core ROI, regardless of Direct CBF value, increases in both Direct and Indirect CBF lead to similar improvements in tissue survival for patients that reperfused. In patients who did not reperfuse, greater Direct CBF led to significantly larger tissue survival fractions than equivalent levels of Indirect CBF.
At the level of the individual, the ability of Indirect CBF to sustain tissue viability prior to reperfusion is highlighted in Figs 1-3. No collateral perfusion is seen in , and the hypoperfused tissue has infarcted at follow up. In contrast, in Figs 2 and 3 the CBF at presentation in some regions is maintained from collateral sources. The Indirect CBF supports tissue survival pending reperfusion by 24-hours, where it occurs. Indirect CBF was observed both from contralateral and posterior circulation sources: in , CBF in the right hemisphere was sustained by blood from the left ICA, before reperfusion from the right ICA at 24-hours; in, where both MCA territories are supplied by the left ICA due to a right ICA occlusion, at presentation blood from the vertebral arteries sustains the tissue prior to reperfusion from the left ICA at 24-hours. www.nature.com/scientificreports www.nature.com/scientificreports/
# Discussion
In this study we have shown that multi-PLD VEPCASL is capable of directly visualizing collateral perfusion and delayed blood arrival serially in acute stroke patients. Unlike conventional imaging approaches, this technique is non-invasive and capable of directly visualizing collateral perfusion at a tissue level, rather than just identifying collateral vessels. A VEPCASL-derived measure this phenomenon, Indirect CBF, measured at presentation was shown to be associated with tissue fate, both within individual patients and across the cohort. In particular, our results suggest that Indirect CBF at presentation can compensate for a lack of Direct CBF early after acute stroke and prior to reperfusion. However, Indirect CBF was not as effective as Direct CBF in sustaining tissue viability in patients who did not reperfuse. Finally, the multi-PLD protocol used allows delayed blood arrival to be measured, yielding information that is complementary to Indirect CBF in acute stroke.
Indirect CBF accounted for less than 20% of total CBF in the contralateral hemisphere ROI in keeping with chronic cerebrovascular disease. In the Surviving Tissue ROI, Indirect CBF contributed around 40% of the CBF for the first 24-hours after stroke onset. The greater proportion of Indirect CBF in the affected hemisphere was most marked in the short-term, but diminished over time. In keeping with all other perfusion techniques it is impossible to know the arrangement of collateral perfusion before the acute event. However, the temporal . Voxel-level tissue survival as a function of collateral perfusion. In Peri-Core voxels greater Indirect CBF at presentation significantly increases the probability of tissue survival at one week. Each bar represents the fraction of voxels, across all patients, that survive within each given range of Indirect CBF. The asterisk (*) represents the significant difference in survival fraction for tissue receiving less than 25 ml/100 g/min of Indirect CBF compared to that receiving more than this value (p < 0.0001). . Voxel-level tissue survival analysis within the Peri-Core ROI, as shown in , plotted separately for patients that reperfuse and those that do not. In both groups, increased Indirect CBF, and thus collateral perfusion, at presentation improves the likelihood of tissue survival (p < 0.0001). However, the probability of tissue survival is significantly increased in patients that reperfuse at all indirect CBF ranges, according to binomial proportion tests (*p < 0.05).
www.nature.com/scientificreports www.nature.com/scientificreports/ changes of the proportion of Indirect CBF within the two hemisphere ROIs is consistent with acute changes in collateral perfusion that have occurred as a result of the stroke, and that regress over time.
As would be expected reperfusion was associated with a greater proportion of tissue survival in the Peri-core ROIs. Indirect CBF increased the chances of tissue survival in the context of reperfusion to a similar extent as Direct CBF. However, in the absence of reperfusion fewer Peri-core voxels survive, and are more likely to do so if they have a greater proportion of Direct rather than Indirect CBF supply. This supports the concept that collateral perfusion is a short-term bridging phenomenon that can sustain tissue viability pending reperfusion of the original feeding vessel in acute stroke. At the level of the individual patients, regression of collateral perfusion was observed once reperfusion occurred. Although evidence of collateral perfusion that persisted over several days was noted in one patient, this pattern appeared to be a longstanding, chronic collateralization, rather than the acute sustaining collateral perfusion observed at a group level. The short-term ability of acute collateral perfusion to maintain tissue viability is consistent with data from pre-clinical work, and from clinical trials demonstrating that patients who have more collateral blood vessels have better responses to treatment, even outside conventional time windows.
While the pattern of delayed ATT at presentation was consistent with findings from previous single time-point studies, the serial and individual patient data point to a more complicated relationship between ATT and collateral perfusion. Established collateral flow, such as that found in response to an ICA occlusion, can be associated with normal ATT, as shown in. Conversely delayed ATT was present even when collateral flow had regressed . Although delayed ATT has been shown to associate with collateral blood flow in specific settings, due to the circuitous route blood may take to reach the tissue, it may also result from incomplete recanalization, or increased vascular resistance to flow due to capillary occlusion, edema or endothelial dysfunction, and caution is required when interpreting delayed blood arrival as a marker of collateral flow. The use of techniques such as VEPCASL to evaluate changes in both collateral perfusion and blood arrival time simultaneously helps to avoid this ambiguity, and could aid the prediction of short and longer term outcomes, particularly in relation to recanalization therapies.
This study is subject to several limitations. Like all ASL studies the results are limited by voxels that have a very prolonged delay to arrival, by which time the signal has decayed considerably, making CBF quantification very challenging. The inherently low signal-to-noise associated with ASL also means that voxel sizes are a compromise between spatial resolution, signal and acquisition times. This in turn can lead to partial volume contamination of each voxel, although the effect on the results will be reduced when using mean values from large ROIs or when comparing with matched contralateral ROIs.
The use of VEPCASL has allowed the observation of collateral flow between the main brain-feeding arteries in this study, but compensatory flow within the vascular territory of a single artery (e.g. from the right anterior cerebral artery to the right MCA) could not be observed. Labeling a larger number of arterial branches more distally has been demonstrated, but this may not be practical in acute patients because of planning time and restricted brain coverage.
The relatively small sample size meant that some of the more subtle trends in the data did not reach significance at the patient level, and we were unable to perform subgroup analyses to assess the impact of comorbidities and current patient medication. The number of data sets available at intermediate time points was also limited by a number of factors, including lack of research scanner availability, changes in the patient's clinical status, and some patients being transferred to rehabilitation facilities or other clinical units where follow-up scans were not possible. Scans at 28 days were more common since the condition of many patients had improved. However, the loss of data due to motion-related artefact and patient dropout may introduce a bias at a group level, with the Analysis was performed in all Peri-Core voxels, for patients who reperfused (a) and patients who did not reperfuse (b) separately. *p < 0.05 using a binomial proportion test.
www.nature.com/scientificreports www.nature.com/scientificreports/ exclusion of more severe stroke syndromes. A larger follow-up study would allow collateral perfusion and delayed blood arrival to be studied in greater detail in the future.
The strengths of this study include not using gadolinium-based contrast, meaning serial data could be acquired to track the dynamics of absolute measures of collateral flow, and making it possible to measure both hypo-and hyperperfusion. Vessel-encoding allows absolute measurement of collateral perfusion across large territories in acute stroke. Multiple post labelling delays allows the independent identification of delayed ATT which may give indications of microcirculatory resistance as well as collateral flow, but this would need validation in larger cohorts. Before such an undertaking could occur further work would be required to improve the reliability of this technique, reduce its sensitivity to motion artefact, and validate it against conventional angiography in acute stroke.
# Conclusions
Multi-PLD VEPCASL offers an opportunity to quantify collateral perfusion and delays to blood arrival serially in acute stroke patients and their relationship to tissue survival. Indirect CBF is an important transient determinant of tissue fate following acute stroke, particularly for patients who reperfuse. Delayed arrival time appears to represent more than collateral perfusion and warrants further investigation.
## Data availability
Summary data that underlie the results presented here are available from the corresponding author upon reasonable request, but individual patient data are not freely available due to the constraints of the consent gained from patients and the restrictions imposed by local regulations at the time of recruitment. |
Pressing needs of biomedical text mining in biocuration and beyond: opportunities and challenges
Citation details: Singhal,A., Leaman,R., Catlett,N., et al. Pressing needs of biomedical text mining in biocuration and beyond: opportunities and challenges.AbstractText mining in the biomedical sciences is rapidly transitioning from small-scale evaluation to large-scale application. In this article, we argue that text-mining technologies have become essential tools in real-world biomedical research. We describe four large scale applications of text mining, as showcased during a recent panel discussion at the BioCreative V Challenge Workshop. We draw on these applications as case studies to characterize common requirements for successfully applying text-mining techniques to practical biocuration needs. We note that system 'accuracy' remains a challenge and identify several additional common difficulties and potential research directions including (i) the 'scalability' issue due to the increasing need of mining information from millions of full-text articles, (ii) the 'interoperability' issue of integrating various text-mining systems into existing curation workflows and (iii) the 'reusability' issue on the difficulty of applying trained systems to text genres that are not seen previously during development. We then describe related efforts within the text-mining community, with a special focus on the BioCreative series of challenge workshops. We believe that focusing on the near-term challenges identified in this work will amplify the opportunities afforded by the continued
# Introduction
The unprecedented advances in high-throughput technology and tools to support bioscience have led to a boom in biological and biomedical science research and an accompanying growth of the scientific literature. Access to the wealth of knowledge embedded in the literature is critical for enabling continued scientific advancements and breakthroughs. For this reason, several efforts over the last decade have focused on improving knowledge reusability through improved storage, representation and curation. These efforts include both public literature resources (e.g. PubMed and PubMed Central/Europe PMC) and biological knowledge bases (e.g. UniProt (1), NCBI Database Resources [bib_ref] Database resources of the National Center for Biotechnology Information, Ncbi Resource Coordinators [/bib_ref]. [fig_ref] Figure 1: Interconnection between literature services and biological databases [/fig_ref] illustrates the interconnection between literature services and biological databases, and their importance in biological research. As can be seen, researchers rely on literature services to keep up with the state of the art on topics of their interest, to generate novel hypotheses, and as a reference for developing research strategies. In addition, today's curated databases are critical in biomedical research by being a firsthand tool for researchers to investigate their hypothesis or research results [bib_ref] The importance of biological databases in biological discovery, Baxevanis [/bib_ref].
Biological knowledge bases rely heavily on expert curation, however, and scaling to accommodate the growth of the scientific literature has been a continued challenge. Automatically annotating biological entities such as genes/ protein and diseases [bib_ref] Europe PMC: a full-text literature database for the life sciences and platform..., Europe [/bib_ref] and other scientific artifacts in biomedical literature, such as investigation techniques or the dataset used [bib_ref] Tools of discovery, Lemberger [/bib_ref] is useful for improving the scalability of biocuration services. Surveys regarding the role of text mining for assisting literature curation were performed during the International Biocuration Conference and Workshop and the BioCreative 2012 Workshop (Washington, DC) [bib_ref] Biocuration workflows and text mining: overview of the BioCreative, Lu [/bib_ref]. The 2012 report indicates that more databases have adopted text mining into their curation workflows in some form than in 2009. A number of studies have indicated improved curation productivity with the assistance of text mining. In [fig_ref] Table 1: A selection of studies demonstrating the benefit of text mining assistance for... [/fig_ref] , we present a subset of studies benchmarking the quantitative significance of text-mining systems in database curation [bib_ref] Text mining in the biocuration workflow: applications for literature curation at WormBase,..., Van Auken [/bib_ref] [bib_ref] Accelerating literature curation with text-mining tools: a case study of using PubTator..., Wei [/bib_ref] [bib_ref] The eFIP system for text mining of protein interaction networks of phosphorylated..., Tudor [/bib_ref] [bib_ref] tagtog: interactive and text-mining-assisted annotation of gene mentions in PLOS full-text articles, Cejuela [/bib_ref]. We also refer the reader to the Interactive Annotation Task (IAT) at BioCreatives III-V [bib_ref] BioCreative III interactive task: an overview, Arighi [/bib_ref] [bib_ref] An overview of the BioCreative, Arighi [/bib_ref] [bib_ref] BioCreative IV interactive task, Matis-Mitchell [/bib_ref] , which investigated some aspects of usability and productivity of the text-mining systems for biocuration.
Given the earlier successes and increasing cost/limited resources in manual curation, we argue that computational approaches such as text mining are essential in the future to provide researchers and medical professionals efficient, comprehensive and up-to-date literature services to manage this growth according to customizable criteria such as clinical relevancy or specific genes or species. Since most of the discoveries and breakthroughs are first made available to the public through scholarly publications, the emphasis of this position article is with regard to text-mining applications in literature search and curation. Specifically, the four real-world applications discussed are (i) Literature search (Europe PMC), (ii) Data search (SourceData), (iii) BEL database curation and (iv) VIROME database curation.
In this article, we first discuss the two applications related to literature services, Europe PMC and SourceData, explaining both their value to the bioscience community and how text mining is essential for their continued progress. We next discuss two recent efforts supporting biological databases, BEL and VIROME, which curate information related to biological cause-effect relationships and microbiomes, respectively. Finally, we summarize the opportunities for text mining in such applications and the multiple challenges that hamper its immediate adoption in these applications. We also provide our understanding of a few strategies to facilitate an increased adoption of text mining in such applications.
## Real world large scale applications
Europe PMC (Johanna McEntyre, EMBL-EBI) Europe PMC (https://europepmc.org/) is a database of abstracts and full text articles [bib_ref] Europe PMC: a full-text literature database for the life sciences and platform..., Europe [/bib_ref]. Partnering with PMC from the National Library of Medicine USA and PMC Canada as a PMC International node, it contains 30 million abstracts (including PubMed) and over 3.6 million full text articles from the life sciences. In addition to serving general life science researchers who use Europe PMC to search the literature and access full text articles, Europe PMC also seeks to serve the specialized subset of users who are database curators. Database curators are professional literature readers, filterers, evaluators and extractors who work with the purpose of adding scientific value and context to public data resources.
Manual literature curation has resulted in many bioinformatics resources of excellent quality. It is clear, however, that some supportive computational approaches will be required in order for curation to scale to the accelerating pace of the biomedical literature while maintaining scientific quality. Since curators often require a wide variety of highly specific information, providing text-mining tools to fill each need may be a complex and never-ending task.
However, text-mined outputs useful for curators are also likely to be useful for others in the broader scientific community; integrating text mining into Europe PMC therefore also opens the possibility for occasional users to contribute to community curation efforts and provide feedback on text-mining results.
Europe PMC is committed to enabling text mining. Currently, it provides features such as 'Highlight terms' to identify core biological entities such as genes/proteins and organisms within the article's abstract view; the entities are also linked to relevant databases. A similar feature is provided for full-text article views. Europe PMC is also being developed as a platform for third-party text-mining algorithms, allowing the output of these algorithms to be displayed in full text articles shown on the Europe PMC website.
In the future, outputs from the text-mining community could further semantically enrich Europe PMC content by including the annotation of additional entities, such as mutations, and/or relationships between various entities., such as genes/proteins and diseases. Search and browse features built on top of these annotations-e.g. references to other articles studying the same relationships (or, perhaps, contradictory relationships)-will help readers to better judge the article in light of related publications.
## Sourcedata (thomas lemberger, embo and ioannis xenarios, sib)
Hypothesis-driven research in molecular and cell biology primarily generates data from small-scale experiments. In scientific publications, such data are visually depicted in figures or tables. However the original data behind the figures-the 'source data'-are almost never available in structured format that would make them findable and reusable.
SourceData (http://sourcedata.embo.org/) (18) is building tools to allow researchers and publishers to generate machine-readable descriptions of data during the publication process and also to make this data searchable. To facilitate generating structured experimental descriptions, SourceData has developed an online tool for computerassisted manual curation of figures and figure legends by data editors. The intention is to integrate a curation step into the publishing workflow to annotate figures of before article publication. Authors then verify and approve the curated information through a validation interface. The result is a machine-readable representation of the data (descriptive metadata) based on the information routinely provided by authors in the text of figure legends, thus respecting the traditional workflow adopted by scientists.
SourceData have also developed a search interface that allows users to search for specific experimental evidence and the articles where these data have been reported. This search function is incorporated into the 'SourceData SmartFigure' viewer, which can easily be embedded in online publications. The SmartFigure application allows a specific figure panel to be linked with figures presenting similar data published elsewhere and therefore makes it possible for users to traverse the web of connected data by following these links across articles. Finally, programmatic access to the SourceData database is provided to the research community through a public API.
Integration of text mining with manual curation in the context of publishing seems to be a promising direction, as [bib_ref] Accelerating literature curation with text-mining tools: a case study of using PubTator..., Wei [/bib_ref] TAIR Genes PubTator 45% increase in productivity [bib_ref] The eFIP system for text mining of protein interaction networks of phosphorylated..., Tudor [/bib_ref] PIR PPI involving protein phosphorylation eFIP 2.5-fold increase curation efficiency [bib_ref] tagtog: interactive and text-mining-assisted annotation of gene mentions in PLOS full-text articles, Cejuela [/bib_ref] Flybase Genes Tagtog 2-fold decrease in curation time it will improve the efficiency and speed of the metadata extraction process and it will allow supervision of the automated results by both data editors and authors. In this context, text-mining methods will be useful for the automated semantic enrichment of figure legends or of the corresponding referring statements in the full text and also for identifying entity relationships that represent tested experimental hypotheses. Text mining is also envisioned to play a complementary role by linking curated figures with interpretative statements made in the article or with reagents listed in 'Materials and Methods' section. Finally, textmining techniques developed for computer science publications [bib_ref] Leveraging web intelligence for finding interesting research datasets, Singhal [/bib_ref] might be useful to automatically prioritize a pool of candidate publications for further extraction of detailed experimental data and metadata.
OpenBEL: computable knowledge bases of cause-effect relationships (Natalie Catlett, Selventa)
Biological Expression Language (BEL) is a knowledge representation developed by Selventa to capture biological cause-and-effect relationships from the scientific literature in a format suitable for computation. BEL and its associated software platform are an open source project (www. openbel.org). BEL knowledge bases have been used to support inference from molecular profiling data [bib_ref] Assessment of network perturbation amplitudes by applying high-throughput data to causal biological..., Martin [/bib_ref] [bib_ref] Quantitative assessment of biological impact using transcriptomic data and mechanistic network models, Thomson [/bib_ref] [bib_ref] Reverse causal reasoning: applying qualitative causal knowledge to the interpretation of high-throughput..., Catlett [/bib_ref] and to construct of network models representing specific biological processes [bib_ref] Causal biological network database: a comprehensive platform of causal biological network models..., Boue [/bib_ref]. These approaches support precision medicine by illuminating the molecular mechanisms of disease, drug mechanisms of action, and supporting patient stratification. BEL is designed to represent experimental observations in molecular biology, providing specific representations of various biological measurements including RNAs, proteins, post-translationally modified proteins, and protein activities, as well as biological processes and pathologies. This granular representation facilitates mapping of biological measurements to BEL networks to drive interpretation of molecular profiling data. BEL also represents the context for these experimental observations, such as the cell line or tissue used for the experiment, as well as a literature citation, allowing the creation of BEL networks that accurately represent the experiment and its context.
Over the last decade, Selventa has built a knowledge base comprised of >500 000 BEL statements primarily through manual curation. Many of these statements resulted from targeted curation efforts to support projects in various disease areas. This approach requires a significant effort from trained scientists to build a comprehensive knowledge base and keep it current.
Text mining promises to greatly improve the efficiency of building BEL knowledge bases. Accurate entity identification from the literature is critical to generating BEL knowledge bases useful for inference or building models. Another computational aspect important for automation is relation identification. Recently, Fluck and colleagues developed BELIEF, a text-mining work flow to improve the efficiency of BEL curation [bib_ref] BELIEF-a Semiautomatic Workflow for BEL Network Creation, Fluck [/bib_ref]. BELIEF includes a UIMAbased text-mining workflow (with several state-of-the-art natural language processing, named entity recognition (NER) and relationship extraction tools) to facilitate a semi-automatic curation pipeline. Use of BELIEF was shown to significantly reduce human curation effort. VIROME and building a knowledge base for microbiomes (Shawn Polson, University of Delaware)
Microbial communities and viral assemblages have been found to be both numerous and important drivers of biological processes globally. Recent research has linked microbiomes, microbes co-existing with a host, to many normal and pathological processes such as co-metabolism of food sources, exclusion of pathogens, fostering of host immune response, obesity, susceptibility to cancer and even mental disorders [bib_ref] Temporal variation selects for diet-microbe co-metabolic traits in the gut of Gorilla..., Gomez [/bib_ref] [bib_ref] Control of brain development, function, and behavior by the microbiome, Sampson [/bib_ref] [bib_ref] Mucosal immunity and the microbiome, Neish [/bib_ref] [bib_ref] The microbiome and regulation of mucosal immunity, Mcdermott [/bib_ref] [bib_ref] High-fat-dietmediated dysbiosis promotes intestinal carcinogenesis independently of obesity, Schulz [/bib_ref]. Research aimed at unraveling the complex community-scale dynamics and functions of microbial communities, and the even more numerous viruses which play important roles in regulating and driving genetic diversity among them, are of paramount importance. Our ability to examine these systems was once limited by factors including the inability to cultivate the vast majority in a laboratory setting, but the advent of increasingly cost-effective platforms for deep sequencing of marker genes (e.g. 16S rRNA) and metagenomes in the past several years have finally opened the door for wide-spread research in this field.
These methods involve generation of raw sequence data elucidating the taxonomic or functional composition of the community at a specific geographic location, time, and environmental condition. Typically a study will include multiple samples varying across some spatial, temporal, or environmental variable allowing for testing of one or more specific hypotheses. The global nature of such data, however, means that its utility could extend far beyond the specific hypotheses it was collected to address. The results of such studies are typically published in peer-reviewed journals with deposition of only the raw data to public repositories such as the NCBI Sequence read archive. Other fields have seen the utility of publishing the analysed results of sequence-based studies (e.g. GEO for gene expression data). Some online tools such as VIROME [bib_ref] VIROME: a standard operating procedure for analysis of viral metagenome sequences, Wommack [/bib_ref] and MG-RAST (33) do provide a route for the analysis results themselves to be made public. Recent work by VIROME (http://virome.dbi.udel.edu/) and others are working to ensure that such results are accompanied by standardized metadata to make them useful when considered in alternate contexts, but the ability to look for trends across projects remains very limited. Leveraging microbial ecology results garnered from disparate projects could prove transformative for the field. Agreements are lacking, however, to populate centralized repositories with analysed data in a manner that would enable the creation of comprehensive microbial ecology resources, similar to what UniProt (1) and the Protein Information Resource (34), among others, provide for proteins. Development of such resources would enable large-scale observations and hypothesis testing, such as to assess the range of conditions under which a given microbe (or microbial protein) has been observed, thus providing key insights into its role, or assessing synergistic relationships by determining the consistency of co-occurrence for two or more microbes.
Text mining should play a key role in future microbiome studies by providing standalone tools to search for specific microbial relationships in the literature and populating databases designed to provide comprehensive views of such global data. Microbes almost always live in mixed communities, and thus cooperation and competition are key features; however, detecting such microbial dependencies is difficult and time-consuming. Similarly, defining the environmental parameters under which certain microbes or guilds of microbes exist can be very informative in understanding their roles. Single studies are rarely comprehensive enough to elucidate such trends, however. Textmining tools may enable a comprehensive understanding of microbiomes by focusing on the NER of specific microbial entities, the extraction of biological conclusions (e.g. organism x can do y, but only in the presence of z), metadata extraction (description of time, place, and conditions of samples at collection), and methodological details of the original sample. The ENVIRONMENTS and EXTRACT tools presented at BioCreative V [bib_ref] EXTRACT: interactive extraction of environment metadata and term suggestion for metagenomic sample..., Pafilis [/bib_ref] [bib_ref] ENVIRONMENTS and EOL: identification of Environment Ontology terms in text and the..., Pafilis [/bib_ref] , are examples of such tools, with emerging capability to extract environmental context and microbial taxonomy from published articles and map them to ontological standards such as Environmental Ontology (37).
## Text-mining needs in large scale applications
The text-mining needs in aforementioned applications can be grouped into three primary tasks: NER, relation extraction (RE) and information visualization. NER involves automatically labeling bio-entities such as dataset name (SourceData), diseases, genes, proteins (BEL, Europe PMC) or microbial proteins (VIROME). Since NER is foundational to most text-mining applications, the availability of accurate application-specific NER tools is critical [bib_ref] TaggerOne: joint named entity recognition and normalization with semi-Markov models, Leaman [/bib_ref] [bib_ref] SR4GN: a species recognition software tool for gene normalization, Wei [/bib_ref]. RE introduces the next higher level of knowledge discovery by automatically extracting relationships between the entities identified by NER. Such relationships may describe cause-effect relations (BEL) or microbe-environment relations (VIROME), and relationships may also involve metadata (such as spatio-temporal variables) to curate complex higher order relations. The final task is visualization of the text mined results. Some applications require visuals-summaries (or visual tags), links to other online databases (EuropePMC) and metadata highlight within text (SourceData)-to enhance knowledge representation. Text mining can help in selecting the most relevant outputs from large scale text-mined results, as not all textmined outputs need be displayed even if they are correctly extracted. Although text-mining roles may be classified broadly into three tasks, the specific entities, relations and representation required for each application may be highly specific.
## Challenges and opportunities in text mining
These domain applications above suggest several areas that remain challenging, namely 'accuracy', 'scalability', 'interoperability' and 'reusability'. These areas represent future opportunities for text mining to address the real world needs of large scale applications.
## Accuracy
Although text-mining systems are rapidly transitioning to real world use, imperfect accuracy remains a limiting factor. Workflows incorporating text-mining systems must design processes that compensate for imperfect output. Although the importance of these considerations tapers as the output quality approaches that of human annotators, there are several limitations with the evaluations typically performed in the text-mining community. First, the evaluation most commonly performed is intrinsic, that is, it compares the output of the system to gold standard annotations performed by human annotators. Although such an evaluation provides several desirable properties, such as being quantifiable and providing a high degree of objectivity, it does miss some important considerations. Notably, it provides no feedback on whether the quality of the output is sufficient to support processes downstream in the workflow. Thus, while intrinsic evaluation of the system is important, the system must also be evaluated extrinsically, i.e. in place in the workflow.
## Interoperability
Because system accuracy is critical and must be evaluated extrinsically in the workflow, each system evaluated must be fully integrated into the workflow. Thus, the difficulty in integrating the system must be kept to a minimum. Unfortunately, many factors reduce system interoperability, such as operating system dependencies and incompatibilities between input and output formats [bib_ref] Web services-based text-mining demonstrates broad impacts for interoperability and process simplification, Wiegers [/bib_ref]. Interoperability could be addressed in several different ways. For instance, UIMA (41) is a software architecture created by IBM in 2003 to provide uniform data formatting standards for different teams working on NLP projects. Although it uses a common analysis system (CAS), the ability to use different semantic tag sets creates an interoperability solution [bib_ref] Abstracting the types away from a UIMA type system, Verspoor [/bib_ref]. Tools written in a systemindependent language such as Java or Python do not require a specific operating system. Format incompatibilities can be addressed by creating a standard data format. The recent BioC project is such an example, which has created an interoperable data format that is both straightforward and sufficiently expressive to represent a wide variety of text-mining tasks [bib_ref] BioC: a minimalist approach to interoperability for biomedical text processing, Comeau [/bib_ref] [bib_ref] BioC interoperability track overview, Comeau [/bib_ref]. Another solution may be web services, which hides all configuration and deployment details from the user by providing an API that can be accessed over the Internet, requiring no system installation or maintenance [bib_ref] Reflect: augmented browsing for the life scientist, Pafilis [/bib_ref] [bib_ref] Beyond accuracy: creating interoperable and scalable text-mining web services, Wei [/bib_ref]. Despite these attempts, integrating text mining into mature database workflows remains difficult due to the complexities of curation workflow and existing infrastructure.
## Scalability
A defining characteristic of large-scale text-mining applications is the requirement to scale to millions of documents. PubMed, e.g. contains over 25 million abstracts-at the relatively high rate of 100 abstracts per second, it therefore requires nearly 3 days of computational time to process; processing an equivalent amount of full text articles requires an order of magnitude longer. Text-mining implementations are therefore frequently paired with a database, allowing the text to be preprocessed and the results cached and indexed. Although this allows the textmining results to be provided on demand for text available beforehand, this approach is insufficient for text that must be processed in real time. Moreover, this approach is also inconvenient for updates to the text-mining system, as all the cached results must be reprocessed. One approach to address scalability is the application of cluster computing: processing multiple documents in parallel on multiple hardware systems. Returning to our PubMed example, a cluster of 10 systems-each processing at the rate of 100 abstracts per second-is sufficient to reduce the processing time to under 7 h, a job which can be completed overnight.
## Reusability
Text-mining systems are commonly applied to text somewhat different than the text used to train and evaluate them, making generalization-the ability to handle text previously unseen-very important. As an example, abstracts describing rare genetic diseases will contain significantly different information than those describing treatments for tropical infectious diseases, even though both will contain disease entities. A particular concern is the ability of the system to handle not only abstracts, but also full text documents [bib_ref] LINNAEUS: a species name identification system for biomedical literature, Gerner [/bib_ref] [bib_ref] BioContext: an integrated text mining system for large-scale extraction and contextualization of..., Gerner [/bib_ref] [bib_ref] Systematic characterizations of text similarity in full text biomedical publications, Sun [/bib_ref] [bib_ref] GeneView: a comprehensive semantic search engine for PubMed, Thomas [/bib_ref] [bib_ref] Intrinsic evaluation of text mining tools may not predict performance on realistic..., Caporaso [/bib_ref] [bib_ref] The structural and content aspects of abstracts versus bodies of full text..., Cohen [/bib_ref]. However, systems for dealing with many of the various nuances (such as figure captions, data in tables, information in supplementary materials, and various text cleaning issues) of full text are still not fully in place. Thus, a large improvement in the robustness of a system against shifts in the textual domain may be significantly more useful for real world applications than incremental improvements in system accuracy.
## Future roles of researchers, publishers and curators
Bridging the gap between text-mining research and its application in real world databases requires a collaborative effort from the various stakeholders involved in advancing biomedical sciences. In this section, we provide a few perspectives which researchers, publishers and curators can use to advance biomedical sciences through text mining.
## Research community
Community run challenges in biomedical text mining such as BioCreative can play a major role in realizing the potential of large scale text-mining applications, both by assessing the state of the art and also helping advance the field [bib_ref] Community challenges in biomedical text mining over 10 years: success, failure and..., Huang [/bib_ref]. The aim of conducting these challenges, in general, is to promote interdisciplinary collaboration, evaluate and advance the NLP techniques to facilitate biological research. Thus, these challenges are conducted as shared tasks where research teams from across the globe participate in fulfilling the goals of specified text-mining tasks. A myriad of such challenges have been organized over the years following the success of CASP in 1994 (54, 55) on protein structure prediction; Huang et al. (2016) (53) provides a comprehensive overview of several challenges conducted within the last decade.
In recent years, the community has introduced challenges that focus on bridging the gap between biomedical text-mining research and new application domains. For example, since 2010, BioCreative has organized workshops at the annual meetings of the International Society for Biocuration (http://biocuration.org/) with a focus on better understanding biocuration workflows [bib_ref] Biocuration workflows and text mining: overview of the BioCreative, Lu [/bib_ref] and promoting the development and deployment of biomedical text-mining tools into production curation pipelines. Several of these have been successfully integrated into existing curation workflows (e.g. [bib_ref] PubTator: a web-based text mining tool for assisting biocuration, Wei [/bib_ref] [bib_ref] tagtog: interactive and text-mining-assisted annotation of gene mentions in PLOS full-text articles, Cejuela [/bib_ref].
Nevertheless, there are several difficulties which must be resolved before community challenges can realize the potential of large scale text-mining applications. The foremost of these difficulties is that challenge tasks are often simplified or abstracted versions of the real-world problems. For example, although biocurators routinely use the full text of an article [bib_ref] Recommending MeSH terms for annotating biomedical articles, Huang [/bib_ref] , challenge tasks often only utilize the abstract due to difficulties in accessing full text articles and processing full text. A consequence of this simplification of the real-world problem is that even systems that perform well on challenge tasks yield significantly lower results when evaluated in practical real-world settings. For example, previous BioCreative Gene Normalization challenges have shown that the task performance dropped significantly when tested on full texts (58) instead of abstracts [bib_ref] Overview of BioCreative II gene normalization, Morgan [/bib_ref]. These difficulties can be addressed by designing challenge tasks that focus on the unique problems presented by real world applications.
The BioCreative Collaborative Biocurator Assistant Task (BioC) and the BioCreative Interactive Text-Mining Task (IAT) serve as examples of such focused efforts. The BioC task centered on creating a text-mining system to support BioGRID curators by developing BioC-compatible text-mining modules complementing each other and integrated into one system. The IAT task involved biocurators in testing text-mining systems. In a similar vein, we describe below a few ideas that can be realized as challenge tasks in BioC workshops in the near term to help realize the opportunities of text-mining research in real-world applications more directly.
i. Creating a wide variety of manually curated benchmarks datasets for various text-mining problems. These benchmarks datasets are critical for text-mining researchers to train, test and compare their algorithms and also for organizations to determine the best fit for their large scale applications. These benchmarks should come from various sources including biomedical literature (both abstract and full text), clinical trials, clinical notes and Electronic Medical Records. ii. Identifying metrics to measure critical system qualities in addition to accuracy. As application needs differ, so do their evaluation criteria for selecting text-mining tools. Identifying or creating metrics addressing performance aspects beyond accuracy, such as scalability, usability, and cost-of-adoption (such as database management and front-end design) will greatly help both researchers and application developers to identify textmining tools that best fit their performance dimensions. In this direction, BioCreative-IAT task has included both performance and usability metrics in the evaluation of the text-mining systems by curators, which were also adopted in the BioC task. These metrics should be extended to include scalability and costof-adoption. iii. Like BioC's focus on BioGRID, challenge tasks can be designed to focus on individual large scale applications such as SourceData, BEL and VIROME. Involving the data indexers and curators in the task design step will enrich the utility of the challenge task for real-world use. Parameters such as evaluation criteria can be designed specifically for the individual application. Moreover, the data bottleneck such as full text access and processing can be addressed with help of literature services such as Europe PMC.
## Publishers' role
The SourceData project provides a good example of how publishers could actively encourage innovative knowledge curation and representation. As described in the SourceData section, the publishers collaborate with researchers to generate machine-readable descriptions of datasets during the publication process and also to make this data searchable. In addition to the role of text-mining expressed earlier, as the databases grow, text-mining systems can be employed in the future to provide automatic recommendations of machine-readable tags or descriptions for the datasets. Similar to SourceData project's initiative to enrich articles during in-publication or pre-publications phase, the publishers' role can be to enrich articles in prepublication phase by employing text-mining systems.
In the future, the curation step may not wait until after publication, as is the current practice. A possibility is to move the curation step 'upstream' i.e. capturing knowledge at the time of peer review and prior to publication. Such an initiative would require development of very high quality and sustainable text-mining systems, and possibly require a greater involvement of the article authors in validating some of the text-mined results.
## Curators' role
It is central to keep the human curators/experts in the loop in any newly proposed text-mining-based curation ecosystem. Curators are critical for defining text-mining requirements, providing annotation guidelines and standards, and providing training data for the initial system development and evaluation. Curators should be involved in evaluating the text-mined results and decide their fitness for curation. Curators should help system developers iteratively improve the text-mining algorithms and make any necessary system customizations for their specific database curation needs. This would be the ideal way to incorporate text mining into curation workflows.
# Conclusions
In this work, we presented four large scale applications of text mining in the biological and life sciences, as showcased during a recent panel at BioCreative V. We used these applications as case studies in the challenges encountered in adopting text-mining solutions into realistic tasks and discussed several areas of opportunity for text mining to support real world services in the near term. Finally, we presented a few actionable steps that the BioCreative community can take to bridge the gap between text-mining research and real world biomedical services.
[fig] Figure 1: Interconnection between literature services and biological databases. [/fig]
[fig] Funding: This work was supported by the National Institutes of Health [R13-GM109648-01A1 to CNA, P20-GM103446 to SWP, Intramural Research Program at National Library of Medicine to A.S., R.L., Z.L.], the US Department of Energy [DE-SC0010838 to C.N.A.], the US National Science Foundation [DBI-1356374 to S.W.P.] for the VIROME project and the Swiss Federal Government through the State Secretariat for Education, Research and Innovation (SERI); SyBIT project of the SystemsX.ch, the Swiss Initiative in Systems Biology (in part) (to IX). The Robert Bosch Foundation and EMBO are acknowledged for funding of the SourceData project. The Wellcome Grant, UK PubMed Central Phase 3 Developments 2012-2016 (grant number 098231/Z/12/Z) for the EuropePMC project. [/fig]
[table] Table 1: A selection of studies demonstrating the benefit of text mining assistance for curation. [/table]
|
Role of Cytokines and Chemokines in Angiogenesis in a Tumor Context
# Introduction
There are two fundamental processes to form blood vessels: vasculogenesis and angiogenesis. Vasculogenesis corresponds to the de novo blood vessel formation, whereas angiogenesis is the formation of new blood vessels from pre-existing vessels. Angiogenesis is required in physiological processes such as embryogenic development and the menstrual cycle. This mechanism is also widely involved in cancer development. The involvement of this process in cancer began to be highlighted in 1800. Indeed, German researchers observed that some tumors were richly vascularized, suggesting that new blood vessel formation happened in some cancers. Later, in 1948, Algire demonstrated in mice that melanoma growth is preceded by blood vessel development [bib_ref] Growth and vascularization of transplanted mouse melanomas, Algire [/bib_ref]. Progressively, research better defined angiogenesis, in 2004 the first anti-angiogenic treatment was approved as a cancer treatment and Hanahan and Weinberg identified it as a hallmark of cancer [bib_ref] Hallmarks of cancer: The next generation, Hanahan [/bib_ref].
The formation of new blood vessels from preexisting vessels is achieved in sequential steps . In a hypoxic environment, angiogenic factors bind to their receptor, present at the surface of endothelial cells, promoting their dilatation and activation. Simultaneously, . Neoangiogenesis in tumor. A tumor needs nutrients and oxygen (O 2 ) to support neoplastic expansion. The provision of these needs requires the establishment of a new vascular network through the process of angiogenesis. Angiogenesis consists of the assembly of endothelial cells in the form of tubes from existing vessels. During hypoxia and tumor growth, the nuclear translocation of HIF1α induces the expression of pro-angiogenic factors such as VEGF, EGF, or FGF... Angiogenic factors are able to activate and stimulate endothelial cells through membrane receptors. Indeed, these signals participate in the proliferation, invasion, migration, survival, and increase in the permeability of the vessels. Inspired from the Cancer Research Product Guide Edition 3, 2015.
## Classical regulators of angiogenesis
## Vascular endothelial growth factor (vegf) family
The most important inducer of angiogenesis is the VEGF family. The VEGF family consists of five members: VEGF-A, VEGF-B, VEGF-C, VEGF-D, and placental growth factor (PlGF). Their biological functions are mediated by three receptors: VEGFR-1, VEGFR-2, VEGFR-3, and 2 co-receptors: neuropilin and heparan sulfate proteoglycans. While VEGF-B, PlGF, and VEGF-A bind to VEGFR-1, VEGF-A, VEGF-C, and VEGF-D bind to VEGFR-2, VEGF-C, and VEGF-D bind to VEGFR-3.
The VEGF-A/VEGFR-2 signaling pathway plays a crucial role in physiological and pathological angiogenesis. Mice with VEGF-A or VEGFR-2 deficiencies are not viable and present an early embryogenic lethality due to abnormal vascular development. The VEGF family has a mitogenic and anti-apoptotic effect on endothelial cells and also induces their migration and proliferation. Furthermore, these growth factors promote blood vessel permeabilization for remodeling blood and lymphatic vessels. The expression of VEGF is enhanced by HIF-1α, the activation of oncogenes such as Ras, growth factors, and cytokines such as IL-1, Tumor Necrosis Factor-α (TNFa), Epidermal Growth Factor (EGF), and Platelet-Derived Growth Factor (PDGF). Various cells produce VEGF-A such as smooth muscle cells, keratinocytes, endothelial cells, platelets, neutrophils, and macrophages. It is believed that approximately 60% of all tumors secrete this molecule. There are many reviews on the VEGF family, so we decided not to go into details [fig_ref] Figure 2: Role of classical regulators of angiogenesis [/fig_ref].
The biological function of PlGF is mediated through VEGFR-1 and two co-receptors, neuropilin-1 and neuropilin-2 [bib_ref] Placenta growth factor. Potentiation of vascular endothelial growth factor bioactivity, in vitro..., Park [/bib_ref]. This growth factor directly induces angiogenesis by increasing tumor vascularity and blood vessel growth and promotes survival, proliferation, and migration of endothelial cells [bib_ref] Placental growth factor is a survival factor for tumor endothelial cells and..., Adini [/bib_ref] [bib_ref] Altered tumor vessel maturation and proliferation in placenta growth factor-producing tumors: Potential..., Taylor [/bib_ref]. In endothelial and vascular cells, the overexpression of HIF-1α induces the expression of PlGF [bib_ref] Hypoxia-inducible factor-1 mediates activation of cultured vascular endothelial cells by inducing multiple..., Yamakawa [/bib_ref]. Contrary to VEGF, PlGF is not required for embryogenic vessel formation but contributes to pathological angiogenesis. Indeed, mice lacking PlGF develop normal blood vessels but tumor growth and angiogenesis are reduced [bib_ref] Synergism between vascular endothelial growth factor and placental growth factor contributes to..., Carmeliet [/bib_ref]. On contrary, Yang et al. have shown that T241 and LLC, two tumor mice models genetically modified to overexpress PlGF, present a slower tumor growth. Furthermore, these tumors present a low density of microvessels, and blood vessels are normalized. Interestingly, they also demonstrated that T241-VEGF-null cells, overexpressing PlGF, grew faster in mice, suggesting that PlGF promotes tumor growth in cells lacking VEGF expression [bib_ref] Vascular endothelial growth factor-dependent spatiotemporal dual roles of placental growth factor in..., Yang [/bib_ref]. Recently, it has been shown that PlGF is secreted by Th17 cells in vitro and in vivo. In turn, PlGF regulates Th17 differentiation through a STAT3-dependent pathway and is able to replace IL-6 functions in th17 differentiation [14].
## The ang-tie system
Angiopoietins (Ang) stimulate angiogenesis and control vascular remodeling and maturation. There are four members: Ang-1 and Ang-2 are well characterized but less is known about the two others: Ang-4 and its mouse ortholog, Ang-3 [15-17]. Their biological functions are mediated by two receptors: Tie-1 and Tie-2 which are nearly exclusively expressed in the endothelium but also in some hematopoietic cells . Tie-1 is an orphan receptor, meaning that it is activated by angiopoietin through its interaction with Tie-2 [20]. This system plays a crucial role in angiogenesis. Indeed, mice with Ang-1, Tie-1, or Tie-2 deficiencies present an abnormal vascular system resulting in embryonic lethality . However, in mice deficient in Ang-2, developmental angiogenesis is mostly unaffected but results in newborns with lymphatic dysfunction and sometimes postnatal death due to chylous ascites . Interestingly, Ang-2 overexpression causes embryonic lethality . Ang-2 can act as an agonist or antagonist of Tie-2, depending on the context. When Ang-2 acts as an antagonist, it induces vascular destabilization and leakiness leading to vascular regression . Under normal conditions, blood vessels are stable and quiescent. Ang-2 is stored in Weibel-Palade bodies [25] and its expression is low. Ang-1 suppresses Ang-2 transcription and its expression dominates. Ang-1 is a constitutive agonist of Tie-2. This molecule is expressed by mural cells, fibroblasts, tumor cells, and non-vascular cells . The Ang-1/Tie-2 signaling pathway increases vascular stability and inhibits vascular permeability by acting on the EC-EC junction and on the actin cytoskeleton . Under inflammatory conditions, Ang-2 is upregulated and competes with Ang-1 for binding to Tie-2. Ang-2 is rapidly released from endothelial cells and its effects are amplified by cytokines such as TNF-α and VEGF . During inflammation, Ang-2 switches to an antagonist function and this mechanism depends on the Tie-1 receptor cleavage . Ang-2 is highly expressed in many types of tumors such as melanoma, RCC, glioblastoma, breast, and colorectal cancer [32-35] and Ang-2 deficient mice show a reduced tumor growth in metastatic colony formation in the lung [bib_ref] G-CSF rescues tumor growth and neo-angiogenesis during liver metastasis under host angiopoietin-2..., Im [/bib_ref] [fig_ref] Figure 2: Role of classical regulators of angiogenesis [/fig_ref].
## Hepatocyte growth factor (hgf)
This growth factor is commonly produced by stromal cells such as fibroblasts and also by colorectal and breast cancer cells due to HGF promoter region mutations [bib_ref] Genomic instability causes HGF gene activation in colon cancer cells, promoting their..., Seneviratne [/bib_ref] [bib_ref] Somatic mutation and functional polymorphism of a novel regulatory element in the..., Ma [/bib_ref] [bib_ref] Identification of the hepatocyte growth factor receptor as the c-met proto-oncogene product, Bottaro [/bib_ref]. HGF is secreted in an inactive proform (pro-HGF) and its activation is mainly due to proteases that are over-expressed in tumor cells [bib_ref] Pericellular activation of hepatocyte growth factor by the transmembrane serine proteases matriptase..., Owen [/bib_ref]. The overexpression of HGF in colorectal cancer stages II and III is associated with poor outcome in patients [bib_ref] Serum hepatocyte growth factor as a prognostic marker for stage II or..., Toiyama [/bib_ref]. This molecule contributes to angiogenesis by promoting endothelial cell growth, survival, and migration and also stimulates epithelial-mesenchymal transition (EMT) by activating its receptor, the mesenchymal-epithelial transition factor (c-MET) [bib_ref] Hepatocyte growth factor is a potent angiogenic factor which stimulates endothelial cell..., Bussolino [/bib_ref] [bib_ref] Scatter factor induces blood vessel formation in vivo, Grant [/bib_ref]. The two molecules, c-MET and HGF have an increased expression in various cancers such as non-small cell lung carcinoma (NSCLC), gastric, ovarian, pancreatic, thyroid, breast, head and neck, colon, and kidney carcinomas [bib_ref] c-MET as a potential therapeutic target and biomarker in cancer, Sierra [/bib_ref]. Furthermore, an in vitro study has demonstrated the ability of HGF to stimulate esophageal squamous cell carcinoma to express VEGF and IL-8 and to enhance the migration and invasion of cancer cells [bib_ref] Hepatocyte growth factor promotes cancer cell migration and angiogenic factors expression: A..., Ren [/bib_ref]. The VEGF expression induced by HGF increases angiogenesis and it has been shown that HGF can induce VEGF transcription through SP1 phosphorylation [bib_ref] Increased Sp1 phosphorylation as a mechanism of hepatocyte growth factor (HGF/SF)-induced vascular..., Reisinger [/bib_ref] [fig_ref] Figure 2: Role of classical regulators of angiogenesis [/fig_ref].
## Fibroblast growth factor (fgf)
The FGF family consists of 22 members: FGF-1 to FGF-23, divided into seven subfamilies: FGF1/2/5, FGF3/4/6, FGF7/10/22, FGF8/17/18, FGF9/16/20, and FGF19/21/23 and FGF 11/12/13/14 [bib_ref] Inhibition of FGF-FGFR and VEGF-VEGFR signalling in cancer treatment, Liu [/bib_ref]. Their biological processes are mediated by four receptors: FGFR1 to 4. There is also a receptor lacking an intracellular kinase domain, FGFR5, that then acts as a coreceptor with FGFR1. These receptors are expressed by endothelial cells. The signaling pathway of FGF/FGFR regulates different biological functions such as endothelial cell proliferation, survival, differentiation, tube formation, protease production, and angiogenesis [bib_ref] Inhibition of FGF-FGFR and VEGF-VEGFR signalling in cancer treatment, Liu [/bib_ref] [bib_ref] Fibroblast growth factor signalling: From development to cancer, Turner [/bib_ref]. However, the contribution of the FGF family to angiogenesis is controversial. It has been shown that FGF-4 and FGF-8 and particularly FGF-1 and FGF-2 have pro-angiogenic properties in different models. In vitro studies have shown that FGF-2, through paracrine and autocrine mechanisms, induces the expression of VEGF by vascular endothelial cells [bib_ref] Fibroblast growth factor-2 (FGF-2) induces vascular endothelial growth factor (VEGF) expression in..., Seghezzi [/bib_ref]. Furthermore, the lack of FGF signaling in endothelial cells downregulates the expression of VEGFR-2 mediated through the activation of Erk1/2 [bib_ref] FGF-dependent regulation of VEGF receptor 2 expression in mice, Murakami [/bib_ref]. There are different cancer types presenting FGFR alteration such as head and neck cancer, non-small cell lung cancer, urothelial cancer, gastric cancer, and breast cancer [bib_ref] Fibroblast growth factor receptors in cancer: Genetic alterations, diagnostics, therapeutic targets and..., Krook [/bib_ref] [fig_ref] Figure 2: Role of classical regulators of angiogenesis [/fig_ref].
## Platelet-derived growth factor (pdgf)
The PDGF/PDGFR signaling pathway plays an important role in angiogenesis and particularly by inducing pericyte recruitment to vessels that allow vessel stability and endothelial cell survival. There are different isoforms of PDGF: PDGF-AA, PDGF-BB, PDGF-CC, PDGF-DD, and PDGF-AB. These molecules are produced by endothelial and epithelial cells and bind to three receptors: PDGFR-αα, PDGFR-αβ, and PDGFR-ββ. PDGFRαα is activated by all PDGF ligands apart from PDGF-DD. PDGFR-αβ is activated by all PDGF ligands except PDGF-AA and PDGFR-ββ is activated by PDGF-BB and PDGF-DD. PDGFR-β is expressed on vascular smooth muscle cells and pericytes and PDGFR-αβ is expressed on endothelial cells. Alterations of these molecules are associated with poor survival, metastatic disease, and tumor angiogenesis. PDGF/PDGFR also induces the stimulation of proangiogenic factors such as VEGF and FGF, endothelial cell proliferation, and recruitment of endothelial precursor cells to vessels. In vivo and in vitro studies have shown that PDGF-D down-regulation in SW480 inhibits tumor growth, migration, and angiogenesis whereas PDGF-D up-regulation in HCT116 is associated with tumor aggressiveness [bib_ref] Ibrutinib Inhibits ERBB Receptor Tyrosine Kinases and HER2-Amplified Breast Cancer Cell Growth, Chen [/bib_ref] [fig_ref] Figure 2: Role of classical regulators of angiogenesis [/fig_ref].
## Interleukines: a link between the immune system and angiogenesis
Angiogenesis is able to modulate the immune system. This mechanism reduces immune cell infiltration by affecting the expression of proteins on endothelial cells. Angiogenesis also induces an immunosuppressive tumor microenvironment. Indeed, it induces the recruitment of immunosuppressive cells such as Treg and MDSC to the tumor, while it reduces DC maturation and CD3 + proliferation and cytotoxicity. Conversely, some immune cells are able to modulate angiogenesis [bib_ref] Vascular Endothelial Growth Factor, a Key Modulator of the Anti-Tumor Immune Response, Geindreau [/bib_ref].
## Interferon family
In humans, there are three subsets of interferon, type I comprising IFNα/β, type II with IFNγ, and also type III with the IFNλs. Type I IFNs are known to inhibit angiogenesis [bib_ref] Inhibition of angiogenesis and vascular tumor growth by interferon-producing cells: A gene..., Albini [/bib_ref] , they prevent the production of proangiogenic factors such as bFGF, VEGF, and IL-8 by tumor cells [bib_ref] Constitutive type I interferon modulates homeostatic balance through tonic signaling, Gough [/bib_ref]. IFNα/β also inhibits the proliferation of endothelial cells and the secretion of endothelial cell chemotaxis molecules. More specifically, IFNα inhibits the production of bFGF and IL-8 by tumor cells in human bladder cancer cells [bib_ref] Inhibition of tumorigenicity and metastasis of human bladder cancer growing in athymic..., Izawa [/bib_ref]. Mice deficient for IFNβ show a faster tumor growth in B16F10 and MCA205 cancer models while wild-type mice show better-developed blood vessels. Mice deficient for IFNβ present an increase in neutrophil infiltration and these cells express higher gene-level expressions of VEGF and MMP9 and CXCR4, a neutrophil tumor-homing factor [bib_ref] Neutrophils responsive to endogenous IFN-beta regulate tumor angiogenesis and growth in a..., Jablonska [/bib_ref]. Type III IFN also has the ability to inhibit angiogenesis [bib_ref] Shared and Distinct Functions of Type I and Type III Interferons, Lazear [/bib_ref]. IFNγ is also known to inhibit angiogenesis by inducing angiostasis, the normal regulation system for the creation of new blood vessels [bib_ref] IFNgamma-responsiveness of endothelial cells leads to efficient angiostasis in tumours involving down-regulation..., Deng [/bib_ref] [fig_ref] Figure 3: Role of cytokines in angiogenesis [/fig_ref].
## The interleukin-1 family
This family is composed of 11 molecules. Among these, there are agonist ligands such as IL-1α, IL-1β, IL-18, IL-33, IL-36α, IL-36β, IL-36γ and antagonist ligands such as IL-1Ra, IL-36Ra, IL-37 and IL-38. This family is able to mediate angiogenesis indirectly or directly by inducing proangiogenic factors such as VEGF. About 30 years ago, it was shown that IL-1 inhibits endothelial cell growth in vitro and in vivo and is able to inhibit the formation of vessels induced by FGF [bib_ref] Interleukin 1 is an autocrine regulator of human endothelial cell growth, Cozzolino [/bib_ref]. In vitro studies have shown that colon, gastric and pancreatic cancer cells can secret IL-1α which in turn enhances angiogenesis [bib_ref] Interleukin-1alpha enhances angiogenesis and is associated with liver metastatic potential in human..., Ma [/bib_ref] [bib_ref] IL-1alpha secreted by colon cancer cells enhances angiogenesis: The relationship between IL-1alpha..., Matsuo [/bib_ref] [bib_ref] Interleukin-1alpha secreted by pancreatic cancer cells promotes angiogenesis and its therapeutic implications, Matsuo [/bib_ref]. Studies demonstrated that IL-1α is able to drive angiogenesis in gastric and prostate cancer [bib_ref] N-myc downstream-regulated gene 1 promotes tumor inflammatory angiogenesis through JNK activation and..., Murakami [/bib_ref] [bib_ref] Mechanism of pro-tumorigenic effect of BMP-6: Neovascularization involving tumor-associated macrophages and IL-1a, Kwon [/bib_ref]. Melanoma cells are able to secrete IL-1α and IL-1β, which in turn upregulate IL-6, IL-8, the intracellular adhesion molecule-1 (ICAM-1), and the vascular cell adhesion molecule-1 (VCAM-1) expression in endothelial cells [bib_ref] Melanoma-derived IL-1 converts vascular endothelium to a proinflammatory and procoagulatory phenotype via..., Strozyk [/bib_ref]. Accordingly, IL-1Ra, which binds to soluble IL-1, reduces angiogenesis [bib_ref] A continuous delivery system of IL-1 receptor antagonist reduces angiogenesis and inhibits..., Bar [/bib_ref]. IL-1β promotes tumor growth in a Lewis lung carcinoma model by upregulating VEGF and CXCL2. IL-1β-deficient mice show no local tumor or lung metastases in a B16 melanoma model injected intravenously or intrafootpad [bib_ref] IL-1 is required for tumor invasiveness and angiogenesis, Voronov [/bib_ref] [fig_ref] Figure 3: Role of cytokines in angiogenesis [/fig_ref].
In the literature, IL-18 is defined as a pro and an anti-angiogenic molecule depending on tissues and context. In the beginning, in vivo studies demonstrated that it negatively regulates neovascularization. In mice subcutaneously injected with T241, IL-18 administration displays an antitumor effect and reduces the microvessel density [bib_ref] Interleukin-18 acts as an angiogenesis and tumor suppressor, Cao [/bib_ref]. Two years later, in vivo and in vitro studies showed that IL-18 can induce endothelial tube formation in rheumatoid arthritis [bib_ref] Evidence of IL-18 as a novel angiogenic mediator, Park [/bib_ref]. In a Lewis lung cancer mice model, IL-18 suppresses tumor growth by down-regulating VEGF-A and VEGF-C expression in tumor tissues. It has also been demonstrated that VEGF increases IL-18 production leading to an increase in gastric cancer cell migration [bib_ref] Interleukin-18 is a critical factor for vascular endothelial growth factor-enhanced migration in..., Kim [/bib_ref]. A recent in vivo study showed that macrophage-derived IL-18 inhibits tumor blood vessel formation [bib_ref] Inhibition of blood vessel formation in tumors by IL-18-polarized M1 macrophages, Xing [/bib_ref] [fig_ref] Figure 3: Role of cytokines in angiogenesis [/fig_ref].
IL-33 is a cytokine with strong angiogenic abilities [bib_ref] Interleukin-33 induces angiogenesis and vascular permeability through ST2/TRAF6-mediated endothelial nitric oxide production, Choi [/bib_ref] [bib_ref] Nuclear interleukin-33 is generally expressed in resting endothelium but rapidly lost upon..., Kuchler [/bib_ref] [bib_ref] Interleukin-33 induces urokinase in human endothelial cells-possible impact on angiogenesis, Stojkovic [/bib_ref]. Its receptor ST2 is highly expressed in colorectal cancer cells, stromal cells, and microvessels of colorectal cancers [bib_ref] Dynamics of the IL-33/ST2 network in the progression of human colorectal adenoma..., Cui [/bib_ref]. IL-33 exhibits a proangiogenic effect on Human Umbilical Vein Endothelial Cells (HUVECs) via the Akt pathway [bib_ref] Interleukin-33 induces angiogenesis and vascular permeability through ST2/TRAF6-mediated endothelial nitric oxide production, Choi [/bib_ref]. Moreover, IL-33 stimulates myofibroblasts to produce the metalloproteases MMP2 and MMP9, involved in the establishment of new vessels [bib_ref] The IL-33/ST2 pathway contributes to intestinal tumorigenesis in humans and mice, Mertz [/bib_ref] [bib_ref] Soluble IL-33 receptor sST2 inhibits colorectal cancer malignant growth by modifying the..., Akimoto [/bib_ref] [bib_ref] Qin, Q. IL-33/ST2 pathway contributes to metastasis of human colorectal cancer, Liu [/bib_ref] [fig_ref] Figure 3: Role of cytokines in angiogenesis [/fig_ref].
## The γc family
This family is based on their shared expression of the cytokine receptor γ c. It is a composed of IL-2, IL-4, IL-7, IL-9, IL-15, and IL-21 [bib_ref] The gammac Family of Cytokines: Basic Biology to Therapeutic Ramifications, Leonard [/bib_ref]. It has been shown that IL-4 is able to block corneal neovascularization through basic fibroblast growth factor and that it inhibits the migration of human microvascular cells [bib_ref] Inhibition of angiogenesis by interleukin 4, Volpert [/bib_ref]. In a spontaneous breast cancer model of mice (PyMT) deficient in IL-4, IL-4 was shown to support vessel remodeling [bib_ref] TGF-beta suppresses type 2 immunity to cancer, Liu [/bib_ref]. In NSCLC, the expression of IL-9 is associated with poor prognosis and promotes angiogenesis via STAT3 [bib_ref] Interleukin-9 promotes tumorigenesis through augmenting angiogenesis in non-small cell lung cancer, He [/bib_ref]. IL-15 reduces the mobility of prostate cancer cells and decreases the number of blood vessels in tumor tissue in vivo in mice [bib_ref] IL-15 regulates migration, invasion, angiogenesis and genes associated with lipid metabolism and..., Rohena-Rivera [/bib_ref]. Finally, in mice that spontaneously develop intestinal tumors, deficiency in IL-21 reduces angiogenesis [bib_ref] Interleukin-21 sustains inflammatory signals that contribute to sporadic colon tumorigenesis, De Simone [/bib_ref] [fig_ref] Figure 3: Role of cytokines in angiogenesis [/fig_ref].
## The interleukin-6 family
This family is composed of different members: IL-6, IL-11, ciliary neurotrophic factor (CNTF), leukemia inhibitory factor (LIF), oncostatin M (OSM), cardiotrophin 1 (CTF1), cardiotrophin-like cytokine factor 1 (CLCF1). Recently IL-27, IL-35, and IL-39 have been added to the interleukin-6 family [bib_ref] Historical overview of the interleukin-6 family cytokine, Kang [/bib_ref].
IL-6 can induce VEGF mRNA expression in A431 cells, a human cell line of epidermoid carcinoma, and also in a rat glioma cell line C6 [bib_ref] Changes in circulating proangiogenic cytokines, other than VEGF, before progression to sunitinib..., Porta [/bib_ref]. Using nude mice, Wei et al. showed that IL-6 promotes tumor growth of a human cervical cancer C33A through VEGF-dependent angiogenesis [bib_ref] Interleukin-6 promotes cervical tumor growth by VEGF-dependent angiogenesis via a STAT3 pathway, Wei [/bib_ref]. In hepatocellular carcinoma, renal cell carcinoma, colorectal cancer, and glioblastoma increased levels of circulating IL-6 are associated with poor response to sunitinib and bevacizumab, a tyrosine kinase inhibitor targeting the VEGF/VEGFR pathway and an anti-VEGF antibody, respectively [bib_ref] Changes in circulating proangiogenic cytokines, other than VEGF, before progression to sunitinib..., Porta [/bib_ref] [bib_ref] Efficacy, safety, and biomarkers of neoadjuvant bevacizumab, radiation therapy, and fluorouracil in..., Willett [/bib_ref] [bib_ref] Modulating antiangiogenic resistance by inhibiting the signal transducer and activator of transcription..., De Groot [/bib_ref] [fig_ref] Figure 3: Role of cytokines in angiogenesis [/fig_ref].
IL-27 inhibits the production of pro-angiogenic factors by A549 cells, in fact, A549 cells treated with IL-27 show a decrease in VEGF, IL-8/CXCL8, and CXCL5 expression in comparison to non-treated cells. Interestingly, the addition of siRNA against STAT1 increases the levels of these proangiogenic molecules, indicating that IL-27 inhibits the production of angiogenic factors through a STAT1 pathway and VEGF production in human NSCLC [bib_ref] IL-27 inhibits epithelial-mesenchymal transition and angiogenic factor production in a STAT1-dominant pathway..., Kachroo [/bib_ref] [bib_ref] Classification of diet-modulated gene signatures at the colon cancer initiation and progression..., Kachroo [/bib_ref] [fig_ref] Figure 3: Role of cytokines in angiogenesis [/fig_ref].
IL-35 is produced in some human cancers such as large B cell lymphoma, nasopharyngeal carcinoma, and melanoma. This interleukin promotes tumor growth by increasing tumor angiogenesis [bib_ref] Tumor-derived IL-35 promotes tumor growth by enhancing myeloid cell accumulation and angiogenesis, Wang [/bib_ref]. IL-35 contributes to the progression of prostate cancer through tumor angiogenesis [bib_ref] Interleukin-35 promotes progression of prostate cancer and inhibits anti-tumour immunity, Zhu [/bib_ref] [fig_ref] Figure 3: Role of cytokines in angiogenesis [/fig_ref].
## The interleukin-17 family
This family of pro-angiogenic molecules is composed of six members: IL-17A, IL-17B, IL-17C, IL-17F, IL-17E (also called IL-25), and IL-17F. These molecules bind to IL-17RA, RB, RC, RD, and RE [bib_ref] IL-17 family: Cytokines, receptors and signaling, Gu [/bib_ref]. Patients with colorectal cancer have a poor prognosis if they present a high IL-17 expression which is associated with high microvessel density in colorectal cancer tissues sample [bib_ref] Increased intratumoral IL-17-producing cells correlate with poor survival in hepatocellular carcinoma patients, Zhang [/bib_ref]. Colorectal carcinoma cell lines express IL-17R and are able to secrete VEGF and IL-6. Interestingly, the stimulation of colorectal carcinoma cells by IL-17 induces the production of angiogenic molecules such as VEGF and IL-6 [bib_ref] IL-17 is associated with poor prognosis and promotes angiogenesis via stimulating VEGF..., Liu [/bib_ref]. Similarly, patients with hepatocellular carcinoma have a poor prognosis when they present an accumulation of Th17 cells in the tumor. The addition of IL-17 in the fibrosarcoma cell line CMS-G4 culture increases the quantity of transcripts of Ang-2 and VEGF [bib_ref] Tumor-infiltrating IL-17-producing gammadelta T cells support the progression of tumor by promoting..., Wakita [/bib_ref]. IL-17 also modulates the production of VEGF by an osteosarcoma cell line. In vivo studies show that IL-17A inhibition at the tumor sites suppressed CD31, MMP9, and VEGF expression in tumor tissues [bib_ref] Inhibition of IL-17A in tumor microenvironment augments cytotoxicity of tumor-infiltrating lymphocytes in..., Hayata [/bib_ref]. IL-17 is also known to promote resistance to VEGF inhibition therapy [bib_ref] An interleukin-17-mediated paracrine network promotes tumor resistance to anti-angiogenic therapy, Chung [/bib_ref]. Contrariwise, IL-17F plays a protective role in colon tumorigenesis because IL-17F-deficient mice show an enhanced tumor development, notably with a downregulated angiogenesis in vivo. Accordingly, an in vivo study shows that IL-17F suppresses the tumor growth in mice bearing the hepatocarcinoma cell line SMMC-7721. In the same study, they show that IL-17F inhibits microvessel formation and that it downregulates VEGF, IL-6, and IL-8 expression in hepatocellular carcinoma [bib_ref] A protective role by interleukin-17F in colon tumorigenesis, Tong [/bib_ref] [fig_ref] Figure 3: Role of cytokines in angiogenesis [/fig_ref].
## The interleukin-12 family
This family includes IL-12 and IL-23. IL-27, IL-35, and IL-39 are also mentioned as members of this family although they are sometimes considered members of the IL-6 family [bib_ref] IL-12 Family Cytokines in Cancer and Immunotherapy, Mirlekar [/bib_ref]. IL-12R is mostly expressed on activated NK and T cells. In vivo studies showed that IL-12 inhibits angiogenesis. One study showed that NK cell neutralization reduces the ability of IL-12 to inhibit angiogenesis in vivo, suggesting that NK cells mediate the inhibition of angiogenesis by IL-12 [bib_ref] Contribution of natural killer cells to inhibition of angiogenesis by interleukin-12, Yao [/bib_ref]. IL-12, by down-regulating angiogenic genes such as CCL2, HIF-1α, VEGF-C, VEGF-D, and IL-6, inhibits the pro-angiogenic activity of human primary lung adenocarcinoma cells [bib_ref] IL-12 can target human lung adenocarcinoma cells and normal bronchial epithelial cells..., Airoldi [/bib_ref]. In a murine breast cancer model, IL-12 also inhibits VEGF and MMP9 expression [bib_ref] IL-12 regulates VEGF and MMPs in a murine breast cancer model, Dias [/bib_ref] [fig_ref] Figure 3: Role of cytokines in angiogenesis [/fig_ref].
## The interleukin-10 family
This family is composed of IL-10, IL-19, IL-20, IL-22, IL-24, and IL-26 [bib_ref] Interleukin-10 Family Cytokines Immunobiology and Structure, Wei [/bib_ref]. IL-10 is an anti-angiogenic molecule. Indeed, SCID mice subcutaneously injected with IL-10 transfected B-cells lymphoma DG75 showed a reduced tumor development in comparison with normal cells. The authors showed that these cells inhibit angiogenesis and that in vitro IL-10 inhibits the proliferation of microvascular endothelial cells induced by VEGF and FGF2 [bib_ref] Abolished angiogenicity and tumorigenicity of Burkitt lymphoma by interleukin-10, Cervenak [/bib_ref]. IL-10 also inhibits angiogenesis in mice injected with an ovarian cancer cell line producing VEGF [bib_ref] Interleukin-10-mediated inhibition of angiogenesis and tumor growth in mice bearing VEGF-producing ovarian..., Kohno [/bib_ref]. It was suggested that IL-10 produced by tumor cells inhibits macrophage-derived angiogenic molecules [bib_ref] Interleukin 10 suppresses tumor growth and metastasis of human melanoma cells: Potential..., Huang [/bib_ref]. In NSCLC, IL-20 is potentially anti-angiogenic because it down-regulates COX-2 and VEGF expression [bib_ref] IL-20, an anti-angiogenic cytokine that inhibits COX-2 expression, Heuze-Vourc'h [/bib_ref] [bib_ref] IL-20 is epigenetically regulated in NSCLC and down regulates the expression of..., Baird [/bib_ref]. IL-22 promotes tumor angiogenesis by stimulating endothelial cell proliferation, survival, and migration. Furthermore, the use of IL-22 neutralizing antibodies inhibits tumor growth, angiogenesis, and microvascular density [bib_ref] Interleukin-22 promotes tumor angiogenesis, Protopsaltis [/bib_ref]. The molecule IL-24, in combination with cisplatin, inhibits tumor growth in the xenografted cervical cancer HeLa in nude mice.
Furthermore, this combination also inhibits angiogenesis by downregulating VEGF, VEGF-C, and PDGF-B expression [bib_ref] Combination of IL-24 and cisplatin inhibits angiogenesis and lymphangiogenesis of cervical cancer..., Wang [/bib_ref] [fig_ref] Figure 3: Role of cytokines in angiogenesis [/fig_ref].
## Chemokines: critical role in tumor angiogenesis
Chemokines are divided into four families: C, C-C, C-X-C, and C-X3-C, depending on the arrangement of the two cysteine residues closest to the N-terminal of chemokines. The C chemokines have only one N-terminal cysteine. The C-C chemokines have these cysteines adjacent, the C-X-C chemokines have an amino acid between these cysteines, and the C-X3-C chemokines have three amino acids between these cysteines. These molecules are mainly known to stimulate leukocyte migration but different studies demonstrated that they also play a role in tumor angiogenesis [bib_ref] A guide to chemokines and their receptors, Hughes [/bib_ref].
## C-c chemokines
As mentioned above, C-C chemokines have two cysteine residues closest to the Nterminal adjacent. This family is composed of 27 chemokines CCL1 to CCL28, with CCL9 and CCL10 being the same chemokine. These molecules bind to 10 different chemokine receptors, CCR1 to CCR10 [bib_ref] Baranowska-Bosiacka, I. CC Chemokines in a Tumor: A Review of Pro-Cancer and..., Korbecki [/bib_ref]. CCL2 is one of the main macrophage chemoattractants and in turn, macrophages recruited by CCL2 secrete proangiogenic molecules such as VEGF. Some patients with glioblastoma multiform (GBM) treated with bevacizumab develop resistance and tumor-associated macrophages notably promote this mechanism. A recent study showed that CCL2 inhibition using mNOX-E38 reduces macrophage migration to CCL2-expressing GBM cells. They also demonstrated that angiogenesis was higher when macrophages and CCL2-expressing cells were cocultured in comparison to CCL2-expressing cells alone. The use of this inhibitor in combination with bevacizumab increases mice survival compared to bevacizumab alone suggesting that CCL2 suppression can increase the efficacy of anti-angiogenic treatments in GBM [bib_ref] Increased Antiangiogenic Effect by Blocking CCL2-dependent Macrophages in a Rodent Glioblastoma Model:..., Cho [/bib_ref]. In patients with endometrial cancer (EC), the expression of CCL4 and VEGF-A is increased in EC tissues in comparison to healthy individuals, and their expressions are positively correlated. In vitro and in vivo studies demonstrated that CCL4 promotes tumor growth by upregulating the expression of VEGF-A, which affected the STAT3 signaling pathway in EC cells [bib_ref] CCL4 promotes the cell proliferation, invasion and migration of endometrial carcinoma by..., Hua [/bib_ref]. In human chondrosarcoma, by down-regulating miR-119, CCL5 promotes VEGF-dependent angiogenesis [bib_ref] CCL5 promotes vascular endothelial growth factor expression and induces angiogenesis by down-regulating..., Liu [/bib_ref]. It also has been shown that CCL11 is able to inhibit angiogenesis by attracting eosinophils in the tumors [bib_ref] CCL11-induced eosinophils inhibit the formation of blood vessels and cause tumor necrosis, Xing [/bib_ref]. In patients with breast cancer, the release of CCL18 by TAMs was positively associated with microvascular density and thus, an increase in angiogenesis. CCL18 also promoted endothelial cell migration and angiogenesis synergistically with VEGF in vitro and in vivo [bib_ref] CCL18 from tumor-associated macrophages promotes angiogenesis in breast cancer, Lin [/bib_ref].
In CRC tissues, CCL19 is low-expressed in comparison to healthy tissues and CCL19 levels are negatively correlated with angiogenesis. In vitro and in vivo studies show that CCL19 suppresses tumor angiogenesis and that it inhibits angiogenesis in CRC by promoting miR-206 [bib_ref] CCL19 suppresses angiogenesis through promoting miR-206 and inhibiting Met/ERK/Elk-1/HIF-1alpha/VEGF-A pathway in colorectal..., Xu [/bib_ref]. In vitro and in vivo studies show that the CCL20/CCR6 axis supports angiogenesis. In vitro, CCL20 promotes endothelial cell migration, tube formation, and angiogenesis [bib_ref] EGFR/Ras-induced CCL20 production modulates the tumour microenvironment, Hippe [/bib_ref]. In HCC tissues, CCL24 is upregulated in comparison to healthy tissue and is correlated with poor prognosis. CCL24 is able to enhance HUVEC tube formation and also contributes to HCC malignancy through the RhoB-VEGF-A-VEGFR-2 signaling pathway [bib_ref] CCL24 contributes to HCC malignancy via RhoB-VEGFA-VEGFR2 angiogenesis pathway and indicates poor..., Jin [/bib_ref]. In an HCC cancer model, hypoxia induces the recruitment of MDSC through CCL26 [bib_ref] Hypoxia induces myeloid-derived suppressor cell recruitment to hepatocellular carcinoma through chemokine (C-C..., Chiu [/bib_ref]. In tumors, CCL28 is induced by hypoxia and is able to promote angiogenesis in lung adenocarcinoma by targeting CCR3 in microvascular endothelial cells. In vitro studies demonstrated that this chemokine is able to promote tube formation, proliferation, and migration of endothelial cells [bib_ref] CCL24 contributes to HCC malignancy via RhoB-VEGFA-VEGFR2 angiogenesis pathway and indicates poor..., Jin [/bib_ref].
## C-x-c chemokines
This family of chemokines presents angiogenic or anti-angiogenic properties based on the ELR (Glu-Leu-Arg) motif. The presence of this motif promotes angiogenesis and its absence inhibits angiogenesis. CXCL1, CXCL6 and CXCL8, CXCLR5 are ELR-positive and promote angiogenesis whereas CXCL4, CXCL10 and CXCL14 are ELR-negative and inhibit angiogenesis. However, CXCL12 is ELR-negative but promotes angiogenesis. CXCL14 transgenic mice injected with Lewis Lung Carcinoma (LLC) or B16 melanoma cells showed a reduced tumor growth and interestingly, CXCL14 transgenic mice injected with LLC showed a decrease in the number and diameters of visible blood vessels in tumors in comparison to WT mice. Furthermore, the percentage of CD31-positive cells in tumors was higher in WT mice. In human CRC tissues, CXCL5 overexpression is positively correlated with the expression of CD31. This chemokine induces the expression of VEGF-A in HUVEC and is also able to promote HUVEC tube formation, migration, and proliferation through CXCR2 [bib_ref] CXCL5 induces tumor angiogenesis via enhancing the expression of FOXD1 mediated by..., Chen [/bib_ref]. CXCL1 is also able to promote angiogenesis in colorectal cancer. Interestingly, the receptor of CXCL1, CXCR2 is elevated in CRC tissue and CXCL1 stimulates tumor growth and increases microvessel density [bib_ref] CXCL1 induced by prostaglandin E2 promotes angiogenesis in colorectal cancer, Wang [/bib_ref]. CXCL10 is able to limit the formation of blood vessels by inhibiting endothelial cell migration. CXCL10 inhibits angiogenesis by binding to CXCR3 expressed on newly forming vessels [bib_ref] An IP-10 (CXCL10)-derived peptide inhibits angiogenesis, Yates-Binder [/bib_ref]. CXCL8 also known as IL-8 was shown to be an inducer of angiogenesis [bib_ref] Lactate influx through the endothelial cell monocarboxylate transporter MCT1 supports an NF-kappaB/IL-8..., Vegran [/bib_ref] [bib_ref] Glioblastoma cell-secreted interleukin-8 induces brain endothelial cell permeability via CXCR2, Dwyer [/bib_ref] [bib_ref] Interleukin-8 and its receptor CXCR2 in the tumour microenvironment promote colon cancer..., Lee [/bib_ref]. Moreover, a study described CXCL8 as a link between tumor metabolism and angiogenesis. Indeed, in a high-lactate-containing tumor microenvironment, tumor cells can release IL-8 that induces angiogenesis by interacting with endothelial cells [bib_ref] Lactate influx through the endothelial cell monocarboxylate transporter MCT1 supports an NF-kappaB/IL-8..., Vegran [/bib_ref].
## C-x3-c chemokines
For now, only one chemokine of this family has been described. This molecule is CX3CL1, also known as Fractalkine (FKN), and binds to CX3CR1. This chemokine regulates angiogenesis. Indeed, in vivo and ex vivo studies showed that FKN simulates angiogenesis and in vitro studies showed that this molecule increases proliferation, migration, and tube formation of human umbilical vein endothelial cells. This study showed that CX3CL1 stimulates angiogenesis through the activation of Raf-1/MEK/ERK and PI3K/Akt/eNos signaling pathways [bib_ref] Fractalkine stimulates angiogenesis by activating the Raf-1/MEK/ERK-and PI3K/Akt/eNOS-dependent signal pathways, Lee [/bib_ref].
## Non-classical pro-angiogenic factors
## Thymidine phosphorylase
Thymidine phosphorylase (TP) is an enzyme of the pyrimidine pathway discovered in 1984. This molecule catalyzes the conversion of thymidine to thymine and 2-deoxy-α-Dribose-1-phosphate. This enzyme is also named the platelet-derived endothelial cell growth factor (PD-ECGF). This molecule is overexpressed in cellular stress conditions such as hypoxia and is expressed by tumoral cells, fibroblast, tumor-associated macrophages, and lymphocytes. TP overexpression is associated with poor clinical outcomes in patients. TP is overexpressed in different cancers such as oral squamous carcinoma, esophageal, gastric, breast, lung, colorectal, bladder, and cervical cancer. This molecule plays an important role in tumor growth by promoting two mechanisms: angiogenesis and apoptosis inhibition. Indeed, TP is an endothelial chemoattractant that stimulates endothelial cell migration as well as angiogenesis factor releases in the tumor microenvironment [bib_ref] The role of thymidine phosphorylase, an angiogenic enzyme, in tumor progression, Akiyama [/bib_ref] [bib_ref] Targeting platelet-derived endothelial cell growth factor/thymidine phosphorylase for cancer therapy, Liekens [/bib_ref]. Therapy targeting TP is a promising strategy. First, this enzyme promotes angiogenesis and inhibits apoptosis. Second, it inactivates deoxynucleoside-based therapy and its inhibition may improve the bioavailability of these therapies [bib_ref] Targeting platelet-derived endothelial cell growth factor/thymidine phosphorylase for cancer therapy, Liekens [/bib_ref] [bib_ref] Recent discovery of non-nucleobase thymidine phosphorylase inhibitors targeting cancer, Bera [/bib_ref]. There are different ways to inhibit TP. The first inhibitor developed are pyrimidine and purine analogs such as 6-aminothymine, 6-amino-5-bromouracile, TPI, TAS-102 (TPI and TFT combination), and KIN59. There are also non-nucleobase-based therapies such as: oxadiazole and imidazolidine derivatives, Pyrazalone, and pyrazolo [1,5-a] [bib_ref] Growth and vascularization of transplanted mouse melanomas, Algire [/bib_ref] [bib_ref] Mechanical Forces in Tumor Angiogenesis, Zanotelli [/bib_ref] [bib_ref] Regulation of redox signaling in HIF-1-dependent tumor angiogenesis, Manuelli [/bib_ref] triazine analogs, Quinazoline and quinoxaline derivatives, Chromone and isocoumarin derivatives, and finally plant glycosides [bib_ref] Recent discovery of non-nucleobase thymidine phosphorylase inhibitors targeting cancer, Bera [/bib_ref].
## Tryptases and chymases
Tryptase and chymase are pro-angiogenic molecules released from mast cell granules. Tryptase is a tetrameric neutral serine protease while chymase is a monomeric endopeptidase. These two molecules promote directly or indirectly angiogenesis. Tryptase contributes to tube formation and endothelial cell growth by upregulating Ang-1 expression. This molecule induces endothelial cell proliferation, interleukin releases, and in vitro angiogenesis and activates matrix metalloproteinases such as MMP-9 and can convert angiotensin I into angiotensin II. It was also shown that tryptase enhances breast cancer angiogenesis through PAR-2 mediated endothelial progenitor cell activation [bib_ref] Tryptase, a novel angiogenic factor stored in mast cell granules, Ribatti [/bib_ref] [bib_ref] Mast Cell Tryptase Contributes to Pancreatic Cancer Growth through Promoting Angiogenesis via..., Guo [/bib_ref] [bib_ref] Tryptase promotes breast cancer angiogenesis through PAR-2 mediated endothelial progenitor cell activation, Qian [/bib_ref].
Three classes of tryptase inhibitors have been reported. The first class corresponds to molecules that can form a covalent bond with the catalytic serine in the active pocket of the tryptase. The second class corresponds to molecules containing a basic P1 group that are able to bind to the active pocket of tryptase. The last class of tryptase inhibitors contains molecules with a non-basic P1 group. Some tryptase inhibitors are under clinical trials [bib_ref] Tryptase inhibitors: A patent review, Ni [/bib_ref].
## Therapies targeting angiogenesis in cancer
## Therapies targeting the vegf family
Therapies targeting the VEGF signaling pathway are the most studied and used in cancer. There are three recombinant proteins approved for cancer treatment: Bevacizumab, Aflibercept, and Ramucirumab. Bevacizumab and Ramucirumab are two humanized monoclonal antibodies targeting, respectively, all VEGF-A isoforms and VEGFR-2. Aflibercept is a protein composed of two recognition domains, VEGFR-1 and VEGFR-2, fused to the Fc portion of a human IgG1. Aflibercept is able to bind to VEGF-A, VEGF-B, and PlGF. There are also tyrosine kinase inhibitors (TKI) approved for cancer: Sorafenib, Sunitinib, Regorafenib, etc. There are many reviews on the anti-VEGF-based therapies, so we decided to not go into much detail. Targeting the VEGF signaling pathway is a promising strategy but, due to many resistances, it appears to be ineffective when used as a single therapy. Indeed, there is a redundancy in the angiogenic signaling pathways, when the VEGF signaling pathway is blocked, other pathways take over to maintain angiogenesis. Therefore, to overcome this resistance, it is of interest to target several angiogenic factors simultaneously.
## Theraiesy targeting angiopoietin
Therapies targeting the Ang-Tie signaling pathway have recently emerged to treat cancer patients . They are monoclonal antibodies directed against Ang-2 such as MEDI3617, Nesvacumab (REGN910), and LY3127804. MEDI3617 significantly inhibits tumor growth in different xenograft tumor models such as colorectal cancer (LoVo and Colo205), renal cell carcinoma (786-0), ovarian carcinoma (HeyA8), and hepatocellular carcinoma (PLCPRF/5) [bib_ref] Anti-angiogenic alternatives to VEGF blockade, Khan [/bib_ref]. Phase I has been achieved to determine its safety in advanced solid tumors (NCT01248949) and another is still under investigation in patients with unresectable Stage III or Stage IV Melanoma (NCT02141542). REGN910 reduces tumor growth and tumor vascularity of different xenograft tumor models such as colorectal cancer (Colo205), prostate cancer (PC3), and epidermoid carcinoma (A431). It also has been shown that REGN910 potentiates the effects of Aflibercept [bib_ref] A Phase I First-in-Human Study of Nesvacumab (REGN910), a Fully Human Anti-Angiopoietin-2..., Papadopoulos [/bib_ref]. ABTAA protein is another strategy that not only neutralizes Ang-2 but also activates TIE-2 to enhance vascular normalization and by this, increase drug delivery. This molecule reduces tumor growth in a subcutaneous LLC tumor model [bib_ref] Normalization of Tumor Vessels by Tie2 Activation and Ang2 Inhibition Enhances Drug..., Park [/bib_ref]. There are also recombinant proteins targeting not only Ang-2 but also the interaction between Ang1/Ang2 with Tie-2. Trebananib is one of these molecules. This molecule is currently under investigation but in combination with paclitaxel, it shows an improved progression-free survival (NCT01204749) for recurrent ovarian cancer. A recent study showed that Bevacizumab plus Trebananib was tolerable and efficient in first-line treatment for patients with metastatic colorectal cancer [bib_ref] Dual Antiangiogenesis Agents Bevacizumab Plus Trebananib, without Chemotherapy, in First-line Treatment of..., Mooi [/bib_ref]. There is also the antibody-targeting VEGF-A and Ang-2, called Vanucizumab. However, a recent study demonstrated that the combination of Vanucizumab/mFOLFLOX-6 did not improve the PFS in comparison to Bevacizumab/mFOLFLOX-6 bitherapy in patients with metastatic colorectal carcinoma [bib_ref] The McCAVE Trial: Vanucizumab plus mFOLFOX-6 Versus Bevacizumab plus mFOLFOX-6 in Patients..., Bendell [/bib_ref].
## Therapies targeting hgf
Targeting the HGF/MET pathway is a promising strategy because it is involved in different cancer types. There are different ways to target this signaling pathway: HGF inhibitors, MET antagonists, MET kinase inhibitors, and HGF activation inhibitors. MET is expressed in several cell types, including epithelial, endothelial, neuronal, and hematopoietic cells and hepatocytes. The activation of the HGF/MET axis is associated with a series of biological responses, such as proliferation, angiogenesis, migration, invasion, metastasis, and survival. HGF/MET signaling is aberrantly activated in different solid tumors and associated with poor prognosis. HGF/MET aberrant activation plays important roles in the development and progression of several human cancers including lung, renal, gastrointestinal, thyroid, and breast carcinomas, as well as sarcomas and malignancies of the nervous system such as GBM among others. Rilotumumab is a fully-humanized monoclonal antibody targeting HGF. A pre-clinical study showed promising results of a combination of Rilotumumab with docetaxel or temozolomide where Rilotumumab decreases tumor growth in nude mice bearing U-87 MG tumor [bib_ref] AMG 102, a fully human anti-hepatocyte growth factor/scatter factor neutralizing antibody, enhances..., Jun [/bib_ref]. In clinical studies, this molecule showed a tolerable profile in patients with mRCC but no effect was identified (NCT00422019). Furthermore, in patients with advanced gastroesophageal adenocarcinoma, there is no benefit to combining Rilotumumab with mFOLFOX6 first-line chemotherapy [bib_ref] FOLFOX alone or combined with rilotumumab or panitumumab as first-line treatment for..., Malka [/bib_ref]. In patients with recurrent malignant glioma, Rilotumumab with Bevacizumab did not improve the response in comparison to Bevacizumab alone. For now, the FDA does not accept this molecule.
Onartuzumab is a fully-humanized monoclonal antibody targeting the extracellular domain of MET. In patients with metastatic triple-negative breast cancer, this molecule did not improve the clinical benefit of paclitaxel in bitherapy or not with bevacizumab (NCT01186991). Furthermore, it did not improve the efficiency of mFOLFOX6 in gastric cancer [bib_ref] A Randomized Phase II Study of FOLFOX With or Without the MET..., Shah [/bib_ref]. In patients with metastatic colorectal cancer, the combination of Onartuzumab with mFOLFOX-6 and bevacizumab did not improve the clinical benefit as well. There are two classes of MET tyrosine kinase inhibitors, class I and class II depending on the MET conformation binding [bib_ref] Targeting receptor tyrosine kinase MET in cancer: Small molecule inhibitors and clinical..., Cui [/bib_ref]. Crizotinib is a type I TKI approved by the European Medical Agency (EMA) and by the Food and Drug Administration (FDA), for patients with NSCLC in particular conditions [bib_ref] FDA approval summary: Crizotinib for the treatment of metastatic non-small cell lung..., Kazandjian [/bib_ref]. This molecule is also approved for patients with anaplastic large cell lymphoma in certain conditions. Cabozantinib is a type II TKI approved by the EMA for patients with medullary thyroid cancer in certain conditions. The FDA has also approved this molecule for patients with locally advanced or metastatic differentiated thyroid cancer.
## Therapies targeting fgf
There are drugs targeting the FGF/FGFR signaling pathway under investigation: monoclonal antibodies targeting FGFR and FGF, and tyrosine kinase inhibitors. Different monoclonal antibodies targeting FGF have shown interesting results. In vivo and in vitro studies showed that antibodies directed against FGF2 and FGF8b have anti-tumor and anti-angiogenic effects [bib_ref] Specific Antibody Fragment Ligand Traps Blocking FGF1 Activity, Chudzian [/bib_ref]. GAL-F2 is a monoclonal antibody targeting FGF2 and was shown to reduce tumor growth in different xenograft mice models of human HCC cell lines: SMMC-7721, HEP-G2, and SK-HEP-1 [bib_ref] A novel monoclonal antibody to fibroblast growth factor 2 effectively inhibits growth..., Wang [/bib_ref]. Different monoclonal antibodies target FGFR such as FPA144, PRO-001, RG7444, and SSR128129E. There are also antibody-drug conjugates targeting this pathway. The molecule BAY 1187982 is a monoclonal antibody directed against FGFR-IIb and FGFR-IIIc conjugated to a microtubule-disrupting auristatin. This molecule reduces tumor growth in different models such as breast, gastric, and ovarian cancer [bib_ref] First-in-Human Phase I Study of Aprutumab Ixadotin, a Fibroblast Growth Factor Receptor..., Kim [/bib_ref]. There are different tyrosine inhibitors targeting FGFR and also other receptors such as VEGFR, FGFR, PDGFR: AZD4547, BAY1163877, BGJ398, AXL1717, Cediranib, Dovotinib, etc.
## Therapies targeting pdgf
There are different therapies targeting the PDGF/PDGFR signaling pathway: inhibitors of PDGF, inhibitors of the interaction between PDGF and PDGFR, and TKI. There are also human monoclonal antibodies targeting PDGFRα. This molecule reduced tumor growth in a xenograft lung carcinoma tumor model Calu-6 and A549 [bib_ref] Targeting the platelet-derived growth factor receptor alpha with a neutralizing human monoclonal..., Loizos [/bib_ref]. It also reduces tumor growth in a xenograft glioblastoma (U118) and leiomyosarcoma (SKLMS-1) [bib_ref] Stromal platelet-derived growth factor receptor alpha (PDGFRalpha) provides a therapeutic target independent..., Gerber [/bib_ref]. In clinical phases Ib and II, the combination of doxorubicin with an antibody targeting PDGFRα increases the overall survival compared to doxorubicin used alone [bib_ref] Olaratumab and doxorubicin versus doxorubicin alone for treatment of soft-tissue sarcoma: An..., Tap [/bib_ref]. This molecule has been approved by the FDA in 2016 for the treatment of soft tissue sarcoma and is under conditional approval by the EMA [bib_ref] The PDGF/PDGFR pathway as a drug target, Papadopoulos [/bib_ref]. In the treatment of glioma, prostate cancer, and ovarian cancer, this molecule is not effective [bib_ref] The PDGF/PDGFR pathway as a drug target, Papadopoulos [/bib_ref]. Finally, there are different TKI targeting the PDGFR signaling pathway clinically approved such as Imatinib, Nilotinib, Dasatinib, Ponatinib, Sunitinib, Axitinib, Sorafenib, etc. [bib_ref] The PDGF/PDGFR pathway as a drug target, Papadopoulos [/bib_ref].
# Conclusions
This review aimed at highlighting the close relationship between angiogenesis and the tumor microenvironment, more specifically the cytokines and chemokines that can be found in tumors. By producing such molecules, cells from the immune system as well as stromal cells, tightly regulate angiogenesis within the tumor.
Inflammation, a key feature of tumorigenesis and a hallmark of cancer, is a strong pro-angiogenic signal. With cytokines such as the IL-1β, IL-6, and TNF all having proangiogenic properties, it is clear that inflammation and angiogenesis are related to cancer. Interestingly, cytokines produced by classical pro-tumor immune cells such as Tregs, which produce IL-10 and IL-35; Th17 cells, which produce IL-17 and IL-22; or Th2 cells with IL-4, are all pro-angiogenic factors. On the other side, known antitumor immune cells are linked to anti-angiogenic molecules such as IFNγ and IL-12 with NK cells, Th1, and cytotoxic CD8 T lymphocytes.
Chemokines serve to attract cells in a gradient-dependent manner and their impact on angiogenesis depends on what cells they attract but also on their direct effect on angiogenesis. Indeed, CCL2 will recruit macrophages to the tumor, and induce their production of VEGF as will CCL4. A direct talk has also been found between CXCL1 and tumor cells where CXCL1 induces the production of VEGF by tumor cells. However, some chemokines can also limit angiogenesis by acting directly on newly formed vessels such as CXCL10 or by promoting the expression of the antiangiogenic MiR206 such as CCL19. Chemokines exerting anti-angiogenic effects are usually associated with the recruitment of antitumor immune cells.
There are currently three recombinant proteins targeting the VEGF/VEGFR pathway approved for the treatment of cancer. However, numerous patients develop a resistance to these treatments due to the many redundant pathways leading to angiogenesis. Considerable effort has been made to develop new therapies targeting these redundant pathways with many still in development or under study in clinical trials. It also clearly appears that targeting angiogenesis alone is not sufficient to trigger a potent immune response. Association between anti-angiogenic treatment and chemotherapies or immunotherapies is starting to give promising results and it is likely that more associations of this sort will appear in the future.
[fig] Figure 2: Role of classical regulators of angiogenesis. [/fig]
[fig] Figure 3: Role of cytokines in angiogenesis. [/fig]
[fig] Author: Contributions: M.G., M.B., F.V.: writing-original draft preparation; M.B., F.V.: writingreview and editing. All authors have read and agreed to the published version of the manuscript. [/fig]
[fig] Funding: This work was supported by a French Government grant managed by the French National Research Agency (ANR) under the program "Investissementsd'Avenir" with reference ANR-11-LABX-0021-01-LipSTIC Labex and by the foundation ARC. 14. Yoo, S.A.; Kim, M.; Kang, M.C.; Kong, J.S.; Kim, K.M.; Lee, S.; Hong, B.K.; Jeong, G.H.; Lee, J.; Shin, M.G.; et al. Placental growth factor regulates the generation of TH17 cells to link angiogenesis with autoimmunity. Nat. Immunol. 2019, 20, 1348-1359. [CrossRef] [PubMed] 15. Valenzuela, D.M.; Griffiths, J.A.; Rojas, J.; Aldrich, T.H.; Jones, P.F.; Zhou, H.; McClain, J.; Copeland, N.G.; Gilbert, D.J.; Jenkins, N.A.; et al. Angiopoietins 3 and 4: Diverging gene counterparts in mice and humans. Proc. Natl. Acad. Sci. USA 1999, 96, 1904-1909. [CrossRef] 16. Davis, M.R.; Zhu, Z.; Hansen, D.M.; Bai, Q.; Fang, Y. The role of IL-21 in immunity and cancer. Cancer Lett. 2015, 358, 107-114. [CrossRef] [PubMed] 17. Maisonpierre, P.C.; Suri, C.; Jones, P.F.; Bartunkova, S.; Wiegand, S.J.; Radziejewski, C.; Compton, D.; McClain, J.; Aldrich, T.H.; Papadopoulos, N.; et al. Angiopoietin-2, a natural antagonist for Tie2 that disrupts in vivo angiogenesis. Science 1997, 277, 55-60. [CrossRef] 18. Hashiyama, M.; Iwama, A.; Ohshiro, K.; Kurozumi, K.; Yasunaga, K.; Shimizu, Y.; Masuho, Y.; Matsuda, I.; Yamaguchi, N.; Suda, T. Predominant expression of a receptor tyrosine kinase, TIE, in hematopoietic stem cells and B cells. Blood 1996, 87, 93-101. [CrossRef] 19. Batard, P.; Sansilvestri, P.; Scheinecker, C.; Knapp, W.; Debili, N.; Vainchenker, W.; Buhring, H.J.; Monier, M.N.; Kukk, E.; Partanen, J.; et al. The Tie receptor tyrosine kinase is expressed by human hematopoietic progenitor cells and by a subset of megakaryocytic cells. Blood 1996, 87, 2212-2220. [CrossRef] 20. Saharinen, P.; Kerkela, K.; Ekman, N.; Marron, M.; Brindle, N.; Lee, G.M.; Augustin, H.; Koh, G.Y.; Alitalo, K. Multiple angiopoietin recombinant proteins activate the Tie1 receptor tyrosine kinase and promote its interaction with Tie2. J. Cell Biol. 2005, 169, 239-243. [CrossRef] 21. Sato, T.N.; Tozawa, Y.; Deutsch, U.; Wolburg-Buchholz, K.; Fujiwara, Y.; Gendron-Maguire, M.; Gridley, T.; Wolburg, H.; Risau, W.; Qin, Y. Distinct roles of the receptor tyrosine kinases Tie-1 and Tie-2 in blood vessel formation. Nature 1995, 376, 70-74. [CrossRef] [PubMed] 22. Suri, C.; Jones, P.F.; Patan, S.; Bartunkova, S.; Maisonpierre, P.C.; Davis, S.; Sato, T.N.; Yancopoulos, G.D. Requisite role of angiopoietin-1, a ligand for the TIE2 receptor, during embryonic angiogenesis. Cell 1996, 87, 1171-1180. [CrossRef] 23. Gale, N.W.; Thurston, G.; Hackett, S.F.; Renard, R.; Wang, Q.; McClain, J.; Martin, C.; Witte, C.; Witte, M.H.; Jackson, D.; et al. Angiopoietin-2 is required for postnatal angiogenesis and lymphatic patterning, and only the latter role is rescued by Angiopoietin-1. Dev. Cell 2002, 3, 411-423. [CrossRef] 24. Benest, A.V.; Kruse, K.; Savant, S.; Thomas, M.; Laib, A.M.; Loos, E.K.; Fiedler, U.; Augustin, H.G. Angiopoietin-2 is critical for cytokine-induced vascular leakage. PLoS ONE 2013, 8, e70459. [CrossRef] [PubMed] 25. Fiedler, U.; Scharpfenecker, M.; Koidl, S.; Hegen, A.; Grunow, V.; Schmidt, J.M.; Kriz, W.; Thurston, G.; Augustin, H.G. The Tie-2 ligand angiopoietin-2 is stored in and rapidly released upon stimulation from endothelial cell Weibel-Palade bodies. Blood 2004, 103, 4150-4156. [CrossRef] 26. Huang, H.; Bhat, A.; Woodnutt, G.; Lappe, R. Targeting the ANGPT-TIE2 pathway in malignancy. Nat. Rev. Cancer 2010, 10, 575-585. [CrossRef] 27. Saharinen, P.; Eklund, L.; Alitalo, K. Therapeutic targeting of the angiopoietin-TIE pathway. Nat. Rev. Drug. Discov. 2017, 16, 635-661. [CrossRef] [PubMed] 28. Oh, H.; Takagi, H.; Suzuma, K.; Otani, A.; Matsumura, M.; Honda, Y. Hypoxia and vascular endothelial growth factor selectively up-regulate angiopoietin-2 in bovine microvascular endothelial cells. J. Biol. Chem. 1999, 274, 15732-15739. [CrossRef] 29. Kim, I.K.; Kim, B.S.; Koh, C.H.; Seok, J.W.; Park, J.S.; Shin, K.S.; Bae, E.A.; Lee, G.E.; Jeon, H.; Cho, J.; et al. Glucocorticoid-induced tumor necrosis factor receptor-related protein co-stimulation facilitates tumor regression by inducing IL-9-producing helper T cells. Nat. Med. 2015, 21, 1010-1017. [CrossRef] 30. Kim, M.; Allen, B.; Korhonen, E.A.; Nitschke, M.; Yang, H.W.; Baluk, P.; Saharinen, P.; Alitalo, K.; Daly, C.; Thurston, G.; et al. Opposing actions of angiopoietin-2 on Tie2 signaling and FOXO1 activation. J. Clin. Investig. 2016, 126, 3511-3525. [CrossRef] 31. Korhonen, E.A.; Lampinen, A.; Giri, H.; Anisimov, A.; Kim, M.; Allen, B.; Fang, S.; D'Amico, G.; Sipila, T.J.; Lohela, M.; et al. Tie1 controls angiopoietin function in vascular remodeling and inflammation. J. Clin. Investig. 2016, 126, 3495-3510. [CrossRef] [PubMed] 32. Sfiligoi, C.; de Luca, A.; Cascone, I.; Sorbello, V.; Fuso, L.; Ponzone, R.; Biglia, N.; Audero, E.; Arisio, R.; Bussolino, F.; et al. Angiopoietin-2 expression in breast cancer correlates with lymph node invasion and short survival. Int. J. Cancer 2003, 103, 466-474. [CrossRef] [PubMed] 33. Helfrich, I.; Edler, L.; Sucker, A.; Thomas, M.; Christian, S.; Schadendorf, D.; Augustin, H.G. Angiopoietin-2 levels are associated with disease progression in metastatic malignant melanoma. Clin. Cancer Res. Off. J. Am. Assoc. Cancer Res. 2009, 15, 1384-1392. [CrossRef] [PubMed] 34. Goede, V.; Coutelle, O.; Neuneier, J.; Reinacher-Schick, A.; Schnell, R.; Koslowsky, T.C.; Weihrauch, M.R.; Cremer, B.; Kashkar, H.; Odenthal, M.; et al. Identification of serum angiopoietin-2 as a biomarker for clinical outcome of colorectal cancer patients treated with bevacizumab-containing therapy. Br. J. Cancer 2010, 103, 1407-1414. [CrossRef] 35. Wang, X.; Bullock, A.J.; Zhang, L.; Wei, L.; Yu, D.; Mahagaokar, K.; Alsop, D.C.; Mier, J.W.; Atkins, M.B.; Coxon, A.; et al. The role of angiopoietins as potential therapeutic targets in renal cell carcinoma. Transl. Oncol. 2014, 7, 188-195. [CrossRef] [/fig]
|
eIF3d is an mRNA cap-binding protein required for specialized translation initiation
Eukaryotic mRNAs contain a 5′ cap structure that is crucial for recruitment of the translation machinery and initiation of protein synthesis. mRNA recognition is thought to require direct interactions between eukaryotic initiation factor 4E (eIF4E) and the mRNA cap. However, translation of numerous capped mRNAs remains robust during cellular stress, early development, and cell cycle progression 1 despite inactivation of eIF4E. Here we describe a cap-dependent pathway of translation initiation in human cells that relies on a previously unknown cap-binding activity of eIF3d, a subunit of the 800-kilodalton eIF3 complex. A 1.4 Å crystal structure of the eIF3d cap-binding domain reveals unexpected homology to endonucleases involved in RNA turnover, and allows modelling of cap recognition by eIF3d. eIF3d makes specific contacts with the cap, as exemplified by cap analogue competition, and these interactions are essential for assembly of translation initiation complexes on eIF3-specialized mRNAs 2 such as the cell proliferation regulator c-Jun (also known as JUN). The c-Jun mRNA further encodes an inhibitory RNA element that blocks eIF4E recruitment, thus enforcing alternative cap recognition by eIF3d. Our results reveal a mechanism of cap-dependent translation that is independent of eIF4E, and illustrate how modular RNA elements work together to direct specialized forms of translation initiation.The rate-limiting step of translation initiation is the recognition of the 5′ cap structure by eIF4E 3,4 . eIF4E activity is highly regulated by extracellular stimuli, predominantly through steric hindrance by eIF4E-binding proteins (4E-BPs) 5,6 . The translational efficiencies of mRNAs range in sensitivity to 4E-BP inhibition 7-9 , and these differences have conventionally been addressed by categorizing translation into cap-dependent versus cap-independent pathways 10 | 5′ end recognition of c-Jun mRNA is eIF4F-independent. a, Distribution of c-Jun or ACTB mRNA-containing initiation complexes in programmed 293T cell in vitro translation extracts. The mRNA abundance (black line) is expressed as the fraction of total recovered transcripts. The results are given as the mean ± s.d. of a representative quantitative RT-PCR experiment performed in duplicate. The polysome profile (grey line) is plotted as relative absorbance at 254 nm versus elution fractions. b, Western blot analysis of initiation factors in 48S translation complexes formed on c-Jun and ACTB mRNAs. 293T, total protein from 293T in vitro translation extracts. rpS19, ribosomal protein S19. For gel source data, see . c, Phosphorimage of SDS-PAGE gel resolving RNase-protected 32 P-internal or 32 P-cap-labelled c-Jun 5′ UTR RNA crosslinked to eIF3 subunits. Recombinant eIF3a migrates at ~100 kDa owing to a C-terminal truncation [bib_ref] Functional reconstitution of human eukaryotic translation initiation factor 3 (eIF3), Sun [/bib_ref]. The results of a-c are representative of three independent experiments. LETTER RESEARCH might also be involved in 5′ cap recognition. In agreement with the previously demonstrated RNA-binding capability of eIF3, the four eIF3 RNA-binding subunits, eIF3a, eIF3b, eIF3d and eIF3g, provide RNase protection to internally 32 P-labelled c-Jun 5′ UTR RNA after UV 254 -induced crosslinking 2 . By contrast, when the 32 P label is placed in the 5′ cap of c-Jun mRNA, RNase protection is observed with a single subunit of eIF3, corresponding to eIF3d , Extended Data [fig_ref] Figure 2 |: Structure of eIF3d reveals a conserved cap-binding domain [/fig_ref]. We confirmed subunit identity by limited proteolysis and mass spectrometry, and defined a C-terminal region of eIF3d that is responsible for protection of the 5′ mRNA terminus (Extended Data [fig_ref] Figure 2 |: Structure of eIF3d reveals a conserved cap-binding domain [/fig_ref]. The mapped C-terminal region of eIF3d is broadly conserved throughout plant, fungal and animal phylogeny [fig_ref] Figure 2 |: Structure of eIF3d reveals a conserved cap-binding domain [/fig_ref] , Extended Data [fig_ref] Figure 3 |: eIF3d cap-binding activity is required for efficient 48S initiation complex formation on... [/fig_ref] , suggesting the apparent 5′ end recognition activity of eIF3d is an evolutionarily preserved function of the eIF3 complex.
To understand how eIF3d recognizes the 5′ RNA terminus, we determined a 1.4 Å crystal structure of the conserved C-terminal domain of eIF3d from Nasonia vitripennis (65% identical, 84% similar to human eIF3d) using sulfur anomalous dispersion for phase determination (Extended Data , Extended Data [fig_ref] Figure 3 |: eIF3d cap-binding activity is required for efficient 48S initiation complex formation on... [/fig_ref]. The structure of eIF3d reveals a complex fold that forms a cup-shaped architecture with a positively charged central tunnel that is negatively charged at its base [fig_ref] Figure 2 |: Structure of eIF3d reveals a conserved cap-binding domain [/fig_ref]. Remarkably, despite no significant sequence homology, the structural topology of eIF3d is nearly identical to the DXO proteins, a recently described family of 5′ cap-endonucleases involved in RNA quality control [bib_ref] Dxo1 is a new type of eukaryotic enzyme with both decapping and..., Chang [/bib_ref] [bib_ref] A mammalian pre-mRNA 5′ end capping quality control mechanism and an unexpected..., Jiao [/bib_ref] [bib_ref] Structure and function of the 5′→3′ exoribonuclease Rat1 and its activating partner..., Xiang [/bib_ref] [fig_ref] Figure 2 |: Structure of eIF3d reveals a conserved cap-binding domain [/fig_ref]. In contrast to DXO, eIF3d contains a unique insertion of ~15 highly conserved amino acids between strand β5 and helix α6. The eIF3d-specific insertion folds down along the front face of the domain, making loosely packed charged interactions that close off the RNA binding tunnel (Extended . We term this insertion an 'RNA gate' , as the sequence clashes with the path of single-stranded RNA (ssRNA) bound to DXO [bib_ref] A mammalian pre-mRNA 5′ end capping quality control mechanism and an unexpected..., Jiao [/bib_ref] and must undergo a conformational change for eIF3d to become competent for RNA recognition [fig_ref] Figure 2 |: Structure of eIF3d reveals a conserved cap-binding domain [/fig_ref]. We determined the structure of eIF3d in two additional crystal forms, and confirmed the RNA gate exhibits a closed conformation regardless of crystal packing (Extended Data . As eIF3d does not bind all capped RNAs 17,18 , we postulate that the RNA gate regulates cap recognition to prevent promiscuous mRNA binding before assembly of eIF3d into the full eIF3 complex. We tested this model using c-Jun mRNA, and verified that eIF3d cap-recognition only occurs in the context of a full eIF3 complex and requires previous eIF3-sequence-specific RNA interactions with the eIF3-recruitment stem-loop (Extended Data . Allosteric communication between eIF3 subunits during initial RNA recruitment likely facilitates eIF3d RNA gate opening to allow 5′ end recognition. The structure of eIF3d therefore reveals a new cap-binding protein and explains the ability of the eIF3 complex to protect the 5′ end of mRNA .
To validate the structural finding that eIF3d is a cap-binding protein, we examined the ability of eIF3 to bind the c-Jun mRNA 5′ cap in the presence of competitor ligands. eIF3d cap recognition is sensitive to m 7 GDP competition but resistant to GDP, indicating that, analogous to eIF4E 4 , eIF3d specifically interacts with the 5′ cap and requires a mature methylated cap structure for recognition [fig_ref] Figure 3 |: eIF3d cap-binding activity is required for efficient 48S initiation complex formation on... [/fig_ref]. Using the DXO-RNA structure as a template [bib_ref] A mammalian pre-mRNA 5′ end capping quality control mechanism and an unexpected..., Jiao [/bib_ref] , we modelled a capped ssRNA along the basic binding groove shared between eIF3d and DXO and identified two conserved helices (α5 and α11) likely to be involved in cap recognition [fig_ref] Figure 3 |: eIF3d cap-binding activity is required for efficient 48S initiation complex formation on... [/fig_ref]. We purified recombinant eIF3 containing helix α5or α11-mutated eIF3d and demonstrated that both mutants have markedly reduced ability to crosslink to the c-Jun mRNA cap [fig_ref] Figure 3 |: eIF3d cap-binding activity is required for efficient 48S initiation complex formation on... [/fig_ref]. eIF3d-mutated complexes retain wild-type-levels of RNAbinding, indicating that these residues specifically coordinate 5′ mRNA cap recognition (Extended Data . We next introduced haemagglutinin (HA) epitope-tagged wild-type or mutant eIF3d into 293T cells, and measured the assembly of 48S initiation complexes on c-Jun [bib_ref] Affinity purification of eukaryotic 48S initiation complexes, Locker [/bib_ref] [bib_ref] Individual overexpression of five subunits of human translation initiation factor eIF3 promotes..., Zhang [/bib_ref]. Mutations to the predicted eIF3d cap-binding surface inhibit c-Jun mRNA incorporation into translation complexes, while the control ACTB mRNA is unaffected [fig_ref] Figure 3 |: eIF3d cap-binding activity is required for efficient 48S initiation complex formation on... [/fig_ref] , Extended Data . These results demonstrate that cap binding by eIF3d is required for efficient initiation complex formation during eIF3-specialized translation. eIF3d recognition of the 5′ cap structure provides an alternative cap-dependent translation mechanism from canonical eIF4F cap recognition. Perplexingly, when the RNA stem-loop element that recruits eIF3 to the c-Jun mRNA is deleted, translation is inhibited even though the mRNA contains a 5′ cap 2 . We proposed that an RNA element within the c-Jun mRNA blocks recruitment of the eIF4F complex. In support, the 5′ cap of c-Jun mRNA crosslinks less efficiently to purified eIF4E than that of the ACTB mRNA (Extended Data . To identify the eIF4F inhibitory element, we constructed luciferase reporters to test deletions in the c-Jun 5′ UTR [fig_ref] Figure 4 |: An RNA element inhibits eIF4F recruitment and directs mRNAs to use an... [/fig_ref]. Deletion of the 5′ 153 nucleotides, but not the initial 67 nucleotides, was sufficient to allow c-Jun mRNA translation to occur independently of the eIF3-recruitment stem-loop, suggesting that canonical cap dependent translation is no longer blocked [fig_ref] Figure 4 |: An RNA element inhibits eIF4F recruitment and directs mRNAs to use an... [/fig_ref]. We confirmed by western blot analysis of the 48S initiation complex formed on c-Jun mRNA lacking the 5′ 153 nucleotides that the eIF4F components are now present [fig_ref] Figure 4 |: An RNA element inhibits eIF4F recruitment and directs mRNAs to use an... [/fig_ref].
Together, we put forth a model of a previously undiscovered capdependent translation initiation pathway controlled by eIF3d recognition of the 5′ mRNA cap [fig_ref] Figure 4 |: An RNA element inhibits eIF4F recruitment and directs mRNAs to use an... [/fig_ref]. We postulate that encoding more than one mechanism of cap-dependent translation allows cells to control protein synthesis specifically in cellular environments in which eIF4E is inactivated. In support, c-Jun mRNA translation is resistant to treatment of cells with chemicals that activate the 4E-BPs 7,8,21 (Extended Data . As modulation of eIF4E cap-binding activity allows cells to incorporate extracellular stimuli into altered translation outputs [bib_ref] Upstream and downstream of mTOR, Hay [/bib_ref] , it will be important to discover whether eIF3d activity is analogously regulated. Furthermore, our data indicates that the c-Jun mRNA encodes an additional cis-acting RNA element that blocks eIF4F to ensure translation can only occur through an eIF3-specialized pathway. RNA elements that block eIF4F recruitment may be a common theme to direct mRNAs into specific translation pathways to ensure controlled protein expression. For example, a subset of homeobox mRNAs contain an RNA element that blocks cap-dependent translation to ensure usage of an internal ribosome entry site and to allow for correct homeobox expression during embryonic development [bib_ref] RNA regulons in Hox 5′ UTRs confer ribosome specificity to gene regulation, Xue [/bib_ref]. Several eIF3-specialized mRNAs encode proteins involved in the control of cell proliferation, suggesting that their translation may also require enhanced regulation [bib_ref] eIF3 targets cell-proliferation messenger RNAs for translational activation or repression, Lee [/bib_ref] [bib_ref] Jun regulates cell cycle progression and apoptosis by distinct mechanisms, Wisdom [/bib_ref].
While considerable advances have been made in the structural understanding of eIF3 bound to the ribosome, direct localization of eIF3d in a 48S complex remains unclear [bib_ref] Structure of mammalian eIF3 in the context of the 43S preinitiation complex, Des Georges [/bib_ref]. Thus, understanding how eIF3d functions and assembles within the full translation initiation complex will have important mechanistic implications in how cap recognition links to mRNA ribosomal recruitment. Our discovery of [bib_ref] A mammalian pre-mRNA 5′ end capping quality control mechanism and an unexpected..., Jiao [/bib_ref]. Positive charge is coloured blue, negative charge is in red, and the RNA gate is removed for clarity. c, Phosphorimage of SDS-PAGE gel resolving RNase-protected 32 P-cap-labelled c-Jun 5′ UTR RNA crosslinked to wild-type (WT) or helix α5or helix α11-mutant eIF3. Helix α5-mutant eIF3d: D249Q/V262I/ Y263A; helix α11-mutant eIF3d: T317E/N320E/H321A. d, Incorporation of c-Jun and ACTB mRNA into initiation complexes by wild-type, helix α5-, or helix α11-mutant eIF3d as measured by quantitative RT-PCR. The mRNA-ribosome association is expressed as the ratio of the quantity of mRNA transcripts to 18S rRNA and normalized to the wild-type sample. The results are representative of three independent experiments and given as the mean ± s.d. from a representative quantitative RT-PCR experiment performed in duplicate.
## Letter research
# Methods
## Cells and transfections.
Human 293T cells were maintained in DMEM (Invitrogen) supplemented with 10% FBS (Seradigm). The cells were obtained from the University of California, Berkeley, Cell Culture Facility, which authenticates cells by STR profiling and tests for mycoplasma contamination. Plasmid transfections were performed using Lipofectamine 2000 (Invitrogen), following the manufacturer's protocol, and polysome or immunoprecipitation analyses were performed at 48 h after transfection. For INK128 (Cayman Chemical) cell treatment, 293T cells were incubated with the indicated concentration of INK128 for ~14-16 h before cell lysis. Plasmids. To generate the eIF3d expression plasmids, eIF3d was amplified from human cDNA and inserted into pcDNA5/FRT. A 39-nucleotide linker followed by the HA epitope tag (YPYDVPDYA) was subsequently inserted before the eIF3d stop codon. The wild-type c-Jun 5′ UTR luciferase reporter plasmid was previously described 2 . To generate the c-Jun in vitro transcription template, the 5′ UTR, ORF and 3′ UTR were separately amplified from human cDNA and stitched together downstream of a T7 promoter by Gibson cloning into pcDNA4. The ACTB in vitro transcription template was constructed by addition of a T7 promoter during amplification of the full mRNA from human cDNA and inserted into pcDNA4. [bib_ref] eIF3 targets cell-proliferation messenger RNAs for translational activation or repression, Lee [/bib_ref] and KOAc. The optimal magnesium and potassium levels to add were determined to be 1.5 mM Mg(OAc) 2 and 150 mM KOAc for c-Jun mRNA, and 1 mM and 150 mM KOAc for ACTB mRNA. For luciferase assays, translation reactions were incubated for 1 h at 30 °C, then luciferase activity was assayed. 48S initiation complex purification. For 48S initiation complex purification from in vitro translation reactions, 180 μl reactions were incubated in the presence of GMP-PNP for 20 min at 30 °C and centrifuged for 6 min at 12,000g at 4 °C. Lysates were purified by size-exclusion chromatography through a 1 ml column packed with Sephacryl S-400 gel filtration resin (GE Healthcare) and the elutant was centrifuged through a 10-25% (w/v) sucrose gradient by centrifugation for 3.5 h at 38,000 r.p.m. at 4 °C in a Beckman SW41 Ti rotor [bib_ref] A ribosome-specialized translation initiation pathway is required for cap-dependent translation of vesicular..., Lee [/bib_ref]. Fractions were collected from the top of the gradient using a peristaltic pump with a Brandel tube piercer. From the appropriate fractions, RNA was purified by phenol-chloroform extraction and ethanol precipitation and protein was precipitated with trichloroacetic acid.
For affinity purification of HA epitope-tagged eIF3d-associated 48S initiation complexes from cells, three 10 cm plates of transfected 293T cells were treated with 100 μg ml −1 cycloheximide for 5 min. Cells were washed with ice-cold PBS (137 mM NaCl, 2.7 mM KCl, 100 mM Na 2 HPO 4 , 2 mM KH 2 PO 4 ) with 100 μg ml −1 cycloheximide and collected in lysis buffer (20 mM HEPES-KOH pH 7.4, 150 mM KOAc, 2.5 mM Mg(OAc) 2 , 1 mM DTT, 100 μg ml −1 cycloheximide, 1% (v/v) Triton X-100). Lysates were centrifuged for 6 min at 12,000g at 4 °C and purified by S-400 size-exclusion chromatography. 80 μl of anti-HA antibody-conjugated agarose beads (Sigma) was added to the elutants, tumbled for 1.5 h at 4 °C, and beads were washed three times with lysis buffer without Triton X-100. Bound complexes were eluted twice with 100 μg ml −1 HA peptide and elutants were centrifuged and analysed the same as for the in vitro purification reactions. Quantitative real-time PCR. cDNA was reverse-transcribed from RNA using random hexamers and Superscript III (Invitrogen), following the manufacturer's protocol. Real-time PCR was performed using DyNAmo HS Sybr Green (ThermoFisher), with a 20 μl reaction volume containing 2 μl cDNA and 0.5 μM of each primer. The following oligonucleotides were used: 18S rRNA forward, 5′-GGCCCTGTAATTGGAATGAGTC-3′, 18S rRNA reverse, 5′-CCAAGA TCCAACTACGAGCTT-3′; ACTB forward, 5′-CTCTTCCAGCCTTCCTTCCT-3′, ACTB-reverse, 5′-AGCACTGTGTTGGCGTACAG-3′; c-Jun forward, 5′-TGACTGCAAAGATGGAAACG-3′, c-Jun reverse; 5′-CAGGGTCATGCT CTGTTTCA-3′. eIF3-RNA crosslinking and gel shift. Recombinant eIF3 was expressed and purified from Escherichia coli as previously described [bib_ref] Functional reconstitution of human eukaryotic translation initiation factor 3 (eIF3), Sun [/bib_ref]. For each crosslinking reaction, 1 μl water, 1 μl 125 nM labelled RNA, 1 μl 10 μg ml −1 heparan sulfate (Sigma), 1 μl 5× binding buffer (125 mM Tris-HCl, pH 7.5, 25 mM Mg(OAc) 2 , 350 mM KCl, 0.5 mM CaCl 2 , 0.5 mg ml −1 BSA, 10 mM TCEP), and 1 μl 1.5 μM purified eIF3 were added, in the listed order, and incubated for 30 min at 25 °C. For competition experiments, the water was substituted with 1 μl of 1 mM m 7 GDP/Mg 2+ or GDP/ Mg 2+ . UV 254 -induced crosslinking was performed using a short-wave UV lamp placed ~4 cm above the samples on ice for 10 min. After treatment with RNase for 10 min at 37 °C, proteins were separated by 12% SDS-PAGE, the gel was dried, and imaged using a phosphorimager. For digestion of internal labelled RNA, 2.5 U benzonase (Novagen) and 250 U RNase T1 (ThermoFisher) were used; for digestion of cap labelled RNA, 4 U RNase R (Epicentre) and 1 U RiboShredder (Epicentre) were used. For eIF3d subunit identification, after RNase treatment, samples were denatured and immunoprecipitation was performed as previously described 2 . For limited proteolysis, after RNase treatment, the reactions were treated with 2 or 20 μg ml −1 sequencing grade trypsin (Promega) for 30 min at 25 °C, before gel electrophoresis. Mass spectrometry samples were prepared as previously described 2 . Gel-shift assays were performed as previously described, using 50 nM labelled c-Jun stem-loop RNA and 300 nM purified eIF3 (ref. 2). Recombinant eIF3d protein purification. Candidate eIF3d cap-binding domain fragments were amplified by PCR and cloned into a modified pET vector to express an N-terminal 6× His (KSSHHHHHHGSS)-MBP-TEV fusion protein as previously described [bib_ref] Structure-guided reprogramming of human cGAS dinucleotide linkage specificity, Kranzusch [/bib_ref]. Extensive expression trials were conducted to determine optimal N-and C-terminal domain boundaries and identified a minimal stable human cap-binding domain S161-F527. Recombinant protein was expressed in BL21-RIL DE3 E. coli cells co-transformed with a pRARE2 tRNA plasmid (Agilent). E. coli was grown in 2× YT media at 37 °C to an OD 600 of ~0.5, cooled at 4 °C for 15 min, induced with addition of 0.5 mM IPTG and then incubated with shaking for ~20 h at 16 °C. Pelleted cells were washed with PBS and then lysed by sonication in lysis buffer (20 mM HEPES-KOH pH 7.5, 400 mM NaCl, 10% glycerol, 30 mM imidazole, 1 mM TCEP) in the presence of EDTA-free Complete Protease Inhibitor (Roche). Following centrifugation for 30 min at 23,000g and 4 °C, clarified lysate was incubated with Ni-NTA agarose resin (QIAGEN) for 1 h at 4 °C with gentle rocking. Resin was washed with lysis buffer supplemented to 1 M NaCl and eluted by gravity-flow chromatography at 4 °C with lysis buffer supplemented to 300 mM imidazole. The eluted fraction was diluted to ~50 mM imidazole and 5% glycerol, concentrated to ~50 mg ml −1 and incubated with Tobacco Etch Virus protease for ~12 h at 4 °C to remove the MBP tag. Recombinant eIF3d was isolated from free MBP by diluting with gel-filtration buffer (20 mM HEPES-KOH pH 7.5, 250 mM KCl, 1 mM TCEP) and passing over a 5 ml Ni-NTA column (QIAGEN) connected in line with a 5 ml MBP-Trap column (GE Life Sciences) before additional purification by size-exclusion chromatography on a Superdex 75 16/60 column. Final purified eIF3d was concentrated to ~20-50 mg ml −1 , used immediately for crystallography, or flash frozen in liquid nitrogen for storage at −80 °C. Crystallization and structure determination. Initial crystals of human eIF3d were grown at 18 °C by hanging drop vapour diffusion, but diffracted poorly. Analogous eIF3d cap-binding domain sequences were cloned from a panel of highly homologous animal sequences, with the equivalent domain from the parasitic wasp N. vitripennis (S172-F537) producing the best crystals. Optimized N. vitripennis eIF3d crystals were grown in 2 μl hanging drops set at a 1:1 ratio over 300 μl of reservoir liquid: 200 mM (NH 4 ) 2 SO 4 , 100 mM Bis-Tris 6.5, 23-27% PEG-3350 (crystal form 1), 1.6-1.8 M ammonium citrate, pH 7.0 (crystal form 2), or 200 mM NaCl, 100 mM Tris 8.5, 25% PEG-3350 (crystal form 3). eIF3d crystals (crystal forms 1 and 2) were cryoprotected by covering the drop with a layer of saturated paratone-N or NVH oil (Hampton) and crystals were transferred into the oil emersion and cleaned using a Kozak cat whisker as previously described [bib_ref] Ancient origin of cGAS-STING reveals mechanism of universal 2′,3′ cGAMP signaling, Kranzusch [/bib_ref] , or cryoprotected by transferring to a reservoir solution supplemented with 20% ethylene glycol (crystal form 3). Crystals were harvested with a nylon loop and then flash-frozen in liquid nitrogen. X-ray diffraction data were collected under cryogenic conditions at the Lawrence Berkeley National Laboratory Advanced Light Source (beamline 8.3.1).
Data were processed with XDS and AIMLESS 30 using the SSRL autoxds script (A. Gonzalez, Stanford SSRL). eIF3d crystals belonged to the orthorhombic space group P2 1 2 1 2 1 , and contained either two copies per asymmetric unit (crystal form 1) or one copy (crystal form 2), or the space group P2 1 and contained two copies per asymmetric unit (crystal form 3). Experimental phase information was collected from a native crystal using sulfur single-wavelength anomalous dispersion. Data were collected at a minimal accessible wavelength (~7,235 eV) and iterative data
[fig] Figure 2 |: Structure of eIF3d reveals a conserved cap-binding domain. a, Cartoon schematic and phylogenetic conservation of eIF3d amino acid sequence according to physiochemical property similarity. Peptides in the cap-binding domain as identified by limited proteolysis are mapped below. b, Structure of the eIF3d cap-binding domain. α-helices are coloured in blue and β-strands in magenta. c, Topological maps of the eIF3d capbinding domain and the DXO cap-endonuclease domain 15 . d, Structures comparing the eIF3d cap-binding domain with its gate insertion to DXO bound to RNA 15 (PDB 4J7L). quantitative RT-PCR [/fig]
[fig] Figure 3 |: eIF3d cap-binding activity is required for efficient 48S initiation complex formation on specific mRNAs. a, Phosphorimage of SDS-PAGE gel resolving RNase-protected 32 P-cap-labelled c-Jun 5′ UTR RNA crosslinked to eIF3 in the presence of competitor ligands (m 7 GDP, GDP). b, Electrostatic surface view of the eIF3d cap-binding domain coloured by charge, with a zoomed view of ssRNA and cap analogue modelled according to their positions bound to DXO [/fig]
[fig] Figure 4 |: An RNA element inhibits eIF4F recruitment and directs mRNAs to use an eIF3-specialized translation pathway. a, Schematic of c-Jun 5′ UTR truncation-luciferase (Luc) reporter mRNAs. SL, stemloop. b, Luciferase activity from in vitro translation of mRNAs containing truncations of the c-Jun 5′ UTR, with or without the internal eIF3recruitment stem-loop sequence. The results are given as the mean ± s.d. of three independent experiments, each performed in triplicate. c, Western blot analysis of initiation factors in 48S translation initiation complexes formed on c-Jun mRNA with a 5′ 153-nucleotide truncation. 293T, total protein from 293T in vitro translation extracts. The result is representative of three independent experiments. For gel source data, see Supplementary Fig. 1. d, Model for eIF3d-directed cap-dependent mRNA translation. An eIF4F-inhibitory RNA element ensures that mRNA translation occurs through an eIF3-specialized pathway. [/fig]
|
Comparing the Effects of Oral HIV Self-Testing With Those of Standard HIV Testing for Men Who Have Sex With Men (MSM): A Systematic Review and Meta-Analysis
# Introduction and background
Regular testing is recognized as a key strategy for HIV control. Awareness of one's HIV-positive status results in the reduction of risky sexual practices, early linkage to care resulting in the early initiation of antiretroviral therapy, and a substantial reduction in the risk of HIV transmission to sexual partners [bib_ref] Barriers and facilitators to past six-month HIV testing among men who have..., Chandler [/bib_ref]. HIV testing services (HTSs) have scaled up significantly in the past decade [bib_ref] Comparing the effects of HIV self-testing to standard HIV testing for key..., Witzel [/bib_ref]. Despite the remarkable progress in the global HIV response, new HIV infections and AIDS-related deaths remain unacceptably high. Only eight countries have achieved the 90-90-90 testing and treatment targets. Globally, in 2020, 1.5 million were newly infected with HIV, 84% (31.6 million) of people living with HIV knew their HIV status, 73% (27.4 million) were accessing treatment, and 66% (24.8 million) were virally suppressed [bib_ref] Usage and acceptability of HIV self-testing in men who have sex with..., Wong [/bib_ref]. [bib_ref] Barriers and facilitators to past six-month HIV testing among men who have..., Chandler [/bib_ref] 1 1 2 Focus now shifts to those who have been missed for various reasons such as existing inequalities, barriers to HIV care and testing services, unjust legal laws, stigma, and more shift to virtual platforms during the COVID-19 pandemic. Noticeable among these are men who have sex with men (MSM) who are largely hidden because of the pervasive stigmatization, discrimination, and criminalization of homosexuality. The fear of being outcasted and ridiculed by family members, relatives, and society prevents them from opening up about their sexual identity and links to the prevention, care, support, and treatment services [bib_ref] Barriers and facilitators to past six-month HIV testing among men who have..., Chandler [/bib_ref]. Enough evidence is available that suggests that psychosocial barriers such as social stigma and discrimination, logistic barriers such as long waiting times, and personal barriers such as the low self-perceived risk of HIV acquisition and inconvenience due to fear of pain in blood tests prevent many MSM from getting tested for HIV at standard of care (SoC) health settings [bib_ref] HIV testing in men who have sex with men: a follow-up review..., Lui [/bib_ref].
New technologies and service models in HIV testing are required to reach the unreached and ensure optimal testing rates. HIV self-testing (HIVST) has the potential to increase accessibility to and uptake of HIV testing, particularly among populations that are unreachable by conventional health services. HIVST is a process whereby a person can collect his/her own sample (oral fluid or blood), conduct the test, and interpret the results, alone or in the company of a trustworthy individual. There are four WHO prequalified HIVST products, of which three (Mylan HIVST, INSTI HIVST, and Sure HIVST) are blood-based and have a sensitivity ranging from 97% to 99.8% and specificity ranging from 99.5% to 100%. The OraQuick HIV selftest is an oral fluid-based HIV testing product that has a sensitivity of 100% and specificity of 99.2%. The WHO in 2016 recommended HIVST as a strategy to expand HIV testing services (HTSs), particularly to highrisk and underserved populations, and came up with an updated policy brief in 2019 where it recommends the distribution of HIVST kits by HIV-positive and HIV-negative clients to their partners and contacts. Policy regarding HIVST and the legal framework to include it in national HTS efforts vary according to country. High-income countries, such as the USA and UK, introduced over-the-counter (OTC) sales in early 2012 and 2014, respectively. France and Ireland subsequently implemented their HIVST policies. Among African countries, Kenya was the first to develop its HTS national policy to include the oral HIVST kit in 2008. In SA, the guidelines were developed in 2018 [bib_ref] HIV self-screening distribution preferences and experiences among men who have sex with..., Radebe [/bib_ref]. As of June 2019, 38 countries only had implemented HIVST of the 77 countries, which had supportive policies for HIVST [bib_ref] Comparing the effects of HIV self-testing to standard HIV testing for key..., Witzel [/bib_ref].
HIVST innovations offer an important opportunity to reduce stigma and confidentiality concerns among hard-to-reach populations. The evidence across the globe suggests that HIVST offers confidentiality to the users and is thus a fear-free and stigma-free HIV testing. Reviews conducted reported that HIVST may increase the uptake of HIV testing MSM [bib_ref] Comparing the effects of HIV self-testing to standard HIV testing for key..., Witzel [/bib_ref] [bib_ref] Examining the effects of HIV self-testing compared to standard HIV testing services:..., Johnson [/bib_ref] [bib_ref] Can self-testing increase HIV testing among men who have sex with men:..., Zhang [/bib_ref]. The reviews included study designs from observational to RCTs.
In India, HIVST is not yet available, and no policy or guideline exists for the use of HIVST. A recent study that conducted mapping and size estimation of MSM in virtual platforms in New Delhi, India, cited that 47% of MSM in India have never been tested for HIV [bib_ref] Mapping and size estimation of men who have sex with men in..., Isac [/bib_ref]. Individuals who are unaware of their HIV status have a transmission rate of 3.5 times higher than individuals who are aware of their status. To improve HIV testing for MSM, HIVST can play a significant role.
This review and meta-analysis aims to understand the effect of oral HIVST on the frequency of HIV testing and the risk behaviors of MSM. The objective of this review is to compare the effects of oral HIVST with those of standard HIV testing. We hope to provide substantial evidence that may help in developing policy guidelines for the introduction of oral HIVST in India.
## Review
We conducted this systematic review in line with the PRISMA guidelines for systematic reviews and metaanalyses [bib_ref] Preferred reporting items for systematic reviews and metaanalyses: the PRISMA statement, Moher [/bib_ref]. The protocol was registered with PROSPERO on August 9, 2021 (registration number CRD42021261875).
## Eligibility criteria
For our review, we followed the PICO question [fig_ref] TABLE 1: Review PICO question [/fig_ref]. Only cluster randomized/randomized controlled trials (RCT) that compared oral or oral-and blood-based HIVST with standard or any other model of HIV testing among MSM in any global setting were included. Only studies that focused on the desired outcomes and provided quantitative results were selected for the review. Full text and abstracts or posters elaborating any one of the desired outcomes were included. No restrictions were placed on the language search. Qualitative studies, modeling studies, trial protocols, reviews, and study designs other than cluster RCT or RCT were excluded. Studies covering key populations (KPs), but not including MSM, were also excluded from the review.
## Pico question
## Population
Men who have sex with men (MSM)
Intervention Intervention that provides oral HIV self-testing or oral-and blood-based HIVST
## Search strategy
The following search strategy was adopted for PubMed and other electronic databases: ((HIV 1) OR (HIV)) OR (HIV2) OR (human immune deficiency virus) OR (HIV type 1) OR (HIV type2 (Title/Abstract)) OR (human immunodeficiency virus (Title/Abstract)) OR (human immune-deficiency virus (Title/Abstract)) OR (acquired immunodeficiency syndrome (Title/Abstract)) OR (acquired immune deficiency syndrome (Title/Abstract)) AND (dried blood spot (Title/Abstract)) OR (dried blood spot self (Title/Abstract)) OR (dried blood spot home (Title/Abstract)) OR (dried blood spot personal (Title/Abstract)) OR (dried blood spot remote (Title/Abstract)) OR (DBS (Title/Abstract)) OR (Oral fluid based test (Title/Abstract)) OR (Saliva based test self (Title/Abstract)) OR (saliva based community (Title/Abstract)) OR (saliva based online (Title/Abstract)) AND (standard of care (Title/Abstract)) OR (Western blot (Title/Abstract)) OR (confirmatory test (Title/Abstract)). For the Cochrane Library, the search strategy was HIV home-based testing in the abstract or HIV rapid test in the abstract OR HIV oral fluid-based test in the title, abstract, and keyword and MSM in the title, abstract, and keyword or men who have sex with men in the title, abstract, and keyword (word variations have been searched). Four reviewers (SV, NT, SJ, and AK) evaluated and assessed citations for eligibility and assessed the quality. A senior reviewer (SK) was consulted on citations for the resolution of discordance.
# Data abstraction
The first author (SV) and the study team members (NT, SJ, and AK) independently abstracted all data on the following items: study design, study population, sample size, outcome measures (uptake and frequency of HIV testing, new HIV infection detected, and risk behavior -condomless sexual intercourse (CAI) and proportion linked to care). A senior reviewer (SK) was consulted for the resolution of discrepancies in data abstraction.
The primary outcomes used to compare the oral HIVST with standard HTS were as follows: uptake of HIV testing services (defined as the proportion of people offered HIV testing who accepted and completed any HIV testing in a specified time frame) and frequency of HIV testing (defined as the mean number of HIV tests conducted during a specified time frame). We also assessed secondary outcomes that included the proportion of newly diagnosed HIV positive (defined as HIV infections detected during the specified time frame), sexual risk behavior (CAI (defined as condomless sex with partners) and multiple male sex partners), and repeat HIV testing (defined as more than two tests conducted in the follow-up period).
# Data analysis
A meta-analysis was conducted where the same or a comparable outcome was reported by two or more studies. It was conducted using random-effects models in RevMan version 5.4. For binary outcomes, the number of events was calculated and pooled. For continuous outcomes, the mean and SD were calculated and pooled. Statistical heterogeneity was evaluated. When outcomes were measured and reported at multiple time points, we used the longest time point or the end of the study period where possible. For two studies that had multiple intervention arms, data from HIVST arms were clubbed together and compared to the control arm as reviewers assessed that the interventions were unlikely to influence the outcome.
## Quality assessment
The quality of the clinical trials was assessed using Cochrane's risk of bias tool version 2.0 (RoB 2.0) [bib_ref] RoB 2: a revised tool for assessing risk of bias in randomised..., Sterne [/bib_ref]. This included evaluation of risk pertaining to six domains: bias arising from the randomization process, due to deviations from intended interventions, due to missing outcome data, in the measurement of the outcome, in the selection of the reported result, and overall bias .
## Figure 1: risk of bias
# Results
We could identify 2,426 records from various electronic databases and from other sources. After the removal of duplicates (n = 37), 2,389 records were screened. We excluded 2,322 records, and 67 abstracts and full texts were screened for eligibility, of which 59 were excluded. Eight studies met the inclusion criteria .
## Figure 2: study selection
HIVST: HIV self-testing; RCTs: randomized controlled trials; MSM: men who have sex with men
The eight RCTs enrolled 5,297 participants, of which 5,212 were MSM and 85 were transgender (TG) women. HIV-negative MSM or those of unknown status only were included in the study. Of the eight studies, seven were conducted in high-income countries (HICs): four in the USA [bib_ref] HIV self-testing increases HIV testing frequency in high-risk men who have sex..., Katz [/bib_ref] [bib_ref] A pilot, randomized controlled trial of HIV self-testing and real-time post-test counseling/referral..., Wray [/bib_ref] [bib_ref] Comparison of home-based oral fluid rapid HIV self-testing versus mail-in blood sample..., Merchant [/bib_ref] [bib_ref] Effect of Internetdistributed HIV self-tests on HIV diagnosis and behavioral outcomes in..., Macgowan [/bib_ref] , two in Australia [bib_ref] Effect of availability of HIV self-testing on HIV testing frequency in gay..., Jamil [/bib_ref] [bib_ref] The longer-term effects of access to HIV self-tests on HIV testing frequency..., Zhang [/bib_ref] , and one in Hong Kong [bib_ref] A randomized controlled trial evaluating efficacy of promoting a home-based HIV self-testing..., Wang [/bib_ref]. One was conducted in a low-middle-income country (LMIC) in Myanmar [bib_ref] High acceptability of HIV self-testing in a randomized trial among transgender women..., Wirtz [/bib_ref]. The characteristics of the included RCTs are shown in [fig_ref] TABLE 2: Characteristics of the included RCTsHIVST [/fig_ref]. In all the studies, HIVST intervention was provided with oral HIVST kits, except in one study in which both blood-based and oral HIVST kits were used [bib_ref] Effect of Internetdistributed HIV self-tests on HIV diagnosis and behavioral outcomes in..., Macgowan [/bib_ref]. All studies compared HIVST with standard HTS. The HIVST kits were either distributed at the study site (n = 4) [bib_ref] HIV self-testing increases HIV testing frequency in high-risk men who have sex..., Katz [/bib_ref] [bib_ref] Effect of availability of HIV self-testing on HIV testing frequency in gay..., Jamil [/bib_ref] [bib_ref] The longer-term effects of access to HIV self-tests on HIV testing frequency..., Zhang [/bib_ref] [bib_ref] High acceptability of HIV self-testing in a randomized trial among transgender women..., Wirtz [/bib_ref] or mailed (n = 3) [bib_ref] A pilot, randomized controlled trial of HIV self-testing and real-time post-test counseling/referral..., Wray [/bib_ref] [bib_ref] Effect of Internetdistributed HIV self-tests on HIV diagnosis and behavioral outcomes in..., Macgowan [/bib_ref] [bib_ref] A randomized controlled trial evaluating efficacy of promoting a home-based HIV self-testing..., Wang [/bib_ref] to the participants. In one study, the participants were provided with a weblink to purchase the kits online [bib_ref] Comparison of home-based oral fluid rapid HIV self-testing versus mail-in blood sample..., Merchant [/bib_ref]. The minimum age of participants was 18 years. Pre-test and post-test counseling were provided to the study participants through various modes [fig_ref] TABLE 3: Counseling of the study participants [/fig_ref].
## Uptake of hiv testing
Five of the eight RCTs reported the uptake of HIV testing. A meta-analysis showed that HIVST increased the uptake of HIV testing by 1.43 times compared to SoC (relative risk (RR) = 1.43; 95% confidence interval (CI) = 1.25, 1.64; Tau2 = 0.02; Chi2 = 19.89; df = 4; I2 = 80%). All the five RCTs showed an increase in the uptake of HIV testing, which was statistically significant (p-value = 0.0005) [fig_ref] FIGURE 3: Forest plot comparing the uptake of HIV testing in oral HIVST as... [/fig_ref]. Wirtz et al. [bib_ref] High acceptability of HIV self-testing in a randomized trial among transgender women..., Wirtz [/bib_ref] reported a 76% increase in testing rates across both arms, relative to baseline lifetime HIV testing history. In the study by Wang et al. [bib_ref] A randomized controlled trial evaluating efficacy of promoting a home-based HIV self-testing..., Wang [/bib_ref] , high uptake of any type of HIV testing was reported in the HIVST group at month 6 (89.8% versus 50.7%; RR = 1.77; 95% CI = 1.54, 2.03; NNT = 2.56; 95% CI = 2.13, 3.20; p = 0.001). The results remained statistically significant in subgroup analysis based on with and without CAI, multiple male sex partners, and experience in HIV testing in the last three years. Wray et al. [bib_ref] A pilot, randomized controlled trial of HIV self-testing and real-time post-test counseling/referral..., Wray [/bib_ref] reported 100% testing for HIV at some point during the study period using any test in the intervention arms as compared with 72% of control participants, a statistically significant between-group effect (F(2, 62) = 7.69; MS = 0.54; p = 0.001).
The study by Merchant et al. [bib_ref] Comparison of home-based oral fluid rapid HIV self-testing versus mail-in blood sample..., Merchant [/bib_ref] reported that 59% of the participants in the HIVST arm completed any type of HIV test (54% completed their assigned test, and 5% used a test they were not assigned), and 41% were not tested. However, in their study, the completion of the assigned HIV test was greater in the oral fluid rapid HIV self-test and the community organization/medical facility arms than in the mail-in blood sample collection HIV test arm (p < 0.01 for both comparisons). MacGowan et al. [bib_ref] Effect of Internetdistributed HIV self-tests on HIV diagnosis and behavioral outcomes in..., Macgowan [/bib_ref] reported a 55.7% increase in annual HIV testing among HIVST participants as compared to only a 6.9% increase in the control arm.
## Frequency of hiv testing
A meta-analysis of four RCTs found that the mean number of HIV tests increased by 2.34 during follow-up in the HIVST arm (mean difference = 2.34; 95% CI = 1.66, 3.02) . Three studies delivered HIVST through facility distribution (with additional optional mail distribution) and had smaller effect sizes at 2 [bib_ref] A pilot, randomized controlled trial of HIV self-testing and real-time post-test counseling/referral..., Wray [/bib_ref] and 1.7 [bib_ref] Comparison of home-based oral fluid rapid HIV self-testing versus mail-in blood sample..., Merchant [/bib_ref] [bib_ref] The longer-term effects of access to HIV self-tests on HIV testing frequency..., Zhang [/bib_ref] , respectively. According to Jamil et al. [bib_ref] Effect of availability of HIV self-testing on HIV testing frequency in gay..., Jamil [/bib_ref] , HIV testing in the HIVST group was significantly greater than in the SoC group (RR = 2.08; 95% CI = 1.82, 2.38; p < 0.0001). Zhang et al. [bib_ref] The longer-term effects of access to HIV self-tests on HIV testing frequency..., Zhang [/bib_ref] reported that the mean overall HIV testing frequency per person in year 2 among men in the intervention arm was higher than in year 1 of the SoC arm (3.7 versus 2.0; RR = 1.71; 95% CI = 1.48, 1.97; p < 0.001). In the study by Katz et al. [bib_ref] HIV self-testing increases HIV testing frequency in high-risk men who have sex..., Katz [/bib_ref] , MSM randomized to the HIVST arm reported significantly more HIV tests during follow-up (mean = 5.3; 95% CI = 4.7, 6.0) than those in the control arm (mean = 3.6; 95% CI = 3.2, 4.0; p < 0.0001). This represented an average increase of 1.7 tests per MSM over 15 months (95% CI = 0.9, 2.5). The study that demonstrated the largest difference at 3.80 was conducted by MacGowan et al. [bib_ref] Effect of Internetdistributed HIV self-tests on HIV diagnosis and behavioral outcomes in..., Macgowan [/bib_ref]. The HIVST participants reported frequent testing as compared to the control arm (mean number of tests over 12 months: 5.3 versus 1.5; p < 0.001).
## Figure 4: forest plot comparing the frequency of hiv testing in oral
HIVST as compared to standard of care HIVST: HIV self-testing; SoC: standard of care; CI: confidence interval
## New hiv infections
A meta-analysis of four RCTs indicated that oral HIVST had positive effect on the detection of new HIV infections among those tested, and this was statistically significant (RR = 2.10; 95% CI = 1.35, 3.28; I2 = 0%, p = 0.001) [fig_ref] FIGURE 5: Forest plot comparing new HIV infections in oral HIVST as compared to... [/fig_ref]. In the study by Katz et al. [bib_ref] HIV self-testing increases HIV testing frequency in high-risk men who have sex..., Katz [/bib_ref] , four in the HIVST arm and two in the SoC arm were HIVpositive. MacGowan et al. [bib_ref] Effect of Internetdistributed HIV self-tests on HIV diagnosis and behavioral outcomes in..., Macgowan [/bib_ref] reported more than twice as many HIV infections identified in the HIVST arm as compared to the SoC arm (1.9% versus 0.8%; p = 0.02). In the follow-up study by Jamil et al., three men were newly diagnosed with HIV during follow-up [bib_ref] Effect of availability of HIV self-testing on HIV testing frequency in gay..., Jamil [/bib_ref] , and the overall incidence was 0.9 per 100 personyears (95% CI = 0.2, 2.6). All new infections were in the HIVST group. HIVST identified 27 (15%) previously undiagnosed HIV infections compared to 14 (9%) identified by SoC [bib_ref] Effect of availability of HIV self-testing on HIV testing frequency in gay..., Jamil [/bib_ref].
## Repeat hiv testing
We conducted a meta-analysis on four of eight RCTs reporting results for repeat testing. The results of the meta-analysis show that the HIVST arm reported 2.80 higher repeat testing as compared to the control arm (RR = 2.04; 95% CI = 1.22, 3.42; I2 = 96%) [fig_ref] FIGURE 6: Forest plot comparing repeat HIV testing in oral HIVST as compared to... [/fig_ref]. The analysis showed a statistically significant result. Jamil et al. [bib_ref] Effect of availability of HIV self-testing on HIV testing frequency in gay..., Jamil [/bib_ref] reported more men with more than two HIV tests (76% versus 38%) during follow-up in the HIVST group as compared with the SoC group (p < 0.0001). In the study by Katz et al. [bib_ref] HIV self-testing increases HIV testing frequency in high-risk men who have sex..., Katz [/bib_ref] , MSM in the intervention arm reported three monthly testing as recommended (76% versus 54%, respectively; p = 0.001). Wray et al. suggested significantly different rates of repeat testing across study conditions (F(2, 62) = 5.33; MS = 1.06; p < 0.007) due to lower rates of repeat testing in the control condition (F(2, 62) = 24.5; p < 0.001) [bib_ref] A pilot, randomized controlled trial of HIV self-testing and real-time post-test counseling/referral..., Wray [/bib_ref].
## Risk behavior: condomless anal intercourse
A meta-analysis of three RCTs found that among MSM, oral HIVST had no statistically significant effect on CAI (odds ratio (OR) = 0.90; 95% CI = 0.67, 1.22; I2 = 2%; p-value = 0.51) [fig_ref] FIGURE 7: Forest plot comparing condomless anal intercourse among MSM in oral HIVST as... [/fig_ref]. Katz et al. [bib_ref] HIV self-testing increases HIV testing frequency in high-risk men who have sex..., Katz [/bib_ref] reported that at the nine-month and end-of-study surveys, MSM randomized to HIVST were 1.07 times more likely to report non-concordant CAI in the prior three months than MSM in the control group (OR = 1.07; 95% CI = 0.61, 1.90). Overall, there was no effect on the reduction of sexual risk behavior. Wang et al. [bib_ref] A randomized controlled trial evaluating efficacy of promoting a home-based HIV self-testing..., Wang [/bib_ref] also reported no significant difference in CAI (intervention: 27.5% versus control: 33.9%; p = 0.237). Statistically significant reductions in the prevalence of CAI (month 6 versus baseline: p = 0.002) were found within the intervention group only. Jamil et al. [bib_ref] Effect of availability of HIV self-testing on HIV testing frequency in gay..., Jamil [/bib_ref] reported no statistical association of reporting CAI with casual partners between the HIVST group and the SoC group (OR = 1.21; 95% CI: 0.62, 2.35; p = 0.575). Wray et al. [bib_ref] A pilot, randomized controlled trial of HIV self-testing and real-time post-test counseling/referral..., Wray [/bib_ref] reported no statistical association between the HIVST group and the SoC group after taking into account repeated measures, reporting condomless anal intercourse with casual partners (OR = 1.21; 95% CI = 0.62, 2.35; p = 0.575). This was not included in the meta-analysis as they reported that the results were not comparable.
## Risk behavior: multiple male sex partners
A meta-analysis was conducted on the three trials reporting results for multiple male partnerships in the follow-up period. The overall RR was 0.89 (95% CI = 0.83, 0.94; I2 = 0%) [fig_ref] FIGURE 8: Forest plot comparing multiple male sex partners of MSM in oral HIVST... [/fig_ref]. Wang et al. [bib_ref] A randomized controlled trial evaluating efficacy of promoting a home-based HIV self-testing..., Wang [/bib_ref] reported a significant between-group difference at month 6 (intervention: 34.2% versus control: 47.7%; RR = 0.72; 95% CI = 0.54, 0.95; p = 0.021). In the study by Katz et al. [bib_ref] HIV self-testing increases HIV testing frequency in high-risk men who have sex..., Katz [/bib_ref] , HIVST was non-inferior to SoC with respect to multiple male sex partners. At the end of the study surveys, MSM in the oral HIVST arm reported 8% fewer male CAI partners than MSM in the control arm in the prior three months (incidence rate ratio = 0.92; 95% CI = 0.64, 1.33). MacGowan et al. [bib_ref] Effect of Internetdistributed HIV self-tests on HIV diagnosis and behavioral outcomes in..., Macgowan [/bib_ref] also reported similar results, where no statistically significant difference between participants was observed. The participants in all the trials who received a reactive result on oral HIVST were linked to care. The participants were contacted, and counseling was provided either by the administrators or on phone by research staff. We also looked for any reported social harms in the trials. We did not find any included trials that studied problems faced with oral HIVST manipulation or unintended harm.
# Discussion
HIVST was introduced way back in 2000 by the Joint United Nations Program on HIV/AIDS [bib_ref] Can self-testing increase HIV testing among men who have sex with men:..., Zhang [/bib_ref] and had only been implemented in 38 countries by June 2019 [bib_ref] Comparing the effects of HIV self-testing to standard HIV testing for key..., Witzel [/bib_ref]. HIVST could play a major role in increasing HIV testing. The studies analyzed in this review were published in the last five years and were limited to MSM. This systematic review and meta-analysis of eight RCTs conducted among MSM found that when compared to SoC, oral HIVST increases the uptake and frequency of HIV testing. The free distribution of HIVST kits can potentially increase the uptake and frequency of HIV testing among MSM. One study, where the participants had purchased the kits, also reported an increase in the uptake of HIV testing [bib_ref] RoB 2: a revised tool for assessing risk of bias in randomised..., Sterne [/bib_ref]. This can result in a reduction in the proportion of MSM unaware of their status due to non-testing for various reasons. The overall outcome may be decreased in new infections as desired for achieving "End AIDS by 2030." In addition to the increase in the uptake of HIV testing, the meta-analysis suggests that repeat testing for HIV also increased oral HIVST use. This is an important feature of the HIVST strategy, as the prevalence of HIV is 6-13 times higher among KPs, including MSM, in India as compared to the adult prevalence due to their highrisk behavior. Repeat testing 3-4 times a year is recommended by the WHO for high-risk groups (HRGs). Oral HIVST has the potential to increase repeat testing among MSM and thus early detection of the HIV infection.
Four RCTs included in this review assessed the impact of oral HIVST in detecting new infections. A statistically significant positive effect was seen in the analysis. More new infections were diagnosed in the oral HIVST groups as compared to the SoC arm. Thus, more MSM were aware of their status by the introduction of oral HIVST. This is one of the main goals of the 95-95-95 targets, and oral HIVST has the potential to achieve this target. Another study by Okoboi et al. distributed oral HIVST kits through peers and reported identifying eight (5.6%) participants with undiagnosed HIV infection during the three months of follow-up compared to only four (2.7%; p = 0.02) in the SoC arm [bib_ref] Peer distribution of HIV self-test kits to men who have sex with..., Okoboi [/bib_ref]. The study was not included in the review as it did not match our eligibility criteria.
This increase in HIV testing uptake, repeat testing, and identification of new infections have important public health implications. Approaches to increase regular HIV testing supplemented by oral HIVST could identify infections at an early stage. This would also help attain the WHO recommendations of quarterly testing by high-risk populations, especially MSM. Oral HIVST has the potential of not only increasing the detection of undiagnosed HIV infection but also expanding HIV testing among non-frequent/delayed testers [bib_ref] The longer-term effects of access to HIV self-tests on HIV testing frequency..., Zhang [/bib_ref].
There was no increase in the sexual risk behavior of MSM as measured by the analysis of CAI with regular or casual partners and multiple male partnerships. None of the studies in the review reported significant differences in CAI and multiple male partnerships between the experimental arm and the control arm. This suggests that the fear of the increase in risk behavior due to oral HIVST is unjustified. It points to the fact that continuous behavioral modification through regular health education and counseling is required to reduce risk behavior among MSM.
# Strengths and limitations
In our study, we focused on the effectiveness of oral HIVST kits only among MSM. Evidence has shown that oral HIVST kits are less painful than finger-prick HIVST kits and are more acceptable to KPs for HIV testing [bib_ref] Acceptability, perceived reliability and challenges associated with distributing HIV self-test kits to..., Okoboi [/bib_ref] [bib_ref] Acceptability of HIV oral self-test among men having sex with men and..., Rao [/bib_ref]. We also included the latest evidence in the literature on oral HIVST.
The limitation of our study is that most of the studies included in the analysis were conducted in HICs.
Limited studies from LMICs limit the scope of generalizability of the results to resource-limited settings. The analysis or review of social harm could not be carried out as well, as none of the included trials covered the topic of social harm. Synthesis of results from qualitative studies is required on this topic.
# Conclusions
Our systematic review and meta-analysis suggests that oral HIVST could increase the HIV testing and detection of new HIV infections among the high-risk population of MSM to optimal levels. HIVST could be a major strategy contributing to the 95-95-95 target of the WHO and bringing on track the efforts to end AIDS by 2030. HIVST can be made available in countries where it is not yet available. Programs best suiting the cultural and economic milieu can be implemented to reach those who are at risk and are not aware of their status.
# Additional information disclosures
## Conflicts of interest:
In compliance with the ICMJE uniform disclosure form, all authors declare the following: Payment/services info: All authors have declared that no financial support was received from any organization for the submitted work. Financial relationships: All authors have declared that they have no financial relationships at present or within the previous three years with any organizations that might have an interest in the submitted work. Other relationships: All authors have declared that there are no other relationships or activities that could appear to have influenced the submitted work.
[fig] FIGURE 3: Forest plot comparing the uptake of HIV testing in oral HIVST as compared to standard of care HIVST: HIV self-testing; SoC: standard of care; CI: confidence interval [/fig]
[fig] FIGURE 5: Forest plot comparing new HIV infections in oral HIVST as compared to standard of care HIVST: HIV self-testing; SoC: standard of care; CI: confidence interval [/fig]
[fig] FIGURE 6: Forest plot comparing repeat HIV testing in oral HIVST as compared to standard of care HIVST: HIV self-testing; SoC: standard of care; CI: confidence interval [/fig]
[fig] FIGURE 7: Forest plot comparing condomless anal intercourse among MSM in oral HIVST as compared to standard of care HIVST: HIV self-testing; SoC: standard of care; CI: confidence interval [/fig]
[fig] FIGURE 8: Forest plot comparing multiple male sex partners of MSM in oral HIVST as compared to standard of care HIVST: HIV self-testing; SoC: standard of care; CI: confidence interval [/fig]
[table] TABLE 1: Review PICO question [/table]
[table] TABLE 2: Characteristics of the included RCTsHIVST: HIV self-testing; CAI: condomless anal intercourse; HIVST-OIC: HIVST kit and online real-time instructions and pre-test/post-test counseling; YMSM: young men who have sex with men; STI: sexually transmitted infection [/table]
[table] TABLE 3: Counseling of the study participants [/table]
|
Re-wiring Guilt: How Advancing Neuroscience Encourages Strategic Interventions Over Retributive Justice
The increasing visibility of neuroscience employed in legal contexts has rightfully prompted critical discourse regarding the boundaries of its utility. High profile debates include some extreme positions that either undermine the relevance of neuroscience or overstate its role in determining legal responsibility. Here we adopt a conciliatory attitude, reaffirming the current value of neuroscience in jurisprudence and addressing its role in shifting normative attitudes about culpability. Adopting a balanced perspective about the interaction between two dynamic fields (science and law) allows for more fruitful consideration of practical changes likely to improve the way we engage in legal decision-making. Neuroscience provides a useful platform for addressing nuanced and multifaceted deterministic factors promoting antisocial behavior. Ultimately, we suggest that shifting normative attitudes about culpability vis-à-vis advancing neuroscience are not likely to promote major changes in the way we assign legal responsibility. Rather, it helps us to shed our harshest retributivist instincts in favor of more pragmatic strategies for combating the most conspicuous patterns promoting mass incarceration and recidivism.
# Introduction
Increasing attention is being devoted to the emerging roles of neuroscience in legal decisionmaking, both in academic settings and in the courtroom. Among these roles is the growing influence neuroscience has in reinforcing more deterministic models of human decision-making and behavior. In a deliberate attempt to (over)simplify this complex landscape, it seems that there are roughly two camps in these conversations. The first includes those who promote the idea that neuroscience has mostly "disproved" the existence of free will, which subverts some of our ordinary notions of accountability. Prominent voices in popular media have indeed heralded the end of free will, calling into question our most basic presumptions about the legitimacy of punitive justice. As a consequence, the criminal justice system, which punishes people on a now-baseless presumption of freedom and agency, has been foundationally undermined and must therefore be replaced immediately with something more enlightened and fair. Expectedly, such claims have incited substantial opposition and motivated counter-arguments aimed at substantiating traditional views of legal responsibility and the enduring value of retributive sanctions against criminal actions. This opposing camp includes those who claim that free will (whether it exists or not) is largely irrelevant to basic notions of legal responsibility, and neuroscience has little to no relevance for assessing guilt or any ordinary sense of civic accountability. As such, the status quo can be safely perpetuated, and the law, as it stands, remains unfettered by trifling nuisances of predetermined actions. Perhaps as a kind of contrecoup effect, these counter arguments often involve rebuffing the very relevance of neuroscience in the legal process more generally. Of course these descriptions are composite caricatures of many subtler perspectives available in this growing academic conversation, e.g.. Still, many of the arguments we read these days overlap substantially and obviously with one of these two extreme positions. Without depreciating the zeitgeist of this revolution or its opponents, we recognize the need for some conservatism in advancing pragmatic attitudes about how the legal system might change in the wake of advancing neuroscience. We therefore set out here to willfully explore the vast middle ground between seemingly extreme perspectives in this conversation. We ultimately promote three main theses related to Neurolaw and its inevitable progress.
## Neuroscience has firmly established its place in jurisprudence
Neuroscience already plays a prominent role in legal proceedings. This trend seems likely to increase rather than decrease; however, this development should not be alarming. As judges and juries are faced with the challenge of weighing this evidence in their decision-making, we have a responsibility to make this process as transparent as possible. This involves promoting better science and better education to judges and legal counsel about these data's interpretation, limitations, and best practices in quality assurance and analytic strategies.
## Our normative understanding of free will and culpability is changing
Our understanding of human behavior and free will has steadily incorporated more contributions from science throughout history. The role of neuroscience only represents a recent and specific extension of this progress. Increasingly deterministic models of behavior need not cripple our aim to hold people responsible for their actions, but they arguably drive changes in our normative view of culpability and what constitutes justice.
## Our legal system is evolving, not static
Recent changes in our legal system highlight evolving standards in normative judgments regarding the relative value of retributive and rehabilitative interventions. Neuroscience provides a platform to re-assess the value of primarily punitive systems that have historically done little to remedy mental health and social issues that perpetuate high incarceration rates. Rather than eroding jurisprudence, this has the potential to inform more effective policies that serve our society in progressive ways.
## Neuroscience has firmly established its place in jurisprudence
Critical evaluations of the role of neuroscience (and particularly neuroimaging) in legal proceedings abound in both academic publications and the popular media. While the tone of these pieces can range from cautionary to insolent, they attend to a common issue of what is frequently described as a meteoric rise in the consideration of neuroscience-based evidence in courtroom decision-making. They frequently highlight perceived negative consequences of this trend, and some suggest the limited relevance of neuroimaging in court overall. Here we argue that neuroscience and neuroimaging in particular have already established their place in legal proceedings, which is unlikely to subside. A more practical approach for exercising caution in its application will be to improve stakeholders' understanding of the strengths and limitations of these techniques which includes educating lawyers, judges, and the general public. Educated adoption of neuroscience in legal settings is a practical and realistic solution to any perceived hazards it engenders.
The characterization of a meteoric rise of neuroscience used in court carries with it a somewhat menacing connotation that may not be wholly justified. While estimates have suggested an approximate doubling of cases that consider neuroscience data as legal evidence in the past decade, this is not out of step with its rise in clinical and research settings over the same period. Further, in contrast to the tone of many commentaries, this steady increase has not occurred unexpectedly, overwhelming courts with claims that its practitioners cannot fairly evaluate. One of the first considerations of brain imaging as evidence in court occurred over 35 years ago in the high profile trial of John Hinckley Jr. for the attempted assassination of President Ronald Reagan (United States vs. Hinckley, 1982 1 ). Brain scans were used in conjunction with other clinical evidence to support Hinckley's diagnosis of schizophrenia. The brain imaging was not foundational to his diagnosis, but served the purpose of grounding his claims of mental illness in a physical domain (as opposed to purely "psychological") -an important educational element for jurors in the early 1980s. This context remains among the most influential roles of neuroscience in contemporary jurisprudence, where judges and juries must inevitably weigh the "legitimacy" of health claims and related assertions that remain difficult to account for objectively (e.g., chronic pain, psychiatric disorders). Hinckley was found not guilty by reason of insanity and was committed to a secure hospital for the mentally ill.
Since then, the application of neuroscience evidence in court has increased, but not on a scale that is out of step with advancing scientific knowledge and improved practical utility. Any notions that the rise in neuroscience has happened too quickly for courts to implement its data sagaciously are likely misguided. Several reports have objectively evaluated this changing landscape. Using broad inclusive criteria, it has been estimated that neuroscience evidence is considered in less than 1% of all criminal proceedings (based on court of appeals data) and only about 5% of murder trials dating back to 2005. Contexts that have seen more pronounced increases include evaluation of competency to stand trial, capital cases, and appeals for mitigation of punishment. An observation made frequently across studies is that the rise in neuroscience evidence appears most striking in high-stakes cases. While these numbers are steadily increasing, they do so in-step with advancing scientific understanding and improved potential to inform judges and juries about a number of relevant aspects of mental health, within standard limits of due process -and this context is important. As with any emerging technology, its relevance in legal proceedings must be evaluated carefully with established evidentiary standardsas courts learn to integrate and adapt to progress in clinical neuroscience. Nonetheless, the utility of neuroscience and neuroimaging data in court is being increasingly realized as counsel, scientists, and practitioners work together to further establish its value in different contexts. Indeed, courts are often recommending, if not requiring, that neuroscience evidence be produced to support arguments being made, when these data are potentially informative.
Understanding the many possible applications of neuroscience in the courts can help broaden this perspective and allay concerns that its arrival in legal proceedings is premature or imprudent. First, neuroscience evidence in the United States has almost never been an essential factor in determining actus reuswhether or not the accused actually committed the criminal act in question. However, for an interesting international example see State of Maharashtra v. Sharma 2 . and commentary by. Neuroscience evidence, in the U.S. and worldwide, is far more commonly determined to be relevant for assessing the competency of an individual to stand trial or in addressing the mens rea element of criminal liability, but for a more nuanced summary of these various contexts see. These latter contexts relating to mens rea often require thorough assessments of a defendant's mental health, which may reasonably include neuroimaging. Defenses built on this reasoning include pleas of insanity, which have evolved substantially over time (see section Our legal system is evolving, not static). Nonetheless, insanity pleas essentially argue that someone was so mentally impaired that s/he was not aware of their own actions, or able to decipher right from wrong. Still, a brain scan may only represent one element of a comprehensive clinical evaluation in such applications, or may be determined to be irrelevant given extant clinical evidence -that is, a brain scan is often not necessary for determining ones mental health status, but may provide supportive evidence. In other cases, neuroimaging may be essential (e.g., brain injury, degenerative disease, tumors). To be sure, insanity pleas overall constitute less than 1% of all felony trials, and their success ordinarily accompanies cases in which a defendant had previously been 2 State of Maharashtra v. Sharma, C.C., No. 508/07, Pune, June 12, 2008 (India). diagnosed with a mental illness.
As noted above, neuroscience data is also increasingly submitted in the sentencing phase of a trial (e.g., after guilt has already been determined), and evidentiary standards are somewhat more permissive and accommodating during these arguments. This is becoming more common in high stakes sentencing decisions, for instance, in capital cases when the convicted offender faces either the death penalty or life in prison. Neuroscience data may then be considered relevant when deciding whether one deserves the harshest possible punishment or a sentence reflecting intervening factors that can include mental health status. Indeed the relevance of mental health in some cases is considered so pertinent that a failure to introduce neuroscience data has been determinative of ineffective counsel and a violation of a defendant's constitutional right to fair representation.
Additional applications of neuroscience arise in civil cases, which may require demonstrating extent of physical injury. Applications of brain imaging in this context include established documentation of gray and white matter injury with structural brain imaging, but may extend to novel applications that provide information on concussion and mild traumatic brain injury, and application of functional imaging techniques addressing chronic pain. These later examples are emerging areas that need to be vetted further by the scientific community; however, their appeal and potential value is undeniable, underscoring the importance of lawyers and judges remaining in touch with advancing technology. In this way, it is imperative that legal counsel is adequately educated about the relevance and interpretation of neuroscience-based evidence that may aid the fair evaluation of each case.
It should be clear that, in any context, the tools of neuroscience are not subject to more lenient standards than other forms of evidence presented in legal arguments. That is, their probative value must be weighed against the potential for introducing a prejudicial impact or confusion among jurors. Commentaries often use this as a linchpin for their arguments, citing (limited) evidence that brain imaging evidence may mislead jurors and/or distract them from primary lines of reasoning. Importantly, this evidence has been critically evaluated by others who have noted that these investigations did not present information in a context matching what juries typically encounter. Other studies accounting for these factors have indicated no evidence that brain imaging carries any additional weight over and above verbal neuroscience-based testimony . Further, when evaluated in a legal context where cross-examination critically evaluates the relevance of information, MRI-based evidence is no more persuasive than other (non-neurosciencebased) evidence. Finally, the role of the judge as a kind of gatekeeper for admissibility of evidence protects the system from more controversial applications of these tools. This has been demonstrated effectively time and again as courts have rejected the use of fMRI, for example, as a form of lie detection ; State v Gary Smith 2012 4 .). This has persisted as there is not sufficient scientific consensus for these purposes, at present. As such, the use of fMRI in this context does not pass established Daubert standards of evidence.
The checks and balances built into the U.S. legal system have largely been effective in the face of expanding neuroscience evidence. It should be clear, however, that these safeguards are not intended to unilaterally prevent change in the legal system. Rather, they are intended to promote adaptive interpretation, reflecting normative standards that shift in step with increasing knowledge, advancing technology, and evolving public attitudes (see section "Our legal system is evolving, not static"). As technology continues to improve and new applications arise, it is essential for practitioners of the law to remain adequately informed in order to best serve their roles. The recognition of this imperative is increasingly evident in the resources and attention being devoted to these objectives in recent years. The MacArthur Foundation Law and Neuroscience project and the Research Network on Law and Neuroscience (MacArthur, 2019) represent large, multimillion dollar investments serving these needs. These efforts accompany many formal educational resources for judges and lawyers that specifically address topics of neuroscience (FJC, 2019), as well as ongoing development of many international conferences and academic societies devoted to increasing scholarship and improving communication within these integrative disciplines.
Our initial assertion in this commentary is intended to be uncontentious. Simply, there are many contexts in which the relevance of neuroscience data is already firmly established, and may in fact be essential for carrying out effective legal decision-making. Most of these applications are not new, but the breadth of their relevance has perhaps widened as their informative value improves in stride with progress in research and the technology itself. The relevance of neuroscience data in jurisprudence shows no evidence of diminishing in the coming years; therefore, we encourage an attitude of integration and motivated legal scholarship. The importance of this is clear even given the limited examples provided here, which leave out additional concerns regarding constitutional principles, moral/ethical considerations, e.g., Pallarés-Dominguez and Gonzalez Esteban (2016);, and Napier (2019), and emerging perspectives in international law. Ongoing critical evaluation of the utility and limits of neuroscience will remain an essential component of this progress. Occasional dismissals of neuroscience's evolving relevance, in our view, are myopic and potentially dangerous. Criticisms on this order are usually intended to reinforce a static view that the law can continue to operate as effectively without neuroscience, simply because it has in the past. However, this attitude offers little guidance for the inevitable progress facing an assuredly dynamic field, which requires close evaluation of evolving technology and evidence. We therefore reinforce a perspective that the best way for the system to adapt to advancing 3 US v. Semrau, 693 F.3d 510 (6th Cir. 2012). Smith v. State, 32 A.3d 59, 423 Md. 573 (2011) (pretrial testimony). technology is by improving education and resources available to legal professionals who are increasingly required to incorporate these data in their arguments.
## Our normative understanding of free will and culpability is changing
Humans have been grappling with the concepts of free will and determinism (or fatalism) for most of recorded history; however, ancient notions of these concepts had little to do with the brain and neuroscience. Instead, philosophers and storytellers alike considered how much of our behavior was controlled by gods, the fates, or other supernatural external forces. Similarly, behavior that was attributable to our own motivations and decisions were not always nested in the brain. Aristotle for instance believed the brain mostly served to cool the blood. Rather, our motivated behavior has historically been attributed to something immaterial like a spirit, soul, or will. As physical sciences improved our understanding of neuromuscular junctions, neurotransmitters, and the role of the brain in organizing behavior based on prior experience, the role of a soul necessarily diminished. In his book Soul Made Flesh, Carl Zimmer develops a vivid history of neuroscience around the idea that advances in physiology, medicine, and psychology have incrementally narrowed the role of an immaterial soul as science has increasingly explained biological systems responsible for cognition and behavior.
In many ways, evolving perspectives about free will represent an extension of this trajectory. As neuroscience offers more detailed and predictive models accounting for human motivations, appetitive drives, and behavioral inhibition, extant descriptions of free will increasingly seem to grasp at something immaterial and elusive. This, somewhat covertly, promotes a paradigm incompatible with natural science, which progresses on a foundation that is fundamentally materialist, reductionist, and determinist in nature. As this represents a predictable extension of prior historical and philosophical progress, the questions neuroscience addresses on this topic are not new ones. However, neuroscience provides an increasingly tangible and convincing platform for demonstrating the limits, proximal antecedents, and illusions that support our subjective sense of free will. The relevance of this for promoting evolving attitudes in jurisprudence relate to how we, as a society, exercise normative judgments about agency, responsibility, and most importantly culpability. Here we illustrate how these attitudes are slowly shifting, and we emphasize the role neuroscience plays in influencing these standards.
Any treatment of how neuroscience has influenced our understanding of free will must address the studies of Benjamin Libet, and perhaps more importantly, contemporary extensions of this work. In the 1980s Libet published research demonstrating the precise timing of one's subjective perception of making a simple decision to freely move one's wrist in relation to other physiological events. The study essentially recorded three events: the movement of the wrist, the time the participant "decided" to move their wrist, and neural activity surrounding these events. The neural activity indicative of preparations to move one's wrist was already known (since the 1960s), and is commonly referred to as the readiness potential. What was striking in Libet's experiments was the demonstration that this neural preparatory activity consistently preceded one's subjective sense of having decided to move, by about 350 ms. Preliminary interpretations of these outcomes suggested that neural activity preceding the decision-point constituted evidence of a deterministic process that had already begun, prior to our subjective awareness of it, undermining conventional notions of agency or free will more generally. These initial conclusions have been rightfully debated for decades, while others have more quietly continued to improve and expand on these methods.
More recent extensions of this work have included the application of machine learning algorithms to accurately predict subjects' movements before they decide to move. This has been carried out using intracranial, intracellular recordings within the supplementary motor cortex. Functional MRI recordings measuring patterns of neural activity across the whole brain have also reliably predicted which of two buttons someone will press up to 7 s before they indicate they've decided. However, other exciting research demonstrates that these behaviors are not determined in such a simplistic way; but rather, they remain influenced by parallel cognitive systems. Executive control systems can essentially veto an intended movement, if given as little as 200 ms warning. That is, a 'stop' signal triggered by a real-time prediction of one's intended movement is sufficient to allow inhibition and eliminate that movement, provided it is delivered at least 200 ms prior to the execution of the event.
Demonstrations like these provide concrete evidence that our decisions and motivations are accompanied by many parallel neural mechanisms that occupy a dimension beyond our conscious, deliberative processes of reasoning. Measuring the corresponding neural activity provides tangible, proximal measurements of these processes, but neuroscience is not the only context in which we are aware of such unconscious influences on behavior. Freud may have been the first to draw public attention to the prominent role of subconscious influences on our otherwise rational behavior. More recently, studies of decision-making in contexts ranging from economics to moral deliberation have made it clear that our choices are strongly guided by implicit emotional influences that often deviate from rational optimization, and the narratives we construct around our decisions are often architected in a post hoc manner. Finally, we are increasingly aware of the predictable consequences of many remote influences that we have no individual control over. These include our genetic makeup, early rearing environments, and complex social systems. These factors all bias our cognition and behavior in predictable ways and their influence impinges on neural systems that guide our behavior directly, in ways that are largely inaccessible to our moment-to-moment conscious, deliberative processes. Better understanding of these influences has, in some ways, also changed the way we reason about culpability.
Challengers to the role of neuroscience in legal contexts will often argue that claims of functional impairments based in neuroscience contribute little to our normative judgments about culpability. This is based, in part, on the rationale that innumerable others with similar impairments have undoubtedly not committed similar crimes, and so the impairment (by itself) is insufficient to predestine the crime. This rebuttal fails to recognize that deterministic influences rarely operate in isolation, and our normative judgments ordinarily consider multiple factors and contextual circumstances. Further, the deterministic limits of isolated factors on criminal behavior are not uniquely reserved for neuroscientific considerations. This same argument, for instance, fails to undermine the relevance of something like faulty brakes influencing our normative judgments about a fatal car accident, given that faulty brakes only sometimes lead to fatalities, see also. This perspective also misses a somewhat more overarching role that neuroscience plays in shifting normative judgments about culpability. That is, neuroscience can help shift our judgments by simply grounding facts about psychological differences in a physical realm, underscoring their contextual relevance among many forms of physical evidence.
If the processes of motivation and decision-making are seen only as imponderable mysteries, inaccessible to reductionistic science, then we are constrained by limited insight into the origins of behaviors we ostensibly wish to diminish in society. We are further bound to make more simplistic normative judgments based on right and wrong, and our interventions will be more unidimensionally focused on reactive punishment. Conversely, integrating deterministic perspectives in explaining behaviors society condemns doesn't prevent us from using punishment as a deterrent, but only highlights additional points of leverage useful for applying more proactive interventions as an added method of diminishing unwanted behavior, see alsoand.
Another interesting context from which to observe this evolving landscape is to consider relatively common forms of pathology that impinge on our ability to choose and behave freely. Fitting examples include obsessive-compulsive disorders and addictions. In both cases, individuals can be said to lose some control over behavior that, in healthy individuals, is attributed to ordinary volitional processes. Normally, washing our hands, going over a mental list, or enjoying a beer are all considered among our ordinary, voluntary, healthy behaviors. Under pathological conditions, however, compulsions to engage in these or other behaviors encroach on (and supersede) other normal motivations. Daily goals, long-term ambitions, and explicit objectives may be at odds with increasingly intrusive thoughts and behaviors that an individual has limited control over. An individual may fully understand, anticipate, and wish to avoid the consequences of certain maladaptive behaviors, while still succumbing to well-worn patterns leading to the undesired behavior. Common understanding about the pathophysiology of these disorders has altered how we address these issues both clinically and interpersonally.
The current accepted model of addiction promoted by the National Institutes of Health is that of a brain disorder instantiated in motivational and inhibitory systems, brought on by exposure to substances that pharmacologically impose lasting physiological changes on these systems (NIDA, 2019). Like other diseases, genetic vulnerabilities, environmental exposures, and variability in physiology all promote individual differences in susceptibility to addiction. Unlike many other diseases, the observable symptoms are almost entirely behavioral. Moreover, these behaviors are often categorically illegal and punishable by law (in the case of illicit drug use), but may also be viewed under a moral lens as a transgression against more virtuous decision making. What makes this acutely relevant to discussions of neurolaw is the nature of arguments for and against the disease model of addiction, and how they reflect philosophical discourse on free will, neuro-determinism, and culpability. Opposition to the disease model can be easily found in publications such as Heyman's Addiction: A Disorder of Choice, Schaler's Addiction is a Choice, and Satel and Lilienfeld's Addiction and the Brain-Disease Fallacy. These arguments make rhetorical appeals to the primary role of choice, agency, volition, and selfcontrol. In doing so, they tacitly place limits on reductionist approaches that examine supportive physiological processes that govern our choices. These arguments seem rooted in the fundamental conservation of free will as something irreducible, and impervious to reductionist, deterministic paradigms.
By contrast, neuroscientific research nested in the disease model of addiction studies elements of motivated behavior in simpler parts, examining individual variability across these dimensions. For example, this research examines shifts in valuation (e.g., the motivational weight of pharmacological reinforcers over natural reinforcers) along with the weakening of inhibitory control. These approaches also examine transitions between behavior governed primarily by executive control systems and behavior carried out by networks governing compulsive, automatic actions. As such, the "disease" aspect of addiction is more fundamentally rooted in the physiological systems that govern our choices and behaviors, rather than in the complex behavioral symptom of drug-taking per se. Adopting this perspective requires a reductionist and determinist paradigm for informing our free will. While not universally accepted (and perhaps still requiring semantic refinement), the progressive contributions of the disease model of addiction include a better understanding of biological influences that culminate in our motivated behavior. Progress on this front further serves to dilute a predominantly moralistic attitude toward addiction that may motivate primarily punitive actions intended to address a very legitimate societal problem. This contribution will feature heavily in our ongoing assessment of the relevance of neuroscience in an evolving criminal justice system.
These arguments are familiar in the context of debates about the nature of free will and responsibility. While our intuitions may still demand the preservation of a concept like free will and agency in our behavior, it has become increasingly necessary to clarify what aspects of our thoughts and behavior remain free, to what extent they are free, and (perhaps most important for our purposes here) what the relevance of this is for our judgments about how to intervene to address pragmatic social needs. After all, moral responsibility is more abstract and partially removed from the practical considerations of punishment and intervention in our justice system. As it turns out, laypeople's judgments about these topics are not always internally consistent, often reflecting shifting attitudes when considered in abstract terms vs. concrete examples. For instance, when considering theoretical arguments, people are more likely to maintain that determinism undermines basic moral responsibility; when considering concrete episodic scenarios, we are more likely to affirm basic accountability for our actions responsibility.
Using the disease model of addiction as an example, opponents do not deny the evidence of biological changes in motivational systems that account for changes in behavior. However, opponents still cling to the relevance of individual agency, free will, and decision making, ostensibly apart from their biological influences, perhaps only because this reaffirms our most basic intuitions about choice, consequences, and our ability to change. This veneration of free will over the biological systems that govern choice may have counterproductive consequences, however. The best methods for intervention arguably improve by understanding the biological systems governing our choices and motivated behavior, particularly in the context of maladaptive behaviors involving substances that impinge directly on these systems. It is the context of intervention that becomes highly relevant for our consideration of the ongoing role of neuroscience in the future of jurisprudence. The influence of neuroscience on these concerns is already evident in a number of contexts discussed in the next section, and it has the potential to continue to improve our practical management of an imperfect but adaptable criminal justice system.
## Our legal system is evolving, not static
The relevance neuroscience has in our current justice system is already firmly established in several contexts outlined in Section "Neuroscience Has Firmly Established Its Place in Jurisprudence." The way neuroscience is promoting progressive changes in our justice system is also evident in a number of ways we address here. We can use recent examples of these changes to help anticipate the ongoing evolution of jurisprudence as informed by advancing neuroscience. Importantly, we reiterate our position that the influence of neuroscience has relatively less to do with any perceived exculpatory extensions of a purely deterministic universe (my brain made me do it), and is more practically relevant for the way we interpret concepts like "justice" and the role of the justice system in promoting a safe, functioning society. Shifting normative attitudes on this front influence how we choose to intervene and hold people accountable for their actions. Neuroscience, after all, has improved our general understanding of motivated human behavior and myriad deterministic influences that converge to promote maladaptive, antisocial behavior. Where there is improved understanding of these influences, we will be better equipped to introduce improved strategies at remedying systemic problems contributing to the behaviors and societal problems we aim to diminish.
Conservative appeals to traditional applications of jurisprudence regularly make the claim that neuroscience need not change anything about the way we interpret legal responsibility. This is true in one sense: if our only motivation is preservation of the status quo. In an article previously published in this series, Criminal Responsibility and Neuroscience: No Revolution Yet, Bigenwald and Chambonestablish that no revolution is necessary for us to continue applying the same normative framework of responsibility that the legal system has always operated on. Several arguments are presented emphasizing the primacy of our intuitions about agency for assigning criminal responsibility. That is, even the reality of a purely deterministic universe does not negate criminal responsibility, which in their view, exists as a mostly pragmatic concept independent of free will. This is true only in that our legal system certainly employs a number of arbitrary rules in order to remain serviceable. It is no great leap in understanding to suggest that it simply operates 'as if ' we are responsible agents. Our objection on this matter is that this reality will become increasingly dissatisfying, even from a normative perspective, as general knowledge increases, providing more insight into the boundaries and limitations of our own agency. Fortunately, the present series of articles is under no obligation to preserve the status quo; but rather, it challenges us to describe how the legal system might be practically changed by discoveries in neuroscience. We therefore submit that these changes may be less visible in the ways we interpret and enforce the law, and more visible in the ways we punish violators of our laws and adapt as a society to preserve (or advance) our most pressing goals.
As Bigenwald and Chambon point out, responsibility has many possible meanings. A tree can certainly be responsible for falling on a wire and causing a power failure, even though it has no real agency. Calling a tree responsible for these consequences doesn't violate any of our intuitions about agency and its value. Calling a tree "guilty" for this, however, feels odd (violates our intuitions), just as wishing to implement retributive harm on the tree would seem senseless. This illustration emphasizes a division between practical considerations of responsibility and the attribution of a kind of value judgment about the tree's actions, and how they align with normative moral values. Even as we are keenly interested in (also) preventing other trees from falling (deterrence), one of the key roles of our justice system remains a punitive one, and this features heavily in how harshly we decide to punish. Our intuitions about agency, free will, and moral judgment play a much larger role in our instinct to punish the guilty. Where neuroscience may play its most significant role is in the space between legal determinations and implementing corrective measures that benefit society. Here there remains a great deal of room for improving strategies aimed at protecting and benefiting society on a large scale. These changes in normative attitudes are evident in the evolving standards we use in legal sentencing and the ways we continue to evaluate the relative efficacy of various punitive strategies.
As noted, the criminal justice system in the United States serves many functions beyond a punitive one. We rely on it to deter flagrant abuses of the law, to protect society at large from the most dangerous individuals, and (ostensibly) to help intervene and rehabilitate those who violate the expectations of their social systems. The current implementation of this system, however, has been heavily biased toward retributivist deterrence strategies, which have demonstrated their own limitations over several decades. Indeed, they have contributed, in part, to the highest incarceration rates, per capita, in the entire world. Public attitudes play an overt role in this as the Supreme Court has endorsed that public desire for retribution is a legitimate basis for establishing harsh, punitive judgments up to and including capital punishment .
Initial steps in adopting more effective strategies may be fostered by increasing numbers of people reconsidering the implicit relevance and meaning of concepts like free will for achieving societies' goals. As the meaning of this concept evolves and our understanding of behavior integrates more deterministic features, we are less compelled to frame maladaptive, antisocial actions within paradigms that embrace elusive immaterial origins (like evil, for instance). Rather, we are better equipped to recognize the influence of pathology, environment, and acquired maladaptive cognitive strategies in promoting antisocial behavior, where the levers of justice have considerably more remedial influence. After all, pathology is a more tractable problem than is evil. Responding to antisocial behavior, then, becomes a more pragmatic issue, and more progressive strategies aimed at addressing objective moderators of such behavior can be readily explored. In this way, even slow shifts in normative judgments are highly relevant to the way we assess culpability as a society, and the degree to which we view punitive measures as achieving their intended purpose as a remedy against undesired, antisocial behavior.
Neuroscience ultimately provides a useful platform for advocating new strategies of social management, where old strategies have perhaps proven ineffective or inefficient. New strategies may be less oriented toward retribution, per se, and more driven by practical concerns serving society with more efficient and productive solutions. Such strategies may, for instance, be aimed at better serving the mental health and social needs of those who come in contact with the justice system, reducing long-term incarceration rates for low risk offenders, and reducing recidivism by improving rehabilitative and reintegration efforts. In the worst scenarios, where rehabilitative interventions may not be a realistic goal, neuroscience also provides a platform for improving our predictions of future dangerousness. Such strategies may be integrated for making better decisions about those who need to be removed from society permanently. Before addressing this further, it will be important to consider a few examples for how advancing science has already changed the way we think about culpability and make decisions about interventions and punishment as a society.
## Limits on capital punishment
Torture and execution have been legally sanctioned forms of punishment since at least the 18th century BCE, as it is indicated in the Code of Hammurabi (c.1750 BCE) for such crimes as burglary, adultery, making false accusations, and poor construction of a house. Even within the history of the United States, the use of capital punishment has changed considerably, formerly implemented in cases of burglary, counterfeiting, and treason among others. Beginning with the adoption of bans on cruel and unusual punishment 6 ., modern societies (including the U.S.) have gradually changed their views on behavior deserving the harshest penalties, limiting its application for the most egregious crimes and even further to individuals most deserving of harsh punishment. Determining who deserves the harshest punishments has a great deal to do with our perceptions of their intentions, malice, and reasonable expectations of self-control. As we will see, these judgments also incorporate the relative utility of the punishment for fulfilling an intended punitive role. The relevance of neuroscience in drawing conclusions about these issues increases as their evaluation increasingly incorporate reductionist, determinist, biological perspectives of motivated behavior.
Prominent examples of this have come in the form of supreme court decisions accompanying restricted applications of the death penalty. For instance, Atkins v. Virginia 7 . ruled to prevent the execution of those with severe intellectual disabilities, citing "evolving standards of decency that mark the progress of a mature society." In cases like these, these evolving standards refer more specifically to what the court witnesses as a consensus among other jurisdictions and the way they have tended to interpret and enforce the law in recent history. Many states, for instance, had previously outlawed the execution of those with severe intellectual disabilities prior to these proceedings. Among the topics discussed in the formal ruling is the sentiment that those with reduced intellectual capacity have limitations in their adaptive functioning, reasoning, communication, and understanding of events around them and the actions of others. Thus, leveraging the most severe of punishments fails to align with the practical concerns of retribution and deterrence.
Similarly, Roper v. Simmons (2005) 8 . abolished capital punishment of juveniles citing similar "evolving standards" and an emerging consensus among other jurisdictions. In this case, however, the decision was also influenced in part by neuroscience research (including fMRI evidence) presented in an amicus brief by the American Psychological Association, suggesting that psychological deficits germane to adolescence (developmental limitations) make young people more prone to impulsive behavior and less capable of the highest order decision-making we ordinarily attribute to adults. That is, opinions informed by progress in neuroscience suggesting a limited capacity for behavioral control are influential for evaluating an individual's culpability (i.e., how harsh a punishment is justified). Implicit in the developmental perspective applied is an acknowledgment of the capacity for ongoing change. MRI evidence was also presented (in amicus brief) for consideration in Graham v. Florida (2011) 9 . which determined it unconstitutional to sentence juveniles to life without parole for crimes not involving homicide.
These decisions can be fully reconciled with normative attitudes about responsibility and determinism. Despite being fully responsible for their behavior, biological limitations on individuals' executive functioning and inherent capacity for change play a prominent role in our consideration of how harsh their punishments ought to be. The Court's decision in Roper v. Simmons affirmed that juveniles have less culpability due to their immature development, making them less deserving of the harshest punishments. These decisions do not imply that, as a society, we are any less interested in protecting ourselves from dangerous people or ensuring the safety of free citizens. What is confirmed in these decisions is a relative diminution in our motivation to levy harsh retributivist judgments in contexts where we recognize deterministic limitations in individual culpability. This, of course, opens the door to consider how we judge those with other biological limitations in cognitive functioning, or those disadvantaged in other ways.
## The insanity defense
Our collective understanding of culpability has almost always included provisions for certain disadvantages. A clear illustration of this endures in the limitations on culpability levied against those with serious mental disorders. This has been a common feature of many ancient legal systems and customs, including elements of Roman law which were carried forward in pre-Norman England. For instance, it was at times customary for juries to find insane criminals guilty, but refer them to the king for subsequent pardoning. More contemporary applications of these provisions give juries specific guidelines for applying these judgments directly. The M'Naghten Rule, for instance, formalized a set of conditions in English law that could be applied more consistently following a controversial acquittal. In 1843, Daniel M'Naghten suffered paranoid delusions and murdered a civil servant, mistaking him for the English Prime Minister. He was acquitted for murder based on substantial evidence that he was mentally ill, and he was forcibly committed to an asylum, where he spent the rest of his life. Despite the very real limits placed on his freedom, the ensuing public dissent following a not-guilty verdict (and official condemnation of the verdict by the queen) compelled establishing a set of explicit requirements for instantiating criminal insanity. These guidelines, in some adapted form, are still prevalent in many jurisdictions across the world today. They essentially require (for an insanity defense) that a defendant be so mentally impaired as to not know what they are doing and/or not know right from wrong.
Following in step with the very impetus for M'Naghten, many subsequent adaptations and amendments to these rules have been applied, usually following controversial rulings. As a result, several variations and alternative defenses have been enacted in state and federal jurisdictions. These either amend the essential language of M'Naghten used to describe what constitutes insanity, or they shift the burden of proof in important ways. For instance, in Parsons v. State of Alabama (1887) 10 ., an appeal was made following the controversial conviction of Nancy Parsons who killed her husband under the delusion that he had cast an evil spell on her. The court established a provision for instances in which a defendant may be deemed insane, despite knowing right from wrong. The ruling described instances where a disease has "destroyed the defendant's free will" and became known as the Irresistible Impulse defense. Other important developments have included modifications that specifically exclude antisocial personality disorder from an exculpatory mental illness, since its symptoms are primarily manifest through repeated criminal conduct 11 . There has also been a formal shift of responsibility from the prosecution -proving beyond a reasonable doubt that a defendant was sane -to the defense, which must prove (by preponderance of evidence) that the defendant was insane (Insanity Defense Reform Act; IDRA, 1984) 12 . A lengthy, stand-alone review would be necessary to adequately review the many important modifications that have been made to these rules over time and across many jurisdictions; however, an overarching pattern is apparent in this complex history. Through many shifts of language and interpretations, we continually re-affirm the preservation of limitations on culpability for those impaired by mental illness. We also betray the cognitive dissonance this instills against the backdrop of our most basic retributive motivations, and our sensitivities to potential abuses of these provisions.
As noted above, many of these changes come on the heels of controversial, high-profile cases. Consider for instance the trial of Dan White for the murder of San Francisco Mayor George Moscone and city supervisor Harvey Milk. Despite substantial evidence of premeditation and malice in the killings, White was ultimately convicted of voluntary manslaughter rather than first-degree murder, and served only 5 years in prison. This outcome was aided by what is still disparagingly referred to as "The Twinkie Defense." To this day, popular retellings of this case often reinforce a narrative that White's defense asserted his behavior was the result of eating sugary snacks, including Twinkies. In reality, psychiatrists testified that White suffered from major depression and had diminished capacity for controlling his behavior due to this pathology. An incidental detail of his diminished capacity included recent poor dietary habits, despite having been extremely health conscious all his life. The public outcry following his alarmingly lenient sentence was instrumental in abolishing the "diminished capacity" defense in California. The relevance of this trial for our present arguments is not so much to draw attention to the trial and defense, but rather what happened afterward. Following a defense which hinged on severe depression, White served a relatively lenient sentence at Soledad State Prison in California (not a secure hospital, or institute known for therapeutic intervention). Two years after his release from prison, Dan White committed suicide.
These events raise many interesting questions from a legal, psychological, and philosophical perspective. Did White ultimately get what he deserved? Did those seeking harsher retributive action find some gratification in his suicide, or is it inherently less satisfying that White's death was not carried out in a punitive context? Was White's case a greater failure in the basic judicial process of determining culpability, or more of a failure in enforcing effective interventions following a determination of his mental illness? Those like White, committing serious offenses (e.g., homicide) in the throes of mental illness, are typically forcibly committed to secure institutions with some focused psychiatric capacity, and do not generally go free in such a short time. From a utilitarian perspective, this form of intervention seems reasonable. It serves the role of protecting society from dangerous people and arguably remains a visible deterrent, while coupling offenders' containment with therapeutic and/or rehabilitative attention. Where this strategy fails is perhaps only limited in satisfying an instinctual urge to serve harsh retributive actions against those that have harmed us personally and/or violated our most sacred moral values.
It matters that this trial has largely been enshrined as a miscarriage of justice, but probably for many of the wrong reasons (as evinced by history's retelling of the "Twinkie Defense"). In some ways eradication of the diminished capacity defense serves as a scapegoat that only distracts us from more fundamental issues in our society and our justice system that are slow to change. Essentially, we still struggle to balance our shifting attitudes of culpability against a stubborn instinct to enact harsh retributive penalties in cases of egregious tragedy. After all, various provisions for mental illness in sentencing still exist in virtually all jurisdictions. Despite fine tuning the language of these rules, we (as a society) have steadfastly acknowledged that criminal actions occurring due to factors outside the ordinary limits of one's control deserve more leniency or a categorically different form of intervention than simple retribution. In order to see the potential benefit of progressive changes in jurisprudence, however, our corrections systems and forensic psychiatric facilities need to be equipped with the tools to enact these changes in ways that demonstrate more satisfying results.
## How the legal system may (continue to) change
The contexts described above illustrate that our normative views of culpability have never been static, but continually adapt to evolving standards ushered in by a more refined understanding of human behavior and the boundaries of our own agency. Neuroscience, surely, does not make this progress simpler. On the contrary, as our understanding of biology's role in promoting pathology and maladaptive behavior increases, this encourages a far more nuanced interpretation of culpability in the face of various advantages and disadvantages. Shifting attitudes on this front have fostered an expansion of contexts that we harbor special provisions for in the law. However, rather than promoting overly exculpatory attitudes (a kind of straw man common in arguments diminishing the role of neuroscience in law), these shifts have largely required changes only in the way we intervene and balance corrective and/or punitive measures in such contexts. We suggest it is reasonable to expect these trends to continue as retributive goals are softened and we aim to integrate more practical solutions for addressing criminogenic needs, improving reintegration, and reducing recidivism.
Using offender age as a model for such a transition, the criminal justice system has recently made provisions to limit harsh sentencing (capital punishment/life in prison) of juvenile offenders in most circumstances, but various jurisdictions still apply somewhat arbitrary rules about the cut-offs for these provisions. Certainly young offenders are not all equal from a neurodevelopmental perspective. So does it make sense to apply a bright line rule allowing capital punishment on someone's 18th birthday? Neuroscience may continue to inform this perspective, expanding a more nuanced evaluation. Recent research, in fact, has demonstrated that a brain-derived measure of gray matter related to age is a better predictor of future antisocial behavior than is chronological age. As such, it may be a more pertinent question to consider the relative advantages in development and mental health with which one is equipped before deciding whether they deserve our harshest punishments. Trends in this direction are encouraged by the bifurcation of guilt and sentencing phases of some criminal trials. Another perspective to consider is what effort and resources are justified in the aim of preserving and enacting capital punishment as an extreme punitive measure for rare circumstances. Studies on the deterrence effects of the death penalty are equivocal at best, and economic scrutiny suggests that we may be incapable of enacting the death penalty in any reasonably efficient manner such that it serves its intended purposes. Despite our enduring retributivist instincts, we may eventually decide that abolition of the death penalty represents a more practical solution, obviating some of our more difficult choices when it comes to punishment. But capital punishment is not the only context within which shifting attitudes may promote more pragmatic strategies.
In the case of less severe sentences, we (as a society) have demonstrated more amenability to the potential value of remedial approaches and possible re-integration of young offenders. Certainly, more aggressive treatment strategies integrating contemporary cognitive-behavioral approaches for improving long-range outcomes have proven both successful and costeffective. Consider for instance progressive treatment programs being instituted at the Mendota Juvenile Treatment Center among high-risk young offenders. Analyses have indicated better outcomes and relatively less economic burden on society by enacting these aggressive treatment strategies. To be sure, there will always be those who are resistant to our best available treatments at any given time, and unable to return safely to free society. However, this further reinforces the value of pursuing new and better strategies informed by ongoing research addressing the origins, development, and maintenance of maladaptive, antisocial behavior.
In assessing how neuroscience may continue to inform judicial decision-making in the future, many possibilities arise. Could brain scans that objectively quantify one's neurodevelopment or functional capacity eventually be used to determine whether one is tried as an adult or juvenile? Could predictive models determine whether one is amenable to therapeutic attention or is likely to remain resistant to available rehabilitative efforts? Could neuroscientific measures reveal specific risk factors for re-offending that are not evident on standard psychiatric assessments? These are difficult questions indeed, and ones that we do not yet have answers for. We simply argue that to ignore them or to undermine their potential value only to reinforce the status quo seems myopic and overtly servile toward an imperfect system. As neuroscience ushers in a more complete bio-psychosocial understanding of maladaptive behavior, and as ongoing incarceration strategies become unsustainable, our prediction is that we will be forced to consider alternative approaches that serve public interest in more pragmatic ways. This will involve wider application of therapeutic, rehabilitative approaches and more aggressive therapeutic and reintegration strategies that reduce the likelihood for recidivism. Such applications may be particularly effective among young offenders. This may also include better risk assessment in making decisions about sentencing and parole. Major advances on these fronts may only require us to first suspend our most basic retributivist instincts when addressing social problems, and remain open minded about the potential for more prudent strategies. Neuroscience doesn't fill these roles on its own, but it provides a platform for advancing each of these goals through empirical research and improved knowledge.
# Author contributions
NA developed the primary theses and arguments presented in this review. KK provided additional insights, commentary, and editorial remarks.
# Acknowledgments
Commentaries contained in this review reflect the thoughts and opinions of the authors only, and do not reflect official strategies or priorities of The Mind Research Network, funding bodies, state/federal organizations, or other supporters of our research.
United States v. Hinckley, 525 F. Supp. 1342 (D.C. 1981).Frontiers in Psychology | www.frontiersin.org
March 2020 | Volume 11 | Article 390
Frontiers in Psychology | www.frontiersin.org
Gregg v. Georgia, 428 U.S. 153, 96 S. Ct. 2909, 49 L. Ed. 2d 859 (1976.Frontiers in Psychology | www.frontiersin.org
U.S. Const. amend. VIII. 7 Atkins v. Virginia, 536 U.S. 304, 122 S. Ct. 2242, 153 L. Ed. 2d 335 (2002). 8 Roper v. Simmons, 543 U.S. 551, 125 S. Ct. 1183, 161 L. Ed. 2d 1 (2005).
Graham v. Florida, 560 U.S. 48, 130 S. Ct. 2011, 176 L. Ed. 2d 825 (2010.Frontiers in Psychology | www.frontiersin.org
Parson v. State, 81 Ala. 577, 2 So. 854, 2 So. 2d 854 (1887). 11 American Law Institute: Model Penal Code. Philadelphia: ALI, (1962) Ref. 11, § 4.01 12 Insanity Defense Reform Act ("IDRA"), 18 U.S.C. § 17(b) (1984).
We would like to thank the staff and administration of the New Mexico Corrections Department and Wisconsin Department of Corrections for their continued cooperation and support with ongoing research at the Mind Research Network. We also thank the volunteers participating in research and the research staff that make our research possible.Conflict of Interest:The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.Copyright © 2020 Anderson and Kiehl. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms. |
HbA1c Levels Are Associated with Chronic Kidney Disease in a Non-Diabetic Adult Population: A Nationwide Survey (KNHANES 2011–2013)
BackgroundMany studies have reported an association between glycated hemoglobin A1c (HbA1c) and metabolic syndrome (MetS) in non-diabetes patients. Each component of MetS is in fact related to chronic kidney disease (CKD) incidence and progression. Therefore, HbA1c in non-diabetic mellitus (DM) may be intrinsically associated with the prevalence of CKD. The hypothesis of the present study was that high HbA1c in non-DM patients is associated with CKD.Patients and MethodsThe total number of participants in this study was 24,594. The participants were divided into three groups according to their HbA1c levels: a Low group (<5.7% or <39 mmol/mol), a Middle group (5.7-6.0% or 39-42 mmol/mol), and a High group (>6.0% or >42 mmol/mol). The estimated glomerular filtration rate (eGFR) was calculated using the Chronic Kidney Disease Epidemiology Collaboration equation.ResultsThe number of participants allocated to the Low, Middle, and High groups was 8,651, 4,634, and 1,387, respectively. Linear regression analyses were performed to evaluate the association between variables. Standardized β ± standard error was 0.25 ± 0.22 for waist circumference, 0.44 ± 0.20 for fasting glucose, -0.14 ± 0.30 for high-density lipoprotein cholesterol levels, 0.15 ± 2.31 for triglyceride levels, 0.21 ± 0.00 for systolic blood pressure, 0.10 ± 0.00 for diastolic blood pressure, and -0.22 ± 0.42 for eGFR (P < 0.001 for all variables). eGFR in non-diabetes participants was inversely associated with the HbA1c level, where eGFR decreased as HbA1c levels increased. Standardized βs were -0.04 ± 0.42 in multivariable PLOS ONE |
analysis (P < 0.001). The proportion of participants with only MetS, only CKD, or both MetS and CKD was higher in the High group than in the Low and Middle groups.
# Conclusion
High HbA1c in non-DM patients may be associated with CKD. Renal function in patients with high HbA1c levels may need to be monitored.
# Background
Chronic kidney disease (CKD) is a widely recognized public health issue and associated with high morbidity and mortality when compared to the non-CKD population [bib_ref] Chronic kidney disease and the risks of death, cardiovascular events, and hospitalization, Go [/bib_ref] [bib_ref] Renal replacement therapy in Korea, Jin [/bib_ref]. The United States Real Data System 2014 Annual Data Report showed that CKD occurs in approximately 13.6% of the general population. Indeed, the prevalence of CKD appears to be rising rapidly with increased life expectancy. Overall Medicare expenditures for CKD were $44,581 million in 2012. Screening for and effective monitoring of CKD are essential for increasing patient quality of life and decreasing the public health burden.
Glycated hemoglobin (HbA1c) is an important indicator for long-term glucose control and has recently been recommended for use in the diagnosis of diabetes mellitus (DM) by the American Diabetes Association (ADA). However, the use of HbA1c for identifying pre-diabetes is a controversial topic [bib_ref] Pre-diabetes, metabolic syndrome, and cardiovascular risk, Grundy [/bib_ref]. In 2015, the ADA suggested that an HbA1c of 5.7-6.4% (39-46 mmol/mol) is reasonable for the diagnosis of pre-diabetes and that patients with HbA1c > 6.0% (>42 mmol/mol) should be considered to be at very high risk for DM. Although the clinical significance of HbA1c as a surrogate marker of metabolic syndrome has not yet been fully examined, many studies have reported an association between HbA1c and MetS in non-DM patients [bib_ref] Is glycosylated hemoglobin A1c a surrogate for metabolic syndrome in nondiabetic, first-degree..., Osei [/bib_ref] [bib_ref] Using glycosylated hemoglobin to define the metabolic syndrome in the United States..., Ong [/bib_ref] [bib_ref] Usefulness of glycated hemoglobin as diagnostic criteria for metabolic syndrome, Park [/bib_ref]. Each component of MetS is in fact related to CKD incidence and progression [bib_ref] The metabolic syndrome and chronic kidney disease in U.S. adults, Chen [/bib_ref]. Therefore, HbA1c in non-DM may be intrinsically associated with the prevalence of CKD. The aim of the present study was to evaluate the clinical association between HbA1c and CKD in non-DM patients. The hypothesis of the present study was that high HbA1c in non-DM patients is associated with CKD.
# Patients and methods
## Study population
Data from the Korean National Health and Nutrition Examination Survey (KNHANES 2011-2013) were used for this analysis. The KNHANES is a nationwide, multi-stage, stratified survey of a representative sample of the South Korean population and is conducted by the Korea Centers for Disease Control and Prevention. The total number of participants from KNHANES analyzed in this study was 24,594. Participants were excluded from the present study based on the following criteria: data could not be provided for HbA1c (n = 2,350) or renal function (n = 2) or participants were younger than 18 years of age (n = 5,385) or had DM (defined as a self-reported history of a DM diagnosis, a fasting glucose level of 126 mg/dL, or HbA1c 6.5% (48 mmol/mol; n = 2,185). As a result, 14,672 participants were ultimately included in this study. Ethical approval for this study was obtained from the institutional review board of Yeungnam University Hospital (2015-04-004). The board waived the need for informed consent, as the subjects' records and information were anonymized and de-identified prior to analysis.
## Study variables
Clinical and laboratory data collected during clinical examination included the following: age, sex, serum creatinine (mg/dL), body mass index (BMI, kg/m 2 ), waist circumference (WC, cm), HbA1c (%, mmol/mol), fasting blood glucose (mg/dL), total cholesterol (mg/dL), high-density lipoprotein (HDL) cholesterol levels (mg/dL), triglyceride levels (mg/dL), systolic blood pressure (mmHg), diastolic blood pressure (mmHg), smoking status, alcohol intake, and levels of physical activity.
HbA1c levels were measured using a high performance liquid chromatography system (HLC-723G7; Tosoh Co., Tokyo, Japan). In the present study, the participants were divided into three groups according to their HbA1c levels: a Low group (<5.7% or <39 mmol/mol), a Middle group (5.7-6.0% or 39-42 mmol/mol), and a High group (>6.0% or >42 mmol/mol). Serum creatinine levels were measured using a Hitachi Automatic Analyzer (alkaline picrate, Jaffé kinetic). The estimated glomerular filtration rate (eGFR) was calculated using the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) equation [bib_ref] A new equation to estimate glomerular filtration rate, Levey [/bib_ref]. CKD was defined as an eGFR <60 mL/min/1.73 m 2 . Urine albumin level was measured from random samples using a turbidimetric immunoassay (Hitachi Automatic Analyzer 7600, Hitachi). Urine creatinine level was measured using a colorimetric method (Hitachi Automatic Analyzer 7600, Hitachi). Urine albumin and creatinine concentrations were measured in the same laboratory for all surveys. The inter-assay coefficient of variation for all laboratory work was consistenly low (<3.1%). The urine albumin-creatinine ratio (UACR) was calculated in mg per g of creatinine (mg/g). Albuminuria was defined as UACR 30 mg/g.
Patients were classified according to smoking status as current smokers, ex-smokers, or non-smokers. Alcohol intake was defined using the Korean version of 'standard drinking' based on the WHO classification [bib_ref] Comparative Quantification of Health Risks: global and regional burden of disease attributable..., Rehm [/bib_ref]. Alcohol intake was classified into 3 categories: abstinence (no consumption of alcohol within the last year); moderate drinking (women: 0.1-19.99 g pure alcohol/day; men: 0.1-39.99 g pure alcohol/day), and heavy drinking (women: 20 g pure alcohol/day; men: 40 g pure alcohol/day). Physical activity was assessed by the presence of exercise. The presence of exercise was defined as moderate activity for more than 30 min/day, for 5 days/week or intense activity for more than 20 min/day, for 3 days/week, or walking more than 30 min/day, for 5 days/week. Coronary artery disease (CAD) was defined as a self-reported history of angina or myocardial infarction. Cerebrovascular accident (CVA) was defined as a self-reported history of stroke.
MetS was defined according to the Adult Treatment Panel III criteria using the modified cutoff values for Asian populations as suggested by the Asia-Pacific guidelines [bib_ref] Diagnostic and management of the metabolic syndrome: an American Heart Association/National Heart,..., Grundy [/bib_ref] [bib_ref] The metabolic syndrome-a new worldwide definition, Alberti [/bib_ref]. Briefly, elevated blood glucose was defined as a fasting blood glucose level 100 mg/dL or a selfreported history of DM. Elevated blood pressure was defined as a systolic or diastolic blood pressure 130/85 mmHg and a self-reported history of hypertension. A low HDL cholesterol level was defined as <40 mg/dL in men and <50 mg/dL in women. Elevated triglyceride levels were defined as a serum triglyceride level of 150 mg/dL. Abdominal obesity was defined as a WC >90 cm in men and >80 cm in women. MetS was defined as the presence of 3 components of MetS.
## Statistical analyses
The data were analyzed using the Statistical package for the Social Sciences software package (SPSS v.21, Chicago, IL., USA). Categorical variables were expressed as both counts and percentages. Continuous variables were expressed as the mean ± standard deviation (SD). UACR was a non-parametric variable expressed as a median (95% CI), and was compared using the Kruskal-Wallis test. The Pearson's χ 2 or Fisher's exact test was used to analyze categorical variables. For continuous variables, means were compared using a one-way analysis of variance. Correlations were analyzed in order to assess the strength of the relationship between continuous variables. Linear regression analysis was performed to assess independent predictors of eGFR or number of MetS components. Variance inflation factor was used to identify multicollinearity for the multivariable linear regression model. Variance inflation factor greater than 10 was not accepted. Logistic regression analyses were used for estimating the odds ratios (OR) and 95% confidence intervals (CI), which were then applied towards determining the relationship between HbA1c and CKD or MetS.
Confounders were defined by their likelihood of preceding or contributing to the development of MetS or CKD. The selection of confounder was based on previous literatures [bib_ref] Dairy consumption and risk of metabolic syndrome: a meta-analysis, Kim [/bib_ref] [bib_ref] KHA-CARI guideline: Early chronic kidney disease: detection, prevention and management, Johnson [/bib_ref]. For MetS, the covariates were HbA1c, age, sex, BMI, alcohol intake, smoking status, and physical activity. For CKD, the covariates were HbA1c, age, sex, BMI, alcohol intake, smoking status, physical activity, CAD, CVA, WC, HDL cholesterol levels, triglyceride levels, systolic blood pressure, and diastolic blood pressure. Discrimination-which is the ability of the model to differentiate between participants who have CKD or MetS and those who do not-was examined using the area under the receiver operating characteristic (AUROC) curve. AUROC analysis was also performed in order to calculate cutoff values, sensitivity, and specificity. Optimal cutoff risk point was defined as the maximum Youden index in the AUROC. The AUROC was calculated using the MedCalc software package (v.11.6.1.0, MedCalc, Mariakerke, Belgium). We also calculated the integrated discrimination improvement (IDI) and the net reclassification improvement (NRI) with a category-free option among models, following the methodology of Penica et al. [bib_ref] Evaluating the added predictive ability of a new marker: from area under..., Pencina [/bib_ref] [bib_ref] Extensions of net reclassification improvement calculations to measure usefulness of new biomarkers, Pencina [/bib_ref]. A restricted cubic spline curve was used to evaluate non-linear relationships between the HbA1c level and CKD, which was adjusted for age and sex. The restricted cubic spline curve was plotted using statistical software SAS version 9.4 (SAS Campus Drive, Cary, NC, USA). A P-value less than 0.05 was considered statistically significant.
# Results
## Clinical characteristics of participants
The number of participants allocated to the Low, Middle, and High groups was 8,651, 4,634, and 1,387, respectively [fig_ref] Table 1: Clinical characteristics of participants by HbA1c level [/fig_ref]. Age, BMI, WC, eGFR, total cholesterol, fasting blood glucose, triglyceride levels, and systolic and diastolic blood pressure were higher in the High group than either the Low or Middle group.
The proportion of participants with only MetS in the Low, Middle, and High groups was 9.1%, 20.4%, and 33.9%, respectively (P < 0.001), whereas the proportion of participants with only CKD in the Low, Middle, and High groups was 0.9%, 2.0%, and 3.5%, respectively (P < 0.001). The proportion of participants with both MetS and CKD in the Low, Middle, and High groups was 0.2%, 0.7%, and 2.0%, respectively (P < 0.001). The proportion of participants with only MetS, only CKD, or both MetS and CKD was higher in the High group than in the Low and Middle groups.
## Association between hba1c level and mets or ckd
We performed univariate linear regression analyses to evaluate the association between HbA1c and each MetS components. Standardized β ± standard error was 0.25 ± 0.22 for WC, 0.44 ± 0.20 for fasting glucose,-0.14 ± 0.30 for HDL cholesterol levels, 0.15 ± 2.31 for triglyceride levels, 0.21 ± 0.00 for systolic blood pressure, and 0.10 ± 0.00 for diastolic blood pressure (P < 0.001 for all variables). There were positive associations between HbA1c levels and WC, fasting glucose, triglyceride levels, systolic blood pressure, and diastolic blood pressure, and negative association between HbA1c levels and HDL cholesterol levels. In addition, HbA1c in non-diabetes participants was associated with the number of MetS components observed (S1 . Numbers of MetS components increased in accordance with increased HbA1c levels.
Univariate and multivariable linear regression analyses were also performed to evaluate the association between HbA1c level and eGFR (S1 . eGFR in non-diabetes participants was inversely associated with the HbA1c level, where eGFR decreased as HbA1c levels increased.
Logistic regression showed that the OR for only MetS with a 1% (11 mmol/mol) increase in HbA1c was 7.53 (95% CI, 6.51-8.70) in univariate analysis and 3.38 (95% CI, 2.85-4.00) in multivariable analysis (S2 . The OR for only CKD with a 1% (11 mmol/mol) increase in HbA1c was 9.32 (95% CI, 6.16-14.11) in univariate analysis and 2.13 (95% CI, 1.33-3.40) in multivariable analysis. The OR for both MetS and CKD with a 1% (11 mmol/mol) increase in the level of HbA1c was 21.49 (95% CI, 10.86-42.52) on univariate analysis and 4.12 (95% CI, 1.80-9.39) on multivariable analysis. A restricted cubic spline curve was plotted, with 5.6% (38 mmol/mol) as the median HbA1c level, and it was adjusted for age and sex (S1 A high HbA1c level was associated with increased OR for CKD.
To estimate the incremental value of HbA1c level to predict only MetS, only CKD or both MetS and CKD, we compared the probabilities of events and nonevents of models using relative IDI and category-free NRI (S3 . The IDI of adding HbA1c level to the multivariable [fig_ref] Fig 1: Receiver operating characteristic curves of HbA1c for the prediction of metabolic syndrome... [/fig_ref]. Sensitivity and specificity for predicting only MetS were 56.7% and 74.2%, respectively. Those for predicting only CKD were 64.2% and 63.8%, respectively, while those for predicting both MetS and CKD were 65.1% and 74.2%, respectively.
## Association between hba1c level and uacr
In participants with eGFR 60 mL/min/1.73 m 2 , the correlation coefficient between UACR and HbA1c was 0.043 (P < 0.001). UACR in the Low, Middle, and High groups was 9.2 (95% CI, 8.1-10.2), 11.5 (95% CI, 9.9-13.1), and 15.9 (95% CI, 12.6-19.1), respectively (P < 0.001). The portion of participants with albuminuria in the Low, Middle, and High groups was 329 (4.3%), 239 (5.6%), and 104 (8.2%), respectively (P < 0.001).
# Discussion
In the present study, a clear association was observed between HbA1c and MetS in non-DM Asian patients, which is in line with numerous other studies that have shown an association between these two variables [bib_ref] Is glycosylated hemoglobin A1c a surrogate for metabolic syndrome in nondiabetic, first-degree..., Osei [/bib_ref] [bib_ref] Using glycosylated hemoglobin to define the metabolic syndrome in the United States..., Ong [/bib_ref] [bib_ref] Usefulness of glycated hemoglobin as diagnostic criteria for metabolic syndrome, Park [/bib_ref] [bib_ref] The metabolic syndrome and chronic kidney disease in U.S. adults, Chen [/bib_ref] [bib_ref] Glycosylated hemoglobin and prevalent metabolic syndrome in nondiabetic multiethnic U.S. adults, Veeranna [/bib_ref] [bib_ref] Distribution and cardiovascular risk correlates of hemoglobin A(1c) in nondiabetic younger adults:..., Nguyen [/bib_ref] [bib_ref] Glycated haemoglobin as a predictor for metabolic syndrome in non-diabetic Korean adults, Sung [/bib_ref] [bib_ref] Association of A1C with cardiovascular disease and metabolic syndrome in Asian Indians..., Dilley [/bib_ref] [bib_ref] Association of hemoglobin A1c with cardiovascular disease risk factors and metabolic syndrome..., Kim [/bib_ref]. Studies aiming to investigate this association should exclude patients with DM as this condition is a critical confounding factor for the prevalence of MetS and certain studies have reported no exclusion of patients with HbA1c 6.5% (48 mmol/mol) [bib_ref] Glycated haemoglobin as a predictor for metabolic syndrome in non-diabetic Korean adults, Sung [/bib_ref] [bib_ref] Association of hemoglobin A1c with cardiovascular disease risk factors and metabolic syndrome..., Kim [/bib_ref]. While a few previous studies did exclude DM patients with HbA1c 6.5% (48 mmol/mol), the majority of these were single-center studies with a possibility of selection bias [bib_ref] Usefulness of glycated hemoglobin as diagnostic criteria for metabolic syndrome, Park [/bib_ref] [bib_ref] Glycated haemoglobin as a predictor for metabolic syndrome in non-diabetic Korean adults, Sung [/bib_ref] [bib_ref] Association of hemoglobin A1c with cardiovascular disease risk factors and metabolic syndrome..., Kim [/bib_ref]. The present study analyzed a nationwide, multi-stage, stratified survey of a representative sample of the South Korean population and excluded patients with HbA1c 6.5% (48 mmol/mol). Results were adjusted for variable confounders and revealed that HbA1c in non-DM patients is associated with the number of MetS components. The linear regression analyses did show an association between HbA1c and each component of MetS as continuous variables.
The present study showed an association between the HbA1c level and CKD with or without MetS. The associations between insulin resistance and CKD are very complex and not clear. Previous studies have shown that each component of MetS is associated with development and progression of CKD. Among the components of MetS, insulin resistance may be the most important related etiological factor for CKD [bib_ref] Metabolic syndrome and chronic kidney disease: Current status and future directions, Prasad [/bib_ref]. HbA1c is an indicator predicting insulin resistance. High HbA1c level in pre-diabetes is associated with insulin resistance or metabolic syndrome, which can lead to development and progression of CKD. Our results suggest that high HbA1c is mainly associated with insulin resistance, which may result in development of CKD. However, CKD results in interference with the intracellular signaling pathway initiated by insulin, which results in insulin resistance [bib_ref] Molecular mechanisms of insulin resistance in chronic kidney disease, Thomas [/bib_ref].
The literature has shown conflict results concerning an association between HbA1c and CKD. Certain studies have shown that HbA1c is associated with development of CKD in non-DM patients [bib_ref] The relationship between dysglycaemia and cardiovascular and renal risk in diabetic and..., Gerstein [/bib_ref] [bib_ref] Epidemiology of chronic kidney disease: results from a population of older adults..., Zhang [/bib_ref] [bib_ref] Prevalence of chronic kidney disease in US adults with undiagnosed diabetes or..., Plantinga [/bib_ref] [bib_ref] Association of HbA1c and cardiovascular and renal disease in an adult Mediterranean..., Hernandez [/bib_ref]. Gerstein et al. conducted a prospective study with an average 4.5-year follow-up and successfully showed that HbA1c is associated with development of overt nephropathy defined by albuminuria or proteinuria [bib_ref] The relationship between dysglycaemia and cardiovascular and renal risk in diabetic and..., Gerstein [/bib_ref]. Zhang et al. evaluated a cross-sectional study using a German cohort and showed an association between HbA1c and eGFR or CKD defined as eGFR < 60 mL/min/1.73 m 2 [bib_ref] Epidemiology of chronic kidney disease: results from a population of older adults..., Zhang [/bib_ref]. Although DM was adjusted for in multivariable analysis, this study did include DM patients. A study by Plantinga et al. enrolled non-DM patients using data from the USA; however, this group evaluated the association between CKD and pre-diabetic status classified by fasting glucose level [bib_ref] Prevalence of chronic kidney disease in US adults with undiagnosed diabetes or..., Plantinga [/bib_ref]. In contrast to the afore-mentioned studies, there have been reports that no association exists between the two variables if variable cardiovascular risk factors are adjusted for [bib_ref] Glycemic status and development of kidney disease: the Framingham Heart Study, Fox [/bib_ref] [bib_ref] Glycated hemoglobin and the risk of kidney disease and retinopathy in adults..., Selvin [/bib_ref] [bib_ref] Prognostic association of HbA1c and fasting plasma glucose with reduced kidney function..., Schöttker [/bib_ref] [bib_ref] Association of prediabetes by fasting glucose and/or haemoglobin A1c levels with subclinical..., Xing [/bib_ref]. Selvin et al. observed no significant difference in the development of CKD in patients with HbA1c 5.7-6.5% (39-48 mmol/mol) when compared to patients with HbA1c <5.7% (<39 mmol/mol) [bib_ref] Glycated hemoglobin and the risk of kidney disease and retinopathy in adults..., Selvin [/bib_ref]. However, that study did not include HbA1c as a diagnostic criterion for DM and could possibly include DM patients. To the best of our knowledge, this is the first study to evaluate the association between HbA1c and CKD in an Asian population. The results of the present study do show an association between HbA1c and eGFR as a continuous variable for renal function or CKD as a categorical variable. In addition, we calculated IDI and NRI as more advanced prediction analyses; these analyses showed that in comparison with multivariable models using only traditional risk factors, the addition of HbA1c to multivariable models improved both IDI and NRI. Unfortunately, cross-sectional study such as this cannot evaluate the causal relationship among these variables. Further prospective studies are needed to identify the causality between two variables.
The present study showed the association between HbA1c and albuminuria as a surrogate marker for early CKD. In participants with eGFR 60 mL/min/1.73 m 2 , UACR and the proportion of participants with albuminuria increased as HbA1c increased. Previous studies demonstrated that HbA1c level is associated with albuminuria in participants with DM [bib_ref] Albuminuria in type 2 (noninsulin-dependent) diabetes mellitus and impaired glucose tolerance in..., Nelson [/bib_ref] [bib_ref] arterial pressure and micro-albuminuria in type 1 (insulin-dependent) diabetes mellitus, Wiseman [/bib_ref] [bib_ref] 10-year follow-up of intensive glucose control in type 2 diabetes, Holman [/bib_ref] [bib_ref] Glucose control and vascular complications in veterans with type 2 diabetes, Duckworth [/bib_ref] [bib_ref] Intensive glucose control improves kidney outcomes in patients with type 2 diabetes, Perkovic [/bib_ref]. Poor glycemic control in DM plays a key role in rapid progression to diabetic nephropathy, which is caused by variable hemodynamic, metabolic, or endothelial dysfunction [bib_ref] Diabetic nephropathy-complications and treatment, Akh [/bib_ref]. Many previous studies have demonstrated pathophysiology or factors associated with progression to albuminuria in DM, but there have been few studies regarding the association between HbA1c and albuminuria in non-DM participants. The present study reveals that high-normal HbA1c levels previously considered to be in the normal range may be associated with albuminuria and may function as a marker for early CKD in non-DM participants.
The pathophysiology of the association between CKD and HbA1c in non-DM participants may be the same as that in high-glucose DM participants. This may ultimately result in subclinical or clinical atherosclerosis in various vessels. Glycemic control is a well-known risk factor for the development of atherosclerosis in DM participants. Previous studies also showed a positive association between prediabetes and atherosclerosis as measured by carotid intimal thickness, subclinical myocardial damage, or coronary artery calcium [bib_ref] The association between A1C and subclinical cardiovascular disease: the multi-ethnic study of..., Mcneely [/bib_ref] [bib_ref] Association of impaired fasting glucose and coronary artery calcification as a marker..., Moebus [/bib_ref] [bib_ref] Diabetes mellitus, prediabetes, and incidence of subclinical myocardial damage, Selvin [/bib_ref]. These pathologic changes can develop in the renal vasculature, which results in CKD with albuminuria.
Very low HbA1c level may be associated with malnutrition, inflammation, and atherosclerosis. However, in our study, a spline curve showed that low HbA1c level is not associated with CKD compared to median HbA1c level. Two factors may be associated with this discordance. First, malnutrition combined with low HbA1c level is common in participants with severe comorbidities, such as advanced cancer or end-stage renal disease, compared with the general population. Participants enrolled in our study may have been healthier than other selected populations who visited hospitals, which could have resulted in selection bias. Second, HbA1c may be used as a nutritional marker, but it mainly reflects glucose intake. The numbers of participants with total cholesterol < 100 mg/dL as another marker of malnutrition were 6, 7, and 0 in the Low, Middle, and High groups, respectively. There were few participants with malnutrition defined by total cholesterol level in our study.
This study has a number of limitations. First, it is a retrospective cross-sectional design and therefore cannot establish causality between two variables. Second, the available data did not include post-prandial blood glucose levels as a criterion for DM and a small number of DM patients could therefore have included. However, all participants have HbA1c < 6.5% (<48 mmol/mol) and a fasting blood glucose <126 mg/dL. Third, KDIGO guidelines define CKD as eGFR < 60 mL/min/1.73 m 2 for >3 months. In our study, CKD was defined using a single serum creatinine or single spot urine sample. However, the effect of these limitations will be reduced by the strength of a nation-wide representative sample.
In conclusion, high HbA1c in non-DM patients may be associated with CKD. Renal function in patients with high HbA1c levels may need to be monitored.
## Supporting information
[fig] Fig 1: Receiver operating characteristic curves of HbA1c for the prediction of metabolic syndrome or chronic kidney disease. A. Only metabolic syndrome. B. Only chronic kidney disease. C. Both metabolic syndrome and chronic kidney disease. [/fig]
[fig] S1: Fig.Adjusted restricted cubic spline curve showing odds ratio and 95% confidence interval (dashed line) for chronic kidney disease associated with HbA1c level (median value = 5.6% or 38 mmol/mol). Spline curve was adjusted for age and sex. (TIF) [/fig]
[table] Table 1: Clinical characteristics of participants by HbA1c level. Data are expressed as numbers (percentages) for categorical variables and mean ± standard deviations for continuous variables. *P values were tested by one-way analysis of variance for continuous variables and Pearson χ 2 test or Fisher exact test for the categorical variables. 693) for only CKD, and 0.760 (95% CI, 0.752-0.768) for both MetS and CKD (P < 0.001). The cutoff value was >5.7% (>39 mmol/mol) for only MetS, >5.6% (>38 mmol/ mol) for only CKD, and >5.7% (>39 mmol/mol) for both MetS and CKD [/table]
[table] Table: Linear regression analyses for the number of metabolic syndrome components or estimated glomerular filtration rate according to HbA1c level. (DOCX) S2 Table. Logistic regression analyses for metabolic syndrome or chronic kidney disease according to HbA1c level. (DOCX) S3 Table. AUCs, IDI, and NRI for multivariable models with or without HbA1c level. (DOCX) [/table]
|
Myricetin 3-O-β-D-Galactopyranoside Exhibits Potential Anti-Osteoporotic Properties in Human Bone Marrow-Derived Mesenchymal Stromal Cells via Stimulation of Osteoblastogenesis and Suppression of Adipogenesis
# Introduction
During osteoporosis progression, it was shown that the differential tendencies of bone marrow mesenchymal stromal cells (MSCs) shift towards adipocytes rather than osteoblasts which lead to a bone formation imbalance [bib_ref] Osteoporosis, differentiation of mesenchymal stem cells (MSCs) improves bone marrow adipogenesis, Pino [/bib_ref]. This imbalance was credited as the reason behind the porous and fragile bone structure seen in patients with osteoporosis. Treating osteoporosis was carried out with a broad range of approaches that target different steps of osteoporotic complications. To date, most of these treatments aimed to increase bone formation employing different types of drug classes such as bisphosphonates, parathyroid hormone, and selective estrogen receptor modulators. Studies have provided evidence that the differential regulation of bone marrow multipotent cells is directly linked with the bone structure and development [bib_ref] Osteoporosis, differentiation of mesenchymal stem cells (MSCs) improves bone marrow adipogenesis, Pino [/bib_ref] [bib_ref] Differentiation of bone marrow mesenchymal stem cells in osteoblasts and adipocytes and..., Wang [/bib_ref]. Considering the correlation between adipogenesis and osteoblastogenesis mechanisms of the MSCs, it was hypothesized that tilting the MSC differentiation towards osteoblasts could result in beneficial effects in osteoporotic provided evidence that the differential regulation of bone marrow multipotent cells is directly linked with the bone structure and development [bib_ref] Osteoporosis, differentiation of mesenchymal stem cells (MSCs) improves bone marrow adipogenesis, Pino [/bib_ref] [bib_ref] Differentiation of bone marrow mesenchymal stem cells in osteoblasts and adipocytes and..., Wang [/bib_ref]. Considering the correlation between adipogenesis and osteoblastogenesis mechanisms of the MSCs, it was hypothesized that tilting the MSC differentiation towards osteoblasts could result in beneficial effects in osteoporotic progression by decreasing adipose tissue and increasing bone formation. Regulation of MSC differentiation is handled by different signaling pathways which, however, have antagonistic relations when it comes to adipogenesis and osteoblastogenesis [bib_ref] Regulation of mesenchymal stem cell differentiation, Cook [/bib_ref]. Inducing MSCs into osteoblastogenesis occurs via activation of osteogenic signaling which at the same time suppresses adipogenic stimuli and vice versa. The main pathways that induce adipogenesis and osteoblastogenesis and in this kind of intertwined relation are peroxisome proliferator-activated receptor γ (PPARγ) and Wnt/β-catenin signaling, respectively [bib_ref] Cross-talking between PPAR and WNT signaling and its regulation in mesenchymal stem..., Xu [/bib_ref]. Inducing MSCs to differentiate into adipocytes through PPARγ activation inhibits the bone morphogenetic protein (BMP)-induced osteoblastogenesis while activation of Wnt/β-catenin for osteogenic differentiation concurrently hinders adipogenesis by suppressing PPARγ transcriptional activities.
Plant-based bioactive substances make up most of the reported natural products with promising beneficial effects. Among these substances, the notable ones that cover the biggest shares were flavonoids, coumarins, and terpenes possessing therapeutic properties for the symptoms of global diseases such as obesity, diabetes, cardiovascular diseases, and osteoporosis [bib_ref] Superfruits: Phytochemicals, antioxidant efficacies, and health effects-A comprehensive review, Chang [/bib_ref]. Flavonoids are very common polyphenol secondary metabolites found in plants and fungus and can be found in various commercial products on the market with medical, nutritional, and cosmeceutical applications [bib_ref] The bioavailability, transport, and bioactivity of dietary flavonoids: A review from a..., Williamson [/bib_ref]. Expectedly, several of these flavonoids were reported to possess anti-adipogenic or pro-osteogenic properties with different action mechanisms and therapeutical targets [bib_ref] Role of flavonoids in the interactions among obesity, inflammation, and autophagy, García-Barrado [/bib_ref] [bib_ref] Flavonoids in bone erosive diseases: Perspectives in osteoporosis treatment, Bellavia [/bib_ref]. Myricetin is a flavonoid from the flavonol sub-class and one of the most common flavonoids in the daily human diet. Myricetin can be found in various plant species and is known for its nutraceutical value [bib_ref] Myricetin: A review of the most recent research, Song [/bib_ref]. Similar to other bioactive flavonoids, myricetin is a good free radical scavenger and its derivatives exert different physiological benefits [bib_ref] Synthesis and antibacterial and antiviral activities of myricetin derivatives containing a 1,2,4-triazole..., Chen [/bib_ref]. Reported nutraceutical properties of myricetin derivatives vary from reducing cardiovascular risks to anti-tumor, anti-diabetic and anti-aging activities [bib_ref] Amelioration of high-fat diet-induced obesity and its associated complications by a myricetin..., Nallappan [/bib_ref] [bib_ref] Methylated derivatives of myricetin enhance life span in Caenorhabditis elegans dependent on..., Büchter [/bib_ref]. In this context, the goal of the present study is to analyze the effect of a myricetin glycoside derivative, myricetin-3-O-β-D-galactopyranoside (M3G) [fig_ref] Figure 1: Chemical structure of M3G [/fig_ref] , on the differentiation of human bone marrow-derived mesenchymal stromal cells (hBM-MSCs) into adipocytes and osteoblasts in order to provide insights towards its anti-osteoporotic potential. The results of the current study are expected to provide data for further studies to utilize M3G as a nutraceutical with MSC differentiation regulatory effects that can be used against metabolic disorders such as obesity and osteoporosis.
# Materials and methods
## M3g isolation and characterization
M3G was isolated from Limonium tetragonum methanolic crude extracts. The isolation and characterization of M3G was carried out as previously described [bib_ref] Antioxidant activity of the halophyte Limonium tetragonum and its major active components, Lee [/bib_ref]. Briefly, 1 H and 13 C NMR spectra were evaluated and compared with published literature for the identification of M3G.
## Cell culture and differentiation
The current study used human bone marrow-derived mesenchymal stromal cells (hBM-MSCs) as in vitro models to analyze MSC differentiation into adipocytes and osteoblast. Cells were purchased from PromoCell (C-12974, Heidelberg, Germany). For the experiments, cells at passage numbers between 3 and 6 were cultured in 6-well plates (1 × 10 6 cells/well) unless otherwise noted. Prior to differentiation, cells were grown to confluency during which they were fed mesenchymal stem cell growth medium (C-28009, PromoCell). Following confluency, hBM-MSCs were induced to differentiate into adipocytes or osteoblasts by swapping the culture medium with mesenchymal stem cell adipogenic differentiation medium 2 (C-28016, PromoCell) or mesenchymal stem cell osteogenic differentiation medium (C-28013, PromoCell), respectively. Cells introduced to differentiation media were incubated for 10 days, swapping the differentiation medium with fresh media every third day. Treatment with M3G was carried out along with initial differentiation inducement only. Incubation of hBM-MSCs before and during the differentiation was in controlled incubators set to 37 - C temperature and 5% CO 2 level.
## Cell viability assay
Any cytotoxic effect that M3G might have on hBM-MSCs was evaluated with standard 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay [bib_ref] Antioxidant activity of the halophyte Limonium tetragonum and its major active components, Lee [/bib_ref]. Briefly, hBM-MSCs in 96-well plates (seeded at 1 × 10 3 cell/well density) were incubated for 24 h prior to the assay. After 24 h, cells were treated with varying doses of M3G, introduced with fresh culture medium. Treatment lasted for 72 h, after which the wells were aspirated and added 100 µL of MTT reagent (1 mg/mL in culture medium) and incubated for 4 more hours. Next, 100 mL dimethyl sulfoxide (DMSO) was added to each well to stop the reaction and the absorbance values were measured at 540 nm. Cell viability was quantified as a relative percentage compared to M3G untreated control group absorbance values after the normalization with the blank group which only contained the reagents but not the cells.
## Staining of lipid droplets in adipocytes by oil red o staining
As a marker of successful adipocyte differentiation and maturation, intracellular lipid accumulation was observed in adipo-induced hBM-MSCs at day 10 differentiation. Observation and quantification of the lipid accumulation of adipocytes were carried out with Oil Red O staining of the intracellular lipid droplets. Briefly, hBM-MSCs were induced to differentiate into adipocytes as described above. At day 10 of differentiation, the wells were aspirated, and the adipocytes were fixed on wells for 1 h by the addition of 10% fresh formaldehyde in phosphate-buffered saline (PBS) (v/v). Fixed cells were stained with 1 mL 0.5% Oil Red O staining solution (m/v in 60% isopropanol) in each well. Staining was carried out for 1 h in incubators. Unbound Oil Red O stain was washed with PBS and the stained lipid droplets were observed under a light microscope (Olympus, Tokyo, Japan). For the quantification of the stain, lipid-bound Oil Red O was eluted from wells by addition of 100% isopropanol to the wells. The eluted stain was then quantified via measuring absorbance values at 500 nm. Lipid accumulation was given as a relative percentage of Oil Red O stain eluted compared to M3G untreated control adipocytes. Absorbance values were normalized against blank wells that contained all the reagents but not the cells.
# Mrna expression analysis
The hBM-MSCs differentiated into adipocytes or osteoblasts were analyzed for the mRNA expressions of adipogenic and osteogenic marker genes. At day 10 of differentiation, total RNA was extracted from non-differentiated and differentiated (M3G-treated and nontreated) hBM-MSC groups by following the given protocols of AccuPrep Universal RNA Extraction Kit (BioNeer, Daejeon, Korea). The cDNA was synthesized from extracted RNA (2 µg) using Cell Script All-in-One cDNA Synthesis Master Mix (CellSafe, Yongin, Korea) according to manufacturer's instructions and following steps: 42 - C for 60 min and 72 - C for 5 min. Subsequently, polymerase chain reaction (PCR) analysis was carried out with Luna ® Universal qPCR Mix (New England Biolabs, Ipswich, MA, USA) in TP800 Thermal Cycler Dice™ Real-Time System (Takara Bio, Ohtsu, Japan). The amplification values were plotted as fold change compared to the differentiated but not M3G treated group. All amplification values were normalized against β-actin as the reference gene. Visualization of the PCR bands was carried out with electrophoresis (30 min at 100 V on a 1.5% agarose gel) of final products and subsequent ethidium bromide (1 mg/mL) staining. Stained gels were pictured with CAS-400SM Davinch-Chemi Imager™ (Davinch-K, Seoul, Korea).
# Protein level analysis
The protein levels of adipogenic and osteogenic markers were analyzed in differentiated hBM-MSCs via Western blotting. At day 10 differentiation, total protein was extracted from cell lysates obtained by addition of 1 mL RIPA buffer and pipetting. The supernatants of the cell lysates after centrifugation (13,000× g, 4 - C, 15 min) were used for Western blotting after assessing their total protein content with the BCA protein assay kit (Thermo Fisher Scientific, Waltham, MA, USA). Furthermore, nuclear protein content was extracted from total protein using the NE-PER TM Nuclear Extraction Kit (#78835; Thermo Fisher Scientific). For the blotting, the same amount of protein from total or nuclear extracts of test groups (20 µg) was separated by sodium dodecyl-sulfate polyacrylamide gel electrophoresis (SDS-PAGE) (4% stacking and 10% separating gels). The transfer of proteins from gels to membranes was carried out with traditional wet transfer protocols for Western blotting using polyvinylidene fluoride membrane (Amersham Bioscience, Westborough, MA, USA). Transfer to membranes was followed by blocking the membranes for 4 h in 5% skim milk (v/v) in TBS-T buffer for 4 h. Then, the proteins were incubated overnight at 4 - C with primary antibodies diluted as suggested by the manufacturer in 1X TBS-T. The primary antibodies used in this study were given in a previous report [bib_ref] Effect of quercetin 3-o-β-d-galactopyranoside on the adipogenic and osteoblastogenic differentiation of human..., Oh [/bib_ref]. Following the primary antibody hybridization, membranes were treated for 2 h with horseradish peroxidase-conjugated secondary antibodies specific to the primary antibody organism. Visualization of the protein bands was carried out by staining membranes with an enhanced chemiluminescent (ECL) kit (Amersham Bioscience) following the manufacturer's protocol. The protein bands were pictured with CAS-400SM Davinch-Chemi imager (Davinch-K).
## Cellular alkaline phosphatase (alp) activity
The activity of ALP obtained from differentiated hBM-MSC osteoblasts was evaluated with a commercial kit. At day 7 of differentiation, osteo-induced hBM-MSCs were lysed with the addition of 0.1% Triton X-100 and 25 mM carbonate buffer and pipetting. The supernatants obtained from cell lysates after centrifugation (12,000× g, 4 - C, 15 min) were first analyzed for their total protein content by the Bradford protein determination method. Next, the ALP activity in the same amount of protein from each test group was calculated with the Alkaline Phosphatase Activity Colorimetric Assay Kit (K412-500; BioVision, Hannover, Germany) following the producer's protocol. The conversion rate of pNP by ALP was plotted after normalization to the total protein amount of cell samples.
## Alizarin red staining
As a marker of osteoblast maturation, extracellular mineralization was analyzed by staining calcified nodules of differentiated hBM-MSC osteoblasts with Alizarin Red staining. At the day 10 of differentiation, hBM-MSC osteoblasts were fixed by swapping culture medium with 4 - C 70% ethanol. Following 1 h of incubation, ethanol was removed from wells and fixed osteoblasts were washed with distilled water. Two percent of Alizarin Red staining solution (m/v) with a pH of 4.2 was then added to the wells for 10 min of staining. Following staining, the solution was aspirated from wells and the dyed cells were washed with distilled water. Stained calcifications were pictured with an Olympus microscope (Tokyo, Japan) fitted with a camera. In order to quantify the staining, the Alizarin Red dye was eluted from stained cells with 10% (m/v) cetylpyridinium chloride in 10 mM sodium phosphate buffer (Sigma-Aldrich, St. Louis, MO, USA). Eluted dyes were measured for their absorbance at 560 nm using a Multiskan GO microplate reader (Tecan Austria GmbH, Grodig, Austria). Values were normalized against the blank group that contained only elution solution and plotted as a relative percentage of differentiated but not M3G-treated hBM-MSCs.
## Immunohistochemical staining
Marker proteins for adipocytes and osteoblasts in differentiated hBM-MSCs were visualized with immunofluorescence staining of the BMP2, and RUNX2 for osteoblasts and PPARγ, and perilipin-1 for adipocytes. The hBM-MSCs were cultured and differentiated as stated earlier except that the immunohistochemical staining wells were fitted with glass coverslips prior to cell seeding. The rest of the differentiation was carried out in the same manner as previous assays. At day 10 of differentiation, differentiated hBM-MSCs were fixed on coverslips and hybridized with antibodies conjugated with Alexa Fluor 488 (A-11008; Thermo Fisher Scientific) against perilipin-1 (ab3526; Abcam, Cambridge, UK), PPARγ (ab9256; Abcam), BMP2, and RUNX2. In the case of nuclei staining as a reference, the ProLong Gold Antifade reagent with DAPI (#8961; Cell Signaling Technology, Danvers, MA, USA) was used. The hBM-MSCs were fixed and stained by using the commercial solutions and protocols of the Immunofluorescence Application Solutions Kit (#12727; Cell Signaling Technology).
# Statistical analysis
Where applicable, all data were given as the average of three independent experiments (n = 3) (run in triplicate except Western blotting) ± SD unless otherwise described. Data groups were subjected to one-way analysis of variance (ANOVA) with post hoc Duncan's multiple range test using SAS software (SAS v9.1, SAS Institute, Cary, NC, USA) and the minimum significant statistical difference was defined at p < 0.05 level.
# Results
## Proliferation, alp activity and extracellular mineralization of osteo-induced hbm-mscs
Treatment with M3G did not result in any viability loss in hBM-MSCs following 72 h treatment up to a concentration of 10 µM [fig_ref] Figure 2: Effect of M3G on the osteoblastogenic differentiation of hBM-MSCs [/fig_ref]. Starting from 25 µM, cell viability was observed to drop. On the other hand, hBM-MSCs treated with M3G with the initial osteogenic differentiation media (3 days) exerted increased viable cell amount in a dose-dependent manner until 10 µM [fig_ref] Figure 2: Effect of M3G on the osteoblastogenic differentiation of hBM-MSCs [/fig_ref]. However, at 25 µM concentration, the M3G-treated group exhibited a loss of viability compared to the untreated differentiated group. Therefore, the following studies were conducted using an M3G dose of up to 10 µM which was deemed safe.
ALP activity of the differentiated hBM-MSCs was analyzed at day 7 of differentiation. ALP activity significantly increased by the osteoblast differentiation from 18.98 U/mL of non-differentiated hBM-MSCs to 65.88 U/mL. At concentrations of 5 and 10 µM, M3G treatment significantly stimulated the ALP activity to the levels of 70.28 U/mL and 75.56 U/mL, respectively, compared to untreated hBM-MSC osteoblasts [fig_ref] Figure 2: Effect of M3G on the osteoblastogenic differentiation of hBM-MSCs [/fig_ref]. The level of ALP activity stimulation was expectedly observed in mineralization as well. M3G treatment dosedependently elevated the extracellular calcification in hBM-MSC osteoblasts [fig_ref] Figure 2: Effect of M3G on the osteoblastogenic differentiation of hBM-MSCs [/fig_ref]. At the highest concentration treated (10 µM), Alizarin Red stained areas were 22.30% higher compared to untreated hBM-MSC osteoblasts. ALP activity of the differentiated hBM-MSCs was analyzed at day 7 of differentiation. ALP activity significantly increased by the osteoblast differentiation from 18.98 U/mL of non-differentiated hBM-MSCs to 65.88 U/mL. At concentrations of 5 and 10 μM, M3G treatment significantly stimulated the ALP activity to the levels of 70.28 U/mL and 75.56 U/mL, respectively, compared to untreated hBM-MSC osteoblasts [fig_ref] Figure 2: Effect of M3G on the osteoblastogenic differentiation of hBM-MSCs [/fig_ref]. The level of ALP activity stimulation was expectedly observed in mineralization as well. M3G treatment dose-dependently elevated the extracellular calcification in hBM-MSC osteoblasts [fig_ref] Figure 2: Effect of M3G on the osteoblastogenic differentiation of hBM-MSCs [/fig_ref]. At the highest concentration treated (10 μM), Alizarin Red stained areas were 22.30% higher compared to untreated hBM-MSC osteoblasts.
## Osteoblast marker gene and protein expression
The effect of M3G on osteoblast marker genes and proteins was analyzed with RT-PCR and Western blotting, respectively, at day 10 differentiation in osteo-induced hBM-MSCs. The mRNA expression levels of ALP and RUNX2 as osteoblastogenesis markers increased with the induction of osteoblastogenesis in untreated hBM-MSCs. The presence of M3G dose-dependently increased the mRNA expression of ALP and RUNX2 compared to the untreated differentiated control group. Similar results were obtained from Western blotting. Protein levels of ALP and RUNX2 were increased by differentiation and further
## Osteoblast marker gene and protein expression
The effect of M3G on osteoblast marker genes and proteins was analyzed with RT-PCR and Western blotting, respectively, at day 10 differentiation in osteo-induced hBM-MSCs. The mRNA expression levels of ALP and RUNX2 as osteoblastogenesis markers increased with the induction of osteoblastogenesis in untreated hBM-MSCs. The presence of M3G dose-dependently increased the mRNA expression of ALP and RUNX2 compared to the untreated differentiated control group. Similar results were obtained from Western blotting. Protein levels of ALP and RUNX2 were increased by differentiation and further stimulated by M3G treatment. In addition, the osteopontin levels were also observed to be increased by M3G treatment [fig_ref] Figure 3: Effect of M3G on osteoblastogenesis marker gene expression [/fig_ref]. Cells 2021, 10, 2690 7 of 16 stimulated by M3G treatment. In addition, the osteopontin levels were also observed to be increased by M3G treatment [fig_ref] Figure 3: Effect of M3G on osteoblastogenesis marker gene expression [/fig_ref].
## Effect of m3g on canonical wnt/bmp signaling during osteoblastogenesis
In order to elucidate the mechanism of action behind the osteoblastogenesis stimulation activity of M3G, the canonical Wnt signaling was investigated. The levels of signaling proteins and their phosphorylation were analyzed by Western blotting. The hBM-MSCs expressed significantly increased levels of Wnt10a and Wnt10b proteins following the inducement of osteoblastogenesis. Expectedly, phosphorylated β-catenin levels were observed along with Axin. The M3G (10 μM)-treated group exhibited increased levels of Wnt10a and Wnt10b compared to the untreated differentiated control group. In a similar trend, M3G treatment also increased the levels of β-catenin phosphorylation and free Axin protein levels [fig_ref] Figure 4: Effect of M3G on Wnt/β-catenin [/fig_ref]. Very similar results were obtained from the BMP signaling results. BMP2 and the phosphorylation of its downstream activator Smad1/5 complex were significantly increased during osteoblastogenesis, and with M3G treatment protein levels of BMP2 and phosphorylated Smad1/5 complex were further increased compared to untreated osteoblasts [fig_ref] Figure 4: Effect of M3G on Wnt/β-catenin [/fig_ref]. As a transcriptional activator signaling cascade, MAPK signaling was also investigated in osteo-induced hBM-MSCs. Phosphorylation of p38, ERK, and JNK MAPKs were all stimulated in osteoblasts [fig_ref] Figure 4: Effect of M3G on Wnt/β-catenin [/fig_ref] and, similar to previous results, treatment with M3G further elevated the phosphorylated levels of p38 and JNK MAPKs but not ERK. In addition, as the downstream effectors of p38 and JNK, respectively, phosphorylation of c-Fos and c-Jun were similarly increased first by osteoblastogenesis inducement and further with M3G treatment.
## Effect of m3g on canonical wnt/bmp signaling during osteoblastogenesis
In order to elucidate the mechanism of action behind the osteoblastogenesis stimulation activity of M3G, the canonical Wnt signaling was investigated. The levels of signaling proteins and their phosphorylation were analyzed by Western blotting. The hBM-MSCs expressed significantly increased levels of Wnt10a and Wnt10b proteins following the inducement of osteoblastogenesis. Expectedly, phosphorylated β-catenin levels were observed along with Axin. The M3G (10 µM)-treated group exhibited increased levels of Wnt10a and Wnt10b compared to the untreated differentiated control group. In a similar trend, M3G treatment also increased the levels of β-catenin phosphorylation and free Axin protein levels [fig_ref] Figure 4: Effect of M3G on Wnt/β-catenin [/fig_ref]. Very similar results were obtained from the BMP signaling results. BMP2 and the phosphorylation of its downstream activator Smad1/5 complex were significantly increased during osteoblastogenesis, and with M3G treatment protein levels of BMP2 and phosphorylated Smad1/5 complex were further increased compared to untreated osteoblasts [fig_ref] Figure 4: Effect of M3G on Wnt/β-catenin [/fig_ref]. As a transcriptional activator signaling cascade, MAPK signaling was also investigated in osteo-induced hBM-MSCs. Phosphorylation of p38, ERK, and JNK MAPKs were all stimulated in osteoblasts [fig_ref] Figure 4: Effect of M3G on Wnt/β-catenin [/fig_ref] and, similar to previous results, treatment with M3G further elevated the phosphorylated levels of p38 and JNK MAPKs but not ERK. In addition, as the downstream effectors of p38 and JNK, respectively, phosphorylation of c-Fos and c-Jun were similarly increased first by osteoblastogenesis inducement and further with M3G treatment. The Wnt and BMP signaling was further analyzed by investigating the nuclear and cytosolic levels of transcriptional activators: phosphorylated β-catenin, Smad1/5, c-Fos, and c-Jun. The nuclear levels of phosphorylated β-catenin, Smad1/5, c-Fos, and c-Jun were all elevated in osteoblasts and in a similar fashion to previous results, M3G presence elevated the nuclear levels of these transcription factors [fig_ref] Figure 4: Effect of M3G on Wnt/β-catenin [/fig_ref].
The effect of M3G on osteoblastogenesis was further examined by immunohistochemical staining of the BMP and RUNX2 proteins. Images of the cells showed that M3G treatment significantly stimulated the BMP and RUNX2 protein expression in osteo-induced hBM-MSCs [fig_ref] Figure 5: Effect of M3G on osteoblastogenesis marker protein expression [/fig_ref]. The Wnt and BMP signaling was further analyzed by investigating the nuclear and cytosolic levels of transcriptional activators: phosphorylated β-catenin, Smad1/5, c-Fos, and c-Jun. The nuclear levels of phosphorylated β-catenin, Smad1/5, c-Fos, and c-Jun were all elevated in osteoblasts and in a similar fashion to previous results, M3G presence elevated the nuclear levels of these transcription factors [fig_ref] Figure 4: Effect of M3G on Wnt/β-catenin [/fig_ref].
The effect of M3G on osteoblastogenesis was further examined by immunohistochemical staining of the BMP and RUNX2 proteins. Images of the cells showed that M3G treatment significantly stimulated the BMP and RUNX2 protein expression in osteo-induced hBM-MSCs [fig_ref] Figure 5: Effect of M3G on osteoblastogenesis marker protein expression [/fig_ref].
## Effect of m3g on the lipid accumulation of hbm-msc adipocytes
As a marker of adipocyte maturation, lipid accumulation of adipo-induced hBM-MSCs was investigated by staining the intracellular lipid droplets. At day 10 of differentiation, adipo-induced hBM-MSCs showed accumulated intracellular lipid droplets which were dose-dependently decreased by M3G treatment [fig_ref] Figure 6: Effect of M3G on adipogenic differentiation of hBM-MSCs [/fig_ref]. The effect of M3G on lipid accumulation is further confirmed by the immunohistochemical staining of perilipin-1, a protein known as lipid droplet-associated protein. Cell images clearly showed that the significant increase in perilipin-1 levels in hBM-MSC adipocytes was reverted by the presence of M3G (10 µM) [fig_ref] Figure 6: Effect of M3G on adipogenic differentiation of hBM-MSCs [/fig_ref].
## Effect of m3g on the expression of adipogenesis marker genes and proteins
The effect of M3G on adipogenesis of hBM-MSCs was further investigated by the mRNA and protein expression of adipogenesis markers: PPARγ, CEBPα, and SREBP1c. The mRNA expression of adipogenic markers was analyzed by RT-qPCR. The expression of all three adipogenic markers was significantly increased by the inducement of adipogenesis [fig_ref] Figure 7: Effect of M3G on the expression of adipogenic marker genes [/fig_ref]. However, treatment with M3G suppressed the expression of PPARγ, CEBPα, and SREBP1c in a dose-dependent manner. Correlative results were obtained from the examination of protein levels. M3G treatment dose-dependently inhibited the protein levels of PPARγ and CEBPα which were stimulated in untreated hBM-MSC adipocytes. However, the effect of M3G on SREBP1c protein levels was not as significant as others [fig_ref] Figure 7: Effect of M3G on the expression of adipogenic marker genes [/fig_ref].
The effect of M3G on adipogenic marker genes was further confirmed by immunohistochemical staining of the adipo-induced hBM-MSCs for PPARγ. Cell images clearly showed that M3G treatment significantly decreased the PPARγ levels in adipo-induced hBM-MSCs [fig_ref] Figure 7: Effect of M3G on the expression of adipogenic marker genes [/fig_ref].
## Effect of m3g on the lipid accumulation of hbm-msc adipocytes
As a marker of adipocyte maturation, lipid accumulation of adipo-induced hBM-MSCs was investigated by staining the intracellular lipid droplets. At day 10 of differentiation, adipo-induced hBM-MSCs showed accumulated intracellular lipid droplets which were dose-dependently decreased by M3G treatment [fig_ref] Figure 6: Effect of M3G on adipogenic differentiation of hBM-MSCs [/fig_ref]. The effect of M3G on lipid accumulation is further confirmed by the immunohistochemical staining of perilipin-1, a protein known as lipid droplet-associated protein. Cell images clearly showed that the significant increase in perilipin-1 levels in hBM-MSC adipocytes was reverted by the presence of M3G (10 μM) [fig_ref] Figure 6: Effect of M3G on adipogenic differentiation of hBM-MSCs [/fig_ref]. [fig_ref] Figure 5: Effect of M3G on osteoblastogenesis marker protein expression [/fig_ref]. Effect of M3G on osteoblastogenesis marker protein expression. Analysis of protein expression was carried out by immunofluorescence staining of osteo-induced hBM-MSCs at day 10 differentiation. M3G was treated with initial differentiation induction (3 days) and M3G was not present in subsequent media changes. β-actin was used as internal loading control. Scale bar: 100 µm.
## Effect of m3g on mapk/ap-1 signaling in adipo-induced hbm-mscs
In order to evaluate the mechanism behind the effect of M3G on adipogenic differentiation of hBM-MSCs, the phosphorylation of MAPK/AP-1 signaling was examined. The hBM-MSCs induced for adipogenesis exhibited significantly increased phosphorylation of p38 and JNK where ERK phosphorylation was suppressed. M3G treatment reverted the effect of adipogenesis on MAPKs, suppressing the p38 and JNK phosphorylation while stimulating that of ERK. Similar results were obtained from the analysis of MAPK downstream activator c-Fos and c-Jun. Adipo-induced hBM-MSCs exhibited suppressed phosphorylation of c-Fos and c-Jun under M3G treatment which was otherwise significantly elevated .
## Effect of m3g on the expression of adipogenesis marker genes and proteins
The effect of M3G on adipogenesis of hBM-MSCs was further investigated by the mRNA and protein expression of adipogenesis markers: PPARγ, CEBPα, and SREBP1c. The mRNA expression of adipogenic markers was analyzed by RT-qPCR. The expression of all three adipogenic markers was significantly increased by the inducement of adipogenesis [fig_ref] Figure 7: Effect of M3G on the expression of adipogenic marker genes [/fig_ref]. However, treatment with M3G suppressed the expression of PPARγ, CEBPα, and SREBP1c in a dose-dependent manner. Correlative results were obtained from the examination of protein levels. M3G treatment dose-dependently inhibited the protein levels of PPARγ and CEBPα which were stimulated in untreated hBM-MSC adipocytes. However, the effect of M3G on SREBP1c protein levels was not as significant as others [fig_ref] Figure 7: Effect of M3G on the expression of adipogenic marker genes [/fig_ref].
The effect of M3G on adipogenic marker genes was further confirmed by immunohistochemical staining of the adipo-induced hBM-MSCs for PPARγ. Cell images clearly showed that M3G treatment significantly decreased the PPARγ levels in adipo-induced hBM-MSCs [fig_ref] Figure 7: Effect of M3G on the expression of adipogenic marker genes [/fig_ref]. hBM-MSCs induced for adipogenesis exhibited significantly increased phosphorylation of p38 and JNK where ERK phosphorylation was suppressed. M3G treatment reverted the effect of adipogenesis on MAPKs, suppressing the p38 and JNK phosphorylation while stimulating that of ERK. Similar results were obtained from the analysis of MAPK downstream activator c-Fos and c-Jun. Adipo-induced hBM-MSCs exhibited suppressed phosphorylation of c-Fos and c-Jun under M3G treatment which was otherwise significantly elevated . . Effect of M3G on the MAPK/AP-1 signaling. Analysis of MAPK and AP-1 activation and was carried out with Western blotting of whole-cell lysates of adipo-induced hBM-MSCs at day 10 of differentiation. M3G was treated with initial differentiation for 3 days. β-actin was used as an internal loading control. * p < 0.05, ** p < 0.01, *** p < 0.001 vs. differentiated untreated group.
# Discussion
Currently, various plant-origin phenolic compounds are being studied and reports show that these compounds are potent nutraceuticals with promising health beneficial effects varying from treating symptoms of several diseases to the prevention of metabolic disorders such as diabetes, obesity, and osteoporosis [bib_ref] Superfruits: Phytochemicals, antioxidant efficacies, and health effects-A comprehensive review, Chang [/bib_ref]. Plant-based nutraceuticals can be found in common dietary sources in various forms and derivatives which may show different and/or more efficient bioactivities. Myricetin is a common phytochemical which is intensively studied for its bioactivities and still provides novel outcomes in terms of efficiency and effectiveness [bib_ref] Myricetin: A review of the most recent research, Song [/bib_ref] [bib_ref] Myricetin bioactive effects: Moving from preclinical evidence to potential clinical applications, Taheri [/bib_ref] [bib_ref] Current pharmacological trends on myricetin, Gupta [/bib_ref]. Therefore, derivatives of myricetin are also receiving increasing interest in this context where their potential bioactivities, action mechanisms, adverse effects, etc., are being continuously explored. On account of myricetin being a promising compound with reported activities against obesity and related complications, this study focused on the evaluation of a myricetin glycoside, M3G, for its effect on adipogenic and osteogenic differentiation of hBM-MSCs to obtain data regarding its antiosteoporotic potential.
Reports showed that myricetin has beneficial effects on osteogenic differentiation, bone formation, and bone repair, while also being an anti-obesity agent with adipogenic inhibitory properties [bib_ref] Myricetin enhances osteogenic differentiation through the activation of canonical Wnt/β-catenin signaling in..., Ying [/bib_ref] [bib_ref] Possible osteoprotective effects of myricetin in STZ induced diabetic osteoporosis in rats, Ying [/bib_ref] [bib_ref] Myricetin inhibits adipogenesis in human adipose tissue-derived mesenchymal stem cells, Bin [/bib_ref]. Myricetin was shown to induce osteoblast differentiation of different origin cell lines via similar mechanisms of BMP-2 and MAPK activation [bib_ref] Myricetin induces human osteoblast differentiation through bone morphogenetic protein-2/p38 mitogen-activated protein kinase..., Hsu [/bib_ref] [bib_ref] Enhancing effects of myricetin on the osteogenic differentiation of human periodontal ligament..., Kim [/bib_ref]. However, the effect of myricetin was also observed to be dependent on the treatment time, . Effect of M3G on the MAPK/AP-1 signaling. Analysis of MAPK and AP-1 activation and was carried out with Western blotting of whole-cell lysates of adipo-induced hBM-MSCs at day 10 of differentiation. M3G was treated with initial differentiation for 3 days. β-actin was used as an internal loading control. * p < 0.05, ** p < 0.01, *** p < 0.001 vs. differentiated untreated group.
# Discussion
Currently, various plant-origin phenolic compounds are being studied and reports show that these compounds are potent nutraceuticals with promising health beneficial effects varying from treating symptoms of several diseases to the prevention of metabolic disorders such as diabetes, obesity, and osteoporosis [bib_ref] Superfruits: Phytochemicals, antioxidant efficacies, and health effects-A comprehensive review, Chang [/bib_ref]. Plant-based nutraceuticals can be found in common dietary sources in various forms and derivatives which may show different and/or more efficient bioactivities. Myricetin is a common phytochemical which is intensively studied for its bioactivities and still provides novel outcomes in terms of efficiency and effectiveness [bib_ref] Myricetin: A review of the most recent research, Song [/bib_ref] [bib_ref] Myricetin bioactive effects: Moving from preclinical evidence to potential clinical applications, Taheri [/bib_ref] [bib_ref] Current pharmacological trends on myricetin, Gupta [/bib_ref]. Therefore, derivatives of myricetin are also receiving increasing interest in this context where their potential bioactivities, action mechanisms, adverse effects, etc., are being continuously explored. On account of myricetin being a promising compound with reported activities against obesity and related complications, this study focused on the evaluation of a myricetin glycoside, M3G, for its effect on adipogenic and osteogenic differentiation of hBM-MSCs to obtain data regarding its antiosteoporotic potential.
Reports showed that myricetin has beneficial effects on osteogenic differentiation, bone formation, and bone repair, while also being an anti-obesity agent with adipogenic inhibitory properties [bib_ref] Myricetin enhances osteogenic differentiation through the activation of canonical Wnt/β-catenin signaling in..., Ying [/bib_ref] [bib_ref] Possible osteoprotective effects of myricetin in STZ induced diabetic osteoporosis in rats, Ying [/bib_ref] [bib_ref] Myricetin inhibits adipogenesis in human adipose tissue-derived mesenchymal stem cells, Bin [/bib_ref]. Myricetin was shown to induce osteoblast differentiation of different origin cell lines via similar mechanisms of BMP-2 and MAPK activation [bib_ref] Myricetin induces human osteoblast differentiation through bone morphogenetic protein-2/p38 mitogen-activated protein kinase..., Hsu [/bib_ref] [bib_ref] Enhancing effects of myricetin on the osteogenic differentiation of human periodontal ligament..., Kim [/bib_ref]. However, the effect of myricetin was also observed to be dependent on the treatment time, treatment period, and the cell line origin. The studies showed that during the onset of osteoporosis, the commitment of the MSCs in favor of adipogenesis occurs at the early stages of differentiation [bib_ref] Osteoporosis and obesity: Role of Wnt pathway in human and murine models, Colaianni [/bib_ref] where antagonistic feedback between PPAR and Wnt signaling takes place to suppress Wnt-mediated osteogenic induction while enhancing PPAR expression [bib_ref] The PPAR-γ/SFRP5/Wnt/β-catenin signal axis regulates the dexamethasone-induced osteoporosis, He [/bib_ref]. Therefore, the current study aimed to analyze the effect of M3G on the early stages of MSC commitment, after the introduction of differentiation medium (osteogenic or adipogenic). The current results reported a similar effect for the M3G when treated only during the initial differentiation inducement (first 3 days). This result suggested that the osteoblast differentiation enhancement effect of M3G appeared through its intervention on the early osteoblastogenesis signaling such as the activation of the transcriptional activities of RUNX2 and canonical Wnt signaling. Results confirmed this as the enhancement of mRNA and protein levels of RUNX2 in differentiated hBM-MSCs in the presence of M3G up to 10 µM. It was also accompanied by the elevation of osteopontin which is a protein expressed as a downstream product of RUNX2 activation. The role of both canonical and non-canonical Wnt signaling and its intertwined relationship with PPAR pathways during osteoporosis have been reported [bib_ref] Osteoporosis and obesity: Role of Wnt pathway in human and murine models, Colaianni [/bib_ref]. Additionally, He and Su [bib_ref] The PPAR-γ/SFRP5/Wnt/β-catenin signal axis regulates the dexamethasone-induced osteoporosis, He [/bib_ref] showed that the antagonist relationship between Wnt and PPAR pathways has regulatory roles in a dexamethasone-induced osteoporosis model. They suggested that the enhanced expression of PPAR by dexamethasone subsequently suppressed osteogenic differentiation which underlies glucocorticoid-induced osteoporosis.
Therefore, Wnt signaling was investigated as the upstream activator of osteoblastogenesis. This canonical pathway regulates osteoblast differentiation via RUNX2. Treatment of the osteo-induced hBM-MSCs with M3G resulted in elevated levels of Wnt10a and Wnt10b levels which are osteoblast differentiation-associated ligands expressed highly to activate the Wnt signaling. Activation of Wnt signaling induces the stabilization and accumulation of β-catenin protein and its consecutive nuclear translocation. Current results showed that following M3G treatment the levels of nuclear β-catenin significantly increased indicating that M3G treatment enhanced the Wnt activation. Similar results were obtained from the BMP-2 signaling cascade. BMP-2 signaling is a mid and late osteoblast differentiation pathway which is activated by canonical Wnt signaling and responsible for the osteoblast maturation and consequent bone formation [bib_ref] Civitelli, R. β-Catenin and BMP-2 synergize to promote osteoblast differentiation and new..., Mbalaviele [/bib_ref] [bib_ref] BMP-2 controls alkaline phosphatase expression and osteoblast mineralization by a Wnt autocrine..., Rawadi [/bib_ref]. M3G treatment significantly enhanced BMP-2 expression and the phosphorylation of its downstream activator Smad1/5 complex compared to untreated differentiated hBM-MSC osteoblasts. In a similar manner to that of Wnt signaling, treatment with M3G also increased the nuclear levels of phosphorylated Smad1/5 complex. Overall, results indicated that M3G presence had beneficial effects on the activation of Wnt and BMP signaling pathways during osteoblast differentiation of hBM-MSCs.
Reports indicated that Wnt10a and Wnt10b also had an antagonistic relationship with adipogenic differentiation of MSCs [bib_ref] Mesenchymal stem cells: Cell fate decision to osteoblast or adipocyte and application..., Hu [/bib_ref] where activation of Wnt signaling inhibits the adipogenic signaling cascade while inducing osteogenic differentiation. Considering that increased adipogenic differentiation of bone marrow MSCs as opposed to osteogenic differentiation is one of the reasons behind porous bones seen in osteoporosis [bib_ref] The relationship between bone marrow adipose tissue and bone metabolism in postmenopausal..., Li [/bib_ref] , the effect of M3G on adipogenic differentiation of hBM-MSCs was also investigated. Reports showed that activation of Wnt signaling during the early stages of osteoblast differentiation suppresses PPARγ-mediated inducement of adipogenic signaling. First, the effect of M3G on the lipid accumulation of adipo-induced hBM-MSCs showed that M3G had a potential inhibitory effect on the differentiation of adipocytes. This was confirmed with the M3G-mediated suppression of PPARγ mRNA and protein expression levels compared to untreated differentiation hBM-MSC adipocytes. Along with PPARγ, M3G treatment also suppressed the expression of downstream adipogenic transcription factors SREBP1c and C/EBPα [bib_ref] C/EBP transcription factors regulate SREBP1c gene expression during adipogenesis, Payne [/bib_ref]. Some studies reported that the regulation of PPARγ gene expression is partly controlled through the MAPK activation and transcriptional activities of its downstream transcription factor AP-1 [bib_ref] IL6 receptor facilitates adipogenesis differentiation of human mesenchymal stem cells through activating..., Deng [/bib_ref] [bib_ref] Fra-2/AP-1 controls adipocyte differentiation and survival by regulating PPARγ and hypoxia, Luther [/bib_ref]. AP-1 is a complex formed by phosphorylated c-Fos and c-Jun, two proteins activated by MAPK signaling. Therefore, the effect of M3G on MAPK signaling was also examined to confirm its role in suppressing PPARγ-mediated adipogenesis. The results showed that the phosphorylation of p38 and JNK MAPKs were significantly increased in adipocytes while the M3G treatment resulted in decreased levels of p38 and JNK phosphorylation. On the other hand, ERK1/2 phosphorylation was suppressed in hBM-MSC adipocytes at day 10 differentiation. Moreover, M3G treatment relieved the adipogenic suppression of ERK1/2. Although ERK activation was reported to be a part of adipogenesis at early stages, it was also shown that mature adipocytes exhibited suppressed ERK1/2 activation [bib_ref] Inhibition of adipocyte differentiation by mechanical stretching through ERK-mediated downregulation of PPARγ2, Tanabe [/bib_ref].
In the current study, adipo-inducement of hBM-MSCs resulted in increased phosphorylation of p38 and JNK MAPKs and reduced ERK phosphorylation. This was suggested to be due to negative regulation of the adipogenic maturation during later stages of adipogenesis via ERK1/2 activation [bib_ref] Inhibition of adipocyte differentiation by mechanical stretching through ERK-mediated downregulation of PPARγ2, Tanabe [/bib_ref]. Overall, treatment with M3G significantly suppressed the adipogenesis in induced hBM-MSCs through the PPARγ pathway. This is in agreement with previous reports where myricetin and other polyphenols with similar structures such as quercetin and kaempferol inhibited the early stages of adipogenesis via suggested interaction with PPARγ [bib_ref] Phenolic compounds inhibit 3T3-L1 adipogenesis depending on the stage of differentiation and..., Aranaz [/bib_ref] [bib_ref] Molecular mechanisms of the anti-obesity and anti-diabetic properties of flavonoids, Hossain [/bib_ref]. Furthermore, the antagonistic relationship between Wnt and PPARγ signaling [bib_ref] Involvement of WNT/β-catenin signaling in the treatment of osteoporosis, Rossini [/bib_ref] might play role in the adipogenesis inhibitory effect of M3G in hBM-MSCs. The therapeutic potential of the Wnt/β-catenin pathway and tilting the differential tendencies of MSCs towards osteoblast while suppressing their adipogenesis as a target for osteoporosis treatment have been reported [bib_ref] Involvement of WNT/β-catenin signaling in the treatment of osteoporosis, Rossini [/bib_ref] [bib_ref] Suppression of EZH2 prevents the shift of osteoporotic MSC fate to adipocyte..., Jing [/bib_ref]. Shifting the differentiation balance of bone marrow stromal cells in favor of osteoblastogenesis had attenuative effects on osteoporosis-mediated damages in bone tissue. Considering all, it was suggested that M3G enhanced osteoblastogenesis while suppressing adipogenesis in hBM-MSCs through Wnt and PPARγ pathways, respectively.
Prior to the results of the current study, the potential of myricetin, of which M3G is derived, has been shown in vivo with several disease models. Fan et al. [bib_ref] Myricetin ameliorates glucocorticoid-induced osteoporosis through the ERK signaling pathway, Fan [/bib_ref] showed that myricetin supplement alleviated dexamethasone-induced osteoporosis in Sprague-Dawley rats through the promotion of osteogenic differentiation. Similar to M3G, myricetin exerted its effects via ERK signaling pathway in the aforementioned study. Additionally, in a study by Ying et al. [bib_ref] Possible osteoprotective effects of myricetin in STZ induced diabetic osteoporosis in rats, Ying [/bib_ref] , myricetin also exhibited osteoprotective effects in diabetic rats as well. The anti-adipogenic effects of myricetin were shown in a study by Su et al. [bib_ref] Myricetin protects against diet-induced obesity and ameliorates oxidative stress in C57BL/6 mice, Su [/bib_ref] where myricetin supplement protected C57BL/6 mice from diet-induced obesity via PPARγlinked signaling pathways. Angelicin, a natural bioactive compound, was also shown to prevent osteoporosis in ovariectomized rat models by regulating Wnt and PPAR signaling which suggests a potential role against postmenopausal osteoporosis [bib_ref] Isopsoralen regulates PPAR-γ/WNT to inhibit oxidative stress in osteoporosis, Wang [/bib_ref]. Parallel to reported in vivo activities of myricetin, M3G was shown to interact with similar pathways in vitro to promote osteoblastogenesis and inhibit adipogenesis.
In conclusion, M3G was shown to be a potential bioactive compound with potential beneficial effects on bone structure. It was hypothesized that M3G may exhibit its activities via enhancing differentiation of bone marrow mesenchymal stromal cells into osteoblasts while hindering adipocyte differentiation to stimulate bone formation. The results suggested M3G as a potential lead natural product as a derivative of myricetin to develop anti-osteoporotic nutraceuticals, although further in vivo and detailed mechanism of action studies are urged for the proper utilization of M3G.
## Institutional review board statement:
The human bone marrow-derived mesenchymal stromal cell line was obtained from PromoCell (C-12974, Heidelberg, Germany). The compound used in this study was isolated from Limonium tetragonum (Thunb.) A.A. Bullock plants which were handpicked in May 2008 from Yulchon-myeon, Yeosu city, Jeollanam-do, South Korea. A voucher specimen (03U-2) was deposited at the Herbarium of Division of Marine Environment and Bioscience, Korea Maritime University, Busan, South Korea. Ethical review and approval were not applicable.
## Informed consent statement: not applicable.
Data Availability Statement: All data used to support the findings of this study are available from the corresponding author upon reasonable request.
[fig] Figure 1: Chemical structure of M3G. [/fig]
[fig] Figure 2: Effect of M3G on the osteoblastogenic differentiation of hBM-MSCs. Effect of M3G on the (a) viability of nondifferentiated hBM-MSCs and (b) proliferation of osteo-induced hBM-MSCs. Viable cell amount was measured by quantification of MTT dye removed from cells at day 3 of differentiation. Proliferation was given as relative viable cell amount (%) of untreated osteo-induced control. (c) Effect of M3G on the activity of cellular ALP. Cellular ALP activity of osteoinduced hBM-MSCs was measured with a spectrophotometric enzymatic activity assay at day 7 of differentiation. M3G was present in the first 3 days of differentiation only. Control: Non-differentiated untreated cells in culture medium. (d) Effect of M3G on the extracellular mineralization of osteo-induced hBM-MSCs. Extracellular mineralization was measured by Alizarin Red staining and quantified by the absorbance values of the retained dye at day 10 of differentiation. Mineralization was given as a relative percentage of untreated osteo-induced hBM-MSCs. M3G was treated in the first 3 days of differentiation only. * p < 0.05, ** p < 0.01 vs. differentiated untreated group. [/fig]
[fig] Figure 3: Effect of M3G on osteoblastogenesis marker gene expression. Analysis of gene expression was carried by measuring mRNA and protein levels in osteo-induced hBM-MSCs via RT-PCR and Western blot, respectively, at day 10 differentiation. Osteo-induced hBM-MSCs were treated with M3G until day 3 of differentiation. * p < 0.05, ** p < 0.01 vs. differentiated untreated group. [/fig]
[fig] Figure 4: Effect of M3G on Wnt/β-catenin (a), BMP (b) and MAPK (c) signaling pathways and their downstream nuclear effectors (d). Analysis of protein expression was carried out by Western blotting of osteo-induced hBM-MSCs at day 10 differentiation. M3G was treated with initial differentiation induction (3 days) and it was not present in subsequent media changes. β-actin (for whole cell and cytosolic fraction) and lamin B1 (for nuclear fraction) were used as internal loading control. * p < 0.05, ** p < 0.01 vs. differentiated untreated group. [/fig]
[fig] Figure 5: Effect of M3G on osteoblastogenesis marker protein expression. Analysis of protein expression was carried out by immunofluorescence staining of osteo-induced hBM-MSCs at day 10 differentiation. M3G was treated with initial differentiation induction (3 days) and M3G was not present in subsequent media changes. β-actin was used as internal loading control. Scale bar: 100 μm. [/fig]
[fig] Figure 6: Effect of M3G on adipogenic differentiation of hBM-MSCs. (a) Effect of M3G on the intracellular lipid accumulation in adipo-induced hBM-MSCs at day 10 of differentiation. Lipid droplets were stained with Oil Red O and the quantification was carried out by measuring absorbance values of retained dye. Lipid accumulation level was given as relative percentage of Oil Red O dye compared to adipo-induced untreated group. * p < 0.05, ** p < 0.01 vs. differentiated untreated group. (b) Effect of M3G on the expression of perilipin-1 in adipo-induced hBM-MSCs at day 10 differentiation analyzed by immunofluorescence staining. DAPI staining was used to highlight the nucleus of viable cells. Scale bar: 50 μm. M3G was treated with initial differentiation induction (3 days) and was not present in subsequent media changes. [/fig]
[fig] Figure 7: Effect of M3G on the expression of adipogenic marker genes (a) and proteins (b). Analysis of gene and protein expression of adipo-induced hBM-MSCs was carried out by RT-PCR and Western blot, respectively, at day 10 of differentiation. β-actin was used as internal loading controls. (c) Effect of M3G on the expression of PPARγ in adipo-induced hBM-MSCs at day 10 differentiation analyzed by immunofluorescence staining. DAPI staining was used to highlight the nucleus of viable cells. Scale bar: 50 μm. M3G was treated with initial differentiation for 3 days. * p < 0.05, ** p < 0.01, *** p < 0.001 vs. differentiated untreated group. [/fig]
[fig] Author: Contributions: Conceptualization and methodology, J.H.O., F.K. and C.-S.K.; validation, formal analysis and investigation, J.H.O., H.J.J. and F.K.; resources, Y.S. and C.-S.K.; data curation, J.H.O., H.J.J. and C.-S.K.; writing-original draft preparation, visualization, F.K.; supervision, and project administration, Y.S. and C.-S.K.; funding acquisition, C.-S.K. All authors have read and agreed to the published version of the manuscript. [/fig]
[fig] Funding: This research was supported by Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Education (2020R1I1A3070750). [/fig]
|
Clinical characteristics and infectivity of asymptomatic carriers of SARS-CoV-2 (Review)
# Introduction
In December 2019, the first case of pneumonia of unknown etiology was identified in Wuhan, China. Subsequently, the pathogen was isolated and officially named Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2) by the International Committee on Taxonomy of Viruses, due to its marked similarity with the agent of Severe Acute Respiratory Syndrome (SARS) [bib_ref] Clinical Characteristics of Coronavirus Disease 2019 in China, Guan [/bib_ref]. SARS-CoV-2 primarily spreads through respiratory droplets and close contact, and all populations are susceptible [bib_ref] A rapid advice guideline for the diagnosis and treatment of 2019 novel..., Jin [/bib_ref]. Coronavirus disease 2019 (COVID-19) caused by SARS-CoV-2 was declared by the World Health Organization (WHO) as a public health emergency of international concern on January 30, 2020 and a pandemic on March 12, 2020. As of September 28, 2020, the number of globally confirmed cases was over 32.7 million and the number of mortalities was >991,000, according to WHO statistics (4).
As the epidemic progresses, the prevention and control of asymptomatic carriers has gradually become the focus of the global epidemic control effort. For example, at present, China and certain other countries have entered the stage of remission, although sporadic new cases, imported cases and asymptomatic carriers remain risk factors for spreading the virus [bib_ref] Asymptomatic carriers of COVID-19 as a concern for disease prevention and control:..., Zhang [/bib_ref]. According to statistics from the National Health Commission of China, there were 375 asymptomatic carriers currently under medical observation by September 29. One study initially estimated that asymptomatic carriers, pre-symptomatic and mild infection of COVID-19 may account for 60% of all infections [bib_ref] Covert coronavirus infections could be seeding new outbreaks, Qiu [/bib_ref]. In addition, the majority of these patients may not self-isolate, and spread the virus to others unconsciously, as they are unaware that they have been infected with SARS-CoV-2 [bib_ref] The incidence of the novel coronavirus SARS-CoV-2 among asymptomatic patients: A systematic..., Nasrallah [/bib_ref]. Therefore, asymptomatic carriers of COVID-19 should be taken seriously because they serve a crucial role in disease transmission. The present review aims to discuss the clinical characteristics and infectivity of asymptomatic carriers and provide suggestions to control these potential hidden sources of infection.
## Demographic characteristics
The studies in the present review identified that the demographic characteristics of asymptomatic carriers are significantly correlated with age. In a study of 147 asymptomatic carriers in Anhui, China, male carriers accounted for 51.7% and carriers <20 years old accounted for 15.6% (9). Hu et al [bib_ref] Clinical characteristics of 24 asymptomatic infections with COVID-19 screened among close contacts..., Hu [/bib_ref] reported the epidemiological and clinical data of 24 asymptomatic infections with COVID-19, and identified that people <15 years old were more likely to be asymptomatic carriers compared with patients >15 years. The study by Wang et al [bib_ref] Clinical Outcomes in 55 patients with severe acute respiratory syndrome coronavirus 2..., Wang [/bib_ref] , which reviewed the clinical information of 55 asymptomatic carriers in Shenzhen, China, confirmed that 27% of asymptomatic carriers were <18 years old. Rawat et al [bib_ref] COVID-19 in newborns and infants-low risk of severe disease: Silver Lining or..., Rawat [/bib_ref] confirmed that children were more likely to have asymptomatic infections compared with adults; their data indicated that 27% of pediatric COVID-19 cases were asymptomatic carriers and that only 7% of adult patients had no symptoms [bib_ref] COVID-19 in newborns and infants-low risk of severe disease: Silver Lining or..., Rawat [/bib_ref]. This result was also verified in Zhejiang, China, where, among 36 children with confirmed COVID-19, 28% were asymptomatic carriers [bib_ref] Clinical and epidemiological features of 36 children with coronavirus disease 2019 (COVID-19)..., Qiu [/bib_ref]. Tan et al [bib_ref] Viral Transmission and Clinical Features in Asymptomatic Carriers of SARS-CoV-2 in Wuhan, Tan [/bib_ref] identified that asymptomatic carriers were relatively young (mean age, 34.5 years old). Current evidence suggests that the asymptomatic carriers with SARS-CoV-2 vary by age, however, to the best of our knowledge, no studies have identified differences in sex between asymptomatic carriers.
Previous data may also explain why asymptomatic carriers are more likely to be young people. For example, it may be that the gene expression of angiotensin converting enzyme II (ACE2) in nasal epithelial cells is age-dependent. ACE2 is the host receptor for SARS-CoV-1 and SARS-CoV-2 (15). Bunyavanich et al [bib_ref] Nasal gene expression of angiotensin-converting enzyme 2 in children and adults, Bunyavanich [/bib_ref] demonstrated that ACE2 gene expression increased with age, with the lowest expression in younger children (<10 years old; n=45). In addition, Do et al [bib_ref] Can data from paediatric cohorts solve the COVID-19 puzzle?, Do [/bib_ref] argued that cytokine storms in adults promote stronger inflammatory responses compared with that in younger people. also demonstrated that older patients (60-85 years old) and middle-aged patients (40-59 years old) were more likely to induce higher neutralizing antibodies titers compared with younger patients (15-39 years old), which has been shown to be associated with disease severity [bib_ref] The convalescent sera option for containing COVID-19, Casadevall [/bib_ref].
## Ratio of asymptomatic carriers among all infected persons
It is not easy to accurately estimate the proportion of asymptomatic carriers among patients with confirmed COVID-19, as the total number of asymptomatic carriers will not be identified unless all people are systematically screened. Moreover, without follow-up observation, it is also difficult to distinguish between pre-symptomatic persons (symptomatic later in the course of the disease) and the asymptomatic carriers (asymptomatic during the whole duration of the disease) [bib_ref] The dilemmas of the classification of SARS-CoV-2 infection without clinical manifestations: Asymptomatic..., Arteaga-Livias [/bib_ref]. Wang et al [bib_ref] Clinical characteristics of non-critically ill patients with novel coronavirus infection (COVID-19) in..., Wang [/bib_ref] identified 1,012 patients with non-severe COVID-19 disease with positive reverse transcriptase polymerase chain reaction (RT-PCR) results in Wuhan, China, of which 30 (3.0%) were asymptomatic [bib_ref] Clinical characteristics of non-critically ill patients with novel coronavirus infection (COVID-19) in..., Wang [/bib_ref]. In Shanghai, China, another study showed that 3.9% of the 328 adults diagnosed with COVID-19 were asymptomatic carriers (13 cases) [bib_ref] Follow-up of asymptomatic patients with SARS-CoV-2 infection, Zhou [/bib_ref]. However, a study in Jinan, China, observed that 11 of 47 confirmed patients (23.4%) were asymptomatic carriers [bib_ref] Characteristics of asymptomatic patients with SARS-CoV-2 infection in Jinan, Ma [/bib_ref]. According to statistics from the Chinese Center for Disease Control and Prevention, on February 11, China recorded a total of 72,314 patients, including 889 asymptomatic cases (1.2%). However, this statistic was measured prior to the large-scale screening of patients with COVID-19 in China, so the proportion may be less compared with the real data. One study estimated that the rate of asymptomatic carriers was 30.8% (95% CI, 7.7-53.8%) among 565 Japanese individuals evacuated from Wuhan, China [bib_ref] Estimation of the asymptomatic qratio of novel coronavirus infections (COVID-19), Nishiura [/bib_ref]. There are also studies using published data for statistical model analysis, estimating that the asymptomatic proportion of COVID-19 cases on the Japanese cruise ship Princess Diamond was 17.9% (95% CI, 15.5-20.2%) [bib_ref] Estimating the asymptomatic proportion of coronavirus disease 2019 (COVID-19) cases on board..., Mizumoto [/bib_ref]. In a cohort of children, Qiu et al [bib_ref] Clinical and epidemiological features of 36 children with coronavirus disease 2019 (COVID-19)..., Qiu [/bib_ref] identified that 10 (28%) of 36 pediatric patients infected with COVID-19 were completely asymptomatic. Similarly, in another study of children with COVID-19, 20 (27%) of 74 cases were asymptomatic infections [bib_ref] Coinfection and other clinical characteristics of COVID-19 in children, Wu [/bib_ref]. In addition, the study by Rawat et al [bib_ref] COVID-19 in newborns and infants-low risk of severe disease: Silver Lining or..., Rawat [/bib_ref] , examining patients in the United States of America, demonstrated that 27% of pediatric patients with COVID-19 had no symptoms, while only 7% of adult patients had no symptoms [bib_ref] COVID-19 in newborns and infants-low risk of severe disease: Silver Lining or..., Rawat [/bib_ref]. Notably, a study conducted in a large hospital in Wuhan, China also found that the asymptomatic infection rate among healthcare workers was 9.7% (88/908) [bib_ref] Asymptomatic infection by SARS-CoV-2 in healthcare workers: A study in a large..., Zhao [/bib_ref]. From these studies, it can be observed that the proportion of asymptomatic carriers varies widely in different studies. In addition, the true prevalence of asymptomatic carriers may be lower compared with current estimates due to the inclusion of pre-symptomatic infections in some cross-sectional studies.
## Radiological and laboratory data
Asymptomatic carriers differ from confirmed patients not only in their symptoms, but also in their laboratory and chest CT results. At present, no uniform pattern in laboratory and radiological manifestations of asymptomatic carriers has been established. Meng et al [bib_ref] CT imaging and clinical course of asymptomatic cases with COVID-19 pneumonia at..., Meng [/bib_ref] collected clinical information on 58 asymptomatic patients with COVID-19 pneumonia and identified that all carriers had a normal laboratory and abnormal chest CT results. The main CT features of asymptomatic patients were ground-glass opacity [(GGO) 55, 94.8%] at admission. However, among the 24 asymptomatic carriers in Nanjing, China, lymphopenia and leukopenia were not obvious, and chest CT images were normal in the majority of young carriers (<15 years old) (10). Yu and Yang (30) also reported 2 asymptomatic carriers with normal lymphocyte count and chest CT images. Similarly, Wang et al [bib_ref] Clinical Outcomes in 55 patients with severe acute respiratory syndrome coronavirus 2..., Wang [/bib_ref] reported increased C-reactive protein and lactate dehydrogenase levels in 10 and 13 carriers, respectively, and only 11 patients (including two with severe infections) exhibited leukopenia, from the follow-up results of 55 asymptomatic carriers [bib_ref] Clinical Outcomes in 55 patients with severe acute respiratory syndrome coronavirus 2..., Wang [/bib_ref]. A study from Zhou et al [bib_ref] Follow-up of asymptomatic patients with SARS-CoV-2 infection, Zhou [/bib_ref] followed up 13 cases of asymptomatic carriers, of which 2 patients had leukocytes below the normal range. A recent study identified that 3 of 11 (27.3%) persistent asymptomatic patients with SARS-CoV-2 infection did not have typical chest CT results and the GGO of other carriers had been absorbed in nearly 5 days (31). Ali and Ghonimy (32) examined 44 asymptomatic cases of COVID-19 and identified that the main features of their CT results were GGO (93%; n=41) and surrounding distribution (77.3%; n=34). In summary, these studies suggest that radiological and laboratory abnormalities vary and laboratory findings and chest CT images can be normal in asymptomatic carriers. Therefore, SARS-CoV-2 infection cannot be ruled out on the basis of normal radiological and laboratory results.
## Infectivity and transmission risk of asymptomatic carriers
A number of studies have confirmed that asymptomatic carriers can transmit the virus. In Anyang, Henan, China, a family cluster of 5 patients with COVID-19 was infected with SARS-CoV-2 by contact with asymptomatic family members [bib_ref] Presumed Asymptomatic Carrier Transmission of COVID-19, Bai [/bib_ref]. In Nanjing, COVID-19 was also transmitted by asymptomatic carriers to cohabiting family members (10). Yu and Yang [bib_ref] COVID-19 transmission through asymptomatic carriers is a challenge to containment, Yang [/bib_ref] described an asymptomatic carrier who had transmitted SARS-CoV-2 to his brother following close contact. Recently, an asymptomatic carrier resulted in 13 secondary cases of infection, reportedly as an asymptomatic super spreader [bib_ref] Unclear but present danger: An asymptomatic SARS-CoV-2 carrier, Yu [/bib_ref]. In summary, it is almost unquestionable that asymptomatic carriers have infectivity.
In earlier cross-sectional studies, no significant differences in viral titers and transmission efficiency were identified between asymptomatic and symptomatic patients. A study showed that the viral load in asymptomatic patients was similar to that of symptomatic patients [bib_ref] SARS-CoV-2 viral load in upper respiratory specimens of infected patients, Zou [/bib_ref]. Chen et al [bib_ref] Epidemiological characteristics of infection in COVID-19 close contacts in Ningbo city, Chen [/bib_ref] followed up and investigated 2,147 close contacts in Ningbo, Zhejiang, China, with a total infection rate of 6.15%, while there was no significant difference in the infection rate between the confirmed cases and asymptomatic cases (6.30 and 4.11%, respectively) [bib_ref] Epidemiological characteristics of infection in COVID-19 close contacts in Ningbo city, Chen [/bib_ref]. Concomitantly, Yin and Jin [bib_ref] Comparison of transmissibility of coronavirus between symptomatic and asymptomatic patients: Reanalysis of..., Yin [/bib_ref] analyzed the transmission rate of 157 symptomatic cases and 30 asymptomatic cases of COVID-19 and identified that there was no significant difference in the transmission rate between symptomatic and asymptomatic patients. However, none of these studies excluded pre-symptomatic patients in the incubation period. Another study in Korea including 3 pre-symptomatic patients and 10 asymptomatic carriers suggested that pre-symptomatic patients in the incubation period may have an increased viral load compared with asymptomatic carriers [bib_ref] Viral kinetics of SARS-CoV-2 in asymptomatic carriers and presymptomatic patients, Kim [/bib_ref]. Therefore, the transmission efficiency of asymptomatic carriers estimated by these studies may be increased compared with the real data. Subsequently, increasing evidence suggests that true asymptomatic carriers may have a decreased risk of transmission than symptomatic patients. The study by Liu et al [bib_ref] The assessment of transmission efficiency and latent infection period in asymptomatic carriers..., Liu [/bib_ref] showed that the secondary attack rate of 131 asymptomatic carriers was 2.6% (24/914), while the secondary attack rate of 16 confirmed cases was 9.7% (23/236), further demonstrating that the transmission risk of asymptomatic carriers was decreased compared with that of the confirmed cases. In China, a study including 12 asymptomatic carriers in Wuhan also identified that only one carrier has transmitted the virus to a close contact [bib_ref] Viral Transmission and Clinical Features in Asymptomatic Carriers of SARS-CoV-2 in Wuhan, Tan [/bib_ref]. Chen et al (39) tracked 9 asymptomatic carriers and identified that none of them had transmitted SARS-CoV-2 to others, so they speculated that the absence of symptoms such as sneezing and coughing may interfere with virus shedding and decrease the risk of infection. In addition, a systematic review showed that the relative risk of asymptomatic carriers transmission was decreased by 42% compared with symptomatic transmission, while there were no differences in viral load. In light of the data from these studies, asymptomatic carriers may have a weaker transmission efficiency compared with symptomatic patients. However, the transmission ability of asymptomatic carriers cannot be ignored, as some of them may develop into super spreaders. There is no doubt that the early identification of asymptomatic carriers and the prevention of the further spread of the virus are critical to disease control.
## Transmission period of asymptomatic carriers
The study in Korea calculated that the median duration of virus shedding (from the first confirmed positive RNA test result to the first negative RNA test result) in asymptomatic carriers was 4.5 days (2.5-9 days), and all asymptomatic carriers were almost non-infectious after isolation of 14 days [bib_ref] Viral kinetics of SARS-CoV-2 in asymptomatic carriers and presymptomatic patients, Kim [/bib_ref]. However, the study from Hu et al [bib_ref] Clinical characteristics of 24 asymptomatic infections with COVID-19 screened among close contacts..., Hu [/bib_ref] showed that the median spread interval was 9.5 days (up to 21 days), and the actual infection period may be longer as the exact date of the first infection is uncertain for some carriers. Tan et al [bib_ref] Viral Transmission and Clinical Features in Asymptomatic Carriers of SARS-CoV-2 in Wuhan, Tan [/bib_ref] demonstrated that the median duration of virus shedding was 11.5 (9-14) days among 12 carriers in Wuhan, 2 of which were found to have positive RNA results lasting up to 2 months [bib_ref] Viral Transmission and Clinical Features in Asymptomatic Carriers of SARS-CoV-2 in Wuhan, Tan [/bib_ref]. also reported a median duration of viral shedding of 19 [bib_ref] Characterization of the receptor-binding domain (RBD) of 2019 novel coronavirus: Implication for..., Tai [/bib_ref] [bib_ref] Nasal gene expression of angiotensin-converting enzyme 2 in children and adults, Bunyavanich [/bib_ref] [bib_ref] Can data from paediatric cohorts solve the COVID-19 puzzle?, Do [/bib_ref] [bib_ref] The convalescent sera option for containing COVID-19, Casadevall [/bib_ref] [bib_ref] The dilemmas of the classification of SARS-CoV-2 infection without clinical manifestations: Asymptomatic..., Arteaga-Livias [/bib_ref] [bib_ref] Clinical characteristics of non-critically ill patients with novel coronavirus infection (COVID-19) in..., Wang [/bib_ref] [bib_ref] Follow-up of asymptomatic patients with SARS-CoV-2 infection, Zhou [/bib_ref] [bib_ref] Characteristics of asymptomatic patients with SARS-CoV-2 infection in Jinan, Ma [/bib_ref] [bib_ref] Estimation of the asymptomatic qratio of novel coronavirus infections (COVID-19), Nishiura [/bib_ref] [bib_ref] Estimating the asymptomatic proportion of coronavirus disease 2019 (COVID-19) cases on board..., Mizumoto [/bib_ref] days in 37 asymptomatic carriers. Meanwhile, Pan et al [bib_ref] Epidemiological and clinical characteristics of 26 asymptomatic severe acute respiratory syndrome coronavirus..., Pan [/bib_ref] retrospectively analyzed 26 persistently asymptomatic SARS-CoV-2 carriers and identified that the median time from diagnosis to negative nucleic acid testing was 7.5 days (2-20 days) for carriers with normal or atypical chest CT findings, compared with 12.5 days (8-22 days) for those with typical GGO [bib_ref] Epidemiological and clinical characteristics of 26 asymptomatic severe acute respiratory syndrome coronavirus..., Pan [/bib_ref]. These studies have shown that SARS-CoV-2 can remain present in asymptomatic carriers for a long period of time as long as 2 months and carriers with abnormal chest CT findings may have a longer transmission period.
## Advice to identify asymptomatic carriers
Asymptomatic carriers can transmit the virus as an unseen source of infection, which is a great challenge for epidemic prevention and control, so the importance of implementing social distancing measures and using masks cannot be overemphasized. It is also of vital importance to identify and isolate asymptomatic carriers at an early stage. The question of how asymptomatic carriers can be identified earlier remains. At present, the detection of SARS-CoV-2 virus infection mainly depends on nucleic acid, serum antibody and chest CT. As the chest CT images of some asymptomatic carriers are entirely normal, particularly in young people, the present study only discussed the detection of nucleic acids and antibodies. Studies have shown that nucleic acid detection has the disadvantages of high false-negative rate and low sensitivity, although it is widely used [bib_ref] Asymptomatic infection by SARS-CoV-2 in healthcare workers: A study in a large..., Zhao [/bib_ref] [bib_ref] Alert for non-respiratory symptoms of coronavirus disease 2019 patients in epidemic period:..., Lu [/bib_ref]. The positive rates of RT-PCR detection of sputum specimens, nasal swabs and throat swabs were 74.4-88.9, 53.6-73.3 and 29.6-61.3%, respectively. In contrast, the sensitivity and specificity of the rapid IgM-IgG combined antibody test were 88.7 and 90.6%, respectively [bib_ref] Alert for non-respiratory symptoms of coronavirus disease 2019 patients in epidemic period:..., Lu [/bib_ref]. There is also evidence that the sensitivity of total antibodies is increased compared with that of IgM or IgG alone [bib_ref] Antibody responses to SARS-CoV-2 in patients of novel coronavirus disease 2019, Zhao [/bib_ref]. A study identified that the asymptomatic infection rate confirmed by RT-PCR alone was 4.2% (38/908), while combined with RT-PCR test and serological detection was 9.7% (88/908) [bib_ref] Asymptomatic infection by SARS-CoV-2 in healthcare workers: A study in a large..., Zhao [/bib_ref].
Considering the available data, the combination of serum antibody testing and nucleic acid testing is recommended to identify asymptomatic carriers in high-risk areas or high-risk populations. Confirmed asymptomatic carriers should be subject to centralized quarantine and observation for at least 14 days until they receive 2 consecutive negative nucleic acid test results (at least 24 h apart).
# Conclusions
Based on the literature, the present review identified that the proportion of asymptomatic carriers of SARS-CoV-2 in all infections varies significantly between different studies, but asymptomatic infection may be correlated with age. Among the young population, the proportion of asymptomatic carriers may be increased. There is no uniform pattern of laboratory and radiological findings in asymptomatic carriers, and some asymptomatic carriers may have completely normal chest CT and laboratory results. It is almost certain that asymptomatic carriers can transmit SARS-CoV-2, but they may have a decreased risk of transmission compared with symptomatic patients. At present, the management of asymptomatic carriers should become a focus of epidemic prevention, particularly in countries with stable epidemics. Dual detection of serum antibodies and nucleic acids in high-risk areas and high-risk groups and quarantining for at least 14 days is recommended following diagnosis as an asymptomatic carrier.
AcknowledgementsNot applicable.FundingNo funding was received.Availability of data and materialsNot applicable.Authors' contributionsWZ and FX made significant contributions to select the topic of the review. WZ was responsible for the collection and collation of documents and writing draft manuscripts. HZ, FX and XW revised the writing of the manuscript. All authors read and approved the final manuscript.Ethics approval and consent to participateNot applicable.Patient consent for publicationNot applicable.Competing interestsThe authors declare that they have no competing interests. |
Quantitative Particle Uptake by Cells as Analyzed by Different Methods
Figure SI-I.2.1: (a) Schematic diagram of a SNARF-1 dextran loaded PEM pH sensor capsule. (b) Fluorescence microscopy images of sensor capsules under acidic (yellow fluorescence, shown in green false colors) and alkaline (red fluorescence) conditions. (c) Dependence of the fluorescence signal of some batches (with 2 and 2.5 bilayers having positive and negative surface charge, respectively) of sensor capsules on ambient pH in terms of the ratio of red to yellow fluorescence intensity I r /I y at different pH values. The hydrodynamic diameters and zeta potentials of SNARF-containing PEM capsules can be seen in Figure SI-I.2.2.
II) Uptake studies of SNARF-loaded capsules by cells using fluorescence microscopy III) Flow cytometer analysis of SNARF-loaded capsules internalized by cells IV) Elemental analysis of gold nanoparticle (GNP)-loaded capsules internalized by cells V) Viability measurements of cells exposed to capsules and inhibitor VI) References Note that excerpts of the results shown in the graphs shown in the Supporting Information are also presented in a different representation/compilation in the main manuscript.
## I) particle tracking of snarf-loaded capsules inside cells i.1) materials, reagents, and equipment i.2) synthesis and characterization of snarf-loaded polyelectrolyte micro-(pem) capsules i.3)
Uptake studies by confocal laser scanning microscope (CLSM) I.4) Particle tracking by CLSM, image processing, and data evaluation I.5) Fractal dimension and average end-to-end scaling exponent I.6) Results
## I.1) materials, reagents, and equipment
Calcium chloride dihydrate (CaCl 2 , #223506), sodium carbonate (Na 2 CO 3 , #S7795), sodium chloride (NaCl, # S7653), poly(sodium 4-styrenesulfonate) (PSS, molar weight M w ≈ 70 kDa, #243051), poly(allylamine hydrochloride) (PAH, M w ≈ 58 kDa, #283223), ethylenediaminetetraacetic acid disodium salt dihydrate (EDTA disodium salt, #E5134), cytochalasin D (#C8273), bafilomycin A1 (#B1793), and sterile dimethyl sulfoxide (DMSO, #D2650) were purchased from Sigma-Aldrich (Germany). Seminaphtharhodafluor-1 dextran (SNARF-1 dextran, M w ≈ 70 kDa, #D3304) was obtained from Life Technologies (USA) and were supplied from Roth (Karlsruhe, Germany). Ultrapure double distilled deionized (Milli-Q) water (DDW) having resistivity of 18.2 MΩ.cm was used for all experiments.
Human cervical carcinoma (HeLa) cells were obtained from American type culture collection (ATCC). For cell culture Dulbecco's modified Eagle's medium (DMEM; #D6546), and penicillin streptomycin (P/S, #P4333) were purchased from Sigma-Aldrich. GlutaMAX TM (#35050-038) and 0.05% Trypsin-EDTA (#25300) were purchased from GIBCO (Life Technologies). Phosphate buffered saline (PBS; Biochrom #L 1825) and fetal bovine serum (FBS; Biochrom, Germany, #S0615) were purchased from Merck Millipore. A Neubauer improved counting chamber (haemocytometer) by Marienfeld Laboratory glassware was used for counting cells. µ-slide 8 well plates (Ibidi #80826) were purchased from Ibidi. Cell culture flasks (25 and 75 cm 2 ; #90025 and #90075, respectively) from TPP were used for culturing cells. Eppendorf (2 mL; # 72.695.500), and falcon tubes (15 and 50 mL; #62.554.502, #62.547.254, respectively) were from Sarstedt. A thermo electron corporation Varifuge 3.0 R by Fisher Scientific was used for centrifuging the cell lines. The hydrodynamic diameter and zeta potential measurements of the capsules were performed in water using a Malvern Zetasizer Nano ZS.
A confocal laser scanning microscope (CLSM 510 Meta) from Zeiss was used for visualizing and live imaging of HeLa cells engulfing SNARF-loaded capsules. For image acquisition the fluorescence was excited at 543 nm using the helium neon laser of the CLSM and samples were observed through a 63X/1.40 oil-immersion DIC M27 objective. ImmersolTM 518F immersion oil (Zeiss) was used during imaging.
## I.2) synthesis and characterization of snarf-loaded polyelectrolyte micro-(pem) capsules
PEM capsules with encapsulated SNARF dextran inside their cavities were fabricated by coprecipitation, followed by layer by layer (LBL) assembly of oppositely charged polymeric layers. Briefly, pH-sensitive PEM capsules having 2 and 2.5 bilayers of non-biodegradable polymers composed of PSS and PAH were fabricated based on LBL assembly of oppositely charged polymers around sacrificial template cores containing the pH sensitive dye SNARF-1 conjugated with dextran. For co-precipitation of SNARF-1 dextran with CaCO 3 microparticles, solutions of CaCl 2 and Na 2 CO 3 were mixed under vigorous stirring in the presence of SNARF-1 dextran at room temperature (RT) in aqueous media. In a glass vial, 4.2 mL SNARF-1 dextran (0.5 mg/mL) was added to 3 mL of 0.33 M CaCl 2 (0.33 M). Under vigorous magnetic stirring (1100 rpm) 3 mL of Na 2 CO 3 (0.33 M) solution was quickly mixed with the above mixture for 30 s, followed by keeping the reaction contents without agitation for 2 min. Calcium carbonate particles were washed three times with DDW and used for LBL assembly of oppositely charged polyelectrolytes (2 mg/mL 0.5 M NaCl). The alternating layers of negatively and positively charged polymers, i.e., PSS and PAH respectively were deposited around the charged sacrificial microparticle templates following a well established protocol. Layer deposition was achieved by alternating immersion of microparticles inside the corresponding polymer solution (5 mL) for 13 min, followed by subsequent washing with DDW to remove excess polymers. Finally the cores of PEM capsules were dissolved by complexion of Ca 2+ ions with EDTA (5 mL, 0.2 M, pH 6.5) and particles were washed with DDW and stored at 4 °C for further use. In order to minimize the artifacts due to size variation of capsules (mean diameter ≈ 3.5-4 µm), the cores were manufactured on large scale, dried under vacuum after washing with acetone, and stored at 4 °C. Latter in all experiments the same cores were used for LBL deposition of polyelectrolyte shells.
SNARF is a ratiometric pH sensitive dye having pK a value of 7.5. After incorporation inside the cavity of capsules, the dye retained its pH sensitive fluorescence characteristics . Upon excitation at 543 nm it has two emission peaks at 580 nm (yellow fluorescence) and 640 nm (red fluorescence). When the ambient pH is low its emission maxima is at 580 nm, while at high pH the emission maximum is at 640 nm. This pH dependent shift in fluorescence intensity makes SNARF-bearing capsules ratiometric indicators for intracellular sensing without the need of additional reference fluorophores . The pH dependent fluorescence response of the SNARF capsules as synthesized for the present work was monitored by adjusting the pH of the capsules by means of immersion in commercially available buffers (pH 3 -10), cf. Figure SI-I.2.1 . The intensity of red (640 nm) and yellow (580 nm, displayed as "green" in false colors of the microscope)) fluorescence of SNARF was measured using confocal laser scanning microscopy (CLSM) and the ratio of the respective fluorescence intensities was plotted versus pH. and zeta potential ζ ± Δζ as measured in water given as average and standard deviation. These data were derived from Figure SI-I.2.2.
## I.3) uptake studies by confocal laser scanning microscope (clsm)
HeLa cells were grown in DMEM supplemented with 10% FBS, 1% P/S, and 1% Glutamax.,000 cells were seeded per well in 8-well µ-slides having growth area of 1 cm 2 per well. Each well was filled with 300 µL of growth medium. Cells were kept inside an incubator set to 37 °C and complemented with 5% CO 2 at constant rate. Capsule tracking and uptake experiments were performed in serum supplemented or in serum deprived media, in the presence or absence of inhibitors using CLSM. In order to track the internalization of PEM sensor capsules, the CLSM was equipped with a portable incubator in order to maintain the µ-slides at 37 °C with 5% CO 2 . Before starting the experiment the cells were provided fresh serum supplemented/ serum deprived media. In addition, the inhibitors (cytochalasin D, 300 nM & bafilomycin A1, 0.25 µM) were added 1 hour before the addition of PEMs and starting the experiments. PEMs were provided to the cells at a concentration of 5 capsules per seeded cell just at the time of start of the imaging using a Plan-Apochromat 63x/1.40 Oil DIC M27 objective. During imaging, time lapse image series were captured . In order to minimize photobleaching of SNARF during image acquisition, very low laser excitation power was used in order to detect fluorescence for long term studies. The fluorophore was excited at λ ex = 543 nm and its fluorescence emission in the yellow and in the red region was detected using a 560 -615 nm band pass filter and a 633 nm long pass filter, respectively. During imaging the lateral resolution (i.e. in the x-y-plane) was adjusted to 0.32 µm, whereas, 120 s temporal resolution was used. In order to acquire two images from the same lateral position, the zposition (i.e. the height of the focus with respect to the substrate) of the maximum scattered photons at the boundary of substrate/medium (µ-dish plate/cell medium) was detected, which helped in determining the absolute axial position of this boundary layer . Two imaging slices were acquired 3 and 2.2 µm above the substrate in order to resolve the extracellular (attached with the cells) and intracellular (internalized by the cells) capsules, respectively. For imaging of both slices the pinhole was adjusted to capture 2 µm thick image sections. A software based autofocus routine was always set to recover the imaging alignment before each time point of measurement . Image dimensions were further reduced after determining the projection along the z-axis of each slice at the end of image acquisition.
## I.4) particle tracking by clsm, image processing, and data evaluation
The experimental protocol and description follows a previous study by Hartmann et al. . Large CLSM images were cut in small segments covering the whole internalization path of given capsules. A Matlab (Mathworks, USA) based particle analyzing toolbox written by Raimo Hartmann was used for visualization, identification, particle (capsules) segmentation, tracking, and automated image processing, as described previously . In this process, by means of median filtering photon shot noise was decreased in the red and yellow (green in false colors) channels of CLSM images and the background was entirely eliminated after defining a threshold. A modified Hough transformation was computed by using the method of Gonzalez, et al., for automatic identification, and masking the PEM capsules from CLSM images. This mask was later used for defining regions of interest (ROIs) around the regions of maximum fluorescence intensity. The dimensions of PEM capsules resembled circles, that is why the centers of all circles were identified by checking the positions of individual pixels whether they match the origin of a circle of given radius. From the classical Hough transformation the radius of the circles cannot be determined. That is why the parameterization of a donut shaped structure (shell thickness ≈ 0.64 µm) was used in order to identify the positions and dimensions of individual capsules. The radius of this structure was varied for each pixel and the sum of the underlying pixel intensities was recorded as a function of this radius. The capsule radius was determined from the first maximum of this function. Individual capsules were trace as described in Hartmann et al . A mask for each capsule's ROIs (in the red and yellow fluorescence channel) was created by using their respective x-y coordinates and radii. The average red/yellow ratio designated as I r /I y corresponding to the ambient pH values was calculated from the noise-corrected raw data within the masked regions. Taking inspiration by the particle tracking algorithms developed by John Crocker et al, which were provided by Danial Blair and Eric Dufresne for Matlab, a capsule tracking function was applied in order to get the progression of I r /I y versus time for individual capsules during their uptake and cellular trafficking (i.e., so called trajectories). A brief description of the workflow for the whole image processing is described in a previous study .
Some representative examples of trajectories of SNARF capsules having different surface charge and bilayers in serum-supplemented and serum-deprived media and in the presence and absence of inhibitors are provided in The trajectories of individual internalized capsules were used to determine their acidification (∆t A ) and processing (∆t P ; e.g., ∆t P10% and ∆t P50% ) times upon internalization . The acidification time ∆t A was determined from the sigmoidal curve as the time interval describing the duration of acidification, i.e. the duration of the transition from high to low pH. The processing time ∆t P50% is defined as the time interval from once a capsule attaches to the cell membrane until it becomes located in acidic endosomes/lysosomes . ∆t P10% is the time interval from the first contact of the PEM capsule with a cell until 10% of the final drop in pH have happened. The time after addition of the capsules to the cells until the first contact of the observed capsule with a cell is defined as time of first contact t C . Examples are shown in
[formula] (I r /I y )(t) = (R a -R b )/(1+exp((t-t 50% )/∆T))+R b (Equation SI-I.4.1) [/formula]
From each fit 4 fit parameters were obtained: R a = (I r /I y ) max , the I r /I y ratio when the capsules are still in the cell medium, i.e. at neutral/slightly alkaline pH R b = (I r /I y ) min , the I r /I y ratio when the capsules are fully internalized ,i.e. at maximum acidic pH in endo/lysosomal compartments t 50% , the time at which upon internalization the pH has dropped, so that 50% of the acidification has been achieved ∆T, describing the length of the acidification period in which the pH drops From these fit parameters the following set of parameters is obtained. t 10% and t 90% are the times when upon internalization 10% and 90% of the pH drop have been achieved:
[formula] (I r /I y )(t 10% ) = (R a -R b )/(1+exp((t 10% -t 50% )/∆T))+R b = (0.9⋅(R a -R b ) + R b ⇒ 1+exp((t 10% -t 50% )/∆T) = (R a -R b )/(0.9⋅(R a -R b ) = 10/9 ⇒ exp((t 10% -t 50% )/∆T) = 1/9 ⇒ (t 10% -t 50% )/∆T = ln(1/9) ⇒ t 10% -t 50% = ln(1/9)⋅∆T [/formula]
In an analogous way one can calculate
[formula] (I r /I y )(t 90% ) = (R a -R b )/(1+exp((t 90% -t 50% )/∆T))+R b = 0.1⋅(R a -R b ) + R b ⇒ t 90% -t 50% = ln(9)⋅∆T [/formula]
This leads to the definition of the acidification time ∆t A , which is the time interval in which the pH drop goes from 10% to 90%: ∆t A = t 90% -t 10% = ln(0.9)⋅∆T -ln(1/9)⋅∆T = 2⋅ln(9)⋅∆T (Equation SI-I..
## 2)
Based on this also the processing time ∆t P10% and ∆t P50% are calculated: t 10% = ln(1/9)⋅∆T + t 50% t 90% = ln(9)⋅∆T + t 50% ⇒ ∆t P10% = t 10% -t C (Equation SI-I.∆t P50% = t 50% -t C These are the time intervals needed from the first contact of a capsule with a cell unto 10% and 90% of the pH drop in the locale capsule environment due to internalization has been achieved. t 50% in contrast refers to the start of incubation t 50% = t C + ∆t P50%. It refers to the time needed from the start of incubation until 50% acidification has been achieved. In addition, the maximum of the absolute slope of the pH response was determined as (|d(I r /I y )/dt|) max . Thus, from the fit of each position/pH trace the following 5 parameters were obtained: t C , t A , t P10% , t P50% , (|d(I r /I y )/dt|) max . Some examples are given in the following. Note, that the bafilomycin A1 is lysomotropic reagent which alkalinizes the intra-lysosomal pH and inhibits the internalization of PEM capsules. The experimental concentration of the reagent was selected that it (partly) inhibited capsule internalization, and the intra-lysosomal pH remained less than pH of the extracellular medium. That is the reason why the internalized capsules turn orange instead of yellow (i.e. less acidic environment than without the presence of bafilomycin A1). Moreover, it was hard to identify the point of inflection in the I r /I y traces, i.e. the parameter t 50% . Thus, only a limited number of experiments were performed with this reagent and the parameters (e.g. acidification, processing time, etc.) are not provided in results.
## I.5) fractal dimension and the average end-to-end scaling exponent
Similarly other parameters, such as fractal dimensions D and the average end-to-end scaling exponent ν, can be determined from this data set. These data are based on the particle trajectories in the CLSM images. The position of each capsule at time t is given by its x-and ycoordinates x(t) and y(t). At images were taken in time intervals of 2 min, discrete coordinates x i = x(t i ) and y i = y(t i ) were obtained with t i+1 -t i = 2 min.
To characterize the trajectories quantitatively one can make use of the fractal dimension D, which is a measure of self-similarity and, thus, remains unchanged when the scale of measurement is changed. D is also a measure of spatial extent, i.e. the space filling properties, and self-affinity. The correct determination of the fractal dimension of a trajectory is a nontrivial problem, several approaches exist in the literature. Adopting the method by Sevcik et al.the fractal dimensions D of the full capsule trajectories were approximated by the fractal dimensions D M of each capsule trajectory consisting of M sample points. In doing so, the trajectory (x,y)(t) in the two-dimensional plane of the M sample points (each at time t i , i = 1,…, M) was mapped into a unit square Where, x/y max/min is the maximum/minimum of x/y i . Then, the fractal dimension D of the trajectory was approximated by
[formula] x i → |x i − x min | |x max− x min | y i →D ≈ D M = 1 + log L log(2M−2) (Equation SI-I.5.2) [/formula]
Where, L is the contour length of the trajectory in the unit square:
[formula] L = ∑ ∆L i M i=1 = ∑ �∆x i 2 + ∆y i 2 M i=1 (Equation SI-I.5.3) with ∆x i = |x i+1 − x i | etc. [/formula]
The average end-to-end distance〈R(L)〉is a strong characteristic of the spatial structure of polymers. Given a polymer of contour length L, the average end-to-end distance <R(L)> scales as
[formula] 〈R(L)〉 ∝ L ν (Equation SI-I.5.4) [/formula]
The exponent ν depends on the dimension of the system, taking values from ν = 1 to ν = 0. Similar to this scaling behavior of polymers one can use the average end-to-end distance scaling exponent to characterize the trajectory of a capsule. The scaling exponent ν is then given by
[formula] ν = log〈R(L)〉 log L (Equation SI-I.5.5) with 〈R(L)〉 = �(x M − x 0 ) 2 + (y M − y 0 ) 2 and L = ∑ (∆x i 2 + ∆y i 2 ) 1/2 M i=1 [/formula]
, where, x M and y M are the last sample points of the trajectories of each capsule along the x-and y-dimension. Note, that for the scaling exponent the trajectories were not mapped into a unit square, but the absolute values were used.
# I.6) results
The results from uptake data of capsules of almost similar sizes in terms of various surface charge, presence and absence of serum, and cytochalasin D inhibitor by determining various parameters from the trajectories of internalized particles are presented in the following. Each data set corresponds to at least 100 different trajectories that were evaluated, the exact number is given in Table SI-I.6.1. Two independent complete set of experiments with two batches of SNARF-loaded capsules were performed in order to validate the effect of cytochalasin D (300 nM, i.e. a potent inhibitor of particle internalization by disrupting actin polymerizationon the uptake and intracellular processing of these capsules. Results are listed individually for both batches. The extracted parameters t C , ∆t A , ∆t P50% , ∆t P10% , t P50% , and (|d(I r /I y )/dt|) max are enlisted in-I.6.1. Number of traces n that were evaluated for SNARF-loaded capsules added to HeLa cells with positive ("+", 2 bilayers) and negative ("-", 2.5 bilayers) charge, with serumsupplemented ("w") and serum deprived ("w/o") conditions, with ("w": two independent experiments "w 1 " and "w 2 ") and without ("w/o") the presence of 300 nM cytochalasin D, as determined from the trajectories of capsules while cellular uptake. Contact time t C of SNARF-loaded capsules added to HeLa cells with positive ("+", 2 bilayers) and negative ("-", 2.5 bilayers) charge, with serum-supplemented ("w") and serum deprived ("w/o") conditions, with ("w": two independent experiments "w 1 " and "w 2 ") and without ("w/o") the presence of 300 nM cytochalasin D, as determined from the trajectories of capsules while cellular uptake. The data for the contact time is provided as median values t C , plus minus the confidence intervals ∆t C .
[formula] + w/o w/o 222 + w/o w 1 141 + w/o w 2 87 - w w/o 120 - w w 1 125 - w w 2 106 - w/o w/o 398 - w/o w 1 121 - w/o w 2 105 [/formula]
## Capsule charge serum cytochalasin d ∆t a [min] ∆∆t a [min]
[formula] + w w/o 45 6 + w w 1 56 8 + w w 2 66 9 + w/o w/o 25 2 + w/o w 1 44 9 + w/o w 2 61 22 - w w/o 43 7 - w w 1 35 4 - w w 2 37 4 - w/o w/o 16 1 - w/o w 1 22 2 - w/o w 2 38 5 [/formula]
Table SI-I.6.3. Acidification time ∆t A of SNARF-loaded capsules added to HeLa cells with positive ("+", 2 bilayers) and negative ("-", 2.5 bilayers) charge, with serum-supplemented ("w") and serum deprived ("w/o") conditions, with ("w": two independent experiments "w 1 " and "w 2 ") and without ("w/o") the presence of 300 nM cytochalasin D, as determined from the trajectories of capsules while cellular uptake. The data for the acidification time is provided as median values ∆t A , plus minus the confidence intervals ∆∆t A . Processing time ∆t P50% of SNARF-loaded capsules added to HeLa cells with positive ("+", 2 bilayers) and negative ("-", 2.5 bilayers) charge, with serum-supplemented ("w") and serum deprived ("w/o") conditions, with ("w": two independent experiments "w 1 " and "w 2 ") and without ("w/o") the presence of 300 nM cytochalasin D, as determined from the trajectories of capsules while cellular uptake. The data for the 50% processing time is provided as median values ∆t P50% , plus minus the confidence intervals ∆∆t P50% . Table SI-I.6.5. Processing time ∆t P10% of SNARF-loaded capsules added to HeLa cells with positive ("+", 2 bilayers) and negative ("-", 2.5 bilayers) charge, with serum-supplemented ("w") and serum deprived ("w/o") conditions, with ("w": two independent experiments "w 1 " and "w 2 ") and without ("w/o") the presence of 300 nM cytochalasin D, as determined from the trajectories of capsules while cellular uptake. The data for the 10% processing time is provided as median values ∆t P10% , plus minus the confidence intervals ∆∆t P10% .-I.6.6. Time t 50% until 50% acidification of SNARF-loaded capsules added to HeLa cells with positive ("+", 2 bilayers) and negative ("-", 2.5 bilayers) charge, with serum-supplemented ("w") and serum deprived ("w/o") conditions, with ("w": two independent experiments "w 1 " and "w 2 ") and without ("w/o") the presence of 300 nM cytochalasin D. The data is provided as median values t 50% , plus minus the confidence intervals ∆t 50% ..7. Maximum of the absolute first derivative of the I r /I y versus t traces of SNARFloaded capsules added to HeLa cells with positive ("+", 2 bilayers) and negative ("-", 2.5 bilayers) charge, with serum-supplemented ("w") and serum deprived ("w/o") conditions, with ("w": two independent experiments "w 1 " and "w 2 ") and without ("w/o") the presence of 300 nM cytochalasin D, as determined from the trajectories of capsules while cellular uptake. The data is provided as median values (|d(I r /I y )/dt|) max , plus minus the confidence intervals, i.e., ∆(|d(I r /I y )/dt|) max . The fractal dimensions D and end-to-end distance scaling exponents ν were calculated for each trajectory for three different regions of the trajectory: before acidification starts (i.e. capsules are still outside cells "out"), during acidification ("uptake"), and after acidification (i.e. when the capsules are fully internalized "in"). Values are provided inHeLa cells with positive ("+", 2 bilayers) and negative ("-", 2.5 bilayers) charge, with serumsupplemented ("w") and serum deprived ("w/o") conditions, with ("w": two independent experiments "w 1 " and "w 2 ") and without ("w/o") the presence of 300 nM cytochalasin D. Trajectories were subdivided in the parts before acidification (D out ), during acidification (D uptake ) and after acidification Average end-to-end distance scaling exponents ν of the trajectories of SNARFloaded capsules added to HeLa cells with positive ("+", 2 bilayers) and negative ("-", 2.5 bilayers) charge, with serum-supplemented ("w") and serum deprived ("w/o") conditions, with ("w": two independent experiments "w 1 " and "w 2 ") and without ("w/o") the presence of 300 nM cytochalasin D. Trajectories were subdivided in the parts before acidification (ν out ), during acidification (ν uptake ) and after acidification (ν in ). Results are presented as mean values ± standard deviations. In the following possible differences in D and ν before, during, and after acidification are discussed, see
[formula] + w w/o D out , D in identical, small difference to D uptake ν equal in all 3 phases + w w 1 D uptake , D in identical, difference to D out ν uptake , ν in identical, difference to ν out + w w 2 D uptake , D in identical, difference to D out ν uptake , ν in identical, difference to ν out + w/o w/o D uptake , D in identical, difference to D out ν all different + w/o w 1 D increasing from D out towards D uptake to D in ν decreasing from ν out towards ν uptake to ν in + w/o w 2 D uptake , D in identical, difference to D out (not significant) ν uptake , ν in identical, difference to ν out - w w/o D out , D in identical, difference to D uptake (maybe not significant) ν equal in all 3 phases - w w 1 D uptake , D in identical, difference to D out (maybe not significant) ν uptake , ν in identical, difference to ν out - w w 2 D uptake , D in identical, difference to D out (maybe not significant) ν uptake , ν in identical, difference to ν out (maybe not significant) - w/o w/o D out , D uptake identical, difference to D in ν out , ν uptake identical, difference to ν in - w/o w 1 D increasing from D out towards D uptake to D in (large error in D uptake ) [/formula]
ν decreasing from ν out towards ν uptake to ν in (large error in ν uptake ) -w/o w 2 D increasing from D out towards D uptake to D in (maybe not significant) ν decreasing from ν out towards ν uptake to ν in However, the scaling exponent ν differs between ~0.67 and 0.45, which is more significant, because it corresponds to a difference of almost 50%. Thus, if one takes into account the correlation between the fractal dimension D and the scaling exponent ν (in general one would expect that when D → 1 (straight line) then also v → 1 and vice versa, which is in general true for the experimental data), then one might be able to extract some information about the different phases and between the different experiments using the data analysis. In that case, one might be able to make the statement that in some cases the existence of the different phases can be predicted from D and ν. It is however open to discussion whether differences are statistically significant. Considering the fractal dimension without the scaling exponent seems to result in rather non significant statements. However, taking both parameters into account may provide a chance to distinguish between the different phases before, during, and after acidification.
## Ii) uptake studies of snarf-loaded capsules by cells using fluorescence microscopy
## Ii.1) materials, reagents, and equipment ii.2) synthesis and characterization of snarf-loaded polyelectrolyte micro-(pem) capsules ii.3) uptake studies based on capsule counting ii.4) results
## Ii.1) materials, reagents, and equipment
Paraformaldehyde (8%; #157-8-100) was purchased from Electron Microscopy Sciences.
## Ii.2) synthesis and characterization of snarf-loaded polyelectrolyte micro-(pem) capsules
The protocol for capsule fabrication is same as described in section §I.2. In addition to SNARFmodified dextran, TRITC-modified dextran was also used as fluorophore, due to the much lower price as compared to SNARF. TRITC-loaded capsules were prepared following the same methodology, by using TRITC-dextran instead of SNARF-dextran at the same concentration.
The hydrodynamic diameter and zeta potential measurement results of TRITC-loaded PEM capsules in water are provided in and zeta potential ζ ± Δζ as measured in water given as average and standard deviation. These data were derived from For SNARF-loaded capsules the uptake study was carried out at various time points. HeLa cells were seeded into 8 well µ-ibidi plates (surface area 1 cm 2 /well) in 0.3 mL of complete growth medium at a density of 20,000 cells per well. After 18 h the growth media was exchanged to fresh growth media. Capsules were then added to cells. Uptake experiments were performed with two types of growth media, either supplemented with 10% FBS or without serum.
Positively and negatively charged capsules were added with two different concentrations, (i) at 10 capsules added per seeded cell, and (ii) at 20 capsules added per seeded cell. Cells were incubated inside an incubator at 37 - C with 5% CO 2 supply. Fluorescence microscopic images were recorded after 2, 4, 6, 12, 24, 30, 36, 42, and 48 hours of incubation with the help of an Axiovert fluorescence microscope using a 63x oil immersion objective. The fluorescence of SNARF was excited at 540 nm and its emission was recorded at 580 nm and 640 nm to detect the yellow and red fluorescence, respectively. Per sample and time point 10 images were captured, covering an average of 80 -100 cells per condition. Experiments were performed in duplicate. A representative fluorescence microscope image of HeLa cells after 2 h incubation with SNARF-loaded capsules is shown in-II.2.2a. Internalized capsules could be identified by their yellow fluorescence, whereas capsules remaining outside cells were fluorescent in red.
In order to evaluate the internalization of TRITC-loaded capsules, HeLa cells were seeded on small cover slips placed inside 24-well plates (surface area 1.82 cm 2 /well) in 0.6 mL complete growth medium at a density of 20,000 cells per well. After 22 h when the cells were firmly attached to the cover slips the growth media was aspirated and fresh growth media containing TRITC-loaded capsules was added. Uptake experiments were performed with two types of growth media, either supplemented with 10% FBS or without serum. Cells were incubated with positively and negatively charged capsules at a concentration of 10 capsules added per seeded cell and were incubated for 24 h inside an incubator at 37 - C with 5% CO 2 supply. In a series of experiments Bafilomycin A1 (0.25 µM) was added in order to investigate its effect as inhibitor on capsule uptake. Bafilomycin A1 was added to cells with fresh growth media 1 h before the addition of capsules. The capsules were subsequently immunostained (cellular cytoskeleton, nuclei, and lysosomes), so that the intracellular localization of TRITC capsules could be detected by means of CLSM images. In this way internalized capsules surrounded by lysosomal membrane, could be distinguished from capsules remaining in the extracellular medium. For the immunostaining procedure, the cells were washed with PBS and fixed with 4% paraformaldehyde solution in PBS (20 min incubation at room temperature), followed by washing with PBS. The cells were washed thrice with Hanks' balanced salt solution (HBSS; PBS would have served the same purpose), and permeabilized (addition of glycin 5 mg/mL and saponin 0.5 mg/mL in PBS, in the following referred to as permeabilization solution, 5 min incubation). Then the cells were incubated at 37 - C in an incubator under 5% CO 2 supply with blocking solution (20 mg/mL BSA in permeabilization solution) for 30 min. Next, the lysosomes were immunostained by LAMP 1 (primary antibody; 5 µg/mL blocking sloution, 1 h incubation at 37 - C followed by 3 times washing with blocking solution) and Dylight 649 (secondary antibody; 1.25 µg/mL PBS, 1 h incubation at 37 - C followed by 3 times washing with PBS). In order to save time the staining agents phalloidin labelled with oregon green 488 (20 µg/mL) and Hoechst reagent (for staining of the cytoskeleton and nuclear membranes, respectively) were added within the secondary antibody solution in PBS to cells. The immunostained cells were washed with PBS, followed by water, dried, and fixed on glass slides by means of fluoromount G. The samples were kept in dark at room temperature for 48 h before analyzing by CLSM. CLSM images were captured using a 63x/1.40 oil-immersion DIC M27 lens. For visualization of different stained cellular compartments and internalized capsules, the fixed samples were excited at 405, 488, 543 and 633 nm, respectively. Hoechst 33342 stained nuclei were excited at 405 nm, and emission of the dye was observed between 420 and 480 nm. The cytoskeleton stained with phalloidin, oregon green 488 was visualized by exciting the fluorophore at 488 nm, and emission was observed between 505 and 550 nm. The fluorescence of TRITC was excited at 543 nm, and emission was recorded using a 560 nm long pass filter. The antibody-labeled lysosomes were excited at 633 nm, and their emission was recorded using a 650 nm long pass filter. Per sample 25 images were captured, covering an average of 100 cells per condition. Experiments were performed in triplicate. A representative CLSM image of TRITC capsules after 24 h incubation with HeLa cells (after immunostaining) is shown in-II.2.2b.
## Figure si-ii.2.2:
Representative a) fluorescence microscopy and b) CLSM images of HeLa cells exposed to SNARF-loaded and TRITC-loaded capsules, respectively. The image in a) was captured after 2 h of incubation with SNARF-loaded negatively charged capsules at 10 capsules added per seeded cell in serum supplemented media. From the color change of the SNARFloaded capsules (red to yellow (shown in green false-colors)), the extracellular capsules can be discriminated from the intracellular capsules. The scale bar corresponds to 10 µm. In b) the HeLa cells were incubated with TRITC-loaded capsules, negatively charged capsules at 10 capsules added per seeded cell in serum supplemented media, and the image was taken after 24 h incubation. From the orthogonal view the presence of TRITC-loaded capsules inside the stained lysosomes can be seen. The scale bar corresponds to 10 µm.
## Ii.3) uptake studies based on capsule counting
In order to evaluate the uptake of capsules inside HeLa cells, first the microscopy images were exported and their format was changed from ".ZVI" to ".JPG". The number of internalized capsules per cell N caps/cell was then manually counted for each time point of incubation for each condition. In case of SNARF-loaded capsules the color change from red to yellow fluorescence was used as indicator of internalization. The internalization of TRITC-loaded capsules verified in the orthogonal CLSM images of immunostained samples by the presence of capsules within stained lysosomal compartments . For each condition by counting, a histogram in which the number of cells f(N caps/cell ) which had internalized N caps/cell capsules was made. From this, cumulative probability / cumulative distribution function (CDFs) p(N caps/cell ) was calculated, see-II.3.1:
[formula] p(N caps/cell ) = � f(i) N caps/cell i=0 � f(i) ∞ i=0 [/formula]
; 0 ≤ p(N caps/cell ) ≤ 1 (Equation SI-II.3.1) p(N caps/cell ) is the probability that a cell has internalized not more than N caps/cell capsules per cell (i.e. less than N caps/cell + 1 capsule). 1 -p(N caps/cell ) is the probability that a cell has internalized more than N caps/cell capsules per cell (i.e. at least N caps/cell + 1 capsule per cell).
## From the histograms the mean number of internalized capsules per cell <n caps/cell > (h) (t) at each time point t can be calculated by summing up the intensities of all capsules.
The normalized fluorescence intensity I inside one cell due to the fluorescence of capsules is proportional to <N caps/cell > (h) (t).
[formula] <I(t)> (h) ∝ <N caps/cell > (h) (t) = � f(i, t) - i ∞ i=0 (Equation SI-II.3.2) [/formula]
Alternatively, from the CDFs the mean number of capsules <N caps/cell > (p) (t) that were internalized with 50% probability (i.e. p(<N caps/cell > (p) (t)) = 0.5) after incubation time t was derived. The way to calculate this number was explained in-III.4.1 from Zyuzin et al.
# Ii.4) results
The charge t up[h] < 1 < 1 < 1 < 1 < 1 < 1 < 1 < 1
[formula] + + + + - - - - serum w w/o w w/o w w/o w w/o N [/formula]
## Iii) flow cytometry analysis of snarf capsules internalized by cells
## Iii.1) materials, reagents, and equipment iii.2) synthesis and calibration of snarf-loaded polyelectrolyte micro-(pem) capsules iii.3) uptake studies by flow cytometry iii.4) results
## Iii.1) materials, reagents, and equipment
4',6-diamidino-2-phenylindole, dilactate (DAPI, #D3571) was purchased from Invitrogen (Germany). All other reagents, chemicals, consumables were the same as described in sections §I.1 and §II.1.
For flow cytometry a BD LSRFortessa™ cell analyzer flow cytometer (Becton, Dickinson and Company, USA) was used. The fluorescence of the SNARF was excited at 561 nm and its emission was captured using 586 nm/15 nm and 670 nm /30 nm band pass filters for detecting the intensity of yellow and red fluorescence, respectively. Data was analyzed with FlowJo, single cell analysis software (Ashland, OR, USA).
## Iii.2) synthesis and calibration of snarf-loaded polyelectrolyte micro-(pem) capsules
The method for the fabrication of SNARF-loaded capsules was the same as described in section §I.2.
Calibration curves of capsules immersed in different pH were recorded with flow cytometry, in order to distinguish between populations of capsules surrounded by medium with different pH. For this positively and negatively charged SNARF-loaded capsules were mixed with buffers having a range of pH values from 3 -10. Solutions were analyzed with flow cytometry and 10,000 events per sample were recorded by exciting SNARF at 561 nm and recording its emissions at 586 nm and 670 nm. Results are plotted in different presentations. In
## Iii.3) uptake studies by flow cytometry
For investigation capsule uptake by cells, HeLa cells were seeded in 24 well plates (growth area 1.82 cm 2 /well) at 40,000 cells seeded per well in 0.6 mL complete cell growth media. Cells were kept inside an incubator set to 37 °C with 5% CO 2 . After 24 h the cells were provided fresh growth media (either serum supplemented or serum deprived) containing positively or negatively charged SNARF-loaded capsules at a density of 10 or 20 capsules added per seeded cell. Cells which were not exposed to capsules served as control. The cells were incubated with the capsules for different times, i.e., 2, 4, 6, 12, 24, 30, 36, 42, and 48 h. After incubation cells were washed with PBS, trypsinized, and resuspended in PBS. For viability assessment, DAPI (2 µL, 1.09 mM) was added to each sample. The samples were then investigated with flow cytometry. The forward (FSC-A) and sideward (SSC-A) scattering signals were used to gate events involving cells, i.e. only events with sufficient scattering signal were further regarded. This gating, which in Figure SI-III.3.1 is referred to as G1, excludes event due to cellular debris and free capsules, which have a much lower scattering signal than cells (compare with. By means of a second gating (G2), cell doublets were removed and excluded from the analysis. This was done in the forward scatter plot in which the area (FSC-A) was plotted against width (FSC-W). In order to detect a single cell population, 5000 events/sample were recorded in G2 and were used for further data processing. Using both gates G1 and G2, 2dimensional density plots of the red and yellow fluorescence signals I r and I y were created. In these plots populations of cells with adherent and internalized SNARF-loaded capsules can be distinguished, as internalized capsules are located in acidic environment. Events without sufficient fluorescence were attributed to cells which had neither capsules adherent to their outer wall, nor internalized capsules. In this way from each plot three different cell populations were identified: cells without associated capsules, cells with adherent capsules, and cells with internalized capsules. From the density plots the fractions of the respective populations were derived according to the amount of detected respective events, see In order to collect the single cell population, 5000 events/sample were recorded in G2, which were used for further data processing. c) 2-dimensional density plots of red and yellow fluorescence signals enable to distinguish between the populations of cells with adherent (N cells w caps(adh) /N cells ) and internalized capsules (N cells w caps(in) /N cells ), by integrating the events above and below the separation line, respectively.
# Iii.4) results
In-III.4.1, the standard uptake curve in which fluorescence intensity per cell due to internalized capsules is plotted versus time is shown. This curve does not used any gating strategy. From this curve the mean intensity <I y > (sat,c) per cell under saturation and the mean time t up(sat,c) until the mean fluorescence intensity <I y > (c) per cell has reached 50% of the saturation value can be determined, see
## Iv) elemental analysis of gold nanoparticle (gnp)-loaded capsules internalized by cells
## Iv.1) materials and reagents iv.2) synthesis and characterization of gold nanoparticle (gnp) -loaded polyelectrolyte micro-(pem) capsules iv.3) uptake studies by inductively coupled plasma mass spectroscopy (icp-ms) iv.4) results
## Iv.1) materials and reagents
6 well cell culture plates (# 83.1839.300), eppendorf tubes (2 mL; # 72.695.500), and falcon tubes (15 and 50 mL; #62.554.502, #62.547.254, respectively) from Sarstedt were used. Hydrochloric acid (HCl; 35 wt%, ultra-pure, #7647010) and nitric acid (HNO 3 ; 67 wt%, ultra pure, #7697372) were purchased from Fisher Chemicals. Cell lysis buffer (5X reagent; #2018-02-12) was used from Promega Corporation. A total protein determination kit (micro-Lowry, Peterson's modification; TPO300-KT, batch# SLBF6513) was purchased from Sigma-Aldrich, consisting of, Lowry reagent powder (2 g; L3540-1VL, SLBD9543), and Folin and Ciocalteu's phenol reagent (F9252-1EA, Lot#SHBB8897V, Pcode: 1001449215). Commercially available gold nanoparticles (GNPs; 15 nm, #EM.GC15) were obtained from BBI solutions. Amino dextran (70 kDa, #D1862) was purchased from Thermo Fisher Scientific. All other reagents, chemicals, and consumables, were the same as described in sections §I.1 and §II.1.
A UV-vis absorption spectrometer (8453 UV-visible spectrophotometer) from Agilent was used for obtaining the absorption values of the protein content in the Lowry tests. An inductively coupled plasma mass spectrometer (ICP-MS) from Agilent 7700 Series was used to determine the concentrations of elemental gold and hence the GNP concentrations.
## Iv.2) synthesis and characterization of gold nanoparticle (gnp) -loaded polyelectrolyte micro-(pem) capsules
In order to fabricate GNP-loaded PEM capsules comparable to SNARF-loaded capsules, the basic strategy for the synthesis was same as that for SNARF-loaded capsules (cf. § I.2). Briefly, amino dextran was co-precipitated with CaCO 3 particles, by mixing the solutions of CaCl 2 and Na 2 CO 3 under vigorous stirring in the presence of amino dextran at room temperature (RT). In a glass vial, 1.4 mL amino dextran (0.5 mg/mL) was added to 1 mL of 0.33 M CaCl 2 (0.33 M). Under vigorous magnetic stirring (1100 rpm) 1 mL of Na 2 CO 3 (0.33 M) solution was quickly mixed with the above mixture for 30 s followed by keeping the reaction contents intact for 2 min. The CaCO 3 particles were washed three times with DDW and used for LBL assembly of oppositely charged polyelectrolytes (2 mg/mL, 0.5 M NaCl). The alternating layers of negatively and positively charged polymers, i.e., PSS and PAH, respectively, were deposited around the charged sacrificial CaCO 3 particle templates. Layer-by-layer deposition was achieved by alternating immersion of the particles inside the respective polyelectrolyte solutions (3 mL) for 13 min, followed by subsequent washing with DDW to remove excess of polymers. Negatively charged GNPs (2.7 mL of the solution as purchased mixed with 0.3 mL 0.5 M NaCl just before layer deposition) were incorporated inside the capsules after the first bilayer (PSS/PAH), followed by the deposition of additional 1 (PSS/PAH) and 1.5 (PSS/PAH/PSS) bilayers for positively and negatively charged capsules, respectively. Finally the cores of the PEM capsules were dissolved by complexion of Ca 2+ ions with EDTA (3 mL, 0.2 M, pH 6.5) and the resulting capsules were washed with DDW and stored in water at 4 °C for further use.
Their hydrodynamic diameter and zeta potentials were recorded and are shown in and zeta potential ζ ± Δζ as measured in water given as average and standard deviation. These data were derived from Figure SI-IV.2.1.
## Iv.3) uptake studies by inductively coupled plasma mass spectroscopy (icp-ms)
HeLa cells were seeded in 6 well plates (8.95 cm 2 surface area/well, 3 mL medium) at a density of 100,000 cells per well. After 24 h incubation fresh growth media, either serum supplemented or serum free, containing GNP-loaded capsules was provided to the cells. Both types of capsules (positively and negatively charged) were added in two different concentrations, at 10 or 20 capsules added per cell. HeLa cells were incubated with capsules for different times , and 48 h) inside an incubator set at 37 - C with 5% CO 2 supply. Afterwards, the cells were washed with PBS and detached from the bottom of the plates by means of 0.05% trypsin-EDTA. The cells were centrifuged with growth media and the pellets were resuspended in PBS. The pellets of cells were washed again with PBS and the maximum of the supernatant was removed (note that remaining trypsin in solution would interfere with determining the protein concentration of the samples as described later). Then, 1x lysis buffer in water (100 µL per sample) was added to each cell pellet and the samples were incubated at room temperature for 30 minutes in order to complete cell lysis. Samples were sonicated and stored at -20 - C for further analysis, which involved determination of the amount of cells (by measuring the protein content) and the amount of capsules (by measuring the amount of elemental gold) in the cell lysates, as described in the following.
In order to determine the number of capsules per cell, the number of cells and the number of capsules in the cell lysate has to be determined. Cells were quantified by detecting their protein content by means of a commercial protein determination kit (Lowry assay; TPO300-KT). Lowry reagent solution was prepared by dissolving 2 g of Lowry reagent powder in 40 mL of water. The labeling solution was prepared by transferring 18 mL of Folin and Ciocalteu's phenol reagent into an amber glass, followed by the addition of 90 mL of water (in order to achieve the working concentration of labeling reagent). First, a calibration curve was plotted, relating the amount of the detected proteins to the number of cells. HeLa cells were detached from cell culture flasks by means of 0.05% trypsin-EDTA, followed twice by washing with PBS. The cells were then dispersed in a small volume of PBS and their number in terms of cells per volume of solution was determined by counting thrice with a haemocytometer. Then 1x lysis buffer was added. After cell lysis serial dilutions of the cell lysates in lysis buffer was performed in order to achieve samples with subsequently smaller cell concentrations. Lysis buffer was used for blank measurements. In sample tubes, 5 µL of cell lysates were added into 100 µL of Lowry reagent solution, mixed well, followed by waiting for 20 minutes for completion of complex formation between proteins of the cell lysates and the Lowry reagent solution. Later, 50 µL of labeling reagent solution was added to the above mixture to develop blue color depending on protein content. After waiting for 30 minutes the absorption of samples was immediately measured by means of UV-visible absorption spectrometry. Spectra were recorded from 550 -800 nm and from the spectra the absorption (A 750 ) at 750 nm was determined, cf. The amount of capsules in the cell lysates was determined in terms of elemental Au (from the GNPs) as measured with ICP-MS. The cell lysates containing the internalized GNP-loaded capsules were first digested with aqua regia. For this first the stock solution for each sample was prepared by adding 50 µL of sample suspension into freshly prepared 150 µL aqua regia (1 HNO 3 : 3 HCl) inside 6 mL perfluoroalkoxy alkane tubes (PFA), followed by mixing for at least 8 h under constant agitation. By doing this the GNP-loaded capsules and remaining organic cell fragments were digested and broken down into small molecular / atomic components. In the second step, 4.6 mL HNO 3 solution (2%) as low matrix was introduced to each digested sample to prevent the aqua regia from digesting the ICP-MS machinery, as well as to provide an ion stable environment with constant background conditions for all samples. During measurements, 5 repetitions/ sample, and 100 sweeps were performed and a peak pattern of 3 peaks was used. The diluted samples were introduced to the ICP-MS set-up through an integrated autosampler coupled to a peltier cooling spray chamber where the samples were nebulized and taken up by the argon gas flow at a speed of ½ m/s. The concentration determination was performed using a calibration curve for Au consisting of 9 measurement points (0 -2500 µg/L) of freshly prepared Au concentrations derived from gold standard solutions from Agilent (1000 mg/L). Results were obtained as the mean of five measurements in parts per billion or µg/L (ppb = µg/L). First, calibration was performed in which the Au content in the samples of GNP-loaded capsules of given number (as determined with a haemocytometer) was determined. From this the mass of elemental gold per capsule was determined. Then the amount of gold in the cell lysates of cells with incorporated GNP-loaded capsules was determined by ICP-MS. Using the calibration curve the number of capsules N caps in the lysate was determined. Finally, the number of internalized capsules per cell was determined as N caps/cell = N caps /N cells .
# Iv.4) results
The data obtained for the determined number of capsules per cell N caps/cell is summarized in-IV.4.1 it can be seen that the internalization of positively charged GNP-loaded capsules was higher than for their negatively charged counterparts. Uptake in serum deprived culture as compared to serum supplemented cell culture was also higher, which is in agreement with some previous findings. In addition, the uptake was dose dependent, i.e., more capsules were found internalized when cells were incubated with 20 capsules per cell as compared to the addition of 10 capsules per cell, which is also in agreement with previous work. Note that for interpretation of the results it is important to realize that the number of capsules added per cell refers to the number of cells seeded. However, the number of internalized capsules per cell, refers to the number of cells which have been found with the Lowry assay after incubation with capsules. After approximately 24 h HeLa cells start to proliferate, which complicates the quantitative analysis. Cell proliferation results in increase in cell number after capsule uptake as compared to the number of seeded cells.
## V) viability measurements of cells exposed to capsules and inhibitor
## V.1) materials and reagents v.2) viability measurements v.3) results
## V.1) materials and reagents
96 well assay plates (clear bottom, # 3603) were purchased from Corning. Resazurin solution (alamar blue; #TOX-8) was purchased from Sigma-Aldrich. All other reagents, chemicals, and consumables were the same as described in sections §I.1 and §II.1.
Fluorescence measurements were performed with a Fluorolog®-3 spectrofluorometer from HORIBA JOBIN YVON. For plate reading a Micromax 384 microwell-plate reader compatible with the Fluorolog® was used.
## V.2) viability measurements
Resazurin based cytotoxicity assays were performed to determine the impact of GNP-loaded PEM capsules and cytochalasin D (which was used in the studies presented in section §I) exposure on cell viability. The test is based on mitochondrial activity of living cells. Active mitochondria of living cells cause bioreduction of the dye, i.e. they convert the nonfluorescent blue dye (resazurin) into its reduced form resorufin which fluoresces in red. In order to perform the studies, HeLa cells were seeded in 96 well transparent bottom plates (7500 cells/well, 0.32 cm 2 area/well) in 100 µL of complete growth media (DMEM supplemented with 10% FBS, 1% glutaMAX TM and 1% P/S) and were incubated for 24 h at 37 - C with a constant supply of CO 2 (5%). After 24 h, the old growth media was replaced by fresh growth media containing capsules or cytochalasin D at different concentrations, i.e. N caps/cell (added), and c(Cytochalasin D). Capsules were added at N caps/cell (added) 10 or 20 capsules per cell for each condition (positive and negative capsules in serum supplemented and serum deprived cell culture media). In case of cytochalasin D serial dilution was performed to examine its toxic effect for a broad range of concentrations (2400 -4.5 × 10 -06 µM). Experiments for each dose (for capsules and cytochalasin D) were performed in triplicate. As negative control in a few wells of the assay plates, fresh growth media was added to the cells without capsules or cytochalasin D. In case of capsules fresh serum supplemented and serum deprived media was provided to the cells as negative control for each sample/condition. Whereas, for cytochalasin D as negative control fresh serum supplemented cell growth media was provided to the cells in the absence of cytochalalsin D. Cells were incubated with capsules for different time points, i.e., 2, 4, 6, 12, 24, 30, 36, 42, 48 h, while 24 h the incubation time with cytochalasin D was used. After incubation for defined time points, the cells were washed with PBS and 10% resazurin solution in growth media was added (100 µL/well) to the cells. In some wells of the assay plate only resazurin solution (10%) was added (without cells) which served as blank and the assay plates were incubated for 4 h under the same conditions as described above. Then, the fluorescence spectra of each well of the assay plates were recorded by means of a microplate reader attached to a spectrofluorometer. Spectra were recorded using an excitation wavelength of 560 nm acquiring the emission spectra from 572 to 650 nm. Background emission of the blank sample containing only resazurin solution was subtracted from all spectra. For getting the viability V of cells the mean of fluorescence value of each sample was normalized with respect to fluorescence of the control samples in which cells were not exposed to capsules or cytochalasin D. Experiments were performed in triplicate and values are expressed as means of 3 independent experiments ± standard deviations.
# V.3) results
The viability (V) of HeLa cells upon exposure to cytochalasin D and PEM capsules with incorporated GNPs is presented in |
Health impacts of parental migration on left-behind children and adolescents: a systematic review and meta-analysis
Background Globally, a growing number of children and adolescents are left behind when parents migrate.We investigated the effect of parental migration on the health of left behind-children and adolescents in low-income and middle-income countries (LMICs).Methods For this systematic review and meta-analysis we searched MEDLINE, Embase, CINAHL, the Cochrane Library, Web of Science, PsychINFO, Global Index Medicus, Scopus, and Popline from inception to April 27, 2017, without language restrictions, for observational studies investigating the effects of parental migration on nutrition, mental health, unintentional injuries, infectious disease, substance use, unprotected sex, early pregnancy, and abuse in left-behind children (aged 0-19 years) in LMICs. We excluded studies in which less than 50% of participants were aged 0-19 years, the mean or median age of participants was more than 19 years, fewer than 50% of parents had migrated for more than 6 months, or the mean or median duration of migration was less than 6 months. We screened studies using systematic review software and extracted summary estimates from published reports independently. The main outcomes were risk and prevalence of health outcomes, including nutrition (stunting, wasting, underweight, overweight and obesity, low birthweight, and anaemia), mental health (depressive disorder, anxiety disorder, conduct disorders, self-harm, and suicide), unintentional injuries, substance use, abuse, and infectious disease. We calculated pooled risk ratios (RRs) and standardised mean differences (SMDs) using random-effects models. This study is registered with PROSPERO, number CRD42017064871.FindingsOur search identified 10 284 records, of which 111 studies were included for analysis, including a total of 264 967 children (n=106 167 left-behind children and adolescents; n=158 800 children and adolescents of non-migrant parents). 91 studies were done in China and focused on effects of internal labour migration. Compared with children of non-migrants, left-behind children had increased risk of depression and higher depression scores (RR 1·52 [95% CI 1·27-1·82]; SMD 0·16 [0·10-0·21]), anxiety (RR 1·85 [1·36-2·53]; SMD 0·18 [0·11-0·26]), suicidal ideation (RR 1·70 [1·28-2·26]), conduct disorder (SMD 0·16 [0·04-0·28]), substance use (RR 1·24 [1·00-1·52]), wasting (RR 1·13 [1·02-1·24]) and stunting (RR 1·12 [1·00-1·26]). No differences were identified between left-behind children and children of non-migrants for other nutrition outcomes, unintentional injury, abuse, or diarrhoea. No studies reported outcomes for other infectious diseases, self-harm, unprotected sex, or early pregnancy. Study quality varied across the included studies, with 43% of studies at high or unclear risk of bias across five or more domains.Interpretation Parental migration is detrimental to the health of left-behind children and adolescents, with no evidence of any benefit. Policy makers and health-care professionals need to take action to improve the health of these young people.Funding Wellcome Trust.
# Introduction
Globally, nearly one in seven individuals are migrants. The majority are labour migrants who originate from lowincome or middle-income countries (LMICs) and relocate in search of employment opportunities either internationally or internally within a country (eg, from rural to urban settings). 1 Some individuals are forced to migrate because of acute drivers such as conflict and disasters. As a result of migration, especially low-skilled labour migration, children are often left behind in the care of other family members or carers. Among labour migrants, a key incentive for migration is the hope of improving the circumstances of their families and children through increased household income and financial stability. International migrants send an estimated US$613 billion per year in remittances to their countries of origin. Although the health and rights of migrant workers is recognised as a priority in the UN Sustainable Development Goals, 3 the health of children of migrants has been largely overlooked in research and policy.
No estimates are available for the number of leftbehind children and adolescents globally, but the figure is thought to be in the hundreds of millions. More than a third of all children residing in rural China are left behind by one or both migrant parents. 4 27% of children in the Philippines, 5 36% in Ecuador, 6 and more than 40% in rural South Africa 7 are estimated to be left behind.
Evidence about the health status of left-behind children is conflicting. On the one hand, material benefits and greater income security from remittances might be expected to confer improvements in health and facilitate access to health care and education. In Pakistan, migration had positive effects on the growth of left-behind children, with girls benefitting more than boys.However, on the other hand, some studies suggest poorer health outcomes among left-behind children. In China, where the most research has been done to date, studies have shown poorer nutritional, 9 developmental 10 and mental health outcomes [bib_ref] The mental health of children left behind in rural China by migrating..., Qin [/bib_ref] in left-behind children than children of non-migrant parents. It is unclear to what extent the health of these children is affected by parental migration, and how the impact might vary according to contextual factors, including sex and age. For example, in China, boys who were left before age 6 years were not as tall as boys whose parents left them at an older age. [bib_ref] Does the timing of parental migration matter for child growth? A life..., Zhang [/bib_ref] Although adolescents might be more independent than younger children, parental absence, and lack of supervision at this crucial age has been associated with risk-taking behaviours, including substance use and physical inactivity, with implications for long-term health. [bib_ref] The impact of parental migration on health status and health behaviours among..., Gao [/bib_ref] Furthermore, effects might vary according to the circumstances of parental migration. For example, maternal absence and the absence of both parents might have more pronounced effects on children's health than paternal absence alone. [bib_ref] Left too early: the effects of age at separation from parents on..., Liu [/bib_ref] To date, to our knowledge, no studies have comprehensively examined the health status of left-behind children and adolescents across all settings and key areas of health. To address this research gap, we did a systematic review and meta-analysis to assess the impact of parental migration on child and adolescent nutrition, mental health, unintentional injuries, infectious disease, substance use, unprotected sex, early pregnancy, and verbal, physical, and sexual abuse in LMICs. We investigated whether parental migration status (one or both parents migrating), type of migration (internal or international; labour or forced) and child characteristics (age, sex) differentially influence the health of left-behind children and adolescents.
# Methods
## Search strategy and selection criteria
For this systematic review and meta-analysis, we searched MEDLINE, Embase, CINAHL, the Cochrane Library, Web of Science, PsychINFO, Global Index Medicus, Scopus, and Popline from database inception to . Full search terms are provided in the appendix. We searched for observational studies reporting the risk of health outcomes done in LMICs
## Research in context
Evidence before this study Migration is increasing globally, which has resulted in a growing number of children and adolescents being left behind when their parents migrate. Before starting this study, we searched the scientific literature for articles on the effect of parental migration on child and adolescent health, and found two narrative reviews: one focused on left-behind children in the Philippines and the other on mental health outcomes in left-behind children in China. These reviews suggested that children benefited from the remittances their parents sent home in terms of improved education and reduced child labour, which could result in improved health, but reported that family separation might have long-term psychological and societal costs. We also identified more than 30 studies, mainly from China, investigating the effects of parental migration on a broad range of health outcomes across different countries. On Nov 26, 2018, we did an updated search of MEDLINE for systematic reviews with no date or language restrictions, using the broad search terms "(child* OR adolescent) AND health AND (migration OR left-behind)". Although we identified 99 systematic reviews, none reviewed the key areas of health of left-behind children and adolescents across all low-income and middle-income countries (LMICs).
## Added value of this study
This is the largest and most comprehensive study to date assessing the impact of parental migration on all key areas of child and adolescent health across all LMICs. Compared with children of non-migrants, left-behind children and adolescents had an increased risk of depression, suicidal ideation, and risk of anxiety. Left-behind children also had smaller increases in risk for wasting, stunting, and substance use. These results highlight a rarely discussed consequence of global migration with implications for global policy making and health-care provision in migrant-sending countries. Although a small number of individual studies found positive health effects of parental migration, overall we found no evidence of benefit across any of the health outcomes.
## Implications of all the available evidence
Our findings highlight the unmet health needs of left-behind children and adolescents. Research to date has focused primarily on China and longitudinal studies in a wider range of LMICs with high rates of emigration are needed to better understand risk and resilience factors within this population, and to inform policy and practice to address unmet health needs in left-behind children, adolescents, and their carers.
(classified according to the World Bank classification) that included children and adolescents (aged 0-19 years) with at least one migrant parent. We defined parental migration as one or more parent moving away from the place their children are living, for a minimum of 6 months. We included studies in which parents had migrated for any reason, such as employment (labour migrants) or armed conflict or disasters (forced migrants). We included internal and international parental migration, defined as migration within and beyond a country's borders, respectively. The comparator group was children of non-migrating parents. We excluded studies in which less than 50% of participants were aged 0-19 years, the mean or median age of participants was more than 19 years, fewer than 50% of parents had migrated for more than 6 months, or the mean or median duration of migration was less than 6 months.
We updated our searches to include all literature published before , to assess whether studies published after our original search might alter the implications of our findings. We tailored search strategies to each database and used controlled vocabulary and search filters where available, or Boolean search methods and free text terms. No date or language restrictions were used. Because of the large volume of research on leftbehind children available in China, we searched the China National Knowledge Infrastructure database and key Chinese public health journals. We also searched reference lists of relevant systematic reviews and grey literature published by key international organisations (UN Children's Fund, International Organization for Migration, and UN High Commissioner for Refugees). The full search strategy is detailed in the appendix. We used Covidence systematic review software (Veritas Health Innovation, Melbourne, VIC, Australia) to organise and screen articles. Two reviewers (KR-C, GF, CZ, LKB, YZ, HS, BE, AB, WL, MO, DK, or DD) independently screened each title and abstract and excluded those that were not relevant, and then independently screened the full text of remaining studies to assess eligibility. Two reviewers (GF, CZ, LKB, YZ, AM, HS, BE, RB, WL, MO, KR, or OM-A) extracted data, and assessed the risk of bias for all included studies. Discrepancies about study inclusion were resolved through discussion with a third reviewer or by contacting study authors. Studies that reported results as mean scores with SDs or as raw proportions or unadjusted odds ratios (ORs) were included in meta-analysis. When insufficient data were reported for inclusion in the meta-analysis, we contacted study authors to request further information.
This study is reported in accordance with the PRISMA guidelines [bib_ref] Preferred reporting items for systematic review and meta-analysis protocols (PRISMA-P) 2015 statement, Moher [/bib_ref] (appendix). The study protocol is available online.
# Data analysis
We extracted data on study design, participant numbers and characteristics, and exposures and outcomes using data extraction sheets designed by the authors (appendix). When duplicate data were identified, only data for the most recent timepoint were extracted.
Outcomes of interest were risk and prevalence of the main causes of disability-adjusted life-years for children aged younger than 5 years, 5-9 years, and 10-19 years, including nutrition (stunting, wasting, underweight, overweight and obesity, low birthweight, and anaemia), mental health (depressive disorder, anxiety disorder, conduct disorders, self-harm, and suicide), unintentional injuries, substance use, physical, emotional, and sexual abuse, [bib_ref] Global and national burden of diseases and injuries among children and adolescents..., Kyu [/bib_ref] and infectious disease. Additional outcomes were unprotected sex and early pregnancy (<18 years; appendix).
We summarised outcomes from all studies included in the review diagrammatically, using adjusted estimates to classify studies according to their effect estimates and provide a visual overview of the evidence. Studies with sufficient data to examine the effect of being left behind on nutrition, mental health, injury, and substance use outcomes were included in random-effects metaanalysis. We estimated pooled risk ratios (RRs) with 95% CIs for binary outcomes and standardised mean differences (SMDs) with 95% CI for continuous outcomes. Binary categorisations indicate the presence or absence of a disorder (caseness), and continuous outcomes were associated with symptom severity. We used unadjusted study outcomes for three main reasons. First, only 15 studies reported adjusted effect estimates. Second, the adjusted effects estimates varied considerably with regard to the covariates included and effects were not directly comparable. Third, a number of studies reported adjusted ORs; due to the so-called noncollapsibility property of ORs, 17 estimates from adjusted ORs can differ significantly from unadjusted estimates even in the absence of confounding. We used metaregression to assess the effect of child age and sex on study-specific effect estimates.
Risk of bias was assessed using an adapted version of the Newcastle Ottawa Scale 18 incorporating items from the National Institute for Clinical Excellence Quality Appraisal 19 (appendix). Studies with a high or unclear risk of bias across five or more domains were defined as being at high risk of bias. This definition was based on consensus between the authors, which acknowledged that any such cutoffs are arbitrary.No studies were excluded on the basis of quality. Funnel plots were used to assess publication bias.
We used the I² statistic to indicate the proportion of total variation between study estimates due to heterogeneity.To identify and assess sources of heterogeneity, we planned a-priori subgroup analyses to assess migration of one versus both parents, and forced versus labour migration. We also planned to do subgroup analyses of internal versus international migration, but due to the predominance of Chinese studies, which were all on internal migration, we decided to do subgroup analyses of studies in China versus the rest of the world. We did a sensitivity analysis to assess the robustness of our conclusions with regard to the assumptions underlying our analytic approach. We did fixed-effects meta-analyses and repeated analyses using only studies with low risk of bias. All statistical analyses were done using Stata (version 13.0) and MetaXL (version 5.3). The study is registered with PROSPERO, number CRD42017064871.
## Role of the funding source
The funder had no role in study design, data collection, data analysis, data interpretation, writing of the report, or the decision to submit. The corresponding author had full access to all the data in the study and had final responsibility for the decision to submit for publication.
# Results
Our systematic search of the literature identified 10 284 records, of which 2265 were duplicates (figure 1). Of the 396 full-text articles retrieved, 111 studies 9,10,13,21-128 done between 1994 and 2017 in 16 countries (n=106 167 leftbehind children and adolescents; n=158 800 children and adolescents of non-migrant parents) were included in the systematic review (appendix). 89 studies (n=78 273 leftbehind children and adolescents; n=88 350 children of non-migrant parents) were included in meta-analyses. Reasons for exclusion at the full-text screening stage and characteristics of included studies are shown in the appendix. Of the 111 included studies, 91 (82%) were done in China, 58 (52%) were published in Chinese, nine (8%) were done in Asia (Thailand, Philippines, Indonesia, Vietnam, Sri Lanka, and India), six (5%) in Latin America (Mexico, Guatemala, and Peru), three (3%) in Africa (Ethiopia, Kenya, and Malawi), two (2%) in the Caribbean (Trinidad and Tobago and Jamaica), and two (2%) in eastern Europe (Romania and Moldova). 101 studies were cross-sectional, seven were cohort studies, and three were case-control studies. All studies included children of labour migrants; none included children of forced migrants. All Chinese studies examined internal migration within China, whereas studies from the rest of the world, with the exception of one study, [bib_ref] Household headship and child nutrition: a case study in western Kenya, Onyango [/bib_ref] focused on international migration. 71 studies included children aged younger than 10 years. Among the 92 studies that reported participant sex, the proportion of male participants ranged from 13·1% to 76·3%. Study quality varied by domain assessed (figure 2). 48 (43%) of the 111 included studies were at high or unclear risk of bias across five or more domains. Funnel plots showed no evidence of publication bias (appendix).
Of the 111 studies included in the systematic review, mental disorders were the most common study outcome (n=64), followed by nutritional status (n=29), substance use (n=14), experience of violence and abuse (n=7), unintentional injury (n=6), and infectious disease (n=5; [fig_ref] Figure 3: Harvest plot of health outcomes among left-behind children and children of non-migrant... [/fig_ref]. Across all outcomes, only 12 studies reported a lower risk of adverse health outcomes among left-behind children and adolescents.
64 of 65 studies reporting mental health outcomes used self-reported screening tools. Meta-analysis showed that left-behind children and adolescents had a significantly higher risk of depression caseness (RR 1·52 [95% CI 1·27-1·82]) and symptoms (SMD 0·16 [95% CI 0·10-0·21]), anxiety caseness (RR 1·85 [1·36-2·53]) and symptoms (SMD 0·18 [0·11-0·26]), and suicidal ideation caseness (RR 1·70 [1·28-2·26]) compared with children of non-migrating parents (figures 4A, . Left-behind children and adolescents had a higher risk of symptoms of conduct disorder (SMD 0·16 [95% CI 0·04-0·28]) but not caseness (RR 1·16 [95% CI 0·88-1·52]). Statistical heterogeneity across mental disorder outcomes was high (I²=67·0-96·9%). In subgroup analyses, no differences were identified in risk of anxiety caseness or symptoms among children and adolescents left behind by one parent or by both parents compared with children and adolescents of non-migrant parents (appendix). Children and adolescents left behind by both parents had a higher risk of depressive symptoms than did those of non-migrants (SMD 0·11 [95% CI 0·01 to 0·22]), but no differences were identified between children or adolescents left behind by one parent compared with children or adolescents of nonmigrants (0·04 [-0·03 to 0·12]; appendix). No significant differences in depression caseness were identified between children and adolescents left behind by one (RR 0·93 [bib_ref] Psychometric characteristics of CES-D in a sample of Mexican rural adolescents in..., Aguilera-Guzman [/bib_ref] Asis and Ruiz-Marave (2013) [bib_ref] Leaving a legacy: parental migration and school outcomes among young children in..., Asis [/bib_ref] Ban et al (2017) [bib_ref] Child feeding and stunting prevalence in left-behind children: a descriptive analysis of..., Ban [/bib_ref] Battistella and Conaco (1998) [bib_ref] The impact of labour migration on the children left behind: a study..., Battistella [/bib_ref] Bi and Oyserman (2015) [bib_ref] Left behind or moving forward? Effects of possible selves and strategies to..., Bi [/bib_ref] Carling and Tonnessen (2013) [bib_ref] Fathers' whereabouts and children's welfare in Malawi, Carling [/bib_ref] Chen and Qu (2010)Chen et al (2010)Chen (2009) [bib_ref] The health status of the left-behind children in rural China, Chen [/bib_ref] Chen et al (2011) [bib_ref] Nutritional status of children during and post-global economic crisis in China, Chen [/bib_ref] Chen et al (2012)Chen et al (2013) [bib_ref] Analysis on the nutritional status of unattended children under 7 years in..., Chen [/bib_ref] Chen and Chan (2016) [bib_ref] Parental absence, child victimization, and psychological well-being in rural China, Chen [/bib_ref] Cheng et al (2008) [bib_ref] Mental health status of left-behind adolescents in rural areas in Anhui, Cheng [/bib_ref] Davis and Brazil (2016) [bib_ref] Migration, remittances and nutrition outcomes of left-behind children: a national-level quantitative assessment..., Davis [/bib_ref] Deng and Li (2014)Fan et al (2010) [bib_ref] Emotional and behavioral problems of Chinese left-behind children: a preliminary study, Fan [/bib_ref] Feng et al (2010) [bib_ref] Analysis on the physical growth and nutrition status of children under 7..., Feng [/bib_ref] Feng (2014) [bib_ref] Investigation and comparison of psychological and behavioral characteristics of rural left-behind children..., Feng [/bib_ref] Frank (2005) [bib_ref] International migration and infant health in Mexico, Frank [/bib_ref] Gao (2008) [bib_ref] A comparative study on mental development characteristics between non-parent-absent children and parent-absent..., Gao [/bib_ref] Gao et al (2010) [bib_ref] The impact of parental migration on health status and health behaviours among..., Gao [/bib_ref] Gao et al (2013) [bib_ref] Parental migration, self-efficacy and cigarette smoking among rural adolescents in south China, Gao [/bib_ref] Ge and Luo (2011) [bib_ref] The study of rural left-behind children's psychological anxiety, Ge [/bib_ref] Graham and Jordan (2011) [bib_ref] Migrant parents and the psychological well-being of left-behind children in southeast Asia, Graham [/bib_ref] Graham and Jordan (2013) [bib_ref] Does having a migrant parent reduce the risk of undernutrition for children..., Graham [/bib_ref] Guo et al (2012) [bib_ref] The relationship between internet addiction and depression among migrant children and left-behind..., Guo [/bib_ref] Guo et al (2015) [bib_ref] Depression among migrant and left-behind children in China in relation to the..., Guo [/bib_ref] He et al (2012) [bib_ref] Depression risk of 'left-behind children' in rural China, He [/bib_ref] Hou et al (2014)Hu et al (2014) [bib_ref] The psychological and behavioral outcomes of migrant and left-behind children in China, Hu [/bib_ref] Jiang et al (2011) [bib_ref] Epidemiological investigation on unintentional injuries of left-behind children in rural area of..., Jiang [/bib_ref] Jiang (2013)Jiang et al (2015) [bib_ref] Alcohol consumption is higher among left-behind Chinese children whose parents leave rural..., Jiang [/bib_ref] Jin et al (2009) [bib_ref] The characteristics and influence factors of smoking behavior of migrated children and..., Jin [/bib_ref] Jin et al (2009)Jones et al (2004) [bib_ref] Children's experiences of separation from parents as a consequence of migration, Jones [/bib_ref] Jordan et al (2013) [bib_ref] Alcohol use among very early adolescents in Vietnam: what difference does parental..., Jordan [/bib_ref] Lan et al (2009)Li et al (2008)Li and Tao (2009) [bib_ref] Survey on mental health status and suicidal ideation among rural children left-behind, Li [/bib_ref] Li et al (2011) [bib_ref] Investigation of physical development among left-behind children in rural areas in western..., Li [/bib_ref] Li et al (2012) [bib_ref] Effects of parental working out-of-home on health risk behavior and psychological status..., Li [/bib_ref] Liao et al (2013)Lin et al (2011) [bib_ref] On the risk to and the future trend of the family environment..., Lin [/bib_ref] Ling et al (2015) [bib_ref] Effects of separation age and separation duration among left-behind children in China, Ling [/bib_ref] Liu et al (2012) [bib_ref] Relationship between emotional problem behavior and social support in left-behind children, Liu [/bib_ref] Liu et al (2012) [bib_ref] Analysis on the current situation of neglected rural children aged 0-6 years..., Liu [/bib_ref] Lu (2015) [bib_ref] Internal migration, international migration, and physical growth of left-behind children: a study..., Lu [/bib_ref] Mo et al (2016) [bib_ref] Do different parenting patterns impact the health and physical growth of 'left-behind'..., Mo [/bib_ref] Mou et al (2009) [bib_ref] Study on the nutritional status and determinants among rural stranded children in..., Mou [/bib_ref] Mu and de Brauw (2015) [bib_ref] Migration and young child nutrition: evidence from rural China, Mu [/bib_ref] Nguyen (2016) [bib_ref] Does parental migration really benefit left-behind children? Comparative evidence from Ethiopia, India,..., Nguyen [/bib_ref] Onyango et al (1994) [bib_ref] Household headship and child nutrition: a case study in western Kenya, Onyango [/bib_ref] Pan and Chen (2014) [bib_ref] Growth and nutritional status among rural left-behind children, Pan [/bib_ref] Pottinger (2005) [bib_ref] Children's experience of loss by parental migration in inner-city Jamaica, Pottinger [/bib_ref] Qiao et al (2008)Qu et al (2015) [bib_ref] Prevalence of mental disorders in 6-16-year-old students in Sichuan province, China, Qu [/bib_ref] Ren and Treiman (2016) [bib_ref] The consequences of parental labor migration in China for children's emotional wellbeing, Ren [/bib_ref] Schmeer (2009) [bib_ref] Father absence due to migration and child illness in rural Mexico, Schmeer [/bib_ref] Schmeer (2013) [bib_ref] Family structure and child anemia in Mexico, Schmeer [/bib_ref] Shen et al (2009) [bib_ref] Non-fatal injury rates among the "left-behind children" of rural China, Shen [/bib_ref] Shen et al (2013) [bib_ref] Agricultural exposures and farm-related injuries among adolescents in rural China, Shen [/bib_ref] Shen et al (2015) [bib_ref] Parental migration patterns and risk of depression and anxiety disorder among rural..., Shen [/bib_ref] Shi et al (2016) [bib_ref] Resilience as moderator of the relationship between left-behind experience and mental health..., Shi [/bib_ref] Shi et al (2016) [bib_ref] Effects of parental migration on mental health of left-behind children: evidence from..., Shi [/bib_ref] Sukamdi and Wattie (2013) [bib_ref] Tobacco use and exposure among children in migrant and non-migrant households in..., Sukamdi [/bib_ref] Sun et al (2015) [bib_ref] Psychological development and educational problems of left-behind children in rural China, Sun [/bib_ref] Tao et al (2016) [bib_ref] Food preferences, personality and parental rearing styles: analysis of factors influencing health..., Tao [/bib_ref] Tomsa and Jenaro (2015) [bib_ref] Children left behind in Romania: anxiety and predictor variables, Tomsa [/bib_ref] Vanore et al (2015) [bib_ref] Left behind' but not left alone: parental migration & the psychosocial health..., Vanore [/bib_ref] Wan et al (2009) [bib_ref] Mental health condition of the left-behind children in rural area of Anhui..., Wan [/bib_ref] Wang and Chen (2010)Wang et al (2011) [bib_ref] Research on children's depression and the influence of left-behind status in rural..., Wang [/bib_ref] Wang et al (2010) [bib_ref] Mental health of the left-behind children in different types in rural Hanchuan, Wang [/bib_ref] Wang et al (2011) [bib_ref] Preliminary study on the health status among the "left-behind" children in the..., Wang [/bib_ref] Wang et al (2012)Wang et al (2012) [bib_ref] Investigation on mental health of left-behind primary and secondary school students in..., Wang [/bib_ref] Wang et al (2006) [bib_ref] Social anxiety and the influencing factors among pupils in rural areas in..., Wang [/bib_ref] Wang (2008) [bib_ref] Investigation on safe sex issues among adolescent left-behind children, Wang [/bib_ref] Wang et al (2014)Wei and Zheng (2007)Wen et al (2008)Wen and Lin (2012) [bib_ref] Child development in rural China: children left behind by their migrant parents..., Wen [/bib_ref] Wickramage et al (2015) [bib_ref] Risk of mental health and nutritional problems for left-behind children of international..., Wickramage [/bib_ref] Wu et al (2015) [bib_ref] Social capital and the mental health of children in rural China with..., Wu [/bib_ref] Wu et al (2016) [bib_ref] The risk and protective factors in the development of childhood social anxiety..., Wu [/bib_ref] Xia et al (2011) [bib_ref] Physical growth and nutrition status among left-behind students in Guangdong, Xia [/bib_ref] Xie et al (2011)Xie et al (2013) [bib_ref] Study on the food intake and medium and severe malnutrition in poor..., Xie [/bib_ref] Xu and Xie (2015) [bib_ref] The causal effects of rural-to-urban migration on children's well-being in China, Xu [/bib_ref] Yan et al (2013) [bib_ref] Investigation and research on physical health of the left-behind and the non..., Yan [/bib_ref] Yang et al (2010) [bib_ref] Depressive symptoms and the influencing factors among left-behind children, Yang [/bib_ref] Yang and Wu (2011) [bib_ref] A survey on the mental health status of rural left-behind children in..., Yang [/bib_ref] Yang et al (2016) [bib_ref] Parental migration and smoking behavior on left-behind children: evidence from a survey..., Yang [/bib_ref] Yao et al (2010)Yu et al (2013) [bib_ref] Malnutrition status and influencing factors in children with migrant worker mothers in..., Yu [/bib_ref] Zhang et al (2011)Zhang et al (2012)Zhang (2013) [bib_ref] Investigation on the mental health status of rural left-behind children, Zhang [/bib_ref] Zhang et al (2015)Zhao et al (2008) [bib_ref] Study on the distribution and risk factors of injuries among home-stranded children..., Zhao [/bib_ref] Zhao and Liu (2010) [bib_ref] Rural left-home-children's depression and antisocial behavior: the protective role of daily pleasures, Zhao [/bib_ref] Zhao et al (2012)Zhao et al (2014) [bib_ref] Left-behind children in rural China experience higher levels of anxiety and poorer..., Zhao [/bib_ref] Zhao et al (2017) [bib_ref] Long-term impacts of parental migration on Chinese children's psychosocial well-being: mitigating and..., Zhao [/bib_ref] Zhou et al (2009) [bib_ref] Characteristics of behavioral and emotional problems of left-behind children in rural areas..., Zhou [/bib_ref] Zhou and Wang (2011)Zhou et al (2015) [bib_ref] China's left-behind children: impact of parental migration on health, nutrition, and educational..., Zhou [/bib_ref] Zhu et al (2012) . Heterogeneity varied across nutrition outcomes (I²=0·0-98·1%), with all outcomes (with the exception of wasting and weightfor-height Z score) varying substantially between studies.
Subgroup analyses of wasting for children left behind by one parent and by both parents showed no statistical evidence of an increased risk in left-behind children compared with children of non-migrants (appendix). Subgroup analyses of stunting revealed a significant increase in risk for children left behind by one parent Each full-height bar represents the health outcomes reported by an individual study included in the systematic review. Half-height bars represent studies which found varying directionality of health outcomes between different population subgroups (eg, a higher risk among girls who were left behind but no difference among boys who were left behind compared with children of non-migrant parents). Numbers refer to study references as cited in the reference list. Nine studies [bib_ref] Leaving a legacy: parental migration and school outcomes among young children in..., Asis [/bib_ref] [bib_ref] The impact of labour migration on the children left behind: a study..., Battistella [/bib_ref] [bib_ref] A comparative study on mental development characteristics between non-parent-absent children and parent-absent..., Gao [/bib_ref] [bib_ref] The study of rural left-behind children's psychological anxiety, Ge [/bib_ref] [bib_ref] Nutritional status of children during and post-global economic crisis in China, Chen [/bib_ref] [bib_ref] Analysis on the physical growth and nutrition status of children under 7..., Feng [/bib_ref] [bib_ref] Migration and young child nutrition: evidence from rural China, Mu [/bib_ref] were excluded from this plot due to the absence of significance testing reported in these studies. Two studies 105,109 were included for those outcomes for which significance testing was reported and excluded for other outcomes. [bib_ref] Psychometric characteristics of CES-D in a sample of Mexican rural adolescents in..., Aguilera-Guzman [/bib_ref] between left-behind children and children of non-migrant parents. Statistical heterogeneity was high across these outcomes (I²=82·8-83·1%). Pooled risk for substance use outcomes in studies outside of China and among children left behind by one parent showed no increased risk compared with children of non-migrants; however, each subgroup included only two studies (appendix). When studies at high risk of bias were excluded, no significant differences were identified in risk of substance use among left-behind children compared with children of nonmigrant parents. Fixed effects meta-analysis revealed a [bib_ref] Parental migration patterns and risk of depression and anxiety disorder among rural..., Shen [/bib_ref] Guo et al (2012) [bib_ref] The relationship between internet addiction and depression among migrant children and left-behind..., Guo [/bib_ref] Zhou and Wang et al (2011)Yang et al (2010) [bib_ref] Depressive symptoms and the influencing factors among left-behind children, Yang [/bib_ref] Lan et al (2009)Wang et al (2011) [bib_ref] Research on children's depression and the influence of left-behind status in rural..., Wang [/bib_ref] Cheng et al (2008) [bib_ref] Mental health status of left-behind adolescents in rural areas in Anhui, Cheng [/bib_ref] Qu et al (2015) [bib_ref] Prevalence of mental disorders in 6-16-year-old students in Sichuan province, China, Qu [/bib_ref] He et al (2012) 36 Subtotal (I = 80·3%, p=0·000)
Shen et al (2015) [bib_ref] Parental migration patterns and risk of depression and anxiety disorder among rural..., Shen [/bib_ref] Wan et al (2009) [bib_ref] Mental health condition of the left-behind children in rural area of Anhui..., Wan [/bib_ref] Zhang (2013) [bib_ref] Investigation on the mental health status of rural left-behind children, Zhang [/bib_ref] Wu et al (2016) [bib_ref] The risk and protective factors in the development of childhood social anxiety..., Wu [/bib_ref] Yao et al (2010)Wang et al (2006) [bib_ref] Social anxiety and the influencing factors among pupils in rural areas in..., Wang [/bib_ref] Wang et al (2012) [bib_ref] Investigation on mental health of left-behind primary and secondary school students in..., Wang [/bib_ref] Qu et al (2015) [bib_ref] Prevalence of mental disorders in 6-16-year-old students in Sichuan province, China, Qu [/bib_ref] Feng (2014) [bib_ref] Investigation and comparison of psychological and behavioral characteristics of rural left-behind children..., Feng [/bib_ref] Qiao et al (2008)Wang and Chen (2010) 60 Subtotal (I = 94·0%, p=0·000) Graham and Jordan (2011) [bib_ref] Migrant parents and the psychological well-being of left-behind children in southeast Asia, Graham [/bib_ref] Graham and Jordan (2011) [bib_ref] Migrant parents and the psychological well-being of left-behind children in southeast Asia, Graham [/bib_ref] Graham and Jordan (2011) [bib_ref] Migrant parents and the psychological well-being of left-behind children in southeast Asia, Graham [/bib_ref] Jiang (2013)Li et al (2008)Adhikari et al (2014) [bib_ref] The impact of parental migration on the mental health of children left..., Adhikari [/bib_ref] Liao et al (2013)Zhang et al (2011)Zhu et al (2012) [bib_ref] Health condition of left-behind children in Dingxi county of Guizhou province, Zhu [/bib_ref] Graham and Jordan (2011) [bib_ref] Migrant parents and the psychological well-being of left-behind children in southeast Asia, Graham [/bib_ref] Wickramage et al (2015) [bib_ref] Risk of mental health and nutritional problems for left-behind children of international..., Wickramage [/bib_ref] Vanore et al (2015) [bib_ref] Left behind' but not left alone: parental migration & the psychosocial health..., Vanore [/bib_ref] Liu et al (2012) [bib_ref] Relationship between emotional problem behavior and social support in left-behind children, Liu [/bib_ref] Qu et al (2015) 50 Subtotal (I = 96·9%, p=0·000)
## Conduct disorder
## Suicidal ideation
Deng et al (2014)Gao et al (2010) [bib_ref] The impact of parental migration on health status and health behaviours among..., Gao [/bib_ref] Lan et al (2009) Meta-regression showed sex and mean age had no significant effects on any health outcomes (appendix).
# Discussion
Although most studies identified by our systematic review focused on internal labour migration in China, our findings suggest that, as a group, left-behind children and adolescents have worse outcomes than children of nonmigrant parents, especially with regard to mental health and nutrition. Compared with children of non-migrants, left-behind children and adolescents had a 52% increased risk of depression, 70% increased risk of suicidal ideation, and an 85% increased risk of anxiety. We found smaller increases in risk for wasting (13%), stunting (12%) and substance use (24%). Left-behind children and adolescents had no increased risk of conduct disorders, being overweight or obese, anaemia, unintentional injury, diarrhoea, or abuse. Although a minority of individual studies [bib_ref] Child development in rural China: children left behind by their migrant parents..., Wen [/bib_ref] [bib_ref] Psychological development and educational problems of left-behind children in rural China, Sun [/bib_ref] [bib_ref] Food preferences, personality and parental rearing styles: analysis of factors influencing health..., Tao [/bib_ref] [bib_ref] Rural left-home-children's depression and antisocial behavior: the protective role of daily pleasures, Zhao [/bib_ref] [bib_ref] Left-behind children in rural China experience higher levels of anxiety and poorer..., Zhao [/bib_ref] [bib_ref] Internal migration, international migration, and physical growth of left-behind children: a study..., Lu [/bib_ref] [bib_ref] Preliminary study on the health status among the "left-behind" children in the..., Wang [/bib_ref] [bib_ref] Investigation and research on physical health of the left-behind and the non..., Yan [/bib_ref] [bib_ref] China's left-behind children: impact of parental migration on health, nutrition, and educational..., Zhou [/bib_ref] reported beneficial health effects, no overall benefits were found across any of the outcomes assessed. We found no studies investigating the effect of forced migration, which might be explained by the fact that leaving children behind in the context of conflict or disaster is unlikely.
We updated our search from April 28, 2017, to Sept 5, 2018, using the same search terms and databases. [bib_ref] Rural left-home-children's depression and antisocial behavior: the protective role of daily pleasures, Zhao [/bib_ref] Wei and Zhang (2007)Aguilera-Guzman et al (2004) [bib_ref] Psychometric characteristics of CES-D in a sample of Mexican rural adolescents in..., Aguilera-Guzman [/bib_ref] Yang and Wu (2011) [bib_ref] A survey on the mental health status of rural left-behind children in..., Yang [/bib_ref] Gao (2008) [bib_ref] A comparative study on mental development characteristics between non-parent-absent children and parent-absent..., Gao [/bib_ref] Wang et al (2010) [bib_ref] Mental health of the left-behind children in different types in rural Hanchuan, Wang [/bib_ref] Guo et al (2015) [bib_ref] Depression among migrant and left-behind children in China in relation to the..., Guo [/bib_ref] Wang et al (2014)Shi et al (2016) [bib_ref] Resilience as moderator of the relationship between left-behind experience and mental health..., Shi [/bib_ref] Zhou et al (2009) [bib_ref] Characteristics of behavioral and emotional problems of left-behind children in rural areas..., Zhou [/bib_ref] Xie et al (2011)Ling et al (2015) [bib_ref] Effects of separation age and separation duration among left-behind children in China, Ling [/bib_ref] Bi and Oyserman (2015) [bib_ref] Left behind or moving forward? Effects of possible selves and strategies to..., Bi [/bib_ref] Wu et al (2015) [bib_ref] Social capital and the mental health of children in rural China with..., Wu [/bib_ref] Tomsa and Jenaro (2015) [bib_ref] Children left behind in Romania: anxiety and predictor variables, Tomsa [/bib_ref] Lan et al (2009)He et al (2012) [bib_ref] Depression risk of 'left-behind children' in rural China, He [/bib_ref] Subtotal (I =67·0%, p=0·000)
## Anxiety
Asis and Ruiz-Marave (2013) [bib_ref] Leaving a legacy: parental migration and school outcomes among young children in..., Asis [/bib_ref] Battistella and Conaco (1998) [bib_ref] The impact of labour migration on the children left behind: a study..., Battistella [/bib_ref] Zhao et al (2012)Ge and Luo (2011) [bib_ref] The study of rural left-behind children's psychological anxiety, Ge [/bib_ref] Wei and Zheng (2007)Shi et al (2016) [bib_ref] Resilience as moderator of the relationship between left-behind experience and mental health..., Shi [/bib_ref] Zhao et al (2014) [bib_ref] Left-behind children in rural China experience higher levels of anxiety and poorer..., Zhao [/bib_ref] Xie et al (2011)Gao (2008) [bib_ref] A comparative study on mental development characteristics between non-parent-absent children and parent-absent..., Gao [/bib_ref] Zhang (2013) [bib_ref] Investigation on the mental health status of rural left-behind children, Zhang [/bib_ref] Wang et al (2010) [bib_ref] Mental health of the left-behind children in different types in rural Hanchuan, Wang [/bib_ref] Zhou et al (2009) [bib_ref] Characteristics of behavioral and emotional problems of left-behind children in rural areas..., Zhou [/bib_ref] Wang et al (2014)Tomsa and Jenaro (2015) [bib_ref] Children left behind in Romania: anxiety and predictor variables, Tomsa [/bib_ref] Zhang et al (2012)Yang and Wu (2011) [bib_ref] A survey on the mental health status of rural left-behind children in..., Yang [/bib_ref] Wang and Chen (2010) 60 Subtotal (I =82·9%, p=0·000)
## Conduct disorder
Lan et al (2009)Zhao et al (2017) [bib_ref] Long-term impacts of parental migration on Chinese children's psychosocial well-being: mitigating and..., Zhao [/bib_ref] Hu et al (2014) [bib_ref] The psychological and behavioral outcomes of migrant and left-behind children in China, Hu [/bib_ref] Fan et al (2010) [bib_ref] Emotional and behavioral problems of Chinese left-behind children: a preliminary study, Fan [/bib_ref] Zhang et al (2011)Ling et al (2015) -0·16 (-0·36 to 0·03) -0·03 (-0·22 to 0·15) 0·04 (-0·19 to 0·27) 0·05 (-0·23 to 0·34) 0·09 (-0·01 to 0·20) 0·10 (0·02 to 0·18) 0·14 (0·06 to 0·21) 0·15 (0·06 to 0·24) 0·15 (0·08 to 0·23) 0·18 (0·08 to 0·28) 0·19 (0·08 to 0·29) 0·19 (0·01 to 0·36) 0·19 (-0·18 to 0·55) 0·23 (0·05 to 0·41) 0·28 (0·06 to 0·50) 0·30 (0·16 to 0·45) 0·50 (0·36 to 0·65) 0·16 (0·10 to 0·21) -0·12 (-0·25 to 0·01) -0·04 (-0·20 to 0·13) -0·02 (-0·11 to 0·07) 0·07 (-0·19 to 0·34) 0·10 (-0·09 to 0·28) 0·14 (0·07 to 0·22) 0·15 (0·07 to 0·22) 0·17 (0·06 to 0·28) 0·18 (0·07 to 0·29) 0·19 (0·04 to 0·34) 0·20 (0·12 to 0·28) 0·23 (0·13 to 0·34) 0·24 (0·15 to 0·33) 0·27 (0·05 to 0·49) 0·50 (0·21 to 0·79) 0·61 (0·32 to 0·90) 0·62 (0·46 to 0·79) 0·18 (0·11 to 0·26) -0·13 (-0·27 to 0·02) 0·07 (-0·02 to 0·15) 0·14 (0·05 to 0·23) 0·18 (0·07 to 0·29) 0·34 (0·18 to 0·50) 0·41 (0·23 to 0·59) 0·16 (0·04 to 0·28) [bib_ref] Does having a migrant parent reduce the risk of undernutrition for children..., Graham [/bib_ref] Graham and Jordan (2013) [bib_ref] Does having a migrant parent reduce the risk of undernutrition for children..., Graham [/bib_ref] Ban et al (2017) [bib_ref] Child feeding and stunting prevalence in left-behind children: a descriptive analysis of..., Ban [/bib_ref] Feng et al (2010) [bib_ref] Analysis on the physical growth and nutrition status of children under 7..., Feng [/bib_ref] Wang et al (2011) [bib_ref] Preliminary study on the health status among the "left-behind" children in the..., Wang [/bib_ref] Mou et al (2009) [bib_ref] Study on the nutritional status and determinants among rural stranded children in..., Mou [/bib_ref] Xie et al (2013) [bib_ref] Study on the food intake and medium and severe malnutrition in poor..., Xie [/bib_ref] Chen et al (2013) [bib_ref] Analysis on the nutritional status of unattended children under 7 years in..., Chen [/bib_ref] Gao et al (2010) [bib_ref] The impact of parental migration on health status and health behaviours among..., Gao [/bib_ref] Onyango et al (1994) [bib_ref] Household headship and child nutrition: a case study in western Kenya, Onyango [/bib_ref] Xia et al (2011) [bib_ref] Physical growth and nutrition status among left-behind students in Guangdong, Xia [/bib_ref] Pan and Chen (2014) [bib_ref] Growth and nutritional status among rural left-behind children, Pan [/bib_ref] Yu et al (2013) [bib_ref] Malnutrition status and influencing factors in children with migrant worker mothers in..., Yu [/bib_ref] Chen et al (2012)Li et al (2011) [bib_ref] Investigation of physical development among left-behind children in rural areas in western..., Li [/bib_ref] Tao et al (2016) 56 Subtotal (I 2 =73·4%, p=0·000) 1 or 2 1 or 2 1 or 2 1 or 2 1 or 2 1 or 2 1 or 2 1 or 2 1 or 2 1 (father) 1 or 2 1 or 2 1 (mother) 1 or 2 1 or 2 1 or 2Mou et al (2009) [bib_ref] Study on the nutritional status and determinants among rural stranded children in..., Mou [/bib_ref] Chen et al (2013) [bib_ref] Analysis on the nutritional status of unattended children under 7 years in..., Chen [/bib_ref] Pan and Chen (2014) [bib_ref] Growth and nutritional status among rural left-behind children, Pan [/bib_ref] Mo et al (2016) [bib_ref] Do different parenting patterns impact the health and physical growth of 'left-behind'..., Mo [/bib_ref] Yan et al (2013) [bib_ref] Investigation and research on physical health of the left-behind and the non..., Yan [/bib_ref] Xia et al (2011) [bib_ref] Physical growth and nutrition status among left-behind students in Guangdong, Xia [/bib_ref] Wang et al (2011) [bib_ref] Preliminary study on the health status among the "left-behind" children in the..., Wang [/bib_ref] Tao et al (2016) [bib_ref] Food preferences, personality and parental rearing styles: analysis of factors influencing health..., Tao [/bib_ref] Wang et al (2011) [bib_ref] Preliminary study on the health status among the "left-behind" children in the..., Wang [/bib_ref] Wen et al (2008)Yan et al (2013) [bib_ref] Study on the food intake and medium and severe malnutrition in poor..., Xie [/bib_ref] Chen et al (2010)Pan and Chen (2014) [bib_ref] Growth and nutritional status among rural left-behind children, Pan [/bib_ref] Mo et al (2016) [bib_ref] Do different parenting patterns impact the health and physical growth of 'left-behind'..., Mo [/bib_ref] Wickramage et al (2015) [bib_ref] Risk of mental health and nutritional problems for left-behind children of international..., Wickramage [/bib_ref] Gao et al (2010) 13 Subtotal (I 2 =70·1%, p=0·001) 1 or 2 1 or 2 1 or 2 1 or 2 1 or 2 1 or 2 1 or 2 1 or 2 1 or 2Wang et al (2011) [bib_ref] Preliminary study on the health status among the "left-behind" children in the..., Wang [/bib_ref] Chen et al (2013) [bib_ref] Analysis on the nutritional status of unattended children under 7 years in..., Chen [/bib_ref] Chen et al (2010) 86 Subtotal (I 2 =98·1%, p=0·000)Onyango et al (1994) [bib_ref] Household headship and child nutrition: a case study in western Kenya, Onyango [/bib_ref] Mou et al (2009) [bib_ref] Study on the nutritional status and determinants among rural stranded children in..., Mou [/bib_ref] Chen et al (2013) [bib_ref] Analysis on the nutritional status of unattended children under 7 years in..., Chen [/bib_ref] Yu et al (2013) [bib_ref] Malnutrition status and influencing factors in children with migrant worker mothers in..., Yu [/bib_ref] Feng et al (2010) [bib_ref] Analysis on the physical growth and nutrition status of children under 7..., Feng [/bib_ref] Xie et al (2013) [bib_ref] Study on the food intake and medium and severe malnutrition in poor..., Xie [/bib_ref] Wickramage et al (2015) [bib_ref] Risk of mental health and nutritional problems for left-behind children of international..., Wickramage [/bib_ref] Li et al (2011) 94 Subtotal (I 2 =83·1%, p=0·000) 1 or 2 1 (father) 1 or 2 1 or 2 1 (mother) 1 or 2 1 or 2 1 or 2 1 or 2 continues on next page) This updated search identified nine additional papers (six published in English and three published in Chinese), all of which focused on internal migration and were done in China. Findings from the studies were consistent with the results from our meta-analyses and provide no evidence of benefits of parental migration for left-behind children and adolescents and support our overall findings in terms of the negative health impact of migration on left-behind children and adolescents. Four studies reported outcomes for depression: three studies 129-131 found a small increase in depressive symptoms or worse depression scores among left-behind children and adolescents and one study 132 found no difference between left-behind children and children of non-migrant parents. Studies 129,133 reporting anxiety outcomes similarly found increased risks among leftbehind children and adolescents compared with children of non-migrants. Consistent with our findings, a large study of adolescents in China (n=13 952) found a significant increase in suicide attempts in left-behind adolescents compared with adolescents who were not left behind (3·75% vs 2·86%, p<0·01). [bib_ref] A comparative analysis of suicide attempts in left-behind children and non-left-behind children..., Chang [/bib_ref] Two studies [bib_ref] Comparison of nutritional status between Han and Yao ethnic left-behind children aged..., Li [/bib_ref] assessed nutrition outcomes. A cohort study [bib_ref] Parental migration, self-efficacy and cigarette smoking among rural adolescents in south China, Gao [/bib_ref] Guo et al (2012) [bib_ref] The relationship between internet addiction and depression among migrant children and left-behind..., Guo [/bib_ref] Guo et al (2012) [bib_ref] The relationship between internet addiction and depression among migrant children and left-behind..., Guo [/bib_ref] Jordan et al (2013) [bib_ref] Alcohol use among very early adolescents in Vietnam: what difference does parental..., Jordan [/bib_ref] Li et al (2012) [bib_ref] Effects of parental working out-of-home on health risk behavior and psychological status..., Li [/bib_ref] Zhang et al (2015)Gao et al (2010) [bib_ref] The impact of parental migration on health status and health behaviours among..., Gao [/bib_ref] Wen and Lin (2012) [bib_ref] Child development in rural China: children left behind by their migrant parents..., Wen [/bib_ref] Sukamdi and Wattie (2013) [bib_ref] Tobacco use and exposure among children in migrant and non-migrant households in..., Sukamdi [/bib_ref] Gao et al (2010) [bib_ref] The impact of parental migration on health status and health behaviours among..., Gao [/bib_ref] Shen et al (2015) [bib_ref] Parental migration patterns and risk of depression and anxiety disorder among rural..., Shen [/bib_ref] Lin et al (2010) [bib_ref] On the risk to and the future trend of the family environment..., Lin [/bib_ref] Jiang et al (2015) [bib_ref] Alcohol consumption is higher among left-behind Chinese children whose parents leave rural..., Jiang [/bib_ref] Li et al (2012) [bib_ref] Effects of parental working out-of-home on health risk behavior and psychological status..., Li [/bib_ref] Yang et al (2016) [bib_ref] Parental migration and smoking behavior on left-behind children: evidence from a survey..., Yang [/bib_ref] Qu et al (2015) [bib_ref] Leaving a legacy: parental migration and school outcomes among young children in..., Asis [/bib_ref] Zhao et al (2014) [bib_ref] Left-behind children in rural China experience higher levels of anxiety and poorer..., Zhao [/bib_ref] Wang (2008) [bib_ref] Investigation on safe sex issues among adolescent left-behind children, Wang [/bib_ref] Shen et al (2015) [bib_ref] Parental migration patterns and risk of depression and anxiety disorder among rural..., Shen [/bib_ref] Chen and Chan (2016) [bib_ref] Parental absence, child victimization, and psychological well-being in rural China, Chen [/bib_ref] Liu et al (2012) [bib_ref] Relationship between emotional problem behavior and social support in left-behind children, Liu [/bib_ref] Pottinger (2005) [bib_ref] Non-fatal injury rates among the "left-behind children" of rural China, Shen [/bib_ref] Chen and Qu (2010)Shen et al (2013) [bib_ref] Agricultural exposures and farm-related injuries among adolescents in rural China, Shen [/bib_ref] Chen and Qu (2010)Shen et al (2009) [bib_ref] Non-fatal injury rates among the "left-behind children" of rural China, Shen [/bib_ref] Jiang et al (2011) [bib_ref] Epidemiological investigation on unintentional injuries of left-behind children in rural area of..., Jiang [/bib_ref] Zhao et al (2008) [bib_ref] Study on the distribution and risk factors of injuries among home-stranded children..., Zhao [/bib_ref] Jiang et al (2011) [bib_ref] Epidemiological investigation on unintentional injuries of left-behind children in rural area of..., Jiang [/bib_ref] Zhao et al (2008) [bib_ref] Study on the distribution and risk factors of injuries among home-stranded children..., Zhao [/bib_ref] Shen et al (2015) : Forest plots of relative risks or standardised mean differences for health outcomes Data are presented for mental health binary outcomes (A), mental health continuous outcomes (B), nutrition binary outcomes (C), nutrition continuous outcomes (D), and substance use, abuse and injury outcomes (E). Weights were assigned by random effects analysis. RR=relative risk. SMD=standardised mean difference.
[formula] 4·52 4·72 3·67 2·77 7·45 8·53 8·77 8·07 8·62 7·54 7·46 5·01 1·88 4·90 3·96 6·05 6·08 100·00 6·29 5·61 7·08 3·78 5·24 7·28 7·30 6·75 6·74 5·82 7·24 6·78 7·03 4·57 3·45 3·43 5·60 100·00 16·00 18·60 18·43 17·59 15·11 14·28 100·00 0 -0·5 [/formula]
weight-for-age Z scores and height-for-age Z scores at baseline and follow-up after migration, although the effect of migration varied according to which parent migrated. Li and Zhang [bib_ref] Comparison of nutritional status between Han and Yao ethnic left-behind children aged..., Li [/bib_ref] found that a higher proportion of left-behind children had stunting and wasting than did children with non-migrant parents. One study [bib_ref] A comparative study of behavior problems among left-behind children, migrant children and..., Hu [/bib_ref] reported a higher risk of any type of unintentional injury (eg, vehicle and traffic injuries and falls) among left-behind children and adolescents than children and adolescents with non-migrant parents (adjusted OR 1·208, p<0·05).
Labour migration is a global trend, shaping families and communities across the world. [bib_ref] Our future: a Lancet Commission on adolescent health and wellbeing, Patton [/bib_ref] Our findings are consistent with previous reviews 5,11 about left-behind children in rural China and the Philippines: although parental labour migration might have economic benefits for families, it might have hidden costs for the health of children and adolescents who are left behind. A previous study [bib_ref] Parental migration and the mental health of those who stay behind to..., Graham [/bib_ref] reported that these negative health consequences extend to other family members. The Child Health and Migrant Parents in South-East Asia study [bib_ref] Parental migration and the mental health of those who stay behind to..., Graham [/bib_ref] showed that left-behind mothers and other carers in transnational migrant households were more likely to have poor mental health than carers in non-migrant households: mental health problems were associated with infrequent contact with the migrant and migrant destinations in the Middle East, whereas receiving remittances in the past 6 months was found to have a protective effect on the mental health of carers. With the exception of age and sex, we were unable to investigate factors mediating poor health outcomes among left-behind children and adolescents, although family structure, community social capital, living conditions, and level of caregiver supervision might be important. [bib_ref] Social capital and the mental health of children in rural China with..., Wu [/bib_ref] [bib_ref] International parental migration and the psychological well-being of children in Ghana, Nigeria,..., Mazzucato [/bib_ref] Future research should consider the circumstances of parental migration.
Children of parents migrating because of extreme poverty, disasters, or oppression are likely to have worse health outcomes than children from wealthier migrant families that are financially stable with access to adequate health care. Residing with siblings and relationships between children and their caregivers could also be important.
82% of studies included in our systematic review were done in China, an upper middle-income country where migration is mainly internal, rural-to-urban labour migration. Our study highlights a major research gap in countries beyond China, potentially limiting the generalisability of our findings to other forms of migration and to other settings, especially low-income countries. Subgroup analyses of studies done across the rest of the world showed no difference in outcomes for left-behind children and adolescents; however, with the exception of conduct disorder, few studies have been done, limiting the conclusions that can be drawn.
Addressing the needs of families who are left behind will be essential for health-care workers and policy makers. [bib_ref] The UCL-Lancet Commission on Migration and Health: the health of a world..., Abubakar [/bib_ref] In China, the health and wellbeing of leftbehind children is a priority and steps are being taken to address this. In 2013, the Chinese Government called on local authorities to take specific responsibility for the education and care of left-behind children. [bib_ref] Announcement on the care and education for rural school aged left-behind children..., Government Of [/bib_ref] The Chinese Women's Federation has taken a lead in most provinces, but action has been inconsistent and it is unclear whether the health of left-behind children is improving as a consequence. Community-based clubs for children have been established to provide left-behind children with educational and recreational opportunities. [bib_ref] Care for left-behind children in rural China: a realist evaluation of a..., Zhao [/bib_ref] sessions, vaccinations, and health checks. [bib_ref] Effect of a conditional cash transfer program on nutritional knowledge and food..., Zhang [/bib_ref] In China until around 10 years ago, the national household registration system limited rural children's access to urban health and educational services, with children forced to attend designated migrant schools, which varied in quality. The situation is now changing, especially in smaller cities, as a result of the national household registration system restrictions being relaxed, migrant children attending mainstream schools, and using rural health insurance to access health care. These changes have led to an increase in the number of children migrating with their parents. Next steps for research and practice require a multifaceted approach, involving clinical, epidemiological, intervention research, and policy perspectives (panel). Focusing on all levels of society, the International Organization for Migration recommends a multidimensional intervention framework that includes the government and business.Clinicians, teachers, and other individuals working with left-behind children and adolescents must be aware of the potential mental health and nutritional needs of this population, and be trained to support and treat them. Increased awareness is particularly important with regard to common mental disorders and risk-taking behaviours that children or adolescents might not present with, or that might be underlying another clinical presentation. Global mental health initiatives should be encouraged to incorporate a focus on left-behind children. However, a one-size-fitsall approach to intervention is likely to be ineffective since left-behind children and adolescents will have different experiences of migration and being left behind. A study in China 68 found that children who were left behind had more depressive symptoms than children residing in rural China with both parents who had never migrated or been left behind, regardless of whether they had previously migrated or not. However, children who were previously left-behind but were now living with their parents had fewer depressive symptoms than children in rural areas without any experience of migration or being left behind. [bib_ref] Social capital and the mental health of children in rural China with..., Wu [/bib_ref] Although sex was not a predictor of health outcomes among left-behind children and adolescents in our study, girls and boys might require different intervention approaches. Interventions are also needed to support caregivers, many of whom might be elderly relatives with health needs of their own. Increasing the evidence base beyond China is essential, as are longitudinal studies investigating the long-term effects of parental migration on children and adolescents. Although familial separation is acutely detrimental for health, children might go on to develop resilience and have potentially better health outcomes.
The comprehensive scope of this review is a strength, since evidence was included across all LMICs, in all languages, across multiple health outcomes, with low publication bias. However, our study has several limitations. Our original systematic search included literature published up to April, 2017, and thus newer studies might alter the conclusions. However, when updating the searches to September, 2018, the studies were consistent with our findings. Statistical heterogeneity was high in the meta-analyses, which persisted in subgroup analyses and meta-regression. This heterogeneity suggests that, despite our use of a strict definition of left-behind children and adolescents, other mediating or moderating factors might influence the results reported in individual studies, including caregiver and contextual factors. Similarly high heterogeneity was identified in a systematic review and meta-analysis 147 of mental disorders among refugees resettled in western countries. Most of the studies included in our systematic review and metaanalysis were from China, focused on internal migration, and were cross-sectional, which means temporal causal inference is limited and might not generalise beyond China. Despite these limitations, our study defines and identifies a global population of young people at risk.
In summary, left-behind children and adolescents have substantial unmet mental health and nutritional needs that have not been well described outside of China. The prevalence of labour migration is increasing, thus interventions that support these young people are urgently needed to prevent long-term negative effects on their health and development.
[fig] Figure 1: Study selection *Some studies included more than one outcome. [/fig]
[fig] Figure 2: Quality assessment of studies included in the systematic reviewScoring was based on an adapted version of the Newcastle Ottawa Scale 18 incorporating items from the National Institute for Clinical Excellence Quality Appraisal. Studies with a high or unclear risk of bias across five or more domains were defined as being at high risk of bias overall. [/fig]
[fig] Figure 3: Harvest plot of health outcomes among left-behind children and children of non-migrant parents included in the systematic review [/fig]
[fig] -: to 0·05) 0.32 (0·69 to 0·05) 33·36 32·09 34·54 100·00 Weight−for−height Z score Chen et al (2013)89 Wang and Chen (2010) 60 Subtotal (I 2 =0·0%, p=0·676) 0·03 (0·11 to 0·05) 0·00 (0·13 to 0·14) 0·02 (0·09 to 0·05) [/fig]
|
Posterior placoid-like maculopathy and macular hole associated with vitamin A deficiency
A B S T R A C TPurpose: To report a case of bilateral posterior placoid-like maculopathy and a macular hole associated with vitamin A deficiency. Observations: A 72-year-old male presented with nyctalopia and progressive vision loss in both eyes. Examination and multimodal imaging were consistent with posterior placoid-like maculopathy bilaterally and a macular hole in the right eye. A workup for infectious, inflammatory, and paraneoplastic etiologies revealed a severely low serum vitamin A level. Two months after initiation of vitamin A repletion, there was improvement in bestcorrected Snellen visual acuity as well as macular hole closure. A diagnosis of posterior placoid-like maculopathy in the setting of vitamin A deficiency (VAD) was made.Conclusions and importance: VAD should be considered when symmetric posterior pole placoid-like lesions are observed and other, more common etiologies have been ruled out.
# Introduction
Posterior placoid maculopathy is a rare clinical finding characterized by large confluent areas of retinal whitening or yellow-white lesions predominantly affecting the outer retinal layers in the posterior pole. [bib_ref] Syphilis: reemergence of an old adversary, Chao [/bib_ref] These lesions have been described to be hyperautofluorescent, sometimes with speckled punctate hypoautofluorescence, and have characteristic early hypofluorescence with late "fill-in" or staining on fluorescein angiography. [bib_ref] Acute syphilitic posterior placoid chorioretinitis: report of a case series and comprehensive..., Eandi [/bib_ref] Optical coherence tomography (OCT) shows thickening of the neurosensory retina with disruption of the ellipsoid zone, thickening and granular hyperreflectivity of the retinal pigment epithelium (RPE), and nodular elevations. [bib_ref] Acute syphilitic posterior placoid chorioretinitis: when the great mimicker cannot pretend any..., Neri [/bib_ref] When discrete placoid lesions are observed in the macula and posterior pole in a patient with visual impairment, infectious causes, particularly syphilitic retinitis, remain high on the differential diagnosis. [bib_ref] Syphilitic retinitis presentations: punctate inner retinitis and posterior placoid chorioretinitis, Devience [/bib_ref] [bib_ref] Ocular syphilis: the re-establishment of an old disease, Wells [/bib_ref] Other conditions to consider include inflammatory, autoimmune, toxic/metabolic, and paraneoplastic etiologies. We report a unique case of a patient with a history of hepatic malignancy in remission who presented with progressive nyctalopia and decreased vision with findings of bilateral posterior placoid-like maculopathy and a macular hole in the right eye due to underlying VAD, with marked improvement once supplementation of vitamin A was initiated.
## Case report
A 72-year-old Caucasian male with hepatocellular carcinoma (HCC) in remission, pseudophakia of both eyes (OU), and strabismic amblyopia OU was referred for a macular hole in the right eye (OD) and age-related macular degeneration OU. He reported experiencing worsening vision over the year prior to presentation, decreased night vision, and severe dry eyes. On examination, his best corrected Snellen visual acuity (BCVA) was 20/60 OD and 20/50 in the left eye (OS). Both eyes had significant punctate epithelial erosions. On fundus examination, both eyes revealed well-circumscribed, stippled hypopigmented placoid-like lesions in the posterior pole. There was a stage 2 full-thickness macular hole with trace cystoid macular edema (CME) OD and a small focal RPE detachment OS seen on OCT, as well as diffuse outer photoreceptor layer attenuation, external limiting membrane disruption, ellipsoid zone loss, and RPE mottling and thickening. Peripherally, there were numerous scattered yellow-white punctate drusenoid deposits OU. OCT, ultrawide-field fundus photos, and autofluorescence corroborated fundus findings, and fluorescein angiogram (FA) showed disc leakage with stippled hyperfluorescence throughout the posterior pole with distinct margins [fig_ref] Figure 1: Fundus photo, OCT macula, and FA of the right eye at presentation [/fig_ref]. Indocyanine green angiography was unremarkable.
The patient had felt unwell for several months and had a new rash of the extremities for three weeks prior to initial presentation. He denied a history of sexually transmitted diseases, autoimmune disorders, or tuberculosis. He was referred to the emergency department for laboratory evaluation to rule out syphilis, although he left against medical advice prior to completing the requested work-up. The initial labs drawn were unremarkable and included: complete blood count, basic metabolic panel, Lyme serologies, and fluorescent treponemal antibody test absorption test/rapid plasma reagin; transaminases were at baseline.
Two months later, the patient was evaluated by a uveitis specialist as his vision and symptoms progressively worsened with a BCVA of 20/125 OD and 20/60 OS. With a stable examination, he was advised to complete the work-up and follow-up with his medical oncologist to ensure stable remission of his HCC. Fortunately, there was no recurrence of hepatic malignancy, although he had ascites that required an abdominal paracentesis. . OCT macula shows a stage 2 full-thickness macular hole with trace CME (B, red arrow), as well as diffuse outer layer attenuation (B, white arrow), ELM disruption, ellipsoid zone loss, and RPE mottling and thickening (B, yellow arrow). After vitamin A supplementation, the OCT findings significantly improved, and the macular hole resolved (E, corresponding arrows). FA in the early phase, displayed in the upper right hand corner, and late phase, in the central picture, shows disc leakage (C, gray arrow) with stippled hyperfluorescence and staining throughout posterior pole (C, navy arrow), with less distinct margins and less disc leakage after supplementation was initiated (F, corresponding arrows). (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.) The patient began to receive vitamin A supplementation (50,000 international units, intramuscular, weekly), and two weeks later, he reported experiencing a subjective improvement in nyctalopia, dry eyes, and visual acuity. Six weeks later, repeat testing of the serum vitamin A level indicated an improvement to 11.9 μg/dL, and his findings dramatically improved with BCVA of 20/50 OD and 20/40 OS. Remarkably, the macular hole in the right eye had closed, and there was decreased disc leakage on FA, as well as less distinct placoid-like macular lesions OU. Three months later, with vitamin A level at 22.3 μg/dL, there was continued improvement to BCVA (20/30 OU), color vision, and a decrease in peripheral scattered yellow-white punctate drusenoid deposits. [fig_ref] Table 1: Timeline of patient's BCVA, color vision, exam findings, and vitamin A levels/supplementation [/fig_ref] highlights the patient course.
The patient will continue with vitamin A supplementation and retinal monitoring.
# Discussion
Posterior placoid maculopathy has been documented in association with many diseases including ocular syphilis, COVID-19, vasculitis, serpiginous choroiditis, birdshot retinochoroidopathy, and acute posterior multifocal placoid pigment epitheliopathy. [bib_ref] Syphilitic retinitis presentations: punctate inner retinitis and posterior placoid chorioretinitis, Devience [/bib_ref] [bib_ref] Unilateral acute posterior multifocal placoid pigment epitheliopathy in a convalescent COVID-19 patient, Olguín-Manríquez [/bib_ref] [bib_ref] Acute posterior multifocal placoid pigment epitheliopathy and cerebral vasculitis, Wilson [/bib_ref] To our knowledge, this is the first documented case of posterior placoid-like maculopathy with macular hole secondary to VAD.
Vitamin A, a fat-soluble vitamin, may become deficient from insufficient intake or impaired absorption. Malnutrition is the most common cause of VAD worldwide. Conditions causing gastrointestinal and hepatobiliary malabsorption, including cirrhosis, inflammatory bowel disease, Celiac's disease, and bariatric surgery, are common causes of VAD in the developed world. [bib_ref] Vitamin A deficiency and clinical disease: an historical overview, Sommer [/bib_ref] Initial signs of VAD include nyctalopia with visual field defects while other associated symptoms include conjunctival xerosis, Bitot's spots, and keratomalacia. [bib_ref] Visual electrophysiology in the assessment of toxicity and deficiency states affecting the..., O'neill [/bib_ref] [bib_ref] Functional observations in vitamin A deficiency: diagnosis and time course of recovery, Mcbain [/bib_ref] Retinal findings in VAD are typically described as multiple round yellow-white lesions in the macula and midperiphery captured on OCT as drusenoid subretinal deposits. [bib_ref] The role of vitamin A in retinal diseases, Sajovic [/bib_ref] [bib_ref] Retinal structure in vitamin A deficiency as explored with multimodal imaging, Aleman [/bib_ref] [bib_ref] Macular thickness analysis and resolution of subretinal drusenoid deposits with optical coherence..., Zatreanu [/bib_ref] These yellow deposits are theorized to represent accumulating shed photoreceptors from a disrupted retinoid cycle layering between the RPE and ellipsoid band, causing blocked hypo-autofluorescence and appearing as hyperreflective deposits on OCT. [bib_ref] Retinal structure in vitamin A deficiency as explored with multimodal imaging, Aleman [/bib_ref] [bib_ref] Outer retina changes on optical coherence tomography in vitamin A deficiency, Berkenstock [/bib_ref] In reports describing VAD retinopathy, treatment with vitamin A supplementation usually improved retinal function, resolved symptoms, and cleared the yellow deposits. [bib_ref] Improvement in visual function and fundus findings for a patient with vitamin..., Apushkin [/bib_ref] [bib_ref] Fundus white spots and acquired night blindness due to vitamin A deficiency, Genead [/bib_ref] In our case, the patient experienced profound subjective improvements in vision within weeks of initiating vitamin A replacement therapy, with notable anatomic improvements as well. Findings of bilateral posterior placoid maculopathy in the retina often points to an underlying systemic etiology, including infectious, inflammatory, toxic/metabolic, or paraneoplastic. Syphilis is the most common cause and required immediate rule-out, but tuberculosis and fungal infection may also cause similar findings. Posterior placoid maculopathy is also associated with inflammation in the eye, as CME is the most frequent complication in uveitis causing visual impairment. [bib_ref] Macular alterations secondary to intraocular inflammatory disease, Nussenblatt [/bib_ref] In paraneoplastic syndromes, autoimmune retinopathy (AIR) and CAR are rare, poorly understood retinal diseases that can also present as bilateral, progressive visual deterioration. [bib_ref] Autoimmune retinopathy: a review, Jr [/bib_ref] The patient's history of HCC drew greater concern for other malignancy. Although no autoimmune or paraneoplastic markers returned positive, AIR and CAR should remain on the differential diagnosis in cases with similar presentations. While our patient refused further workup, he agreed to do so in the event that visual and retinal decline continued even after Vitamin A levels normalized.
Macular holes and CME, however, have not been previously described as an association with VAD. Since VAD has been linked to atrophic changes in the RPE, [bib_ref] Acute syphilitic posterior placoid chorioretinitis: when the great mimicker cannot pretend any..., Neri [/bib_ref] [bib_ref] C20D3-Vitamin A prevents retinal pigment epithelium atrophic changes in a mouse model, Zhang [/bib_ref] it is possible that a similar process could have caused a full-thickness macular hole to form over time. Although spontaneous closure of macular holes is common, given the speed of closure after starting supplementation in our patient without other ocular intervention, we strongly believe that treatment of the underlying hypovitaminosis likely resulted in resolution of the macular hole.
Our report is limited by the patient's hepatologic comorbidities which may impact his metabolism of vitamin A, although he was deemed stable and in remission. It is not possible to prove a causative relationship between therapy and treatment effect, but the temporal association is strongly supportive.
# Conclusions
Posterior placoid maculopathy is classically associated with infectious and inflammatory etiologies, particular ocular syphilis. We report a unique case of a patient with hepatobiliary disease with severely low vitamin A levels that presented with bilateral posterior placoid-like lesions and a macular hole of the right eye that resolved with vitamin A supplementation.
## Patient consent
The patient consented to publication of the case in writing.
# Funding
No funding or grant support.
## Authorship
All authors attest that they meet the current ICMJE criteria for Authorship.
## Declaration of competing interest
The following authors have no financial disclosures: EWL, RD, BKD, SAS.
[fig] Figure 1: Fundus photo, OCT macula, and FA of the right eye at presentation (A-C) compared to two months post-vitamin A supplementation (D-F). Fundus photo demonstrates well-circumscribed, placoid-like lesion (A, black arrows) in posterior pole as well as peripheral yellow-white lesions (A, green arrow) with improvement after vitamin A supplementation (D, corresponding line & arrow) [/fig]
[fig] Figure 3: Fundus autofluorescence photography of the right eye (A) and left eye (C) at presentation compared to two months post-vitamin A supplementation (B, D) shows improvement in hyperautoflourescence in the placoid-like lesion. [/fig]
[table] Table 1: Timeline of patient's BCVA, color vision, exam findings, and vitamin A levels/supplementation. Vitamin A was dosed at 50,000 units by intramuscular injection once per week. [/table]
|
The impact of the COVID lockdown on alcohol consumption in the Netherlands. The role of living arrangements and social isolation
A B S T R A C TBackground: The COVID-19 pandemic and the subsequent lockdown have a strong impact on health and health behaviours, such as alcohol consumption. Although there is some evidence of an overall decline in alcohol consumption during the lockdown, studies also show an increase in risky drinking patterns, e.g. solitary drinking, and differences between subgroups of individuals, e.g. depending on their living arrangement. Yet most studies rely on cross-sectional designs with retrospective questions, and small samples. Methods: A longitudinal study was conducted using 13 waves of the COVID-Questionnaire within the Lifelines cohort from the northern Netherlands (n = 63,194). The outcome was alcohol consumption (glasses per week) between April 2020 and July 2021. Linear fixed-effects models were fitted to analyse trends in alcohol consumption, and these were compared with pre-COVID drinking levels. Moreover, the role of living arrangement and feelings of social isolation as potential moderators was tested. Results: Alcohol consumption during the pandemic was lower than in previous years, and the seasonal pattern differed from the pre-COVID one, with levels being lower when lockdown measures were stricter. Moreover, the seasonal pattern differed by living arrangement: those living alone saw a relative increase in drinking throughout tight lockdown periods, whereas those living with children showed the strongest increase during the summer. Social isolation showed a weaker moderation effect. Conclusions: Overall alcohol levels were down in the pandemic, and in particular during strict lockdowns. Those living on their own and those who felt more isolated reacted more strongly to the lockdown, the longer it lasted.
# Introduction
The COVID-19 pandemic, as well as the regulations to control the spread of the virus, represent a "natural experiment", as they abruptly affected a vast majority of the population world-wide [bib_ref] Accumulation of economic hardship and health during the COVID-19 pandemic: Social causation..., Bierman [/bib_ref] , who was suddenly exposed to unprecedented restrictions (i.e. lockdown) and fear of viral infection. Although its consequences in terms of impact on health (behaviours) are still to be fully grasped, there is some evidence that almost two years of lockdown -defined as all preventive measures implemented to control the spread of the virus-had undesirable effects on individuals, such as increased feelings of loneliness, stress and mental health issues [bib_ref] The mental health impact of COVID-19 and lockdown-related stressors among adults in..., Chandola [/bib_ref] [bib_ref] Follow-up Survey of US Adult Reports of Mental Health, Substance Use, and..., Czeisler [/bib_ref] [bib_ref] COVID-19 bereavement, depressive symptoms, and binge drinking, Grace [/bib_ref]. Moreover, there is evidence showing changes in health behaviours, such as alcohol consumption [bib_ref] Health Behavior Changes During COVID-19 Pandemic and Subsequent "Stay-at-Home" Orders, Knell [/bib_ref] [bib_ref] Alcohol and other substance use during the COVID-19 pandemic: A systematic review, Roberts [/bib_ref].
Alcohol consumption, particularly in high doses, has large public health consequences, as it is associated with chronic medical conditions, such as cardiovascular diseases [bib_ref] Inequalities in Alcohol-Related Mortality in 17 European Countries: A Retrospective Analysis of..., Mackenbach [/bib_ref] , cancer, and mental disorders [bib_ref] The relation between different dimensions of alcohol consumption and burden of disease:..., Rehm [/bib_ref]. Alcohol consumption is responsible for 6% of all deaths worldwide, and for more than 25% of deaths among young men aged 15-29 years in the E.U [bib_ref] Associations between socioeconomic factors and alcohol outcomes, Collins [/bib_ref]. Moreover, it has been rated as one of the four most harmful drugs on a population level [bib_ref] Drug harms in the UK: a multi-criteria decision analysis, Nutt [/bib_ref] [bib_ref] Ranking the harm of alcohol, tobacco and illicit drugs for the individual..., Van Amsterdam [/bib_ref] , as it is also associated with undesirable social outcomes, such as higher rates of violent behaviour and traffic accidents [bib_ref] The relation between different dimensions of alcohol consumption and burden of disease:..., Rehm [/bib_ref] [bib_ref] How economic crises affect alcohol consumption and alcoholrelated health problems: a realist..., De Goeij [/bib_ref].
Several studies have analysed changes in alcohol consumption during the first few months of the pandemic, showing mixed results: while some show an overall decrease in alcohol levels compared with the pre-COVID period [bib_ref] The psychological impact of COVID-19 pandemic lockdowns: a review and meta-analysis of..., Panagiotidis [/bib_ref] [bib_ref] Assessing international alcohol consumption patterns during isolation from the COVID-19 pandemic using..., Sallie [/bib_ref] [bib_ref] Drinking to Cope During COVID-19 Pandemic: The Role of External and Internal..., Wardell [/bib_ref] [bib_ref] What is associated with the increased frequency of heavy episodic drinking during..., Valente [/bib_ref] [bib_ref] The impact of quarantine on mental health status among general population in..., Wang [/bib_ref] -in some cases relying on objective measurements, such as wastewater analyses [bib_ref] Changes in alcohol consumption associated with social distancing and self-isolation policies triggered..., Bade [/bib_ref] -, others show an increase [bib_ref] Rebound of Severe Alcoholic Intoxications in Adolescents and Young Adults After COVID-19..., Grigoletto [/bib_ref] [bib_ref] Psychological factors associated with substance use initiation during the COVID-19 pandemic, Rogers [/bib_ref] [bib_ref] Alcohol and other substance use during the COVID-19 pandemic: A systematic review, Roberts [/bib_ref]. These contrasting results may be due to different pathways leading to alcohol consumption. On the one hand, alcohol consumption has been shown to be associated with positive affect and social gatherings. From this perspective, a decrease in alcohol consumption could be explained by lower alcohol availability due to the closure of bars and restaurants [bib_ref] Alcohol and other substance use during the COVID-19 pandemic: A systematic review, Roberts [/bib_ref] , as well as restrictions on social gatherings that kept "social drinking" -and, therefore, peer pressure to drink as well-at minimum levels. In turn, a possible increase could be viewed as alcohol acting as a coping mechanism [bib_ref] Alcohol use in times of the COVID 19: Implications for monitoring and..., Rehm [/bib_ref] [bib_ref] Psychological factors associated with substance use initiation during the COVID-19 pandemic, Rogers [/bib_ref] for increased feelings of loneliness and psychological distress.
Moreover, studies that looked beyond the amount of alcohol consumed during the COVID-19 pandemic found an increase in specific drinking patterns, such as higher drinking frequency [bib_ref] Drinking to Cope During COVID-19 Pandemic: The Role of External and Internal..., Wardell [/bib_ref] [bib_ref] Changes in Alcohol Consumption Among College Students Due to COVID-19: Effects of..., White [/bib_ref] [bib_ref] Mental health and health behaviours before and during the initial phase of..., Niedzwiedz [/bib_ref] [bib_ref] Changes in alcohol consumption during the COVID-19 pandemic: exploring gender differences and..., Thompson [/bib_ref] , binge drinking [bib_ref] Rebound of Severe Alcoholic Intoxications in Adolescents and Young Adults After COVID-19..., Grigoletto [/bib_ref] [bib_ref] Mental health and health behaviours before and during the initial phase of..., Niedzwiedz [/bib_ref] and solitary drinking [bib_ref] The psychological impact of COVID-19 pandemic lockdowns: a review and meta-analysis of..., Panagiotidis [/bib_ref] [bib_ref] Alcohol consumption and alcohol-related problems during the COVID-19 pandemic: a narrative review, Ramalho [/bib_ref] [bib_ref] Drinking to Cope During COVID-19 Pandemic: The Role of External and Internal..., Wardell [/bib_ref]. The latter has been shown to lead to the development of alcohol-related problems, such as alcohol abuse and dependence [bib_ref] Associations between solitary drinking and increased alcohol consumption, alcohol problems, and drinking..., Skrzynski [/bib_ref] , especially when drinking is used as a coping mechanism against stress [bib_ref] The association between stress and drinking: modifying effects of gender and vulnerability, Dawson [/bib_ref] [bib_ref] Relations among stress, coping strategies, coping motives, alcohol consumption and related problems:..., Corbin [/bib_ref] , anxiety [bib_ref] The relation between different dimensions of alcohol consumption and burden of disease:..., Rehm [/bib_ref] , or loneliness [bib_ref] Mental health and health behaviours before and during the initial phase of..., Niedzwiedz [/bib_ref] [bib_ref] Psychological Stressors Predicting Increased Drinking During the COVID-19 Crisis: A Longitudinal National..., Oksanen [/bib_ref].
However, there are important gaps in the literature that need to be addressed. First, the vast majority of studies use cross-sectional data and rely on retrospective questions [bib_ref] The psychological impact of COVID-19 pandemic lockdowns: a review and meta-analysis of..., Panagiotidis [/bib_ref] [bib_ref] Self-reported alcohol, tobacco, and cannabis use during COVID-19 lockdown measures: results from..., Vanderbruggen [/bib_ref] [bib_ref] Drinking to Cope During COVID-19 Pandemic: The Role of External and Internal..., Wardell [/bib_ref] , and therefore do not capture changes over time. This is particularly relevant because time-related issues, such as the duration or the strictness of the lockdown, as well as the period of the year, may play a role in shaping drinking patterns [bib_ref] The psychological impact of COVID-19 pandemic lockdowns: a review and meta-analysis of..., Panagiotidis [/bib_ref].
Second, most studies do not examine heterogeneity in the impact of the lockdown on drinking patterns. Yet, there may be differences between groups of individuals. For instance, the living arrangements in which individuals faced the lockdown -i.e. whether they lived alone or a shared a household with partner and/or family-may be relevant [bib_ref] Associations between solitary drinking and increased alcohol consumption, alcohol problems, and drinking..., Skrzynski [/bib_ref]. Accordingly, studies analysing the impact of the lockdown on mental health have shown that the effect was stronger for individuals who lived alone [bib_ref] Impact of COVID-19 lockdown on mental health in Germany: longitudinal observation of..., Ahrens [/bib_ref] , whereas others have shown that those with children in the household during the lockdown suffered higher stress levels, which may have exposed them to increased alcohol consumption [bib_ref] Age and living situation as key factors in understanding changes in alcohol..., Villanueva-Blasco [/bib_ref].
Moreover, the subjective element of whether individuals feel lonely or isolated may be a key element in explaining differences in the response to the lockdown [bib_ref] The mental health impact of COVID-19 and lockdown-related stressors among adults in..., Chandola [/bib_ref] [bib_ref] Alcohol consumption and alcohol-related problems during the COVID-19 pandemic: a narrative review, Ramalho [/bib_ref] [bib_ref] Drinking to Cope During COVID-19 Pandemic: The Role of External and Internal..., Wardell [/bib_ref]. Although some cross-sectional studies have pointed in that direction [bib_ref] COVID 19 fear, stress, anxiety, and substance use among Russian and Belarusian..., Gritsenko [/bib_ref] [bib_ref] Acute mental health responses during the COVID-19 pandemic in Australia, Newby [/bib_ref] [bib_ref] Assessing international alcohol consumption patterns during isolation from the COVID-19 pandemic using..., Sallie [/bib_ref] , to the best of our knowledge no longitudinal study so far has provided evidence on how living arrangements and feelings of social isolation have affected alcohol consumption during the lockdown.
This study relies on a large sample and a longitudinal design and aims at testing: (1) whether the lockdown had an effect on alcohol consumption levels; (2) whether this effect was different for those who lived alone compared with those who lived with others, and (3) whether this effect differed between individuals who felt isolated and those who did not.
# Methods
The Lifelines COVID-19 Questionnaire was launched within the Lifelines Cohort Study, a large prospective population-based prospective cohort study and biobank in the three northern provinces of the Netherlands, examining in a three-generation design the biomedical, socio-demographic, behavioural, physical and psychological factors contributing to the health and disease of 167,729 individuals living in the north of the Netherlands [bib_ref] Understanding socioeconomic differences in incident metabolic syndrome among adults; What is the..., Hoveling [/bib_ref] -composition and characteristics of the sample have been discussed elsewhere [bib_ref] Representativeness of the lifelines cohort study, Klijs [/bib_ref] [bib_ref] Cohort Profile: Lifelines, a Three-Generation Cohort Study and Biobank, Scholtens [/bib_ref].
In order to assess the effects of the pandemic, the attitudes towards the COVID-19 regulations, and the (health) behaviours of the population of study, the Lifelines COVID-19 cohort was developed (n = 76,795) [bib_ref] Lifelines COVID-19 cohort: investigating COVID-19 infection and its health and societal impacts..., Mc Intyre [/bib_ref]. Participants were asked to fill out detailed web-based questionnaires about their physical and mental health, living situation, and health behaviours between late March 2020 and July 2021 -first, on a biweekly basis, after June 2020 on a monthly basis-, with a total of 24 waves. For the purpose of our study, 13 waves of COVID-19 questionnaire panel data were used (those with questions on alcohol consumption), covering the period between April 2020 and July 2021. Our final sample consists of 451,128 observations nested in 63, 194 individuals (61.2% female; mean age 57.2 years). Moreover, data of the same participants from three previous waves of the Lifelines cohort study that included information on alcohol consumption -waves 1, 4 and 5-were used for comparative purposes.
## Measurements
Outcome. Alcohol consumption: Alcohol consumption was assessed from wave 5 (April 2020) until wave 24 (July 2021), with 13 measurements in total. Questions referred to the amount of alcohol consumed: "How many glasses of alcohol did you drink in the past 7 days?" ("in the past 14 days" from wave 7 onwards). In the Netherlands, a standard glass is defined as containing roughly 10 g of alcohol (Gezondheidsraad, Health Council of the Netherlands, 2015). In order to harmonize the data, the average weekly alcohol consumption score (in glasses per week) was calculated (the same variable was created for previous waves of the Lifelines cohort). The scores range from 0 to 70 (all values higher than 70 were recoded to 70). Independent variable. COVID lockdown: a variable "Days since first lockdown (15th of March, 2020)" was created and converted into a categorical variable with categories coinciding with every month of the observation period (category 1 =April 2020; category 24 = July 2021). Due to the timing of Lifelines assessments, some months had very few observations (e.g. August 2020), or no observations at all (e.g. February 2021). These categories were merged with the previous month (e.g. "July/August 2020 ′′ ) or dropped (February 2021).
Time-varying covariates. Living arrangements were captured by a dichotomous variable: "lives alone" / "lives in a shared household", based on the question "do you have one or more housemates?". Additionally, sensitivity analyses distinguishing between "adult(s) with children (<18) living at home", and "adults without children living at home" were carried out.
Feelings of social isolation were measured by the following question: "How socially isolated have you felt in the last 14 days?", with responses ranging from 1 ("not isolated") to 10 ("extremely isolated"). For the descriptive analyses, a dichotomous version of this variable was created, assessing "low isolation" (scores 1-6) and "high isolation" (≥7). This question was not asked in wave 11 -mid-end of July-, which created a large number of missing values. These were imputed by taking the last observation available (n = 33,867) from wave 10 -early to mid-July. Alternatively, they were also imputed by means of multiple imputation (MICE). Yet, as results were practically identical, the models presented rely on the first method.
Employment status was a categorical variable, comparing "employed"
-including full-and part-time as well as freelancers-(reference category), "retired", "unemployed", "disabled", and "others" -which included students, homemakers, those on maternity leave, etc.-. Time-constant confounders. Although they drop out of the fixedeffects models, we use the following additional variables to describe the sample: Gender (male/female); Age at baseline, categorized in age groups (<40, 41-50, 51-60, 61-70 and >70); and Socioeconomic status (SES), assessed through educational level: a variable, based on the Dutch educational system, was created with three categories: "low" (up to general secondary education), "middle" (secondary vocational education, or higher general and pre-university education), and "high" (higher professional education or university education).
## Statistical analyses
Our analytical strategy was based on the following steps: first, linear regression models (OLS) accounting for fixed-effects (FE) were fitted, with weekly alcohol consumption as the outcome variable and the lockdown period as a main independent variable. Additionally, employment status was accounted for as a potential time-varying confounder (whereas time-constant confounders dropped out of the models). FE models focus on the changes within individuals throughout the observation period, net of time-constant unobserved confounding. Thus, they allow us to infer the impact of the lockdown on alcohol consumption. Based on this FE model, predicted margins estimated alcohol consumption levels at every time point. Sensitivity analyses with FE Poisson models were carried out, due to the large number of zero values of the outcome. As shown in the appendix -see table A2 and figure A2 -, the patterns are practically identical, although the coefficients differ because Poisson models are expressed different units, which makes them harder to interpret. In contrast, an advantage of OLS models is that they allow to interpret the coefficients in the same units of the outcome. For this reason, we decided to stick to OLS models.
Second, alcohol consumption during the lockdown was compared with alcohol consumption in previous years among the same population. For that purpose, data from the Lifelines Cohort (observations collected from 2007 until 2018) were used. Alcohol consumption during this period was estimated by means of cross-sectional pooled OLS models, in which alcohol consumption was regressed on the variable "month of the observation", using the same categories as the independent variable "lockdown period", for comparative purposes. Additionally, "year of observation" was accounted for in the models. Posterior margins predicted the estimated average alcohol consumption for each month, thus making results comparable with the ones from the COVID cohort.
Third, back to the COVID cohort, a potential role of living arrangement as moderator was tested by adding interactions between living arrangement and each time point of the lockdown period to the maineffects FE model. Additionally, the moderating role of subjective feelings of isolation was tested by adding an interaction between social isolation and the lockdown period. Again, based on this model, the predicted alcohol consumption at each time point for these subgroups of individuals was estimated. All analyses were carried out with Stata 13.
# Results
The descriptive analyses of the main variables of interest by alcohol consumption at baseline are shown in [fig_ref] Table 1: Main variables of interest by alcohol consumption at baseline [/fig_ref]. Women report much lower alcohol consumption than men (3.14 versus 5.68 glasses/week on average, respectively). Alcohol consumption is higher among older age groups (with the exception of those over 70), and individuals under 40 report the lowest consumption. The lower educated report the lowest drinking levels, whereas the higher educated report the highest (3.88 and 4.55 glasses/week respectively). As for employment status, the retired report the highest drinking (4.34), followed by the employed (4.26) and the unemployed and the disabled (3.67 and 3.05 respectively). Finally, those living alone, as well as those reporting higher levels of social isolation report lower alcohol consumption (3.77 and 4.06 respectively, compared to 4.18 and 4.14 among their counterparts).
## Alcohol consumption during the lockdown
An overview of the most relevant preventive measures implemented in the Netherlands during the observation period is shown in [fig_ref] Figure 1: Timeline of COVID-19 preventive measures implemented in the Netherlands [/fig_ref] (for a more exhaustive timeline see [fig_ref] Table 1: Main variables of interest by alcohol consumption at baseline [/fig_ref] in the appendix). The main effects of the lockdown period on alcohol consumption are shown in (Model 0). Based on this model, predicted drinking levels during the whole lockdown period were estimated, as shown in [fig_ref] Figure 2: Alcohol consumption during the COVID lockdown [/fig_ref].
Alcohol consumption levels tend to decrease during the first months of the so-called "intelligent lockdown", when bars and restaurants were closed and social interactions reduced to a minimum. At the beginning of June, there is an inflection point and a second phase starts, in which alcohol consumption shows a steep increase, with a peak around July-August, which could be attributed to the summer holidays season. However, it is also remarkable that, on June the 1st, most lockdown measures were relaxed and bars and restaurants were allowed to open again, as shown in [fig_ref] Figure 1: Timeline of COVID-19 preventive measures implemented in the Netherlands [/fig_ref].
By the end of August, a third phase starts with alcohol consumption steadily decreasing again. At that time, the Dutch government announced a renewed tightening of the measures due to the increasing number of infections (#3 in [fig_ref] Figure 1: Timeline of COVID-19 preventive measures implemented in the Netherlands [/fig_ref]. By mid-October, a ban on alcohol sales between 8 pm and 7 am was enacted (#4), as alcohol consumption was considered to be hindering the compliance with social distancing rules. Results show that alcohol consumption decreased throughout the autumn, reaching its lowest levels after the Christmas season. In January 2021 the lockdown measures were tightened again and a curfew was imposed (#6), due to the rising concern about the spread of new virus variants.
Finally, a fourth phase starts during Spring 2021, with an increase in alcohol consumption before the strict lockdown measures were relaxed by the end of April. Alcohol consumption steadily increases during early summer, when restrictions are lifted and vaccination starts to be massive, until the end of the observation period (end of July 2021). Although a new set of measures were implemented by mid-July (#8), its effects are unlikely to be captured by our data, since the observation period ended shortly after.
## Alcohol seasonality
In order to disentangle potential effects of the lockdown from the typical seasonal pattern in alcohol consumption, results have been compared with previous waves of Lifelines data. It is worthwhile to mention that alcohol consumption in the Netherlands is pretty average in European terms -e.g. the rate of binge drinking is the same as the European average (Eurostat, 2019), which is slightly lower than in the US (NIAAA, 2019), and there are no significant differences in terms of seasonality either.
Thus, [fig_ref] Figure 3: Seasonality in alcohol consumption [/fig_ref] compares alcohol consumption during the first year of the lockdown (blue line) with a cross-sectional analysis of pooled observations of waves 1, 4 and 5 of the Lifelines cohort study among the same individuals (red line) in the decade before the pandemic.
First, results clearly show that average alcohol consumption was lower during the lockdown period than in previous years . Second, although the seasonal pattern is roughly similar, there are some differences: while in previous years alcohol consumption steadily increased during spring and peaked in mid-summer, drinking levels decreased during spring 2020, which made the summer peak somewhat abrupter. After the summer period, the decline in alcohol consumption was steeper and steadier during the lockdown period. Last but not least, results from previous years show a small peak around Christmas -although rather small compared with the usual "January effect" reported in the literature [bib_ref] Seasonal variation in survey and sales estimates of alcohol consumption, Lemmens [/bib_ref] [bib_ref] Seasonal variation in Alcohol use, Uitenbroek [/bib_ref] [bib_ref] Seasonal variation in self-reports of recent alcohol consumption: racial and ethnic differences, Carpenter [/bib_ref] -, whereas in January 2021 not only no peak is observed but the lowest drinking levels of the whole observation period were reported.
In sum, results suggest that the lockdown had an effect in decreasing alcohol consumption. Moreover, it also modified alcohol seasonality, with decreased drinking levels during the initial "intelligent lockdown" and during the even more stringent lockdown in the winter of 2021, including the Christmas period.
## The moderating role of living arrangements and feelings of isolation
Our next aim was to test whether the effect of the pandemic on alcohol consumption differed by living arrangements and/or feelings of social isolation. For that purpose, interaction terms between these variables and time were added to the main effects model in a stepwise fashion.
First, as shown in [fig_ref] Figure 4: Alcohol consumption during the lockdown, by living arrangement [/fig_ref] , individuals who lived alone reported lower drinking levels throughout the whole observation period. Second, interaction coefficients are significant, as shown in Model 1 suggesting that the patterns differ between groups: the gap between those living alone and those with partner or family became smaller in relative terms during autumn and particularly winter, as those living Effects of the lockdown period on alcohol consumption. Main effects (Model 0) and interaction terms [fig_ref] Figure 1: Timeline of COVID-19 preventive measures implemented in the Netherlands [/fig_ref]. OLS fixed-effects models a . * p-value < 0.05; ** p-value < 0.01. a All models account for employment status. Model 0 accounts for "lockdown period"; Model 1 adds the interaction between "lockdown period" and "living arrangement"; Model 2 adds the interaction between "lockdown period" and "feelings of isolation". alone showed a steady increase in alcohol consumption between November 2020 and May 2021 -i.e. during the strictest lockdown period when bars were closed and bans on alcohol sales were imposed-, whereas those living with others show a clear decrease until February 2021, and only increase afterwards. Furthermore, those living in a shared household report a higher increase in their consumption during the summer periods. Sensitivity analyses with a distinction among the latter between those who lived with other adults only and those who lived with children were performed. As shown in [fig_ref] Figure 1: Timeline of COVID-19 preventive measures implemented in the Netherlands [/fig_ref] in the Appendix, the subgroup with children reports higher seasonal fluctuation, with very low consumption during autumn and winter but a steep increase in summer. That suggests that those living with others -and particularly those who lived with children-participated more in summer celebrations than those who lived alone.
Next, we tested the potential moderating role of social isolation by adding the interaction between "feelings of social isolation" and time to the model (Model 2 in . As the significant interaction coefficients show, differences between groups were smaller than by living arrangements, and mainly significant during the summer periods. As shown in [fig_ref] Figure 5: Alcohol consumption during the lockdown by feelings of social isolation [/fig_ref] , individuals who felt extremely socially isolated -i.e. isolation score ≥ 7-reported a lower increase in their alcohol consumption during both summer periods than those with no feelings of social isolation -score= 1-, whereas in the winter period they show similar trends.
Moreover, the interaction coefficients for living arrangement and time remained intact after adding the interaction with social isolation, showing that their moderation effects follow independent patterns. Taken together, results show that the living arrangement, in which individuals lived during the lockdown, played a somewhat bigger role overall, particularly in the autumn -winter period. In turn, social isolation played a somewhat bigger role in summer, probably because the ones who felt most isolated did not participate so much in the summer celebrations, suggesting that fluctuations in alcohol consumption during the summer months seem more related with social drinking than with alcohol being a "coping mechanism". Last but not least, results suggest that the patterns differ more by living arrangement during the second year of the lockdown: the fluctuation in alcohol consumption is very similar until November 2020 (although drinking levels differ). Yet, from November onwards those living alone show a slow but steady increase in the pattern, whereas those living with partner/family have a much greater variation in the slope. Consistently, as shown in [fig_ref] Table 1: Main variables of interest by alcohol consumption at baseline [/fig_ref] , the interaction term for those who lived alone in May 2020 was β = 0.02 (CI95% − 0.08 to 0.12), whereas in the same month in 2021 it was β = 0.22 (CI95% 0.10-0.35).
# Discussion
Our study, relying on a robust longitudinal design and a large sample, showed that overall alcohol consumption levels were far lower during the pandemic than in the years before. However, alcohol levels were not constant throughout the whole observation period but varied during the pandemic, in ways that partially differed from 'normal' seasonal patterns. Moreover, patterns differed by living arrangement and subjective feelings of social isolation, suggesting that different subgroups reacted to the pandemic and its related lockdown measures in different ways.
First, results show that alcohol consumption levels during the lockdown were lower than in previous years, which is probably due to the restrictions in social gatherings, and the closure of bars and restaurants, in line with what previous studies reported [bib_ref] The psychological impact of COVID-19 pandemic lockdowns: a review and meta-analysis of..., Panagiotidis [/bib_ref] [bib_ref] Changes in alcohol consumption associated with social distancing and self-isolation policies triggered..., Bade [/bib_ref].
Second, the seasonal pattern of alcohol consumption during the pandemic differs from the one in previous years among the same population. Thus, our results show lower drinking levels during the strictest lockdown periods, compared to the equivalent periods before the pandemic. To the best of our knowledge, no other study has compared the seasonality of alcohol consumption during the pandemic with the periods prior to it. Moreover, our results contrast with one consistent finding in the literature, namely: the overall increase in alcohol consumption levels around the Christmas season. While this has been widely reported in different contexts [bib_ref] Seasonal variation in survey and sales estimates of alcohol consumption, Lemmens [/bib_ref] [bib_ref] Seasonal variation in Alcohol use, Uitenbroek [/bib_ref] -sometimes called the "January effect" [bib_ref] Seasonal variation in self-reports of recent alcohol consumption: racial and ethnic differences, Carpenter [/bib_ref] -, our study shows that drinking levels during Christmas 2020 were at its lowest point, probably due to the strong restrictions in family gatherings that led many individuals to cancel or postpone their celebrations.
Third, we observe that changes in drinking patterns are not homogeneous across the whole population but differ by living arrangement and, to a lesser degree, by feelings of social isolation. Although it could be argued that both variables are highly related, it is interesting that their role as moderators is independent from each other, and the relative importance of one or the other changes throughout the observation period. Thus, the long covid winter of 2021 made a different impact among those living alone, who increased their drinking earlier than those sharing a household, in line with studies reporting increased solitary drinking during the lockdown [bib_ref] Drinking to Cope During COVID-19 Pandemic: The Role of External and Internal..., Wardell [/bib_ref]. On the other hand, those living with partner and/or family, particularly those with children, reported higher alcohol consumption in summer, probably in the context of outdoor activities and social gatherings during the Consistently, the subjective assessment of feeling socially isolated was associated with lower alcohol consumption in both summer periods, when most restrictions were lifted. Furthermore, social isolation shows a negative association with alcohol use, with those who feel most socially isolated drinking the least. This finding, at odds with the predictions that isolation during the lockdown would lead individuals to increase their alcohol consumption [bib_ref] Alcohol consumption and alcohol-related problems during the COVID-19 pandemic: a narrative review, Ramalho [/bib_ref] , suggests that, in the context of our study, drinking is mainly recreational and related with social events rather than a coping mechanism against loneliness, as previous evidence suggested [bib_ref] COVID 19 fear, stress, anxiety, and substance use among Russian and Belarusian..., Gritsenko [/bib_ref] [bib_ref] Drinking to Cope During COVID-19 Pandemic: The Role of External and Internal..., Wardell [/bib_ref]. These contrasting results may be due to different factors, ranging from study design -previous studies often rely on small sample sizes and cross-sectional designs-and different characteristics of the sample, e.g. in terms of age [bib_ref] COVID 19 fear, stress, anxiety, and substance use among Russian and Belarusian..., Gritsenko [/bib_ref] [bib_ref] Drinking to Cope During COVID-19 Pandemic: The Role of External and Internal..., Wardell [/bib_ref]. It is noteworthy that our sample is mainly composed by older adults, and those who report higher alcohol consumption in our study are mostly medium-educated males in their late fifties.
The question is, then, why would those who spent the strictest periods of the lockdown on their own report different drinking patterns, if it is not due to feelings of social isolation? Results suggest that other mechanisms may operate, e.g. higher family support may buffer the negative effects of the lockdown that could otherwise lead to higher alcohol consumption [bib_ref] Drinking to Cope During COVID-19 Pandemic: The Role of External and Internal..., Wardell [/bib_ref] [bib_ref] What is associated with the increased frequency of heavy episodic drinking during..., Valente [/bib_ref]. Alternatively, other mediators may be at play, e.g. boredom, which has been shown to be one of the main stressors during the pandemic [bib_ref] The relationship between perceived stress and emotional distress during the COVID-19 outbreak:..., Yan [/bib_ref] , as well as a risk factor for alcohol abuse [bib_ref] Comparing daily drivers of problem drinking among older and younger adults: An..., Kuerbis [/bib_ref]. The fact that parents with children at home reported the lowest alcohol consumption during the strictest periods of the lockdown seems to point in that direction, as they certainly did not have time to be bored.
Last but not least, the finding that differences between these subgroups became larger in the second year of the pandemic suggests that some of the lockdown-related stress may have accumulated over time. Moreover, the potential buffering effect of, e.g. family support, may become less effective over time. Further longitudinal research is needed in order to disentangle these potential mechanisms.
This study has several limitations. First, information on alcohol consumption did not contain specific information on drinking patterns such as solitary drinking and, for that reason, the study design was based on "proxy variables", such as living arrangement or feelings of isolation that, in a context of restricted social interactions, allowed us to approach this issue. However, interpretation of the results must be done being aware of this limitation. Second, although for those sharing a household, we had information about the household members and their ages, no information about their relationship status was available. However, given the age of our sample, and the very low prevalence of students, we assumed that most of them would be living with partner and/or family.
# Conclusions
This longitudinal study shows that the Covid pandemic and subsequent lockdown had an impact in reducing overall alcohol consumption levels in the northern provinces of the Netherlands. However, this effect was not homogeneous but differed by living arrangement and subjective feelings of social isolation: those living alone and/or feeling more socially isolated reacted more strongly to the lockdown. Moreover, the impact of the lockdown was not constant over time but it was stronger, the longer it lasted, suggesting an accumulative effect and a reduction of resilience of some groups of individuals to cope with it.
## Credit authorship contribution statement
Lluís Mangot-Sala: Data cleaning and harmonization, Statistical analyses, Drafting of the manuscript. Khoa A. Tran: Literature review, Drafting of the manuscript. Aart C. Liefbroer, Nynke Smidt: Obtained funding and study supervision. Lluís Mangot-Sala, Aart C. Liefbroer, Nynke Smidt: Study concept and design, Acquisition and interpretation of data, Critical revision of the analyses and the manuscript.
# Funding
This work was supported by the Research Fund of the Royal Netherlands Academy of Arts and Sciences (KNAW-Institutes).
# Author disclosure
Lluís Mangot-Sala was responsible for the data cleaning and harmonization, performed the statistical analyses and wrote the manuscript. Khoa A. Tran carried out the literature review and wrote a first draft of the introduction. Aart C. Liefbroer and Nynke Smidt obtained funding for the study, supervised it and reviewed the final versions of the manuscript. The study concept and design, acquisition and interpretation of data, and the critical revision of the analyses and the manuscript were carried out by Lluís Mangot-Sala, Aart C. Liefbroer and Nynke Smidt. All authors have read and approved the manuscript for submission. None of the original material contained in the manuscript has been submitted for consideration nor will any of it be published elsewhere.
## Conflicts of interest
None.
# Data availability statement
The data underlying this article were provided by Lifelines under licence. Access to the data can be granted under licence by Lifelines and the authors will share their codes used to produce the results presented in this paper upon request.
[fig] Figure 1: Timeline of COVID-19 preventive measures implemented in the Netherlands (March 2020 -July 2021). 11 [/fig]
[fig] Figure 2: Alcohol consumption during the COVID lockdown (April 2020 -July 2021). Predictive margins based on fixed-effects linear regression models. [/fig]
[fig] Figure 3: Seasonality in alcohol consumption. Comparison between the lockdown period (blue line) and previous observations from the Lifelines cohort (2007-2018) (dashed red line). [/fig]
[fig] Figure 4: Alcohol consumption during the lockdown, by living arrangement. Predictive margins based on fixed-effect models including the interaction term. [/fig]
[fig] Figure 5: Alcohol consumption during the lockdown by feelings of social isolation. Predictive margins based on fixed-effect models including the interaction term. summer holidays season. [/fig]
[table] Table 1: Main variables of interest by alcohol consumption at baseline (n = 45,384). [/table]
|
Tension pneumothorax as a severe complication of endobronchial ultrasound-guided transbronchial fine needle aspiration of mediastinal lymph nodes
StreszczenieW pracy zaprezentowano opis przypadku odmy opłucnowej prężnej, która wystąpiła u pacjenta z rozedmą pęcherzową i przewlekłą obturacyjną chorobą płuc po wykonaniu przezoskrzelowej biopsji aspiracyjnej cienkoigłowej węzłów chłonnych śródpiersia wykonanej pod kontrolą ultrasonografii wewnątrzoskrzelowej (endobronchial ultrasound-guided transbronchial needle aspiration -EBUS-TBNA). Odma prężna jest ciężkim, jednak rzadkim powikłaniem EBUS-TBNA. Może być wynikiem zarówno uszkodzenia płuca igłą biopsyjną, jak iu pacjentów z rozedmą pęcherzową -samoistnego pęknięcia pęcherza rozedmowego na skutek wzrostu ciśnienia w klatce piersiowej podczas kaszlu wywołanego wprowadzeniem bronchofiberoskopu do drzewa oskrzelowego. Autorzy zwracają uwagę na konieczność ścisłego monitorowania chorych po wykonaniu biopsji oraz zapewnienia dostępu do natychmiastowego leczenia powikłań w warunkach szpitalnych. Pacjentów o zwiększonym ryzyku powikłań należy zidentyfikować przed przystąpieniem do zabiegu i szczególnie uważnie monitorować po jego wykonaniu. Słowa kluczowe: odma prężna, bronchofiberoskopia, przezoskrzelowa biopsja aspiracyjna cienkoigłowa pod kontrolą ultrasonografii wewnątrzoskrzelowej (EBUS-TBNA).CASE REPORTSAbstractThis article presents a case report of a patient suffering from bullous emphysema and chronic obstructive pulmonary disease, who was diagnosed with tension pneumothorax after undergoing endobronchial ultrasound-guided transbronchial needle aspiration (EBUS-TBNA). Tension pneumothorax is a severe but rare complication of EBUS-TBNA. It can result from lung injury caused by the biopsy needle or, in patients suffering from bullous emphysema, from spontaneous rupture of an emphysematous bulla resulting from increased pressure in the chest cavity during cough caused by bronchofiberoscope insertion. The authors emphasize that patients should be carefully monitored after the biopsy, and, in the case of complications, provided with treatment immediately in proper hospital conditions. Patients burdened with a high risk of complications should be identified before the procedure and monitored with extreme care after its completion.
# Introduction
The most widely accepted method of minimally invasive diagnostics for mediastinal lymph node enlargement in the course of lung cancer and sarcoidosis is endobronchial ultrasonography (EBUS), which has been in use for over a decade [bib_ref] Real-time endobronchial ultrasound guided transbronchial needle aspiration for sampling mediastinal lymph nodes, Herth [/bib_ref] [bib_ref] Endosonography vs conventional bronchoscopy for the diagnosis of sarcoidosis: the GRANULOMA randomized..., Von Bartheld [/bib_ref]. In cases in which imaging of the posterior mediastinum and the left adrenal gland is required, the examination can be supplemented with endoesophageal ultrasonography (EUS) [bib_ref] Transesophageal endoscopic ultrasound-guided fine-needle aspiration (EUS-FNA) and endobronchial ultrasound-guided transbronchial needle aspiration..., Vilmann [/bib_ref]. The joint examination, known as combined ultrasonography (CUS), was initially performed using two devices (a bronchofiberoscope and a gastrofiberoscope) equipped with a sector ultrasound transducer. At present, the most popular method is CUSb, which employs a single bronchofiberoscope; this significantly reduces the duration of the examination and the cost associated with the procedure without having a significant impact on Tension pneumothorax as a severe complication of endobronchial ultrasound-guided transbronchial fine... the diagnostic efficacy of the method [bib_ref] Combined endoscopicendobronchial ultrasound-guided fine-needle aspiration of mediastinal lymph nodes through a single..., Herth [/bib_ref] [bib_ref] A comparison of the combined ultrasound of the mediastinum by use of..., Szlubowski [/bib_ref]. The most important advantage of the described diagnostic methods is that needle aspiration can be performed in real time under ultrasound control, providing cytological material for histopathological assessment.
The combined ultrasound-needle aspiration (CUSb-NA) examination is considered relatively safe and is well tolerated by the patients when short-term analgesia and sedation is administered intravenously [bib_ref] Endosonography vs conventional bronchoscopy for the diagnosis of sarcoidosis: the GRANULOMA randomized..., Von Bartheld [/bib_ref] [bib_ref] A comparison of the combined ultrasound of the mediastinum by use of..., Szlubowski [/bib_ref]. The relevant complications that are most often described in the literature are infectious complications after endobronchial ultrasound-guided transbronchial needle aspiration (EBUS-TBNA) [bib_ref] A rare constellation of empyema, lung abscess, and mediastinal abscess as a..., Huang [/bib_ref] [bib_ref] Mediastinal abscess after endobronchial ultrasound with transbronchial needle aspiration: a case report, Moffatt-Bruce [/bib_ref].
The incidence of EBUS-TBNA complications in the largest survey study conducted by the Japan Society for Respiratory Endoscopy was estimated at 1.23%; the most common complications were harmless self-limiting bleeding (0.68%) and infectious complications (0.14%). The latter included mediastinitis, pneumonia, pericarditis, infections after cyst puncture, and even sepsis [bib_ref] Complications associated with endobronchial ultrasound-guided transbronchial needle aspiration: a nationwide survey by..., Asano [/bib_ref]. Pneumothorax as a complication of EBUS-TBNA was diagnosed in only 2 patients out of over seven thousand examined (0.03%), and only one of these patients required pleural drainage [bib_ref] Complications associated with endobronchial ultrasound-guided transbronchial needle aspiration: a nationwide survey by..., Asano [/bib_ref]. In the largest American study, ACCP AQuIRE (American College of Chest Physicians Quality Improvement Registry, Education), which included 1317 patients, the incidence of complications after EBUS-TBNA was 1.44%; pneumothorax occurred in as many as 7 patients, in most cases after lung biopsy with the EBUS-TBNA method. One case of pneumothorax has been described as a major complication after CUS-NA performed during the only randomized study comparing the efficacy of echosonography with mediastinoscopy [bib_ref] Mediastinoscopy vs endosonography for mediastinal nodal staging of lung cancer: a randomized..., Annema [/bib_ref].
## Case study
The 63-year-old male patient with diagnosis of infiltrative lesions of the left lung and hilum as well as mediastinal lymphadenopathy [fig_ref] Figure 1: Chest computed tomography [/fig_ref] and suspicion of hyperplasia was admitted to the Clinical Department of Thoracic Surgery of the John Paul II Specialist Hospital in Krakow for histopathological verification. The patient's medical history included chronic obstructive pulmonary disease (COPD) and bullous emphysema of both lungs; seven years earlier, he underwent left-sided upper lobectomy due to the presence of non-neoplastic cysts. Infiltration of the left hilar area and the mediastinal lymph node packets were revealed by a chest computed tomography (CT) scan performed in July, 2013 at the Department of Pulmonology, where the patient was hospitalized due to COPD exacerbation with bronchogenic inflammation of the right lung. In the course of further diagnostics, an out-patient positron emission tomography-computed tomography (PET-CT) examination was performed, revealing moderate metabolic stimulation (SUV 3.5-6.2) of the infiltrative-atelectatic lesions of the left lung and mediastinal lymph nodes: group 4R (right lower paratracheal) and group 7 (subcarinal).
The CUSb examination performed in August, 2013 confirmed the presence of perihilar infiltration of the lower lobe of the left lung (visible on both the EBUS and the EUS image). However, due to the fibrosis and rotation of the lower lung bronchial openings, EBUS-TBNA was not performed as the lesion was outside the reach of the biopsy needle. Group 7 lymph node biopsy was performed, and the procedure was uneventful. Cytological assessment did not demonstrate the presence of neoplastic cells.
In September 2013, the patient was readmitted to the Clinical Department of Chest Surgery due to persisting radiological changes in the left lung and mediastinal lymph node enlargement. The patient again underwent a CUSb-NA examination, and material for cytological investigation was obtained from the lymph nodes (group 7 -with EUS-FNA, and group 4R -with EBUS-TBNA; [fig_ref] Figure 2: Endobronchial ultrasound-guided transbronchial needle aspiration [/fig_ref].
During the examination and immediately after, the patient reported no complaints. No saturation drop was observed in pulse oximetry. After the examination, the patient was transferred from the Endoscopy Unit to the ward. Approximately 30 min after the end of the endoscopic examination, the patient reported sudden pain in the right half of the chest, accompanied by dyspnea. Physical examination revealed tachypnea, no alveolar murmur above the right side of the chest, and tympanitic resonance. The obtained chest radiogram visualized right-sided pneumothorax and mediastinal displacement to the left [fig_ref] Figure 3: X-ray of the chest -right-sided tension pneumothorax [/fig_ref]. During the diagnostic examinations, the patient's condition deteriorated: he lost consciousness, his arterial blood pressure dropped to 50/20 mmHg, and his heart rate slowed to 30 bpm (bradycardia). A drain was immediately inserted into the right pleural cavity, and suction drainage was applied, resulting in swift improvement of the patient's condition. Further radiograms confirmed full expansion of the right lung [fig_ref] Figure 4: X-ray of the chest after removal of the drain -normally expanded lungs [/fig_ref]. The drain was removed on the 4 th day; on the next day, the patient was discharged home in good general condition.
# Discussion
Methods of endoscopic echosonography are safe for patients [bib_ref] Real-time endobronchial ultrasound guided transbronchial needle aspiration for sampling mediastinal lymph nodes, Herth [/bib_ref] [bib_ref] Endosonography vs conventional bronchoscopy for the diagnosis of sarcoidosis: the GRANULOMA randomized..., Von Bartheld [/bib_ref] [bib_ref] Transesophageal endoscopic ultrasound-guided fine-needle aspiration (EUS-FNA) and endobronchial ultrasound-guided transbronchial needle aspiration..., Vilmann [/bib_ref] [bib_ref] Combined endoscopicendobronchial ultrasound-guided fine-needle aspiration of mediastinal lymph nodes through a single..., Herth [/bib_ref] [bib_ref] A comparison of the combined ultrasound of the mediastinum by use of..., Szlubowski [/bib_ref] [bib_ref] Mediastinoscopy vs endosonography for mediastinal nodal staging of lung cancer: a randomized..., Annema [/bib_ref]. According to the largest publications, the incidence of all complications is approximately 1-1.5% [bib_ref] Complications associated with endobronchial ultrasound-guided transbronchial needle aspiration: a nationwide survey by..., Asano [/bib_ref]. Most major complications are associated with the needle aspiration. The most common ones include inflammatory complications (mediastinitis, pneumonia, pericarditis), bleeding complications (bleeding into the lumen of the bronchial tree or into the mediastinum), pneumothorax, and mediastinal emphysema. Most often, the complications pose no threat to the patient's life [bib_ref] A rare constellation of empyema, lung abscess, and mediastinal abscess as a..., Huang [/bib_ref] [bib_ref] Mediastinal abscess after endobronchial ultrasound with transbronchial needle aspiration: a case report, Moffatt-Bruce [/bib_ref] [bib_ref] Complications associated with endobronchial ultrasound-guided transbronchial needle aspiration: a nationwide survey by..., Asano [/bib_ref] [bib_ref] Mediastinoscopy vs endosonography for mediastinal nodal staging of lung cancer: a randomized..., Annema [/bib_ref].
Pneumothorax may result from injury of the pulmonary pleura of a previously unchanged lung or from an accidental puncture of an emphysematous bulla during the needle biopsy. In patients with bullous emphysema, this complication may occur as a result of a spontaneous rupture of an emphysematous bulla, caused by an increase of pressure in the chest resulting from strenuous cough during bronchofiberoscopy.
In the present case, the emphysematous bullae adhered directly to the group 4 lymph nodes biopsied during the second CUSb examination, as shown in [fig_ref] Figure 5: Chest tomography [/fig_ref].
The management strategy in view of diagnosed emphysema depends on the lesion's size and the clinical condition of the patient. If the emphysema is small, and the patient's condition is stable, it is sufficient to employ passive oxygen therapy and monitor the size of the emphysematous chamber. Patients with larger emphysemas require suction drainage of the pleural cavity. In the case of tension pneumothorax, introducing a drain into the pleural cavity is a life-saving procedure and must be performed immediately after diagnosis.
Pneumothorax is a particularly rare complication after echosonography [bib_ref] Complications associated with endobronchial ultrasound-guided transbronchial needle aspiration: a nationwide survey by..., Asano [/bib_ref] [bib_ref] Mediastinoscopy vs endosonography for mediastinal nodal staging of lung cancer: a randomized..., Annema [/bib_ref]. Nonetheless, the possibility of its occurrence must be taken into account; patients should be monitored during the first hours after the procedure, and their complaints, such as chest pain, dyspnea, or cough, should be addressed immediately. Delaying the diagnosis and, consequently, the implementation of appropriate treatment, may have dire consequences for the patient. Therefore, it appears justified to monitor patients after EBUS/EUS procedures for at least 24 hours in order to minimize the risk to their lives. It is also of paramount importance to identify patients burdened with a higher risk of complications (e.g. with advanced emphysema, as in the case of the present patient) before the procedure and to carefully monitor such patients after the procedure is completed.
# Disclosure
Authors report no conflict of interest.
[fig] Figure 2: Endobronchial ultrasound-guided transbronchial needle aspiration (EBUS-TBNA) -aspiration biopsy of group 4R lymph nodes [/fig]
[fig] Figure 1: Chest computed tomography (CT) (mediastinal window)enlarged mediastinal lymph nodes (N4R) and infiltrative-atelectatic lesions of the left lung and hilum CASE REPORTS [/fig]
[fig] Figure 4: X-ray of the chest after removal of the drain -normally expanded lungs [/fig]
[fig] Figure 3: X-ray of the chest -right-sided tension pneumothorax [/fig]
[fig] Figure 5: Chest tomography (pulmonary window) -emphysematous bullae adhering to the mediastinal lymph nodes Tension pneumothorax as a severe complication of endobronchial ultrasound-guided transbronchial fine... [/fig]
|
Premorbid traumatic stress and veteran responses to the COVID‐19 pandemic
The COVID-19 pandemic has had unprecedented effects on lifestyle stability and physical and mental health. We examined the impact of preexisting posttraumatic stress disorder (PTSD), alcohol use disorder (AUD), and depression on biopsychosocial responses to the pandemic, including psychiatric symptoms, COVID-19 exposure, and housing/financial stability, among 101 U.S. military veterans enrolled in a longitudinal study of PTSD, a population of particular interest given veterans' trauma histories and defense-readiness training. Participants (83.2% male, 79.2% White, M age = 59.28 years) completed prepandemic, clinicianadministered psychiatric diagnostic interviews and a phone-based assessment between May and September 2020 using a new measure, the Rapid Assessment of COVID-19-Related Experiences (RACE), which was used to assess pandemic responses and its effects on mental and physical health; COVID-19 diagnosis and testing were also extracted from electronic medical records. Multivariate regressions showed that, controlling for demographic characteristics, prepandemic PTSD, β = .332; p = .003, and AUD symptoms, β = .228; p = .028, were associated with increased pandemic-related PTSD symptoms. Prepandemic AUD was associated with increased substance use during the pandemic, β = .391; p < .001, and higher rates of self-reported or medical record-based COVID-19 diagnosis, β = .264; p = .019. Minority race was associated with pandemic-related housing/financial instability, β = −.372; p < .001, raising concerns of population inequities. The results suggest that preexisting PTSD and AUD are markers for adverse pandemic-related psychiatric outcomes and COVID-19 illness. These findings carry implications for the importance of targeting prevention and treatment efforts for the highest-risk individuals.
In late 2019, the first cases of the coronavirus disease 2019 from the SARS-CoV-2 virus were reported in China. COVID-19 rapidly grew to the scale of a global pandemic, and by March 2020, there was widespread community transmission necessitating masking, social distancing, quarantines, school closures, and other major lifestyle changes to mitigate exposure and disease spread. As the only global pandemic since the Spanish Flu of 1918−1919, COVID-19 has had an unprecedented impact on society, the economy, and health care. The effects of COVID-19 on physical health and mortality have been devasting for the global population, especially for the elderly, racial/ethnic minorities, and individuals with preexisting chronic health conditions [bib_ref] Health equity and COVID-19: Global perspectives, Shadmi [/bib_ref]. The pandemic also poses risks to mental health and well-being [bib_ref] Mental health and the COVID-19 Pandemic, Pfefferbaum [/bib_ref] , including anxiety about threats to health and safety, depression and loneliness related to social isolation and loss, and the potential for pandemic-related increases in substance misuse and housing and financial instability. There is also stress associated with the need for increased vigilance (e.g., to avoid coronavirus exposure) and adaptability (e.g., to follow changing guidelines and community regulations).
Numerous studies have demonstrated the detrimental effects of the COVID-19 pandemic on mental health. In a self-report study of 2,485 college students, participants who lived in areas with higher rates of COVID-19 infection experienced more severe symptoms of PTSD and depression compared with students living in less affected areas [bib_ref] Prevalence and correlates of PTSD and depressive symptoms one month after the..., Tang [/bib_ref]. Similarly, in a sample of 7,143 undergraduates, participants who had family or friends who contracted the coronavirus reported higher levels of anxiety symptoms compared with those without proximal exposure to the virus [bib_ref] The psychological impact of the COVID-19 epidemic on college students in China, Cao [/bib_ref]. Increases in anxiety were also reported among students who experienced pandemicrelated disruptions to daily life or economic standing [bib_ref] The psychological impact of the COVID-19 epidemic on college students in China, Cao [/bib_ref]. Thus, indirect effects of the pandemic, such as employment or financial instability, may also contribute to worsened mental health [bib_ref] Psychological aspects of employment instability during the COVID-19 pandemic, Mimoun [/bib_ref]. In light of this, the identification of individuals with the highest risk of mental health concerns carries important public health implications such that it could inform prevention and intervention efforts aimed at mitigating the deleterious long-term effects of the pandemic.
Individuals with preexisting mental health conditions are likely among those at increased risk for detrimental outcomes in response to the pandemic. Preexisting PTSD may be a vulnerability given that exposure to new-onset stressors (e.g., the COVID-19 pandemic) has been shown to increase PTSD symptom severity associated with prior trauma exposure [bib_ref] Posttraumatic stress in aging World War II survivors after a fireworks disaster:..., Bramsen [/bib_ref] [bib_ref] Impact of new traumatic or stressful life events on preexisting PTSD in..., Schock [/bib_ref]. In addition, the omnipresent threat and mandates to social distance stemming from the COVID-19 pandemic may reinforce or even exacerbate symptoms of PTSD, such as emotional disconnect from others, hypervigilance about safety, behavioral avoidance of crowds and other social situations, and maladaptive cognitions about the basic safety of the world. In a web-based self-report study using a large, cross-sectional sample of veterans in the United Kingdom with preexisting mental health conditions (i.e., PTSD, anxiety, depression, anger difficulties, alcohol misuse), [bib_ref] Exploring the impact of COVID-19 and restrictions to daily living as a..., Murphy [/bib_ref] found that the pandemic exacerbated preexisting psychiatric symptoms, especially symptoms of depression, anxiety, and PTSD. This effect was moderated by social support such that individuals with low levels of social support experienced steeper declines in preexisting mental health conditions [bib_ref] Exploring the impact of COVID-19 and restrictions to daily living as a..., Murphy [/bib_ref].
Health-related anxiety about the virus and isolation resulting from social distancing also has the potential to exacerbate other disorders that are often comorbid with PTSD, such as depression and alcohol use disorder (AUD). Given the association between loneliness and depression [bib_ref] Relationship between depression and loneliness in elderly and examination of influential factors, Aylaz [/bib_ref] [bib_ref] A mixed-methods approach to understanding loneliness and depression in older adults, Barg [/bib_ref] , pandemicrelated social isolation may worsen depression. Similarly, drinking alcohol has been shown to be a coping mechanism for stress [bib_ref] The relationship between coping strategies, alcohol expectancies, drinking motives, and drinking behaviour, Hasking [/bib_ref] [bib_ref] PTSD symptoms mediate academic stress and drinking to cope in college students, Woolman [/bib_ref] and trauma [bib_ref] Trauma-related drinking to cope: A novel approach to the self-medication model, Hawn [/bib_ref] , and a recent study demonstrated an association between COVID-19-related stress and increased alcohol use in the general population [bib_ref] Drinking to cope with the pandemic: The unique associations of COVID-19-related perceived..., Rodriguez [/bib_ref]. Thus, for individuals who have preexisting depression and AUD, especially in conjunction with PTSD, pandemic-related stress may worsen these symptoms.
Preexisting mental health conditions may also increase the risk of poorer physical health outcomes related to the pandemic. In their recent review, [bib_ref] Alcohol use disorder: A preexisting condition for COVID-19?, Bailey [/bib_ref] posited that preexisting AUD might increase the risk of contracting COVID-19 and increase disease severity. A study bysupported this suggestion: The authors conducted a retrospective study of 73,099,850 patients using electronic health database records from hospitals and health care providers across the United States to (n = 12,030 with a COVID-19 diagnosis, n = 7,510,380 with a substance use disorder [SUD] diagnosis) and found that those with SUDs had an increased risk of contracting COVID-19. Additionally, in a retrospective study of 7,348 adult COVID-19 patients, schizophrenia spectrum disorders were associated with an increased risk of mortality [bib_ref] Association of psychiatric disorders with mortality among patients with COVID-19, Nemani [/bib_ref] , further demonstrating the potential impact of psychiatric conditions on COVID-19 illness risk and severity.
Although numerous studies have examined the effects of both the COVID-19 pandemic broadly and the viral illness specifically on mental health, there has been relatively less research to date concerning how a range of preexisting psychiatric diagnoses and symptoms might predict biopsychosocial responses to the pandemic, including psychiatric symptoms, exposure to the disease, and financial and housing stability during the pandemic. The studies that do exist have been limited by cross-sectional and retrospective designs and the use of web-based self-report data. Temporal separation of predictors and outcomes using a longitudinal design and clinician-based assessments may offer a more valid and nuanced approach to evaluating mental health predictors of psychiatric symptoms in response to the pandemic. This would allow for targeted prevention, monitoring, and intervention to be made available to the highest-risk individuals. Our research group had the opportunity to address these questions as we were in the midst of a longitudinal study of veterans with PTSD and comorbid psychiatric conditions when the pandemic began. The primary aim of this study was to examine how baseline (i.e., prepandemic) clinician-assessed traumatic stress (i.e., PTSD, AUD, and depression) predicted three biopsychosocial responses to the pandemic: (a) changes in mental health symptoms during the COVID-19 pandemic, (b) exposure to the SARS-COV-2 virus and associated illness, and (c) changes in housing and financial stability. We developed a new measure, the Rapid Assessment of COVID-19-Related Experiences (RACE), to evaluate these outcomes, and, thus, a secondary study aim was to conduct a preliminary evaluation of the measure's psychometric characteristics (i.e., reliability and factor structure). To our knowledge, no other measure to date assesses changes in a range of pandemic-related experiences among individuals with preexisting trauma-related and psychiatric symptoms.
We examined these issues in a longitudinal sample of 101 trauma-exposed United States military veterans. Traumaexposed veterans are a particularly unique population to study for these purposes because social distancing mandates and the threat of COVID-19 may reinforce or exacerbate preexisting PTSD symptoms, while, at the same time, veterans have unique military and disaster preparedness training that might also influence responses to the pandemic. We hypothesized that more severe psychiatric symptoms at baseline would predict poorer pandemicrelated psychiatric and housing and financial outcomes. We also hypothesized that baseline psychiatric disorders, specifically AUD and PTSD, would increase the risk of COVID-19 illness given that preexisting SUDs have been shown to increase the risk of COVID-19 infectionand that PTSD is associated with adverse physical health outcomes [bib_ref] Posttraumatic stress disorderrelated cardiovascular disease and accelerated cellular aging, Wolf [/bib_ref]. Because demographic characteristics have been shown to be differentially associated with PTSD (e.g., minorities, women, and younger individuals tend to have more severe symptoms; [bib_ref] The combined effect of gender and age on post-traumatic stress disorder: Do..., Ditlevsen [/bib_ref] [bib_ref] Age differences in posttraumatic stress disorder, psychiatric disorders, and healthcare service use..., Frueh [/bib_ref]
# Method participants
Participants (N = 101) were drawn from a cohort of veterans who were enrolled in an ongoing longitudinal study of PTSD and common comorbid psychiatric conditions, including depression and AUD, prior to the start of the COVID-19 pandemic. To be eligible for the longitudinal study, veterans 18 years of age or older had to endorse a history of trauma exposure. Veterans were excluded if they were at immediate risk of self-or other-harm or if they were under the acute influence of substances. To be eligible for the follow-up COVID-19-related phone assessment, participants were required to have completed the most recent follow-up assessment before the start of the pandemic; this assessment was used as the baseline data for the present analyses. Of the 175 initially eligible participants, 102 individuals were reached by telephone for this assessment; of these veterans, one declined to participate. Participant characteristics are provided in . The resulting study sample was primarily male (83.2%), White (79.2%), and middle-aged (M = 59.28 years, SD = 12.92). The most common types of trauma exposure were combat or warfare (43.6%), unwanted sexual contact as an adult (11.9%), and an assault by an acquaintance or stranger (5.9%). Most participants had experienced multiple types of traumatic experiences (M = 7.01, SD = 4.40). Analyses revealed no significant differences with respect to age; sex; minority status; educational attainment; or PTSD, AUD, and depression diagnosis and severity among eligible participants who did versus did not participate in the COVID-19 phone assessment, ps = .159-.950.
## Procedure
Eligible veterans were contacted by phone and asked to participate in a brief phone-based assessment related to the COVID-19 pandemic. Elements of informed consent were discussed at the start of the phone call. Phone assessments took place between May and September 2020. On average, the assessments occurred 522.05 days (SD = 234.17, range: 138−1005) after the participant's most recent in-person study evaluation. Prepandemic assessments were completed in-person and consisted of self-report assessments, structured diagnostic interviews, a brief neuropsychological assessment, and a blood draw; each prepandemic assessment lasted approximately 4.5 hr. The diagnostic assessments were videotaped, with approximately 30% coded by a second rater to determine diagnostic reliability. Participants had scheduled breaks for food and were able to take additional breaks as needed. The protocol was approved by the VA Boston Healthcare System Institutional Review Board. Participants received $135 (USD) for their participation in the prepandemic in-person assessment and $20 for their participation in the COVID-19-related phone assessment. In addition to asking participants about whether they had been diagnosed with COVID-19 as part of the phone assessment, we also had permission to view the VA electronic medical record (EMR) to determine participants' history of COVID-19 testing and diagnoses. Medical record information was evaluated as of April 13, 2021.
## Measures
## Pandemic responses
The RACEwas used to assess a range of biopsychosocial pandemic responses. This questionnaire was developed for this study and intended to efficiently measure a range of both adverse and adap-tive biopsychosocial responses to the COVID-19 pandemic, including changes in psychiatric symptoms; to date, no other measures of which we are aware have been established for this purpose. The RACE was rationally derived after reviewing the literature concerning mental health questionnaires developed in response to prior SARS viruses. The instrument consists of 26 items across five subscales that are meant to capture recent self-reported pandemic-related changes in (a) housing and financial stability (two items), (b) mood and anxiety symptoms (three items), (c) substance use (i.e., alcohol use, nonalcohol substance use, and prescription drug use; three items), (d) PTSD symptoms (eight items), and (e) the respondent's COVID-19 exposure and exposure among their friends and family (five items). In addition, the RACE includes descriptive items that are used to assess social distancing habits (three items), concern over contracting COVID-19 (one item), and a rating of change in sense of personal resilience during the pandemic (one item). The personal resilience item was included to ensure that the RACE covered both adaptive and maladaptive psychiatric responses, as there is evidence that resilience and psychiatric symptoms are at opposite ends of the same underlying construct [bib_ref] A classical twin study of PTSD symptoms and resilience: Evidence for a..., Wolf [/bib_ref]. Items related to psychiatric symptoms and personal resilience anchor response options to the past 2 weeks and followed a standard structure with Likert-like response options. For example, the item related to pandemic-related increases in PTSD-associated nightmares was phrased as follows: "In the past 2 weeks, to what extent has the COVID-19 pandemic affected your nightmares about past trauma?" Responses were rated on a 5-point scale ranging from 1 (I am much less bothered by trauma nightmares than usual) to 5 (I am a lot more bothered by trauma nightmares than usual), with no change in symptoms at the midpoint of the scale. The COVID-19 exposure items (e.g., "Have you been diagnosed with COVID-19 by a healthcare professional?") and questions about social distancing (e.g., "Were you able to practice social distancing starting around mid-March?") used dichotomous (i.e., "yes" or "no") response options. Affirmative responses to the initial COVID-19 items related to personal infection and illness among friends and family were followed by additional questions used to assess the degree of illness (e.g., history of hospitalization, intubation). For the present paper, we refer to self-exposure to COVID-19 illness as "exposure," based on self-report and medical record data, whereas exposure of a participant's family or close friend to COVID-19 is referred to as "proximal exposure." The psychiatric symptom and resilience response options were keyed such that higher scores indicate more pathological outcomes, and the Housing/Financial subscale was keyed so higher scores indicated higher levels of stability.
The Mood/Anxiety and PTSD Symptom subscales demonstrated adequate, Cronbach's α = .64, and good internal consistency, Cronbach's α = .83, respectively. Reliability was not assessed for the other subscales, as they were either composed of a single item or of items that were not expected to covary because they were assessing distinct phenomena. Items were read to participants over the phone along with the possible response options. The RACE is included in the Supplementary Materials.
## Ptsd
The Clinician-Administered PTSD Scale for DSM-5 (CAPS) for the fifth edition of the Diagnostic and Statistical Manual of Mental Disorders (DSM-5; i.e., CAPS-5;is the gold standard diagnostic tool for PTSD. The CAPS-5 was administered by interviewers, ranging from trained bachelor's-level psychology technicians to licensed clinical psychologists, during the prepandemic in-person assessment and was used to assess current (i.e., past-month) PTSD diagnosis and symptom severity. PTSD symptoms were anchored to the participant's self-identified prepandemic most distressing traumatic experience. The diagnostic reliability based on a subset of about 30% of the interviews from the initial data collection demonstrated good interrater reliability, PTSD diagnosis: k = .78, intraclass correlation coefficient (ICC) for PTSD severity: r = .78.
## Other mental health disorders
Sections of the Structured Clinical Interview for DSM-5 (SCID;were administered at the prepandemic in-person assessment following standard SCID administration rules. Administered modules included those on major depressive disorder (MDD), substance use disorders, generalized anxiety disorder, panic disorder, agoraphobia, and antisocial personality disorder from the SCID-PD. For the present study, we examined common diagnoses on the SCID (i.e., MDD and AUD; see and symptom summary scores on these modules as predictors of COVID-19-related outcomes. The diagnostic reliability from the prepandemic data collection demonstrated good interrater reliability for SCID-based diagnoses, depression diagnosis: k = .94, AUD diagnosis: k = .93.
# Data analysis
As the RACE was newly developed for the present study, we first conducted a confirmatory factor analysis (CFA) of items included in the Mood/Anxiety and PTSD subscales using the weighted least squares estimator (WLSMV) to account for the categorical nature of the response options. The CFA was conducted using the Mplus (Version 8.5) statistical modeling software and evaluated using standard fit indices and guidelines [bib_ref] Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus..., Hu [/bib_ref] , including root mean square error of approximation (RMSEA; values less than .06 are indicative of good model fit), standardized root mean square residual (SRMR; (values less than .08 are consistent with good fit), and the comparative fit index (CFI) and Tucker-Lewis index (TLI; values of .95 or greater on both suggest good model fit). These fit statistics were evaluated together such that a single fit statistic that fell outside these guidelines in a model that otherwise demonstrated a good fit to the data would not negate the overall acceptability of the model. We also tested a competing model based on the initial CFA results and compared the fit using a nested chi-square test, adjusting for the use of the WLSMV estimator using the DIFFTEST function in Mplus.
Next, we conducted bivariate correlations in SPSS (Version 26) to assess the associations among baseline psychiatric symptoms, pandemic-related psychiatric symptoms, and COVID-19 exposure. We then ran five linear regression equations to examine how prepandemic psychiatric symptom severity predicted pandemic-related changes in psychiatric symptoms (i.e., PTSD, mood and anxiety symptoms, substance use) and experiences (i.e., COVID-19 exposure, housing and financial stability). In each regression, prepandemic PTSD, AUD, and depressive symptom severity were included as predictors together in the model, controlling for age, sex, race (minority vs. nonminority), and educational attainment. For the analyses predicting COVID-19 exposure, we used a variable that reflected participants who either self-reported a diagnosis of COVID-19 or who had a positive test result in their EMR. Prepandemic PTSD severity was calculated by summing the severity scores for each CAPS item per the standard scoring algorithm. Prepandemic AUD and depressive symptom severity were calculated by summing the scores (i.e., reflecting threshold, subthreshold, or negative ratings for each DSM-5 criterion) of the items from the SCID. If a SCID module was discontinued due to a lack of initial item endorsement per standard SCID administration guidelines, severity scores of 0 were assigned for unassessed items from that module. In follow-up analyses, we replaced these three severity scores with prepandemic PTSD, MDD, and AUD current diagnoses to see if significant results remained when diagnostic determinations were used as the predictors of COVID-19-related outcomes. Additional follow-up analyses were performed to determine if significant effects remained after controlling for proximal COVID-19 exposure and, separately, housing and financial stability, to evaluate if COVID-19-related stress resulting from illness among family and friends or housing instability better accounted for the effects attributed to psychopathology. There were no missing data in these analyses.
# Results
## Sample characteristics with respect to covid-19 exposure and impact
Sample characteristics with respect to COVID-19 exposure and impact are summarized in . All participants reported that they were able to socially distance beginning in March 2020. Over one quarter of the sample (27.0%) reported a worsened financial situation as a result of the pandemic, whereas 7.9% of participants reported increased housing instability due to the pandemic. A small proportion of individuals (2.0%) self-reported that they were diagnosed with COVID-19 by a health care professional, whereas 23.8% believed they had symptoms of COVID-19 but were not tested. Nearly one third of the sample (31.3%) reported that a close family member or friend was diagnosed with COVID-19 by a health care professional, and 25.0% of participants had a close family member or friend who thought they had COVID-19 symptoms but did not get tested. Over one quarter of the sample (26.7%) reported that a close family member or friend died as a result of COVID-19. Per EMR review, as of April 13, 2021, six (5.9%) participants tested positive for COVID-19, 28 (27.7%) tested negative, and 67 (66.3%) were not tested.
## Cfa of race items and bivariate correlations
The results of the CFA supported the use of the RACE Mood/Anxiety and PTSD subscales. The two-factor model fit the data well, χ 2 (43, n = 101) = 70.23, p = .005, RMSEA = .079, SRMR = .061, CFI = .973, TLI = .966. All items loaded significantly onto their respective latent variables, ps < .001; the standardized factor loadings for the mood/anxiety items ranged from β = .61 to β = .86, and the standardized factor loadings for the PTSD items ranged from β = .56 to β = .92. The two factors were highly correlated with each other, r = .90. Based on this high factor correlation, we compared the two-factor model to a more parsimonious single-factor model. We found that the more restrictive single-factor model was associated with significantly degraded fit compared to the two-factor model, Δχ 2 = 4.46, Δdf = 1, p = .035. Thus, the two-factor model was preferred. Correlations revealed expected relations among prepandemic psychiatric symptom severity and self-reported pandemic-related changes in psychiatric symptoms. Prepandemic PTSD severity was significantly associated with pandemic-related increases in symptoms of PTSD, r = .38, p < .001, and mood/anxiety, r = .33, p = .001. Prepandemic AUD severity was significantly associated with pandemic-related increases in PTSD symptoms, r = .32, p = .001; substance use, r = .47, p < .001; and mood/anxiety symptoms, r = .20, p = .047, as well as COVID-19 exposure, r = .27, p = .006. Prepandemic depressive symptom severity was significantly positively correlated with pandemic-related increases in PTSD symptoms, r = .26, p = .009 and mood/anxiety symptoms, r = .33, p = .001. In addition, a sense of reduced personal resilience during the pandemic, as indicated by high scores on the resilience item, was correlated with pandemic-related increases in mood/anxiety symptoms, r = .33, p = .001, and PTSD symptoms, r = .26, p = .009, although alterations in one's sense of personal resilience were not associated with any prepandemic variables. TA B L E 3 Descriptive characteristics and bivariate associations among prepandemic psychiatric conditions and pandemic-related psychiatric symptoms and experiences Note: Variables not identified as "prepandemic" were derived from the Rapid Assessment of COVID-19-Related Experiences (RACE), which was administered several months after the start of the pandemic. PTSD = posttraumatic stress disorder; AUD = alcohol use disorder; MDD = major depressive disorder. a Observed minimum and maximum. * p < .05; ** p < .01.
Additional correlations among the RACE subscales are shown in .
## Longitudinal multivariate regression models
The results of the regression models are summarized in . Significant predictors of pandemic-related changes in PTSD symptoms were prepandemic PTSD severity, β = .332, p = .003, and prepandemic AUD severity, β = .228; p = .028. Significant predictors of increased substance use during the pandemic included age, β = -.317, p = .002, and prepandemic AUD severity, β = .391; p < .001. Prepandemic AUD severity also significantly predicted self-reported or EMR-defined exposure to COVID-19, β = .264; p = .019. Minority race and ethnicity was the only significant predictor of housing and financial instability during the pandemic, β = −.372; p < .001. All effects remained significant (i.e., p < .05) after including both proximal COVID-19 exposure and housing and financial stability as covariates, in separate analyses, to account for the possible confounding effects of other COVID-19-related stressors. Models with significant effects for prepandemic psychiatric symptoms on pandemic out-comes were rerun using the prepandemic diagnostic variables in place of symptom severity scores. The pattern of results was unchanged with respect to significant effects (details available from the corresponding author). Sex, educational attainment, and prepandemic depression did not significantly predict psychiatric outcomes, COVID-19 exposure, or lifestyle stability during the pandemic.
# Discussion
Identifying individuals at heightened risk for adverse outcomes during the pandemic is critical for leveraging resources for those with the highest level of need. In a sample of veterans who had experienced a range of trauma types, including combat and sexual and physical assault, we found that prepandemic mental health conditions, especially PTSD and AUD, were associated with psychological and health responses to the pandemic. In particular, our results demonstrated that baseline AUD predicted later exposure to the COVID-19 virus, which is consistent with recent literature. Given the association between SUDs and both increased risk-taking propensity (LaSpada et al., 2020) and reduced harm avoidance [bib_ref] Multidimensional Personality Questionnaire profiles of veterans with traumatic combat exposure: Externalizing and..., Miller [/bib_ref] , individuals with SUDs may be less Note: Demographic characteristics were included in Step 1 and psychiatric variables in
Step 2 of each model. Bolded values are statistically significant. PTSD = posttraumatic stress disorder; AUD = alcohol use disorder; MDD = major depressive disorder. a Overall model fit statistics for model predicting pandemic-related PTSD:
Step 1: R 2 = .034, F(4, 95) = 1.873, p = .121;
Step 2: ∆R 2 = .172, ∆F(3, 92) = 6.993, ∆p < .001;
Overall fit statistics for model predicting pandemic-related mood/anxiety:
Step 1: R 2 = .035, F (4, 95) = .866, p = .487;
Step 2: ∆ R 2 = .133, ∆ F (3, 92) = 4.889, ∆ p = .003; Overall fit statistics for model predicting pandemic-related substance use:
Step 1: R 2 = .139, F(4, 95) = 3.844, p = .006;
Step 2: ∆R 2 = .138, ∆F (3, 92) = 5.870, ∆p = .001; Overall fit statistics for model predicting COVID-19 self-exposure:
Step 1: R 2 = .046, F(4, 95) = 1.140, p = .342;
Step 2: ∆R 2 = .067, ∆F(3, 92) = 2.308, ∆p = .082; Overall fit statistics for model predicting pandemic-related housing/financial stability:
Step 1: R 2 = .146, F(4, 95) = 4.052, p = .004;
Step 2: ∆R 2 = .001, ∆F (3, 92) = .033, ∆p = .992.
able to accurately judge risks related to physical health and safety, especially while they are under the influence. Furthermore, individuals with SUDs may be using substances in social settings, which could increase their risk of COVID-19 exposure. SUDs also impact immunological responses [bib_ref] Substance use disorders: Psychoneuroimmunological mechanisms and new targets for therapy, Loftis [/bib_ref] , which may play a role in the increased risk of COVID-19 diagnosis and related symptoms. Prepandemic AUD was also associated with pandemicrelated increases in PTSD symptoms, which carries implications for understanding PTSD comorbidity and its associations with broad, underlying dimensions representing internalizing (e.g., unipolar mood disorders, anxiety, somatization disorders) and externalizing (e.g., SUDs, antisocial personality disorder) psychopathology [bib_ref] The structure of common DSM-IV and ICD-10 mental disorders in the Australian..., Slade [/bib_ref]. PTSD is often thought of as an internalizing disorder [bib_ref] The structure of common DSM-IV and ICD-10 mental disorders in the Australian..., Slade [/bib_ref]. However, there is evidence that PTSD may arise through genetic liability to either internalizing or externalizing psychopathology [bib_ref] Posttraumatic stress disorder and the genetic structure of comorbidity, Wolf [/bib_ref] , and research has demonstrated phenotypic associations between PTSD and both psychopathology spectra [bib_ref] Posttraumatic stress disorder in DSM-5: New criteria and controversies, Miller [/bib_ref] [bib_ref] Posttraumatic stress disorder and the genetic structure of comorbidity, Wolf [/bib_ref]. The results of our longitudinal research further suggest that externalizing conditions, such as SUDs, may be associated with worsening PTSD symptom severity after controlling for baseline PTSD severity. Thus, although PTSD is more strongly associated with internalizing disorders and, in the present study, shared more variance in common with internalizing versus externalizing symptoms, the externalizing presentation may be a marker for a particularly unique and problematic PTSD symptom course. The importance of this is further highlighted by the present finding that AUD was associated with an increased risk of COVID-19 exposure over time.
In addition, our results demonstrate an increase in preexisting psychiatric symptoms during the pandemic. Prepandemic PTSD predicted pandemic-related increases in PTSD symptoms, which is consistent with prior literature concerning PTSD and psychiatric responses to the pandemic [bib_ref] Risk factors for depression, anxiety, and PTSD symptoms in perinatal women during..., Liu [/bib_ref] [bib_ref] Exploring the impact of COVID-19 and restrictions to daily living as a..., Murphy [/bib_ref]. This finding suggests that the COVID-19 pandemic, an acuteonset stressor, exacerbates PTSD symptoms related to prior traumatic experiences. One possibility is that behaviors that are important for reducing viral exposure during the pandemic, such as quarantining, social distancing, and remaining vigilant about masking, may reinforce existing PTSD symptoms, such as avoidance, estrangement from others, and hypervigilance. In addition, participants with prepandemic AUD were more likely to report increased substance use during the pandemic. Thus, individuals with PTSD and AUD are at a higher risk for poorer pandemicrelated psychiatric outcomes, suggesting targeted treatment efforts among these individuals may be particularly useful in mitigating the long-term consequences of the COVID-19 pandemic.
We found that 5.9% of the sample had tested positive for COVID-19, which was lower than the cumulative prevalence of the disease in the northeast region of the United States, where the study was conducted, at the time of study completion. [bib_ref] Association of PTSD with COVID-19 testing and infection in the Veterans Health..., Haderlein [/bib_ref] found that veterans with PTSD were less likely to test positive for COVID-19 than veterans without PTSD. The authors suggested that veterans with PTSD were already more socially isolated before the pandemic, potentially resulting in lower infection rates in this population [bib_ref] Association of PTSD with COVID-19 testing and infection in the Veterans Health..., Haderlein [/bib_ref]. Veterans may also have military training that lends itself to increased hypervigilance and preparedness (e.g., isolating and stocking resources for long periods). Thus, the low rates of infection within this clinical veteran sample could suggest that the characteristics of PTSD (e.g., avoidance, heightened perceived sense of threat) associated with an increased risk of worsening mental health symptoms are protective with respect to avoiding exposure to the virus. This requires careful clinical consideration in how to prevent COVID-19 exposure among this population without reinforcing PTSD symptoms.
An alarming proportion of participants (26.7%) reported that a close family member or friend died from COVID-19. [bib_ref] Exploring the impact of COVID-19 and restrictions to daily living as a..., Murphy [/bib_ref] also reported that a high proportion (15.1%) of individuals in their veteran sample knew someone who had died from COVID-19. Given that the current assessment was administered during the summer of 2020, relatively early in the pandemic, it is particularly troubling that such a high proportion of the present sample had experienced a COVID-19-related loss. These losses may meet the definition of a DSM-5defined traumatic experience, potentially contributing to psychiatric symptoms in response to new trauma exposure. Another concerning trend in the data was that racial and ethnic minorities, a variable included in all analyses along with age, educational attainment, and sex, were more likely to lose their housing as a result of the pandemic. This is consistent with known race-related health disparities [bib_ref] Racial and ethnic health disparities related to COVID-19, Lopez [/bib_ref] [bib_ref] The disproportionate impact of COVID-19 on racial and ethnic minorities in the..., Tai [/bib_ref] and the disproportionate rise in unemployment [bib_ref] Early evidence of the impacts of COVID-19 on minority unemployment, Couch [/bib_ref] among minority individuals during the pandemic. Notably, the factors that were associated with changes in psychiatric symptoms during the pandemic were distinct from those associated with housing changes. This finding highlights the growing need for programs designed to prevent housing loss among veterans, such as the U.S. Department of Housing and Urban Development-VA Supportive Housing (HUD-VASH) Program.
These results should be interpreted in consideration of the study's strengths and limitations. The main strengths of the study include its longitudinal design, the use of clinician-administered diagnostic interviews, and the breadth of assessment of both prepandemic and pandemic-related mental health. Limitations include the small sample size and the predominantly male, all-veteran sample, which limits the generalizability of the results. An additional limitation is that our pandemic assessment was based on a newly derived measurement self-report tool, the RACE, although this initial examination of its psychometric properties, including internal consistency and the results of the CFA, supports its use. The RACE was administered relatively early in the pandemic, before the spike of COVID-19 cases in the United States in late 2020 and early 2021, and, thus, an additional limitation is that we may not have fully captured veterans' responses to the worst of the pandemic to date. Additionally, the current version of the RACE does not assess mask usage, and this would be a useful addition to future revisions to the measure. Another limitation of the measure is that there are strong demand characteristics when inquiring about social distancing practices; thus, respondents may have been reluctant to report that they did not follow these guidelines. Furthermore, due to the small number of participants who either self-reported a COVID-19 diagnosis or had a positive test result in the EMR (5.9%), we considered both participants who were diagnosed with COVID-19 and those who believed they had COVID-19 symptoms but were not tested as having been exposed to COVID-19. It is important to note that the phone assessment was conducted in the initial 6 months of the pandemic, when testing was not easily accessible. In addition, we did not administer the CAPS-5 or SCID over the phone, so our assessment of pandemic-related symptom change was limited by the self-report nature of this evaluation. We also did not assess mental health treatment at the time of the follow-up, so we were not able to control for intervention effects. We did not assess if participants' exposure to COVID-19-related loss or illness met the DSM-5 definition of a traumatic experience or if individuals had PTSD symptoms specific to this potentially traumatic event. In addition, we were underpowered to evaluate potential demographic or other moderators of the associations of interest, so it is unclear if associations between prepandemic psychiatric symptoms and pandemic-related outcomes might differ by participant characteristics. Finally, although the study benefited from a longitudinal design, we cannot determine causal associations among the data given the potential for unmeasured, confounding variables.
Overall, the present results demonstrate that individuals with preexisting conditions, particularly PTSD and AUD, are uniquely affected by the pandemic and are at heightened risk for both adverse psychiatric outcomes and exposure to the virus. These findings carry implications for targeting prevention and treatment efforts for these highrisk individuals. For instance, brief PTSD and AUD assessments administered during routine primary care visits, followed by the provision of brief psychoeducation related to drinking guidelines and risks factors for COVID exposure or treatment referrals may have clinically significant downstream effects. Just as resources are leveraged for individuals with an increased risk of poor outcomes from the virus itself (i.e., the elderly, individuals with preexisting cardiac conditions), targeted prevention and intervention efforts should be made for those with the highest risk of adverse psychiatric outcomes.
## O p e n p r a c t i c e s s tat e m e n t
The study reported in this article was not formally preregistered. Neither the data nor the materials have been made available on a permanent third-party archive; qualified investigators may apply for access to deidentified data via a data repository that we maintain by sending an email to the corresponding author at [email protected]. |
Discovery of N-(benzo[1,2,3]triazol-1-yl)-N-(benzyl)acetamido)phenyl) carboxamides as severe acute respiratory syndrome coronavirus (SARS-CoV) 3CLpro inhibitors: Identification of ML300 and noncovalent nanomolar inhibitors with an induced-fit binding
## A b s t r a c t
Herein we report the discovery and SAR of a novel series of SARS-CoV 3CLpro inhibitors identified through the NIH Molecular Libraries Probe Production Centers Network (MLPCN). In addition to ML188, ML300 represents the second probe declared for 3CLpro from this collaborative effort. The Xray structure of SARS-CoV 3CLpro bound with a ML300 analog highlights a unique induced-fit reorganization of the S 2 -S 4 binding pockets leading to the first sub-micromolar noncovalent 3CLpro inhibitors retaining a single amide bond.
Ó 2013 Elsevier Ltd. All rights reserved.
Coronaviruses (CoV) are enveloped, large plus-strand RNA viruses associated with mild to severe respiratory symptoms, including the common cold and the Severe Acquired Respiratory Syndrome (SARS)-CoV.Identified as the etiological agent responsible for the global pandemic in 2003, SARS presents an atypical pneumonia that during the first major outbreak led to progressive respiratory failure in over 8000 individuals and about 800 deaths by July of that year.With the cooperation of leading nations, a rigorous public healthcare campaign was fortunately successful in controlling this outbreak. However, a reemergence of the SARS-CoV is considered a potential pandemic risk and new strains of human coronavirus continue to be identified. Since 2003, two additional human coronaviruses, NL63 and HKU1, have been identified in patients and the viruses have been characterized and found to be significantly less lethal than SARS-CoV.Most recently in 2012, a new SARS-like virus, designated the Middle East respiratory syndrome coronavirus (MERS-CoV), has been identified in 144 patients so far, 54 of whom died.There is now evidence for person-to-person transmission of MERS-CoV. Now, nearly a decade later, the possibility of another SARS-like pandemic appears even more palpable based upon the lethality and properties of the newly identified MEV-HCoV strain. Effective vaccines and small molecule antiviral agents to prevent or treat SARS-like infections still do not exist, thus tailored antiviral therapies are urgently needed in order to treat potential future outbreaks of SARs and related human coronaviruses.
The SARS and MERS coronaviruses encode two proteases, a papain-like protease (PLpro) and a 3-chymotrypsin-like protease (3CLpro), in their genome that are essential for viral replication. The viral polyprotein is cleaved at three unique sites by PLpro and 11 unique sites by 3CLpro. Initial reports of 3CLpro inhibitors in the literature focused on peptidomimetics, often four to five residues in length, bearing a reactive 'warhead' group, such as an aldehyde, halo-methyl ketone, or Michael acceptor at the terminus with several demonstrating a covalent interaction with the active site Cys-145 residue.Until recently, the majority of efforts to develop nonpeptidic 3CLpro inhibitors also relied on 'warhead' based design strategies [fig_ref] Scheme 1: Synthesis of P 1 analogs 9a-c and 10a-c [/fig_ref] 17-21 and a number of these nonpeptidic inhibitors achieved sub-micromolar activity. In the case of pyridyl ester 4, 20 this potent nanomolar mechanism-based enzyme inactivator led to cell based inhibition below 10 lM in SARS-CoV infected Vero E6 cells. Recently, we reported N-(tert-butyl)-2-(N-arylamido)-2-(pyridin-3-yl) acetamide 6 [fig_ref] Scheme 1: Synthesis of P 1 analogs 9a-c and 10a-c [/fig_ref] and its X-ray complex with 3CLpro (PDB: 3V3M) as a rare example of a noncovalent SARS-CoV 3CLpro inhibitor of moderate molecular weight with good enzyme and antiviral inhibitory activity.Herein, we describe the continuation of efforts to develop potent, noncovalent SARS-3CLpro inhibitors based upon a second chemical class of triazoles from our MLPCN screening campaign (7, and progression of this lead series to a second generation probe ML300 (16e, [fig_ref] Scheme 1: Synthesis of P 1 analogs 9a-c and 10a-c [/fig_ref] and beyond to arrive at sub-100 nM inhibitors. We propose from crystallography data that ML300 and related triazoles in this series inhibit 3CLpro via a novel mechanism of action and provide a new direction for additional noncovalent inhibitor design and refinement.
Using a designed expression construct which produces the post-proteolytic and authentic 3CLpro dimer, a screen against the NIH molecular libraries sample collection (293 K compounds) at the Scripps Research Institute Molecular Screening Center (SRIMSC) was undertaken. In addition to the diamide acetamide series represented by ML188 (6, [fig_ref] Scheme 1: Synthesis of P 1 analogs 9a-c and 10a-c [/fig_ref] ,a related diamide series, represented by SID 24808289 (7, , was identified demonstrating a 3CLpro IC 50 of 6.2 lM and good selectivity versus PLpro (IC 50 > 60 lM) which is used as a control for cysteine-protease activity. Fortunately, quite early in the chemistry campaign an Xray crystal structure of diamide 7 bound to 3CLpro was determined to 1.85 Å resolution. A solvent accessible surface depiction of 7 in the 3CLpro active site along with a wall-eye stereo view with key contact resides and hydrogen bonding contacts in depicted in [fig_ref] .Scheme 3: Synthesis of P 2 -P 1 analogs 12 and 13a-l [/fig_ref]. Interestingly, in contrast to the ML188-3CLpro crystal structure in which ML188 accommodates substrate sub-pockets in the enzyme active-site traditionally occupied by peptidomimetics, diamide 7 engenders an induced-fit complex resulting in a new surface dictated largely by a rearrangement of the Gln-189 and Met-49 residue side-chains. [bib_ref] For an example of 'induced-fit' and discussion on active-site flexibility in the..., Lee [/bib_ref] This induced fit accommodates the syn N-methyl pyrrole and anilido acetamide moieties of the inhibitor within subpockets that can be characterized as S 2 -S 4 and S 2 -S 1 0 subpockets, respectively. schematically illustrates the inhibitor-active site interactions oriented in a manner similar as depicted in [fig_ref] .Scheme 3: Synthesis of P 2 -P 1 analogs 12 and 13a-l [/fig_ref]. In addition to the P 2 -P 4 and P 2 -P 1 0 groups the inhibitor partially occupies the S 3 subpocket with a terminating 2-methylbutylamide. Key hydrogen bonding interactions can be found near the catalytic site with His-163 and the benzotriazole N-(3) engaged in a key interaction, with an interatomic distance of 2.9 Å. In addition a backbone Glu-166 NH interaction is evident with the central acetamide oxygen (N-O distance 2.8 Å).
Flexibility of the diamide scaffold (RotBon 7) coupled with the observed induced-fit within the active site of 3CLpro presents an added challenge with respect to in silico inhibitor approaches. Thus, our structure-activity-relationship (SAR) studies focused initially on three key areas within the diamide scaffold: (1) benzotriazole replacements with alternative hydrogen bond acceptor functionality to interact with His-163, (2) acetamide modifications within the P 2 -P 1 0 region, and (3) minimum pharmacophore deletion studies of the P 3 2-methylbutylmide. The P 2 -P 4 group was held constant for this investigation and based upon HTS and reconfirmation results (data not shown) the N-methyl pyrrole was replaced with an equipotent 3-thienyl moiety.
In parallel with efforts to obtain the 3CLpro-7 crystal structure, synthesis of first generation analogs to survey diversity of the benzotriazole unit were initiated using a modified version of our 4CC-Ugi strategy (Scheme 1) to allow for late stage azole introduction. Thus, Ugi reaction using t-butyl isocyanide, chloroacetic acid, thiophene-3-carbaldehyde, and N-(4-aminophenyl)acetamide proceeded smoothly to give chloride 8, which could be isolated in good yield after chromatography. Displacement of chloride 8 with azole NH heterocycles provided 9a-c. Alternatively, displacement of 8 with sodium azide and subsequent Huisgen cycloaddition reaction with an appropriate acetylene furnished 1,2,3-triazoles 10a-c in good overall yield.
Synthesis of P 2 -P 1 0 amide analogs within the elaborated diamide were similarly prepared in an Ugi reaction using Boc-protected 4-(amino) aniline (Scheme 2) as the amine component. Deprotection of 11 using trifluoroacetic gave aniline 12 which was reacted with a variety of carboxylic acid derivatives under HATU coupling conditions or reacted with an acid chloride or sufonyl chloride in the presence of TEA to give final examples 13a-l.
Synthesis of P 3 truncated analogs began with reductive amination using thiophene-3-carbaldehyde with either 4-bromoaniline or Boc-protected 4-(amino) aniline to give intermediates 14a-b in good yield. Amide coupling with HATU using benzotriazol-1yl-acetic acid installed the requisite P 1 groups to afford 15a-b. Initial efforts focused on preparing the identical amide library prepared in the elaborated series 13 (Scheme 3). This was readily accomplished as before; Boc-deprotection of 15a followed by amide coupling or acylation/sulfonylation, gave 16a-k. Subsequent synthesis of a series of biaryls as amide replacements commenced using a Suzuki cross-coupling with 15b and a variety of boronic acids to afford target molecules 17a-e. SAR for 1,3-azole P 1 replacements (9a-c, [fig_ref] Figure 4: Representative azole replacements [/fig_ref] indicated a strict requirement for the 1,2,3-triazole unit; benzimidazoles 9a-b and 2-methyl-1-imidozyl derivative 9c were uniformly inactive. Since the N-(3) nitrogen of 7 appeared to be involved in a hydrogen bond with His-163, it was somewhat surprising that 9a was not tolerated since the imidazole has the potential to maintain a N-(3)-His-163 hydrogen bond interaction. However, within the 3CLpro-7 structure the catalytic Cys-145 residue is located within 3.3 Å of the N-(2) nitrogen, indicating potential for a weak hydrogen bond and/or dipole-dipole stabilization interaction. This potential interaction may perhaps be responsible for the 1,2,3-triazole preference. Interestingly, 4-phenyl 1,2,3-triazole 10c was tolerated with an IC 50 of 11 lM, suggesting additional avenues for optimization. Accommodation of the phenyl moiety of 10c within the active site S 1 subpocket is not entirely clear at this time. Based on the 3CLpro-7 structure, Glu-166, Phe-140, and Glu-166 are predicted to be within close proximity. Unsubstituted triazole 10a and trimethyl silyl triazole 10b were inactive, demonstrating the importance of maintaining a proper aromatic ring in this subpocket.
Amide library 13a-l [fig_ref] Figure 5 50: are the average of three independent determinations and represent a co-efficient... [/fig_ref] within the elaborated diamide series displayed a range of potency from moderate micromolar activity (13a, 13b-d, 13f-g) comparable to the HTS hit 7, to weakly active or inactive. Cyclic and acyclic acetamide congeners related to HTS lead 7 showed consistent activities below 10 lM with branched i-propyl derivative 13d and cyclobutyl amide 13g having the greatest activity below 5 lM. Modification to sulfonamide 13b resulted in a three-fold loss in inhibition relative to acetamide 13a. The smaller cyclopropyl (13f) or larger cyclohexyl (13h) cyclic amide generally resulted in loss of inhibition. Incorporating a sterically hindered t-butyl amide 13e also led to a modest three-fold loss in activity. Lastly, aromatic and heteroaromatic amides (13ik) in addition to iso-butyl carbamate 13l were weak or inactive as 3CLpro inhibitors. Collectively, these data appear to be consistent with the 3CLpro-7 structure whereby a short helix-loop-helix motif (Val-42-Ile-43-Cys-44-Thr-45-Ala-46) and a proximal b-turn element (Thr-24-Thr-25) define the edge of this pocket with minimal volume for larger groups beyond acetamide 7.
With limited success from the above S 1 and S 2 -S 1 0 studies we turned to P 3 truncation to potentially identify a minimum pharmacophore to reduce overall MW and improve ligand efficiency (LE).Examination of the P 3 group within the 7-3CLpro structure suggested this group was unfavorably solvent exposed relative to the t-butylamide-S 3 interaction found within the ML188-3CLpro structure.Initial efforts led to 16a-k [fig_ref] Table 1: 3CLpro activity 16a-k, 17a-e [/fig_ref]. Gratifyingly truncated amides proved to have comparable activity versus the elaborated diamide counterparts [fig_ref] Scheme 1: Synthesis of P 1 analogs 9a-c and 10a-c [/fig_ref] -g, vs [fig_ref] Table 1: 3CLpro activity 16a-k, 17a-e [/fig_ref] see 16a-c, 16e-f). Interestingly, truncated series 16 appeared to better tolerate larger substituents, perhaps suggesting additional changes in the shape of the active site within this subpocket. For example cyclohexyl amide 16g was found to be a weak inhibitor and similarly carbamate 16i had moderate inhibitory activity of 10.3 lM while its related counterpart 13l [fig_ref] Figure 5 50: are the average of three independent determinations and represent a co-efficient... [/fig_ref] was inactive.
At this stage in the project with efforts focused on P 3 truncated analogs bearing a putative S 2 -S 1 0 interaction, we elected 16e for further characterization and probe declaration (ML300, [fig_ref] Figure 6: Profiles of 3CLpro inhibitors 6 [/fig_ref]. Relative to probe ML188 (6) and the equipotent diamide 13d, ML300 proved to offer improvements in several areas [fig_ref] Figure 6: Profiles of 3CLpro inhibitors 6 [/fig_ref]. SARS 3CLpro inhibitor ML300 is 100 amu lower MW (MW = 431) relative to 13d with moderate ligand efficiency (LE). Deletion of the lipophilic P 3 group reduces c Log P an order of magnitude (c Log P = 3.2) and thus greatly improves ligand-efficiency-dependent lipophilicity (LELP) versus ML188 and 13d. Probe molecules ML188 and ML300 were evaluated in an in-housein vitro DMPK panel including plasma protein binding, P450 enzyme inhibition, and intrinsic clearance using liver microsomes [fig_ref] Figure 6: Profiles of 3CLpro inhibitors 6 [/fig_ref]. Both ML188 and ML300 possess good free fraction with ML300 being superior (1.5 and 4.0-fold improved human and rat fraction unbound, respectively); however, intrinsic clearance (CL HEP normalized to liver blood flow, Q h = 21 mL/min/kg, Q r = 70 mL/min/kg) indicates ML188 and both ML300 are both predicted to be highly cleared. ML188 and ML300 possess modest P450 enzyme inhibition, with ML300 maintaining 5-10 lM activity across four major CYP enzymes [fig_ref] Figure 6: Profiles of 3CLpro inhibitors 6 [/fig_ref]. Probe ML300 was found to be highly selective in a Eurofins lead-profiling screenwith only modest activity for melatonin MT 1 receptor in a radioligand binding assay.
Based on the absence of key hydrogen bonding interactions of the P 2 -P 1 0 amide of 7 with the 3CLpro active site, in addition to the poor metabolic instability and CYP profile of ML300, we opted to explore more diverse amide replacements as a means to improve metabolic stability, P450 activity, and 3CLpro inhibitory potency. Initial efforts identified representative N-methyl (16j) and N-benzyl (16k) anilines with potency comparable to probe ML300 (16e). The lack of activity for benzamide 16h versus the reduced benzylamine 16k is striking and indicates the enhanced flexibility of the N-benzyl group is permitting a productive interaction to occur where previously aromatic amides were not tolerated (see [fig_ref] Figure 5 50: are the average of three independent determinations and represent a co-efficient... [/fig_ref] , 13i-j). A subsequent survey of biaryls was explored [fig_ref] Table 1: 3CLpro activity 16a-k, 17a-e [/fig_ref] , 17a-e) and on the basis of the 3CLpro-7 X-ray, we targeted 3-pyridyl (17b-c) and 4-pyrimidyl (17e) heterocycles as means to potentially engage a side-chain interaction from the hydroxyl groups of Thr-24 or Thr-25. These modifications afforded inhibitors with micromolar activity and in the case of 2-methoxypyridyl 17c submicromolar activity (IC 50 = 700 nM). Unexpectedly the parent simple phenyl biaryl 17a proved to have a major impact on activity with a 7-10-fold increase relative to 17b-c. 3CLpro inhibitor 17a represented the first sub-100 nM inhibitor for the series and to our knowledge one of the most potent nonwarhead based SARS 3CLpro inhibitors to date. At this time inhibitor 17a is relatively unoptimized and thus current efforts are focused on targeted biaryl congeners to understand DMPK, cellular activity, as well as potential broad spectrum activity against other coronavirus strains including MERS-CoV.
In summary, we have described the identification and binding orientation and interactions for a second class of diamide SARS 3CLpro inhibitors, culminating in probe ML300 and subsequently improved inhibitors such as 17a, which possess LE > 0.3 and an LELP approaching 10. The X-ray crystal structure of HTS hit 7 bound to 3CLpro 27 was instrumental in guiding optimization and the induced-fit of this inhibitor 3CLpro complex illustrates the challenges of divergent SAR and the limitations of virtual based screens. The four component Ugi reaction was utilized once more to rapidly generate SAR for the putative P 2 -P 1 0 and P 1 subgroups. Importantly, P 3 truncation was possible for this triazole series of 3CLpro inhibitors, allowing for significant MW reduction without diminishing potency. Collaborative efforts in these laboratories continue towards the identification active inhibitors within the truncated biaryl class. Integrated efforts with DMPK assessment continue in order to improve intrinsic clearance and diminish P450 activity, which are issues to be addressed within the series prior to in vivo proof-of-mechanism studies. ML300 is an MLPCN probe and is freely available upon request.
[fig] ,Figure 1, Figure 2, Figure 3: ML188 (2013)22 IC 50 = 1.5 µM SARS-CoV EC 50 = 12.9 µM pdb code: 3V3M Representative nonpeptidic 3CLpro inhibitors utilizing a warhead and noncovalent mechanism of inhibition (Binding orientation and properties of MLPCN 3CLpro HTS hit 7 and evolved probe molecule ML300(16e). (A) Solvent accessible surface view of 7-3CLpro complex (PDB code: 4MDS, PubChem SID 24808289); (B) X-ray crystal structure of 7 (capped sticks in orange carbon) with SARS 3CLpro in wall-eye stereo view with key residues and hydrogen bonds. [/fig]
[fig] Scheme 1: Synthesis of P 1 analogs 9a-c and 10a-c. Reagents and conditions: (a) MeOH, 50°C, 4 h, 95%, (b) (i) NaH, HetNH, DMF, (ii) 9, DMF, 65-80%, (c) (i) NaN 3 , DMF, 100°C lwave 30 min, 95%, (ii) acetylene (R = Ph, TMS), DCE, 120°C 16 h, 85-98%, (iii) R = TMS, TBAF, HOAc, 0°C-rt, 45%. Final library compounds were purified by UV prep or mass-directed prep HPLC. [/fig]
[fig] .Scheme 3: Synthesis of P 2 -P 1 analogs 12 and 13a-l. Reagents and conditions: (a) MeOH, 50°C, 4 h, 94%, (b) TFA, 95% (c) (i) HATU, DIPEA, DMF, RCO 2 H, 55-73% (d) RCO 2 Cl or RSO 2 Cl, TEA, DCM, 51-64%. Final library compounds were purified by UV prep or mass-directed prep HPLC. Synthesis of P 2 -P 1 0 analogs 16a-k and 17a-e within truncated series. Reagents and conditions: (a) NaHB(OAc) 3 , DCE, rt, 80% (b) benzotriazol-1-yl-acetic acid, HATU, TEA, DMF, rt, 74% (c) (i) 15a, TFA, DCM, 95%, (ii) HATU, DIPEA, DMF, RCO 2 H, 65-80%; RCO 2 Cl or RSO 2 Cl, TEA, DCM, 90-95%; NaHB(OAc) 3 , RCHO, DCE, 45-95% (d) 15b, Ar/HetB(OH) 2 , 1 M aq Na 2 CO 3 , 5 mol % Pd(PPh 3 ) 4 , THF, 30-65%. Final library compounds were purified by UV prep or mass-directed prep HPLC. [/fig]
[fig] Figure 4: Representative azole replacements (9a-c and 10a-c). [/fig]
[fig] Figure 5 50: are the average of three independent determinations and represent a co-efficient of variation (CV) < 0.10. P2-P1' Exploration: amide, urea, sulfonamide 3CLpro activity from library 13. [/fig]
[fig] Figure 6: Profiles of 3CLpro inhibitors 6 (ML188), 13d, and 163 (ML300). [/fig]
[table] Table 1: 3CLpro activity 16a-k, 17a-e [/table]
|
Genome-Wide Association Study of Relative Telomere Length
Telomere function is essential to maintaining the physical integrity of linear chromosomes and healthy human aging. The probability of forming proper telomere structures depends on the length of the telomeric DNA tract. We attempted to identify common genetic variants associated with log relative telomere length using genome-wide genotyping data on 3,554 individuals from the Nurses' Health Study and the Prostate, Lung, Colorectal, and Ovarian Cancer Screening Trial that took part in the National Cancer Institute Cancer Genetic Markers of Susceptibility initiative for breast and prostate cancer. After genotyping 64 independent SNPs selected for replication in additional Nurses' Health Study and Women's Genome Health Study participants, we did not identify genome-wide significant loci; however, we replicated the inverse association of log relative telomere length with the minor allele variant [C] of rs16847897 at the TERC locus (per allele b = 20.03, P = 0.003) identified by a previous genome-wide association study. We did not find evidence for an association with variants at the OBFC1 locus or other loci reported to be associated with telomere length. With this sample size we had .80% power to detect b estimates as small as 60.10 for SNPs with minor allele frequencies of $0.15 at genome-wide significance. However, power is greatly reduced for b estimates smaller than 60.10, such as those for variants at the TERC locus. In general, common genetic variants associated with telomere length homeostasis have been difficult to detect. Potential biological and technical issues are discussed.
# Introduction
Two major challenges are encountered in maintaining the physical integrity of linear chromosomes. First, cells must prevent the recognition of chromosome ends as double-stranded DNA breaks [bib_ref] Functional links between telomeres and proteins of the DNA-damage response, Di Fagagna [/bib_ref]. Second, telomeric ends are lost with each cell division as a result of the end replication problem [bib_ref] A theory of marginotomy. The incomplete copying of template margin in enzymic..., Olovnikov [/bib_ref]. On average, human telomeres lose 50 to 100 bp per mitotic division, which limits the replicative capacity of a cell [bib_ref] Telomere length predicts replicative capacity of human fibroblasts, Allsopp [/bib_ref]. To overcome these issues, eukaryotic cells evolved the telomere maintenance system, the almost universal mechanism used to protect chromosome ends [bib_ref] T-loops and the origin of telomeres, De Lange [/bib_ref].
Telomeres are complex dynamic nucleoprotein structures that consist of a long stretch of hexameric (TTAGGG) n DNA repeats, a single-stranded G-rich 39 overhang, the telomerase enzyme complex, six telomere-specific proteins, and other telomere-associated proteins including those of the DNA repair machinery [bib_ref] Shelterin: the protein complex that shapes and safeguards human telomeres, De Lange [/bib_ref]. In normal mitotic cells, telomeres switch between capped and uncapped states [bib_ref] Telomere states and cell fates, Blackburn [/bib_ref]. Telomere-associated proteins assist the singlestranded G-rich 39 overhang in invading the double-stranded telomeric tract to form a large terminal loop. This conformation contributes to the functional capped structure of the telomeric end, inhibiting the DNA damage response. Telomerase, a highly conserved enzyme consisting of protein [telomerase reverse transcriptase (TERT), dyskerin (DKC1)] and RNA [telomerase RNA component (TERC)] components [bib_ref] Protein composition of catalytically active human telomerase from immortal cells, Cohen [/bib_ref] , appends chromosome ends with hexameric repeats to restore telomere length [bib_ref] Regulation of telomerase by telomeric proteins, Smogorzewska [/bib_ref] and plays a role in stabilizing telomeres in the capped state [bib_ref] Telomere states and cell fates, Blackburn [/bib_ref].
The likelihood of a telomere existing in an uncapped vs. a capped state is dependent on telomere length. Longer telomeres are more likely to form the protective capped structure than shorter telomeres [bib_ref] Telomere states and cell fates, Blackburn [/bib_ref]. In germ cells, strong telomerase expression maintains long telomere lengths [bib_ref] Specific association of human telomerase activity with immortal cells and cancer, Kim [/bib_ref]. Telomerase activity has been detected at lower levels in fibroblasts, peripheral blood leukocytes, skin, hair follicles, intestinal crypts, and endometrium. In these tissues, expression levels are not sufficient to prevent replication-associated telomere attrition [bib_ref] Telomerase maintains telomere structure in normal human cells, Masutomi [/bib_ref] [bib_ref] Telomerase activity in normal and malignant hematopoietic cells, Broccoli [/bib_ref] [bib_ref] Telomerase activity in normal leukocytes and in hematologic malignancies, Counter [/bib_ref] [bib_ref] Telomerase activity in human endometrium, Kyo [/bib_ref] [bib_ref] Telomerase activity in normal human epithelial cells, Yasumoto [/bib_ref]. As a result, telomere length declines with age [bib_ref] Telomere reduction in human colorectal carcinoma and with ageing, Hastie [/bib_ref] [bib_ref] Genetic determination of telomere size in humans: a twin study of three..., Slagboom [/bib_ref] [bib_ref] Comparison of chromosome telomere integrity in multiple tissues from subjects at different..., Butler [/bib_ref] [bib_ref] The rate of telomere sequence loss in human leukocytes varies with age, Frenck [/bib_ref] [bib_ref] Telomere length in different tissues of elderly patients, Friedrich [/bib_ref] [bib_ref] Telomere lengths are characteristic in each human individual, Takubo [/bib_ref]. As telomeres shorten and become dysfunctional, the uncapped chromosomal ends undergo nucleolytic decay, chromosomal end-to-end fusions, and atypical recombination triggering senescence or apoptosis [bib_ref] Telomeres: cancer to human aging, Stewart [/bib_ref].
Diseases characterized by telomere dysfunction highlight the importance of telomere maintenance for healthy human aging. Patients with dyskeratosis congenita, a rare inherited bone marrow failure and cancer predisposition syndrome, have very short germline telomeres (,1 st percentile for age), and approximately one-half have a mutation in a known telomere biology gene [bib_ref] Dyskeratosis congenita, Savage [/bib_ref]. Patients with the shortest telomeres exhibit the most severe phenotypes [bib_ref] Mutations in dyskeratosis congenita: their impact on telomere length and the diversity..., Vulliamy [/bib_ref]. Telomere attrition, genomic instability, and premature senescence are also features of Hutchinson-Gilford progeria syndrome and Werner syndrome [bib_ref] Progeria syndromes and ageing: what is the connection?, Burtner [/bib_ref]. Shorter leukocyte telomere length has been implicated in a number of common aging-related diseases such as cancer [bib_ref] Telomere dysfunction: a potential cancer predisposition factor, Wu [/bib_ref] [bib_ref] Telomere length, cigarette smoking, and bladder cancer risk in men and women, Mcgrath [/bib_ref] [bib_ref] Genetic variation in telomere maintenance genes, telomere length, and lung cancer susceptibility, Hosgood [/bib_ref] [bib_ref] Telomere length and the risk of lung cancer, Jang [/bib_ref] [bib_ref] Leukocyte telomere length in a population-based case-control study of ovarian cancer: a..., Mirabello [/bib_ref] and cardiovascular disease [bib_ref] Telomere shortening in atherosclerosis, Samani [/bib_ref] [bib_ref] White cell telomere length and risk of premature myocardial infarction, Brouilette [/bib_ref] [bib_ref] Leukocyte telomere length and cardiovascular disease in the cardiovascular health study, Fitzpatrick [/bib_ref] [bib_ref] Telomere length, risk of coronary heart disease, and statin treatment in the..., Brouilette [/bib_ref] [bib_ref] Association of shorter mean telomere length with risk of incident myocardial infarction:..., Zee [/bib_ref] [bib_ref] Cellular aging reflected by leukocyte telomere length predicts advanced atherosclerosis and cardiovascular..., Willeit [/bib_ref] as well as increased mortality [bib_ref] Association between telomere length in blood and mortality in people aged 60..., Cawthon [/bib_ref] [bib_ref] Telomere length and mortality: a study of leukocytes in elderly Danish twins, Kimura [/bib_ref] [bib_ref] The rate of leukocyte telomere shortening predicts mortality from cardiovascular disease in..., Epel [/bib_ref] [bib_ref] Telomere length predicts survival independent of genetic influences, Bakaysa [/bib_ref] [bib_ref] Telomere length is paternally inherited and is associated with parental lifespan, Njajou [/bib_ref].
Despite the essential role telomeres play in the maintenance of normal cellular function, little is known about the common genetic determinants of telomere length. Heritability estimates from family and twin studies for telomere length range from 36% to 86% [bib_ref] Genetic determination of telomere size in humans: a twin study of three..., Slagboom [/bib_ref] [bib_ref] Telomere length predicts survival independent of genetic influences, Bakaysa [/bib_ref] [bib_ref] Telomere length is paternally inherited and is associated with parental lifespan, Njajou [/bib_ref] [bib_ref] Telomere length inversely correlates with pulse pressure and is highly familial, Jeanclos [/bib_ref] [bib_ref] Mapping of a major locus that determines telomere length in humans, Vasa-Nicotera [/bib_ref] [bib_ref] Mapping genetic loci that determine leukocyte telomere length in a large sample..., Andrew [/bib_ref] [bib_ref] Evolution in health and medicine Sackler colloquium: Genetic variation in human telomerase..., Atzmon [/bib_ref]. Telomere lengths in different tissues are significantly correlated, with far less variation between tissues from the same individual compared to variation between individuals of a particular tissue type [bib_ref] Comparison of chromosome telomere integrity in multiple tissues from subjects at different..., Butler [/bib_ref] [bib_ref] Telomere length in different tissues of elderly patients, Friedrich [/bib_ref] [bib_ref] Telomere lengths are characteristic in each human individual, Takubo [/bib_ref] [bib_ref] Telomere length in the newborn, Okuda [/bib_ref] [bib_ref] Synchrony in telomere length of the human fetus, Youngren [/bib_ref] [bib_ref] Short telomeres on human chromosome 17p, Martens [/bib_ref]. Quantitative trait linkage analyses identified significant linkage to chromosomes 12p11.2 [bib_ref] Mapping of a major locus that determines telomere length in humans, Vasa-Nicotera [/bib_ref] [bib_ref] A regulatory SNP of the BICD1 gene contributes to telomere length variation..., Mangino [/bib_ref] , and 14q23.2 [bib_ref] Mapping genetic loci that determine leukocyte telomere length in a large sample..., Andrew [/bib_ref] , but these loci have not been replicated in independent studies [bib_ref] Mapping genetic loci that determine leukocyte telomere length in a large sample..., Andrew [/bib_ref] [bib_ref] A genomewide association study identifies a novel locus on chromosome 18q12.2 influencing..., Mangino [/bib_ref] [bib_ref] Common variants near TERC are associated with mean telomere length, Codd [/bib_ref] [bib_ref] Genome-wide association identifies OBFC1 as a locus involved in human leukocyte telomere..., Levy [/bib_ref]. Common genetic variants in known telomere maintenance genes do not exhibit strong influences on telomere length [bib_ref] Evolution in health and medicine Sackler colloquium: Genetic variation in human telomerase..., Atzmon [/bib_ref] [bib_ref] Polymorphisms in telomere-associated genes, breast cancer susceptibility and prognosis, Varadi [/bib_ref] [bib_ref] The association of telomere length and genetic variation in telomere biology genes, Mirabello [/bib_ref]. Genome-wide association studies (GWAS) have identified and replicated associations at the TERC locus, which account for no more than 1% of variation in telomere length [bib_ref] Common variants near TERC are associated with mean telomere length, Codd [/bib_ref] [bib_ref] Genome-wide association identifies OBFC1 as a locus involved in human leukocyte telomere..., Levy [/bib_ref]. Variants at the oligonucleotide/ oligosaccharide-binding fold containing 1 (OBFC1) gene locus, which codes for a protein that interacts with telomere-specific proteins and is implicated in telomere length regulation [bib_ref] OB fold-containing protein 1 (OBFC1), a human homolog of yeast Stn1, associates..., Wan [/bib_ref] , were also recently identified using a meta-analysis of GWAS cohorts [bib_ref] Genome-wide association identifies OBFC1 as a locus involved in human leukocyte telomere..., Levy [/bib_ref]. To identify additional variants associated with telomere length, we conducted a GWAS using data from two large cohorts that took part in the National Cancer Institute Cancer Genetic Markers of Susceptibility (CGEMS) initiative for breast and prostate cancer.
# Results
We used a two-stage GWAS to identify common genetic variants associated with log relative telomere length. Age, smoking, and mean log relative telomere length values of each population are shown in [fig_ref] Table 1: Characteristics of GWAS and replication populations [/fig_ref]. The GWAS discovery set included genotypes for 519,076 SNPs on a total of 3554 individuals from the Nurses' Health Study (NHS) and the Prostate, Lung, Colorectal, and Ovarian (PLCO) Cancer Screening Trial. After adjusting for the top principal components of genetic variation from each cohort, the distribution of the observed P values does not suggest any inflation in Type 1 error rates due to population stratification or other sources of bias [fig_ref] Figure 1: Log Quantile-Quantile [/fig_ref].
None of the P values observed in the first stage of our GWAS reached the genome-wide significance level of P,5.0610 28 [bib_ref] Estimation of the multiple testing burden for genomewide association studies of nearly..., Pe'er I [/bib_ref].
The SNP most significantly associated with log relative telomere length in this discovery set was rs1978423, located upstream of the pyroglutamyl-peptidase I (PGPEP1) and LSM4 homolog, U6 small nuclear RNA associated (LSM4) genes on chromosome 19 (P = 3.8610 27 ). We selected the 64 independent SNPs most significantly associated with log relative telomere length to genotype in additional samples in the NHS (N = 1345) and the Women's Genome Health Study (WGHS) (N = 1115) replication populations. Estimates for all SNPs exhibited moderate to high between-study heterogeneity (I 2 $0.50; [fig_ref] Table 1: Characteristics of GWAS and replication populations [/fig_ref]. rs1978423 was not associated with log relative telomere length in either replication population (P$0.12). Two SNPs, rs4073054 at the alipoprotein A-II (APOA2)/nuclear receptor subfamily 1, group I, member 3 (NR1I3) locus on chromosome 1 and rs975504 near dehydrogenase/reductase (SDR family) member 12 (DHRS12) and a predicted non-coding RNA on chromosome 13, displayed estimates similar to the associations from the initial GWAS population with P values ,0.05 in the NHS replication population, but not within the WGHS [fig_ref] Table 1: Characteristics of GWAS and replication populations [/fig_ref] ; joint P values of GWAS and replication populations are 1.7610 24 and 9.5610 24 , respectively). rs7210405, ,24 kb downstream of the cytohesin 1 (CYTH1) gene on chromosome 17, displayed the smallest P value (0.01) within the WGHS population with a similar estimate. However, this SNP was not associated with log relative telomere length in the NHS replication population (joint P = 1.4610 24 ).
We examined SNPs at the TERC and OBFC1 loci, which were recently identified by published GWAS of telomere length [bib_ref] Common variants near TERC are associated with mean telomere length, Codd [/bib_ref] [bib_ref] Genome-wide association identifies OBFC1 as a locus involved in human leukocyte telomere..., Levy [/bib_ref]. Within our discovery GWAS and WGHS replication populations, we examined associations with genotyped and imputed SNPs (rs12696304, rs10936599, rs3772190, rs16847897) at the TERC locus reported as significantly associated with telomere length in prior GWAS. In our study populations, we found evidence for associations between log relative telomere length and SNPs at this locus, with the most significant association observed for the minor allele [C] of rs16847897 (per allele b = 20.03, joint P = 0.003). Meta-analysis of our results for rs16847897 compared with studyspecific estimates from published GWAS generated a P = 1.6610 213 and displayed virtually no between-study heterogeneity [fig_ref] Table 2: Relative telomere length associations with SNPs at loci identified by published GWAS [/fig_ref]. Whereas Levy et al. [bib_ref] Genome-wide association identifies OBFC1 as a locus involved in human leukocyte telomere..., Levy [/bib_ref] reported genome-wide significant associations with SNPs in the OBFC1 gene, we did not find evidence for associations between genotyped or imputed SNPs at the OBFC1 locus and log relative telomere length in our study (joint P$0.12). Meta-analysis of our results for SNPs at the OBFC1 locus with those reported by Levy et al. demonstrated high between-study variability [fig_ref] Table 2: Relative telomere length associations with SNPs at loci identified by published GWAS [/fig_ref]. As we previously reported, we did not find associations between log relative telomere length and common genetic variants at the TERT-CLPTM1L, BICD1, and DDX11 loci [bib_ref] The association of telomere length and genetic variation in telomere biology genes, Mirabello [/bib_ref].
# Discussion
Despite the high heritability of telomere length [bib_ref] Genetic determination of telomere size in humans: a twin study of three..., Slagboom [/bib_ref] [bib_ref] Telomere length predicts survival independent of genetic influences, Bakaysa [/bib_ref] [bib_ref] Telomere length is paternally inherited and is associated with parental lifespan, Njajou [/bib_ref] [bib_ref] Telomere length inversely correlates with pulse pressure and is highly familial, Jeanclos [/bib_ref] [bib_ref] Mapping of a major locus that determines telomere length in humans, Vasa-Nicotera [/bib_ref] [bib_ref] Mapping genetic loci that determine leukocyte telomere length in a large sample..., Andrew [/bib_ref] [bib_ref] Evolution in health and medicine Sackler colloquium: Genetic variation in human telomerase..., Atzmon [/bib_ref] , common variants associated with large effects have remained elusive. We were unable to identify novel variants associated with log relative telomere length with a high degree of confidence among a total of 6,014 participants. The two most promising SNPs identified by our study were rs4073054 (NHS GWAS, b = 20.04, P = 1.6610 24 ; PLCO GWAS, b = 20.04, P = 0.041; NHS replication, b = 20.03, P = 0.01) and rs7210405 (NHS GWAS, b = 0.03, P = 0.006; PLCO GWAS, b = 0.05, P = 0.005; WGHS replication, b = 0.03, P = 0.01), as 3 of our 4 study populations displayed evidence of an association with log relative telomere length. We had .80% power to detect b estimates as small as 60.10 for SNPs with minor allele frequencies (MAF) of $0.15 in our discovery GWAS population (N = 3554). However, published associations with relative telomere length suggest effect estimates may be quite a bit smaller than 60.10. Our study was underpowered to detect genome-wide significant associations between log relative telomere length and SNPs of much weaker effect. For example, adequate power to detect genome-wide significant associations between SNPs at loci such as TERC that have a MAF ,0.25, b,0.03, and log relative telomere length as measured in our study would require a sample size of ,25,000 unrelated individuals. Even to reach the suggestive threshold of a = 1610 25 , to detect an association with such a SNP would require a sample size of 18,000 individuals, roughly 5 x greater than the size of our initial GWAS population. Discovery of additional loci with small effect sizes will be feasible once metaanalyses of existing and future GWAS of telomere length are conducted.
To date, only the TERC locus has been identified and replicated in multiple independent populations for its association with telomere length [bib_ref] Common variants near TERC are associated with mean telomere length, Codd [/bib_ref] [bib_ref] Genome-wide association identifies OBFC1 as a locus involved in human leukocyte telomere..., Levy [/bib_ref]. Although SNPs within the TERC locus did not reach genome-wide significance in the discovery GWAS population (N = 2917) of Codd et al. [bib_ref] Common variants near TERC are associated with mean telomere length, Codd [/bib_ref] , rs16847897 fell marginally outside of their pragmatic significance threshold (P,1610 25 ) for replication. Due to the biological significance of the locus, the authors examined and confirmed associations between rs16847897 and a second SNP at the locus, rs12696304, with telomere length in 3 replication populations. Levy et al. found additional evidence for associations between SNPs at the TERC locus and telomere length [bib_ref] Genome-wide association identifies OBFC1 as a locus involved in human leukocyte telomere..., Levy [/bib_ref]. We found evidence for associations with SNPs at the TERC locus, observing the strongest association between the minor allele [C] of rs16847897 and log relative telomere length (joint P = 0.003) with an effect size and direction consistent with that of Codd et al. Similar to Codd et al., we did not find associations between log relative telomere length and SNPs at loci previously reported to be associated with telomere length [bib_ref] Mapping of a major locus that determines telomere length in humans, Vasa-Nicotera [/bib_ref] [bib_ref] A regulatory SNP of the BICD1 gene contributes to telomere length variation..., Mangino [/bib_ref] [bib_ref] A genomewide association study identifies a novel locus on chromosome 18q12.2 influencing..., Mangino [/bib_ref] [bib_ref] Sequence variants at the TERT-CLPTM1L locus associate with many cancer types, Rafnar [/bib_ref] , including the OBFC1 locus recently identified by a similarly sized GWAS of telomere length [bib_ref] Genome-wide association identifies OBFC1 as a locus involved in human leukocyte telomere..., Levy [/bib_ref].
If yeast genetic mapping studies are any indication [bib_ref] A genome-wide screen for Saccharomyces cerevisiae deletion mutants that affect telomere length, Askree [/bib_ref] [bib_ref] Telomere length as a quantitative trait: genome-wide survey and genetic mapping of..., Gatbonton [/bib_ref] , hundreds of genes regulate human telomere length homeostasis. Given that known telomere maintenance genes are highly conserved [bib_ref] Genetic variation, nucleotide diversity, and linkage disequilibrium in seven telomere stability genes..., Savage [/bib_ref] and mutations have deleterious consequences [bib_ref] Dyskeratosis congenita, Savage [/bib_ref] , most functional common variants (. 5%) will likely have small effects. Additionally, recent studies demonstrate that while telomere length of offspring is significantly correlated with maternal [bib_ref] Telomere length and possible link to X chromosome, Nawrot [/bib_ref] [bib_ref] Telomere length in small-for-gestational-age babies, Akkad [/bib_ref] and paternal [bib_ref] Telomere length is paternally inherited and is associated with parental lifespan, Njajou [/bib_ref] [bib_ref] Telomere length and heredity: Indications of paternal inheritance, Nordfjall [/bib_ref] [bib_ref] Large-scale parent-child comparison confirms a strong paternal influence on telomere length, Nordfjall [/bib_ref] telomere lengths, the paternal contribution appears to be much stronger [bib_ref] Telomere length is paternally inherited and is associated with parental lifespan, Njajou [/bib_ref] [bib_ref] Telomere length and heredity: Indications of paternal inheritance, Nordfjall [/bib_ref] [bib_ref] Large-scale parent-child comparison confirms a strong paternal influence on telomere length, Nordfjall [/bib_ref] and positive correlations have been observed between paternal age at birth and offspring telomere length [bib_ref] Telomere length is paternally inherited and is associated with parental lifespan, Njajou [/bib_ref] [bib_ref] Paternal age is positively linked to telomere length of children, Unryn [/bib_ref] [bib_ref] Paternal age at birth is an important determinant of offspring telomere length, Meyer [/bib_ref] [bib_ref] Offspring's leukocyte telomere length, paternal age, and telomere elongation in sperm, Kimura [/bib_ref] ]. An influence of imprinted genes in regulating telomere length could account for these observations. By not controlling for parental origin of alleles at imprinted loci, SNP effects, which are likely to be small to begin with, would be underestimated and the power to detect associations reduced [bib_ref] Parental origin of sequence variants associated with complex diseases, Kong [/bib_ref]. For many GWAS populations, including ours, determining the parental origin of alleles is not feasible.
Evidence suggests cis-acting factors in subtelomeric regions regulate telomere length. While chromosome-specific telomeres tend to be similar between different individuals [bib_ref] Short telomeres on human chromosome 17p, Martens [/bib_ref] [bib_ref] Differences in telomere length between homologous chromosomes in humans, Londono-Vallejo [/bib_ref] [bib_ref] The pattern of chromosome-specific variations in telomere length in humans is determined..., Graakjaer [/bib_ref] [bib_ref] The effect of TERC haploinsufficiency on the inheritance of telomere length, Goldman [/bib_ref] [bib_ref] Structural stability and chromosome-specific telomere length is governed by cis-acting determinants in..., Britt-Compton [/bib_ref] , allele specific lengths are inherited [bib_ref] Allele-specific relative telomere lengths are inherited, Graakjaer [/bib_ref] and may differ by more than 6 kb at the same telomere [bib_ref] Extensive allelic variation and ultrashort telomeres in senescent human cells, Baird [/bib_ref]. A substantial amount of genetic diversity is provided by subtelomeric regions, which are riddled with repetitive DNA, segmental duplications, and large copy number polymorphisms including variable copy numbers of functional genes. Knowledge of subterminal sequences is still incomplete and the complexity within these regions has complicated sequencing efforts [bib_ref] The complex structure and dynamic evolution of human subtelomeres, Mefford [/bib_ref] resulting in low coverage of such regions on commercially available genome-wide genotyping arrays [bib_ref] Evaluation of coverage variation of SNP chips for genome-wide association studies, Li [/bib_ref].
Technical challenges further complicate the search for genetic variants of telomere length. While the quantitative PCR-based assay is the most economical, high-throughput method for telomere length measurements in large epidemiologic studies [bib_ref] Telomere measurement by quantitative PCR, Cawthon [/bib_ref] , values are relative representations of the average telomere length. The telomere restriction fragment (TRF) assay provides absolute values for the average, shortest, and longest telomere lengths. However, TRF measurements include variable amounts of the subtelomeric region, are time and labor intensive, and require much more genomic DNA than the PCR-based assay [bib_ref] The telomere length dynamic and methods of its assessment, Lin [/bib_ref]. The single telomere length analysis (STELA) assay is currently the most sensitive assay, developed with the potential to provide allele-specific telomere lengths [bib_ref] Extensive allelic variation and ultrashort telomeres in senescent human cells, Baird [/bib_ref]. To date, telomere and allele-specific STELA assays have only been developed for a very small fraction of chromosome ends due to incomplete knowledge of subtelomeric regions [bib_ref] Structural stability and chromosome-specific telomere length is governed by cis-acting determinants in..., Britt-Compton [/bib_ref]. The inability to capture the complexity of telomere length regulation with a single measurement is a major obstacle to GWAS efforts. Even the well established decline in telomere length with increasing age is not a simple linear relationship. Relatively fast rates of telomere attrition are observed during childhood and adolescence. Attrition rates slow creating a plateau from ,30 to 50 years of age followed by another decline in telomere length from about age 50 to 80 [bib_ref] The rate of telomere sequence loss in human leukocytes varies with age, Frenck [/bib_ref] [bib_ref] Telomere length and heredity: Indications of paternal inheritance, Nordfjall [/bib_ref] [bib_ref] Telomeric length and telomerase activity vary with age in peripheral blood cells..., Iwama [/bib_ref] [bib_ref] Dietary patterns, food groups, and telomere length in the Multi-Ethnic Study of..., Nettleton [/bib_ref]. It is unknown whether temporally regulated genes and/or environmental exposures are responsible for the change in attrition rates.
The age range of individuals in our GWAS population [fig_ref] Table 1: Characteristics of GWAS and replication populations [/fig_ref] and those of previously published studies [bib_ref] A genomewide association study identifies a novel locus on chromosome 18q12.2 influencing..., Mangino [/bib_ref] [bib_ref] Common variants near TERC are associated with mean telomere length, Codd [/bib_ref] [bib_ref] Genome-wide association identifies OBFC1 as a locus involved in human leukocyte telomere..., Levy [/bib_ref] span these different age periods potentially diluting telomere length associations with variants that influence temporally regulated genes.
Oxidative stress is thought to accelerate telomere attrition as a result of damage to telomeric DNA, which is less efficiently repaired [bib_ref] Oxidative stress shortens telomeres, Von Zglinicki [/bib_ref]. Systemic exposures that contribute to oxidative stress such as smoking [bib_ref] Telomere length and possible link to X chromosome, Nawrot [/bib_ref] [bib_ref] Human diseases of telomerase dysfunction: insights into tissue aging, Garcia [/bib_ref] [bib_ref] Obesity, cigarette smoking, and telomere length in women, Valdes [/bib_ref] and obesity [bib_ref] Obesity, cigarette smoking, and telomere length in women, Valdes [/bib_ref] [bib_ref] Telomere length is associated with obesity parameters but with a gender difference, Nordfjall [/bib_ref] [bib_ref] The association between physical activity in leisure time and leukocyte telomere length, Cherkas [/bib_ref] [bib_ref] Obesity and weight gain in adulthood and telomere length, Kim [/bib_ref] have been associated with shorter telomeres. Whereas, healthy lifestyle choices promote a more stable telomere length [bib_ref] Dietary patterns, food groups, and telomere length in the Multi-Ethnic Study of..., Nettleton [/bib_ref] [bib_ref] The association between physical activity in leisure time and leukocyte telomere length, Cherkas [/bib_ref] [bib_ref] The association between leukocyte telomere length and cigarette smoking, dietary and physical..., Mirabello [/bib_ref] , possibly by increasing telomerase activity in cells [bib_ref] Increased telomerase activity and comprehensive lifestyle changes: a pilot study, Ornish [/bib_ref]. Evidence suggests environmental factors may take on a more prominent role as determinants of telomere length with increasing age [bib_ref] Large-scale parent-child comparison confirms a strong paternal influence on telomere length, Nordfjall [/bib_ref] [bib_ref] Shared environmental factors associated with telomere length maintenance in elderly male twins, Huda [/bib_ref] , but the relationship with these exposures are not yet clearly defined as many studies fail to find significant correlations between telomere length and smoking [bib_ref] Leukocyte telomere length and cardiovascular disease in the cardiovascular health study, Fitzpatrick [/bib_ref] [bib_ref] Telomere length predicts survival independent of genetic influences, Bakaysa [/bib_ref] [bib_ref] Mapping genetic loci that determine leukocyte telomere length in a large sample..., Andrew [/bib_ref] [bib_ref] Telomere length is associated with left ventricular function in the oldest old:..., Collerton [/bib_ref] , obesity [bib_ref] Telomere length and cardiovascular risk factors in a middle-aged population free of..., Bekaert [/bib_ref] [bib_ref] A prospective study of relative telomere length and postmenopausal breast cancer risk, Vivo [/bib_ref] , and/or physical activity [bib_ref] Telomere length and cardiovascular risk factors in a middle-aged population free of..., Bekaert [/bib_ref] [bib_ref] Associations between diet, lifestyle factors, and telomere length in women, Cassidy [/bib_ref]. Since pack-years of smoking were significantly correlated with relative telomere length among men of the PLCO GWAS population, we adjusted for smoking characteristics in all of our regression models. Although relative telomere length in our prospectively collected blood samples was not associated with breast [bib_ref] A prospective study of relative telomere length and postmenopausal breast cancer risk, Vivo [/bib_ref] or prostate cancer [bib_ref] The association between leukocyte telomere length and cigarette smoking, dietary and physical..., Mirabello [/bib_ref] , we adjusted for case status to control for any potential effects of pre-clinical disease on telomere length.
In summary, improved sequence maps and technical capabilities are necessary to increase success in identifying and validating common genetic variants associated with telomere length homeostasis. Efforts will likely require meta-analyses of existing and future telomere length GWAS to increase power to detect common variants of small effects while stratifying by age groups defined by attrition rate and controlling for environmental factors.
# Materials and methods
# Ethics statement
The NHS and WGHS study protocols were approved by the Committee on Use of Human Subjects of the Brigham and Women's Hospital, Boston, MA. Institutional review boards at the U.S. National Cancer Institute and the 10 screening centers approved the PLCO protocol.
## Nhs breast cancer gwas population
The NHS is a prospective cohort study of 121,700 female registered nurses in 11 states in the United States who were 30-55 years of age at enrollment. In 1976 and biennially thereafter, detailed information from participants was collected by selfadministered questionnaires. Participants in this study were selected for a nested case-control study of telomere length and postmenopausal breast cancer risk from the subcohort of 32,826 women who donated a blood sample in 1989-90 [bib_ref] A prospective study of relative telomere length and postmenopausal breast cancer risk, Vivo [/bib_ref]. Eligible cases consisted of postmenopausal women of European ancestry with pathologically confirmed incident invasive breast cancer diagnosed anytime after blood collection up to with no prior diagnosis of cancer. Controls were randomly selected postmenopausal women free of cancer and matched to cases according to age, blood collection, and ethnicity. Completion of the questionnaire and submission of the blood sample was considered to imply informed consent. No significant difference was observed for log relative telomere length between 1,122 cases and 1,147 controls and no significant association with postmenopausal breast cancer risk [bib_ref] A prospective study of relative telomere length and postmenopausal breast cancer risk, Vivo [/bib_ref] ; therefore, we included telomere length data from both breast cancer cases and controls for the current study.
## Plco prostate cancer gwas population
The PLCO Cancer Screening Trial is an ongoing randomized trial with 154,942 persons aged 55 to 74 enrolled between September 1993 and July 2001 from 10 screening centers nationwide [bib_ref] Design of the Prostate, Lung, Colorectal and Ovarian (PLCO) Cancer Screening Trial, Prorok [/bib_ref]. Detailed questionnaire data was collected from all subjects at baseline. Participants provided blood and tissue samples for etiologic studies of cancer [bib_ref] The Prostate, Lung, Colorectal and Ovarian (PLCO) Cancer Screening Trial of the..., Gohagan [/bib_ref] and all participants provided written informed consent. Participants in this study were male subjects selected for a nested case-control study of telomere length, prostate cancer risk, and life-style variables [bib_ref] The association between leukocyte telomere length and cigarette smoking, dietary and physical..., Mirabello [/bib_ref]. Eligible cases and controls consisted of non-Hispanic white men aged 55 to 74 who had been screened for prostate cancer (PSA test) prior to October 1, 2003, completed a baseline questionnaire of cancer risk factors, provided a blood sample 1 month to 3 years prior to prostate cancer diagnosis for cases, and did not have a personal history of cancer prior to study entry. All cases had pathologically confirmed incident aggressive prostate cancer and a Gleason score of $7. No significant difference was observed for relative telomere length between 616 cases and 1,061 matched controls and no significant association with aggressive prostate cancer risk [bib_ref] The association between leukocyte telomere length and cigarette smoking, dietary and physical..., Mirabello [/bib_ref] ; therefore, we included telomere length data from both prostate cancer cases and controls for the current study.
## Replication populations
NHS participants in the replication phase of this study were selected for a nested case-control study of telomere length and skin cancer risk from the blood subcohort [bib_ref] A prospective study of telomere length and the risk of skin cancer, Han [/bib_ref]. Eligible cases were women of European ancestry with skin cancer diagnosed anytime after blood collection up to June 1, 2000 with no prior diagnosis of skin cancer. A common control series was randomly selected from participants who gave a blood sample and were free of diagnosed skin cancer up to and including the questionnaire cycle in which the case was diagnosed. Controls were matched to cases by age and ethnicity. The nested case-control study consisted of 218 melanoma cases, 285 cases with squamous cell carcinoma, 300 cases with basal cell carcinoma, and 870 matched controls. The study protocol was approved by the Committee on Use of Human Subjects of the Brigham and Women's Hospital, Boston, MA.
The Women's Genome Health Study (WGHS) is a prospective cohort of female North American health care professionals representing participants in the Women's Health Study (WHS) who provided a blood sample at baseline and consent for bloodbased analyses [bib_ref] Rationale, design, and methodology of the Women's Genome Health Study: a genomewide..., Ridker [/bib_ref]. The WHS was a 262 trial beginning in 1992-1994 of vitamin E and low dose aspirin in prevention of cancer and cardiovascular disease with about 10 years of follow-up. Follow-up continues in observational mode. Participants in the WHS were 45 or older at enrollment and free of cardiovascular disease, cancer or other major chronic illness and were followed prospectively for the influence of random allocation of vitamin E on cancer [bib_ref] Low-dose aspirin in the primary prevention of cancer: the Women's Health Study:..., Cook [/bib_ref] [bib_ref] Vitamin E in the primary prevention of cardiovascular disease and cancer: the..., Lee [/bib_ref] [bib_ref] A randomized trial of low-dose aspirin in the primary prevention of cardiovascular..., Ridker [/bib_ref]. Additional information related to health and lifestyle were collected by questionnaire throughout the WHS trial and continuing observational follow-up. Participants in the current analysis are individuals selected for a breast cancer case-control study nested within the WGHS. Eligible cases consisted of women of European ancestry diagnosed with pathologically confirmed incident invasive breast cancer until March 7, 2000. Controls were randomly selected participants who had given a blood sample, and were free of diagnosed cancer. Controls were matched to cases according to age, menopausal status, postmenopausal hormone use at time of blood draw, and race/ethnicity. Written informed consent was obtained from all women before their entry into the trial.
## Genotyping and quality control
As part of the CGEMS initiative, genotyping of the NHS samples was conducted for the first stage of a three-stage GWAS of breast cancer susceptibility using the Illumina HumanHap550 Infinium assay (Illumina, San Diego, CA), which contains SNPs derived from the HapMap phase I and II data [bib_ref] A genomewide association study identifies alleles in FGFR2 associated with risk of..., Hunter [/bib_ref]. Genotyping of PLCO samples for the prostate cancer susceptibility project occurred in two parts [bib_ref] Genome-wide association study of prostate cancer identifies a second risk locus at..., Yeager [/bib_ref]. Phase 1A used Illumina's Sentrix HumanHap300 assay and Phase 1B used the Sentrix Human-Hap240 assay. For each cohort, samples with call rates ,90% and single nucleotide polymorphism (SNP) assays with call rates under 90% were removed. Polymorphisms with a minor allele frequency of ,1% were removed.
Genotyping of WGHS samples was performed using the Illumina HumanHap300 Duo ''+'' chips or the combination of the HumanHap300 Duo and iSelect chips. In either case, the custom SNP content was the same; these custom SNPs were chosen without regard to minor allele frequency (MAF) to saturate candidate genes for cardiovascular disease as well as to increase coverage of SNPs with known or suspected biological function, e.g. disease association, non-synonymous changes, substitutions at splice sites, etc. For quality control, all samples were required to have successful genotyping using Illumina's BeadStudio v. 3.3 software for at least 98% of the SNPs. Self-reported European ancestry was verified on the basis of multidimensional scaling analysis of identity by state using 1443 ancestry informative markers in PLINK v. 1.06. The final data set retained SNPs with MAF .1%, successful genotyping in 90% of the subjects, and deviations from Hardy-Weinberg equilibrium not exceeding P = 10 26 in significance. Among the final 23,294 individuals of verified European ancestry, genotypes for a total of 2,608,509 SNPs were imputed from the experimental genotypes and LD relationships implicit in the HapMap r. 22 CEU samples.
Genotyping of SNPs selected for replication in the NHS nested skin cancer case-control data set was performed at the Dana Farber/Harvard Cancer Center High-Throughput Genotyping Core. Whole genome amplified DNA was genotyped using the TaqmanH OpenArrayH Real-Time qPCR system (Applied Biosystems Inc, Foster City, CA). Of the 64 SNPs selected for replication, one failed to genotype. We removed 1 SNP with a call rate ,90%. All SNPs were tested for deviation from Hardy-Weinberg equilibrium within the entire skin cancer case-control data set. We observed significant deviations for 3 SNPs at the Bonferroni-adjusted P,0.0008, which were excluded from analyses. We included 5% blinded quality control samples to validate genotyping procedures; concordance for blinded samples was .99%.
## Relative telomere length measurement
Genomic DNA was extracted from peripheral blood leukocytes using the QIAmp (Qiagen, Chatsworth, CA) 96-spin blood protocol. DNA was quantified using either the Molecular Devices 96-well spectrophotometer (NHS, WGHS) or the Nanodrop SD-1000 spectrophotometer (PLCO), and subsequently dried down and resuspended to ensure accurate and uniform DNA concentrations. Relative telomere length was measured using a previously described modified version [bib_ref] Telomere length, cigarette smoking, and bladder cancer risk in men and women, Mcgrath [/bib_ref] of the quantitative PCR-based telomere assay [bib_ref] Telomere measurement by quantitative PCR, Cawthon [/bib_ref]. Briefly, 5 ng of genomic DNA was dried down in a 384-well plate and resuspended in 10 mL of either the telomere (T) or 36B4 (S; single copy gene) PCR reaction mixture. The Telomere reaction mixture consists of 1x QuantiTectH SYBRH Green PCR Master Mix (Qiagen), 2 mM of DTT, 270 nM Tel-1b primer, and 900 nM Tel-2b primer. The Telomere thermal cycling profile proceeds as follows: 95uC for 10 minutes then 30 cycles consisting of 95uC for 15 seconds and 54uC for 2 minutes. The 36B4 reaction mixture consists of 1x QuantiTectH SYBRH Green PCR Master Mix, 300 nM 36B4u primer, and 500 nM 36B4d primer. The 36B4 thermal cycling profile proceeds as follows: 95uC for 10 minutes then 30 cycles consisting of 95uC for 15 seconds and 58uC for 1 minute and 10 seconds. The threshold cycle (Ct) value for each reaction represents the number of PCR cycles required to detect a signal over background fluorescence and is inversely proportional to the amount of starting DNA. Assuming 100% PCR efficiency, the amount of PCR product exactly doubles with each cycle. Triplicate reactions of each assay were performed on each sample and the average of the 3 measurements was used for analyses. Relative telomere length is calculated as the exponentiated ratio of the telomere repeat copy number to single-gene copy number (2 -T/S ) and represents the average telomeric DNA signal per genome of an individual. Coefficients of variation (CV) for the telomere and single-gene assay ranged from 0.66% to 3.02% and 0.56% to 2.07%, respectively.
## Statistical analyses
Continuous relative telomere length was natural logarithm transformed to satisfy the assumption of normality. We used multivariable linear regression to analyze the additive effect of each SNP (0, 1, or 2 copies of minor allele) on log relative telomere length using the pooled NHS and PLCO GWAS data set. As log relative telomere length is inversely associated with age in all populations within this study [bib_ref] The association between leukocyte telomere length and cigarette smoking, dietary and physical..., Mirabello [/bib_ref] [bib_ref] A prospective study of relative telomere length and postmenopausal breast cancer risk, Vivo [/bib_ref] [bib_ref] A prospective study of telomere length and the risk of skin cancer, Han [/bib_ref] and significantly inversely associated with smoking in the PLCO population, which included a higher proportion of heavier smokers [fig_ref] Table 1: Characteristics of GWAS and replication populations [/fig_ref] [bib_ref] The association between leukocyte telomere length and cigarette smoking, dietary and physical..., Mirabello [/bib_ref] , the regression model included age as a continuous variable, smoking status, and pack-year categories of smoking (never, ,10, 10 to ,30, 30+) in addition to disease status. Since most published studies find shorter telomere lengths among adult men compared to women [bib_ref] Telomere length inversely correlates with pulse pressure and is highly familial, Jeanclos [/bib_ref] [bib_ref] Mapping of a major locus that determines telomere length in humans, Vasa-Nicotera [/bib_ref] [bib_ref] Telomere length and possible link to X chromosome, Nawrot [/bib_ref] [bib_ref] Large-scale parent-child comparison confirms a strong paternal influence on telomere length, Nordfjall [/bib_ref] [bib_ref] Telomere length and cardiovascular risk factors in a middle-aged population free of..., Bekaert [/bib_ref] [bib_ref] The individual blood cell telomere attrition rate is telomere length dependent, Nordfjall [/bib_ref] [bib_ref] Telomere length and risk of incident cancer and cancer mortality, Willeit [/bib_ref] as well as potentially faster rates of telomere attrition among men [bib_ref] Paternal age is positively linked to telomere length of children, Unryn [/bib_ref] [bib_ref] Paternal age at birth is an important determinant of offspring telomere length, Meyer [/bib_ref] [bib_ref] Telomere length and cardiovascular risk factors in a middle-aged population free of..., Bekaert [/bib_ref] , we included gender and an age-gender interaction term in the initial GWAS linear regression analyses. To control for potential confounding by population stratification in the NHS GWAS, PLCO GWAS, and WGHS replication data sets, we additionally adjusted for the top principal components of genetic variation chosen for each study after excluding any admixed individuals clearly not of European descent. Principal components of genetic variation were calculated with EIGENSTRAT software [bib_ref] Principal components analysis corrects for stratification in genome-wide association studies, Price [/bib_ref] as described in Hunter et al, 2007 [bib_ref] A genomewide association study identifies alleles in FGFR2 associated with risk of..., Hunter [/bib_ref]. All P values are two-sided. Statistical analyses were performed with SAS (version 9.1; SAS Institute, Cary, NC) and PLINK [bib_ref] PLINK: a tool set for whole-genome association and population-based linkage analyses, Purcell [/bib_ref]. Power calculations were performed using QUANTO.
## Supporting information
Table S1 Relative telomere length association results from the GWAS and replication populations. (XLSX)
[fig] Figure 1: Log Quantile-Quantile (Q-Q) P value plot. The observed -log10 P values (Y-axis) of 519,076 SNPs from the pooled NHS and PLCO GWAS data set adjusted for the principal components of genetic variation plotted against the expected -log10 quantile (X-axis). The dashed line represents imputed P values. doi:10.1371/journal.pone.0019635.g001 [/fig]
[table] Table 1: Characteristics of GWAS and replication populations. [/table]
[table] Table 2: Relative telomere length associations with SNPs at loci identified by published GWAS.{ b estimates and P values derived from linear regression adjusted for age at blood collection, smoking status, pack-year categories of smoking, disease status, and principal components of genetic variation; N/A = SNP could not be imputed. Combined effect sizes and P values are calculated using a fixed-effect meta-analysis. Estimates in bold are based upon meta-analysis of the current study populations along with those published by Codd et al. and Levy et al. Study-specific values from the published studies were used where available. doi:10.1371/journal.pone.0019635.t002 [/table]
|
The emergence of neoadjuvant therapy in advanced melanoma
The discovery of immunotherapy and targeted therapy has introduced new and effective treatment options for advanced melanoma, providing therapeutic options where none existed before. The natural extension of these novel therapies is to identify their role in the neoadjuvant setting. Neoadjuvant therapy for advanced melanoma is still in its infancy, with a wealth of clinical trials underway. Early results are promising, allowing for management of a disease that previously had few options. We review the current literature and interim results from several ongoing investigations to understand the current state of neoadjuvant treatment options and what is to come. These studies pave the way for further advancements in melanoma therapy.Past experience with neoadjuvant biochemotherapy Early neoadjuvant trials investigated biochemotherapy for the treatment of stage III disease, which is the combination of cytotoxic chemotherapeutic agent(s) with IL-2 and/or IFN-α[5]. At that time, dacarbazine was the standard of care for metastatic melanoma despite reported response rates of approximately 20% and no proven overall survival (OS) benefit. While some regimens demonstrated increased response rates, no improvement in survival was shown[6]. Nevertheless, these studies paved the way for further investigations[7][8][9][10].Subsequently, a number of studies were published investigating neoadjuvant biochemotherapy using a regimen of multiagent chemotherapy (cisplatin, vinblastine and dacarbazine), IFN-α and IL-2. All of these studies investigated neoadjuvant biochemotherapy in patients with stage III melanoma and will be briefly discussed in chronological
Recent advances in melanoma treatment have drastically changed the therapeutic landscape for patients with advanced disease. These patients may have unresectable tumors, whether due to size or extent of involvement, or resectable disease with satellite, in-transit, regional or distant metastatic disease. The implementation of checkpoint inhibitors and targeted therapies now provide effective adjuvant therapies for these patients [bib_ref] Longer follow-up confirms relapse-free survival benefit with adjuvant dabrafenib plus trametinib in..., Hauschild [/bib_ref] [bib_ref] Systemic treatments for metastatic cutaneous melanoma, Pasquali [/bib_ref]. These encouraging results have led to a number of investigations evaluating the utility of the agents in the neoadjuvant setting.
Neoadjuvant therapy has the potential to further improve patient outcomes in melanoma. Neoadjuvant therapy can reduce the size and involvement of locally advanced tumors, rendering inoperable disease operable and reducing the extent and morbidity of large resections. Additionally, the risk of disease recurrence may be lowered through early treatment of occult metastatic disease. Due to the presence of the primary tumor during treatment, systemic therapies that induce immunomodulatory effects may have stronger and long-lasting responses when applied in the neoadjuvant setting. Preclinical murine models of advanced-stage cancer suggest that neoadjuvant immunotherapy provides a survival benefit over adjuvant immunotherapy; mice treated with preoperative immunotherapy display elevated levels of tumor-specific CD8 + T cells that are sustained postoperatively [bib_ref] Improved efficacy of neoadjuvant compared with adjuvant immunotherapy to eradicate metastatic disease, Liu [/bib_ref]. While benefits of neoadjuvant therapy are well established in other malignancies, the role in melanoma is in its infancy. We review early investigations of neoadjuvant therapy in advanced melanoma and then discuss published and ongoing trials with the novel agents.
order. Buzaid et al. reported a histologic response rate 50%, with 4/65 (6%) patients achieving a pathologic complete response (pCR) [bib_ref] Phase II study of neoadjuvant concurrent biochemotherapy in melanoma patients with local-regional..., Buzaid [/bib_ref]. At median follow-up of 27 months, they reported recurrence free survival (RFS) of 44%, which was significantly higher in patients that demonstrated a response. The OS was 58%, with no survival benefit in patients that demonstrated response to therapy. Dose reduction was required for 37.5% of patients due to treatment toxicity. Gibbs et al. reported a Phase II study of 36 patients with overall response rate (ORR) of 38.9% including pCR in 4/36 (11%) patients [bib_ref] A Phase II study of neoadjuvant biochemotherapy for stage III melanoma, Gibbs [/bib_ref]. At median follow-up of 31 months, RFS was 64.6% and OS 79.2%. Only four patients required a dose reduction in this study due to toxicity. Lewis et al. conducted a Phase II study of 92 patients from four institutions [bib_ref] Phase II multicenter study of neoadjuvant biochemotherapy for patients with stage III..., Lewis [/bib_ref]. They reported clinical ORR of 26%. At median follow-up of 40.4 months, RFS was 64% and OS was 78%. A total of 34 (38%) patients required a dose reduction due to drug toxicity. Finally, Kounalakis et al. published a retrospective review of a single-institutional experience with neoadjuvant biochemotherapy [bib_ref] A neoadjuvant biochemotherapy approach to stage III melanoma: analysis of surgical outcomes, Kounalakis [/bib_ref]. A total of 154 patients were included in this study. Median follow-up time was 3.4 years with 5-year event free survival of 61% and 5-year OS of 81%. This study reported the highest rate of toxicity in 71 (46%) patients. In summary, neoadjuvant biochemotherapies carried a relatively low response rate with high rates of adverse events. Furthermore, a prospective randomized trial of adjuvant biochemotherapy compared with IFN, the standard of care at that time, was terminated early due to futility of treatment [bib_ref] A randomized Phase III trial of biochemotherapy versus interferon-alpha-2b for adjuvant therapy..., Kim [/bib_ref]. Biochemotherapy would prove to be ineffective therapy without improvement in survival and the introduction of effective immunotherapy agents would soon revolutionize melanoma treatment.
## Era of immunotherapy & targeted therapy
Prior to the discussion of modern checkpoint inhibitors, the earliest immunotherapeutic agent was IFN-α. highdose IFN-α (HDI) was also the standard adjuvant agent, which led to evaluation as a neoadjuvant agent. Moschos et al. evaluated 20 patients treated with HDI before and after surgery [bib_ref] Neoadjuvant treatment of regional stage IIIB melanoma with high-dose interferon alfa-2b induces..., Moschos [/bib_ref]. Although they reported an ORR of 55%; only 3/20 patients demonstrated pCR. At a median follow-up of 18.5 months, RFS was 50% and OS was 65%; 25% of patients required a dose reduction due to toxicity.
The current standard of care for stage III melanoma is resection of the primary tumor followed by adjuvant therapy. Previous neoadjuvant therapies were largely ineffective with significant treatment toxicities; however, the introduction of checkpoint-inhibitors and targeted therapy demonstrate early promising results.
## Neoadjuvant checkpoint-inhibitor therapy
The two major mechanisms of checkpoint inhibition are inhibition of CTLA-4 and PD-1 protein and its ligand. Neoadjuvant immunotherapy is under investigation for both breast cancer and non-small-cell lung cancer; there is limited conclusive evidence on its efficacy in advanced melanoma [bib_ref] Improved efficacy of neoadjuvant compared with adjuvant immunotherapy to eradicate metastatic disease, Liu [/bib_ref] [bib_ref] Nivolumab plus ipilimumab as first-line treatment for advanced non-small-cell lung cancer (CheckMate..., Hellmann [/bib_ref].
The first investigation of neoadjuvant checkpoint-inhibitor therapy was published by Tarhini et al. in 2014 [bib_ref] Immune monitoring of the circulation and the tumor microenvironment in patients with..., Tarhini [/bib_ref]. This study investigated the utility of neoadjuvant ipilimumab in patients with surgically operable regionally advanced melanoma ranging from stage IIIB to IV. A total of 35 patients were enrolled and treated with two cycles of ipilimumab 10 mg/kg before and two cycles after definitive surgery with the same dose. Preoperative imaging by PET/CT 6-8 weeks after initiation of neoadjuvant ipilimumab demonstrated objective response in three patients (9%; two complete response [CR], one partial response [PR]). A total of 21 patients (64%) had stable disease and eight patients (24%) progressed despite treatment. No patients had achieved a pCR, as all patients had histologically documented residual melanoma of the surgical specimen. Median follow-up was 17.6 months with progression free survival of 10.8 months. A total of 14 (40%) patients experienced grade 3 adverse events (AEs), but there were no grade 4 or higher toxicities.
Three randomized trials were published in 2018: one investigating ipilimumab in combination with HDI, two investigating combination ipilimumab and three nivolumab neoadjuvant therapy. Tarhini et al. conducted a trial investigating safety and efficacy of combination immunotherapy with concurrent HDI [bib_ref] Neoadjuvant ipilimumab (3 mg/kg or 10 mg/kg) and high dose IFN-alpha2b in..., Tarhini [/bib_ref]. A total of 28 patients with locally or regionally advanced melanoma were randomized to ipilimumab at 3 or 10 mg/kg for two doses followed by definitive surgery. High-dose interferon (20 MU/m 2 /day, 5 days/week for 4 weeks, followed by 10 MU/m 2 /day subcutaneously 3 days/week) was given concurrently for 2 weeks prior to definitive surgery. Ipilimumab was continued for up to four doses after surgery, while high-dose interferon was resumed with the same subcutaneous regimen for 46 additional weeks. A total of 15 patients completed the intended treatment course; all other patients were limited by AEs, except for nine patients who had progressive disease. As expected, there were more grade 3/4 AEs with the higher dose of ipilimumab therapy; AEs from interferon therapy resolved after holding doses. The pCR was 36% (9/28) and was not significantly different between the dosing regimens with two additional patients classified as minimal residual disease (one cancer cell or minute clusters of cancer cells on histologic evaluation of surgical specimen). At median follow-up of 32 months, 10/11 patients with either pCR or minimal residual disease remained disease free.
Blank et al. reported the results of the OpACIN trial (NCT02437279), a randomized Phase Ib trial [bib_ref] Neoadjuvant versus adjuvant ipilimumab plus nivolumab in macroscopic stage III melanoma, Blank [/bib_ref]. A total of 20 patients with palpable stage III melanoma were equally randomized to four cycles of adjuvant or neoadjuvant ipilimumab 3 mg/kg plus nivolumab 1 mg/kg treatment (two cycles before and after surgery). In the neoadjuvant arm, all patients underwent complete lymph node dissection after at least one course of neoadjuvant therapy although only one patient completed all four intended courses of treatment. One patient in the adjuvant arm discontinued therapy due to disease progression. All the other patients stopped therapy due to grade 3/4 AEs, except one patient in the neoadjuvant arm who wished to discontinue therapy for grade 2 dermatitis. Nine of ten patients were evaluated for a pathologic response, of which seven patients achieved a response: three patients achieved pCR, three patients achieved 'near' pCR, defined as ≤10% viable tumor cells; one patient experienced a partial pathologic response (pPR), defined as ≤50% viable tumor cells. The two patients without a pathology response relapsed. At median follow-up of 21.6 months in this group, none of the seven patients with a pathologic response relapsed. In terms of AEs, 9/10 (90%) patients stopped therapy due to grade 3/4 AEs; the last patient elected to discontinue treatment for grade 2 dermatitis.
Amaria et al. reported a randomized Phase II study of neoadjuvant nivolumab 3 mg/kg monotherapy compared with combination ipilimumab 3 mg/kg and nivolumab 1 mg/kg [bib_ref] Neoadjuvant immune checkpoint blockade in high-risk resectable melanoma, Amaria [/bib_ref]. A total of 23 patients with stage IIIB and IIIC disease were evaluated; 12 in the monotherapy arm and 11 in the combination arm. The trial was ended early for early disease progression in the monotherapy arm and high rates of grade 3 AEs in the combination arm. The pCR rate was 25% (3/12) in the monotherapy arm, compared with 45% (5/11) in the combination arm. In terms of AEs, 1/12 patients had grade 3 AEs with nivolumab monotherapy, while 8/11 patients had grade 3 AEs on combination ipilimumab and nivolumab treatment. There were no grade 4 or 5 AEs in this trial. Notably, combination neoadjuvant therapy was associated with improved progression-free survival, RFS and distant metastasis-free survival; however, these were not statistically significant.
Based on the current randomized trials, early data demonstrate promising results from neoadjuvant immunotherapy that is unfortunately limited by significant grade 3 or higher AEs. There are many ongoing investigations of neoadjuvant immunotherapy. Tarhini et al. also presented interim results from an ongoing trial investigating combination pembrolizumab 200 mg with high-dose interferon (NCT02339324) [bib_ref] Neoadjuvant combination immunotherapy with pembrolizumab and high dose IFN-α2b in locally/regionally advanced..., Tarhini [/bib_ref]. Patients were treated with two doses of pembrolizumab before surgery and then every 3 weeks for up to 1 year afterward. The same high-dose interferon regimen as the combination ipilimumab and high-dose interferon study was used [bib_ref] Neoadjuvant ipilimumab (3 mg/kg or 10 mg/kg) and high dose IFN-alpha2b in..., Tarhini [/bib_ref]. A total of 20 patients were treated and the investigators reported significant AEs, including grade 5 AEs. At the time of presentation, the pCR was 35%.
The OpACIN-neo trial (NCT02977052; [fig_ref] Table 2: Ongoing clinical investigations of neoadjuvant therapy for advanced cutaneous melanoma as of... [/fig_ref] is an ongoing follow-up to the OpACIN trial discussed above, with the goal of investigating alternative scheduling of combination ipilimumab (ipi) and nivolumab (nivo) to reduce AEs and preserve efficacy [bib_ref] Neoadjuvant versus adjuvant ipilimumab plus nivolumab in macroscopic stage III melanoma, Blank [/bib_ref] [bib_ref] OpACIN-neo -a multicenter Phase II Study to identify the optimal neo-adjuvant combination..., Rozeman [/bib_ref]. The study enrolled 86 patients with resectable stage III melanoma randomized 1:1:1 to: arm A with tow cycles of ipi 3 mg/kg + nivo 1 mg/kg; arm B with two cycles of ipi 1 mg/kg + nivo 3 mg/kg and arm C with two cycles of ipi 3 mg/kg followed by two cycles of nivo 3 mg/kg, with a scheduled lymph node dissection after neoadjuvant treatment. The pCR rates were reported at 43, 57 and 24% respectively for the three arms. The investigators concluded that arm B with 1 mg/kg ipi + 3 mg/kg nivo is the optimal dose with grade ≥ 3 AEs in 40% of patients and 43% pCR in 30 patients. The trial will examine this dose of neoadjuvant therapy compared with adjuvant PD-1 blockade in a Phase III trial.
All the previously discussed trials investigated neoadjuvant therapy of patients with advanced disease (stage III or IV). We would like to highlight NCT03757689, which is an ongoing trial currently in the recruitment stage that will be the first trial to investigate immunotherapy (pembrolizumab) in node-negative (stage IIB or IIC), nonmetastatic melanoma. The primary outcome is to determine whether neoadjuvant pembrolizumab decreases rate of positive SLN in high-risk stage II patients. Another study (S1512) is exploring neoadjuvant pembrolizumab in early-stage patients with a rare type of melanoma (desmoplastic) that may be uniquely sensitive to immunotherapy [bib_ref] High response rate to PD-1 blockade in desmoplastic melanomas, Eroglu [/bib_ref]. [fig_ref] Table 2: Ongoing clinical investigations of neoadjuvant therapy for advanced cutaneous melanoma as of... [/fig_ref].
## Neoadjuvant braf/mek-inhibitor therapy
Early experience with neoadjuvant targeted therapy for BRAF V600-mutated melanomas were published in case reports and retrospective analyses. In 2012, Fadaki et al. successfully treated a patient with inoperable stage III future science group www.futuremedicine.com disease with neoadjuvant vemurafenib [bib_ref] Inoperable bulky melanoma responds to neoadjuvant therapy with vemurafenib, Fadaki [/bib_ref]. After 4 months of therapy, the patient's tumor shrank to less than 50% of its original size and the patient then successfully underwent surgery. Notably, despite clinically bulky nodal disease, only 1/40 nodes harbored metastatic cells and the patient had over 98% pathological response and remained disease free 6 months after surgery. At a similar time, Koers et al. successfully treated a patient in the Netherlands with a similar clinical response to vemurafenib [bib_ref] Vemurafenib as neoadjuvant treatment for unresectable regional metastatic melanoma, Koers [/bib_ref]. None of the 20 resected lymph nodes contained metastatic melanoma, and the patient remained disease free 5 months after surgery. Several larger case series also reported favorable results. The first reported use of neoadjuvant vemurafenib or combination dabrafenib and trametinib for 15 patients, six of which underwent surgical resection [bib_ref] BRAF inhibition for advanced locoregional BRAF V600E mutant melanoma: a potential neoadjuvant..., Sloot [/bib_ref]. Four of these patients demonstrated at least a pPR; three patients that received surgery survived for greater than 2 years and remained free of disease. Another study in Israel reported treatment of 13 consecutive patients with either vemurafenib alone, or dabrafenib with or without combination trametinib; twelve patients demonstrated clinical response and underwent successful excision, all of which demonstrated at least PR on pathologic examination [bib_ref] Perioperative BRAF inhibitors in locally advanced stage III melanoma, Zippel [/bib_ref]. While all the aforementioned studies examined stage III patients alone, one series also included stage IV patients (11/20 examined patients) [bib_ref] Patterns of histologic response to neoadjuvant targeted therapy in patients with BRAF..., Eroglu [/bib_ref]. The investigators also included patients treated with neoadjuvant combination encorafenib and binimetinib, in addition to the other targeted therapy regimens. Seven patients had pCR and none of these patients had subsequent disease recurrence. The definition of pCR used in this analysis is the same as detailed in the methods described by Tetzlaff et al. [bib_ref] Pathological assessment of resection specimens after neoadjuvant therapy for metastatic melanoma, Tetzlaff [/bib_ref].
These previous studies paved the way for the only published randomized clinical trial investigating neoadjuvant targeted therapy, published by Amaria et al. [bib_ref] Neoadjuvant plus adjuvant dabrafenib and trametinib versus standard of care in patients..., Amaria [/bib_ref]. This study included patients with resectable stage III or IV melanoma with confirmed BRAF V600E or V600K mutation. Patients were randomized 1:2 to standard of care (surgery and standard adjuvant therapy, including IFN-α2b, ipilimumab or a biochemotherapy regimen) or 8 weeks of neoadjuvant dabrafenib and trametinib followed by surgery and continued adjuvant therapy. A total of 21 patients were analyzed, seven to standard of care and 14 to neoadjuvant therapy. The study was terminated early at a prespecified interim safety analysis when the neoadjuvant arm demonstrated significantly longer event-free survival compared with standard of care. At median follow-up of 18.6 months, 10/14 patients that received neoadjuvant therapy were alive without disease progression, whereas all patients that received standard of care therapy progressed. Seven of 12 patients in the neoadjuvant arm achieved pCR after surgery and 2/12 achieved pPR and 3/12 patients had no response to neoadjuvant therapy. In terms of AEs, all patients experienced grade 2 AEs, and 2/14 patients experienced grade 3 diarrhea; there were no grade 4 or higher AEs. All patients in the neoadjuvant arm had resectable disease after neoadjuvant therapy. Completed targeted therapy studies are summarized in [fig_ref] Table 1: Summary of completed neoadjuvant immunotherapy and targeted therapy studies [/fig_ref].
Interim results from an ongoing clinical trial (NCT01972347; [fig_ref] Table 2: Ongoing clinical investigations of neoadjuvant therapy for advanced cutaneous melanoma as of... [/fig_ref] of combination dabrafenib and trametinib neoadjuvant therapy reported a high rate of pCR for resectable stage III disease [bib_ref] Phase II study of neoadjuvant dabrafenib + trametinib (D+T) for resectable stage..., Saw [/bib_ref]. Updated results of this study were presented at ESMO 2017 [bib_ref] Phase II study of neoadjuvant dabrafenib + trametinib (D+T) for resectable stage..., Menzies [/bib_ref]. At this time, 35 patients have received neoadjuvant dabrafenib and trametinib, of which 33 completed resection. Total 17 of 33 (52%) patients had pCR, though the authors note discordant pathologic response with RECIST response in 7 (21%) patients. Once again, no patients progressed during neoadjuvant treatment and no patients discontinued therapy. At median 12.1 month follow-up post resection, 12 (36%) patients had recurred; six of these patients had prior pCR and eight patients had distant recurrence.
Another neoadjuvant trial by a Dutch group, the REDUCTOR trial (EudraCT: 20134-002616-28), reported interim results from a trial of 8 weeks of neoadjuvant therapy with dabrafenib and trametinib in unresectable BRAF-mutated, locally advanced stage III or oligometastatic stage IV melanoma [bib_ref] Neoadjuvant cytoreductive treatment with BRAF/MEK inhibition of prior unresectable regionally advanced melanoma..., Blankenstein [/bib_ref]. A total of 17 patients have been included in this analysis: two progressed during neoadjuvant therapy; 14 of 15 remaining patients were restaged with resectable disease, 13 of which underwent R0 resections. A pathologic CR rate of 35% was reported. Median RFS was 9 months, with median follow-up of 22 months; the number of patients that recurred was not reported. Most patients experienced at least a grade 1 AE, but only two experienced grade 3 AEs. Of note, this is the only study that specifically investigated neoadjuvant targeted therapy in unresectable primary tumors.
## Intralesional therapy
Intralesional therapy for melanoma is a relatively new discovery that is primarily used in the setting of advanced disease, especially in patients with limited locoregional spread [bib_ref] Talimogene laherparepvec improves durable response rate in patients with advanced melanoma, Andtbacka [/bib_ref] [bib_ref] High response rate after intratumoral treatment with interleukin-2: results from a Phase..., Weide [/bib_ref]. These agents are injected directly into tumor lesions to limit systemic exposure, thus improving tolerability of treatment. Furthermore, abscopal effects have been observed, which is a systemic response to local treatment of a tumor lesion resulting in response in untreated distant lesions. A number of intralesional therapies have been investigated, but the newest agents that are under investigation for neoadjuvant therapy are talimogene laherparepvec (T-VEC) and Daromun .
T-VEC is an oncolytic herpes simplex virus type 1 which has been genetically modified to remove its virulence factor (ICP34.5) such that it selectively replicates in tumor cells and expresses GM-CSF [bib_ref] Suppression of the phenotype of gamma(1)34.5-herpes simplex virus 1: failure of activated..., He [/bib_ref] [bib_ref] ICP34.5 deleted herpes simplex virus with enhanced oncolytic, immune stimulating, and anti-tumour..., Liu [/bib_ref]. Once the engineered virus infects tumor cells, it replicates and destroys the host cell, releasing GM-CSF to initiate a tumor-specific systemic response. This agent was approved by the US FDA in 2015 based on the OPTiM trial [bib_ref] OPTIM trial: a Phase III trial of an oncolytic herpes virus encoding..., Kaufman [/bib_ref]. A subgroup analysis of Stage IIIB-IVM1a patients demonstrated ORR of 40.5% and durable response rate of 25.2% with TVEC therapy, significant improved over GM-CSF [bib_ref] Efficacy and safety of talimogene laherparepvec versus granulocyte-macrophage colony-stimulating factor in patients..., Harrington [/bib_ref]. These patients received more benefit from TVEC than other patients.
Interim results from the first neoadjuvant trial of T-VEC were presented at ASCO 2018 and SSO 2019 [bib_ref] Interim analysis of a randomized, open-label Phase II study of talimogene laherparepvec..., Andtbacka [/bib_ref]. A total of 150 patients with resectable stage IIIB-IVM1a disease were randomized 1:1 to neoadjuvant T-VEC (6 doses/12 weeks) followed by surgery versus surgery upfront. In the neoadjuvant arm, T-VEC was administered until surgery, no injectable lesions or treatment intolerance. Total 19/76 (25%) of patients in the neoadjuvant arm did not undergo surgery, 11/19 due to progressive disease. The remaining patients completed surgery with a 21% pCR. Negative margin (R0) resection was achieved in 56.1% in the neoadjuvant arm compared with 40.6% in the surgery arm. In terms of adverse events, 93% of patients in the neoadjuvant arm experienced treatment emergent AEs and 45% in the surgery arm. Severe AEs were observed in 17.8 versus 2.9% of patients in the neoadjuvant and surgery arms, respectively. The authors concluded that 12 weeks of neoadjuvant T-VEC leads to a higher pCR rate than ORR and higher rate of R0 resection. This study is ongoing [fig_ref] Table 2: Ongoing clinical investigations of neoadjuvant therapy for advanced cutaneous melanoma as of... [/fig_ref].
Daromun is the combination of two monoclonal antibody-cytokine fusion proteins (immunocytokines) darleukin (L19IL2) and fibromun (L19TNF). This allows for selective delivery of the immunocytokine to tumor sites. Previous studies of each agent alone only led to delayed progression of disease; however, murine melanoma models demonstrated synergistic effect when administered in combination, leading to complete melanoma remission [bib_ref] Preclinical evaluation of IL2-based immunocytokines supports their use in combination with dacarbazine,..., Pretto [/bib_ref] [bib_ref] The immunocytokine L19-IL2 eradicates cancer when used in combination with CTLA-4 blockade..., Schwager [/bib_ref]. In 2015, Danielli et al. published results of Daromun therapy in stage IIIC-IVM1a melanoma patients with unresectable disease [bib_ref] Intralesional administration of L19-IL2/L19-TNF in stage III or stage IVM1a melanoma patients:..., Danielli [/bib_ref]. While this study was not primarily intended to evaluate neoadjuvant therapy with Daromun, the protocol allowed for resection of tumor if feasible after 12 weeks of therapy. Eight of 20 (40%) evaluable patients became eligible for surgery and were withdrawn from the study. These results led to an ongoing international, randomized trial investigating Daromun in the neoadjuvant setting for stage IIIB and IIIC disease compared with surgery upfront (NCT03567889; [fig_ref] Table 2: Ongoing clinical investigations of neoadjuvant therapy for advanced cutaneous melanoma as of... [/fig_ref].
# Conclusion
The introduction of immunotherapy and targeted therapy has accelerated melanoma research, and has provided much needed therapies for advanced disease. Early investigations of neoadjuvant therapy have already shown promising results and are paving the way for further advancements in understanding the mechanisms of these novel agents. Effective and safe neoadjuvant therapy is on the horizon for the treatment of advanced melanoma and will allow for surgical treatment of previously unresectable disease.
## Future perspective
Early reports and interim results show that systemic neoadjuvant therapy for advanced melanoma is promising. Both classes of systemic agents seem to be moderately efficacious and preserve resectability of primary disease after completion of the treatment. Early evidence suggests that neoadjuvant targeted therapy may have a different effect on the tumor microenvironment, abrogating the rapid development of resistance as seen in adjuvant therapy. Also, as previously suggested, the pattern of pathologic response to neoadjuvant therapy may have implications for patient outcomes [bib_ref] Patterns of histologic response to neoadjuvant targeted therapy in patients with BRAF..., Eroglu [/bib_ref]. Further investigation is required to determine how strongly pathologic responses, as defined by Tetzlaff et al., correspond to improvement in clinical outcomes and long-term survival [bib_ref] Pathological assessment of resection specimens after neoadjuvant therapy for metastatic melanoma, Tetzlaff [/bib_ref] ; for example, it is not yet clear if achieving a pCR versus a non-pCR may correlate to longer relapse-free and overall survival in melanoma. Currently, neoadjuvant immunotherapy is primarily limited by AEs that are frequently dose limiting or require treatment cessation. The ongoing OpACIN-neo trial is directly addressing this and interim results indicate that there may be an optimal neoadjuvant dose that retains therapeutic benefit while balancing AEs. Optimal duration of neoadjuvant therapy is also not yet clear as different trials utilize different lengths of therapy. Ongoing and future trials are also exploring combinations of targeted and immunotherapies, and are incorporating intralesional therapies as well. There are also ongoing investigations into the role of adjuvant versus neoadjuvant treatment in melanoma; a recently opened large SWOG trial is randomizing patients to either three cycles of neoadjuvant pembrolizumab followed by adjuvant therapy or standard adjuvant pembrolizumab for 1 year (NCT03698019; Currently, there are nearly 20 ongoing clinical trials [fig_ref] Table 2: Ongoing clinical investigations of neoadjuvant therapy for advanced cutaneous melanoma as of... [/fig_ref] examining combinations of existing and novel agents, the results of which are highly anticipated to help establish the application and utility of neoadjuvant therapy in advanced melanoma. As evidence becomes available, we will better understand which group of patients will best benefit from neoadjuvant therapy.
## Executive summary
- The introduction of immunotherapy and targeted therapy has improved the adjuvant treatment of advanced melanoma. The neoadjuvant application of these agents is the natural extension.
## Past experiences with neoadjuvant biochemotherapy
- Early use of neoadjuvant biochemotherapy is briefly discussed to provide historical perspective and context for the evolution of neoadjuvant treatment in advanced melanoma.
## Era of immunotherapy & targeted therapy
- The majority of available data is from neoadjuvant treatment of advanced resectable melanoma. In general, both classes offer high treatment response that does not sacrifice resectability of the primary disease. Neoadjuvant checkpoint-inhibitor therapy - Checkpoint inhibitors and targeted agents (BRAF/MEK-inhibitors) are effective adjuvant agents. Results of the earliest trials demonstrate high initial response rates that are limited by high rates of adverse events. Ongoing trials are investigating alternative doses without sacrificing efficacy.
## Neoadjuvant braf/mek-inhibitor therapy
- There is the least amount of data regarding targeted agents, found in case series and a single completed randomized trial. Results from these early studies suggest that neoadjuvant therapy is well tolerated with high-response rates. Intralesional therapy - Intralesional therapies are ideal for the treatment of advanced melanoma because a higher dose of medication can be delivered to localized disease, limiting systemic toxicity. In the neoadjuvant setting, two agents are under investigation: talimogene laherparpvec (T-VEC) and Daromun (L19IL2 + L19TNF). T-VEC is an engineered herpes simplex virus that expressed GM-CSF in host cells, while Daromun is a combination immunocytokine that uses monoclonal antibodies to deliver IL-2 and TNF directly to tumor cells. Both have demonstrated promising local effect without significant adverse events. Investigation of these two agents for neoadjuvant therapy is in the early phases and results from ongoing clinical trials are anticipated in 2022.
# Conclusion
- Early results suggest that neoadjuvant therapy will become a mainstay of the management of advanced melanoma. There are many ongoing studies which are eagerly awaited to determine the optimal neoadjuvant treatment strategy to maximize patient outcomes. Future perspective - More follow-up time is needed to understand the effect of neoadjuvant treatments, including long-term survival and disease recurrence. Comparisons of neoadjuvant and adjuvant treatment are forthcoming. Ideally, it will be possible to identify patients that are most likely to benefit from neoadjuvant treatment; patients with unresectable tumors are ideal candidates for neoadjuvant therapy yet there is insufficient data to guide this treatment at this time. Finally, the identification of new neoadjuvant agents will provide more options for patients. or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.
No writing assistance was utilized in the production of this manuscript.
[table] Table 1: Summary of completed neoadjuvant immunotherapy and targeted therapy studies.Terminated early. ‡ Two patients with minimal residual disease (single cancer cell or minute clumps of cancer cells). § Of 14 patients, only 12 underwent surgery. One patient withdrew consent prior to initiation of protocol, stage of disease not specified. OS: Overall survival; pPR: Pathologic partial response; pCR: Pathologic complete response; RFS: Recurrence free survival. [/table]
[table] Table 2: Ongoing clinical investigations of neoadjuvant therapy for advanced cutaneous melanoma as of January 2019. AJCC: American Joint Committee on Cancer; HDI: High dose interferon; PD-1: Programmed death protein; TLR: Toll-like receptor; T-VEC: Talimogene laherparepvec. [/table]
|
Right to health and social justice in Bangladesh: ethical dilemmas and obligations of state and non-state actors to ensure health for urban poor
Background: The world is urbanizing rapidly; more than half the world's population now lives in urban areas, leading to significant transition in lifestyles and social behaviours globally. While offering many advantages, urban environments also concentrate health risks and introduce health hazards for the poor. In Bangladesh, although many public policies are directed towards equity and protecting people's rights, these are not comprehensively and inclusively applied in ways that prioritize the health rights of citizens. The country is thus facing many issues that raise moral and ethical concerns. Methods: A narrative literature review was conducted between October 2016 and November 2017 on issues related to social justice, health, and human rights in urban Bangladesh. The key questions discussed here are: i) ethical dilemmas and inclusion of the urban poor to pursue social justice; and ii) the ethical obligations and moral responsibilities of the state and non-state sectors in serving Bangladesh's urban poor. Using a Rawlsian theory of equality of opportunity to ensure social justice, we identified key health-related ethical issues in the country's rapidly changing urban landscape, especially among the poor. Results: We examined ethical dilemmas in Bangladesh's health system through the rural-urban divide and the lack of coordination among implementing agencies. The unregulated profusion of the private sector and immoral practices of service providers result in high out-of-pocket expenditures for urban poor, leading to debt and further impoverishment. We also highlight policy and programmatic gaps, as well as entry points for safeguarding the right to health for Bangladeshi citizens. Conclusions: The urban health system in Bangladesh needs a reform in which state and non-state actors should work together, understanding and acknowledging their moral responsibilities for improving the health of the urban poor by engaging multiple sectors. The social determinants of health should be taken into account when formulating policies and programs to achieve universal health coverage and ensure social justice for the urban poor in Bangladesh.
# Background
As early as 1948, the World Health Organization (WHO), described health as "a state of complete physical, mental and social well-being and not merely the absence of disease or infirmity" [bib_ref] The preamble of the constitution of the World Health Organization, Grad [/bib_ref]. The 1978 Alma-Ata Declaration further emphasized its broader social importance by conceiving of health as a "social goal whose realization requires the action of many other social and economic sectors in addition to the health sector" [bib_ref] The preamble of the constitution of the World Health Organization, Grad [/bib_ref]. In short, equity in health and healthcare were firmly established as central to the pursuit of social justice [bib_ref] Ethics of the social determinants of health, Ruger [/bib_ref] [bib_ref] Why health equity?, Sen [/bib_ref].
The notion of health as a human right and essential to ensuring human welfare and development [bib_ref] Reflections on emerging frameworks of health and human rights, Freedman [/bib_ref] dates back to the Universal Declaration of Human Rights (UDHR) in 1948and the International Covenant on Economic, Social and Cultural Rights (ICESCR) in 1966. The ICESCR established particular objectives for improving health provision, which emphasized people's right to access health services without discrimination and the State's duty to provide essential medicines and ensure equitable distribution of health facilities, goods, and services.
Bangladesh is a signatory to most of these international declarations and ratified international agreements. The Constitution of Bangladesh also gives high priority to the development of the social sector, including health and education. Constitutional provisions guarantee employment with reasonable wage, the right to social security and good quality of life, and the protection, promotion, and respect of healthcare as a constituent of human rights in Bangladesh [bib_ref] Human rights, health and the state in Bangladesh, Rahman [/bib_ref]. In recent years, the country has witnessed extraordinary gains in health and has been applauded for its deliberate policy and programmatic focus on health equity [bib_ref] National Institute of Population Research and Training (NIPORT), International Centre for Diarrhoeal..., Sen [/bib_ref]. Although many public policies in Bangladesh are directed towards equity and social justice, implementation tends to lack a comprehensive and inclusive approach that prioritizes the health rights of its citizens. This is particularly evident in the latest urban health survey, which revealed the existence of large-scale inequity and differentials in terms of health service use and health outcomes between slum and non-slum dwellers in the urban context of Bangladesh.
[10] These discrepancies are even greater for people with disabilities and/or living in gender discriminating environments.
With rapid global urbanization, more than half of the world's population currently lives in urban areas. While city living offers many opportunities and services, urban environments also concentrate health risks and introduce health hazards and disadvantages for some segments of the population. Inadequate housing and food insecurity, coupled with lack of social protection, increase the burden of disease among the poor, especially those living in slums. In addition, low-income urban dwellers are the most susceptible to natural disasters and the negative effects of unplanned and unregulated growth, such as increased air pollution, carbon emissions by industries, and unsafe work conditions leading to disability and death.
In the past few decades, the South Asian countries have experienced rapid development and urbanization, accompanied by widening inequalities in wealth and health [bib_ref] The state of urban health in India: comparing the poorest quartile to..., Agarwal [/bib_ref]. In Bangladesh, the rate of migration into Dhaka and other urban areas [bib_ref] Public health, urban governance and the poor in Bangladesh: policy and practice, Osman [/bib_ref] continues to increase, driven by perceived economic opportunity and the negative impact of climate change on coastal livelihoods [bib_ref] Climate change and health in Bangladesh: a baseline cross-sectional survey, Kabir [/bib_ref] [bib_ref] Climateinfluenced migration in Bangladesh: the need for a policy realignment, Martin [/bib_ref]. Dhaka, the capital, has become the most densely populated city in the world, and over 30% of its residents, including female garment workers, live in urban slums, or on streets, rail stations, and railroad tracks. Studies have revealed that street dwellers remain out of the formal health service delivery mechanism and suffer continuously from various diseases [bib_ref] Health needs and health-care-seeking behaviour of street-dwellers in Dhaka, Uddin [/bib_ref]. It was also found that the formal health system of the Bangladesh government lacks adequate resources and support to address the vulnerability of pregnant garment workers in relation to the physical and mental stress they face in their workplace [bib_ref] What makes pregnant workers sick: why, when, where and how? An exploratory..., Akhter [/bib_ref]. This heightened vulnerability and social exclusionof this large segment of the population raise serious moral and ethical concerns.
In this paper, we explore issues related to social justice, health, and human rights in the context of rapid urban growth in Bangladesh and discuss the present situation from the right-to-health perspective. The key questions discussed here are: i) ethical dilemmas and inclusion of the urban poor with a view to achieving social justice; and ii) the ethical obligations and moral responsibilities of the state and non-state sectors in serving Bangladesh's urban poor.
# Methods
A narrative review was carried out of the literature published in English and available in the databases of PubMed, Google Scholar, WHO, and the United Nations. In addition to these, manual searching was conducted to identify and review the relevant articles in the organizational database and library of International Centre for Diarrhoeal Disease Research (icddr,b). Articles covered a broad range that included philosophical debates, public health ethics, and descriptive reports, as well as qualitative and quantitative studies. The key words used were: "health and human rights", "social justice", "urban health", "urban poor", and "Bangladesh". The literature search was conducted between October 2016 and November 2017; during this period the collected literature was reviewed and synthesized for analysis. The review involved two stages: we first conducted an extensive search of the existing literatures, and then we screened the collected literatures in terms of their relevance to issues of social justice. During the review process efforts were made to synthesize the relevant materials to gain a comprehensive understanding.
Applying the social justice theory of American philosopher John Rawls, we identified key health-related ethical issues in the country's rapidly changing urban landscape, especially among the poor. In this paper, we begin by presenting a brief outline of Rawls' theory of "equality of opportunity" to ensure social justice. We then illustrate Norman Daniels' approach of extending Rawls' theory into the arena of public health, applying that theory particularly to addressing the social determinants of health. Subsequently, we examine the ethical dilemmas and obligations related to healthcare in Bangladesh context, with special attention to poor and marginalized populations in urban areas. Finally, we discuss how Rawls' theory could serve as a useful guide for addressing the ethical challenges of formulating and implementing policy decisions in Bangladesh to ensure equitable healthcare, with a focus on the most vulnerable population.
In his book A Theory of Justice, John Rawls describes "justice as fairness" and emphasizes the role justice needs to play within the social contract. In his view, principles of social justice should guide how society organizes to enable the equitable distribution of resources. To ensure fair distribution, Rawls proposes the idea of "original position"-a hypothetical situation in which people imagine themselves in a blank state, where they know nothing about their place in society (based on class, gender, ethnicity, birthplace, socio-economic background, and other characteristics, etc.). This blank state, in which individuals are unaware of their social position, is termed "a veil of ignorance". Rawls argues this veil of ignorance is important for formulating the set of principles that will keep society functioning effectively without undue unfairness to any particular individual or group. The equality principle seeks a level playing field so that a person's social position-which is beyond an individual's control-does not influence their outcomes.
The concept of social justice emphasizes the rational disbursement of common benefits and the sharing of collective problems. It encompasses two moral impulses that are relevant to public health: a) the improvement of human well-being by advancing health; and b) the obligation to focus on the health of the most underprivileged. Extending Rawls' theory of justice, Norman Daniels argued that justice is a necessary requirement for population health. He emphasized the moral significance of health, as it contributes significantly to opening up a variety of opportunities for people. Applying Rawls' theory to the social determinants of health, Daniels advocated that, from a policy perspective, governments should implement policies aimed at equalizing and distributing life opportunities to individuals, e.g. basic education and affordable housing, which could ultimately reduce health inequities. According to the fair equality-of-opportunity principle, therefore, "place" should not determine or restrict an individual's quality of life. Taking this argument further, we consider in this article how disparities in access to health services in Bangladesh are determined by "place", and in particular, how inequities in health and access to healthcare are distributed according to the rural-urban divide and, within urban areas, how being located in poor urban settlements limits opportunities for health and quality of life.
# Results
Ethical dilemmas in the Bangladesh health system: the rural-urban divide The current health system of Bangladesh is complex, involving several government ministries, the private sector, non-government organizations (NGOs), and development partners, each of which plays a crucial role. The Ministry of Health and Family Welfare (MOHFW) is responsible for formulating health policy and regulation and for ensuring all citizens have access to comprehensive healthcare. The MOHFW supports an extensive country-wide network of health services that includes district hospitals, upazila health complexes at the subdistrict level, union health and family welfare centres, and community clinics in rural areas. Although rural health provision is mainly the responsibility of the MOHFW, the urban health system is governed simultaneously by two ministries: the MOHFW and the Ministry of Local Government, Rural Development and Cooperatives (MOLGRDC). While public tertiary care is provided by the MOHFW in urban areas, according to the local government act primary healthcare (PHC) is the administrative responsibility of local governments through city corporations and municipalities. However, the lack of coordination between the two implementing ministries around service provision, coverage, and referral poses critical challenges to ensuring quality and accessible healthcare, especially for the urban poor.
Due to insufficient human and financial resources at the level of urban local bodies, the MOLGRDC has been providing PHC services in urban areas through an Asian Development Bank (ADB) supported project that contracts out the services to NGOs [bib_ref] Contracting urban primary healthcare services in Bangladesh -effect on use, efficiency, equity..., Heard [/bib_ref]. As in rural areas, the essential package of services offered is largely focused on maternal, neonatal, and child health, with limited capacity for the prevention and control of emerging and re-emerging diseases or for providing specialized care for men, adolescents, and the elderly [bib_ref] Who serves the urban poor? A geospatial and descriptive analysis of health..., Adams [/bib_ref]. Also, few urban PHC services are offered at convenient times for the working population, and there are no referral linkages from PHC centres to public secondary and tertiary healthcare facilities, as those are governed by the MOHFW [bib_ref] Who serves the urban poor? A geospatial and descriptive analysis of health..., Adams [/bib_ref].
Urban disadvantage and vulnerability lead to various forms of health inequalities. The absence of basic amenities in low-income settlements in urban areas, together with unsanitary environments and overcrowding, creates a vicious cycle of infections, malnutrition, and poor health . The report State of the World's Mothers 2015 asserts that the urban slum is one of the worst places to be a mother. Not surprisingly, health indicators are not only far worse in urban slums than in non-slum urban areas, but are well below the national average . While under-five mortality rate is 46 (per 1000 live births) at the national level, the rate is 57 among residents in urban slums . Similarly, neonatal and infant mortality rates in urban slum settlements are double those in non-poor urban areas. Although there are official provisions for antenatal care, skilled birth attendance, and full childhood vaccine coverage from the PHC services, service coverage remains low in urban slums . In the past decade, many infectious diseases have re-emerged in urban areas and are more widespread in urban than in rural areas [bib_ref] Sexually transmitted infections prevalence rates in slum communities of Dhaka, Bangladesh, Sabin [/bib_ref]. Overweight and obesity-risk factors for non-communicable diseases (NCDs)-are also increasing over time even among urban poor women [bib_ref] Trends of under-and overweight among rural and urban poor women indicate the..., Shafique [/bib_ref]. Hypertension, diabetes mellitus, cancer, and other NCDs are also more common in urban poor communities [bib_ref] Prevalence of non-communicable disease risk factors among poor shantytown residents in Dhaka,..., Khalequzzaman [/bib_ref].
Although the informal sector contributes the most to the economy of Bangladesh, workers in this sector, who mostly reside in urban poor communities, experience various forms of exploitation in the workplace but are unable to exercise their right to safe and healthy working conditions. Low pay, long hours, and poor nutrition make many workers vulnerable to health problems, including malnutrition and micronutrient deficiencies [bib_ref] Morbidity patterns, nutritional status, and healthcare-seeking behavior of female garment workers in..., Hasnain [/bib_ref]. At the same time, unsafe work conditions occasionally lead to disabilities and death from injuries.
Because of inadequate public service provision in urban areas, many urban dwellers seek care from the formal and informal private sectors, which are rapidly increasing in size and importance. More than 70% of the urban poor seek care from the informal private sector (i.e., pharmacies, drug sellers, and traditional healers) as a first point of care [bib_ref] Who serves the urban poor? A geospatial and descriptive analysis of health..., Adams [/bib_ref]. Unregulated over-the-counter drug selling by untrained informal providers is of obvious concern, as is the high cost of formal private sector care, differentially impacting the poorer segment of the population and leading to high out-of-pocket expenditures. Unnecessary diagnostic tests and caesarean sections are also common and impose a substantial economic burden on the poor. High out-of-pocket expenditures for these items may be catastrophic for slum dwellers and poor households, leading to deeper impoverishment [bib_ref] Reducing out-of-pocket expenditures to reduce poverty: a disaggregated analysis at rural-urban and..., Garg [/bib_ref]. Studies in similar settings showed that the poorest households are the most heavily burdened and often resort to borrowing and selling assets to meet expenses, in the absence of provisions for financial and social protection that would mitigate these impacts [bib_ref] Catastrophic health expenditure and impoverishment in Turkey, Yardim [/bib_ref].
In Bangladesh, physicians' dual practice in both public hospitals and private clinics leads to frequent absenteeism from public sector services. Unethical interactions between pharmaceutical industries and physicians also lead to negative outcomes that compromise patients' well-being [bib_ref] Relationship between doctors and pharmaceutical industry: an ethical perspective, Mandal [/bib_ref]. Aggressive pharmaceutical promotion poses an ethical threat to physicians' professionalism by influencing prescribing behaviours that are not in patients' best interests despite endorsement of 1994 Code of Pharmaceutical Marketing Practices [bib_ref] Qualitative insights into promotion of pharmaceutical products in Bangladesh: how ethical are..., Mohiuddin [/bib_ref]. In the absence of regulation and accountability, serious incidents of medical malpractice and exploitation have been reported, especially in the private sector, that disproportionately affect the poor and less educated. These instances of exploitation, dehumanization, and lack of ethical professionalism in health service delivery are also apparent in clinical and public health research dealing with human subjects [bib_ref] Poverty and health ethics in developing countries, Begum [/bib_ref]. Disadvantaged and illiterate patients with limited access to health information are often not informed of potential risks and adverse effects prior to medical procedures.
A review of bioethics in South Asia further notes distinctions between local practices and Western principles, pertaining to activities such as discussing informed consent, applying norms in clinical decision-making that perpetuate physician paternalism, involving families in decision-making, and providing precise information to patients [bib_ref] Perspectives from south and East Asia on clinical and research ethics: a..., Pratt [/bib_ref]. The issue of obtaining informed consent is of particular importance in social settings where patients belong to disadvantaged populations [bib_ref] Beyond informed consent, Bhutta [/bib_ref]. Lack of trust or communication between healthcare providers and patients is another source of stress and anxiety for the urban poor. Due to low educational status and social position, the urban poor are often unable to voice their concerns or understand medical advice. Being unaware of their right to respectful and timely quality care, many of the urban poor are incapable of asserting themselves or demanding quality and fairness.
## Ethical obligations of state and non-state actors
As mentioned earlier, the Bangladesh government has committed constitutionally to provide health services to all citizens equally. Bangladesh has been acclaimed by the international community for its significant success in terms of the Millennium Development Goals (MDGs), especially in reducing maternal and child mortality. Recently, Bangladesh has signed the United Nations Sustainable Development Goals (SDGs), which, among many other goals, obliges the country to provide healthcare to all, irrespective of social class, and to make cities and human settlements inclusive, safe, and resilient. Bangladesh has also set the target to achieve universal health coverage (UHC) by 2032. However, unless action on the social determinants of health is prioritized and good quality service to the underserved segment of the population is ensured, achieving UHC and SDGs will be challenging.
Urban health service delivery is hindered by a lack of clear and functional mechanisms for coordination and planning within and between health and local government ministries, as well as insufficient implementation capacity at the local level. There are both substantial overlap and fragmentation in service delivery, which need to be addressed through better coordination and referral linkages between public and private sectors. Given its constitutional obligation, the government is responsible for the rights to health of its citizens and, by extension, the regulation of both state and non-state actors in health. Towards this goal, the MOHFW has made significant progress in improving accountability and quality of service delivery through the implementation and scale-up of health management information systems (HMIS). However, HMIS data on urban areas are scant, with the exception of large hospitals and donor-funded NGO networks. Data from the private sector are still not transmitted as a regular practice despite that sector's overwhelming presence in the healthcare system. This is particularly the case among smaller clinics serving poor urban settlements, where capacity and coordination to collect and use routine health information are limited. The recently migrated poor who settle in urban slums are often absent in the system, a situation that jeopardizes their health entitlement. Despite the overwhelming presence of private healthcare service delivery in urban areas, this dimension remains completely undocumented.
On a positive note, the NGO sector has played, and can continue to play, an important role in service delivery in collaboration with government and other stakeholders. In terms of supporting the goal of UHC [bib_ref] Innovation for universal health coverage in Bangladesh: a call to action, Adams [/bib_ref] , evidence suggests that urban NGOs contribute importantly to improved coverage, equity, quality of care, and efficiency, although they remain a relatively minor player in the urban healthcare landscape. Also noteworthy, on the other hand, are criticisms of the NGO sector's susceptibility to donor directives and biomedical approaches that ignore the broader context of development [bib_ref] Comprehensive versus selective primary health care: lessons for global health policy, Magnussen [/bib_ref]. Evidence of duplication in the sector similarly shows a failure to abide by the Paris Declaration, which emphasizes the coordination of aid in the health sector as an ethical obligation.
# Discussion
This paper identifies the emerging health challenges facing the urban poor in Bangladesh as a consequence of rapid and unplanned urbanization and of the absence of coordinated urban health governance in a pluralistic health system. Referring to the Rawlsian theoryof "equality of opportunity" to ensure social justice, we have identified the existing ethical dilemmas and moral obligations of the state and non-state sectors in serving the urban poor in Bangladesh.
In his Theory of Justice, Rawls put forward the idea that, to ensure social justice, the basic institutions of society need to function in compliance with the principles of justice. Rawls thus placed particular emphasis on the mechanism through which the principles of justice are formulated. His views on "original position" and "veil of ignorance" suggest ways of formulating the principle of justice to support the building of a just society. Rawls' position was in opposition to classical utilitarianism, which advocates maximum well-being for the maximum number of people. In his view, classical utilitarianism seems to favour the majority over the minority and is ethically unfair, as it results in depriving society's poorest and weakest of access to basic rights and liberties. Therefore, Rawls argued mainly for a fair distribution of goods. His idea of fairness would see individuals in a society having access to the services they need; in his justice criterion of the "Maximin principle", Rawls focused particularly on ensuring the well-being of disadvantaged groups in society.
In Bangladesh, the MOHFW has achieved notable success in recent years through its sector-wide approach, especially in terms of service delivery in rural areas [bib_ref] Fifteen years of sector-wide approach (SWAp) in Bangladesh health sector: an assessment..., Ahsan [/bib_ref]. However, its stewardship is constrained by a weak legal framework and institutional inadequacies, particularly around regulation and coordination. The government's capacities to regulate, plan, and provide health services have been further challenged by rapid urbanization [bib_ref] Who serves the urban poor? A geospatial and descriptive analysis of health..., Adams [/bib_ref]. Given that Bangladesh has committed to provide health services for all its citizens and to ensure their right-to-health, decision-makers need to be cognizant of the many ethical dilemmas inherent in the urban health space. In this context, the Rawlsian approach could provide useful guidance for identifying the current status of justice in healthcare policies, especially in terms of how the existing system treats the disadvantaged groups in society. Rawls' idea of principles of justice stipulates that individuals should not be discriminated against on the basis of their place in society. The urban poor in Bangladesh, however, are facing more daunting challenges in terms of healthcare outcomes in comparison with their rural counterparts. Considering the urban poor as disadvantaged group, the Rawlsian approach, with a view to establishing the fair distribution of goods, would suggest that national health policies and strategies need to incorporate urban health as a distinctive area of focus. In particular, special attention should be given to addressing deep inequities in access to healthcare and reducing the disproportionate exposure of the urban poor to adverse social, economic, and environmental conditions that provoke ill health. Important in this response is recognition of the heterogeneity of the urban poor and the context-specific challenges they face. To improve slum conditions, integrated urban development efforts are especially needed that address issues of housing, tenure security, water and sanitation, green space, education, and healthcare [bib_ref] Brazil's early urban transition: what can it teach urbanizing countries? London: International..., Martine [/bib_ref].
In Bangladesh, the urban population's contribution to the Gross Domestic Product (GDP) is estimated as 50% [bib_ref] Public health, urban governance and the poor in Bangladesh: policy and practice, Osman [/bib_ref]. Urban slum dwellers having migrated from rural areas in search of better economic opportunities are an important part of this economic growth. As a matter of human rights, therefore, slum dwellers should have the same opportunity as others in the country to access basic amenities as full citizens. To achieve this, the State will need to give special attention to policy planning for the poor and disadvantaged in urban Bangladesh. Moreover, there is evidence that, in developing countries, focusing on urbanization in a positive and proactive way can result in a large portion of the urban poor population becoming strong contributors to overall economic growth, rather than being a social burden [bib_ref] Brazil's early urban transition: what can it teach urbanizing countries? London: International..., Martine [/bib_ref]. In light of these examples, Bangladesh should adopt an inclusive strategy for urban areas targeting the health issues of the urban poor.
In discussions around implementing Rawls' approach in healthcare, one key observation has been that Rawls did not mention health and the social determinants of health as primary goods. However, Norman Daniels has further enhanced Rawls' theory to include health services. Starting from a "normal function" premise, Daniels argued that health has a moral significance, in that it opens up other life opportunities to an individual, and a society will be unjust if it allows health inequalities to limit the individual's basic life opportunities. From this ethical standpoint, Daniels advocated that, apart from the biomedical concept of health, all the social determinants that affect health should be considered for fair distribution according to Rawls' principle of justice [bib_ref] Enhancing John Rawls's theory of justice to cover health and social determinants..., Ekmekci [/bib_ref]. Daniels' proposition on social determinants has been helpful in developing specific implementable actions for achieving health that have mitigated the deliberate exclusion of health in Rawls' theory.
Displacing the biomedical paradigm of health and its focus on the discovery, treatment, and cure of disease and illness, the broader concept of social determinants of health and their contribution to sustainable development and health is widely accepted [bib_ref] Building of the global movement for health equity: from Santiago to Rio..., Marmot [/bib_ref]. The increasing interactions among health, sustainable development, environment, and gender have underscored the importance of the cumulative effect of interventions across various components of society to ensure social justice [bib_ref] Engagement of sectors other than health in integrated health governance, policy, and..., De Leeuw [/bib_ref]. In Bangladesh, policy-makers also need to acknowledge that health is not an output of the health sector alone, but an outcome of many other factors and sectors beyond the confines of health services provision. The government of Bangladesh should prioritize a "Health in All Policies" (HiAP) approach to public policies across sectors that systematically takes into account the broader social determinants of health. As theorized by Rawls, this approach will provide the foundation for greater equality of opportunityfor the urban poor and disadvantaged and thereby ensure social justice in health and other entitlements. HiAP is particularly important for urban settings, where health problems are concentrated and their determinants are complex. In short, without successful integration and coordination between the health and non-health sectors, the health of the urban poor will not be improved [bib_ref] The importance of intersectoral factors in promoting equity-oriented universal health coverage: a..., Huda [/bib_ref] [bib_ref] Implementing health in all policies -time and ideas matter too! Comment on..., Clavier [/bib_ref].
Investments in social protection and health insurance may reduce out-of-pocket expenditures and help prevent catastrophic medical expenditures, while greater coordination among urban service providers around hours and locations, community outreach, and referral may improve services coverage and use [bib_ref] Health care for poor people in the urban slums of Bangladesh, Afsana [/bib_ref]. A strengthened referral system joining different levels of care and spanning private and public sectors is particularly crucial.
Urban health financing is often fragmented because of aid modalities that focus on time-limited projects, while fragmentation and poor coordination among implementing agencies in both state and non-state sectors undermine quality, equity and efficiency of health services. These inefficiencies in urban health services provision might be overcome by strengthening governance at both national and local levels. This would require energizing the public health mandates of local governments and ensuring coordination among implementing bodies. While coordination is crucial to address bottlenecks in governance that hinder the availability, accessibility, and utilization of urban health services, it requires policy support at the highest level to be effective.
# Conclusion
In sum, health is neither an isolated process output nor the sole concern and responsibility of the health sector. Multiple sectors and actors-including those within national and local governments, private agencies, NGOs, and international partners-need to work together on the multiple social determinants of health. This calls for gaining a deep understanding of the needs of the urban poor through intersectoral forums and then implementing needs-based social interventions.
Although the purpose of this paper is to examine the health of the urban poor from a rights perspective, our major focus has been on the obligations of the supply side, including both the health and related non-health sectors. Government should not only implement regulatory mechanisms to establish the governance of both state and non-state sectors, but also ensure every citizen's right to obtain good quality healthcare services from all sectors when needed. At the same time, both the state and non-state sectors should ensure a supportive environment is created to address demand-side issues that concern the public, such as improving health literacy and increasing citizens' awareness of their legal rights and responsibilities. It might also be useful to create a social solidarity mechanism by developing a mutual cooperation system among the urban poor, which could be particularly beneficial during emergency situations often faced by this group. While better coordination is crucial to strengthen the quality, coverage, and equity of urban health services delivery, optimal strategies will be those that take into account the heterogeneous needs of the urban poor and disadvantaged. The particular needs of vulnerable groups such as the extreme poor, women, children, and the disabled, warrant in-depth investigation and the implementation of specialized service delivery approaches to ensure health equity. Decision-makers in Bangladesh should explore policies and programs undertaken in similar settings in the global South and tackle this challenge by identifying best practices and learning from both successes and failures.
Ultimately, progress in the health of the urban poor will depend on how well the Bangladesh government can exercise good governance and ensure accountability. Strong stewardship and commitment to strengthening systems in order to provide equitable and ethical health services for the urban poor is a crucial step towards ensuring all citizens enjoy their right-to-health and reach their full potential. |
Exploring the applications of hyaluronic acid‐based nanoparticles for diagnosis and treatment of bacterial infections
Hyaluronic acid (HA) has become a topic of significant interest in drug delivery research due to its excellent properties, including biosafety, biodegradability, and nonimmunogenicity. Moreover, due to its ease of modification, HA
can be used to prepare several HA-based nanosystems using various approaches. These approaches involve conjugating/grafting of hydrophobic moieties, polyelectrolytes complexation with cationic polymers, or surface modification of various nanoparticles using HA. These nanoparticles are able to selectively deliver antibacterial drugs or diagnostic molecules into the site of infections. In addition, HA can bind with overexpressed cluster of differentiation 44 (CD44) receptors in macrophages and also can be degraded by a family of enzymes called hyaluronidase (HAase) to release drugs or molecules. By binding with these receptors or being degraded at the infection site by HAase, HA-based nanoparticles allow enhanced and targeted antibacterial delivery. Herein, we present a comprehensive and up-to-date review that highlights various techniques of preparation of HA-based nanoparticles that have been reported in the literature. Furthermore, we also discuss and critically analyze numerous types of HA-based nanoparticles that have been employed in antibacterial delivery to date. This article offers a critical overview of the potential of HA-based nanoparticles to overcome the challenges of conventional antibiotics in the treatment of bacterial infections. Moreover, this review identifies further avenues of research for developing multifunctional and biomimetic HA-based nanoparticles for the treatment, prevention, and/or detection of pathogenic bacteria.
This article is categorized under: Therapeutic Approaches and Drug Discovery > Nanomedicine for Infectious Disease Nanotechnology Approaches to Biology > Nanoscale Systems in Biology Therapeutic Approaches and Drug Discovery > Emerging Technologies
To date, several types of nanodrug delivery systems, including liposomes, polymersomes, micelles, solid lipid nanoparticles (SLNs), and others, using various types of organic and inorganic materials such as polymers, lipids, or hybrid systems, have been reported in the literature for their enhanced antibacterial potential [bib_ref] Nanoengineered drug delivery systems for enhancing antibiotic therapy, Kalhapure [/bib_ref] [bib_ref] Nanoantibiotics: Future nanotechnologies to combat antibiotic resistance, Muzammil [/bib_ref] [bib_ref] Nano-technology for targeted drug delivery to combat antibiotic resistance, Sharma [/bib_ref]. Owing to their biocompatibility, ease of surface and chemical modifications, high drug loading capacity, and microenvironment responsiveness, polymer-based NPs have gained significant attention in antibacterial applications [bib_ref] An update on polysaccharide-based nanomaterials for antimicrobial applications, Arora [/bib_ref] [bib_ref] Hyaluronic acid nanoparticles as nanomedicine for treatment of inflammatory diseases, Rao [/bib_ref]. An interesting natural polymer that has generated increasing interest as a component of antibacterial nanodrug delivery systems over the last two decades is hyaluronic acid (HA; [bib_ref] Development of poly-L-lysine multi-functionalized muco-penetrating self-emulsifying drug delivery system (SEDDS) for improved..., Arshad [/bib_ref] [bib_ref] Ionic coupling of hyaluronic acid with ethyl N-lauroyl L-arginate (LAE): Structure, properties..., Gamarra [/bib_ref] [bib_ref] Hyaluronic acid-based nanogels improve in vivo compatibility of the anti-biofilm peptide DJK-5, Kłodzi Nska [/bib_ref] [bib_ref] Evaluation of hyaluronic acid nanoparticle embedded chitosan-gelatin hydrogels for antibiotic release, Özkahraman [/bib_ref]. Hyaluronic acid is a linear, naturally occurring mucopolysaccharide composed of alternating N-acetyl glucosamine and D-glucuronic acid units, which is synthesized in the plasma membrane by hyaluronan synthases [bib_ref] Hyaluronidases of gram-positive bacteria, Hynes [/bib_ref]. It is also a versatile material for nanodrug delivery systems as it exists in a wide range of molecular weights starting from 6.1 kDa up to 107 kDa. Additionally, it plays a crucial role in the configuration and organization of extracellular matrix, wound healing, cell adhesion control, the elasticity of connective tissue, and inflammatory modulation. Very importantly, HA is considered harmless and is rapidly degraded by hyaluronidase enzymes (HAase), making it an ideal polymer for targeted drug delivery systems . Furthermore, the favorable characteristics of HA, such as biodegradability, nonimmunogenicity, waterretaining activity, and selectivity toward specific receptors, have contributed to its popular use in drug delivery systems [bib_ref] Hyaluronan/tannic acid nanoparticles via catechol/Boronate Complexation as a smart antibacterial system, Montanari [/bib_ref]. The cluster of differentiation, or CD proteins, is a group of glycoproteins that are abundant throughout the body and are considered the primary HA receptors [bib_ref] Hyaluronic acid for anticancer drug and nucleic acid delivery, Dosio [/bib_ref]. Normally, these receptors (like CD44) are responsible for various cellular actions, including inflammation and cellular adhesion responses. In contrast, it was found to be overexpressed on several cancer cells; this phenomenon captivated the interest of researchers working on anticancer medicine. Consequently, several cancers, including ovarian, colon, breast, and squamous cancer have been extensively studied using HA nanoparticles (HA-NPs; J. H. [bib_ref] Hyaluronic acid-based nanomaterials for cancer therapy, Kim [/bib_ref] [bib_ref] Cancer nanotechnology: Application of nanotechnology in cancer therapy, Misra [/bib_ref].
Nanodrug delivery systems that consist of HA as a component have also exhibited activity against bacteria due to their intrinsic antibiofilm and bacteriostatic activity against certain bacteria [bib_ref] Antiadhesive and antibiofilm activity of hyaluronic acid against bacteria responsible for respiratory..., Drago [/bib_ref] [bib_ref] Bacteriostatic effects of hyaluronic acid, Pirnazar [/bib_ref]. Recently, CD44 receptors have been revealed to be overexpressed on human macrophages and could therefore be selectively targeted by HA-NPs to eliminate intracellular bacterial infections [bib_ref] Hyaluronan/tannic acid nanoparticles via catechol/Boronate Complexation as a smart antibacterial system, Montanari [/bib_ref]. Moreover, wound healing and tissue regeneration properties of HA may aid in the treatment of cutaneous infections and enable rapid recovery [bib_ref] Hyaluronic acid and polyethylene glycol hybrid hydrogel encapsulating Nanogel with hemostasis and..., Zhu [/bib_ref]. These exceptional abilities of HA have led to the growing interest in the formulation and development of HA-NPs to target bacterial infections.
There have been reviews that cover in-depth the use of HA-based nanomaterials in the treatment of cancer, inflammatory diseases, and nucleic acids delivery [bib_ref] Hyaluronic acid for anticancer drug and nucleic acid delivery, Dosio [/bib_ref] [bib_ref] Hyaluronic acid-based nanomaterials for cancer therapy, Kim [/bib_ref] [bib_ref] Hyaluronic acid nanoparticles as nanomedicine for treatment of inflammatory diseases, Rao [/bib_ref] [bib_ref] Hyaluronan-modified nanoparticles for tumor-targeting, Sakurai [/bib_ref]. To the best of our knowledge, this is the first review that highlights and discusses all reported nanosystems in the literature that incorporate HA using different approaches for enhanced delivery of antibiotics to prevent, detect, and/or eradicate bacterial infections. Herein, we present an in-depth and up-to-date review that highlights different techniques of preparation of HA-NPs that have been reported in the literature. Furthermore, after an extensive analytical search of several scientific databases, we discuss and critically analyze numerous types of HA-NPs that have been employed in antibacterial delivery to date. We have systematically categorized these HA-NPs according to their approach for application: (i) HA-based NPs; (ii) HA-capped NPs; (iii) Drug-Conjugated HA-NPs. The benefits and limitations of these nanomaterials in the treatment of bacterial infections are thoroughly discussed and assessed. Finally, this review highlights the gaps, challenges, and future perspectives of HA-NPs for antibacterial drug delivery.
## | hyaluronic acid and its applications
Hyaluronic acid, conjointly termed hyaluronan, is a linear natural, polyanionic, and FDA-approved biocompatible polymer , synthesized in mammals by three hyaluronan synthases (HAS1, HAS2, and HAS3; [bib_ref] Hyaluronan: Preparation, structure, properties, and applications, Lapčík [/bib_ref]. These enzymes elongate the HA backbone by repeatedly adding glucuronic acid and N-acetyl-D-glucosamine groups to the growing sugar, resulting in the formation of HA with a range of molecular weights (2 Â 10 5 -2 Â 10 6 Da) [bib_ref] Mammalian hyaluronan synthases, Itano [/bib_ref]. The estimated quantity of HA in the human body is 15 g, of which one-third is replenished daily, with an average half-life varying from 2.3 to 5.5 min. Approximately 7-8 g of this HA is found in the skin; moreover, HA is also located in connective tissue, vitreous humor, synovial fluids, and others [bib_ref] Hyaluronic acid (hyaluronan): A review, Necas [/bib_ref]. Hydrolysis of HA is predominantly by the HAase family, which may be categorized based on the producing organism into (i) mammalian HAases, which cleave the β1-4 glycosidic linkage; (ii) leech HAases, which target the β1-3 glycosidic bond; and (iii) bacterial HAase (B-HAase), also known as hyaluronan lyase, which targets β1-4 bond like mammalian enzyme; however, the end product here is unsaturated disaccharides [bib_ref] Hyaluronidases of gram-positive bacteria, Hynes [/bib_ref]. For biomedical applications, HA has been historically obtained by isolation from rooster combs; however, due to the overuse of animal rooster combs in the production of biomaterials, microbial fermentation, particularly from Streptococci species, has arisen as a new source of HA (L.. Hyaluronic acid has attracted increased interest in the development of various drug delivery systems; however, owing to its short blood circulation time and poor enzymatic stability, several HA conjugates are prepared to overcome these limitations without affecting the value inherent characteristics of HA [bib_ref] Modified hyaluronic acid based materials for biomedical applications, Tiwari [/bib_ref].
The backbone of HA consists of several carboxyl, hydroxyl, and N-acetyl groups , which can be easily modified, resulting in a variety of HA derivatives. The primary approaches for modifying HA include the creation of ester and amide bonds at the carboxyl groups of the polymer in the presence of condensing agents, as well as the formation of ester and ether bonds at the hydroxyl groups of the polymer [bib_ref] Chemical modifications of hyaluronic acid for the synthesis of derivatives for a..., Schanté [/bib_ref]. The resultant conjugates can be used to construct HA-based delivery systems with enhanced physicochemical characteristics, including increased enzymatic stability, altered viscoelastic behavior, or controlled release [bib_ref] Crosslinking method of hyaluronic-based hydrogel for biomedical applications, Khunmanee [/bib_ref]. Additionally, they are suitable for drug delivery applications by exploiting the inherent targeting characteristics of HA, minimizing immune recognition, or prolonging drug circulation [bib_ref] Hyaluronic acid as potential carrier in biomedical and drug delivery applications, Prajapati [/bib_ref].
Hyaluronic acid and its conjugates have been used to formulate several types of nanocarriers, which can be prepared using three main strategies (J. H. [bib_ref] Hyaluronic acid-based nanomaterials for cancer therapy, Kim [/bib_ref] [bib_ref] Hyaluronan-modified nanoparticles for tumor-targeting, Sakurai [/bib_ref] , as shown in . The first strategy involves conjugating/crosslinking hydrophobic moieties (including drugs) to the HA backbone, which results in the formation of amphiphilic derivatives that can be employed to formulate various types of NPs, including nanogels [bib_ref] Delivery of LLKKK18 loaded into self-assembling hyaluronic acid nanogel for tuberculosis treatment, Silva [/bib_ref] , polymersomes, micelles [bib_ref] Hyaluronic acid-tocopherol succinate-based self-assembling micelles for targeted delivery of rifampicin to alveolar..., Gao [/bib_ref] , and others (J. H. [bib_ref] Hyaluronic acid-based nanomaterials for cancer therapy, Kim [/bib_ref]. Owing to the anionic nature of HA, the second strategy is via ionotropic gelation with cationic polymers; this method accounts for approximately 30% of formulated HA-based NPs [bib_ref] Hyaluronan-modified nanoparticles for tumor-targeting, Sakurai [/bib_ref]. Finally, the third strategy involves surface modification of various nano-delivery systems, which is achieved by depositing an auxiliary layer(s) on the surface of the nanocarriers, which alters the properties of the nanocarrier with or without the formation of covalent bonds (K. [bib_ref] Hyaluronic acid-coated nanomedicine for targeted cancer therapy, Kim [/bib_ref]. The produced HA-based NPs exhibit several desirable features, including nonimmunogenicity, biosafety, and anti-inflammatory activity [bib_ref] Hyaluronic acid-based nanocarriers for intracellular targeting: Interfacial interactions with proteins in cancer, Choi [/bib_ref]. HA-based NPs have been thoroughly investigated in the field of drug delivery to enhance the biocompatibility of the material and facilitate drug delivery via passive and active targeting. Recently, the number of published articles on "hyaluronic acid" and "nanoparticles" have grown remarkably [bib_ref] Hyaluronan-modified nanoparticles for tumor-targeting, Sakurai [/bib_ref] ; interestingly, the number of published articles has reached 2019 and 3095 in PubMed and Scopus, respectively, as of the date of this review. However, compared to anticancer delivery, the application of HA in the nano delivery of antibiotics is still in its infancy.
The application of HA to develop nanomaterials for antibacterial delivery may demonstrate enhanced and synergistic activity owing to its intrinsic bacteriostatic and antibiofilm properties against certain bacterial strains [bib_ref] Antiadhesive and antibiofilm activity of hyaluronic acid against bacteria responsible for respiratory..., Drago [/bib_ref] [bib_ref] Bacteriostatic effects of hyaluronic acid, Pirnazar [/bib_ref]. Hyaluronic acid is also known to have wound healing, tissue regeneration, and antiinflammatory characteristics, which may aid in treating cutaneous infections and promoting rapid recovery [bib_ref] Hyaluronic acid and polyethylene glycol hybrid hydrogel encapsulating Nanogel with hemostasis and..., Zhu [/bib_ref]. Furthermore, numerous human macrophages have been shown to overexpress CD44 receptors, suggesting that HA-NPs might be used to target and destroy intracellular bacterial infections selectively [bib_ref] Hyaluronic acid (HA) presentation as a tool to modulate and control the..., Almalik [/bib_ref].
[formula] F I G U R E 1 [/formula]
The structure of HA as well as potential locations for chemical modification of the polymer It has been reported that polycarboxylic acids like HA are found to reduce the pH of infection sites, thereby generating an environment where bacteria find it hard to survive. Additionally, the concept of developing HAase-responsive antibiotic-loaded NPs has been employed to enable on-demand release of the encapsulated antibiotics, resulting in safe and effective antibiotic delivery [bib_ref] Enzyme responsive hyaluronic acid nanocapsules containing polyhexanide and their exposure to bacteria..., Baier [/bib_ref]. All these benefits make HA-based nanomaterials an excellent candidate for eradicating both extracellular and intracellular bacteria.
The following sections review in detail HA-based nanomaterials that have been used for antibacterial drug delivery. Herein, we systematically classified these nanomaterials into three broad groups based on the rationale behind the use of HA: (i) HA-based nanocarriers that are conjugated, crosslinked, or ionically complexed with other molecules (fatty acids, polymers, etc.); (ii) HA-coated NPs, where HA is utilized to modify the surface of NPs; and (iii) HA-drug conjugate, which contains a variety of antibiotics covalently bonded to HA.
[formula] F I G U R E 3 [/formula]
The total number of publications on HA-based NPs. When the terms "hyaluronic acid" and "nanoparticles" were entered in PubMed (https://www.ncbi.nlm.nih.gov/pubmed/) and Scopus (https://www.scopus.com/search), the number of publications that appeared year after year was graphed F I G U R E 2 Nano-delivery systems based on HA. Both (a) and (b) illustrate conjugation of HA with a hydrophobic moiety or drug, which self-assemble into NPs; (c) ionic gelation of HA with cationic polymer; and (d) HA-coated NP 3 | HYALURONIC ACID-BASED NANOCARRIERS Hyaluronic acid and its conjugates have been used to formulate various types of nanocarriers. These nanocarriers exhibited enhanced biological activity, including anticancer , anti-inflammatory [bib_ref] Hyaluronic acid nanoparticles as nanomedicine for treatment of inflammatory diseases, Rao [/bib_ref] , and most importantly, antimicrobial activity [bib_ref] Antimicrobial hyaluronic acid-Cefoxitin sodium thin films produced by Electrospraying, Ahire [/bib_ref]. As it is an ideal biopolymer, HA has been widely used to develop HA-NPs as targeted antibacterial nano delivery systems [bib_ref] Biopolymer nanogels improve antibacterial activity and safety profile of a novel lysine-based..., Kłodzi Nska [/bib_ref] [bib_ref] Hyaluronan-based Nanohydrogels for targeting intracellular S. aureus in human keratinocytes, Montanari [/bib_ref] [bib_ref] Biodistribution and intracellular localization of hyaluronan and its nanogels. A strategy to..., Montanari [/bib_ref]. Therefore, this section will highlight and critically analyze all previously reported HAbased nanocarriers used for antibacterial delivery to date.
3.1 | Nanogels Nanogels (NGs) are spherical nanosized networks produced when polymers are chemically or physically crosslinked. They exhibit the distinct properties of both hydrogels and NPs. Nanogels can be prepared using a wide variety of polymers. At the top of the list is HA, owing to its simple conjugation methods, excellent mechanical properties, and biosafety (H. [bib_ref] New progress and prospects: The application of nanogel in drug delivery, Zhang [/bib_ref]. The review on NGs shows that various HA-NGs have been formulated with as well as without drug loading to target specific objectives (P. D. [bib_ref] Hyaluronic acid-based nanogels improve in vivo compatibility of the anti-biofilm peptide DJK-5, Kłodzi Nska [/bib_ref] [bib_ref] Biodistribution and intracellular localization of hyaluronan and its nanogels. A strategy to..., Montanari [/bib_ref].
Various antibiotics have been encapsulated into HA-based NGs (HA-NGs) to eliminate bacterial infections (S. N. [bib_ref] Utilizing nanoparticles for improving anti-biofilm effects of azithromycin: A head-to-head comparison of..., Kłodzi Nska [/bib_ref] [bib_ref] Chasing bacteria within the cells using levofloxacin-loaded hyaluronic acid nanohydrogels, Montanari [/bib_ref] [bib_ref] Hyaluronan-based Nanohydrogels for targeting intracellular S. aureus in human keratinocytes, Montanari [/bib_ref]. In this regard, reported two studies on HA-NGs encapsulated with levofloxacin (LVF) and gentamycin (GM). The HA-NGs in these studies were prepared using the nanoprecipitation technique of cholesterol-conjugated HA. In the first study, [bib_ref] Chasing bacteria within the cells using levofloxacin-loaded hyaluronic acid nanohydrogels, Montanari [/bib_ref] formulated LVF-loaded self-assembled HA-NGs (LVF-NGs) targeting intracellular bacteria . Levofloxacin was loaded into the NGs with a drug loading efficiency of 5% and average size of 155 nm. The antibacterial studies assessed both extracellular and intracellular activity against Staphylococcus aureus, MRSA, and Pseudomonas aeruginosa. The extracellular in vitro results showed no improvement in the minimum inhibitory concentration (MIC) values of LVF-NGs compared with the free LVF. While the intracellular studies on the human ovarian cancer cell line (HeLa) showed that LVF-NGs significantly killed intracellular bacteria as compared to free LVF (which exhibited no activity). Interestingly, this study represents the first-ever report of antibiotic-loaded self-assembled NGs [bib_ref] Chasing bacteria within the cells using levofloxacin-loaded hyaluronic acid nanohydrogels, Montanari [/bib_ref]. In their second study, they explored the use of HA-NGs to specifically target infected human keratinocytes. The NGs were preloaded with either GM or LVF. The mean diameters of GM-loaded HA-NGs (GM-NGs) and LVF-NGs were 250 and 350 nm, respectively, with a drug loading efficiency of 40% and 11.4%, respectively. Moreover, the antibacterial studies showed that both formulations had the same MIC and minimum bactericidal concentration (MBC) values as free antibiotics against extracellular S. aureus. However, intracellularly, NGs significantly enhanced the antibacterial activity of LVF, while GM and GM-NGs showed comparable activity. This was explained by the ability of HA-NGs to shift the intracellular fate of LVF from the cytosol to the lysosomes, therefore enhancing their activity. In contrast, GM, an antibiotic mainly accumulates in lysosomes, exhibits considerable intracellular activity even without loading into NGs [bib_ref] Hyaluronan-based Nanohydrogels for targeting intracellular S. aureus in human keratinocytes, Montanari [/bib_ref]. Although the drug loading efficiency in the first study was lower than in the second, both studies demonstrated a substantial improvement in intracellular antibacterial activity F I G U R E 4 Schematic illustration of LVF-NGs for targeting intracellular infections [bib_ref] Chasing bacteria within the cells using levofloxacin-loaded hyaluronic acid nanohydrogels, Montanari [/bib_ref] with no change in extracellular activity. Furthermore, unlike their former study, this study investigated biosafety and drug release studies, which are critical in nanoantibiotics characterization. Further in vivo studies are required in order to provide a step forward toward reaching the clinical stage. Another antibiotic azithromycin (AZ) was explored in a study by P. D. [bib_ref] Hyaluronic acid-based nanogels improve in vivo compatibility of the anti-biofilm peptide DJK-5, Kłodzi Nska [/bib_ref] where they evaluated and compared HA-NGs and coated poly(lacticco-glycolic acid) NPs (T-PLGA NPs) in a head-to-head comparison to determine the relative benefits and drawbacks of each as promising delivery system . The microfluidic technique was used to prepare the HA-NGs, followed by AZ loading. Compared to T-PLGA NPs, AZ-loaded NGs were larger in size and had a higher EE %; nevertheless, both had a diameter of less than 200 nm. The in vitro cytotoxicity assay revealed that loaded HA-NGs are relatively nontoxic to HepG2 and lung epithelial cells (Calu-3), but T-PLGA NPs exhibit some toxicity. Although both delivery systems enhanced the antibacterial and anti-virulence activity of AZ, AZ-loaded NGs enhanced the penetration of AZ into biofilms and also eliminated preformed biofilms more effectively than T-PLGA NPs (S. N. [bib_ref] Utilizing nanoparticles for improving anti-biofilm effects of azithromycin: A head-to-head comparison of..., Kłodzi Nska [/bib_ref]. Nevertheless, this study lacked in vitro drug release and in vivo antibacterial studies which may add further biological validation to such promising nanosystem. A recent study by Yuda [bib_ref] Composite inclusion complexes containing hyaluronic acid/chitosan nanosystems for dual responsive enrofloxacin release, Liu [/bib_ref] further expanded the performance of HA-NGs in antibacterial delivery. They reported novel targeted "on-demand" delivery of composite nanosystems based on the triple controlled release of inclusion composites (IC), polymeric NPs, and HA-NGs. The ultimate objective of this investigation was to efficiently eliminate S. aureus. In this regard, enrofloxacin was incorporated into IC and then dispersed into poloxamer 188 coating NGs formulated by ionic complexation between chitosan (CS) and HA. Nanosystems with mean EE%, sizes, and PDI of 95.4%, 118.8 nm, and 0.26, respectively, were prepared. By incorporating IC into the HA-NGs, the nanosystem gained multifunctional characteristics by releasing enrofloxacin in a dual pH/HAaseresponsive manner at the infection site, therefore reducing premature release of the drug. Furthermore, the composite nanosystems were able to be absorbed onto the surface of S. aureus, resulting in increased antibacterial activity [bib_ref] Composite inclusion complexes containing hyaluronic acid/chitosan nanosystems for dual responsive enrofloxacin release, Liu [/bib_ref]. Unlike the above HA-NGs, this is the first stimuli-responsive HA-NGs encapsulated with an antibacterial drug. This study established that by cleverly integrating IC with HA-NGs stabilized with poloxamer 188, the developed composite nanosystems may reduce premature drug release and improve targeting to S. aureus. Therefore, this study may serve as a fruitful approach to tackle the treatment difficulties associated with S. aureus as well as other bacterial infections. In conclusion, the studies described in this paragraph could lay the ground for expanding the use of HA-NGs to enhance antibiotics delivery due to their ability to enhance intracellular antibacterial activity, alter the fate of several antibiotics, provide on-site antibiotic release, and, most importantly, eliminate side effects associated with loaded antibiotics.
Antimicrobial peptides are molecules that have been identified as among the most promising therapeutic candidates for the treatment of bacterial infections. However, these agents showed low stability and bioavailability as well as a high toxicity profile [bib_ref] Nanomedicines for the delivery of antimicrobial peptides (amps), Teixeira [/bib_ref]. Hyaluronic acid-based nanogels have been studied recently as a promising nanoplatform for the delivery of antimicrobial peptides and peptidomimetics (P. D. [bib_ref] Hyaluronic acid-based nanogels improve in vivo compatibility of the anti-biofilm peptide DJK-5, Kłodzi Nska [/bib_ref] [bib_ref] Biopolymer nanogels improve antibacterial activity and safety profile of a novel lysine-based..., Kłodzi Nska [/bib_ref] [bib_ref] Delivery of LLKKK18 loaded into self-assembling hyaluronic acid nanogel for tuberculosis treatment, Silva [/bib_ref]. To this end, [bib_ref] Delivery of LLKKK18 loaded into self-assembling hyaluronic acid nanogel for tuberculosis treatment, Silva [/bib_ref] reported HA-NGs loaded with LLKKK18 F I G U R E 5 Head-to-head comparison of HA-NGs and coated poly (lactic-co-glycolic acid) NPs for AZ delivery (Kłodzi nska, [bib_ref] Hyaluronic acid-based nanogels improve in vivo compatibility of the anti-biofilm peptide DJK-5, Kłodzi Nska [/bib_ref] peptide to cure pulmonary mycobacterial infections . HA-NGs were prepared via cross-linking of HA with a thiolated alkyl chain and produced NGs with a large diameter (533 nm), small PDI (0.1), and high EE% (70%). The biocompatibility of the loaded NGs in bone marrow-derived macrophages (BMMΦ) was evaluated using two distinct assays. Both assays revealed that loaded NGs had no harmful impact on BMMΦ at concentrations up to 100 μM, which is more than 20-fold the toxic concentration of free peptide. Furthermore, activated macrophages overexpress the CD44 receptor, allowing infected macrophages to successfully uptake loaded HA-NGs, resulting in selective targeting of LLKKK18 to mycobacteria that reside in intracellular compartments. In vitro incubation of macrophages with LLKKK18-loaded NGs reduced Mycobacterium avium and Mycobacterium tuberculosis intracellular levels. Notably, in vivo studies with LLKKK18-loaded NGs demonstrated a substantial reduction in infection levels in mice infected with M. avium or M. tuberculosis [bib_ref] Delivery of LLKKK18 loaded into self-assembling hyaluronic acid nanogel for tuberculosis treatment, Silva [/bib_ref]. This study established a strong in vitro-in vivo correlation, thereby validating the potential of LLKKK18-loaded HA-NGs to eliminate mycobacteria efficiently. Moreover, HA-NGs enhanced stability and minimized cytotoxicity, while potentiating peptides targeting the main sites of infection. Continuing the research study into the delivery of antimicrobial peptides and peptidomimetics, Kłodzi nska and his group developed HA-NGs by conjugating HA with octenyl succinic anhydride. In their first study, they explored the ability of the HA-NGs to enhance the antibacterial activity and biosafety of antimicrobial peptidomimetics (lysine-based α-peptide/β-peptoid). This was the first reported biopolymer NGs incorporating antimicrobial peptidomimetics. The NGs were prepared using the microfluidic chip technique. Design of Experiments (DoE) was used to assess the optimal parameters for preparation of the NGs; accordingly, 12 formulations were prepared (F1-F12). F10 was the most encouraging in this study with size, zeta potential (ZP), and EE% of 175 nm, À16 mV, and 88%, respectively. The HA-NGs encapsulating the peptidomimetic LBP-3 exhibited no hemolysis and reduced LBP-3 cytotoxicity in hepatocytes (HepG2 cells) when compared to free peptidomimetic, whereas the MIC and minimum biofilm inhibitory concentration (MBIC) values of LBP-3 against P. aeruginosa remained unchanged. Only F10 exhibited the potential to delay the regrowth of P. aeruginosa [bib_ref] Biopolymer nanogels improve antibacterial activity and safety profile of a novel lysine-based..., Kłodzi Nska [/bib_ref]. Despite the significant enhancements, this study lacked investigations into morphology, the peptidomimetic release from the NGs, and in vivo studies, which would have further validated the ability of HA-NGs to effectively deliver LBL-3 and other peptidomimetics. Another peptide-loaded HA-NG was formulated by the same group, this time focusing mainly on improving peptide biosafety. Kłodzi nska et al. (2018) evaluated the ability of HA-NGs to encapsulate the antibiofilm peptide DJK-5 with the overall goal of improving the in vivo biosafety of DJK-5s . For this purpose, NGs with sizes ranging from 174 to 194 nm encapsulating 33%-60% of DJK-5 were prepared. This EE% was relatively low compared to their prior LBP-loaded HA-NGs, where the EE% reached 88%. The in vitro release studies showed that 80% of the loaded peptide was released within 5 h. The in vivo biosafety of DJK-5 and DJK-loaded NGs was evaluated using a P. aeruginosa abscess model following both IV and subcutaneous administration. Encapsulation of DJK-5 in HA-NG reduced its toxicity by 2-fold following IV administration, and by 4-fold after subcutaneous injections, compared to the free peptide, without compromising the antiabscess efficacy of DJK-5 (P. D. [bib_ref] Hyaluronic acid-based nanogels improve in vivo compatibility of the anti-biofilm peptide DJK-5, Kłodzi Nska [/bib_ref]. In comparison to their previous study, all essential physiochemical and biological characterization stages (in vitro and in vivo) were carried out perfectly in this study, hence supporting the use of HA-NGs to improve the safety of antibacterial and anti-biofilm peptides. Taken together, the findings of these studies support the application of HA-NGs to enhance antimicrobial peptides and peptidomimetics delivery owing to their ability to improve antibiofilm activity, antibiotic localization on-site, and to eliminate the F I G U R E 6 Schematic illustration of LLKKK18-loaded HA-NGs targeting intracellular mycobacteria [bib_ref] Delivery of LLKKK18 loaded into self-assembling hyaluronic acid nanogel for tuberculosis treatment, Silva [/bib_ref] cytotoxicity of loaded drugs. Finally, various tailored HA-NGs and formulation conditions can be explored to determine the most appropriate delivery system for peptide and peptidomimetics and allow their efficient use in the clinic.
In addition to the above-mentioned applications of HA-NGs, HA-NGs have been also formulated without drug loading. Since S. aureus is able to persist within keratinocytes and phagocytic cells being protected from both the immune system and a variety of antibacterials, to overcome these obstacles, nano-formulations that enable targeted therapies against intracellular S. aureus might be developed. In this regard, [bib_ref] Biodistribution and intracellular localization of hyaluronan and its nanogels. A strategy to..., Montanari [/bib_ref] formulated blank HA-NGs (with no drug) to investigate the biodistribution and intracellular localization of HA-NGs (compared to HA) following intravenous (IV) injections (in mice) and topical administrations (in ex vivo human skin). All HA and NGs used in this study were labeled using rhodamine B isothiocyanate without loading of any drug. Following IV injection, HA accumulated considerably in the skin, which corroborates the findings of [bib_ref] Targetability of hyaluronic acid nanogel to cancer cells: In vitro and in..., Pedrosa [/bib_ref] , who showed that HA-NGs exhibited high accumulation in the skin but with a blood clearance rate slower than HA [bib_ref] Targetability of hyaluronic acid nanogel to cancer cells: In vitro and in..., Pedrosa [/bib_ref]. Furthermore, after topical administration in an ex vivo human skin model, neither HA nor its NGs penetrated the intact skin. In contrast, in both barrier-disrupted skin and mechanically produced wounds, HA was taken up by dermis cells in a higher percentage than HA-NGs; however, in both cases, the uptake is mediated by the CD44 receptor. Then, using human dermal fibroblasts and murine macrophages, they investigate the intracellular fate of HA and its NGs in dermal cells. Interestingly, the data obtained suggest the potential of both HA and HA-NGs to accumulate in dermal cell lysosomes, which is in accordance with their previous work [bib_ref] Hyaluronan-based Nanohydrogels for targeting intracellular S. aureus in human keratinocytes, Montanari [/bib_ref] [bib_ref] Biodistribution and intracellular localization of hyaluronan and its nanogels. A strategy to..., Montanari [/bib_ref].
From all the papers discussed in this subsection, the loaded molecule, molecular weight of HA, cross-linked hydrophobic moiety, targeted bacteria, and in vitro/in vivo status information were extracted and then presented in . So far, only four types of hydrophobic moieties have been used to formulate HA-NGs. Therefore, in the future, other hydrophobic moieties such as methacrylate molecules, fatty acids, acetic anhydride, as well as active antibacterial compounds, could be investigated. In summary, the findings presented in this subsection confirm the potential of HA-NGs as a promising nanosystem for antibacterial delivery, with the capability to offer selective targeting while also improving bioavailability, penetration, and antibacterial activity. Moreover, HA-NGs provide controlled release behavior and reduce the toxicity of loaded compounds.
## | hyaluronic acid-ion complex nps
Owing to its anionic nature, HA tends to form ion complex NPs with cationic small molecules or polymers through polyelectrolyte complexation, which is based on the ability of polyelectrolytes to cross-link in the presence To date, two such nanosystems have been reported in the literature for potential applications in the prevention as well as therapy of bacterial infections. Electrostatic assembly of HA with an arginine-based compound, called ethyl-α-N-lauroyl-L-arginate hydrochloride (LAE), to build antimicrobial nanofilms with potential application in food preservation and designing of antiseptic medical devices has been reported by [bib_ref] Ionic coupling of hyaluronic acid with ethyl N-lauroyl L-arginate (LAE): Structure, properties..., Gamarra [/bib_ref]. Thin films prepared from HA-LAE by casting exhibited a smectic-like structure based on an ordered arrangement of LAE and HA layers with a nanometric periodicity of 3.8-3.9 nm. In such systems, HA acts as a matrix for holding and regulating the antimicrobial release from LAE, therefore reducing its potential toxicity and prolonging its activity time compared to the neat LAE. Thereafter, the rate of LAE release from the nanofilm was studied at two pH levels (5.5 and 7.4). Within 4-8 h of incubation, the dissociation equilibrium was reached. At that time, the concentration of LAE was much greater at pH 7.4 than at pH 5.5, and the quantity of free LAE accumulated significantly increased with the LAE: HA ratio in the nanofilm. The antibacterial activity of LAE-HA at pH 7.4 was explored using the gram-positive (L. monocytogenes and S. aureus) and gram-negative (S. enterica and E. coli) bacteria. The results indicated significant bactericide activity against both gram-positive and gramnegative bacteria, with the highest activity observed against S. enterica. Although further in vitro biocompatibility and in vivo antibacterial studies are required to support their hypothesis, the findings of this research demonstrate that HA has the potential to form arginine-based nanofilms with inherent antibacterial activity as an alternative antiseptic and preservation agent.
Current TB therapies yield low drug loading and/or accumulation in the lung alveoli. Therefore, [bib_ref] Aerodynamic properties and in silico deposition of isoniazid loaded chitosan/thiolated chitosan and..., Mukhtar [/bib_ref] developed dry powder inhalers containing CS/thiolated CS (TC) in conjugation with HA as an innovative drug delivery strategy that will allow the antibacterial agent to penetrate deeper lung tissues. The ionic gelation approach was used to produce CS-HA and TC-HA NPs, which were subsequently loaded with isoniazid (INH). Several formulations were developed using DoE, and the average particle size of optimized formulations was 300.2 nm and 342.1 nm with EE% of 90.23% and 92.58% for CS-HA and TC-HA NPs, respectively. The NPs carried a cationic ZP, making them favorable to target the negatively charged surface of macrophages. Additionally, incorporating HA into NPs enables targeting of macrophage CD44 surface receptors, therefore activating a synergistic strategy. Later the in vitro drug release of both NPs, a controlled release pattern was revealed over time, with approximately more than 50% of INH released from both. The MTT cell viability assay was carried out on adenocarcinoma human alveolar basal epithelial (A549) cells. Isoniazid-loaded NPs were found to be biocompatible with the A549 cell line and had no cytotoxicity at concentrations ranging from 0.1 to 0.5 mg/ml. Moreover, the in vitro drug accumulation followed by Andersen Cascade measurement was promising and aligned with the in silico particle deposition model for TB. Both formulations exhibited a greater amount of particle accumulation in the alveolar area, which was considered the site of action. Collectively, the results of this study proposed a novel HA-NC that actively delivers anti-TB medication to the site of action while also improving F I G U R E 8 Scheme illustrating the ionic complexation of LAE-HA and its antibacterial activity [bib_ref] Ionic coupling of hyaluronic acid with ethyl N-lauroyl L-arginate (LAE): Structure, properties..., Gamarra [/bib_ref] the pulmonary drug deposition; however, antibacterial studies are still required to confirm the antibacterial effects of this nanosystem [bib_ref] Aerodynamic properties and in silico deposition of isoniazid loaded chitosan/thiolated chitosan and..., Mukhtar [/bib_ref].
The above-mentioned data demonstrates that the complexation of cationic molecules with HA for targeting bacteria can lead to controlled drug release and consequent inhibition of bacterial growth in infected areas [bib_ref] Ionic coupling of hyaluronic acid with ethyl N-lauroyl L-arginate (LAE): Structure, properties..., Gamarra [/bib_ref] [bib_ref] Aerodynamic properties and in silico deposition of isoniazid loaded chitosan/thiolated chitosan and..., Mukhtar [/bib_ref].
## | hyaluronic acid-polymersomes
In recent decades, polymersomes have garnered remarkable attention for antibacterial delivery owing to their colloidal stability, flexible membrane characteristics, and capacity to encapsulate or integrate a wide variety of drugs and molecules. Polymersomes are defined as self-assembling nano-vesicles composed of amphiphilic block copolymers. Along with its application in antibacterial delivery, polymersomes are also efficiently utilized to encapsulate therapeutic molecules such as anti-inflammatory, anticancer, antifungal, as well as DNA and RNA fragments either in the aqueous core or integrated into their membrane (J. S. [bib_ref] Polymersomes for drug delivery: Design, formation and characterization, Lee [/bib_ref] [bib_ref] Emerging era of "somes": Polymersomes as versatile drug delivery carrier for cancer..., Sharma [/bib_ref].
The application of HA-based polymersomes as diagnostic tools is recognized as a popular area of research. In this regard, [bib_ref] Enzyme degradable Polymersomes from hyaluronic acid-block-poly(ε-caprolactone) copolymers for the detection of enzymes..., Haas [/bib_ref] developed smart enzyme-sensitive polymersomes to detect pathogenic bacteria such as S. aureus in wounds and burn environments. They describe a novel HAase-responsive amphiphilic block copolymer system based on HA and polycaprolactone (PCL). These polymersomes were constructed using an inversed solvent-shift technique using chloroform and water followed by loading with various fluorescent dyes and antibacterial agents, resulting in HA-PCL assemblies varying in size from 50 to 400 nm, depending on the loaded molecules. The polymersomes, as demonstrated by two different imaging microscopes, are vesicular in nature. In the presence of HAase, the in vitro release of the loaded molecule was studied using the dynamic light scattering technique; a reduction in diameter was observed as the reaction time increased, indicating that the polymersomes were successfully degraded by HAase. Additionally, when loaded with a fluorogenic dye, fluorescence spectroscopy revealed a spontaneous light-up in the presence of HAase and chymotrypsin. However, the degradation of the HA-PCL vesicles was studied using bovine HAase instead of bacterial HAase, which are different. This work supports the potential of HA-PCL polymersomes as a smart indicator system for identifying harmful bacteria in wounds. This approach may be exploited to produce innovative detective and therapeutic HA-polymersomes targeting various HAase-producing bacteria [bib_ref] Enzyme degradable Polymersomes from hyaluronic acid-block-poly(ε-caprolactone) copolymers for the detection of enzymes..., Haas [/bib_ref].
In addition to the above-mentioned application in infection diagnosis, HA-polymersomes have been recently explored for antibacterial delivery. For this purpose, our research group developed novel HA-oleylamine (HA-OLA) polymersomes to enhance the antibacterial activity of vancomycin (VCM) against MRSA . Following the synthesis of novel HA-OLA, the biosafety of the conjugates was confirmed using the MTT assay on three different human cell lines: embryonic kidney cells, cervix adenocarcinoma cells, and human breast adenocarcinoma cells. A probe ultrasonication technique was used to prepare the polymersomes, which were subsequently loaded with VCM. The optimized VCM-loaded HA-OLA polymersomes (VCM-PS6) had an average diameter of less than 250 nm and an EE% of 43.12%. These polymersomes showed slow and sustained release of VCM over 3 days. The in vitro MIC studies against MRSA showed that VCM-PS6 had fourfold the activity of bare VCM. Furthermore, the synergistic action of VCM and HA-OLA against MRSA was reported. Additionally, when samples were treated at an MIC of 1.95 g/ml, flow cytometry demonstrated that VCM-PS6 resulted in 1.8-fold more dead MRSA cells than free VCM. A bacterial membrane rupture test revealed that VCM-loaded polymersomes were more effective in disrupting MRSA membranes than free VCM. These findings suggest that VCM-PS6 was more effective against MRSA than free VCM in all antibacterial tests. In summary, these findings confirm that HA-OLA polymersomes have the potential to be promising antibacterial nanosystems to tackle resistant bacterial strains.
Of the HA-NPs studied, and to the best of our knowledge, only the abovementioned polymersomes have been designed, constructed, and studied for their potential in antibacterial therapy. These polymersomes have considerable therapeutic and diagnostic potential for bacterial infections, which can be significantly increased by mimicking the natural substrate of B-HAase, thus enabling the self-regulated release of antimicrobials. Accordingly, there is a gap in research for designing innovative multifunctional HA-based polymersomes for dual detection and treatment of HAaseproducing microorganisms.
## | other hyaluronic acid-based nanocarriers
In addition to all the HA-NPs mentioned above, others have been reported to either eradicate or prevent bacterial infections. These include HA-based micelles, nanocapsules, nanoplexes, nanofibers (NFs), and nanofilms [bib_ref] Enzyme responsive hyaluronic acid nanocapsules containing polyhexanide and their exposure to bacteria..., Baier [/bib_ref] [bib_ref] Hyaluronic acid-tocopherol succinate-based self-assembling micelles for targeted delivery of rifampicin to alveolar..., Gao [/bib_ref] [bib_ref] Electrospun PVA/-hyaluronic acid/L-arginine nanofibers for wound healing applications: Nanofibers optimization and in..., Hussein [/bib_ref] [bib_ref] Hyaluronan/tannic acid nanoparticles via catechol/Boronate Complexation as a smart antibacterial system, Montanari [/bib_ref].
Nanometer-sized HA systems can selectively deliver antibacterial drugs into infection sites via interaction with overexpressed CD44 receptors in certain cells, such as macrophages, which allow effective and enhanced localization of loaded antibacterial drugs in the site of infection. For this purpose, two studies have reported HA-NPs for targeted delivery of antibacterial agents, namely rifampicin (RIF) and tannic acid (TA), to eradicate intracellular bacterial infections [bib_ref] Hyaluronic acid-tocopherol succinate-based self-assembling micelles for targeted delivery of rifampicin to alveolar..., Gao [/bib_ref] [bib_ref] Hyaluronan/tannic acid nanoparticles via catechol/Boronate Complexation as a smart antibacterial system, Montanari [/bib_ref]. [bib_ref] Hyaluronic acid-tocopherol succinate-based self-assembling micelles for targeted delivery of rifampicin to alveolar..., Gao [/bib_ref] presented a TB-targeted drug delivery system based on HA-functionalized micelles loaded with RIF; a drug widely used for the treatment of TB. Tocopherol succinate was grafted onto HA to generate a hydrophobic moiety enabling self-assembly. They synthesized an HA-tocopherol succinate conjugate (HA-TS), which self-assembled into HA-micelles with average particle diameter ranging from 212 to 294.6 nm. The RIF was incorporated with high drug loading capacity exceeding 70%. The in vitro release studies conducted in two different environments (PBS at pH 7.4 and acetate buffer at pH 5.2) revealed the sustained release of RIF at both pH levels; however, a slower release rate of RIF was observed in acidic media compared to the release rate at physiologic pH. The in vitro biocompatibility on murine alveolar macrophage (MH-S cells) was examined using the MTT assay. The results showed that the loaded micelles exhibited slightly higher cytotoxicity to MH-S cells compared to free RIF within the first 2 days, while no apparent cytotoxicity was observed for blank HA-TS micelles with concentrations between 25 and 250 μg/ml. As expected, the loaded micelles were more efficiently taken up by MH-S cells than the free drug. This uptake occurred via phagocytosis and CD44-mediated endocytosis, and it was time-and dose-dependent. Additionally, the HA-TS micelles may promote the release of pro-inflammatory cytokines, which could improve the antibacterial action of RIF. In another study, [bib_ref] Hyaluronan/tannic acid nanoparticles via catechol/Boronate Complexation as a smart antibacterial system, Montanari [/bib_ref] designed and synthesized HA-aminophenyl (HA-APBA) conjugated NPs with a pH-cleavable linkage for smart delivery of TA to intracellular infection sites. A reaction between HA-APBA and TA yielded NPs with a size smaller than 400 nm, which was dependent on the molecular weight of HA. The boronate ester bond made the NPs stable at physiological pH, while hydrolyzing in an acidic environment allowed TA release within endosomal compartments. The presence of HA enhanced the selectivity of these NPs toward bacteria that colonize within macrophages. Therefore, RAW 246.7 macrophages were used as the model to assess the cytotoxicity of the NPs. The results indicated that HA-APBA alone exhibited slight cytotoxicity, TA significantly reduced cell viability at 56 μg/ml only when combined with AA, but neither AA nor TA appeared to have any impact when used alone. The loaded NPs exhibited similar toxicity to the TA-AA combination. Finally, three bacterial models were employed in the antibacterial study: S. aureus, MRSA, and E. coli. Both the loaded NPs and TA had no significant action against E. coli, while activity against sensitive and resistant S. aureus were significantly greater. Interestingly, the NPs demonstrated nearly comparable antibacterial activity to the TA/ascorbic acid combination, confirming that the boronate complexation retains the reduced (catechol) form of TA and that the NPs indeed release this active form of TA in an acidic medium [bib_ref] Hyaluronan/tannic acid nanoparticles via catechol/Boronate Complexation as a smart antibacterial system, Montanari [/bib_ref]. The outcomes of these two studies confirm the potential of HA-NPs to eradicate intracellular bacteria; however, in vivo antibacterial studies of both nanocarriers are required to validate their theory.
Hyaluronic acid-NPs can also act as carriers for the purpose of wound dressing and tissues regeneration. Wound dressing has been widely studied in order to develop an "optimal" technique that promotes rapid recovery while minimizing scarring. For this reason, HA-NPs, with their diverse physicochemical properties, are a promising research domain for wound dressing and healing [bib_ref] Hyaluronic acid as potential carrier in biomedical and drug delivery applications, Prajapati [/bib_ref]. In this regard, [bib_ref] Enzyme responsive hyaluronic acid nanocapsules containing polyhexanide and their exposure to bacteria..., Baier [/bib_ref] developed novel HA-based nanocapsules encapsulating an antibacterial agent that cleaves specifically in the presence of HAase. The inverse miniemulsion approach was used to prepare stable cross-linked HA-based nanocapsules loaded with polyhexanide (PH). The obtained nanocapsules had diameters of 360-380 nm and a negative surface. Furthermore, the release of the entrapped PH was quantified by measuring the absorbance of the model dye that was released in response to HAase. As expected, control nanocapsules composed of starch derivative or only PH as a shell material did not exhibit any enzyme-responsive release. The in vitro antibacterial studies in the presence of S. aureus and E. coli revealed that both the PH containing HAbased (HA-PH-NCs) and the PH-based nanocapsules (PH-NCs) had antibacterial activity. Additionally, these nanocapsules were found to be bactericidal against both S. aureus and E. coli. However, the MIC for S. aureus was lower, reaching 62.5 μg/ml for both nanocapsules, while their activity against E. coli was significantly reduced with MIC values of 250 and 125 μg/ml for HA-PH-NCs and PH-NCs, respectively. This is explained by the capacity of S. aureus to generate and release HAase, which cleaves HA and promotes bacterial dissemination [bib_ref] Enzyme responsive hyaluronic acid nanocapsules containing polyhexanide and their exposure to bacteria..., Baier [/bib_ref]. Using another strategy for wound healing, [bib_ref] Electrospun PVA/-hyaluronic acid/L-arginine nanofibers for wound healing applications: Nanofibers optimization and in..., Hussein [/bib_ref] reported novel L-arginine-loaded HA-based NFs as a bioactive wound dressing with inherent wound healing properties. To address the poor mechanical properties of HA-based NFs, they strengthened them by including cellulose nanocrystals (CNC) as a bio-composite, which were then loaded with L-arginine. An electrospinning technique was used to fabricate the HA-based NFs. Thereafter, HA-based NFs showed round-shaped nanoparticles with an average size of 120 nm. It was observed that this size fell to 60 nm after incorporation of the CNC, then increased again to 100 nm after incorporation of L-arginine. This result was in accordance with Reesi et al., who observed that the size of L-arginine-loaded lignin NFs increased from $20-50 nm to $100-250 nm [bib_ref] A novel lignin-based nanofibrous dressing containing arginine for wound-healing applications, Reesi [/bib_ref]. The loaded HA-CNC NFs (PVA/HA/CNC/L-arginine NFs) showed slow-release behavior, which led to a sustained elevation in the level of arginine in plasma, thus reducing its adverse effects. Interestingly, the slow release of arginine was found to be a trigger for nitric oxide (NO) signaling, thus accelerating wound closure. Using normal human skin melanocyte (HFB-4) and lung fibroblast (WI38) cell lines, in vitro bio-evaluation of NFs revealed that PVA/HA/CNC/L-arginine NFs displayed favorable hemocompatibility, elevated protein adsorption, and excellent proliferation and adhesive capability on HFB-4 cells. Regarding the antibacterial studies, PVA/HA/CNCs/L-arginine demonstrated acceptable antibacterial action, particularly against Klebsiella pneumonia, a common acute pathogen that causes skin infection [bib_ref] Electrospun PVA/-hyaluronic acid/L-arginine nanofibers for wound healing applications: Nanofibers optimization and in..., Hussein [/bib_ref]. Overall, the data presented in this paragraph confirms that such functionalization (using HA) can represent a robust technique for developing multifunctional HA-nanosystems for wound dressing and infection control.
Infections associated with biomedical devices contribute significantly to the growing problem of nosocomial infections, imposing a considerable societal and economic impact [bib_ref] Nosocomial infections: Epidemiology, prevention, control and surveillance, Khan [/bib_ref]. Therefore, a HA-nanofilm was developed to reduce this bioburden. Hernandez-Montelongo et al. (2016) described the optimal fabrication of layer-by-layer HA-CS nanofilms (HA-CS) as an antibacterial coating technique for preventing nosocomial infections from biomedical devices. Hyaluronic acid created a soft, well-hydrated, nontoxic coating for this nanosystem, whereas CS demonstrated antibacterial properties. The optimization of HA-CS nanofilm synthesis was based on altering the pH values of the biopolymer stem solutions and, as a result, the degree of ionization. Moreover, by increasing the number of HA-CS bilayers, the surface density of amino groups, which is associated with antibacterial activity, was enhanced. The antibacterial activity of HA-CS nanofilms was then investigated against S. aureus and P. aeruginosa using the spread plate counting method. The HA-nanofilms demonstrated promising results against S. aureus, which was highly sensitive to the surface of nanofilms fabricated at various pH values compared to the control sample (silicon substrate). This sensitivity was attributed to the free ammonium group of CS on the top HA-CS layer of the nanofilms. However, when P. aeruginosa cultures were cultured for 4 h at pH 4.5 and pH 3, the HA-CS nanofilms had no significant antibacterial action, even though the bacteria were grown at slightly lower rates when compared to the control. Thus, HA-CS nanofilms in the optimized conditions of synthesis shown here are an excellent alternative as antibacterial surfaces against S. aureus. Nevertheless, their use against P. aeruginosa would require the addition of antimicrobial agent, such as bioactive peptides, silver nanoparticles, and other suitable biocides [bib_ref] Hyaluronan/chitosan nanofilms assembled layer-by-layer and their antibacterial effect: A study using Staphylococcus..., Hernandez-Montelongo [/bib_ref].
The outcomes of all the studies mentioned in this section validate the potential application of HA as a promising polymer to formulate HA-NPs for preventing and eradicating bacterial infections. These NPs can enhance the localization of antibacterial agents by providing smart targeted antibacterial therapy, lowering adverse effects of loaded drugs, and regulating the release of the loaded antibacterial agents.
## | hyaluronic acid-coated nanocarriers
The surface of nanocarriers can be functionalized with a variety of polymers depending on the surface characteristics of nanocarriers [bib_ref] Hyaluronic acid coated chitosan nanoparticles reduced the immunogenicity of the formed protein..., Almalik [/bib_ref] [bib_ref] Current trends in using polymer coated gold nanoparticles for cancer therapy, Muddineti [/bib_ref]. Coating nanocarriers with HA may improve their biological characteristics and biocompatibility, boost their antibacterial activity, and enable controlled site-specific drug release. Additionally, coating with HA can facilitate the ability of nanocarriers to evade immune cell detection and uptake by the reticuloendothelial system, while also significantly improving target recognition and localization. Therefore, HA-coated nanocarriers have been widely applied for targeted drug delivery in the treatment and diagnosis of cancer. However, the use of HA to coat antibacterial-loaded NPs is still in its infancy [bib_ref] An update on polysaccharide-based nanomaterials for antimicrobial applications, Arora [/bib_ref] [bib_ref] Hyaluronic acid-coated nanomedicine for targeted cancer therapy, Kim [/bib_ref] [bib_ref] Surface modification of nano-drug delivery systems for enhancing antibiotic delivery and activity, Osman [/bib_ref] [bib_ref] Hyaluronan-modified nanoparticles for tumor-targeting, Sakurai [/bib_ref]. In this section, we discuss and critically elaborate on the antibacterial nanocarriers that have been surface-modified using HA, which are summarized in . We have systematically categorized these nanocarriers according to the function of the HA layer into (i) CD44 targeting nanocarriers; (ii) HAase-responsive nanocarriers; and (iii) others.
## | hyaluronic acid-coated nanocarriers targeting cd44 receptors
The primary receptor for HA, which is CD44, is overexpressed in a subset of cells, such as the majority of cancer cells and activated inflammatory cells (including macrophages). This can boost the selectivity of HA-NPs through CD44-mediated uptake, allowing HA to serve as a homing device (C. [bib_ref] The biology and role of CD44 in cancer progression: Therapeutic implications, Chen [/bib_ref] [bib_ref] Hyaluronic acid-based nanomaterials for cancer therapy, Kim [/bib_ref] [bib_ref] Emerging role of CD44 receptor as a potential target in disease diagnosis:..., Pandey [/bib_ref]. Among the several HA-coated nanocarriers, inorganic NPs are gaining considerable interest for antibacterial delivery due to their high stability, simplicity of functionalization, and inertness. Apart from improving the targeting of these nanocarriers, HA also regulates and controls drug release and increases the colloidal stability (Y. [bib_ref] Inorganic nanoparticle-based drug codelivery nanosystems to overcome the multidrug resistance of cancer..., Chen [/bib_ref] [bib_ref] Layer-by-layer decorated nanoparticles with tunable antibacterial and Antibiofilm properties against both gram-positive..., Ivanova [/bib_ref] [bib_ref] Hyaluronic acid-templated Ag nanoparticles/graphene oxide composites for synergistic therapy of bacteria infection, Ran [/bib_ref] [bib_ref] Inorganic nanomedicine-part 2, Sekhon [/bib_ref].
Macrophages, the predominant site of intracellular infection, shield intracellular bacteria from the immune cells, therefore limiting the action of antibiotics. These macrophages express CD44 receptors. Thus, modifying nanocarriers with HA enables the development of therapeutic NPs that may actively target these macrophages, enhancing the release and localization of the loaded antibiotics [bib_ref] The role of CD44 in disease pathophysiology and targeted treatment, Jordan [/bib_ref] [bib_ref] Hyaluronic acid targeting of CD44 for cancer therapy: From receptor biology to..., Mattheolabakis [/bib_ref] [bib_ref] Recent advances in polymeric nanoparticle-encapsulated drugs against intracellular infections, S Anchez [/bib_ref]. To this end, X. [bib_ref] The highly efficient elimination of intracellular bacteria: Via a metal organic framework..., Zhang [/bib_ref] and designed pH-responsive HA-coated metal-organic frameworks (MOFs) loaded with antibiotics for targeted elimination of intracellular bacteria. Zeolitic imidazolate framework-8 (ZIF-8) was chosen as an appropriate MOF owing to its acceptable stability, minimal cytotoxicity, and acid sensitivity, which facilitates drug release at the site of infection. The antibiotics were loaded into the ZIF-8 formed by the self-assembly of the zinc ion and organic bond via an aqueous phase self-assembly approach. Then, HA was added to decorate the surface of the nanocarrier via the coordination effect between its carboxylic acid groups and Zn 2+ in ZIF-8. In both studies, the nanocarriers can be taken up by cells via an HA-mediated pathway owing to CD44 receptors on the macrophage surface [bib_ref] The highly efficient elimination of intracellular bacteria: Via a metal organic framework..., Zhang [/bib_ref]. X. Zhang et al. encapsulated tetracycline in HA-coated ZIF-8 to create tetracycline-loaded HA-ZIF-8 (TZH) as shown in . The in vitro drug release study was carried out at pH 5.5 and 7.4. As expected, both tetracycline and Zn 2+ were released rapidly from TZH at pH 5.5 compared to the physiological pH. Both in vitro and in vivo antibacterial results demonstrated that ZIF-8 and tetracycline encapsulated in TZH had a synergistic effect, with an intracellular bacterial eradication rate exceeding 98% at a safe low dose, which was significantly greater than the effect of equal amounts of antibacterial agents alone. Similarly, Liu et al. loaded their HA-coated ZIF-8 (HA-ZIF8) with VCM, and the formed nanocarrier was abbreviated as ZVH . The release studies were carried out in PBS solutions at pH values of 5.5, 6.5, and 7.4. The results indicated that a larger amount of VCM was released at pH 5.5 than at the other two pH levels. The MTT assay on RAW 264.7 cells displayed slight dose-dependent cytotoxicity; nevertheless, the formulation was biocompatible at In vitro [bib_ref] Layer-by-layer decorated nanoparticles with tunable antibacterial and Antibiofilm properties against both gram-positive..., Ivanova [/bib_ref] concentrations below 60 μg/ml. Both in vitro and in vivo antibacterial studies confirmed the superiority of VCM loaded in the HA-ZIF8 compared to the same dose of bare VCM. These results confirmed that TZH and ZVH are unique biosafe nanoplatforms that are able to selectively deliver antibiotics to pathogenic sites and achieve responsive and controllable release, thus improving the therapeutic effect against resistant infections and providing a new platform for further overcoming antibiotic resistance [bib_ref] The highly efficient elimination of intracellular bacteria: Via a metal organic framework..., Zhang [/bib_ref]. reported another HA-ZIF8, wherein they elaborated the use of HA-coated MOFs to deliver silver NPs (AgNPs) rather than conventional antibacterial agents. In this nanosystem, ZIF-8 accompanied by glucose oxidase was loaded with sliver nanocubes and subsequently coated with HA to enhance their biocompatibility and capability to target CD44. Interestingly, this nanosystem can function as a pool for releasing silver ions (Ag + ) and tiny Ag NPs (5 nm) from large silver nanocubes (30 nm) in response to bacterial microenvironments. The nanocomposite formed, termed GOD/Ag@ZIF-HA, had significant antibacterial activity against the studied bacteria with an MIC value of less than 10 μg/ml; this is concomitant with the controlled release profile of Ag + and the small size of AgNPs in the presence of bacteria. Additionally, at the MIC concentration, GOD/Ag@ZIF-HA inhibited the growth of tested bacteria completely . Even though biocompatibility studies are still outstanding for this research, this study revealed a promising Ag nanoplatform with both size-switchable and selective targeting ability. While all the above-mentioned studies investigated MOFs, a recent study expanded the pool and explored HAcoated nanocages. [bib_ref] Hyaluronic acid and antimicrobial peptide-modified gold/silver hybrid nanocages to combat bacterial multidrug..., Yang [/bib_ref] presented gold-silver hybrid nanocages coated with antimicrobial peptide and HA (HA-P(Au/Ag)). This study aimed to treat pneumonia by eradicating multidrug-resistant bacteria-Acinetobacter baumannii (MDR-AB). These HA-P(Au/Ag)s were biosafe with negligible cytotoxicity and hemolytic effects. The functionalization of the surface of the nanocages with HA facilitates the adhesion of HA-P(Au/Ag) to the bacteria, enabling the loaded antibacterial agents to penetrate and kill the bacteria. Thus, in vitro antibacterial tests revealed that the MIC and MBC values of uncoated nanocages were reduced to 3 and 12 μg/ml, respectively, after coating with HA and the peptide. Furthermore, combining photothermal therapy with the coated nanocages further potentiates their antibacterial activity and entirely eradicates MDR-AB. Similarly, the in vivo antibacterial studies using a pneumonia mice model revealed that treatment with HA-P(Au/Ag) dramatically improved the survival rate of tested mice while restoring the normal structure of the alveoli. This study proved that HA-P(Au/Ag) in combination with photothermal treatment is a viable approach for potentially overcoming MDR-AB multiple mechanisms .
Collectively, these findings elucidate that surface coating of nanocarriers with HA for targeting CD44 receptors on the surface of phagocytic cells can improve their uptake and internalization, resulting in targeted drug release and subsequent eradication of bacteria within infected macrophages.
## | hyaluronic acid-coated nanocarriers for hyaluronidase responsiveness
Bacterial infection sites are characterized by acidic pH as well as the presence of numerous virulent enzymes, including B-HAase. Therefore, exploiting these enzymes for diagnostics, drug targeting, and drug release is remarkably promising F I G U R E 1 0 Schematic illustration of (a) tetracycline-loaded HA-ZIF8 and (b) VCM-loaded HA-ZIF8 targeting intracellular bacteria [bib_ref] The highly efficient elimination of intracellular bacteria: Via a metal organic framework..., Zhang [/bib_ref]. Since HA can be easily degraded by B-HAase, HA has been widely employed as a targeting and capping agent. As previously reported, many species of bacteria, including S. aureus, Clostridium difficile, and others, were capable of generating HAase [bib_ref] Enzyme responsive hyaluronic acid nanocapsules containing polyhexanide and their exposure to bacteria..., Baier [/bib_ref] [bib_ref] Hyaluronidases of gram-positive bacteria, Hynes [/bib_ref]. Hence, by taking advantage of this phenomenon, HA (as a coating agent) can be used to detect and/or eradicate these bacteria [bib_ref] Layer-by-layer (LBL) self-assembled biohybrid Nanomaterials for efficient antibacterial applications, Wu [/bib_ref] [bib_ref] Integration of diagnosis and treatment in the detection and kill of S.aureus..., Xu [/bib_ref].
In recent years, mesoporous silica nanoparticles (MSNs) have been introduced as a promising drug delivery system. Mesoporous silica nanoparticles have essential parts of nanomedicines owing to their chemical stability, large encapsulation capacity, favorable biosafety, ease of preparation and modification, and tunable pore size [bib_ref] Recent advances toward the use of Mesoporous silica nanoparticles for the treatment..., Castillo [/bib_ref]. Thus, coating MSNs with HA could efficiently provide anti-adhesion properties in physiological fluids, as well as on-demand drug release in response to B-HAase [bib_ref] Layer-by-layer (LBL) self-assembled biohybrid Nanomaterials for efficient antibacterial applications, Wu [/bib_ref] [bib_ref] Integration of diagnosis and treatment in the detection and kill of S.aureus..., Xu [/bib_ref]. In this regard, [bib_ref] Layer-by-layer (LBL) self-assembled biohybrid Nanomaterials for efficient antibacterial applications, Wu [/bib_ref] explored the potential of layer-by-layer coated MSNs to release encapsulated amoxicillin in response to the HAase of S. aureus. In this study, amoxicillin, an example of a broad-spectrum antibiotic, was encapsulated into MSNs and subsequently coated with antimicrobial lysozyme, followed by a second layer of HA, and then a third layer of modified cationic polyglycerol methacrylate . The prepared NPs, abbreviated as MSN-Lys-HA-PGMA, had a diameter of 174 nm and a drug encapsulation capacity of 8.5%. The presence of the HA layer enables enzyme-responsive release behavior in the presence of B-HAase, which leads to the enzymolysis of MSN-Lys-HA-PGMA, followed by the release of lysozyme and the loaded amoxicillin at the site of action. The biosafety evaluation of MSN-Lys-HA-PGMA was carried out using the CCK-8 assay and hemolysis studies. Both attested to the biocompatibility of the coated nanosystem. In vitro antibacterial action was assessed by determining MIC, zone of inhibition, and bacterial cell viability. These studies showed that MSN-Lys-HA-PGMA had higher antibacterial activity than the equal dose of the loaded antibacterial agents alone. However, their efficacy against gram-negative bacteria was lower due to the presence of the lipopolysaccharide layers, which may inhibit the synergistic action of the lysozyme. Furthermore, the in vivo antibacterial test confirmed the superior activity of the MSN-Lys-HA-PGMA against with a 99.9% inhibition rate, which was consistent with the in vitro findings [bib_ref] Layer-by-layer (LBL) self-assembled biohybrid Nanomaterials for efficient antibacterial applications, Wu [/bib_ref].
Magnetic MSN, in contrast to the nonmagnetic MSN discussed above, enables integrated diagnosis and treatment of infection owing to its excellent magnetic property. As a step toward this direction, [bib_ref] Integration of diagnosis and treatment in the detection and kill of S.aureus..., Xu [/bib_ref] reported on-demand magnetic MSNs for the detection and treatment of S. aureus. In this study, magnetic MSNs were loaded with VCM, and F I G U R E 1 1 Schematic representation of (a) layer-by-layer coating of MSN-Lys-HA-PGMA and the possible antibacterial mechanism for each layer. (b) Synthesis of an "on-demand" integrated Ab@S-HA@MMSNs for diagnosis and treatment toward S. aureus [bib_ref] Layer-by-layer (LBL) self-assembled biohybrid Nanomaterials for efficient antibacterial applications, Wu [/bib_ref] [bib_ref] Integration of diagnosis and treatment in the detection and kill of S.aureus..., Xu [/bib_ref] then coated with sulfonated HA, followed by functionalization of the surface of magnetic MSNs (Ab@S-HA@MMSNs) with S. aureus antibodies by chemical reaction. Here, sulfonated-HA, instead of bare HA, was chosen owing to superior anticoagulant and release behavior. The prepared Ab@S-HA@MMSNs were spherical with a size of 240 nm and a maximum VCM loading capacity of 10.7%; thereafter these NPs were functionalized on the surface of magnetic glassy carbon electrode (MGCE). Therefore, since these NPs have both HA (which exerts anticoagulant action) and S. aureus antibodies, they can be directly applied to detect the amount of pathogenic S. aureus in whole blood. In addition, with the increased amount of S. aureus arriving at the MGCE, the coat of Ab@S-HA@MMSNs was degraded by B-HAase leading to the release of the encapsulated VCM and ultimately, the effective eradication of S. aureus. At a concentration below 500 μg/ml, Ab@S-HA@MMSNs were hemocompatible with a hemolysis rate of less than 5%. Owing to the ondemand controlled release of VCM, the antibacterial studies (in vitro) showed significant bacteriostatic action with a rate reaching 98% against S. aureus [bib_ref] Integration of diagnosis and treatment in the detection and kill of S.aureus..., Xu [/bib_ref]. Therefore, the finding presented in this paragraph displayed that such integrated nanosystems can be promising multifunctional platforms for achieving accurate detection and efficient eradication of S. aureus bloodstream infections.
Recently, multifunctional nanomedicines derived from a combination of chemo and photothermal treatments have garnered considerable interest due to their cooperatively enhanced bactericidal activity [bib_ref] Emerging photothermal-derived multimodal synergistic therapy in combating bacterial infections, Huo [/bib_ref] [bib_ref] Phototherapy-based combination strategies for bacterial infection treatment, Wei [/bib_ref]. In this regard, [bib_ref] Hyaluronic acid-templated Ag nanoparticles/graphene oxide composites for synergistic therapy of bacteria infection, Ran [/bib_ref] presented an HAase-triggered photothermal nanoplatform for the eradication of bacteria based on AgNPs and graphene oxide (GO). First, The AgNPs were coated with HA to form nanocomposites with a size of 30 nm; subsequently, these NPs were integrated with GO to form GO-HA-AgNPs. Due to the presence of the HA layer, the nanocomposites exhibited low toxicity to host cells. Furthermore, the release behavior of HAase-responsive AgNPs enabled AgNPs to be protected by the HA layer without affecting host cells and offer controlled release in the presence of bacteria. After HA was degraded, the sheetlike GO could accumulate on the surface of bacteria via physical interactions, and upon illumination of NIR light, the GO-based sheet locally raised the temperature, resulting in significant bacteriostatic and bactericidal actions thus resulting in synergistic therapy. This hypothesis was confirmed by performing both in vitro and in vivo antibacterial studies, which showed that GO-HA-AgNPs had greater antibacterial activity compared to AgNPs [bib_ref] Hyaluronic acid-templated Ag nanoparticles/graphene oxide composites for synergistic therapy of bacteria infection, Ran [/bib_ref]. Two years later, another HAase-responsive nanocomposite, AA@Ru@HA-MoS2, with a synergistic chemo-photothermal property was reported by [bib_ref] Enzyme-responsive Mesoporous ruthenium for combined chemo-Photothermal therapy of drug-resistant bacteria, Liu [/bib_ref] to treat bacterial infections. In this study, HA-coated mesoporous ruthenium nanoparticles (MRNs) were used as nanocarriers, ascorbic acid was loaded as a prodrug, and then ciprofloxacin-molybdenum disulfide was integrated as a catalyst at the outer surface. The formulated nanocomposites were pompon-like with a diameter of 206 nm. When AA@Ru@HA-MoS2 reached the infection site, B-HAase degraded the HA layer, releasing ascorbic acid and subsequently generating hydroxyl radicals in situ through molybdenum disulfide catalysis. Additionally, by exploiting the excellent photothermal properties of MRNs, a combination of chemo-and photo-thermal antibacterial therapy could be accomplished. A set of in vitro antibacterial studies on two bacterial models, namely S. aureus and P. aeruginosa, indicated that AA@Ru@HA-MoS2 not only had bactericidal activity, but also significant antibiofilm activity. The same finding was also obtained in the in vivo antibacterial experiment. Furthermore, the mouse infection model showed faster wound healing activity and excellent short and long-term biosafety profiles. Overall, these studies indicate that HA-coated antibacterial nanocarriers could further enhance the synergy of chemo-photothermal treatments. Such functionalization represents a robust strategy for designing enzyme-responsive multifunctional nanomedicines to improve targeted antibacterial delivery.
In summary, as HA is made up of enzyme-degradable disaccharide units, it is considered an ideal candidate for the fabrication of nanocarriers with an HAase-responsive crown. These HAase-responsive NPs hold great potential for more precise detection and more effective therapy of bacterial infections.
## | other hyaluronic acid-coated nanocarriers
In addition to the aforementioned rationales for coating various antibacterial nanocarriers with HA, HA can also be used as a coating material to reduce the toxicity of metallic NPs as well as an antiadhesive agent to prevent adsorption of biomolecules on the surface of nanocarriers [bib_ref] Layer-by-layer decorated nanoparticles with tunable antibacterial and Antibiofilm properties against both gram-positive..., Ivanova [/bib_ref].
The application of colloidal Ag-based materials has been rising owing to their potent antimicrobial action. Therefore, AgNPs have been used in several areas, including manufacturing cosmetics, medical devices, and pharmaceutical products [bib_ref] New strategies in the development of antimicrobial coatings: The example of increasing..., Knetsch [/bib_ref] [bib_ref] Silver nanoparticles: Synthesis, characterization, properties, applications, and therapeutic approaches, Zhang [/bib_ref]. However, due to its toxicity, these applications are limited. Silver NPs have been widely studied to reach biocompatible and environmentally friendly products by capping AgNPs with various bioactive polymers. Thus, coating AgNPs with HA could be a promising strategy for biomedical applications. In this regard,designed and prepared AgNPscapped with HA fibers (HA-AgNPs) as a potential wound dressing material. Silver NPs with diameters of 10 and 40 nm were prepared, then coated with HA fibers. In this study, HA served as a capping agent for Ag + as well as stabilizing agent during and after the synthesis of AgNPs. Several experiments were conducted to investigate the characterization, biocompatibility and cell viability, thermal profiles, and antibacterial activity of HA-AgNPs. Using mouse fibroblast 3T3 cells, the MTT assay indicated that HA, HA fibers, and HA-AgNPs were nontoxic to the cell line. Moreover, HA-AgNPs showed potent in vitro antibacterial activity against S. aureus and E. coli. This work showed that modification of AgNPs with HA can minimize their cellular toxicity. Further in vivo evaluation is necessary; however, AgNPs embedded in HA fibers can be employed for a variety of biomedical applications, most notably wound healing and dressing.
Hyaluronic acid has been widely employed for surface engineering of nanomaterials to enhance their stability and functionalities for biomedical applications. One of the primary obstacles that face the optimized nanoformulations when administered in vivo is the biofouling effect. Biofouling is defined as nonspecific interactions between biomolecules and the engineered nanocarriers leading to in situ changes in their pharmacokinetics, pharmacodynamics, and toxicity. Among the polysaccharides studied as anti-biofouling agents, HA is the most employed polysaccharide. Its antifouling actions appeared to be promising in terms of inhibiting nonspecific protein adsorption and enhancing nanocarrier stability in complex serum-containing media [bib_ref] Polysaccharide-decorated nanoparticles, Lemarchand [/bib_ref] [bib_ref] Impact of anti-biofouling surface coatings on the properties of nanomaterials and their..., Li [/bib_ref] [bib_ref] Carbohydrate-based nanocarriers and their application to target macrophages and deliver antimicrobial agents, Mosaiab [/bib_ref]. [bib_ref] Layer-by-layer decorated nanoparticles with tunable antibacterial and Antibiofilm properties against both gram-positive..., Ivanova [/bib_ref] accomplished this goal by coating bioinert polymeric NPs with several bilayers of biocompatible and antifouling HA and antibacterial aminocellulose. After five cycles of layer-by-layer coating technique, polyelectrolyte-coated NPs (HA-AC NPs) with a size below 50 nm were formed. The surface charge of the nanosystems is governed by its outermost layer. All coated NPs were shown to be biosafe when tested on human fibroblast cells. A series of in vitro antibacterial studies against E. coli and S. aureus were conducted. The results indicated that the coated NPs were capable of eradicating planktonic bacteria and inhibiting biofilm formation. For an exposure period of less than 1 h, the NPs with aminocellulose as the outermost layer demonstrated high antibacterial efficacy against grampositive and gram-negative bacteria. On the other hand, when HA was used as the outermost layer of NPs, a significant increase in antibiofilm activity was found against tested bacteria (including E. coli). Although the exact antibiofilm mechanism of HA-terminated NPs is not fully understood; it has been justified by the contribution of the HA layer (as anti-biofouling) to the improved colloidal stability of the nanocarrier in tryptic soy growth media and consequently enhanced efficacy of NPs toward gram-negative strains. This work demonstrated the anti-biofouling property of HA in protecting nanocarriers from nonspecific interactions with serum proteins [bib_ref] Layer-by-layer decorated nanoparticles with tunable antibacterial and Antibiofilm properties against both gram-positive..., Ivanova [/bib_ref]. However, further experiments (in vivo) should be pursued to study the potential of these NPs for treatment of infectious diseases.
Overall, surface functionalization with HA endows nanoparticles with bacterial targeting ability, enhances the colloidal stability and biosafety, protects the NPs from biofouling effects, and even serves as a gatekeeper to control the drug release. summarizes all HA-coated nanocarriers that have been used to date for antibacterial delivery. The table provides comprehensive information such as the size of the nanosystem, bacteria species, antibacterial agents(s), release triggering strategy, mode of therapy, and mode of study.
## | hyaluronic acid-drug conjugates
Hyaluronic acid is a large biocompatible hydrophilic polymer of repeating sugar units that can be covalently attached to various drugs [bib_ref] An update on polysaccharide-based nanomaterials for antimicrobial applications, Arora [/bib_ref] [bib_ref] Hyaluronic acid as potential carrier in biomedical and drug delivery applications, Prajapati [/bib_ref]. In 1991, the concept of drug conjugation to HA was introduced for the first time [bib_ref] Hyaluronan in drug delivery, Drobnik [/bib_ref]. In antibacterial delivery, direct conjugation of HA to various drugs could yield novel nanosystems with promising antibacterial effects [bib_ref] A hyaluronic acid functionalized self-nano-emulsifying drug delivery system (Snedds) for enhancement in..., Arshad [/bib_ref] [bib_ref] Hyaluronic acid-based levofloxacin nanomicelles for nitric oxidetriggered drug delivery to treat bacterial..., Lu [/bib_ref] [bib_ref] Synergistic clearance of intracellular pathogens by hyaluronan-streptomycin micelles encapsulated with rapamycin, Qiu [/bib_ref]. Such straightforward yet effective nanoformulations can be employed to enhance therapeutic efficacy against intracellular pathogens, as many inflammatory cells overexpress HA-targeted receptors (CD44). Furthermore, the covalent bonds between HA and the conjugated antibiotics can be tailored so they are not easily broken in the blood, but they are disrupted by hydrolysis and/or enzymatic degradation once they reach the target, releasing the drug. Thus, providing merits in terms of prolonging circulation time, increasing drug stability and solubility, and controlling release behavior (S. N. Kłodzi nska & Nielsen, 2021).
Instead of conjugating macromolecular hydrophobic polymers, small lipophilic drug molecules can be grafted onto the N-acetyl, hydroxyl, or carboxyl groups of HA. The synthesized HA-drug conjugates can be self-assembled into nanomicelles; thus, the conjugated drugs can be safely delivered to their destination (B. [bib_ref] Hyaluronic acid-based drug conjugates: State-of-the-art and perspectives, Chen [/bib_ref]. Even though antibiotics are powerful agents against pathogenic bacteria, more than 70% of them are ineffective against intracellular pathogens [bib_ref] Antibiotics in the clinical pipeline in 2011, Butler [/bib_ref]. Accordingly, to combat these threatening infections, there is a dire need to invent novel antibacterial agents that can penetrate inside mammalian macrophages and efficiently eliminate these pathogens. In this regard, two studies by Jinyou Duan research group have been reported. In both studies, HAbased micelles were constructed by conjugating lipophilic antibiotics with HA. Hyaluronic acid was used to maximize the uptake of the conjugated drug via a CD44-mediated mechanism leading to improved antibacterial action [bib_ref] Synergistic clearance of intracellular pathogens by hyaluronan-streptomycin micelles encapsulated with rapamycin, Qiu [/bib_ref]. In the former study, [bib_ref] Synergistic clearance of intracellular pathogens by hyaluronan-streptomycin micelles encapsulated with rapamycin, Qiu [/bib_ref] designed and synthesized pH-sensitive HA-based micelles containing HA-streptomycin conjugate and rapamycin, an autophagy activator . The novel selfassembled compound was synthesized by grafting streptomycin and decylamine on HA (via acid-labile bond) and subsequently loading with rapamycin via a spontaneous self-assembly approach. The prepared HA-based micelles had a hydrodynamic diameter of 179 nm with an EE% of 69.3%. At pH 5.5, the drug-HA bond was cleaved, releasing streptomycin and the loaded rapamycin at the infection site. Moreover, using RAW 264.7 and HK-2 cell lines, MTT assay validated the biocompatibility of blank and loaded micelles at concentrations below 400 μg/ml. The rapamycin-loaded HA-micelles demonstrated a high intracellular eradication rate against Salmonella typhimurium and S. aureus, even more than the equivalent mixture of streptomycin and rapamycin. As mentioned in literature, bare streptomycin cannot penetrate mammalian cells, thus exhibiting poor antibacterial action against intracellular microorganisms [bib_ref] Intracellular distribution and activity of antibiotics, Tulkens [/bib_ref]. Accordingly, conjugation of streptomycin with HA facilitates its uptake via CD44-mediated endocytosis. In addition, the rapamycin could further potentiate the action of the system via activation of autophagy . In the subsequent study, [bib_ref] Hyaluronic acid-based levofloxacin nanomicelles for nitric oxidetriggered drug delivery to treat bacterial..., Lu [/bib_ref] synthesized NO-triggered micelles (HA-NO-LVF) by conjugating an HA chain and a lipophilic LVF followed by self-assembly in aqueous conditions. It is known that when macrophages become infected, they generate a large amount of NO. Thus, exploiting endogenous NO to cleave HA-conjugated nanocarriers increases the potential for not only responsive drug release but also for scavenging redundant NO radicals to reduce tissue injury. In that sense, the in vitro release study in the presence of NO demonstrated that the micelles F I G U R E 1 2 Schematic illustration of rapamycin-loaded HA-based micelles were cleaved, releasing LVF in the pathogenic environment. Meanwhile, the NO-responsive HA-micelles retained hemocompatibility as well as biosafety profiles. Similar to their former study, HA-NO-LVF demonstrated significant antibacterial action against intracellular S. aureus, which could be due to CD44-mediated endocytosis of HA-NO-LVF and subsequent on-demand release of LVF, resulting in a higher eradication rate against intracellular bacteria. Moreover, in vivo evaluations acknowledged the ability of HA-NO-LVF to safely eliminate intracellular bacteria and reduce inflammatory mediators. In both studies, the assumption of CD44-mediated uptake of nanosystems was validated via performing a competitive inhibition assay. The findings of the assay revealed that when the CD44 receptors were occupied by HA, the intracellular uptake of the conjugated HA-micelles was dramatically reduced. Taken together, both studies proved the synergistic action of HA-mediated uptake and the bactericidal action of the antibiotic in eradicating intracellular pathogens, hence lowering the required doses of used antibiotics [bib_ref] Synergistic clearance of intracellular pathogens by hyaluronan-streptomycin micelles encapsulated with rapamycin, Qiu [/bib_ref].
Self-emulsifying drug delivery systems are isotropic mixtures of drugs, oil, surfactants, and co-surfactants that form oil-in-water (O/W) nanoemulsions upon contact with water and mild agitation. The formed nanoemulsions are thermodynamically stable with nanosized droplets, high solubilization ability for hydrophobic drugs, and enhanced biocompatibility and pharmacokinetic profiles [bib_ref] Formulation and in vitro evaluation of self-nanoemulsifying liquisolid tablets of furosemide, Dalal [/bib_ref] [bib_ref] Evaluation of self-nanoemulsifying drug delivery systems (SNEDDS) for poorly water-soluble Talinolol: Preparation,..., Kazi [/bib_ref]. Due to the overexpression of CD44 receptors on intestinal inflammatory cells, HA is an ideal choice for the preparation of HA-based nanoemulsions targeting intestinal infections [bib_ref] CD44-regulated intracellular proliferation of listeria monocytogenes, Eriksson [/bib_ref] [bib_ref] CD44 and its role in inflammation and inflammatory diseases, Johnson [/bib_ref]. In this regard, [bib_ref] A hyaluronic acid functionalized self-nano-emulsifying drug delivery system (Snedds) for enhancement in..., Arshad [/bib_ref] successfully conjugated ciprofloxacin to the HA chain via an amidation reaction. The synthesized conjugate was delivered as a self-nano emulsifying drug delivery system (HA-CIP-SNEDDS). The aim of this study was to improve ciprofloxacin mucopenetration, biopharmaceutical parameters, and hence intracellular antibacterial activity. The optimized nanoemulsion was found to be hemocompatible and biocompatible with a droplet size of 50 nm and maximum drug encapsulation efficiency of 85%. Moreover, at conditions mimicking the endosomal environment, HA-CIP-SNEDDS demonstrated superior and sustained release for 72 h in comparison to bare ciprofloxacin. As predicted, the fluorescence assay demonstrated that HA enhanced the internalization of ciprofloxacin by scavenger inflammatory cells, resulting in efficient targeting and eradication of S. typhimurium. Consequently, antibacterial studies showed that HA-CIP-SNEDDS had a strong bactericidal and antibiofilm activity against S. typhimurium compared to bare ciprofloxacin. When administered in vivo, HA-CIP-SNEDDS exhibited enhanced anti-salmonella activity, confirming the in vitro results. Therefore, this study showed that HA-CIP-SNEDDS seems to be a promising antibacterial agent against S. typhimurium with a strong targeting ability [bib_ref] A hyaluronic acid functionalized self-nano-emulsifying drug delivery system (Snedds) for enhancement in..., Arshad [/bib_ref].
These studies confirmed the potential use of HA as an antibacterial drug carrier to enhance targeted delivery and activity of conjugated antibiotics. This approach can significantly enhance antibiotic concentrations at infection sites, as well as cellular uptake and recognition of antibiotics, resulting in highly effective antibiotic targeted therapy.
## | conclusion and future perspectives
Polymer-based nanoantibiotics have been thoroughly explored to improve the efficacy of conventional antibiotics in various biomedical applications. Among all polymers used, HA has garnered considerable attention in the development of nanocarriers that can enhance the delivery of antibacterial agents to infection sites. In this review, we have highlighted and critically analyzed all reported nanosystems in literature that incorporate HA using different strategies for antibacterial applications such as treating, preventing, and/or diagnosing bacterial infections. After an extensive analytical search of various scientific databases, we have systematically categorized these HA-NPs according to their approach for application into HA-based nanocarriers, HA-capped NPs, and drug-conjugated HA-NPs. Therefore, HA and its conjugates are considered promising candidates for developing targeted nanocarriers of therapeutic or imaging agents for bacterial infections.
Based on the reported studies to date, we quantitatively analyzed and compared the various types of HA-based nanocarriers that have been employed for improving antibacterial properties . It is apparent that HA-based nanocarriers have been formulated and evaluated for their antibacterial potential, more than the other two strategies; and that the antibacterial activity of these nanocarriers has been tested against S. aureus and E. coli, more than other bacteria. This will aid future researchers in identifying the best strategy for preparing HA-NPs as well as the infectious microorganisms on which to focus.
Sites of bacterial infection have hallmarks not present in uninfected tissues; they often include cells overexpressing CD44 receptors and have unique microenvironments such as acidic pH and high levels of specific enzymes (including B-HAase). Therefore, HA-NPs can enhance antibacterial delivery to eradicate intracellular infections such as TB, typhoid, brucellosis, and others. Moreover, the polymeric nature of HA enables prolonged circulation of loaded (or conjugated) antibiotic(s) in the bloodstream, resulting in sustained release behavior. After reaching the targeted area, initial burst release of antibacterial agents and activation of their functions at the infection site may occur in response to B-HAase, thus enabling on-demand release of the loaded drug(s). In addition, surface functionalization using HA could increase the biocompatibility, intracellular uptake, and colloidal stability of inorganic nanoparticles. Therefore, this strategy has a bright future in advancing the field of inorganic nanomedicines drug delivery to treat and diagnose bacterial infections. On the other hand, HA-drug conjugates are considered as prodrugs synthesized by covalently attaching small antimicrobial drugs to HA. These linkages are not easily broken in the blood; however, they are breakable through acidic or enzymatic hydrolysis after reaching the target and releasing the drug. Therefore, such conjugates can serve as a targeting ligand for CD44 receptors on the surface of macrophages, allowing antibiotics to be safely delivered to eliminate intracellular infections.
While considerable efforts have been invested in proving the enhanced potential of HA-NPs for bacterial eradication and detection, these nanosystems are still in their laboratory research stages, and numerous further studies are necessary to obtain regulatory approval. The production, synthetic modification, and precise characterization and optimization of HA-NPs, as well as their biosafety and in vivo pharmacokinetics, should be carried out accurately according to clinical requirements. Moreover, more efforts are required to overcome obstacles associated with the process of scaling up these formulations into cost-effective products. Additional investigations, including in silico molecular dynamics and in vitro binding affinity calculations (using microscale thermophoresis), are required to fully understand the interactions between these nanocarriers and the targeted biomolecules. Furthermore, the in silico studies can help to determine the stability of the synthesized HA-NPs using virtual conditions that mimic experimental conditions.
Coating the surface of nanocarriers with HA is an effective strategy for enhancing antibacterial delivery as it provides anti-adhesion properties, colloidal stability, and selective targeting ability. In addition, the secretion of HAase by various bacteria can be exploited by creating multifunctional HA-coated NPs that can detect and treat HAase-secreting bacteria. These multifunctional nanocarriers have the potential to revolutionize the diagnosis and treatment of HAaserelated infections. As a result, more research efforts should be focused on the surface modification of nanocarriers with HA. Furthermore, targeting bacterial virulence factors via mimicking cell signaling pathways may enhance antibacterial efficacy of loaded drugs and minimize their adverse effects and thereby, contribute to tackling the F I G U R E 1 3 Hyaluronic acid-based nanoparticles (right) that have been developed for the detection and/or eradication of various bacteria (left) over the last two decades pathogenesis of infection. Therefore, by synthesis of modified HA-NPs targeting B-HAase (as a virulent factor), it is possible to inhibit bacterial pathogenicity. However, up to the date of this review, no research has been done in order to develop HA-NPs with this biomimetic action, leaving huge research areas to be explored. Generally, formulating HAdrug conjugates is the simplest form in terms of preparation and validation; however, in antibacterial delivery, only three studies have been reported. Therefore, it is expected that the investigations particularly in the area of intracellular antibacterial delivery will be expanded, since HA can facilitate the internalization of conjugated antibiotics via CD44 receptors and provide "on-demand" release at the infection site. Based on the potential shown in this review, we envisage an explosion in the use of HA for transforming diagnosis and treatment of bacterial infections. Hyaluronic acid as a key excipient in nanomedicines has the potential to save lives and improve quality of life. |
Clinical valve thrombosis post-transcatheter aortic valve implantation with hypoattenuating leaflet thickening in computed tomography: anticoagulation is the answer
A 59-year-old woman presented to the emergency room in acute pulmonary oedema. Her past medical history included aortic valve replacement in 2005 for severe aortic insufficiency with a mechanical prosthesis, which was followed 2 years later by an aortic homograft replacement due to prosthetic endocarditis. In 2017, the patient developed homograft dysfunction with severe stenosis and worsening left ventricular (LV) systolic dysfunction, with a left ventricular ejection fraction (LVEF) of 29%. After heart team discussion, it was decided to perform transfemoral transcatheter aortic valve implantation (TAVI) -Corevalve Evolut PRO no. 29. Because LV dysfunction persisted after TAVI (LVEF 31%) and the patient had left bundle branch block (QRS 150 ms), a cardiac resynchronization therapydefibrillator was implanted.After admission, the patient responded well to medical therapy which included non-invasive ventilation for 24 h. Device interrogation denoted over 98% biventricular pacing, and no atrial fibrillation was detected. Transthoracic and transoesophageal echocardiography showed TAVI dysfunction with leaflet thickening and an increased mean transvalvular pressure gradient of 38 mmHg with no significant regurgitation (Video 1) and a LVEF of 26%. Leaflet lesions were further evaluated with cardiac computed tomography (CT), which showed bioprosthesis dysfunction with extensive hypoattenuating leaflet Video 1 Transesophageal echocardiography denoting obstructive TAVI dysfunction with leaflet thickening.Video 2 Cardiac CT showing extensive leaflet thickening suggestive of thrombosis.
A 59-year-old woman presented to the emergency room in acute pulmonary oedema. Her past medical history included aortic valve replacement in 2005 for severe aortic insufficiency with a mechanical prosthesis, which was followed 2 years later by an aortic homograft replacement due to prosthetic endocarditis. In 2017, the patient developed homograft dysfunction with severe stenosis and worsening left ventricular (LV) systolic dysfunction, with a left ventricular ejection fraction (LVEF) of 29%. After heart team discussion, it was decided to perform transfemoral transcatheter aortic valve implantation (TAVI) -Corevalve Evolut PRO no. 29. Because LV dysfunction persisted after TAVI (LVEF 31%) and the patient had left bundle branch block (QRS 150 ms), a cardiac resynchronization therapydefibrillator was implanted.
After admission, the patient responded well to medical therapy which included non-invasive ventilation for 24 h. Device interrogation denoted over 98% biventricular pacing, and no atrial fibrillation was detected. Transthoracic and transoesophageal echocardiography showed TAVI dysfunction with leaflet thickening and an increased mean transvalvular pressure gradient of 38 mmHg with no significant regurgitation (Video 1) and a LVEF of 26%. Leaflet lesions were further evaluated with cardiac computed tomography (CT), which showed bioprosthesis dysfunction with extensive hypoattenuating leaflet Video 1 Transesophageal echocardiography denoting obstructive TAVI dysfunction with leaflet thickening. thickening compatible with thrombosis (Panels A-C, red arrow; Video 2), despite having been treated with aspirin since TAVI was performed. Warfarin was started, which was associated with clinical improvement and significant resolution of the leaflet lesions after 7 weeks of therapy (Panels D-F, red arrow; Video 3), and reduction of mean transprosthetic gradient to 11 mmHg.
Multiple mechanisms have been proposed for TAVI thrombosis, including (i) small valve size and under-expansion, (ii) aggressive postdilation, (iii) geometric deformation of aortic valve stent, (iv) suboptimal antiplatelet/antithrombotic therapy, or (v) underlying thrombophilia, which was not investigated in this patient.
This case highlights the importance of cardiac CT in the initial assessment and follow-up of patients with symptomatic bioprosthetic valve dysfunction. Moreover, it demonstrates that they can have a significant clinical improvement with anticoagulation therapy and illustrates the challenge of optimizing the antithrombotic strategy in patients with no previous formal indication for anticoagulation.
## Consent:
The authors confirm that written consent for submission and publication of this case report including images and associated text has been obtained from the patient in line with COPE guidance.
## Panel
Video 3 Cardiac CT confirming significant resolution of thrombosis after 7 weeks of anticoagulation. |
Urinary Potassium and Kidney Function Decline in the Population—Observational Study
Citation: Cirillo, M.; Bilancio, G.; Cavallo, P.; Palladino, R.; Zulli, E.; Villa, R.; Veneziano, R.; Laurenzi, M. Urinary Potassium and Kidney Function Decline in the Population-Observational Study.Nutrients 2021, 13, 2747. https://
# Introduction
Low kidney function is a highly prevalent disorder in the population and implies the risk of kidney failure or premature death. The control of hypertension is considered pivotal to reduce glomerular dysfunction and to slow down kidney disease progression although there is not unanimity about blood pressure targets and preferred antihypertensive drugs. Inhibitors or blockers of the renin-angiotensin system are generally considered preferable due to specific effects on glomerular hemodynamics. Similar favorable effects on glomerular dysfunction are ascribed also to antidiabetic drugs inhibiting the sodium-glucose transporter-2and to reduced protein intake.
The relation in the general population between dietary potassium and glomerular filtration rate is an unanswered question. Although daily potassium intake is close to the whole extracellular potassium pool, significant hyperkalemia does not occur after potassium ingestion because absorbed potassium is rapidly translocated into the intracellular Nutrients 2021, 13, 2747 2 of 13 space or excreted by the kidney. The intracellular potassium translocation is due to the activity of the sodium-potassium-ATPase and/or of potassium channels. The urinary potassium excretion is activated not only via stimulation of aldosterone secretion but also due to a cascade of events in the distal convoluted kidney tubule which include potassium channel stimulation in the basolateral cell membrane, down-regulation of sodium-chloride cotransporter, and increased potassium excretion due to increased distal sodium delivery. The strict relation of potassium intake with urinary potassium excretion is the rationale which supports the assessment of urinary potassium as index of dietary potassium intake. With the use of urine potassium as objective index of dietary potassium intake, a higher 24-h urinary potassium did not independently relate to 10-year development of estimated glomerular filtration rate (eGFR) < 60 mL/min per 1.73 m 2 in analyses controlled for baseline eGFR on 5315 adult residents in Groningen, the Netherlands. Similarly, a higher spot urine potassium did not relate to 5-year eGFR decline in a population sample of 4141 adult residents in Lauzanne, Switzerland. Results for urine potassium per se were not included in a study that investigated the relation of urine sodium/potassium ratio to 5-year decline of kidney function in 7063 adults from the Japanese general population. With the use of food frequency questionnaires, a higher potassium intake related to a less frequent 14-year mortality rate due to kidney disease or dialysis in 544,635 retired community-dwelling US individualsbut did not relate to the 6-year incidence of eGFR < 60 mL/min per 1.73 m 2 in 1780 Iranian adults. Inconsistencies exist also among clinical studies as recently reviewed. On the one hand, a better prognosis for kidney function was predicted by a higher potassium intake with food frequency questionnaire in Korean patients with chronic kidney diseaseor by a higher urinary potassium in Japanese diabetics, in Dutch outpatients, in Korean patients with chronic kidney disease, and in post hoc analyses of multi-center controlled clinical trials. On the other hand, a higher urinary potassium did not relate to kidney failure in a post hoc analysis of the Modification of Diet in Renal Disease cohortand related positively to kidney failure or eGFR halving in a prospective study on 3939 US patients with CKD defined as eGFR < 70 mL/min per 1.73 m 2.
The Gubbio study is an ongoing, observational, longitudinal, population-based study investigating also on the relation of dietary factors with kidney diseases assessed as eGFR. Measurements of glomerular filtration rate are inadvisable in populationbased studies due to invasivity, costs, complexity, and lack of international standardization. The use of eGFR has progressively spread in epidemiological and clinical studies after the development of international standards for creatinine assay. In particular, this was true for the CKD-Epi equation, which consistently had good accuracy and low bias over the whole range of kidney function. Within the Gubbio study cohort, the relation of dietary indices with kidney function differed between cross-sectional analyses and longitudinal analyses. Therefore, the present analysis was designed to investigate cross-sectionally and longitudinally the relation of urinary potassium as index of dietary potassium with eGFR and eGFR changes in adult examinees of the Gubbio study.
# Materials and methods
The Gubbio study is a population-based project ongoing in Gubbio, Italy. The study adheres to the Declaration of Helsinki and included an informed consent and the approval by the institutional committee (CEAS-Umbria #2850/16). Study design, involvement of the invited population, response rates, and characteristics of the Gubbio cohort have been reported previously. Three main exams were performed: exam-1 , , and exam- . Study protocol of all exams comprised the administration of standardized questionnaires by trained physicians and a medical visit with the measurements of anthropometry and blood pressure. With regard to collection of biological samples, the exam-1 protocol included a daytime, untimed, spot urine sample and a venous blood sample after a fast of at least two hours. The exam-2 protocol included a timed urine collection under fed condition from the first Nutrients 2021, 13, 2747 3 of 13 void after the evening meal to the first void at the morning wake-up included, an early morning blood sample collected under fasting conditions after the completion of the overnight urine collection, and a morning timed urine sample under fasting conditions collected after blood sampling. The exam-3 protocol included the collection of an early morning blood sample under fasting conditions only. Data about initiation of kidney failure replacement treatment and mortality were collected after exam-1 from the local sections of national registries. Target cohort of the present analyses were the individuals with age ≥ 18 years at exam-1 that participated also in exam-2 and exam-3.
## Measurements
The present analyses included data collected at exam-1, exam-2, exam-3 together with mortality data from after exam-1 up to the completion of exam-3 (30 June 2007). Urinary potassium was assessed as the ratio of urinary potassium concentration to urinary creatinine concentration to exclude the bias due to errors in completeness or timing of urine collectionand was used as the main independent variable (from here on abbreviated as uK/Cr). Exam-1 uK/Cr in spot samples and exam-2 uK/Cr in overnight collections were separately analyzed to control whether findings were affected by circadian rhythms in urinary potassium. Serum potassium was used as index of extracellular potassium. eGFR was calculated by the Chronic Kidney Disease-Epidemiology Collaboration equation using gender, age, and serum creatinine. eGFR, eGFR change between exams, and incidence of decreased eGFR were used as separate, alternative, dependent variables. eGFR changes were expressed in absolute units and with data normalization per the time interval between the exams that was calculated as the difference between the exam dates.
The list of possible confounders selected for the analysis included gender, age, body mass index (weight/height 2 ), estimated 24-h urinary creatinine, blood pressure, serum glucose, urinary sodium to creatinine ratio, urinary urea nitrogen to creatinine ratio, and data reported at questionnaires on drug treatments, smoking, habitual alcohol intake, habitual intake of water and other beverages. Urinary albumin as albumin/creatinine ratio was measured and included in the analyses only for examinees with age 45-64 years at exam-2.
Blood pressure was measured by trained physicians after participants had been seated quietly for 5 min, on the right arm, with the use of mercury sphygmomanometers and cuffs of appropriate size. Three measurements were taken one minute apart and the mean of the second and third measurement were used in analyses. Serum creatinine was measured in frozen samples by automated biochemistry (Express Plus, Bayer Diagnostic) using a kinetic alkaline picrate assay with IDMS-traceable standardization. Urinary potassium and other lab variables were measured in fresh samples using automated biochemistry and quality controls. Decreased eGFR was defined as eGFR < 60 mL/min × 1.73 m 2. Hypertension was defined as mean systolic pressure ≥ 140 mm Hg or mean diastolic pressure ≥ 90 mm Hg or regular antihypertensive drug treatment. Obesity was defined as body mass index ≥ 30 kg/m 2 . Diabetes was defined as serum glucose ≥ 7.0 mmol/L or regular anti-diabetic treatment.
## Statistical analyses
Descriptive data in the whole cohort were reported as prevalence for categorical variables, mean ± standard deviation (SD) for non-skewed numerical variables, and median with interquartile range (IQR) for skewed numerical variables (skewness > 1). Comparisons of skewed variables between exam-1 and exam-2 were done by Wilcoxon signed-rank test and/or by correlation analyses and paired t-test with log-transformed data. The crosssectional relations of uK/Cr to serum potassium and to covariates were investigated using quintile analyses, separately for exam-1 and at exam-2. To reduce the effect of sex and agethat is to obtain quintiles with different uK/Cr but with similar sex and age-quintiles were defined separately for men and women in seven age-strata 55-64, 65-74 and ≥75 years). eGFR and eGFR changes were investigated along sex-and age-controlled uK/Cr quintiles without adjustments for covariates. After that, the relation of uK/Cr to eGFR or eGFR change as dependent variable was investigated in multi-variable regression models with uK/Cr and other skewed variables as log-transformed data. Five multi-variable linear models were investigated: cross-sectional relation of exam-1 uK/Cr to exam-1 eGFR (Model 1); cross-sectional relation of exam-2 uK/Cr to exam-2 eGFR (Model 2); longitudinal relation of exam-1 uK/Cr to eGFR change from exam-1 to exam-2 (Model 3); longitudinal relation of exam-2 uK/Cr to eGFR change from exam-2 to exam-3 (Model 4); longitudinal relation of the mean of uK/Cr at exam-1 and exam-2 to eGFR change from exam-1 to exam-3 (Model 5). The list of covariates in the cross-sectional Models 1-2 included gender, age, body mass index as overweight index, estimated 24-h urinary creatinine as index of urinary creatinine excretion, systolic pressure, diabetes, urinary sodium to creatinine ratio as index of dietary sodium intake, urinary urea nitrogen to creatinine ratio as index of dietary protein intake, smoking, habitual alcohol intake, habitual intake of water and other beverages, and treatment with antihypertensive drugs, diuretics, potassium-sparing diuretics, inhibitors/blockers of renin-angiotensin system. The list of covariates in the longitudinal Models 3-5 included also eGFR at initiation of the observation period that is a key predictor of eGFR changes over time in the Gubbio study cohort. For sensitivity analyses, Model 5 was analyzed also in selected subgroups. Lastly, multi-variable logistic regression models investigated the relation of the mean of uK/Cr at exam-1 and exam-2 to after exam-2 mortality and to the incidence from exam-1 to exam-3 of decreased eGFR. For direct comparability among models, the results of linear regression were reported as standardized regression coefficient (beta) that indicates the fraction of the dependent variable standard deviation that is explained by a difference of one standard deviation of the independent variable (uK/Cr). Results included 95% confidence interval (95%CI). Statistical procedures were performed by IBM-SPSS Statistics 19.
# Results
## Descriptive statistics
From the original cohort of the 4554 participants with age ≥ 18 years at exam-1, the analyses excluded 97 examinees with missing data at exam-1, 350 examinees who had died before exam-2, 1008 non-responders to exam-2 (exam-2 mortality-corrected response rate = 75.5%), 39 examinees with missing data at exam-2, 615 examinees who had died before exam-3, and 418 non-responders to exam-3 (exam-3 mortality-corrected response rate = 82.9%). No examinee was with missing data at exam-3. Thus, the study cohort comprised 2027 examinees (56.9% women, age at exam-1 18-74 years). in Supplementary Material shows the flowchart for study cohort selection.
The distribution of uK/Cr was positively skewed both at exam-1 and at exam-2 in the Supplementary Material). Daytime uK/Cr in spot sample of exam-1 was significantly higher than and positively correlated with overnight uK/Cr in timed collection of exam-2 (Table S1 in the Supplementary Material). eGFR as mL/min per 1.73 m 2 decreased from 90.5 ± 18.2 at exam-1 to 87.6 ± 14.0 at exam-2, and to 76.7 ± 14.5 at exam-3. Absolute GFR change as mL/min per 1.73 m 2 was −2.81 ± 15.08 from exam-1 to exam-2, −10.93 ± 9.82 from exam-2 to exam-3, and −13.74 ± 15.11 from exam-1 to exam-3. Mean ± SD for time interval was 6.0 ± 1.0 years from exam-1 to exam-2, 13.3 ± 2.1 years from exam-2 to exam-3, and 19.3 ± 2.0 years from exam-1 to exam-3. With data normalization for time interval, eGFR change per year was −0.49 ± 2.64 from exam-1 to exam-2, −0.85 ± 0.76 from exam-2 to exam-3, and −0.72 ± 0.79 from exam-1 to exam-3. Prevalence of decreased eGFR was 3.2% at exam-1 (n = 64), 3.9% at exam-2 (n = 79), and 11.4% at exam-3 (n = 231). Incidence of decreased eGFR was 10.2% from exam-1 to exam3 (n = 200).
## Analyses by sex-and age-controlled uk/cr quintiles
At exam-1, higher uK/Cr quintile cross-sectionally associated with higher body mass index, higher diabetes prevalence, higher alcohol intake, and higher urinary sodium . . Descriptive statistics for data of exam-1 in whole cohort (prevalence for categorical variables, mean ± SD for non-skewed numerical variables, and median with interquartile range for skewed variables) and trend analyses by quintiles of uK/Cr (prevalence or mean). At exam-2, higher uK/Cr quintile cross-sectionally associated with higher serum potassium, higher systolic pressure, higher urinary sodium, and higher urinary urea nitrogen . . Descriptive statistics for data of exam-2 in whole cohort (prevalence for categorical variables, mean ± SD for non-skewed numerical variables, and median with interquartile range for skewed variables) and trend analyses by quintiles of uK/Cr (prevalence or mean). Regarding the relation with eGFR, higher uK/Cr quintile at exam-1 did not associate cross-sectionally with eGFR at exam-1 but it associated longitudinally with higher eGFR at exam-2 and with lesser eGFR decline from exam-1 to exam-2 (upper section of. Findings were similar without or with data normalization for time interval from exam-1 to exam-2. Higher uK/Cr quintile at exam-2 did not associate cross-sectionally with eGFR at exam-2 but it associated longitudinally with higher exam-3 eGFR and with lesser eGFR decline from exam-2 to exam-3 (central section of. Findings were similar without or with data normalization for time interval from exam-2 to exam-3. Higher quintile of the uK/Cr mean between exam-1 and exam-2 associated longitudinally with higher exam-3 eGFR and with lesser eGFR decline from exam-1 to exam-3 (lower section of. Findings were similar without or with data normalization for time interval from exam-1 to exam-3.
## Whole
## Whole
## Multivariable regression analyses
Log-transformed uK/Cr did not independently relate to eGFR of the same exam either at exam-1 and at exam-2, Models 1 and 2). Log-transformed uK/Cr, both at exam-1 and at exam-2, positively and independently related to eGFR change at the subsequent exam, Models 3-4). Mean of log transformed uK/Cr at exam-1 and exam-2 positively and independently related to eGFR change from exam-1 to exam-3, Model 5). ns not significant (p > 0.05), * p < 0.03, ** p < 0.01, *** p ≤ 0.001. a covariates in the model: gender and exam-1 data for age, estimated 24-h urinary creatinine, body mass index, systolic pressure, log-transformed urinary sodium/creatinine ratio, log alcohol intake, and categorical values for diabetes, treatment with antihypertensive drugs, diuretics, potassium-sparing diuretics, inhibitors/blockers of renin-angiotensin system (yes/no = 1/0). b covariates in the model: gender and exam-2 data for age, estimated 24-h urinary creatinine, body mass index, systolic pressure, log-transformed urinary sodium/creatinine ratio, log alcohol intake, log urinary urea nitrogen/creatinine ratio, and categorical values for diabetes, treatment with antihypertensive drugs, diuretics, potassium-sparing diuretics, inhibitors/blockers of renin-angiotensin system (yes/no = 1/0). c covariates in the model: same as in Model 1 with addition of exam-1 eGFR. d covariates in the model: same as in Model 2 with addition of exam-2 eGFR. e covariates in the model: gender, exam-1 age, exam-1 eGFR, mean of exam-1 and exam-2 data for estimated 24-h urinary creatinine, body mass index, systolic pressure, log alcohol intake, log urinary sodium/creatinine ratio, exam-2 log urinary urea nitrogen/creatinine ratio, and categorical values for the prevalence at exam-1 and the incidence at exam-2 of diabetes, treatment with antihypertensive drugs, diuretics, potassium-sparing diuretics, inhibitors/blockers of renin-angiotensin system (yes/no = 1/0).
Findings of Model 5 were similar in sensitivity analyses for the following groups: men and women, age 18-44, 45-64 and 65 and over, eGFR < 90 and eGFR ≥ 90 mL/min per 1.73 m 2 , hypertensive and non-hypertensive, obese and non-obese, diabetic and nondiabetic, drinker and non-drinker, and with urinary sodium or urinary urea nitrogen below or above the median .
## Multivariable regression analyses
Log-transformed uK/Cr did not independently relate to eGFR of the same exam either at exam-1 and at exam-2, Models 1 and 2). Log-transformed uK/Cr, both at exam-1 and at exam-2, positively and independently related to eGFR change at the subsequent exam (Table 4, Models 3-4). Mean of log transformed uK/Cr at exam-1 and exam-2 positively and independently related to eGFR change from exam-1 to exam-3, Model 5).. Standardized regression coefficient of log transformed uK/Cr to eGFR at the same exam and to eGFR change over time at subsequent exams (beta and 95%CI). ns not significant (p > 0.05), * p < 0.03, ** p < 0.01, *** p ≤ 0.001. a covariates in the model: gender and exam-1 data for age, estimated 24-h urinary creatinine, body mass index, systolic pressure, log-transformed urinary sodium/creatinine ratio, log alcohol intake, and categorical values for diabetes, treatment with antihypertensive drugs, diuretics, potassium-sparing diuretics, inhibitors/blockers of renin-angiotensin system (yes/no = 1/0). b covariates in the model: gender and exam-2 data for age, estimated 24-h urinary creatinine, body mass index, systolic pressure, log-transformed urinary sodium/creatinine ratio, log alcohol intake, log urinary urea nitrogen/creatinine ratio, and categorical values for diabetes, treatment with antihypertensive drugs, diuretics, potassium-sparing diuretics, inhibitors/blockers of renin-angiotensin system (yes/no = 1/0). c covariates in the model: same as in Model 1 with addition of exam-1 eGFR. d covariates in the model: same as in Model 2 with addition of exam-2 eGFR. e covariates in the model: gender, exam-1 age, exam-1 eGFR, mean of exam-1 and exam-2 data for estimated 24-h urinary creatinine, body mass index, systolic pressure, log alcohol intake, log urinary sodium/creatinine ratio, exam-2 log urinary urea nitrogen/creatinine ratio, and categorical values for the prevalence at exam-1 and the incidence at exam-2 of diabetes, treatment with antihypertensive drugs, diuretics, potassium-sparing diuretics, inhibitors/blockers of renin-angiotensin system (yes/no = 1/0).
## Type of model
Findings of Model 5 were similar in sensitivity analyses for the following groups: men and women, age 18-44, 45-64 and 65 and over, eGFR < 90 and eGFR ≥ 90 mL/min per 1.73 m 2 , hypertensive and non-hypertensive, obese and non-obese, diabetic and non-diabetic, drinker and non-drinker, and with urinary sodium or urinary urea nitrogen below or above the median .
## Figure 1.
Multivariable standardized regression coefficient (beta) with 95% confidence interval and P value for the relation of the mean of uK/Cr at exam-1 and exam-2 with yearly eGFR change from exam-1 to exam-3: whole study cohort and selected subgroups. The dotted line indicates beta = 0. Covariates included in the multivariable model are listed in the footer of. Subgroups were defined on the basis of data collected at exam-1. Findings of multivariable Model 5 were similar when the analysis was repeated with control also for urinary albumin/creatinine ratio in the 943 examinees with age 45-64 at exam-2 and measured urinary albumin/creatinine ratio (beta = 0.046, 95%CI = 0.01/0.091, p = 0.044; median and IQR of urinary albumin/creatinine ratio = 3.40 and 1.08/6.04 mg/g).
In the logistic multivariable model with control for the same covariates included in Model 5, one SD higher mean of log-transformed uK/Cr at exam-1 and exam-2 (+0.16 log mmol/g) did not relate to after exam-2 mortality (odds ratio = 0.99, 95%CI = 0.85/1.16, p = 0.889) but it related to 22% lower incidence of decreased eGFR from exam-1 to exam-3 (odds ratio = 0.78, 95%CI = 0.61/0.98, p = 0.039).
# Discussion
The results of the present long-term observational study indicated that, in a sample of the adult general Italian population, higher uK/Cr ratio did not relate cross-sectionally to eGFR but related to a lesser eGFR decline during an observation period ranging approximately from 6 to 20 years, independently of gender, age, and other variables.
The main limitation of the study was the lack of 24-h urine collection which is considered the best marker of dietary potassium intake. Thus, the possibility cannot be excluded that the findings reflected the influence of circadian rhythms of urinary potassium rather than the influence of dietary potassium intake. If this were the case, the study results would indicate that the circadian rhythm in urinary potassium is an independent predictor of kidney function decline in the population. However, two observations were against a possible key role of circadian rhythms. First, the relation of uK/Cr with eGFR change over time was substantially identical in analyses for daytime uK/Cr at exam-1 and for overnight uK/Cr at exam-2. Second, quintiles with higher overnight uK/Cr should have lower morning serum potassium at the end of the overnight urine collection if higher overnight uK/Cr was not sustained by a higher potassium intake in the previous meal and was caused only by an increased excretion due to circadian rhythms. Vice versa, exam-2 data in indicated that quintiles with higher overnight uK/Cr had higher morning serum potassium, a finding that could be explained only by parallel influences of a higher dietary potassium intake on urine potassium and serum potassium. In accordance with the interpretation that higher uK/Cr reflected higher dietary K intake there were the positive associations of uK/Cr with other diet-related indices both at exam-1 (body mass index, alcohol intake, and urinary sodium) and at exam-2 (urinary sodium and urinary urea nitrogen). Due to the day-to-day intra-individual variability in potassium intake and urinary potassium, a single measurement of urinary potassium was expected to imply a regression dilution bias and an underestimate of the true strength of the association. The observation that the highest beta was found with the use of the mean of exam-1 uK/Cr and of exam-2 uK/Cr was in accordance with this expectation. Another limitation was the lack of data on serum cystatin C that could have improved the accuracy of eGFR.
An observational, population-based study can hardly clarify the mechanism(s) underlying the observed associations. At present, the mechanisms remain hypothetical for the relation of higher urinary potassium with lesser eGFR decline over time. A direct favorable effect of potassium intake could be involved if high potassium intake would play a direct protective role against kidney damage also in humans, as described in experimental models. If potassium had direct favorable effects, a reno-protective role could be conceivable also for drugs affecting potassium homeostasis. The effects of high potassium intake on kidney function decline could be mediated also by indirect, sodium-dependent mechanism(s). As recently reviewed by Wey et al., high potassium intake stimulates natriuresis reducing the activity of the thiazide-sensitive sodium-chloride cotransporter in the kidney distal convoluted tubule. Given that high sodium intake related to greater kidney function decline in clinical studies and in population-based studies, the natriuretic effects of high potassium intake could be an indirect pathway linking potassium intake to kidney function decline. Other possibilities cannot be excluded.
Urinary potassium did not relate to eGFR decline in the seminal paper of Kieneker et al. when analyses were controlled also for baseline eGFR and albuminuria, nor in the paper of Deriaz et al. when analyses were done for urinary potassium by itself. Therefore, this is the first population-based report of an independent relation of higher urinary potassium with lesser eGFR decline over time. Regarding previous clinical studies on urinary potassium, the present results agreed with the observations on Japanese diabetics, on Dutch outpatients with eGFR ≥ 60 mL/min per 1.73 m 2, on Korean non-dialysis patients with chronic kidney disease stages 1-5, and on high cardiovascular risk patients of multicenter trialsbut disagreed with observations on US patients of the Modification of Diet in Renal Disease study and of the Chronic Renal Insufficiency Cohort Study. Given that potassium content is high not only in fruits like banana but also in meat, the disagreement among studies in different countries could be due also to the predominant sources of habitual dietary potassium in the specific country under study. According to this possibility, the relation of dietary potassium with kidney function decline could differ between populations with high potassium intake due to a steak-rich diet and population-samples with high potassium intake due to veg-rich diet.
Regarding practical implications, the study results support the suggestion of high potassium diet in individuals at risk of kidney function decline on the basis of the significant and independent relation of higher urinary potassium with lesser eGFR decline. Although the relation was consistent in this study also in the subgroup with eGFR < 90 mL/min × 1.73 m 2 , two points recommend a word of caution for the expansion of the suggestion in chronic kidney disease stage 3 or higher: the present observation of higher serum potassium in people with higher urinary potassium and the previous report of higher mortality in people with serum potassium higher than 5 mmol/L.
Briefly, this observational cohort study reported that, in a sample of the adult general population, a higher urinary potassium related to a lesser eGFR decline during an observation period approximately ranging from 6 to 20 years, and independently of gender, age, and several other variables. Altogether, the results supported the hypothesis that a high dietary potassium intake could have favorable effects against the progressive decline in kidney function associated with ageing. Moreover, study results underlined the need of investigations about the relation of dietary potassium on kidney function with focus also on the source of dietary potassium.
Supplementary Materials: The following are available online at https://www.mdpi.com/article/10 .3390/nu13082747/s1, : Flow chart for selection of the study cohort; : Frequency distribution and skewness of uK/Cr at exam-1 (daytime spot sample) and and at exam-2 (overnight timed collection); : uK/Cr at exam-1 and uK/Cr at exam-2: descriptive statistics, differences between exams, and correlation between exams.
Author Contributions: The present analysis was planned by M.C. in collaboration with M.L. The first draft of the manuscript was prepared by M.C. and M.L. take responsibility for the overall content of the work, and/or the conduct of the study, had access to the data, and controlled the decision to publish. G.B., P.C., R.P., E.Z., R.V. (Rachele Villa) and R.V. (Rosangela Veneziano), participated in the acquisition, analysis, or interpretation of data for the work. All authors gave substantial contributions to the followings: drafting the work or revising it critically for important intellectual content; final approval of the version to be published; agreement to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. All authors have read and agreed to the published version of the manuscript.
Funding: Economic support to the Gubbio study was given in the past by: Merck, Sharp & Dohme, Italy (MSD); the U.S. National Heart, Lung and Blood Institute (Grant R01HL-40397-02); the Ministero Italiano di Università e Ricerca . None of the sponsor(s) or funder(s) had any role in the study design; in the collection, analysis, and interpretation of data; in the writing of the report; and in the decision to submit the article for publication.
## Institutional review board statement:
The study was conducted according to the guidelines of the Declaration of Helsinki and approved by the Institutional Ethics Committee of Regione Umbria (protocol code: CEAS-Umbria #2850/16; date of approval: 14 July 2016).
Informed Consent Statement: Informed consent was obtained from all subjects involved in the study.
Data Availability Statement: M.C. and M.L. had access to the data at the Centro Studi Epidemiologici di Gubbio. |
COVID-19 pandemic: health impact of staying at home, social distancing and ‘lockdown’ measures—a systematic review of systematic reviews
A B S T R A C TBackground To systematically review the evidence published in systematic reviews (SR) on the health impact of staying at home, social distancing and lockdown measures. We followed a systematic review approach, in line with PRISMA guidelines.MethodsIn October 2020, we searched the databases Cochrane Database of Systematic Reviews, Ovid Medline, Ovid Embase and Web of Science, using a pre-defined search strategy.ResultsThe literature search yielded an initial list of 2172 records. After screening of titles and abstracts, followed by full-text screening, 51 articles were retained and included in the analysis. All of them referred to the first wave of the coronavirus disease 2019 pandemic. The direct health impact that was covered in the greatest number (25) of SR related to mental health, followed by 13 SR on healthcare delivery and 12 on infection control. The predominant areas of indirect health impacts covered by the included studies relate to the economic and social impacts.Only three articles mentioned the negative impact on education.Conclusions The focus of SR so far has been uneven, with mental health receiving the most attention. The impact of measures to contain the spread of the virus can be direct and indirect, having both intended and unintended consequences.
# Introduction
In response to the coronavirus disease 2019 (COVID- [bib_ref] The psychosocial impact of flu influenza pandemics on healthcare workers and lessons..., Barello [/bib_ref] pandemic, governments worldwide adopted policies that aimed to reduce transmission, culminating in March and April 2020 in many countries in staying at home and physical (or 'social') distancing measures, often referred to as 'lockdown'. While these measures helped to bring down the number of new infections, gaining valuable time for the health sector to shore up its capacity and expertise for dealing with infected patients, it has become clear that the policy response had wideranging impacts on the health and well-being of populations across all sectors of society and affecting all health determinants.
Faced with new waves of infections in autumn 2020 and winter 2020/2021 and the imposition of new lockdowns in many countries, it is important to understand the positive and negative impacts of lockdowns on the health and well-being of populations to inform future policy responses.
A Health Impact Assessment conducted by Public Health Wales April-May 2020 found that there was a scarcity of academic peer-reviewed research literature regarding the impacts of prolonged quarantine periods and social distancing on health and well-being. [bib_ref] A Health Impact Assessment of the 'Staying at Home and Social Distancing..., Green [/bib_ref] However, the academic literature on COVID-19 is evolving rapidly and so a renewed assessment of the academic literature was appropriate.
The overarching aim of this study was to systematically review the evidence published in systematic reviews on the health impact of staying at home, social distancing and lockdown measures.
# Methods
A systematic review of systematic reviews was conducted following the Prepared Items for Systematic Reviews and Meta-Analysis (PRISMA) guidelines. [bib_ref] Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement, Moher [/bib_ref] Relevant publications were identified by systematically searching the scientific literature, with the search undertaken on 20 October 2020. We searched the scientific databases Cochrane Database of Systematic Reviews, Ovid Medline, Ovid Embase and Web of Science, using a pre-defined search strategy (detailed search strategies are provided in the Supplementary material).
Inclusion and exclusion criteria for study selection were defined a priori, after piloting them on a sample of 70 articles. Articles were included if they were published in English, were systematic reviews and focused on the health impact of staying at home, social distancing and lockdown measures related to the COVID-19 pandemic or other pandemics. There was no limitation set on the date of publication or the country of study implementation.
Articles published in languages other than English, not concerned with humans, not following a systematic review study design, or not concerned with the health impact of measures were excluded.
Identified studies were reviewed independently for eligibility in a two-step process: a first screen was performed on title and abstract, followed by the screening of full texts. Data were extracted using a standardized data extraction spreadsheet. In cases of doubt, studies were discussed within the research group and consensus reached. Because of the heterogeneity of included studies, no meta-analysis could be undertaken, and the results of our systematic review are presented in the form of a narrative synthesis.
# Results
The literature search yielded an initial list of 2172 records that provided 450 relevant articles after the first screening of title and abstract. Papers were screened and selected, as illustrated in [fig_ref] Figure 1: PRISMA diagram of systematic article selection [/fig_ref]. After the second screening based on full texts, 51 articles were retained. [bib_ref] Systematic rapid living review on rehabilitation needs due to COVID-19: update to..., Andrenelli [/bib_ref] [bib_ref] Telepsychotherapy: a leaflet for psychotherapists in the age of COVID-19. A review..., Poletti [/bib_ref] [bib_ref] Effects of the COVID-19 pandemic on supply and use of blood for..., Stanworth [/bib_ref] [bib_ref] Applications of e-health to support personcentered health care at the time of..., Tebeje [/bib_ref]
## General description of included articles
The overall characteristics of the articles included in the systematic review are shown in [fig_ref] Table 1: Main characteristics of the studies included [/fig_ref]. All of them referred to the first wave of the COVID-19 pandemic. April and March 2020 represent the time limits for almost half of the systematic reviews included (n = 25). Overall, eight systematic reviews were performed with a metaanalysis. 3,5,6, [bib_ref] How do funeral practices impact bereaved Relatives' mental health, grief and bereavement?..., Burrell [/bib_ref] [bib_ref] Mental health effects of infection containment strategies: quarantine and isolation-a systematic review..., Henssler [/bib_ref] [bib_ref] Prevalence of anxiety in medical students during the COVID-19 pandemic: a rapid..., Lasheras [/bib_ref] [bib_ref] The psychological and mental impact of coronavirus disease 2019 (COVID-19) on medical..., Luo [/bib_ref] [bib_ref] Mental health during the Covid-19 outbreak in China: a meta-analysis, Ren [/bib_ref] Almost one third of articles included (n = 16) describes other outbreaks or pandemics in addition to the COVID-19 pandemic, including Severe acute respiratory syndrome (SARS), Middle East Respiratory Syndrome (MERS), Influenza A (H1N1), Ebola, Chikungunya, Zika, Multiple drug resistance (MDR) bacteria, Methicillinresistant Staphylococcus aureus (MRSA), and human immunodeficiency virus (HIV). [bib_ref] The potential impact of the COVID-19 pandemic on child growth and development:..., Araujo [/bib_ref] [bib_ref] The psychological impact of quarantine and how to reduce it: rapid review..., Brooks [/bib_ref] [bib_ref] The potential impact of COVID-19 on psychosis: a rapid review of contemporary..., Brown [/bib_ref] [bib_ref] Travel-related control measures to contain the COVID-19 pandemic: a rapid review, Burns [/bib_ref] [bib_ref] The psychological impact of COVID-19 and other viral epidemics on frontline healthcare..., Cabarkapa [/bib_ref] [bib_ref] PTSD symptoms in healthcare workers facing the three coronavirus outbreaks: what can..., Carmassi [/bib_ref] [bib_ref] Mental health effects of infection containment strategies: quarantine and isolation-a systematic review..., Henssler [/bib_ref] [bib_ref] Mental health outcomes of quarantine and isolation for infection prevention: a systematic..., Hossain [/bib_ref] [bib_ref] Psychological burden of quarantine in children and adolescents: a rapid systematic review..., Imran [/bib_ref] [bib_ref] Suicidal behaviors and ideation during emerging viral disease outbreaks before the COVID-19..., Leaune [/bib_ref] [bib_ref] Rapid systematic review: the impact of social isolation and loneliness on the..., Loades [/bib_ref] [bib_ref] Quarantine alone or in combination with other public health measures to control..., Nussbaumer-Streit [/bib_ref] [bib_ref] A systematic review of covid-19 epidemiology based on current evidence, Park [/bib_ref] [bib_ref] A rapid review of pandemicrelated behaviours and psychological outcomes, Usher [/bib_ref] [bib_ref] School closure and management practices during coronavirus outbreaks including COVID-19: a rapid..., Viner [/bib_ref] [bib_ref] How to improve adherence with quarantine: rapid review of the evidence, Webster [/bib_ref]
## Characteristics of included articles
The majority of systematic reviews included focused on the impact of lockdown measures, with only nine articles focussing mostly on the impact of the pandemic.
Concerning the type of lockdown restrictions, the majority of the systematic reviews was focused on isolation, quarantine and social isolation, with many articles discussing multiple restrictive measures.
As regards other lockdown measures, four articles described the impact of school closures, [bib_ref] Quarantine alone or in combination with other public health measures to control..., Nussbaumer-Streit [/bib_ref] [bib_ref] A systematic review of covid-19 epidemiology based on current evidence, Park [/bib_ref] [bib_ref] Non-pharmaceutical interventions for containment, mitigation and suppression of COVID-19 infection, Patino-Lugo [/bib_ref] [bib_ref] School closure and management practices during coronavirus outbreaks including COVID-19: a rapid..., Viner [/bib_ref] seven systematic reviews explored the impact of travel restrictions, two examined the impact of workplace distancing, [bib_ref] A systematic review of covid-19 epidemiology based on current evidence, Park [/bib_ref] [bib_ref] Non-pharmaceutical interventions for containment, mitigation and suppression of COVID-19 infection, Patino-Lugo [/bib_ref] and one explored the impact of restrictions of funeral practices. [bib_ref] How do funeral practices impact bereaved Relatives' mental health, grief and bereavement?..., Burrell [/bib_ref] With regard to the impact on health services, two systematic reviews 1,4 explored the rescheduling of nonurgent outpatient visits, non-urgent surgery interventions, the putting on hold of 'non-essential' activities and the limitations in accessing hospitals. The indirect effect of restrictions of health services, and lockdown more generally, is represented by telemedicine, which is described by the 10 systematic reviews. 1,3,5-11,13 The health impact of lockdown measures can be direct or indirect . The direct health impact that has been covered in the greatest number of included articles relates to mental health, [bib_ref] Domestic violence and substance abuse during COVID19: a systematic review, Abdo [/bib_ref] [bib_ref] The potential impact of the COVID-19 pandemic on child growth and development:..., Araujo [/bib_ref] [bib_ref] Impact of the COVID-19 pandemic on psychosocial health and well-being in South-Asian..., Banerjee [/bib_ref] [bib_ref] The psychosocial impact of flu influenza pandemics on healthcare workers and lessons..., Barello [/bib_ref] [bib_ref] The psychological impact of quarantine and how to reduce it: rapid review..., Brooks [/bib_ref] [bib_ref] The potential impact of COVID-19 on psychosis: a rapid review of contemporary..., Brown [/bib_ref] [bib_ref] How do funeral practices impact bereaved Relatives' mental health, grief and bereavement?..., Burrell [/bib_ref] [bib_ref] The psychological impact of COVID-19 and other viral epidemics on frontline healthcare..., Cabarkapa [/bib_ref] [bib_ref] PTSD symptoms in healthcare workers facing the three coronavirus outbreaks: what can..., Carmassi [/bib_ref] [bib_ref] Child protection and resilience in the face of COVID-19 in South Africa:..., Fouche [/bib_ref] [bib_ref] Mental health effects of infection containment strategies: quarantine and isolation-a systematic review..., Henssler [/bib_ref] [bib_ref] Mental health outcomes of quarantine and isolation for infection prevention: a systematic..., Hossain [/bib_ref] [bib_ref] Psychological burden of quarantine in children and adolescents: a rapid systematic review..., Imran [/bib_ref] [bib_ref] Effectiveness of preventive measures against COVID-19: a systematic review of in silico..., Lahiri [/bib_ref] [bib_ref] Prevalence of anxiety in medical students during the COVID-19 pandemic: a rapid..., Lasheras [/bib_ref] [bib_ref] Suicidal behaviors and ideation during emerging viral disease outbreaks before the COVID-19..., Leaune [/bib_ref] [bib_ref] COVID-19 in older people: a rapid clinical review, Lithander [/bib_ref] [bib_ref] Rapid systematic review: the impact of social isolation and loneliness on the..., Loades [/bib_ref] [bib_ref] The psychological and mental impact of coronavirus disease 2019 (COVID-19) on medical..., Luo [/bib_ref] [bib_ref] Video calls for reducing social isolation and loneliness in older people: a..., Noone [/bib_ref] [bib_ref] Mental health during the Covid-19 outbreak in China: a meta-analysis, Ren [/bib_ref] [bib_ref] Violence against women during the COVID-19 pandemic: an integrative review, Sanchez [/bib_ref] [bib_ref] Studies of novel coronavirus disease 19 (COVID-19) pandemic: a global analysis of..., Tran [/bib_ref] [bib_ref] A rapid review of pandemicrelated behaviours and psychological outcomes, Usher [/bib_ref] [bib_ref] COVID-19: review of a 21st century pandemic from etiology to neuro-psychiatric implications, Yamamoto [/bib_ref] followed by systematic reviews on healthcare delivery, 1-13 and those on infection control. [bib_ref] Travel-related control measures to contain the COVID-19 pandemic: a rapid review, Burns [/bib_ref] [bib_ref] The current status and perspectives for the emerging pandemic: Covid-19, Kumari [/bib_ref] [bib_ref] Spread and impact of COVID-19 in China: a systematic review and synthesis..., Lin [/bib_ref] [bib_ref] COVID-19 in older people: a rapid clinical review, Lithander [/bib_ref] [bib_ref] Quarantine alone or in combination with other public health measures to control..., Nussbaumer-Streit [/bib_ref] [bib_ref] A systematic review of covid-19 epidemiology based on current evidence, Park [/bib_ref] [bib_ref] Non-pharmaceutical interventions for containment, mitigation and suppression of COVID-19 infection, Patino-Lugo [/bib_ref] [bib_ref] Secondary attack rate of COVID-19 in household contacts: systematic review, Shah [/bib_ref] [bib_ref] Studies of novel coronavirus disease 19 (COVID-19) pandemic: a global analysis of..., Tran [/bib_ref] [bib_ref] School closure and management practices during coronavirus outbreaks including COVID-19: a rapid..., Viner [/bib_ref] [bib_ref] How to improve adherence with quarantine: rapid review of the evidence, Webster [/bib_ref] The predominant areas of indirect health impacts covered by the included studies relate to the economic 9, [bib_ref] The psychological impact of quarantine and how to reduce it: rapid review..., Brooks [/bib_ref] [bib_ref] Travel-related control measures to contain the COVID-19 pandemic: a rapid review, Burns [/bib_ref] [bib_ref] The current status and perspectives for the emerging pandemic: Covid-19, Kumari [/bib_ref] [bib_ref] Child protection and resilience in the face of COVID-19 in South Africa:..., Fouche [/bib_ref] [bib_ref] Psychological burden of quarantine in children and adolescents: a rapid systematic review..., Imran [/bib_ref] [bib_ref] Prevalence of anxiety in medical students during the COVID-19 pandemic: a rapid..., Lasheras [/bib_ref] [bib_ref] Non-pharmaceutical interventions for containment, mitigation and suppression of COVID-19 infection, Patino-Lugo [/bib_ref] [bib_ref] Mental health during the Covid-19 outbreak in China: a meta-analysis, Ren [/bib_ref] [bib_ref] Spreading of SARS-CoV-2 in West Africa and assessment of risk factors, Tinto [/bib_ref] [bib_ref] Studies of novel coronavirus disease 19 (COVID-19) pandemic: a global analysis of..., Tran [/bib_ref] [bib_ref] A rapid review of pandemicrelated behaviours and psychological outcomes, Usher [/bib_ref] [bib_ref] School closure and management practices during coronavirus outbreaks including COVID-19: a rapid..., Viner [/bib_ref] [bib_ref] How to improve adherence with quarantine: rapid review of the evidence, Webster [/bib_ref] [bib_ref] COVID-19: review of a 21st century pandemic from etiology to neuro-psychiatric implications, Yamamoto [/bib_ref] and social impacts. 9, [bib_ref] Travel-related control measures to contain the COVID-19 pandemic: a rapid review, Burns [/bib_ref] [bib_ref] Psychological burden of quarantine in children and adolescents: a rapid systematic review..., Imran [/bib_ref] [bib_ref] Non-pharmaceutical interventions for containment, mitigation and suppression of COVID-19 infection, Patino-Lugo [/bib_ref] [bib_ref] Mental health during the Covid-19 outbreak in China: a meta-analysis, Ren [/bib_ref] [bib_ref] School closure and management practices during coronavirus outbreaks including COVID-19: a rapid..., Viner [/bib_ref] [bib_ref] COVID-19: review of a 21st century pandemic from etiology to neuro-psychiatric implications, Yamamoto [/bib_ref] Only 3 articles mentioned the negative impact on education. [bib_ref] The potential impact of the COVID-19 pandemic on child growth and development:..., Araujo [/bib_ref] [bib_ref] Prevalence of anxiety in medical students during the COVID-19 pandemic: a rapid..., Lasheras [/bib_ref] [bib_ref] School closure and management practices during coronavirus outbreaks including COVID-19: a rapid..., Viner [/bib_ref] - Quarantine: hospital employees had a high degree of post-traumatic stress symptoms which were strongly associated with exposure to SARS, quarantine and a relative or friend acquiring SARS. They also had the greatest risk for PTSD symptoms one-month later, and, this risk was increased even after home quarantine. Home quarantined HCWs had poorer sleep and a heightened degree of numbness than those who were not quarantined.
- Social isolation: a lack of family support and social isolation had a negative psychological impact on nurses who chose to isolate away from their families (Continued) To assess the psychological effects in both quarantined and isolated persons compared to non-quarantined and non-isolated persons
[formula] - Taiwan - USA - UK - Hong Kong - Canada - China - South Korea - Turkey - France - Singapore - Spain - NL - Australia - General population - Healthcare workers - Students - Home-based setting - Community environment - Inpatient level - Isolation - Quarantine [/formula]
- Isolation and quarantine: individuals experiencing isolation or quarantine were at increased risk for adverse mental health outcomes, particularly after containment duration of 1 week or longer. Effect sizes were summarized for depressive disorders, anxiety disorders and stress-related disorders. Elevated levels of anger were reported most consistently. There is compelling evidence for adverse mental health effects of isolation and quarantine, in particular depression, anxiety, stress-related disorders, and anger 29 Hossain, M. M., et al.
To synthesize the evidence on mental health outcomes of quarantine and isolation for preventing infectious diseases
[formula] - UK - USA - Hong Kong - Canada - Sweden - Netherlands - New Zealand - Ireland - Brazil - China Taiwan - Australia - Korea - Liberia - Sierra Leone - Senegal - Spain - Turkey, - Singapore - France - Patients with a pandemic infection - Providers - Students - Institutional stakeholders - Community members - Home-based setting - Community environment - Inpatient level - Isolation - Quarantine [/formula]
- Isolation and quarantine: it was reported a high burden of mental health problems among patients, informal caregivers and healthcare providers who experienced quarantine or isolation. Prevalent mental health problems among the affected individuals include depression, anxiety, mood disorders, psychological distress, posttraumatic stress disorder, insomnia, fear, stigmatization, low self-esteem, lack of self-control, and other adverse mental health outcomes 30 Imran, N., et al.To assess the impact of quarantine on mental health of children and adolescents, and proposes measures to improve psychological outcomes of isolation (Continued) To describe which non-pharmaceutical interventions used different countries and a when they use them. It also explores how Non-pharmaceutical interventions impact the number of cases, the mortality and the capacity of health systems - Telepsychotherapy is a trustworthy alternative to be adopted, which can be used efficaciously to treat common mental-health disorders such as anxiety, depression and post-traumatic distress. As well as in the traditional setting, a higher number of sessions and the proper management of patients' expectations seem to be asso-ciated with better outcomes. 11 - Quarantine, isolation and social distancing: rapid implementation of these public health strategies is the most effective, and indeed necessary, for containing viruses in pandemics, they also have many potentially negative sequelae and lead to a higher level of distress, fear and anxiety, and drive an increase in levels of panic and uncertainty. These measures are implemented very quickly without very much time for preparation. The rapidity of the change can (in itself) cause community alarm and anxiety. - Quarantine: evidence suggests a link between post-traumatic stress disorder (PTSD) and/or depression and quarantine. There is a positive correlation between length of quarantine and symptoms of PTSD. The psychological symptoms were higher among health-care workers relative to others - Social isolation: working from home, physical distancing, job loss and critical illness from the virus could induce long-term psychological effects in many individuals. Social isolation has been linked to a heightened risk of suicide attempts and suicide and several studies address the connection between job loss and a heightened risk of depression, anxiety and increased substance abuse. Social isolation has also been linked to domestic abuse and violence-related behaviours in the home - Other effects: economic and social impact [bib_ref] Preliminary trajectories in dietary behaviors during the COVID-19 pandemic: a public health..., Zupo [/bib_ref] To analyze the preliminary effects of the quarantine lifestyle from the standpoint of dietary habits
[formula] - Canada - Norway - Mexico - Finland - Sierra Leone - Denmark - USA - China - Italy - - Argentina - Australia - Brazil - Canada - Chile - China - Colombia - Cuba - Germany - Iran - Italy - Japan - Mexico - Norway - Russia - South Korea - Spain - United(Continued)- UK - Italy - China - Switzerland - USA - Brazil - Spain - Germany - Kenya - Canada - Australia - India - Netherlands - South Africa - Egypt - - Poland - India - Italy - Spain - China - Chile - Colombia - Brazil - General population - Home-based setting - Lockdown - Quarantine [/formula]
- Lockdown and quarantine: these results identified: i) a rise in consuption of carbohydrates; ii) more numerous snacks; iii) an high intake of fruits, vegetables and protein and iv) a decreased alcohol intake and fresh fish/seafood consumption. Data were scant on the consumption of junk foods.
## Direct health impact
## Mental health
Overall, almost half of the studies explore the impact of lockdown measures on mental health. While the rapid implementation of quarantine, isolation and social distancing measures seems to be the most effective strategy to contain the spread of the virus, these measures, when implemented at short notice, can produce alarm and anxiety. [bib_ref] A rapid review of pandemicrelated behaviours and psychological outcomes, Usher [/bib_ref] The studies reported a high burden of mental health problems among several groups of the population who experienced quarantine or isolation: patients, the general population and health workers. Prevalent mental health issues include anxiety, [bib_ref] The potential impact of the COVID-19 pandemic on child growth and development:..., Araujo [/bib_ref] [bib_ref] Impact of the COVID-19 pandemic on psychosocial health and well-being in South-Asian..., Banerjee [/bib_ref] [bib_ref] The psychological impact of quarantine and how to reduce it: rapid review..., Brooks [/bib_ref] [bib_ref] Mental health effects of infection containment strategies: quarantine and isolation-a systematic review..., Henssler [/bib_ref] [bib_ref] Mental health outcomes of quarantine and isolation for infection prevention: a systematic..., Hossain [/bib_ref] [bib_ref] Psychological burden of quarantine in children and adolescents: a rapid systematic review..., Imran [/bib_ref] [bib_ref] Prevalence of anxiety in medical students during the COVID-19 pandemic: a rapid..., Lasheras [/bib_ref] [bib_ref] Rapid systematic review: the impact of social isolation and loneliness on the..., Loades [/bib_ref] [bib_ref] Mental health during the Covid-19 outbreak in China: a meta-analysis, Ren [/bib_ref] [bib_ref] A rapid review of pandemicrelated behaviours and psychological outcomes, Usher [/bib_ref] [bib_ref] COVID-19: review of a 21st century pandemic from etiology to neuro-psychiatric implications, Yamamoto [/bib_ref] depression, [bib_ref] The potential impact of the COVID-19 pandemic on child growth and development:..., Araujo [/bib_ref] [bib_ref] Impact of the COVID-19 pandemic on psychosocial health and well-being in South-Asian..., Banerjee [/bib_ref] [bib_ref] Mental health effects of infection containment strategies: quarantine and isolation-a systematic review..., Henssler [/bib_ref] [bib_ref] Mental health outcomes of quarantine and isolation for infection prevention: a systematic..., Hossain [/bib_ref] [bib_ref] Rapid systematic review: the impact of social isolation and loneliness on the..., Loades [/bib_ref] [bib_ref] Mental health during the Covid-19 outbreak in China: a meta-analysis, Ren [/bib_ref] [bib_ref] COVID-19: review of a 21st century pandemic from etiology to neuro-psychiatric implications, Yamamoto [/bib_ref] post-traumatic stress disorder (PTSD), stress, [bib_ref] The potential impact of the COVID-19 pandemic on child growth and development:..., Araujo [/bib_ref] [bib_ref] The psychosocial impact of flu influenza pandemics on healthcare workers and lessons..., Barello [/bib_ref] [bib_ref] The psychological impact of quarantine and how to reduce it: rapid review..., Brooks [/bib_ref] [bib_ref] The potential impact of COVID-19 on psychosis: a rapid review of contemporary..., Brown [/bib_ref] [bib_ref] The psychological impact of COVID-19 and other viral epidemics on frontline healthcare..., Cabarkapa [/bib_ref] [bib_ref] PTSD symptoms in healthcare workers facing the three coronavirus outbreaks: what can..., Carmassi [/bib_ref] [bib_ref] Mental health effects of infection containment strategies: quarantine and isolation-a systematic review..., Henssler [/bib_ref] [bib_ref] Mental health outcomes of quarantine and isolation for infection prevention: a systematic..., Hossain [/bib_ref] [bib_ref] Psychological burden of quarantine in children and adolescents: a rapid systematic review..., Imran [/bib_ref] [bib_ref] Rapid systematic review: the impact of social isolation and loneliness on the..., Loades [/bib_ref] [bib_ref] A rapid review of pandemicrelated behaviours and psychological outcomes, Usher [/bib_ref] [bib_ref] COVID-19: review of a 21st century pandemic from etiology to neuro-psychiatric implications, Yamamoto [/bib_ref] and stigmatization. In particular among children, older people and health workers the evidence suggests a link between PTSD and quarantine or isolation. [bib_ref] The psychological impact of quarantine and how to reduce it: rapid review..., Brooks [/bib_ref] [bib_ref] The psychological impact of COVID-19 and other viral epidemics on frontline healthcare..., Cabarkapa [/bib_ref] [bib_ref] Mental health outcomes of quarantine and isolation for infection prevention: a systematic..., Hossain [/bib_ref] [bib_ref] Psychological burden of quarantine in children and adolescents: a rapid systematic review..., Imran [/bib_ref] [bib_ref] Rapid systematic review: the impact of social isolation and loneliness on the..., Loades [/bib_ref] [bib_ref] COVID-19: review of a 21st century pandemic from etiology to neuro-psychiatric implications, Yamamoto [/bib_ref] Stigma is linked both to quarantine and isolation [bib_ref] Mental health outcomes of quarantine and isolation for infection prevention: a systematic..., Hossain [/bib_ref] and particularly experienced by health workers 21 and children [bib_ref] Psychological burden of quarantine in children and adolescents: a rapid systematic review..., Imran [/bib_ref] [bib_ref] Studies of novel coronavirus disease 19 (COVID-19) pandemic: a global analysis of..., Tran [/bib_ref] ; the two groups experienced stigma due to quarantine even if they had been confirmed to be negative. [bib_ref] Psychological burden of quarantine in children and adolescents: a rapid systematic review..., Imran [/bib_ref] [bib_ref] Studies of novel coronavirus disease 19 (COVID-19) pandemic: a global analysis of..., Tran [/bib_ref] Health care delivery
The pandemic and the subsequent lockdown measures had a negative impact on health care delivery, resulting in limitations to available health care services. These restrictions included: the postponement of non-urgent outpatient visits and of non-urgent surgical interventions, the reduction of non-essential services, and restrictions in accessing hospitals for both patients and their caregivers. [bib_ref] Systematic rapid living review on rehabilitation needs due to COVID-19: update to..., Andrenelli [/bib_ref] The included studies find that restrictions of health care services posed enormous challenges to patients and health care providers, and telemedicine has been proposed by several authors as a potential solution to overcoming the barrier in accessing health care services, especially for outpatient care. 3,5-11, [bib_ref] Applications of e-health to support personcentered health care at the time of..., Tebeje [/bib_ref] Tele-psychotherapy 8,11 has been evaluated in treating common mental-health disorders such as anxiety, depression and PTSD. The use of telemedicine has also been investigated in orthopaedic care. 3,7 The resulting reduction in inpatient and outpatient orthopaedic care and the increase in remote orthopaedic care was associated with high patient satisfaction related to convenience and reduced waiting and travelling times. Evidence suggests that telemedicine in orthopaedic care can be safe and cost-effective, with high patient and clinician satisfaction. 7
The restrictions of rehabilitation services due to lockdown measures increased, especially among older people, the risk of frailty, sarcopenia, dementia, cognitive decline and depression, in particular among those quarantined at home or with restricted mobility. [bib_ref] Systematic rapid living review on rehabilitation needs due to COVID-19: update to..., Andrenelli [/bib_ref] Yet, a systematic review on telerehabilitation identified 53 challenges in the literature (e.g.: on sustainability and privacy) that could affect the development of tele-rehabilitation. Finally, a systematic review on the delivery of urogynaecology care using telemedicine 6 identified the clinical situations that would allow virtual settings and those that should be managed with an in-person visit despite the risks of COVID-19 transmission.
## Infection control
The effect of lockdown measures on infection control was investigated in 12 systematic reviews. [bib_ref] Travel-related control measures to contain the COVID-19 pandemic: a rapid review, Burns [/bib_ref] [bib_ref] The current status and perspectives for the emerging pandemic: Covid-19, Kumari [/bib_ref] [bib_ref] Spread and impact of COVID-19 in China: a systematic review and synthesis..., Lin [/bib_ref] [bib_ref] COVID-19 in older people: a rapid clinical review, Lithander [/bib_ref] [bib_ref] Quarantine alone or in combination with other public health measures to control..., Nussbaumer-Streit [/bib_ref] [bib_ref] A systematic review of covid-19 epidemiology based on current evidence, Park [/bib_ref] [bib_ref] Non-pharmaceutical interventions for containment, mitigation and suppression of COVID-19 infection, Patino-Lugo [/bib_ref] [bib_ref] Secondary attack rate of COVID-19 in household contacts: systematic review, Shah [/bib_ref] [bib_ref] Studies of novel coronavirus disease 19 (COVID-19) pandemic: a global analysis of..., Tran [/bib_ref] [bib_ref] School closure and management practices during coronavirus outbreaks including COVID-19: a rapid..., Viner [/bib_ref] [bib_ref] How to improve adherence with quarantine: rapid review of the evidence, Webster [/bib_ref] According to Chandana et al ., [bib_ref] The current status and perspectives for the emerging pandemic: Covid-19, Kumari [/bib_ref] quarantine is one 'of the most misunderstood and feared methods of controlling COVID-19, because it may affect both infected and noninfected individuals with psychological, economical and emotional complications such as post-traumatic stress disorder, depression, insomnia, mood swings'. They continue that the lockdown of a city 'was proved to be effective when a study reported 72% drop in the number of infected people'. [bib_ref] The current status and perspectives for the emerging pandemic: Covid-19, Kumari [/bib_ref] A systematic review conducted in China 35 emphasises that the lockdown of a city reduced the reproduction number (R0) from 3.77 to a controlled reproduction number (Rc) of 1.88 after lockdown. Other public health measures implemented, apart from citywide lockdowns and, encompassing contact tracing, intensification of screening, quarantine and mask utilisation, may also be contributing to containing the spread of the virus. [bib_ref] Spread and impact of COVID-19 in China: a systematic review and synthesis..., Lin [/bib_ref] In fact, some systematic reviews suggest that combinations of different control measures are the most effective way to reduce transmission of the disease, prevent the collapse of health care services and reduce mortality. [bib_ref] Quarantine alone or in combination with other public health measures to control..., Nussbaumer-Streit [/bib_ref] [bib_ref] Non-pharmaceutical interventions for containment, mitigation and suppression of COVID-19 infection, Patino-Lugo [/bib_ref] Concerning travel restrictions, a systematic review on COVID-19, SARS and MERS suggested that travel restrictions leading to reduced transmissibility can be highly effective in containing the spread. [bib_ref] A systematic review of covid-19 epidemiology based on current evidence, Park [/bib_ref] In line with these results are those retrieved by the Cochrane Systematic Reviews developed by Burns et al ., [bib_ref] Travel-related control measures to contain the COVID-19 pandemic: a rapid review, Burns [/bib_ref] which found that travel-related control measures during the COVID-19 pandemic may have a positive impact on infectious disease outcomes. In particular, travel restrictions may limit the spread of disease across national borders, while entry and exit symptom screening measures on their own are not likely to be effective. The evidence is scant on the effectiveness of travel-related quarantine [bib_ref] Travel-related control measures to contain the COVID-19 pandemic: a rapid review, Burns [/bib_ref] and there is very low-certainty evidence on the effect of quarantine of travellers from a country with a declared outbreak on reducing incidence and death. [bib_ref] Quarantine alone or in combination with other public health measures to control..., Nussbaumer-Streit [/bib_ref] Finally, systematic reviews on the impact of school closures found that they do not seem to be effective [bib_ref] A systematic review of covid-19 epidemiology based on current evidence, Park [/bib_ref] and do not contribute to the control of the epidemic. 50
## Children, child development and desire for parenthood
Six systematic reviews on children and their development [bib_ref] The potential impact of the COVID-19 pandemic on child growth and development:..., Araujo [/bib_ref] [bib_ref] Practical recommendations for maintaining active lifestyle during the COVID-19 pandemic: a systematic..., Bentlage [/bib_ref] [bib_ref] Child protection and resilience in the face of COVID-19 in South Africa:..., Fouche [/bib_ref] [bib_ref] Rapid systematic review: the impact of social isolation and loneliness on the..., Loades [/bib_ref] [bib_ref] A systematic review of covid-19 epidemiology based on current evidence, Park [/bib_ref] [bib_ref] School closure and management practices during coronavirus outbreaks including COVID-19: a rapid..., Viner [/bib_ref] have been included in our study. The focus on the limited effect of school closures on pandemic control, 42,50 as discussed above, and on adverse effects of school closures on issues including: increased risk of transmission from children to grandparents, harms to child welfare particularly among the most vulnerable pupils, nutritional issues and the loss of teaching/learning and socialization processes. Importantly, children miss out on public policies taking place in schools, such as balanced and free food programs, guidance about personal hygiene, physical activity and citizenship initiatives. [bib_ref] School closure and management practices during coronavirus outbreaks including COVID-19: a rapid..., Viner [/bib_ref] Social isolation in children may increase the risk for cardiovascular disease, reduce physical activity and have negative effects on mental health, 20,50 such as an increased likelihood of high rates of depression and anxiety during and after enforced isolation. [bib_ref] Rapid systematic review: the impact of social isolation and loneliness on the..., Loades [/bib_ref] Quarantine in children is linked to anxiety, stress and depression and can become a risk factor for child growth and development. [bib_ref] The potential impact of the COVID-19 pandemic on child growth and development:..., Araujo [/bib_ref] Isolation and quarantine together are related to an increased risk of requiring mental health services and to higher levels of post-traumatic stress. [bib_ref] Rapid systematic review: the impact of social isolation and loneliness on the..., Loades [/bib_ref] A systematic review found that during quarantine, despite a reduction in the quality of life, there was an increased desire for parenthood, although it is unknown if these changes are associated with an increase in terms of birth rates. 39
## Older people
Despite quarantine and isolation being the most effective strategies for prevention of the secondary transmission of disease, the evidence suggests a greater vulnerability of older people for secondary transmission than other household members. [bib_ref] Secondary attack rate of COVID-19 in household contacts: systematic review, Shah [/bib_ref] Other negative consequences were also experienced, particularly if quarantine and isolation were in place for an extended period, and the loneliness caused by social isolation has been associated with impaired cognitive function in older adults. [bib_ref] COVID-19 in older people: a rapid clinical review, Lithander [/bib_ref] Lockdown in older people with a subsequent reduction in social participation and physical activity during home confinement was identified as a serious concern, as they are typically more inactive and more disposed to chronic disease. [bib_ref] Impact of the COVID-19 pandemic on psychosocial health and well-being in South-Asian..., Banerjee [/bib_ref] [bib_ref] Practical recommendations for maintaining active lifestyle during the COVID-19 pandemic: a systematic..., Bentlage [/bib_ref] Finally, a systematic review on older people in nursing homes emphasized that older people suffer from social distancing due to isolation and confinement. The evidence on this however was limited because only few studies with a small sample size and using unreliable methods were included in this systematic review. 40
## Well-being and quality of life
Only five systematic reviews were retrieved on well-being and quality of life (QOL). [bib_ref] Impact of the COVID-19 pandemic on psychosocial health and well-being in South-Asian..., Banerjee [/bib_ref] [bib_ref] The psychosocial impact of flu influenza pandemics on healthcare workers and lessons..., Barello [/bib_ref] [bib_ref] The psychological impact of quarantine and how to reduce it: rapid review..., Brooks [/bib_ref] [bib_ref] Reported quality of life in countries with cases of COVID19: a systematic..., Melo-Oliveira [/bib_ref] [bib_ref] Mental health during the Covid-19 outbreak in China: a meta-analysis, Ren [/bib_ref] Importantly, four systematic reviews explored the impact of lockdown measures on health workers in terms of well-being and QOL. [bib_ref] Impact of the COVID-19 pandemic on psychosocial health and well-being in South-Asian..., Banerjee [/bib_ref] [bib_ref] The psychosocial impact of flu influenza pandemics on healthcare workers and lessons..., Barello [/bib_ref] [bib_ref] The psychological impact of quarantine and how to reduce it: rapid review..., Brooks [/bib_ref] [bib_ref] Mental health during the Covid-19 outbreak in China: a meta-analysis, Ren [/bib_ref] According to the evidence summarised in these studies, healthcare professionals who had been quarantined had more severe symptoms of post-traumatic stress than the general population, felt stigmatised, presented more avoidance behaviours, reported huger lost income and were more affected at the psychological level. [bib_ref] The psychological impact of quarantine and how to reduce it: rapid review..., Brooks [/bib_ref] Quarantine in the general population was linked to a reduction of the mean wellbeing scores, [bib_ref] Reported quality of life in countries with cases of COVID19: a systematic..., Melo-Oliveira [/bib_ref] work-related stress, burnout, [bib_ref] The psychosocial impact of flu influenza pandemics on healthcare workers and lessons..., Barello [/bib_ref] frustration, fears of infection, boredom, inadequate supplies and inadequate information. [bib_ref] The psychological impact of quarantine and how to reduce it: rapid review..., Brooks [/bib_ref] Finally, lockdown and social distancing were linked in the general population to a negative psychosocial impact, an increased prevalence of depression, anxiety, sleep, alcohol use disorders and the fear of being infected. People were also worried about their loved ones. [bib_ref] Impact of the COVID-19 pandemic on psychosocial health and well-being in South-Asian..., Banerjee [/bib_ref] [bib_ref] The psychosocial impact of flu influenza pandemics on healthcare workers and lessons..., Barello [/bib_ref] [bib_ref] Mental health during the Covid-19 outbreak in China: a meta-analysis, Ren [/bib_ref]
## Substance abuse
The four systematic reviews [bib_ref] Domestic violence and substance abuse during COVID19: a systematic review, Abdo [/bib_ref] [bib_ref] Impact of the COVID-19 pandemic on psychosocial health and well-being in South-Asian..., Banerjee [/bib_ref] [bib_ref] Child protection and resilience in the face of COVID-19 in South Africa:..., Fouche [/bib_ref] [bib_ref] COVID-19: review of a 21st century pandemic from etiology to neuro-psychiatric implications, Yamamoto [/bib_ref] focussed on the correlation of infection control measures and substance abuse found that lockdown was associated with increased alcohol use disorders in the general population, [bib_ref] Impact of the COVID-19 pandemic on psychosocial health and well-being in South-Asian..., Banerjee [/bib_ref] and social isolation and quarantine were identified as potential contributors to the aggravation of substance abuse. [bib_ref] Domestic violence and substance abuse during COVID19: a systematic review, Abdo [/bib_ref] [bib_ref] COVID-19: review of a 21st century pandemic from etiology to neuro-psychiatric implications, Yamamoto [/bib_ref] Violence and abuse A link between lockdown and domestic violence and abuse was identified in four systematic reviews, [bib_ref] Domestic violence and substance abuse during COVID19: a systematic review, Abdo [/bib_ref] [bib_ref] Child protection and resilience in the face of COVID-19 in South Africa:..., Fouche [/bib_ref] [bib_ref] Violence against women during the COVID-19 pandemic: an integrative review, Sanchez [/bib_ref] [bib_ref] COVID-19: review of a 21st century pandemic from etiology to neuro-psychiatric implications, Yamamoto [/bib_ref] with three of them [bib_ref] Domestic violence and substance abuse during COVID19: a systematic review, Abdo [/bib_ref] [bib_ref] Child protection and resilience in the face of COVID-19 in South Africa:..., Fouche [/bib_ref] [bib_ref] COVID-19: review of a 21st century pandemic from etiology to neuro-psychiatric implications, Yamamoto [/bib_ref] also exploring substance abuse (see previous section).
Social isolation was linked to domestic abuse and violencerelated behaviour in the home. [bib_ref] COVID-19: review of a 21st century pandemic from etiology to neuro-psychiatric implications, Yamamoto [/bib_ref] A systematic review identified that some factors increasing women's vulnerabilities to violence were exacerbated during the social distancing and lockdown period. [bib_ref] Violence against women during the COVID-19 pandemic: an integrative review, Sanchez [/bib_ref] Even quarantine can increase the power and control abusers hold over victims and trigger violence. [bib_ref] Domestic violence and substance abuse during COVID19: a systematic review, Abdo [/bib_ref] [bib_ref] Violence against women during the COVID-19 pandemic: an integrative review, Sanchez [/bib_ref] To overcome this issue with regard to children, South Africa's strict lockdown offered protective pathways, including a policy to protect children at risk of abuse. 28
## Lifestyle and dietary habits
Among the 51 systematic reviews included in our study, only two 20,53 focussed on lifestyle and dietary habits. Lockdown and quarantine were found to be associated with an increase of carbohydrate consumption, as well as more frequent consumption of snacks, although together with a high consumption of fruits and vegetables, and protein sources. [bib_ref] Practical recommendations for maintaining active lifestyle during the COVID-19 pandemic: a systematic..., Bentlage [/bib_ref] [bib_ref] Preliminary trajectories in dietary behaviors during the COVID-19 pandemic: a public health..., Zupo [/bib_ref] Social isolation was found to cause a decrease in physical activity and, for children, a decrease in the time devoted to sports, and an increase in time sleeping and spent in front of screens, potentially increasing overweight and obesity among children. [bib_ref] Practical recommendations for maintaining active lifestyle during the COVID-19 pandemic: a systematic..., Bentlage [/bib_ref] [bib_ref] Preliminary trajectories in dietary behaviors during the COVID-19 pandemic: a public health..., Zupo [/bib_ref]
## Indirect health impact
The areas of indirect health impact 9, [bib_ref] The potential impact of the COVID-19 pandemic on child growth and development:..., Araujo [/bib_ref] [bib_ref] The psychological impact of quarantine and how to reduce it: rapid review..., Brooks [/bib_ref] [bib_ref] Travel-related control measures to contain the COVID-19 pandemic: a rapid review, Burns [/bib_ref] [bib_ref] The current status and perspectives for the emerging pandemic: Covid-19, Kumari [/bib_ref] [bib_ref] Child protection and resilience in the face of COVID-19 in South Africa:..., Fouche [/bib_ref] [bib_ref] Psychological burden of quarantine in children and adolescents: a rapid systematic review..., Imran [/bib_ref] [bib_ref] Prevalence of anxiety in medical students during the COVID-19 pandemic: a rapid..., Lasheras [/bib_ref] [bib_ref] Non-pharmaceutical interventions for containment, mitigation and suppression of COVID-19 infection, Patino-Lugo [/bib_ref] [bib_ref] Mental health during the Covid-19 outbreak in China: a meta-analysis, Ren [/bib_ref] [bib_ref] Spreading of SARS-CoV-2 in West Africa and assessment of risk factors, Tinto [/bib_ref] [bib_ref] Studies of novel coronavirus disease 19 (COVID-19) pandemic: a global analysis of..., Tran [/bib_ref] [bib_ref] A rapid review of pandemicrelated behaviours and psychological outcomes, Usher [/bib_ref] [bib_ref] School closure and management practices during coronavirus outbreaks including COVID-19: a rapid..., Viner [/bib_ref] [bib_ref] How to improve adherence with quarantine: rapid review of the evidence, Webster [/bib_ref] [bib_ref] COVID-19: review of a 21st century pandemic from etiology to neuro-psychiatric implications, Yamamoto [/bib_ref] identified in the included studies concern the economic and social impact, the impact on education and the lack of supplies and food .
Overall, the non-pharmaceutical interventions implemented to contain the virus, such as quarantine, isolation, social distancing and community containment, were noted to have important economic [bib_ref] The psychological impact of quarantine and how to reduce it: rapid review..., Brooks [/bib_ref] [bib_ref] The current status and perspectives for the emerging pandemic: Covid-19, Kumari [/bib_ref] [bib_ref] Child protection and resilience in the face of COVID-19 in South Africa:..., Fouche [/bib_ref] [bib_ref] Psychological burden of quarantine in children and adolescents: a rapid systematic review..., Imran [/bib_ref] [bib_ref] Non-pharmaceutical interventions for containment, mitigation and suppression of COVID-19 infection, Patino-Lugo [/bib_ref] [bib_ref] Studies of novel coronavirus disease 19 (COVID-19) pandemic: a global analysis of..., Tran [/bib_ref] [bib_ref] A rapid review of pandemicrelated behaviours and psychological outcomes, Usher [/bib_ref] [bib_ref] How to improve adherence with quarantine: rapid review of the evidence, Webster [/bib_ref] [bib_ref] COVID-19: review of a 21st century pandemic from etiology to neuro-psychiatric implications, Yamamoto [/bib_ref] and social consequences. [bib_ref] The current status and perspectives for the emerging pandemic: Covid-19, Kumari [/bib_ref] [bib_ref] Psychological burden of quarantine in children and adolescents: a rapid systematic review..., Imran [/bib_ref] [bib_ref] Non-pharmaceutical interventions for containment, mitigation and suppression of COVID-19 infection, Patino-Lugo [/bib_ref] [bib_ref] Mental health during the Covid-19 outbreak in China: a meta-analysis, Ren [/bib_ref] [bib_ref] COVID-19: review of a 21st century pandemic from etiology to neuro-psychiatric implications, Yamamoto [/bib_ref] In particular, quarantine was associated with the necessity to work, the fear of loss of income, the lost income itself and a reduction in overall productivity resulting in a decline of economic growth. [bib_ref] The psychological impact of quarantine and how to reduce it: rapid review..., Brooks [/bib_ref] [bib_ref] The current status and perspectives for the emerging pandemic: Covid-19, Kumari [/bib_ref] Moreover, some systematic reviews [bib_ref] The psychological impact of quarantine and how to reduce it: rapid review..., Brooks [/bib_ref] [bib_ref] Child protection and resilience in the face of COVID-19 in South Africa:..., Fouche [/bib_ref] [bib_ref] Psychological burden of quarantine in children and adolescents: a rapid systematic review..., Imran [/bib_ref] [bib_ref] A rapid review of pandemicrelated behaviours and psychological outcomes, Usher [/bib_ref] identified other fundamental issues, such as the lack or insecurity of supplies and food, and inadequate information, particularly linked to quarantine.
School closures were associated with a loss in teaching/learning and education, as well as with wider social impact and economic harm on working parents, health workers and other key workers being forced from work to care for children at home. [bib_ref] The potential impact of the COVID-19 pandemic on child growth and development:..., Araujo [/bib_ref] [bib_ref] School closure and management practices during coronavirus outbreaks including COVID-19: a rapid..., Viner [/bib_ref] Moreover, a systematic review 33 on the prevalence of anxiety in medical students during the pandemic identified concerns on economic impact, academic delay, curricular factors and impact on their daily life.
Travel-related control measures related to quarantine had far-reaching economic, social, legal, ethical and political implications. [bib_ref] Travel-related control measures to contain the COVID-19 pandemic: a rapid review, Burns [/bib_ref] Some populations, such as in west Africa, [bib_ref] Spreading of SARS-CoV-2 in West Africa and assessment of risk factors, Tinto [/bib_ref] had difficulties complying with certain measures, such as travel limitations and the closure of markets and places of worship, as the majority of people work in the informal sector, including trading, other businesses, transport and restoration and these jobs are not subject to social protection.
# Discussion
This systematic review set out to systematically review the evidence published in systematic reviews on the health impact of staying at home, social distancing and lockdown measures. A number of important findings emerged.
The first relates to the areas that have been studies so far. We intentionally kept a broad focus on all policy areas that are associated with the social determinants of health. Surprisingly, almost half of the studies (25 of 51) explore the impact of lockdown measures on mental health, with the common finding that these measures put a strain on the mental health of patients, the health workers and the general population. The second most commonly studied area, explored in 14 of the 51 included studies, was concerned with health care delivery. Many of these 14 systematic reviews explore the issue of telemedicine, with only indirect references to the Coronavirus pandemic. The impact of lockdown measures on containing the spread of the virus was explored in 12 studies, with the overall finding that these measures are successful and most promising when used in combination. In general, lockdown measures are enacted to contain the virus, but often discontinued for economic or political rather than purely epidemiological reasons. Other areas of the health impact of lockdown measures have received far less attention so far and warrant further research.
A second key finding of our study highlights that the complex and multifactorial nature of the health impact of lockdown measures, which can be both direct and indirect. While the closure of schools, for example, will have a direct impact on the education, mental and physical health of children, an indirect impact is that parents will have to stay at home to look after young children, preventing them from going to work. While our primary interest was on the impact of lockdown measures, it was sometimes difficult to ascertain whether the impact was due to these measures or the pandemic itself. We found that many studies were struggling with the same challenge. Causal pathways are often blurred, as mental health, for example, can be affected by both, policy measures and the pandemic itself. Policy measures aimed at containing the spread of the virus will have to mindful of direct and indirect impacts and intended and unintended consequences.
A third key finding relates to the strength of evidence gathered by October 2020. Unsurprisingly, the evidence on the topic was still mainly focused on the first wave of the COVID-19 pandemic that occurred in spring 2020 and a renewed search of the literature is needed to capture more up-to-date evidence. We also identified methodological and terminological challenges. With regard to the methods used, some narrative reviews are defined by the authors as systematic reviews and vice versa. Furthermore, in many systematic reviews, conclusions are drawn based on a very limited number of papers with often low quality. In addition, in some systematic reviews, the impact of lockdown measures is mainly described in the introduction and the conclusions, rather than in the results section. There is also a need for more terminological clarity. Some authors misuse the terms 'isolation' and 'quarantine' and confuse 'social isolation' with 'isolation'.
## Supplementary data
Supplementary data are available at the Journal of Public Health online.
[fig] Figure 1: PRISMA diagram of systematic article selection. [/fig]
[table] Table 1: Main characteristics of the studies included [/table]
|
Exercise and Atrial Fibrillation: Some Good News and Some Bad News
Atrial fibrillation (AF) is considered as the most common sustained arrhythmia in adults, whose incidence rate is on the rise due to the increase in the mean age of the global population. In recent years, many efforts have been made to identify effective factors in the incidence of AF to prevent them and thereby reduce the consequences of AF. Physical activity is one of the topics that attracted much attention in the last two decades. According to some findings, extreme and prolonged exercise itself can be considered as a risk factor for the onset of AF; however, other studies have shown that exercise can be regarded a protective factor against AF in the general population. The present study reviews the findings of studies on the relationship between AF and exercise and discusses possible mechanisms for this relationship. Additionally, we present some recommendations for researchers and physicians about exercise management in association with AF prevention.
# Introduction
A trial fibrillation (AF) is the most commonly reported sustained arrhythmia in adults, with an estimated prevalence of 1-2% in the general population [bib_ref] A roadmap to improve the quality of atrial fibrillation management: proceedings from..., Kirchhof [/bib_ref] [bib_ref] Screening for atrial fibrillation: a European Heart Rhythm Association (EHRA) consensus document..., Mairesse [/bib_ref]. Due to the aging population of the world and the increasing prevalence of risk factors associated with AF, we are expected to witness an AF epidemic in the coming decades [bib_ref] ESC Guidelines for the management of atrial fibrillation developed in collaboration with..., Kirchhof [/bib_ref]. Therefore, the identification of possible mechanisms and triggers of AF, especially the detection of preventable agents, and appropriate preventive programs can play a significant role in promoting community health and reducing the costs associat-ed with disease management [bib_ref] Cost of an emerging epidemic: an economic analysis of atrial fibrillation in..., Stewart [/bib_ref]. The general belief is that regular exercise can improve cardiovascular health and reduce cardiac events [bib_ref] A roadmap to improve the quality of atrial fibrillation management: proceedings from..., Kirchhof [/bib_ref] [bib_ref] Is exercise becoming a danger for our health? The complex relationship between..., Seccia [/bib_ref]. In fact, physical activity is considered as an appropriate measure in preventing the risk of heart diseases by controlling weight gain and blood pressure as well as enhancing cardiometabolic efficacy [bib_ref] ESC Guidelines for the management of atrial fibrillation developed in collaboration with..., Kirchhof [/bib_ref] [bib_ref] Emerging risk factors and the dose-response relationship between physical activity and lone..., Calvo [/bib_ref]. Indeed, studies have shown that regular exercise leads to a reduction of 30-50% in all-cause mortality [bib_ref] Exercise and Atrial Fibrillation: Prevention or Causation?, Elliott [/bib_ref]. However, recent evidence suggests that extreme and prolonged exercise itself can be considered as a risk factor for the onset of AF [bib_ref] Preparticipation cardiovascular evaluation for athletic participants to prevent sudden death: Position paper..., Mont [/bib_ref].
On the other hand, many studies have shown that exercise can be a protective factor for AF GMJ.2018;7:e1401 ww.gmj.ir in the general population [bib_ref] Risk factor management in atrial fibrillation, Brandes [/bib_ref]. Over the past two decades, more than 40 different studies have attempted to provide evidence of how physical exercise is linked to the risk of developing AF [bib_ref] Is exercise becoming a danger for our health? The complex relationship between..., Seccia [/bib_ref] [bib_ref] Exercise and Atrial Fibrillation: Prevention or Causation?, Elliott [/bib_ref] [bib_ref] The role of exercise in atrial fibrillation prevention and promotion: finding optimal..., Elliott [/bib_ref]. Most of these studies are design-oriented in two overall categories: 1) studies that have taken place in the general population with the purpose of showing the effects of physical activity or inactivity on the risk of AF; and 2) studies that have been performed in athletes, sometimes compared with the control group, aimed at revealing the effects of intense exercise on the risk of AF. In this paper, we review the findings of a variety of studies in this area and discuss possible mechanisms regarding the relationship between exercise and AF. Lastly, we present some recommendations for further research in this field.
## Studies in the general population
More than ten population-based studies have been conducted so far, to investigate the impact of physical activity on the risk of AF incidence [bib_ref] Is exercise becoming a danger for our health? The complex relationship between..., Seccia [/bib_ref] [bib_ref] The role of exercise in atrial fibrillation prevention and promotion: finding optimal..., Elliott [/bib_ref]. In the majority of these studies, evidence suggests that mild to moderate physical activity can reduce the risk of AF in the community, whereas only few studies provide evidence of the ineffectiveness of regular exercise in attenuating the risk of AF [bib_ref] Exercise and Atrial Fibrillation: Prevention or Causation?, Elliott [/bib_ref] [bib_ref] Relation of vigorous exercise to risk of atrial fibrillation, Aizer [/bib_ref] [bib_ref] Resting heart rate and physical activity as risk factors for lone atrial..., Thelle [/bib_ref]. Results of the Aizer et al. study in a population of more than 16,000 people showed that habitual vigorous exercise for 5-7 days per week in people under 50 years of age increases the risk of AF, while there is no association between the incidence of AF and exercise in less time duration and at ages older than 50 years [bib_ref] Relation of vigorous exercise to risk of atrial fibrillation, Aizer [/bib_ref]. Similarly, in another study in Sweden, higher level of exercise in men of lower age was associated with an increased risk of AF, while in older men, cycling/walking was associated with reduced AF risk [bib_ref] Atrial fibrillation is associated with different levels of physical activity levels at..., Drca [/bib_ref]. Two studies from Norway and Finland showed that physical activity in women [bib_ref] Resting heart rate and physical activity as risk factors for lone atrial..., Thelle [/bib_ref] and men [bib_ref] Long-Term Change in Cardiorespiratory Fitness in Relation to Atrial Fibrillation and Heart..., Khan [/bib_ref] was not associated with a decrease in AF incidence.
In contrast with the studies showing a non-significant or reversed relationship between physical activity and AF burden, other studies have suggested that daily walking or cycling can reduce the risk of AF [bib_ref] Is exercise becoming a danger for our health? The complex relationship between..., Seccia [/bib_ref] [bib_ref] Exercise and Atrial Fibrillation: Prevention or Causation?, Elliott [/bib_ref] [bib_ref] Extreme Exercise Hypothesis": Recent Findings and Cardiovascular Health Implications, Eijsvogels [/bib_ref]. In a study in Sweden, it was found that moderate physi-cal activity could be effective in reducing the risk of AF in women (related risk: 0.85 for ≥4 h/week vs. <1 h/week) [bib_ref] Physical activity is associated with a reduced risk of atrial fibrillation in..., Drca [/bib_ref]. In the Cardiovascular Health Study, it was concluded that the incidence of AF in older adults with more regular physical is around a half lower than those with less physical activity [bib_ref] Physical Activity and Incidence of Atrial Fibrillation in Older Adults: The Cardiovascular..., Mittal [/bib_ref].
In recent years, two Asian studies were also conducted to complete the Asian section of the puzzle of the relationship between exercise and AF. In both of these studies, which were collectively performed in a population of about 375,000 people, it was shown that moderate physical activity plays a protective role against the onset of AF [bib_ref] Association between modifiable lifestyle and the prevalence of atrial fibrillation in a..., Yang [/bib_ref] [bib_ref] 136-75: Gender difference of life style patterns including smoking, alcohol, and physical..., Park [/bib_ref].
Recently, some other evidences have approved what was shown by most of the population-based studies [bib_ref] Exercise and Atrial Fibrillation: Prevention or Causation?, Elliott [/bib_ref] [bib_ref] Cardiorespiratory fitness and risk of incident atrial fibrillation: results from the Henry..., Qureshi [/bib_ref]. The findings of these studies, which include three cohort studies by 2018, suggest that better cardiorespiratory fitness is associated with a 30-60% lower risk of AF during follow-up of 5-19 years [bib_ref] Exercise and Atrial Fibrillation: Prevention or Causation?, Elliott [/bib_ref]. Another study also found that any 1-metabolic equivalent (MET) increase in cardiorespiratory fitness could reduce the risk of AF by 7% (1 MET = a whole-body oxygen consumption of 3.5 mL O2/kg/min) [bib_ref] Cardiorespiratory fitness and risk of incident atrial fibrillation: results from the Henry..., Qureshi [/bib_ref]. Also, in other studies in the United States and Norway, results are also in favor of the effectiveness of exercise in reducing the incidence of AF [bib_ref] A roadmap to improve the quality of atrial fibrillation management: proceedings from..., Kirchhof [/bib_ref] [bib_ref] Is exercise becoming a danger for our health? The complex relationship between..., Seccia [/bib_ref] [bib_ref] Physical activity volume in relation to risk of atrial fibrillation. A non-linear..., Ricci [/bib_ref] [bib_ref] Empowerment of athletes with cardiac disorders: a new paradigm, Providencia [/bib_ref]. Although many population-based studies have been carried out so far, we still cannot see a very precise model for estimating the effect of minimal physical activity, depending on its condition and type, in relation to AF risk; however, the beneficial effects of exercise are unobtainable on the overall pattern of heart disease risk. By examining the sum of evidence, it is possible to claim that mild to moderate physical activity in the general population can be matched with the expected decrease in the risk of AF incidence, whereas physical inactivity or heavy exercise, especially in young men, increases the risk of AF.
## Endurance exercise
As described above, according to population-based studies, vigorous physical activity, especially in young men, could be accompanied by an increased risk of AF. Never-theless, in vitro studies on animal models, clinical trials with heavy exercise induction, case-control studies on AF patients, and cohort research on professional athletes have reinforced the hypothesis of the risk factor nature of intense exercise for AF in recent years. The first evidence related to the risk factor nature of extreme exercise for AF was reported in the late 1990s [bib_ref] Lone atrial fibrillation in vigorously exercising middle aged men: casecontrol study, Karjalainen [/bib_ref] , while in recent years there have been other studies that uncovered a U-Shape and/or J-Shape relationship between exercise and risk of AF [bib_ref] Emerging risk factors and the dose-response relationship between physical activity and lone..., Calvo [/bib_ref] [bib_ref] Physical activity, resting heart rate, and atrial fibrillation: the Tromso Study, Morseth [/bib_ref]. Studies have reported that athletes are about 2-4 times more likely to have AF than the "normal" population [bib_ref] Long-lasting sport practice and lone atrial fibrillation, Mont [/bib_ref] [bib_ref] Sport practice and the risk of lone atrial fibrillation: a case-control study, Elosua [/bib_ref]. Other investigations in marathon runners, skiers, and cyclists, especially in long-distance cross-country skiing and cycling, have underlined that there is a higher incidence of AF than the general population; and every 10 years of participation in professional sports results in a 20% increase in the AF risk [bib_ref] Relation of vigorous exercise to risk of atrial fibrillation, Aizer [/bib_ref] [bib_ref] Increased risk of atrial fibrillation among elderly N orwegian men with a..., Myrstad [/bib_ref] [bib_ref] Increased risk of atrial fibrillation among elderly Norwegian men with a history..., Myrstad [/bib_ref]. It was also found that moderate to severe activity in excess of 2000 hours during lifetime was significantly associated with a higher incidence of AF (odds ratio=3.88; 95% confidence interval= 1.55-9.73) [bib_ref] Physical activity, height, and left atrial size are independent risk factors for..., Mont [/bib_ref]. Another study on 52.755 long-distance cross-country skiers reported that the incidence of AF was significantly higher among those who had reached the finish line faster and completed more matches [bib_ref] Risk of arrhythmias in 52 755 long-distance crosscountry skiers: a cohort study, Andersen [/bib_ref]. A meta-analysis exhibited that the AF risk rate in athletes was 5.3 times more than in non-athletes [bib_ref] Is the risk of atrial fibrillation higher in athletes than in the..., Abdulla [/bib_ref]. This odds ratio of around five is impressive when compared with the odds ratio of about 1.5 for the proven relationship between high blood pressure and increased risk of AF [bib_ref] Is exercise becoming a danger for our health? The complex relationship between..., Seccia [/bib_ref]. Other meta-analyses have also been published in recent years, which reported that the risk of AF in relation to the extent of exercise follows a J-shaped pattern [bib_ref] Physical activity volume in relation to risk of atrial fibrillation. A non-linear..., Ricci [/bib_ref]. However, among studies conducted in athletes, two studies with a sample size of less than 100 people have shown that endurance exercise is not necessarily associated with structural changes in the heart and thus the risk of AF [bib_ref] Troponin release following endurance exercise: is inflammation the cause? a cardiovascular magnetic..., O'hanlon [/bib_ref] [bib_ref] Right and Left Ventricular Function and Mass in Male Elite Master Athletes:..., Bohm [/bib_ref]. These two studies may not question the overall orientation of the other studies for the direct association between intense exercise and increased risk of AF, but they can provide new hypotheses about the mechanisms that cause structural changes in the heart, and development of AF associated with extreme exercise. For example, can any kind of intense exercise be associated with increased risk for AF development or is this association restricted only to some particular types of severe exercise? Can the duration of exercise and its continuity be recognized as a risk factor in the incidence of AF?
## Possible mechanisms
To investigate the possible mechanisms in relation to the effect of exercise on preventing or increasing the risk of AF, it is better to consider the mechanisms by assuming the J-Shape association of exercise with AF in two general categories: 1) the mechanisms of AF induction through exercise, and 2) protective mechanisms of exercise against AF.
## Mechanisms explaining an increased risk for af
To better understand the mechanisms that are associated with AF induction through exercise, we categorized them in two groups: 1) for athletes with physical endurance activity, and 2) mechanisms associated with the lack of physical activities (sedentary lifestyle).
Although there is still no precise mechanism for increasing AF among athletes with severe physical activity, some important pathways for the underlying pathophysiology have been proposed. Ectopic triggers (originated from the pulmonary veins), modulators as well as altered atrial substrate are three potential mediators suggested for the athlete's atrial arrhythmogenesis [bib_ref] Atrial remodeling, autonomic tone, and lifetime training hours in nonelite athletes, Wilhelm [/bib_ref] [bib_ref] P296Exercise-induced atrial fibrillation is associated with a pro-inflammatory status, Guasch [/bib_ref]. More precisely, elevated autonomic activity, bilateral atrial dilatation, atrial fibrosis (possibly due to increased aldosterone following exercise), abrupt shifts between vagal dominance and sympathetic drive, increased atrial premature beats, changes in the ion channels of the pacemaker cells, and cell junction loss and collagen accumulation (possibly due to increased vascular thickness, similar to what happened for blood pressure) are among the suggested mechanisms for increasing the risk of AF following endurance sports [bib_ref] The role of exercise in atrial fibrillation prevention and promotion: finding optimal..., Elliott [/bib_ref] [bib_ref] Atrial fibrillation progression and outcome in patients with young-onset atrial fibrillation, De With [/bib_ref] [bib_ref] Vagal thirddegree atrioventricular block in a highly trained endurance athlete, Vidal [/bib_ref] [bib_ref] Atrial fibrillation and gastroesophageal reflux disease: the cardiogastric interaction, Linz [/bib_ref]. On the other hand, the lack of physical activity also increases systemic inflammation and oxidative stress; inflammation can induce atrial remodeling and eventually the incidence of AF [bib_ref] Exercise and Atrial Fibrillation: Prevention or Causation?, Elliott [/bib_ref] [bib_ref] The role of exercise in atrial fibrillation prevention and promotion: finding optimal..., Elliott [/bib_ref]. Autonomic dysfunction and sympathetic tone enhancement are other results of sedentary lifestyle that can increase afterdepolarization-dependent triggered activity and thereby increase the risk of AF [bib_ref] The role of exercise in atrial fibrillation prevention and promotion: finding optimal..., Elliott [/bib_ref] [bib_ref] Physical activity is associated with a reduced risk of atrial fibrillation in..., Drca [/bib_ref].
## Mechanisms explaining a decreased risk for af
There is still insufficient information available on the protective mechanisms of exercise against AF [bib_ref] Is exercise becoming a danger for our health? The complex relationship between..., Seccia [/bib_ref]. However, increased physical activity and better cardiorespiratory fitness may contribute to reducing the risk of AF by helping to prevent other AF-related illnesses such as obesity (by decreasing visceral fat), hypertension, diabetes (by improving glycemic control), and obstructive sleep apnea [bib_ref] The impact of lifestyle intervention on atrial fibrillation, Hong [/bib_ref] [bib_ref] PREVEntion and regReSsive Effect of weight-loss and risk factor modification on Atrial..., Middeldorp [/bib_ref] [bib_ref] The evaluation of right atrial temporary pacing for preventing postoperative atrial fibrillation..., Mohammad [/bib_ref] [bib_ref] Association between sleep disordered breathing and electrocardiographic markers of atrial abnormalities: the..., Kwon [/bib_ref] [bib_ref] Pacemaker-detected severe sleep apnea predicts new-onset atrial fibrillation, Mazza [/bib_ref] [bib_ref] Cardiac pauses in competitive athletes: a systematic review examining the basis of..., Senturk [/bib_ref]. Further, improving systolic and diastolic function, reducing arterial stiffness, lower sympathetic drive, and making positive changes in the structure of the heart (such as decreasing the left atrial size) are among the other possible mechanisms in relation to the effect of physical activity on reducing the risk of AF.
## What to do next
## Suggestions for practice
Based on existing evidence, mild to moderate-intensity physical activity for 150 to 200 minutes per week (about half an hour per day), or aerobic exercise for 90-150 minutes per week (about 15-20 minutes per day), or achievement of cardiorespiratory fitness over 8 seems to be associated with a lower risk of AF [bib_ref] Risk factor management in atrial fibrillation, Brandes [/bib_ref]. However, it is strongly recommended to determine the desired duration and intensity of exercise in coordination with the medical practitioner [bib_ref] Empowerment of athletes with cardiac disorders: a new paradigm, Providencia [/bib_ref] , especially in individuals with other AF risk factors.
## Suggestions for further research
Perhaps the most significant limitation of current studies is the presence of heterogeneity in the methodologies used to measure the level of physical activity associated with decreasing or increasing the risk of AF. On the other hand, the criteria used to detect AF do not fit in many studies, and this makes it hard to draw a comprehensive conclusion. In relation to measurements of activity level, there have been hitherto several methods, including self-reported data (based on qualitative variables such as low, moderate, severe, and very severe), daily exercise time report in hour, cardiorespiratory fitness status and calories intake measurement. Self-reporting of daily activities of participants in studies can create a report-bias, resulting in errors in the conclusion, which is suggested to be replaced in subsequent studies by other methods. Therefore, it seems that conducting further investigations, with a higher focus on more precise methods for measuring activity and AF, can provide a better evidence basis for clinical decision-making. Also, while many studies have been conducted to uncover the mechanisms of AF induction by heavy exercise, the protective mechanisms of exercise against AF have not yet been well understood, and further studies are needed in this area.
# Conclusions
According to published results, it seems that the level of physical activity and risk of AF have a nonlinear relationship. Regular "mild to moderate exercise" may have to be considered as an appropriate strategy to reduce the risk of AF in the general population. Moreover, avoiding a sedentary lifestyle or very heavy exercise may be also ways to prevent the onset of AF, especially in young men. Lastly, besides the exact potential mechanisms, one of the questions that remains partially unanswered is the effect of different types of sports, their frequencies and durations on AF incidence. So, further studies are needed in order to gain a better understanding of the association between sports with different properties and risk of AF.
GMJ.2018;7:e1401 www.gmj.ir
GMJ.2018;7:e1401 www.gmj.ir
AcknowledgmentsNoneConflict of InterestsNone |
Improved power for TB Phase IIa trials using a model-based pharmacokinetic–pharmacodynamic approach compared with commonly used analysis methods
Background:The demand for new anti-TB drugs is high, but development programmes are long and costly. Consequently there is a need for new strategies capable of accelerating this process.Objectives: To explore the power to find statistically significant drug effects using a model-based pharmacokineticpharmacodynamic approach in comparison with the methods commonly used for analysing TB Phase IIa trials.Methods: Phase IIa studies of four hypothetical anti-TB drugs (labelled A, B, C and D), each with a different mechanism of action, were simulated using the multistate TB pharmacometric (MTP) model. cfu data were simulated over 14 days for patients taking once-daily monotherapy at four different doses per drug and a reference (10 mg/ kg rifampicin). The simulated data were analysed using t-test, ANOVA, mono-and bi-exponential models and a pharmacokinetic-pharmacodynamic model approach (MTP model) to establish their respective power to find a drug effect at the 5% significance level.Results: For the pharmacokinetic-pharmacodynamic model approach, t-test, ANOVA, mono-exponential model and bi-exponential model, the sample sizes needed to achieve 90% power were: 10, 30, 75, 20 and 30 (drug A); 30, 75, 245, 75 and 105 (drug B); 70, .1250, 315, .1250 and .1250 (drug C); and 30, 110, 710, 170 and 185 (drug D), respectively.Conclusions:A model-based design and analysis using a pharmacokinetic-pharmacodynamic approach can reduce the number of patients required to determine a drug effect at least 2-fold compared with current methodologies. This could significantly accelerate early-phase TB drug development.
# Introduction
The currently recommended treatment for drug-susceptible TB consists of rifampicin, isoniazid, pyrazinamide and ethambutol.Under trial conditions the regimen provides a cure rate of approximately 95%. [bib_ref] Four-month moxifloxacin-based regimens for drug-sensitive tuberculosis, Gillespie [/bib_ref] [bib_ref] A four-month gatifloxacin-containing regimen for treating tuberculosis, Merle [/bib_ref] [bib_ref] High-dose rifapentine with moxifloxacin for pulmonary tuberculosis, Jindani [/bib_ref] For MDR TB, however, cure rates are much lowerand few drugs have been developed for use in these patients. Although bedaquiline 6 and delamanid [bib_ref] Delamanid for multidrugresistant pulmonary tuberculosis, Gler [/bib_ref] have recently reached the market through conditional approval there remains a paucity of candidates in early clinical development. Thus, there is a pressing need for the development of new agents to address this gap in our therapeutic armamentarium.
The primary endpoint for Phase III TB trials is relapse-free cure after treatment. [bib_ref] Four-month moxifloxacin-based regimens for drug-sensitive tuberculosis, Gillespie [/bib_ref] The choice of Phase III combinations is guided by the results of Phase IIb trials where drug combinations are studied for the first 8 weeks of treatment. [bib_ref] Assessment of the efficacy of new antituberculosis drugs, Mitchison [/bib_ref] These late-stage clinical trials are preceded by a Phase IIa methodology adapted to TB when a novel drug is given to TB patients alone or in combination. This is designed to provide information on the bactericidal effects [bib_ref] 14-Day bactericidal activity of PA-824, bedaquiline, pyrazinamide, and moxifloxacin combinations: a randomised..., Diacon [/bib_ref] [bib_ref] Early phase evaluation of SQ109 alone and in combination with rifampicin in..., Heinrich [/bib_ref] [bib_ref] Bactericidal activity of pyrazinamide and clofazimine alone and in combinations with pretomanid..., Diacon [/bib_ref] and may aid in dose selection for Phase IIb. [bib_ref] The early bactericidal activity of antituberculosis drugs, Diacon [/bib_ref] Phase IIa TB studies are usually monotherapy trials analysed by comparing the changes in mycobacterial load in sputum for 7-14 days of treatment compared within or between dose arms. A traditional approach used to analyse such trials is by calculating the early bactericidal activity (EBA) defined as the fall in log 10 cfu/day. [bib_ref] The early bactericidal activity of antituberculosis drugs: a literature review, Donald [/bib_ref] The EBA is non-model-based and is obtained using only two timepoints, such as the first and last cfu measurements in a patient. The mean EBA is compared between drugs and doses by ANOVA or t-test. [bib_ref] The early bactericidal activity of isoniazid related to its dose size in..., Donald [/bib_ref] Conventionally, an empirical model approach is also used which gives model-based estimates of the change in cfu using simple empirical models including mono-and bi-exponential regression models. [bib_ref] The early bactericidal activity of drugs in patients with pulmonary tuberculosis, Jindani [/bib_ref] [bib_ref] Bactericidal and sterilizing activities of antituberculosis drugs during the first 14 days, Jindani [/bib_ref] These empirical models are simultaneously fitted to all available data points from a patient. The model-based estimates of change in cfu are compared between drugs and doses. Phase IIa TB studies typically include 10-15 patients per dose arm and many fail to demonstrate statistically significant differences between drugs and between different doses of the same drug. [bib_ref] The early bactericidal activity of antituberculosis drugs: a literature review, Donald [/bib_ref] [bib_ref] A multicentre comparison of a novel surrogate marker for determining the specific..., Gosling [/bib_ref] Model-based analysis using a pharmacokineticpharmacodynamic approach has been shown to increase the power of Phase II studies in indications such as HIV, acute stroke and diabetes when compared with traditional statistical analyses. [bib_ref] Model based design and analysis of phase II HIV-1 trials, Reki C [/bib_ref] [bib_ref] Comparisons of analysis methods for proof-of-concept trials, Karlsson [/bib_ref] Thus, in this study, we adapt this approach to TB to evaluate the power of a pharmacokinetic-pharmacodynamic approach using the multistate TB pharmacometric (MTP) model [bib_ref] A multistate tuberculosis pharmacometric model: a framework for studying anti-tubercular drug effects..., Clewe [/bib_ref] [bib_ref] Application of the multistate tuberculosis disease model for studying pharmacokinetics and pharmacodynamics..., Chen [/bib_ref] [bib_ref] Application of the multistate tuberculosis pharmacometric model in patients with rifampicin-treated pulmonary..., Svensson [/bib_ref] in comparison with traditional statistical analyses including empirical model approaches (mono-and bi-exponential regression) and traditional approaches (t-test and ANOVA).
# Methods
To explore the impact of different analysis techniques we postulated four drugs (labelled A, B, C and D) with different mechanisms of action, but similar pharmacokinetics. Firstly, individual cfu versus time data were simulated using a previously described pharmacokinetic-pharmacodynamic model for pre-clinical and clinical TB, the MTP model. [bib_ref] A multistate tuberculosis pharmacometric model: a framework for studying anti-tubercular drug effects..., Clewe [/bib_ref] [bib_ref] Application of the multistate tuberculosis disease model for studying pharmacokinetics and pharmacodynamics..., Chen [/bib_ref] [bib_ref] Application of the multistate tuberculosis pharmacometric model in patients with rifampicin-treated pulmonary..., Svensson [/bib_ref] As a second step, the simulated data were analysed using the different approaches in order to determine the statistical power for each respective approach for drugs A-D.
## Study design
The Phase IIa study design included four study arms receiving 100, 200, 300 or 400 mg of the putative drugs and 10 mg/kg rifampicin (reference arm) administered orally once daily as monotherapy for 14 consecutive days. Subjects were divided equally across arms. We simulated sputum collections at baseline and days 1-7, 9 and 14 after start of treatment 23 between 8 p.m. and 8 a.m. with assumed sputum volumes of 10 mL. As patients were assumed to have established infections, all treatments were started 150 days after time of infection. [bib_ref] Application of the multistate tuberculosis pharmacometric model in patients with rifampicin-treated pulmonary..., Svensson [/bib_ref] Patients were assumed to weigh 56 kg and to be smear-positive, newly diagnosed adults with uncomplicated previously untreated pulmonary TB taking no other medications.
## Phase iia cfu simulations
The MTP model 22 was linked to a pharmacokinetic model in addition to exposure-response parameters and was used to simulate individual cfu data over time for four different drugs (A, B, C and D). The MTP model parameters described cfu without drug, the pharmacokinetic model parameters described exposure and the exposure-response parameters described the drug effect [fig_ref] Table 1: Pharmacokinetic and MTP model parameters used in the simulations of cfu versus... [/fig_ref]. [bib_ref] Application of the multistate tuberculosis pharmacometric model in patients with rifampicin-treated pulmonary..., Svensson [/bib_ref] The MTP model described three mycobacterial states including fast-, slow-and non-multiplying states with growth present on the fast-multiplying state. The numbers in the different states at any timepoint were defined by the following differential equations:
[formula] dF dt ¼ k G Á log B max F þ S þ N Á F þ k SF Á S À k FS Á F À k FN Á F dS dt ¼ k FS Á F þ k NS Á N À k SN Á S À k SF Á S dN dt ¼ k SN Á S þ k FN Á F À k NS Á N where k FS ¼ k FS lin Á t [/formula]
where F, S and N are the model-predicted bacterial numbers in fast-, slowand non-multiplying states, respectively. Rate constants (k) with two-letter subscripts describe transfer rates between fast (F), slow (S) and nonmultiplying (N) states; the first letter indicates the transfer origin and the second letter the destination. The parameter k FSlin describes a timedependent transfer from fast-to slow-multiplying state. Time (t) is in days after infection, k G is the fast-multiplying bacterial growth rate and B max is the system carrying capacity. We assumed that only the fast-and slow-multiplying bacteria are detectable as cfu and those in the non-multiplying state are not. The sputum sampling compartment method was included representing the clinical sampling procedure performed over a given time interval. For the sputum sampling compartment method, bacteria accumulate in a sputum sample compartment (Sample) during the sampling interval described by:
[formula] dSample dt ¼ k prod Á F þ S ð Þ where k prod ¼ V sample D sample [/formula]
where k prod is the sputum production rate. [bib_ref] Application of the multistate tuberculosis pharmacometric model in patients with rifampicin-treated pulmonary..., Svensson [/bib_ref] The parameter V sample is the sputum sample volume (mL) and D sample is the sampling duration (h). The cfu in the sample compartment was calculated by:
[formula] cfu ¼ Sample V sample [/formula]
at the end of each sampling period to get cfu (mL #1 ). The drug pharmacokinetic parameters for drugs A-D were set to the same hypothetical values, reflecting a drug with one-compartment disposition with first-order absorption and rapid elimination [fig_ref] Table 1: Pharmacokinetic and MTP model parameters used in the simulations of cfu versus... [/fig_ref] and . The pharmacokinetics of rifampicin were generated using a previously developed population pharmacokinetic model. [bib_ref] A semimechanistic pharmacokinetic-enzyme turnover model for rifampin autoinduction in adult tuberculosis patients, Smythe [/bib_ref] Rifampicin was assumed to inhibit growth of the fast-multiplying bacteria in addition to killing the slow-and non-multiplying bacteria. [bib_ref] Application of the multistate tuberculosis pharmacometric model in patients with rifampicin-treated pulmonary..., Svensson [/bib_ref] Drugs A and B had the same mechanism of action as rifampicin, i.e. inhibition of fast-multiplying bacterial growth and killing of the slow-and nonmultiplying bacteria . Drug C was defined as being able to kill non-multiplying bacteria and drug D was assumed to kill slow-multiplying bacteria. The different mechanisms of action for drugs A-D are shown in .
The different mechanisms of action were represented by three different exposure-response models:
[formula] EFG ¼ 1 À FG on=off ESD ¼ SD k Á C p Svensson et al. END ¼ ND k Á C p [/formula]
where EFG, ESD and END are drug effects, namely inhibition of fastmultiplying growth and killing of slow-and non-multiplying bacteria, respectively. The parameter FG on/off is the fractional inhibition of growth of the fast-multiplying state, where the drug effect is present once the drug concentration (C p ) is above 0. The parameters SD k and ND k are the death rates of the slow-and non-multiplying states, respectively. Different values of the exposure-response parameters (FG on/off , SD k and ND k ) were assumed for drugs A-D [fig_ref] Table 1: Pharmacokinetic and MTP model parameters used in the simulations of cfu versus... [/fig_ref]. Drug A was 20% more potent than rifampicin, 22 except for inhibition of fast-multiplying growth which was similar to rifampicin, i.e. 100%. Drug B was 50% less potent than drug A. Drug C had the same potency as drug A for killing of the non-multiplying state. Drug D had the same potency as drug A for killing of slow-multiplying bacteria. The drug effects were included in the differential equations for the MTP model as follows:
[formula] dF dt ¼ k G Á log B max F þ S þ N Á EFG Á F þ k SF Á S À k FS Á F À k FN Á F dS dt ¼ k FS Á F þ k NS Á N À k SN Á S À k SF Á S À ESD Á S dN dt ¼ k SN Á S þ k FN Á F À k NS Á N À END Á N [/formula]
# Power analysis
A large cfu dataset was simulated and subsets of the dataset were sampled repeatedly at several sample sizes. Each subset was analysed using five different methods: pharmacokinetic-pharmacodynamic model Power for TB Phase IIa trials JAC approach (MTP model), traditional approaches (t-test and ANOVA) and empirical model approaches (mono-and bi-exponential regression). The proportion of the subsets at each sample size where a drug effect was detected was defined as the power. Statistical significance was accepted at the 5% significance level. An additional criterion for clinical significance was included for each approach (outlined below). The approaches are described briefly below. For a detailed description see Appendix S1 (available as Supplementary data at JAC Online).
## Pharmacokinetic-pharmacodynamic model approach using the mtp model
For the pharmacokinetic-pharmacodynamic model approach using the MTP model, the Monte Carlo mapped power (MCMP) method was applied. [bib_ref] Rapid sample size calculations for a defined likelihood ratio test-based power in..., Vong [/bib_ref] For each drug a reduced model without drug effect corresponded to H 0 and a full model including drug effect corresponded to H 1 . The clinical significance criterion was included as a constraint on the exposure-response parameters SD k and ND k ; only positive values were allowed for these parameters. Negative values would mean that the drug directly increases the bacterial number.
## Empirical model approaches
For the empirical approaches the power to find a drug effect was obtained using mono-and bi-exponential regression models applied to the highest (400 mg) dose group. Both models assumed that cfu changed over time and in order to test if these changes were statistically significant the mono-and bi-exponential regression models were compared with a reduced empirical model assuming no change in cfu over time. Testing was done using an F-test under the null hypothesis that the mono-or bi-exponential model did not provide a better fit than the reduced empirical model (H 0 ). The clinical significance criterion was implemented to only allow declines in cfu.
## Traditional statistical approaches
For the traditional approaches using t-test or ANOVA, the EBA between days 0 and 14 (EBA 0-14 ) was used as the test statistic. The power to detect a drug effect was determined using a two-sided t-test under the null hypothesis that the mean EBA 0-14 from patients in the highest dose group (400 mg) did not differ from 0 (H 0 ). A clinical significance criterion was introduced where only positive values for EBA 0-14 contributed to the power since negative EBA 0-14 s represent drugs that increase cfu over time, which was considered a clinically irrelevant drug effect. The power was also calculated for finding differences in EBA 0-14 between study arms using ANOVA under the null hypothesis that there was no difference in EBA 0-14 between study arms (H 0 ). . Schematic representation of the MTP model linked to a pharmacokinetic model. The four drugs were assumed to have identical pharmacokinetics, but different mechanisms of action, indicated by the broken lines connecting the drug plasma concentration (C p ) to either killing of slowmultiplying state (S) or non-multiplying state (N) bacteria or inhibition of the growth rate (k G ) of bacteria in the fast-multiplying state (F). The broken lines indicate drug effect on each of the three possible effect sites. The letter of the drug (A-D) is shown on the broken line for the mechanism of action included for each hypothetical drug. Abs, absorption compartment; k a , absorption rate constant; CL/F, apparent oral clearance; V/F, apparent volume of distribution; B max , system carrying capacity; k FS , time-dependent linear rate parameter describing transfer from fast-to slow-multiplying state; k SF , transfer rate between slow-and fast-multiplying state; k FN , transfer rate between fast-and non-multiplying state; k SN , transfer rate between slow-and non-multiplying state; k NS , transfer rate between non-and slow-multiplying state; k prod , sputum production rate constant; Sample, sputum sample compartment; FG on/off , on/off-effect as inhibition of fast-multiplying bacterial growth; SD k , second-order slow-multiplying death rate; ND k , non-multiplying death rate.
## Software
Svensson et al.
# Results
Simulated typical cfu versus time after treatment with different doses of drugs A-D in monotherapy are shown in [fig_ref] Figure 2: Typical simulated log 10 cfu change from baseline versus time after start... [/fig_ref]. The required sample sizes to reach 90% power for finding a statistically significant drug effect (P 0.05) for treatment with drugs A-D using different analysis approaches, including a pharmacokineticpharmacodynamic model approach (MTP model), empirical approaches (mono-and bi-exponential regression) and traditional approaches (t-test and ANOVA), are shown in [fig_ref] Table 2: Total sample size required for finding a drug effect at 90% power... [/fig_ref] with corresponding power curves in . The required sample size to achieve 90% power for finding a statistically significant drug effect for Phase IIa TB trials was at least two times lower for the pharmacokinetic-pharmacodynamic model approach (MTP model) in comparison with other tested approaches for all drugs. The ANOVA approach required higher sample sizes for all drugs compared with all other analyses except drug C, where all other analyses except the ANOVA and pharmacokinetic-pharmacodynamic approaches failed to reach 90% power. Mono-exponential regression required lower sample sizes compared with bi-exponential regression for all drugs except drug C where all approaches except the pharmacokinetic-pharmacodynamic model approach and the ANOVA approach failed to reach 90% power at the studied sample size range. The t-test required lower sample sizes than the empirical approaches for drug D whilst it performed similarly to bi-exponential regression for drug A and similarly to mono-exponential regression for drug B.
# Discussion
This work shows the advantage of adopting a pharmacokineticpharmacodynamic model approach when designing and analysing Phase IIa TB trials. For all the hypothetical drugs studied here the sample size required to achieve a significant result at 90% power was smaller using the MTP model than all of the other approaches investigated. The reasons for this sample size reduction are that a pharmacokinetic-pharmacodynamic analysis includes all longitudinal data simultaneously, including data from all dose groups. [bib_ref] Model based design and analysis of phase II HIV-1 trials, Reki C [/bib_ref] [bib_ref] Comparisons of analysis methods for proof-of-concept trials, Karlsson [/bib_ref] The empirical approaches (mono-and bi-exponential regression) are less informative as they include longitudinal data only from the highest dose group. The traditional approaches (t-test and ANOVA) only include the difference between cfu at day 0 and day 14.
Phase IIa trials within TB typically include 10-15 patients per arm; [bib_ref] The early bactericidal activity of antituberculosis drugs: a literature review, Donald [/bib_ref] hence the five arms used in the current study design correspond to a total study size of 50-75 patients. For drugs A-D, most analysis methods except the pharmacokinetic-pharmacodynamic model approach failed to attain 90% power at total sample sizes of 50-75 patients, in line with the known difficulty of finding statistically significant drug effects for Phase IIa TB trials [bib_ref] A reiterative method for calculating the early bactericidal activity of antituberculosis drugs, Gillespie [/bib_ref] and the even more difficult to detect differences between dose arms (i.e. exposure-response) which is assessed using ANOVA. Defining exposure-response is crucial for Phase IIa in general, but is not feasible for TB when using traditional methods and including 10-15 patients per arm. The pharmacokineticpharmacodynamic model approach also assesses differences between arms by estimating exposure-response parameters. By using the MTP model, the sample size needed to reach 90% power was reduced several-fold [fig_ref] Table 2: Total sample size required for finding a drug effect at 90% power... [/fig_ref] compared with the other approaches studied, enabling detection of exposure-response when including as few as 10-15 patients per arm. This allows for more robust dose selection for future trials based on exposure-response defined in Phase IIa.
The empirical approaches (mono-and bi-exponential regression) are expected to have higher power than traditional approaches since the empirical approaches include more data. However, when comparing the empirical approaches with the t-test approach, our results do not favour one over the other. Regression-based methods are expected to have higher power if the tested model fits the data well. In contrast, low power is likely to be seen for a model which is unable to describe the data. The cause of the unexpectedly low powers for the empirical analyses seen in this work is unknown, but our analysis shows that the relative power of empirical versus traditional approaches appears drug dependent. Svensson et al.
The ANOVA approach had lower power than the t-test approach except for drug C (increased killing of non-multiplying bacteria) where ANOVA had higher power than the t-test . The reason is that cfu increased in the highest dose group for day 14 compared with day 0 [fig_ref] Figure 2: Typical simulated log 10 cfu change from baseline versus time after start... [/fig_ref] which was not considered clinically significant for the t-test analysis and therefore did not contribute to the power. The ANOVA approach also had higher power than the mono-and bi-exponential analyses due to the criterion for clinical significance only allowing a decrease in cfu. The criterion for clinical significance was implemented as constraints on the rate constant parameters for bacterial elimination to only allow decrease in cfu which stopped the mono-or bi-exponential models from describing the increases in cfu well, resulting in low power. The ANOVA approach only looks at statistical significance Power at 5% significance level (%) . Predicted power at 5% significance level versus total sample size for four hypothetical anti-TB drugs (drugs A-D) using a pharmacokineticpharmacodynamic model approach (MTP model), mono-exponential regression, t-test, ANOVA and bi-exponential regression.
Power for TB Phase IIa trials JAC and not clinical significance, resulting in higher power for drug C. Exposure-response parameters were also constrained to be positive (i.e. only allowed to increase bacterial killing) for the pharmacokinetic-pharmacodynamic model approach, but the pharmacokinetic-pharmacodynamic model approach included disease-specific parameters governing bacterial growth (k G and B max ) which could allow an increase in cfu as an indirect consequence of drug exposure. The B max parameter acts to constrain growth at high bacterial densities, resulting in stationary phase growth. Drugs that kill bacteria reduce the bacterial density, resulting in growth proportional to the reduction in bacterial density, which caused regrowth for drug C [fig_ref] Figure 2: Typical simulated log 10 cfu change from baseline versus time after start... [/fig_ref]. It was only possible using a pharmacokinetic-pharmacodynamic approach to support statistically and clinically significant exposure-response relationships despite regrowth since this approach includes diseasespecific parameters governing growth. If monotherapy data from drugs that increase cfu over time (due to regrowth as seen for drug C) are analysed using traditional or empirical approaches, this will lead to the conclusion that the drug is clinically insignificant, although this might not always be the case according to our results. It should be noted that the risk of this happening is low for drugs that either cause substantial inhibition of growth such as rifampicin [bib_ref] Comparisons of analysis methods for proof-of-concept trials, Karlsson [/bib_ref] or for drugs that strongly kill growing bacteria, effectively preventing regrowth. This finding is interesting in the context of the Phase IIa trial of SQ109 where daily monotherapy doses of 75 mg resulted in a greater decline in cfu over 14 days than doses of 150 and 300 mg. [bib_ref] Early phase evaluation of SQ109 alone and in combination with rifampicin in..., Heinrich [/bib_ref] If SQ109 kills only the non-multiplying bacteria it would not be expected that the highest dose necessarily results in the greatest reduction in cfu over 14 days due to the regrowth phenomenon seen for drug C . Although it would mean that if SQ109 acts to kill non-multiplying bacteria it would most likely enhance decline of cfu in combination with other drug(s) able to inhibit growth or otherwise prevent regrowth. Interestingly, SQ109 was combined with rifampicin in Phase IIa 10 and an enhanced decline in cfu was seen when rifampicin was combined with 150 mg of SQ109 compared with rifampicin monotherapy. We speculate that if the Phase IIa trial of SQ109 had been designed and analysed using a pharmacokinetic-pharmacodynamic approach, SQ109 might have been interpreted as a drug that only kills nonmultiplying bacteria. If this speculation is true SQ109 has potential for treating TB efficiently. But further refinement of the dose of SQ109 and choice of companion drugs are required as the reported efficacy after 12 weeks of treatment with 300 mg SQ109 combined with 10 or 20 mg/kg rifampicin, 5 mg/kg isoniazid and 25 mg/kg pyrazinamide was not better than the standard regimen consisting of 10 mg/kg rifampicin, 5 mg/kg isoniazid, 25 mg/kg pyrazinamide and 15-20 mg/kg ethambutol. [bib_ref] High-dose rifampicin, moxifloxacin, and SQ109 for treating tuberculosis: a multi-arm, multi-stage randomised..., Boeree [/bib_ref] In conclusion, a pharmacokinetic-pharmacodynamic model approach using the MTP model [bib_ref] A multistate tuberculosis pharmacometric model: a framework for studying anti-tubercular drug effects..., Clewe [/bib_ref] [bib_ref] Application of the multistate tuberculosis disease model for studying pharmacokinetics and pharmacodynamics..., Chen [/bib_ref] [bib_ref] Application of the multistate tuberculosis pharmacometric model in patients with rifampicin-treated pulmonary..., Svensson [/bib_ref] was able to reduce the sample sizes required to reach 90% power to find statistically significant drug effects for Phase IIa TB trials compared with traditional and empirical approaches. This has the potential to make early TB drug development less expensive and more robust as fewer patients are required to show a statistically significant drug effect.
[fig] Figure 2: Typical simulated log 10 cfu change from baseline versus time after start of treatment of four hypothetical anti-TB drugs (drugs A-D) following 100, 200, 300 and 400 mg (black continuous lines) and 10 mg/kg rifampicin (grey broken lines) given orally once daily (OD). [/fig]
[table] Table 1: Pharmacokinetic and MTP model parameters used in the simulations of cfu versus time after different doses of drugs A, B, C and D for different study sample sizes (1000 replicates per sample size) ND k (L mg #1 days #1 ) second-order non-multiplying state death rate CV, coefficient of variation.The pharmacokinetic parameters were the same for all four different drugs. [/table]
[table] Table 2: Total sample size required for finding a drug effect at 90% power and 5% significance level for drugs A, B, C and D using a pharmacokineticpharmacodynamic model approach with the MTP model, traditional statistical approaches and empirical model approaches Analysis Total sample size (ratio compared with pharmacokinetic-pharmacodynamic model approach) [/table]
|
Identifying and prioritizing strategies for comprehensive liver cancer control in Asia
Background: Liver cancer is both common and burdensome in Asia. Effective liver cancer control, however, is hindered by a complex etiology and a lack of coordination across clinical disciplines. We sought to identify strategies for inclusion in a comprehensive liver cancer control for Asia and to compare qualitative and quantitative methods for prioritization.Methods: Qualitative interviews (N = 20) with international liver cancer experts were used to identify strategies using Interpretative Phenomenological Analysis and to formulate an initial prioritization through frequency analysis. Conjoint analysis, a quantitative stated-preference method, was then applied among Asian liver cancer experts (N = 20) who completed 12 choice tasks that divided these strategies into two mutually exclusive and exhaustive subsets. Respondents' preferred plan was the primary outcome in a choice model, estimated using ordinary least squares (OLS) and logistic regression. Priorities were then compared using Spearman's Rho.Results: Eleven strategies were identified: Access to treatments; Centers of excellence; Clinical education; Measuring social burden; Monitoring of at-risk populations; Multidisciplinary management; National guidelines; Public awareness; Research infrastructure; Risk-assessment and referral; and Transplantation infrastructure. Qualitative frequency analysis indicated that Risk-assessment and referral (85%), National guidelines (80%) and Monitoring of at-risk populations (80%) received the highest priority, while conjoint analysis pointed to Monitoring of at-risk populations (p < 0.001), Centers of excellence (p = 0.002), and Access to treatments (p = 0.004) as priorities, while Risk-assessment and referral was the lowest priority (p = 0.645). We find moderate concordance between the qualitative and quantitative methods (rho = 0.20), albeit insignificant (p = 0.554), and a strong concordance between the OLS and logistic regressions (rho = 0.979; p < 0.0001). Conclusions: Identified strategies can be conceptualized as the ABCs of comprehensive liver cancer control as they focus on Antecedents, Better care and Connections within a national strategy. Some concordance was found between the qualitative and quantitative methods (e.g. Monitoring of at-risk populations), but substantial differences were also identified (e.g. qualitative methods gave highest priority to risk-assessment and referral, but it was the lowest for the quantitative methods), which may be attributed to differences between the methods and study populations, and potential framing effects in choice tasks. Continued research will provide more generalizable estimates of priorities and account for variation across stakeholders and countries.
# Background
Hepatocellular carcinoma (HCC), the predominant form of liver cancer, is the sixth most common cancer and the third most frequent cause of cancer-related death worldwide [bib_ref] The incidence and epidemiology of hepatocellular carcinoma: a global and regional perspective, Venook [/bib_ref]. At least two thirds of the people who die each year from HCC live in the Asia-Pacific region. The majority of patients with HCCs are diagnosed in the advanced stages of presentation due to the relative paucity of symptoms in the early stages [bib_ref] Hepatocellular carcinoma: an Asian perspective, Teo [/bib_ref]. Because of the multifocal and advanced stage of disease at time of diagnosis, potentially curative treatment for HCC is not feasible in 80% of patients.
Chronic liver disease is closely associated with HCC. In areas where hepatitis B virus (HBV) is endemic, the incidence of HCC is high. It has been estimated that about 75% of the world's chronic HBV carriers are in Asia [bib_ref] Liver transplantation for hepatocellular carcinoma in Asia, De Villa [/bib_ref]. However, the etiology of HCC in Japan is different as hepatitis C virus (HCV) is more prevalent than HBV. Ninety percent of the HCC in Japan is HCV related. As stated in a recent report by the United States Institute of Medicine, both HBV and HCV can be prevented and controlled, which would reduce the incidence of HCC and liver disease [bib_ref] Palmer Beasley R: Institute of Medicine recommendations for the prevention and control..., Mitchell [/bib_ref].
The relative burden and complexity of liver cancer, especially in Asia, lends itself to a comprehensive cancer control plan. However, there is a paucity of data or experience to design such a policy response. While comprehensive cancer control plans regularly target lung, colorectal, breast and cervical cancer, such approaches have not been applied to liver cancer [bib_ref] Public policy action and CCC implementation: benefits and hurdles, Steger [/bib_ref]. The WHO guidance for the development of national cancer programs offers some guidance for implementation. The WHO conceptualizes its model around disease progression and is focused around six dimensions: prevention, early detection, diagnosis/treatment, pain relief/palliative care, cancer control research, and surveillance. One of the limitations of this approach is that it distinguishes between appropriate strategies that should be used in countries with low, middle and high levels of resourcesa barrier to a common policy framework that would be appropriate for a pan-Asian response [bib_ref] Identifying Important Breast Cancer Control Strategies in Asia, Latin America and the..., Bridges [/bib_ref].
This paper reports the findings of a study aimed at identifying strategies appropriate for inclusion in a comprehensive liver cancer control plan and at assessing the relative priorities among these strategies. We also sought to compare the implied priorities in the qualitative data (i.e. via semi-quantification using frequency analysis) to those found using a quantitative stated-preference methodology (conjoint analysis)-with a particular focus on Asia.
Our research is of interest to those focused on liver cancer control, especially in Asia, for three reasons. First, beyond clinical guidelines, there is very little in the way of comparative research on liver cancer policy internationally. This paucity of data extends even to basic epidemiological data on HCC, which are fragmented and come from diverse populations, using different methodologies and from studies performed at different times [bib_ref] The incidence and epidemiology of hepatocellular carcinoma: a global and regional perspective, Venook [/bib_ref]. Second, while Japan and Taiwan have demonstrated successful strategies to combat HCC, especially through HBV vaccination and control, there is an absence of models of best practice for comprehensive liver cancer control beyond HBV vaccination in most countries of Asia [bib_ref] Cancer prevention by vaccination against hepatitis B. Recent Results, Chang [/bib_ref] [bib_ref] Optimal treatment increased survival of hepatocellular carcinoma patients detected with community-based screening, Tseng [/bib_ref]. Third, there are few templates available for the development of comprehensive cancer control plans for liver cancer, and it is uncertain if general cancer-control frameworks, such as the one proposed by the WHO, are relevant for liver cancer control.
We also make an important methodological contribution that is relevant to a wider audience of policy makers and health services researchers. Specifically, we demonstrate that while qualitative methods are valuable in identifying strategies [bib_ref] Identifying Patient-Relevant Endpoints Among Individuals with Schizophrenia: An Application of Patient Centered..., Kinter [/bib_ref] , semi-quantification methods such as frequency analysis [bib_ref] Content analysis in empirical social research, Bos [/bib_ref] may be less desirable for prioritization [bib_ref] The Challenge of Qualitative Content Analysis, Kracauer [/bib_ref]. We demonstrate this by comparing our frequency data with the results of a conjoint analysis-a qualitative stated-preference method [bib_ref] Stated preference methods in health care evaluation: an emerging methodological paradigm in..., Bridges [/bib_ref] that is increasingly used to identify priorities for health care policy [bib_ref] Patients' Preferences for Health Care System Reform in Hungary: A Conjoint Analysis..., Akkazieva [/bib_ref] [bib_ref] Using discrete choice modelling in priority setting: an application to clinical service..., Farrar [/bib_ref] and more broadly in health services research [bib_ref] Things are looking up since we started listening to patients: Recent trends..., Bridges [/bib_ref] [bib_ref] Conjoint Analysis Applications in Health-How are studies being designed and reported? An..., Marshall [/bib_ref].
# Methods
The study utilized both qualitative and quantitative research methods. First, in-depth, open-ended interviews were used to identify possible strategies and to explore possible priorities using frequency analysis, a common semi-quantification method [bib_ref] Content analysis in empirical social research, Bos [/bib_ref]. Qualitative methods are an important method for identifying complex issues in health care, including priority setting [bib_ref] Setting priorities in Canadian regional health authorities: a survey of key decision..., Mitton [/bib_ref] [bib_ref] Priority-setting decisions for new cancer drugs: a qualitative case study, Martin [/bib_ref] , and are an important way to include clinical stakeholders in decision making processes [bib_ref] The role of clinical opinion leaders in guideline implementation and quality improvement, Borbas [/bib_ref] [bib_ref] The state of the art versus the state of the science. The..., Greer [/bib_ref] [bib_ref] Accelerating the diffusion of innovations using opinion leaders, Valente [/bib_ref] , including the study of cancer care and coordination [bib_ref] Controlling liver cancer internationally: A qualitative study of clinicians' perceptions of current..., Bridges [/bib_ref] [bib_ref] What are the current barriers to effective cancer care coordination? A qualitative..., Walsh [/bib_ref] [bib_ref] Assessing palliative care needs: views of patients, informal carers and healthcare professionals, Mcilfatrick [/bib_ref]. Second, quantitative stated-preference methods were used to focus more on the priorities for Asia and to compare the implied priorities based on the qualitative data.
All participants were informed about the study and its potential risks and benefits. Participation in the study was voluntary and respondents were not reimbursed for participation. The study was deemed exempt from human subjects consideration from the Johns Hopkins University, Bloomberg School of Public Health Institutional Review Board (IRB). All respondents were guaranteed anonymity and confidentially.
## Identification (qualitative)
Respondents for the qualitative study were purposively sampled to constitute a geographically and professionally diverse sample of clinical experts in liver cancer and related disease [bib_ref] Checklists for improving rigour in qualitative research: a case of the tail..., Barbour [/bib_ref]. Potential respondents were actively involved in clinical practice, academic medical centers and/or policy relating to the prevention, detection and/ or management of liver cancer. Potential respondents were identified through published literature, medical societies and peer referral. Respondents were included if they were i) working in liver disease and liver cancer in their country; ii) involved in HCC clinical practice and policy; iii) active members in national and international liver associations and/or published extensively in peer review journals, and were excluded if they were not board certified or licensed to practice medicine in their countries and with at least three years of clinical experience or were unwilling or unable to complete the interview within the period required for completion of all interviews.
It is clear that our focus on clinical expertise is restrictive, but we wanted to ground our results in those who actually implement liver cancer control. This said, other stakeholders, including patients, family members, nursing staff or community leaders, may have given important insights. While this certainly is a limitation in our study, we aimed at assessing complex clinical issues, and as such needed experts who were experienced with discussing national liver cancer policies.
Information about the study and an invitation to participate was sent to respondents via mail or email in English, and in the respondents' native language where necessary. If no response was received within two weeks, follow-up included a second email and/or telephone call.
Open-ended qualitative interviews were conducted via face-to-face interviews or, in a limited number of instances, as telephone interviews. Multiple interviewers were used so as to accommodate multiple languages, with many of the interviews completed by the study leaders (JB, BB), an important triangulation method.
After respondents were informed about the study and consented to participate, they were asked about their country's "strategies to promote liver cancer prevention, treatment and research" and then "the main gaps in public policy." Finally, respondents were asked "if you had an opportunity to develop a comprehensive liver cancer control strategy, what elements would it cover?"
Interviews were recorded, transcribed (translated where necessary) and systematically analyzed in conjunction with any field notes. Respondents were allowed and encouraged to discuss other factors via open-ended questioning, but conversations were facilitated through the use of a comprehensive aide memoire based on previous research [bib_ref] Identifying Important Breast Cancer Control Strategies in Asia, Latin America and the..., Bridges [/bib_ref]. While saturation of themes was achieved after 16 interviews, we completed 20 interviews to facilitate semi-quantification.
Analysis was guided by Interpretive Phenomenological Analysis (IPA)in order to capture respondents' experiences, perceptions, practices, and processes associated with liver cancer control. Data were initially reviewed and coded by two researchers, including one who participated in data collection and one who did not. To ensure reliability, coding was compared and discussed with senior study members (JB and BB), and a final selection and appropriate labeling of identified themes was determined.
Triangulation methods included the use of multiple interviewers and analysts, geographical and professional heterogeneity of respondents, the comparison of transcripts with field notes, and comparison of results to the published literature via a targeted literature review.
Content experts (MK, KO, K-HH and S-LY) were consulted to ensure the validity of interpretation and to resolve any ambiguity in the data. After this, the two researchers reviewed the data to identify representative quotes and to ensure the reliability of the coding. Finally, to ensure that this manuscript reported all relevant information, we utilized the RATS guidelines [bib_ref] How to peer review a qualitative manuscript, Clark [/bib_ref].
## Prioritization (qualitative)
The use of numeration and/or semi-quantification in qualitative research remains controversial [bib_ref] Using Numbers in Qualitative Research, Maxwell [/bib_ref] [bib_ref] Real Qualitative Researchers Do Not Count: The Use of Numbers in Qualitative..., Sandelowski [/bib_ref]. This said, such methods are frequently used in health care research [bib_ref] Geovisual evaluation of public participation in decision making: The grapevine, Aguirre [/bib_ref] [bib_ref] Patient experiences with oily skin: The qualitative development of content for two..., Arbuckle [/bib_ref] [bib_ref] Understanding physicians' skills at providing end-of-life care: Perspectives of patients, families, and..., Curtis [/bib_ref] , and are called for by the RATS guidelines. Within the framework of IPA, numeration through an analysis of the frequency with which a theme is supported can be used as an indicator of its importance. As noted by Smith and colleagues:
"...it makes sense to think of the frequency with which emergent themes appear as one (though not the only) indication of the relative importance and relevance..."
To examine the potential relative importance of the identified strategies, we examined the frequency with which these strategies (and any sub-ordinate concepts) were discussed [bib_ref] Three Approaches to Qualitative Content Analysis, Hsieh [/bib_ref]. Rather than examine the frequency within a respondent, we report the percentage of the respondents making any reference to each of the indentified themes.
## Prioritization (quantitative)
As a means of offering a more quantitative assessment of importance, we developed and implemented conjoint analysis to examine the importance that respondents placed on the identified strategies. While it would have been beneficial to draw such data from the same respondents who participated in the qualitative research, it was decided to recruit new respondents from a single geographic region.
Conjoint-analysis methods, and more specifically discrete choice experiments, are grounded in both mathematical psychology and economic theory [bib_ref] A law of comparative judgement, Thurstone [/bib_ref] [bib_ref] Conditional Logit analysis of qualitative choice behavior, Mcfadden [/bib_ref]. They are based on the notion of the assessment of multiple stimuli (referred to as objects or attributes) that are combined to create vignettes or profiles that are presented to respondents in order to evoke an action, choice, or valuation [bib_ref] A Checklist for Conjoint Analysis Applications in Health: Report of the ISPOR..., Bridges [/bib_ref]. While such methods are widely used in health care [bib_ref] Things are looking up since we started listening to patients: Recent trends..., Bridges [/bib_ref] [bib_ref] Conjoint Analysis Applications in Health-How are studies being designed and reported? An..., Marshall [/bib_ref] , they have more recently been applied to examine issues associated with liver cancer control [bib_ref] Computer-based decision making in medicine: A model for surgery of colorectal liver..., Langenhoff [/bib_ref] [bib_ref] Patients' preferences for treatment of hepatitis C, Fraenkel [/bib_ref] [bib_ref] Understanding Surgical Decision Making in Early Hepatocellular Carcinoma, Nathan [/bib_ref].
Our approach to conjoint analysis is similar to that of Bridges et al., where conjoint analysis cards are developed to present a number of attributes (or objects) that do not vary across levels. Hence, for any given scenario in a conjoint analysis task, the attribute is either turned on or off [bib_ref] Source Design and Analysis of Simulated Consumer Choice or Allocation Experiments: An..., Louviere [/bib_ref]. Rather than identifying the best object in each profile, we present competing plans that represent mutually exclusive and exhaustive subsets of the 11 attributes identified in the qualitative section.
Our experimental design utilized a 2^11 main-effects orthogonal design from a catalogue of designs. This design consisted of a 12 × 11 matrix, with each row representing a specific experiment and each column representing the 11 strategies identified from the qualitative method. Each cell in the matrix was either a 0 or 1 and in developing the pair tasks we interpreted 0 as implying that the strategy should be assigned to the left plan and 1 as assigning the strategy to the plan on the right. The properties of the design were rigorously tested and the results cards did constitute a balanced, orthogonal, and minimal (i.e. zero) overlap design [bib_ref] The Importance of Utility Balance in Efficient Choice Designs, Huber [/bib_ref]. An example of the conjoint analysis task is provided in [fig_ref] Figure 1: An example of a conjoint analysis task [/fig_ref] , where a respondent is asked to identify which of two national liver cancer control plans would have the most impact in their own country.
Potential respondents for the conjoint analysis were identified in China, Japan and South Korea by country experts (MK, KO, KH and SY) and the inclusion/exclusion criteria from our qualitative analysis were used, as were the recruitment procedures. Again, we did not recruit stakeholders other than clinicians, so our results may be biased towards their viewpoint. As the aim of this analysis was to compare the results of the frequency and conjoint analyses, we thought that it was appropriate to use a similar sample size (n = 20). While this is small for a conjoint analysis, it is similar to mixed methods preference studies found in the literature [bib_ref] Identifying Patient-Relevant Endpoints Among Individuals with Schizophrenia: An Application of Patient Centered..., Kinter [/bib_ref] , and many commercial and legal applications of conjoint analysis methods have used similar sample sizes (especially when the focus is on the preferences of experts). Given this sample size, the results should not be interpreted as being widely generalizable, but comparable in scope to the qualitative research.
The quantitative survey instrument was administered to the respondents through a face-to-face interview or, where this was not possible, the survey was sent to the respondent and administered via a telephone interview. Respondents were guided through the questionnaire and answered the questions in the presence of the researcher. Respondents were asked to select the set of strategies they thought would be most important in a liver cancer control plan. No other answers or justifications were sought, and this process was repeated 12 times per participant. While some applications of conjoint analysis follow each task with an open or closed question regarding either strength of preference, ease of task or confidence in the answer [bib_ref] A Checklist for Conjoint Analysis Applications in Health: Report of the ISPOR..., Bridges [/bib_ref] , we did not include such questions so as to minimize the time burden on respondents. This said, notes were taken if respondents made any comments on the conjoint tasks.
The primary outcome in the analysis was the liver cancer control plan selected by the participant for each task, which was coded as a zero if the left-hand-side was chosen and one if the plan on the right was chosen. An identical method was used to code the placement of the strategies on the left and right of the choice tasks. For the purposes of comparison, we utilized both a linear probability model (via ordinary least squares) and logistic regression to estimate choice models using SAS (Version 9.13, Cary, NC, USA), but substantive conclusions are draw from the latter. For both estimation methods, robust standard errors are estimated to account for clustering of multiple choice tasks within each respondent [bib_ref] The Behavior of Maximum Likelihood Estimates Under Nonstandard Conditions, Huber [/bib_ref] [bib_ref] A Heteroskedasticity-consistent Covariance Matrix Estimator and a Direct Test for Heteroskedasticity, White [/bib_ref]. Hypothesis testing was based on the null that respondents' choices were not affected by each strategy (i.e. the importance weight is zero). The natural alternative hypothesis was that the importance weights were positive (given that all factors were identified as having priorities), however, we allowed for strategies to have a negative sign (as was found in some previous research, and utilized a two-tailed test.
To compare the implied priorities drawn from the qualitative and quantitative analyses, the estimated rank of the eleven strategies is presented graphically and in the results table. Further, the prioritization is compared between the qualitative and quantitative methods (and among the two quantitative estimation techniques) using the Spearman's Rho [bib_ref] The proof and measurement of association between two things, Spearman [/bib_ref].
# Strengths and weaknesses
This is the first paper to focus on the development of strategies for inclusion in a comprehensive liver cancer control program, and in doing so we demonstrate three important issues. First, there is a paucity of robust scientific research to inform the development of evidencebased cancer control plans. Second, preferences-based methods, both qualitative and quantitative, are valuable in identifying and prioritizing control strategies from the perspective of local stakeholders. Finally, such methods offer an important alternative to consensus methods that can be driven by "strong personalities", rather than generalizable data.
There are also several weaknesses in the research underpinning this paper. First, while it is clear that this is subjective research, it is somewhat unclear who the best subjects to recruit are. In some respects our respondents are too homogeneous (i.e. clinicians with a national or international profile), and we have omitted many important viewpoints (other clinical experts, policy makers/leaders and patients/advocates). On the other side, our respondents are heterogeneous, spanning many countries that may have different priorities, which may be biasing our results towards the null. Second, while this method is focused on the comparison of two methods (one qualitative, one quantitative), they have different samples-the former being more international to identify a range of possible strategies, the latter focusing on only three, albeit geographically close, countries. Finally, we have used a rather small sample size to illustrate conjoint analysis, and a much larger sample would be required to ensure generalizability of our results.
Given these weaknesses, there are several limitations in our research that must be addressed. First, while the study was primarily focused on the identification of possible strategies, this should not be considered as an exhaustive set. Second, while this study presents data on priorities, the primary purpose is to demonstrate the limitation of qualitative methods in identifying priorities and to illustrate the benefit of conjoint analysis, not to offer a definitive prioritization of strategies for Asia. Finally, although the data presented here are somewhat novel, more research is needed to see how priorities vary across countries and other stakeholders and to identify which priorities are common in Asia, and which are specific to individual countries in the region.
# Results
## Identification (qualitative)
Invitations to participate in the open-ended interviews were sent to 25 possible respondents, all of who met the eligibility criteria. One respondent refused to participate, and a further four consented, but a mutually agreeable time to schedule the interview could not be identified before the desired number of respondents was reached. Twenty interviews were conducted between February and June 2010 with experts based in eleven different countries (Australia, China, France, Germany, Italy, Japan, Spain, South Korea, Taiwan, Turkey and United States). The average duration of the interviews was 34 minutes (range 16-80 minutes).
Many respondents found the discussion of comprehensive liver cancer control a complicated task. As one respondent put it "My gosh, that is a 40 hour discussion, it would cover many things", and another cautioned at the end of a detailed discussion "Those are some points [but] I am not being complete." An example of the range of problems that need to be addressed by a comprehensive liver cancer control program was conveyed by one respondent:
"We will need to start with identifying the patients at-risk, we would then, after identifying those patients, need to come out with a surveillance strategy to monitor these patients regularly to minimize the chance that we overlooked the development of liver cancer in these patients, then we will need to have a general guideline on who should treat these patients meaning that they should be treated in specialized liver cancer centers that should be part of comprehensive cancer centers, and then we would need to have a study program using new drugs for the adjuvant treatment of those patients that have been treated and also palliative strategies to provide the best level of care for patients with incurable liver cancer."
Based on these interviews, 11 possible strategies of a comprehensive liver cancer control plan emerged as key themes, including Access to treatments; Centers of excellence; Clinical education; Measuring social burden;
Monitoring of at-risk populations; Multidisciplinary management; National guidelines; Public awareness; Research infrastructure; Risk-assessment and referral; and Transplantation infrastructure. Rather than focus on the presentation of key quotes, we analyzed the data and worked with content experts (MK, KO, KH and SY) to elaborate a description of each of the 11 strategies (see [fig_ref] Table 1: Strategies for comprehensive liver cancer control [/fig_ref]. This ensured that the findings constituted both a grounded and coherent interpretation of the data.
## Prioritization (qualitative)
Initial prioritization was based on the frequency with which the 11 strategies were discussed by the 20 respondents (but not accounting for multiple references within a single interview). The frequency and rank ordering of priorities are presented in table 2. The frequency of discussion across the key themes varied between 20-85%.
The most discussed items were Risk-assessment and referral (85%), National guidelines (80%) and Monitoring of at-risk populations (80%) implying that they are potential priorities. Research infrastructure (20%), Centers of excellence (25%), Measuring social burden and Transplantation infrastructure (both 30%) were strategies that were discussed with the lowest frequency, implying a lower priority.
## Prioritization (quantitative)
Invitations were sent to 42 potential respondents. Of these 23 (55%) consented to participate and 20 were eligible to participate. Field workers noted that after the first interviews in each country, which were all supervised by a senior investigator (JB or BB), respondents reported some difficulty with the choice tasks, mainly due to a lack of familiarity with conjoint analysis methods. Based on these concerns, all field workers discussed these difficulties, and strategies to overcome this problem were discussed. Here an example question, that was completed and explained, was added to ensure that all respondents were comfortable with the survey instrument and that all respondents were managed in a way that was consistent with these early interviews. This resolved the issue, with the remaining responders reporting no difficultly with the tasks. [fig_ref] Table 2: Importance of strategies for liver cancer controlRobust standard errors in parentheses [/fig_ref] presents the importance weights (i.e. parameter estimates) from choice models estimated from the conjoint analysis data using both a linear probability model (estimated via ordinary least squares) and logistic regression. Robust standards errors, p-values (based on a two tailed test) and rankings of priorities are also shown. Statistical significance (p < 0.05) was achieved on six strategies for both methods, with both methods in agreement on the significance on the top five factors. Here National statistics was significant based on OLS (p = 0.025), but not based on the logistic model (p = 0.056). Likewise, Clinical education was significant when considering logistic estimation (p = 0.019), but not when using OLS (p = 0.055). Both methods identified Measuring social burden and Risk-assessment and referral as having negative importance weights, but neither aversion reach statistical significance. Overall, there was a very-high level of agreement between the two methods (rho = 0.979; p < 0.0001), so substantive findings are drawn only from the logistic estimation.
As seen in table 2 the highest priority as estimated using the conjoint analysis was Monitoring of at-risk populations (p < 0.001), followed by Access to treatment (p = 0.004), Centers of excellence (p = 0.002), Multidisciplinary management (p = 0.004), Public awareness (p = 0.018), National guidelines (p = 0.056) and Clinical education (p = 0.019).
## Comparison of qualitative and quantitative priorities
When comparing the priorities from the conjoint analysis to the frequency analysis based on the qualitative data, there was some positive correlation (rho = 0.20), but this relationship was not significant (p = 0.554). When considering the priority given to individual attributes (see [fig_ref] Figure 2: A comparison of priorities using qualitative and quantitative methods [/fig_ref] , similar importance (as indicated by their rank) was given to Monitoring of at-risk populations (qual = 2/quant = 1), Public awareness (qual = 4/ quant = 5), Access to treatment (qual = 5/quant = 2), Clinical education (qual = 6/quant = 7), Multidisciplinary management (qual = 7/quant = 4), Transplantation infrastructure (qual = 8/quant = 8), Measuring social burden (qual = 8/quant = 10), and Research infrastructure (qual = 11/quant = 9). Differences in priority between the two methods were found for Risk-assessment and referral (qual = 1/quant = 11) and Centers of excellence (qual = 10/quant = 2), and to a lesser extent National guidelines (qual = 2/quant = 6).
# Discussion
When one considers the 11 strategies for comprehensive liver cancer control identified in this paper, we can see that they cover factors associated with facilitating, providing and integrating care into a single system. To facilitate the possible implementation of these strategies, one can conceptualize them into three categories: antecedents; better care; and connection. As seen in , this can lead to a model that relates to the ABCs of comprehensive liver cancer control. Here Antecedents include Clinical education, Measuring social burden and Public Awareness, all factors that can motivate the adoption of comprehensive liver cancer control. Access to treatments, Monitoring of at-risk populations, Riskassessment and referral and Transplantation infrastructure are all factors aimed at providing Better care, a vital component of any comprehensive cancer control plan. "Creating access to treatment-screening is a waste of effort if you don't link it to care" "The national insurance system does not fully cover payment" "Patients ask for new treatments, however, they are not covered by insurance" "It is important to eradicate drug lag and make good medication available as soon as possible"
## Centers of excellence
Specialized liver cancer centers to provide coordinated surveillance, treatment and research within a national liver cancer program.
"Transfer patients with a HCC diagnosis to a tertiary hospital to receive state-of-the-art treatment" "There is no organization that brings all liver cancer research together under one roof" "Build a large center, experienced with international techniques, with a large number of patients" "We need to continue to preach to establish centers of excellence with multidisciplinary efforts"
Clinical education Improve primary care provider's awareness of the benefits of screening and early treatment, and necessary skills in risk assessment.
"Most of the educational resources need to go into educating healthcare professionals" "Increase awareness among general practitioner, most are not aware" "Education of general practitioners concerning the screening of HCC, and gastroenterologists too" "We need to focus on the general education for primary care physicians so they will become vigilant"
## Measuring social burden
Accurate measures of risk factors, cirrhosis, liver cancer, the societal costs of illness and the benefits of improving liver cancer care.
"Research the epidemiology of liver cancer, I think that we underestimate liver cancer by 50%" "Prevalence, surveillance, burden of disease, effective and cost-effective strategies" "Know the epidemiological trend for nonalcohol fatty liver disease and its impact on HCC incidence" "We need to have some comparison about how many lives we can save if we improve"
Monitoring of atrisk populations National surveillance programs for at-risk patients through expert services to diagnose HCC in early stages and improve outcomes.
"Get at-risk patients into adequate screening programs at appropriate intervals and tested by experts" "Of cause surveillance programs are important to prevent or to detect early HCC" "There should be a national surveillance program for liver cirrhosis" "Monitor high-risk patients so if they develop HCC they can be diagnosed at an early stage and treated"
## Multidisciplinary management
Diagnosis, treatment decisions and follow-up of all HCC patients through collaborative teams of all relevant specialists.
"Follow-up of HCC patients should be in a multidisciplinary team of different specialists" "Collaboration among physicians, surgeons, radiologists and oncologists is very poor" "Create an appropriate interdisciplinary board where every single patient is evaluated by this team" "It is very important to appreciate that this disease is heterogeneous with regards to the etiology"
National guidelines National standards for diagnosis and guidelines for screening, surveillance, treatment and palliation related to liver cancer.
"There are no national guidelines on how to deal with patients with liver cancer" "There should be a national treatment strategy recognized and outcomes captured" "There is a lack of standardization of clinical diagnosis and treatment" "Information exchange among world leaders to prepare a global standard for prevention and treatment" Finally, a well functioning system must have its components well connected. In our model, Centers of excellence, Multidisciplinary management, National guidelines and Research infrastructure are important Connections of a comprehensive liver cancer control plan. The strategies identified here parallel some of the strategies embedded in the WHO guidelines for general comprehensive cancer control with two exceptions. First, strategies for pain relief/palliative care were not identified as an important cancer control strategy by our clinical respondents. This may have been different if more variety in the types of stakeholders were included in our sample (e.g. we had no nurses, patients or advocates), but there may be a lack of advocacy for liver cancer more generally [bib_ref] Prevention and control of viral hepatitis: the role and impact of patient..., Fitzsimons [/bib_ref]. This said, pain relief/palliative care can be seen as belonging to our Access to treatment strategy. Second, we do not differentiate strategies for implementation in low, middle and high income countries [bib_ref] Identifying Important Breast Cancer Control Strategies in Asia, Latin America and the..., Bridges [/bib_ref] , nor did we examine such heterogeneity in priorities. These two differences highlight the need for further research to differentiate priorities across different stakeholders (including advocates, where they may exist) and across different countries in Asia and beyond.
While this paper identified certain priorities for implementation in an Asian comprehensive liver cancer control plan, it is important to compare these to current policies in Asia. Highest priority was given to Monitoring of at-risk populations, which has been shown to facilitate early diagnosis [bib_ref] Regular surveillance by imaging for early detection and better prognosis of hepatocellular..., Noda [/bib_ref]. Such surveillance programs are related to surveillance in primary care [bib_ref] Impact of surveillance on survival of patients with initial hepatocellular carcinoma: a..., Toyoda [/bib_ref] , which may account for the low value given to Risk- "Greater public awareness of liver disease, risk factors and the fact that good treatments are available" "There is an absolute ignorance among the public and there is a clear need for education" "Patient groups are limited to popular types of cancer, but HCC is mainly the cancer of the poor" "Support experts to handle the details of patient advocacy so prevention and treatment could benefit"
## Research infrastructure
Funding, personnel, and facilities to conduct relevant basic, clinical and translational liver cancer research throughout the health system.
"There is no specific program for HCC with public funding ... research infrastructure is always needed" "Train physicians who can lead clinical trials ... we also need research nurses" "Get thorough scientific research for HCC, genetics, biology, the pathways, it is very important" "There is an uneven distribution of research funding and the lack of grass-roots research funding"
## Risk-assessment and referral
Risk stratification conducted by primary care providers who refer patients to appropriate surveillance provided regularly by experts.
"Identify at-risk patients, encourage them to be screened, and link them to appropriate care" "Primary doctors should not be treating viral hepatitis, they should be detecting it" "GPs don't consider it necessary and don't perform screening in patients with diagnosed cirrhosis" "We have very inefficient tools for identifying the high risk patients"
## Transplantation infrastructure
Improve awareness and capacity for organ donation, more capacity for transplantation, and alternatives to cadaveric transplantation "The situation cannot be altered without donors, but there is not much social infrastructure to support it" "It has been a major necessity to promote more cadaveric liver transplantation for more than decade" "The only shortcoming is transplantation, cadaveric transplantation is standard in other countries" "Real awareness of organ donation. There are some examples in the media, but still nothing happens".
Representative quotes remain unidentified to ensure anonymity and confidentiality of respondents assessment and referral (i.e. respondents found the former more beneficial to the latter). This said, risk stratification may be important to target surveillance strategies [bib_ref] Hepatocellular carcinoma, Marrero [/bib_ref]. While Centers of excellence were only discussed by a minority of respondents in the qualitative interviews, this strategy was considered a priority in the conjoint analysis. Such specialized centers have been shown to be of value in early surveillance and improved outcomes for HCC in Japan [bib_ref] Surveillance program for early detection of hepatocellular carcinoma in Japan: results of..., Ando [/bib_ref]. Priority was also placed on Access to treatment by respondents in both the qualitative and quantitative portions of this study. While lack of robust financing systems is a major barrier in many Asian countries, barriers to access persist in those countries with national health insurance. Other barriers include a lack of reimbursement, high copayments, a lack of specialty centers, the availability of specialists and awareness of the disease among primary care physicians and the general public [bib_ref] Hepatitis B: Overview of the burden of disease in the Asia-Pacific region, Lesmana [/bib_ref]. A shortage of organ donors and subsequent waiting lists also pose barriers to access to transplantation [bib_ref] The continuing challenge of hepatic cancer in Asia, Lai [/bib_ref].
There are some omissions from our set of strategies. For example, in addition to the absence of pain management and palliative care, our study did not specifically characterize hepatitis control as a strategy. However, prevention (e.g. through HBV vaccination), treatment and control (which would include treatment of hepatitis associated with HCC) are within the descriptors for Access to treatment. While quality hepatitis control exists in many Asian countries [bib_ref] The management of hepatocellular carcinoma in Asia: a guideline combining quantitative and..., Song [/bib_ref] [bib_ref] Nationwide Hepatitis B Vaccination Program in Taiwan: Effectiveness in the 20 Years..., Chien [/bib_ref] , hepatitis control must be a priority in many countries not included in this study [bib_ref] Sobue T: Cancer epidemiology and control in peninsular and island South-East Asia-past,..., Moore [/bib_ref]. It is also important to consider some alternative interpretations of the data in this study. One interpretation is that valuation of some strategies in the qualitative analysis may be as a result of framing effects in the presentation of the choice tasks. For example, Measuring social burden, which received a negative value, was described as "Measuring the social burden of liver cancer" (see [fig_ref] Figure 1: An example of a conjoint analysis task [/fig_ref]. Here respondents may have found this label ambiguous, or as implying factors that were not implicit in the qualitative analysis. This label may have been better described with terms such as measuring incidence or prevalence, terms that are more familiar to the respondents.
Risk-assessment and referral received a negative valuation despite being the most frequently discussed strategy in the qualitative analysis. Here several factors may have contributed to this aversion. First, the factor was described in the conjoint tasks as "Improved risk-assessment and referral by primary care" (see [fig_ref] Figure 1: An example of a conjoint analysis task [/fig_ref] , and the "improved" may have unduly framed the factor (especially for countries that have good risk-assessment mechanisms) or made the factor ambiguous (especially for those who do not have such mechanisms). Second, it may have been more accurate to refer to this as "continuous surveillance" of at risk populations. Third, there was some confusion between "risk-assessment and referral"-mechanisms to stratify those at-risk of developing HCC and referring them to appropriate care-and "monitoring of at-risk populations"-mechanisms of surveillance for patients identified as being at-risk, preferably in specialty care. Finally, there may be heterogeneity in the valuation of this factor across the study countries, i. e. this may be a priority in some countries, but not in others, potentially because such services are already provided or because systems are not based upon primary care providers originating risk assessment and diagnosis.
While this study is motivated by a need for comprehensive liver cancer control in Asia [bib_ref] Identifying Important Breast Cancer Control Strategies in Asia, Latin America and the..., Bridges [/bib_ref] , it also highlights a more general need for more quantitative research methods to guide priority setting in health care. The prioritization of limited resources across competing demands presents an "economic challenge and a political puzzle" [bib_ref] Explicit and implicit rationing: taking responsibility and avoiding blame for health care..., Ham [/bib_ref] , but is a vital element of systematic planning in public health [bib_ref] Basic priority rating model 2.0: current applications for priority setting in health..., Neiger [/bib_ref]. While some health care planners utilize multiple evidence sources (both qualitative and quantitative) for the purposes of priority setting [bib_ref] Evidence-based priority-setting: what do the decisionmakers think?, Mitton [/bib_ref] , stakeholder engagement in policy often is limited to nominal groups or consensus-based approaches (e.g. Delphi techniques) [bib_ref] A checklist for health research priority setting: nine common themes of good..., Viergever [/bib_ref] [bib_ref] Risk perception and priority setting for intervention among hepatitis C virus and..., Schwarzinger [/bib_ref]. As such, this study makes a significant contribution towards demonstrating the value of conjoint analysis in prioritization of health care policy strategies and challenges the soundness of consensusbased approaches.
# Conclusions
Identified strategies can be conceptualized as the ABCs of comprehensive liver cancer control as they focus on Antecedents, Better care and Connections within a national strategy. Some concordance was found between the qualitative and quantitative methods (e.g. Monitoring of at-risk populations), but substantial differences were also identified (e.g. qualitative methods gave highest priority to risk-assessment and referral, but it was the lowest for the quantitative methods), which may be attributed to differences between the methods and study populations, and potential framing effects in choice tasks. Continued research will provide more generalizable estimates of priorities and account for variation across stakeholders and countries.
# Funding
This study was funded, in part, by Bristol-Myers Squibb. The funders had no role in study design, data collection and analysis, selection of respondents, decision to publish, or preparation of the manuscript.
[fig] Figure 1: An example of a conjoint analysis task. [/fig]
[fig] Figure 2: A comparison of priorities using qualitative and quantitative methods. [/fig]
[table] Table 1: Strategies for comprehensive liver cancer control [/table]
[table] Table 2: Importance of strategies for liver cancer controlRobust standard errors in parentheses. [/table]
|
Primary Pulmonary Hodgkin Lymphoma
Primary pulmonary Hodgkin lymphoma (PPHL) is a rare disease. Herein, we report a case of PPHL with diagnostic concerns encountered during initial evaluation which is of paramount importance to keep the differential diagnosis in cases with high index of suspicion for this rare entity.
# Introduction
Primary pulmonary Hodgkin lymphoma (PPHL) is a rare disease. It occurs when the clonal lymphoid proliferation affect the lungs, and has no extra pulmonary spread at time of diagnosis or within the following 3 months. PPHL without peripheral lymphadenopathy or hepatosplenomegaly is exceptionally uncommon. [bib_ref] Giant thoracic mass: an unusual presentation of primary pulmonary Hodgkin's lymphoma, Mcelnay [/bib_ref] [bib_ref] Primary Hodgkin's disease of the lung, Chetty [/bib_ref] Pulmonary involvement is reported to occur in 15-40% of patients with Hodgkin's disease. However, the presentation of PPHL is very rare, with just over 70 cases recorded internationally between 1927 and 2006. [bib_ref] Giant thoracic mass: an unusual presentation of primary pulmonary Hodgkin's lymphoma, Mcelnay [/bib_ref] A case of PPHL is herein reported with a brief description, as a diagnostic challenge. We assessed the clinical, radiological, diagnostic and pathological features along with treatment and prognosis.
## Case report
A 15-year-old female of Saudi origin presented with persistent fever for more than one month duration. Over a period of one year, she had productive cough, weight loss and night sweats. She had no relative past medical history and was a non-smoker.
Respiratory examination revealed marked decrease air entry with dull percussion of the left upper lung zone. Physical examination did not reveal clubbing, peripheral lymphadenopathy or hepatosplenomegaly.
Laboratory investigation revealed anemia and severe leucopenia. Chest X-ray showed a persistent homogenous opacity occupying left upper lung zone for more than 5 weeks [fig_ref] Figure 1: A [/fig_ref]. Chest computed tomography (CT) scan images showed consolidation and collapse of the left upper lung lobe with air bronchogram and pleural thickening [fig_ref] Figure 1: A [/fig_ref]. CT guided biopsy was insignificant.
Bronchoscopy revealed normal macroscopic appearance of airways and absence of endobronchial lesions or foreign body.
The internal medicine and infectious disease team suggested tuberculosis (TB) pneumonitis as an initial diagnosis and started the patient on anti TB treatment in addition to antibiotics without improvement over a period of one month (if TB is suspected then acid-fast bacilli stain and mycobacterial culture should have been done).
Patient's conditions deteriorated and she became severely ill with persistent fever, significant weight loss and marked leucopenia. Based on clinical picture and failure to response to medical treatment, we thought that this is a case of lung gangrene that will soon develop cavitation into the consolidated lobe. We decided to perform exploration and lobectomy. The patient underwent left posterolateral thoracotomy which revealed a hepatized and necrotic left upper lung lobe with extensive adhesions to chest wall, left upper lobectomy was carried out. There was neither mediastinal nor hilar lymphadenopathy. No pleural effusion and no pleural nodules or any other lesions.
## Histopathology
Histopathological examination showed Hodgkin lymphoma (HL), nodular sclerosis type. Images [fig_ref] Figure 2: A [/fig_ref] of lung tissue showed sheets of polymorph cells separated by fibrous septa; [fig_ref] Figure 2: A [/fig_ref] showed the sheets composed of small lymphocytes, histiocytes, eosinophils, neutrophils, plasma cells, large multiloculated giant cells and some Reed Sternberg cells with occasional lacunar cells. [fig_ref] Figure 2: A [/fig_ref] ,D shows the neoplastic Reed Sternberg cells interacting with CD15 and CD30 in an immunohistochemical analysis.
## Outcome and follow-up
After the left upper lobectomy the patient markedly improved, fever subsided and her general condition and laboratory investigation improved. She received postoperative chemoand radiotherapy. She remains well without evidence of recurrence 6 years later.
She only had a postoperative eventration of the left diaphragm, due to intraoperative injury of the left phrenic nerve [fig_ref] Figure 3: Follow up chest x-ray, six years after the surgery and finishing the... [/fig_ref]. Patient did not have any respiratory symptoms and refused to perform plication of the diaphragm.
# Discussion and conclusions
Primary pulmonary lymphomas are uncommon, comprising less than 1% of lung cancers and fewer than 1% of malignant lymphomas, and accounting for only 3.6% of extranodal lymphomas. [bib_ref] Primary pulmonary classical Hodgkin lymphoma with two recurrences in the mediastinum: a..., Homma [/bib_ref] Nodular sclerosis (NS) PPHL is the most common variety and comprises 60-70% of PPHL. [bib_ref] Primary pulmonary classical Hodgkin lymphoma: a case report, Schild [/bib_ref] The criteria for the diagnosis of PPHL include: i) histological features of Hodgkin's lymphoma, ii) restriction of the disease to the lung with or without minimal hilar lymph node involvement and iii) adequate clinical and/or pathological exclusion of the disease at distant sites. 1,4-7 Our case meets each criteria as mentioned above.
In the largest report, PPHL showed a slight female preponderance (1.4:1 F:M), with a bimodal age distribution (<35 and >60 years). The most common presenting symptoms are weight loss, fever, night sweats and dry cough. Dyspnea and hemoptysis are also common. Radiologically, PPHL typically involves the superior portions of the lungs, whereas secondary pulmonary involvement from Hodgkin's lymphoma shows a more random miliary distribution without zonal predilection. Many present as a solitary mass, alveolar consolidation or multiple nodules. Cavitary pulmonary lesions have a wide differential diagnosis. [bib_ref] Primary pulmonary Hodgkin's lymphoma and a review of the literature since, Cooksley [/bib_ref] The present case emphasizes that no radiological appearance is pathognomic for PPHL. The principal radiographic differential diagnosis of primary Hodgkin's lymphoma includes pseudolymphoma, lymphocytic interstitial pneumonia, lymphoid granulomatosis, bronchioloalveolar carcinoma, metastasis, and cryptogenic organizing pneumonia. [bib_ref] Unusual radiologic and histologic manifestations of primary pulmonary lymphoma, Baccari-Ezzine [/bib_ref] The Ann Arbor pulmonary lymphoma staging system was used for classification. Stage IE: lung only, could be bilateral; Stage II 1E: lung and hilar lymph nodes; Stage II 2E: lung and mediastinal lymph nodes; Stage II 2EW: Lung and chest wall or diaphragm; Stage III: lung and lymph nodes below the diaphragm; Stage IV: diffuse. [bib_ref] Forty years literature review of primary lung lymphoma, Parissis [/bib_ref] According to Ann Arbor staging system, a patient with PPHL would be either stage IE (involvement of a single extra nodal site) or IIE (localized involvement of extra nodal site and its contiguous lymph node chain.Hence the clinical presentation in primary pulmonary HL is non-specific, open-lung biopsy is needed in the majority of cases. Occasionally, FNA coupled with IHC stains is helpful. The microscopic appearances are typical of classic HL usually of either nodular sclerosis or mixed cellularity type, and IHC staining's are confirmatory. [bib_ref] Primary pulmonary classic Hodgkin's lymphoma, Binesh [/bib_ref] Because of the low incidence of primary pulmonary lymphoma, a high index of suspicion is required to diagnose on time and avoid unnecessary delay. As demonstrated by our case, even in the presence of classical B-symptoms, the clinical and radiological data was inconclusive for accurate diagnosis of pulmonary lymphoma. The condition suggested acute necrotizing pneumonia and lung gangrene with obstruction of blood supply to the affected lobe as a cause of failure to response to medical treatment. This was actually our indication for surgical management and lobec-Case Report tomy. The correct diagnosis was made after surgical resection. Our experience demonstrates that suspicious lesions should undergo open lung biopsy or surgical resection for ultimate diagnosis as supported in literature as well. [bib_ref] Primary pulmonary Hodgkin's disease and tuberculosis in an 11-year-old boy: case report..., Codrich [/bib_ref] [bib_ref] Primary pulmonary hodgkin lymphoma simulating a mediastinal tumour: an uncommon occurrence, Fratoni [/bib_ref] [bib_ref] Primary pulmonary Hodgkin's disease: report of two cases, Boshnakova [/bib_ref] [bib_ref] Management of acute necrotizing lung infections: the role of surgery, El-Baz [/bib_ref] Learning points
The key points of the present paper are: PPHL is a rare disease; nodular sclerosis PPHL is the most common variety; open lung biopsy is often required for definitive diagnosis.
[fig] Figure 1: A) Chest x-ray showing a homogenous opacity occupying the left upper lung zone. B-C) Axial and reconstruction computed tomography images are showing consolidation and collapse of the left upper lung lobe and lingula with air bronchogram. [/fig]
[fig] Figure 3: Follow up chest x-ray, six years after the surgery and finishing the chemo and radiotherapy. [/fig]
[fig] Figure 2: A) Lung Hodgkin lymphoma, large atypical Reed Sternberg like cells in background of inflammatory cells infiltration surrounded by fibrous septa on ×200 magnification. B) Lung Hodgkin Lymphoma, diagnostic binuclear Reed Sternberg cells with eosinophillic inclusions like nucleoli on ×400 magnification. C) Hodgkin lymphoma, immunohistochemical analysis showing clusters of Reed Sternberg cell react (cells react) with CD15 on ×400 magnification. D) Hodgkin lymphoma, immunohistochemical analysis showing clusters of Reed Sternberg cell reacts (cells react) with CD30 on ×400 magnification. [/fig]
|
Biotechnical paving of recombinant enterocin A as the candidate of anti-Listeria agent
Background: Enterocin A is a classic IIa bacteriocin isolated firstly from Enterococcus faecium CTC492 with selective antimicrobial activity against Listeria strains. However, the application of enterocin A as an anti-Listeria agent has been limited due to its very low native yield. The present work describes high production of enterocin A through codon optimization strategy and its character study. Results: The gene sequence of enterocin A was optimized based on preferential codon usage in Pichia pastoris to increase its expression efficiency. The highest anti-Listeria activity reached 51,200 AU/ml from 180 mg/l of total protein after 24 h of induction in a 5-L fermenter. Recombinant enterocin A (rEntA), purified by gel filtration chromatography, showed very strong activity against Listeria ivanovii ATCC 19119 with a low MIC of 20 ng/ml. In addition, the rEntA killed over 99% of tested L. ivanovii ATCC19119 within 4 h when exposed to 4 × MIC (80 ng/ml). Moreover, it showed high stability under a wide pH range (2-10) and maintained full activity after 1 h of treatment at 80°C within a pH range of 2-8. Its antimicrobial activity was enhanced at 25 and 50 mM NaCl, while 100-400 mM NaCl had little effect on the bactericidal ability of rEntA. Conclusion: The EntA was successfully expressed in P. pastoris, and this feasible system could pave the pre-industrial technological path of rEntA as a competent candidate as an anti-Listeria agent. Furthermore, it showed high stability under wide ranges of conditions, which could be potential as the new candidate of anti-Listeria agent.
# Background
Bacteriocins are antimicrobial peptides synthesized in the ribosome and secreted into medium to establish a competitive advantage in their environment by eliminating competitors to gain resources [bib_ref] Development of Class IIa bacteriocins as therapeutic agents, Lohans [/bib_ref]. Bacteriocins are generally classified in terms of size, structure, and modifications. Class I bacteriocins are lantibiotics. Class II bacteriocins consist of small peptides that do not contain modified residues. Class III bacteriocins usually are large and heat-labile proteins [bib_ref] Bacteriocins: developing innate immunity for food, Cotter [/bib_ref]. The well-known bacteriocin is nisin, a class I bacteriocin, which is widely used in commerce [bib_ref] Anti-Listeria effect of enterocin A, produced by cheese-isolated Enterococcus faecium EFM01, relative..., Ennahar [/bib_ref]. Recently, many reports clearly indicate that bacteriocins of class IIa have greater potential as antimicrobial agents [bib_ref] Bacteriocins-a viable alternative to antibiotics?, Cotter [/bib_ref] with a narrower inhibitory spectrum to Listeria strains than nisin [bib_ref] In vitro inhibition activity of nisin A, nisin Z, pediocin PA-1 and..., Blay [/bib_ref]. Listeria, the most common pathogen in food, can lead the host to suffer from serious diseases such as enteritis, sepsis, meningitis and abortion [bib_ref] A novel PrfA-regulated chromosomal locus, which is specific for Listeria ivanovii, encodes..., Engelbrecht [/bib_ref]. The mortality rate caused by listeriosis is between 15 and 30% [bib_ref] Listeriosis in patients receiving biologic therapies, Bodro [/bib_ref] [bib_ref] Foodborne illness acquired in the United States-major pathogensl, Scallan [/bib_ref]. Additionally, some strains of L. monocytogenes easily acquire resistance to many antibiotics [bib_ref] Occurrence and antibiotic resistance profiles of Listeria monocytogenes isolated from seafood products..., Fallah [/bib_ref]. To control food contamination and listeriosis effectively, more or better anti-listerial drugs are needed.
Enterocin A (EntA), with many antimicrobial merits, is a class IIa bacteriocin that was first isolated from Enterococcus faecium CTC492 in the mid-1990s. Its mature form is composed of 47 amino acids with two disulfide bridges [bib_ref] Biochemical and genetic characterization of enterocin A from Enterococcus faecium, a new..., Aymerich [/bib_ref]. It shows high activity, particularly against Listeria species at nanomolar concentrations [bib_ref] Enterococcus faecium P21: a strain occurring naturally in dry-fermented sausages producing the..., Herranz [/bib_ref]. The native EntA has proven to effectively inhibit L. monocytogenes in fermented foods [bib_ref] Controlling Listeria monocytogenes in cottage cheese through heterologous production of enterocin A..., Liu [/bib_ref] [bib_ref] Technological performance of the enterocin A producer Enterococcus faecium MMRA as a..., Rehaiem [/bib_ref]. However, the low levels of bacteriocins secreted from natural strains do not meet the requirements of the industrial scale and have limited its application to study stages thus far. Therefore, various heterologous expressions were attempted in lactic acid bacteria, Escherichia. coli (E.coli) and yeast [bib_ref] Controlling Listeria monocytogenes in cottage cheese through heterologous production of enterocin A..., Liu [/bib_ref] [bib_ref] Cloning, production and functional expression of enterocin P, a sec-dependent bacteriocin produced..., Gutiérrez [/bib_ref] [bib_ref] A versatile system for the expression of nonmodified bacteriocins in Escherichia coli, Ingham [/bib_ref] [bib_ref] Protein secretion in Lactococcus lactis: an efficient way to increase the overall..., Loir [/bib_ref] , but their actual production levels were not desirable and left room for improvement. Pichia pastoris is considered to be a promising system because the target protein can be directly secreted into culture medium. It was reported that the production and bactericidal titer of enterocin P expressed by P. pastoris X-33 was 3.7-and 16-fold higher (28.2 μg/ml and 1,024 BU/ml), respectively, than that from the native E. faecium P13 [bib_ref] Production of enterocin P, an antilisterial pediocin-like bacteriocin from Enterococcus faecium P13,..., Gutiérrez [/bib_ref] ; in fact, even though the level of 45.1 μg/ml of recombinant enterocin A expressed by P. pastoris [bib_ref] Cloning, production, and functional expression of the bacteriocin enterocin A, produced by..., Borrero [/bib_ref] was still too low for its industrial production and end application, it demonstrates the potential to increase its productivity to be as high as possible and to further easily characterize its purification and properties. However, there are only few studies at the modification of bacteriocin genes, such as gene synthesis or codon optimization, which is considered as a promising technique for increasing protein expression level [bib_ref] Improvement of Aspergillus sulphureus endo-β-1,4-xylanase expression in Pichia pastoris by codon optimization..., Li [/bib_ref] ; thus, further work with this system is necessary to achieve an increased protein expression level of target gene.
Due to the high anti-Lister activity of EntA and its low yield either in native strain and recombinant expression system, the EntA gene was optimized by the preferential codon usage of P. pastoris and was expressed into medium as recombinant EntA (rEntA). The purification of rEntA from ferment supernatant was tried by four methods including gel filtration chromatography, then the antimicrobial activity, proteolytic sensibility and stabilities of heat, pH and salt of purified rEntA were examined.
# Results
## Construction and transformation of the expression vector
Compared to naturally occurring EntA, the base codons coding for 37 residues (78.72%) in total 47 amino acids were optimized by the preferential codon usage of P. pastoris [fig_ref] Figure 1: Construction of the expression plasmid pPICZαA-EntA [/fig_ref]. The GC content of the full target sequence increased from 41.13% to 41.9%. The gene sequence of the optimized EntA was synthesized and inserted into pPICZαA between XhoI and XbaI sites [fig_ref] Figure 1: Construction of the expression plasmid pPICZαA-EntA [/fig_ref] , C). The expression vector pPICZαA-EntA was transferred into competent E. coli DH5α cells. Resulting transformants were confirmed by PCR and DNA sequencing. Correct plasmid and control vector pPICZαA were linearized by PmeI and transferred into competent P. pastoris X-33 cells by electroporation. Positive transformations were screened and confirmed by colony PCR.
Expression of rEntA in shake flasks and at the fermenter level
The heterologous expression of rEntA in P. pastoris X-33 was induced by methanol at the concentration of 0.5% and analyzed by agar diffusion and Tricine-SDS-PAGE. P. pastoris X-33 containing the empty pPICZαA vector was used as a negative control. As shown in [fig_ref] Figure 2: Expression and purification of rEntA [/fig_ref] , after 12 h of methanol induction, the antibacterial activity of the supernatants of P. pastoris X-33 (pPICZαA-EntA) was observed. Its antibacterial activity reached maximum with 6,400 AU/ml after 24 h of methanol induction. However, the antimicrobial activity decreased from 48 to 72 h. No antibacterial activity was detected in the supernatants of P. pastoris X-33 (pPIC-ZαA). The results of the MALDI-TOF MS for fermentation supernatants indicated that the molecular weight of rEntA was 4,830.1 Da, which was consistent with its theoretical value of 4,829 Da [fig_ref] Figure 2: Expression and purification of rEntA [/fig_ref].
To increase the production of rEntA, high-density fermentation of the recombinant yeast was performed using a 5-L fermenter. Although the total supernatant protein and biomass reached 365 mg/l and 343 g/l after induction for 90 h, the maximal antimicrobial activity was 51200 AU/ml (180 mg/l) after induction for 24 h [fig_ref] Figure 2: Expression and purification of rEntA [/fig_ref] , which was 8-fold higher than that found at the shake-flask level. Figures 2B and D clearly showed that rEntA was rapidly degraded after 72 h of induction. Moreover, the expression of rEntA in the fermenter could be detected directly by Coomassie blue staining [fig_ref] Figure 2: Expression and purification of rEntA [/fig_ref] , while its expression in the shake-flask could only be detected by silver staining (data not shown).
## Purification of renta
The rEntA was purified from the ferment supernatant after a 24-h induction in a 5-L fermenter. The bacteriocin activity of 6.40 × 10 5 AU/mg with a 2.25-fold increase was obtained after gel filtration. The purified rEntA was analyzed by Tricine-SDS-PAGE and showed a band at 4.8 kDa representing the target protein band [fig_ref] Figure 2: Expression and purification of rEntA [/fig_ref] , corresponding with its theoretical molecular weight.
## Antimicrobial spectrum of renta
Only L. ivanovii ATCC19119, E. faecalis CGMCC1.130 and E. faecalis CGMCC1.2024 were sensitive to rEntA in the 16 tested strains. Other Gram-positive bacteria, such as E. faecium CGMCC1.2136, S. aureus ATC-C25923, S. epidermidis ATCC26069, B. licheniformis CGMCC1.265, and B. coagulans CGMCC1.2407, were found to be resistant to rEntA. All of the Gram-negative bacteria strains were resistant to rEntA in this assay [fig_ref] Table 1: Antimicrobial spectrum of rEntA [/fig_ref]. The MIC and MBC of rEntA against L. ivanovii ATCC19119 were 20 ng/ml and 80 ng/ml, respectively, and were lower than those of ampicillin (390 ng/ml and 1560 ng/ml, respectively).
## In-vitro killing curve assay
The time-killing kinetics curve showed that the amount of L. ivanovii ATCC19119 increased from 6.63 log 10 CFU/ml to 9.48 log 10 CFU/ml within 10 h in the absence of rEntA. The decrease in the counts of L. ivanovii ATCC19119 varied considerably depending on the concentration of rEntA. For example, the maximum viability loss (MVL), which was approximately 0.44 log 10 CFU/ml (~60% reduction in CFU), was reached within 2 h in 1 × MIC of rEntA. The 2 × MIC of rEntA could cause approximately 1.42 log 10 CFU/ml viability loss (96% reduction) within 6 h. Moreover, the MVL of L. ivanovii treated by rEntA at 4 × MIC was approximately 2.03 log 10 CFU/ml (>99% reduction in CFU) within 4 h. Although rEntA could inhibit the growth of L. ivanovii ATCC19119, the survivors resumed growth at 1× and 2 × MIC of rEntA and 2 × MIC ampicillin for L. ivanovii ATCC19119 after MVL was achieved [fig_ref] Figure 3: Time-kill curves of rEntA [/fig_ref]. However, L. ivanovii ATCC19119 treated by 4 × MIC of rEntA did not show re-growth within 10 h, revealing that 80 ng/ml rEntA could effectively inhibit the growth of pathogenic bacteria for an extended time.
Effects of pH, temperature, proteolytic enzymes and NaCl on the activity of rEntA As shown in , rEntA was highly stable at a wide range of pH values. The activity of rEntA was maintained completely within a pH range of 2-8 at 37°C for 12 h and was 75% retained even after incubation at a pH of 10 for 12 h. Furthermore, the antimicrobial activity of rEntA was not affected by heat treatment at 37, 60, 80 and 100°C for 1 h under acid conditions (pH 2 and 4) . The residual activity decreased to 20% at a pH of 10 at 80°C, to 50% at a pH of 6, 8 at 100°C, and to 10% at a pH of 10 at 100°C. In addition, the antimicrobial activity of rEntA was completely abolished by pepsin and trypsin treatment, but it retained 16.7% of initial antimicrobial activity after papain treatment at 37°C for 1 h .
The antimicrobial activity of rEntA against L. ivanovii ATCC19119 was slightly enhanced by the addition of 25 and 50 mM NaCl [fig_ref] Figure 5: Effect of NaCl concentration on the activity of rEntA [/fig_ref]. The lowest amount of 2.43 log 10 CFU/ml was observed with a treatment of rEntA (12,800 AU/ml) in 25 mM NaCl (44.52% of that at 0 mM NaCl). The other treatments, from 100 -400 mM NaCl, had no significant effect on the bactericidal ability of rEntA [fig_ref] Figure 5: Effect of NaCl concentration on the activity of rEntA [/fig_ref]. In the controls without rEntA, growth was not influenced by NaCl (0 -400 mM) [fig_ref] Figure 5: Effect of NaCl concentration on the activity of rEntA [/fig_ref].
# Discussion
Bacteriocin has attracted attention in recent years for its potential application as a food preservative and therapeutic antimicrobial agent [bib_ref] Natural antimicrobial peptides from bacteria: characteristics and potential applications to fight against..., Hassan [/bib_ref]. However, low production of these bacteriocins by native strains cannot meet the requirements of commercial applications. Moreover, some Enterococci strains were recognized as opportunistic pathogens associated with lots of infections [bib_ref] Diversity of enterococcal bacteriocins and their grouping in a new classification scheme, Franz [/bib_ref]. Attempts to produce bacteriocins by using safe heterologous hosts have been undertaken in recent years [bib_ref] Production of enterocin P, an antilisterial pediocin-like bacteriocin from Enterococcus faecium P13,..., Gutiérrez [/bib_ref] [bib_ref] Heterologous coproduction of enterocin A and pediocin PA-1 by Lactococcus lactis: detection..., Martínez [/bib_ref] [bib_ref] Heterologous expression of enterocin A, a bacteriocin from Enterococcus faecium, fused to..., Klocke [/bib_ref] , including some typical expression systems such as E. coli, L. lactis, and P. pastoris. Although E. coli and L. lactis are widely used in heterologous protein expression because of their easy operation and safety [bib_ref] Cloning, production and functional expression of enterocin P, a sec-dependent bacteriocin produced..., Gutiérrez [/bib_ref] [bib_ref] Protein expression vector and secretion signal peptide optimization to drive the production,..., Borrero [/bib_ref] , they are not suitable for bacteriocins due to toxicity to the host [bib_ref] Hafner-Bratkovic I, Jerala R: Expression, purification and structural studies of a short..., Zorko [/bib_ref] and low recovery percentages from the fusion protein [bib_ref] High-level expression of an antimicrobial peptide histonin as a natural form by..., Kim [/bib_ref].
Many bacteriocins, such as enterocin P [bib_ref] Production of enterocin P, an antilisterial pediocin-like bacteriocin from Enterococcus faecium P13,..., Gutiérrez [/bib_ref] , hiracin JM79 [bib_ref] Cloning and heterologous production of hiracin JM79, a Sec-dependent bacteriocin produced by..., Sánchez [/bib_ref] , enterocin L50 [bib_ref] Use of the yeast Pichia pastoris as an expression host for secretion..., Basanta [/bib_ref] , pediocin PA-1 [bib_ref] Production of pediocin PA-1 in the methylotrophic yeast Pichia pastoris reveals unexpected..., Beaulieu [/bib_ref] and EntA [bib_ref] Cloning, production, and functional expression of the bacteriocin enterocin A, produced by..., Borrero [/bib_ref] , have been expressed as active forms in P. pastoris, but their expression levels remained low (below 280 mg/l). It is known that codon optimization is a useful strategy to increase the yield of target protein during heterogeneous expression. Many antimicrobial peptides, such as plectasin [bib_ref] Expression of plectasin in Pichia pastoris and its characterization as a new..., Zhang [/bib_ref] , NZ2114 [bib_ref] High expression of a plectasin-derived peptide NZ2114 in Pichia pastoris and its..., Zhang [/bib_ref] and AgPlectasin [bib_ref] Design, expression, and characterization of a novel targeted plectasin against methicillin-resistant Staphylococcus..., Mao [/bib_ref] , were expressed with high production through codon-usage optimization in our laboratory. In addition, Divercin V41, a class IIa bacteriocins was also expressed through this system [bib_ref] Heterologous expression and purification of active Divercin V41, a Class IIa bacteriocin..., Richard [/bib_ref]. These cases encouraged us to use codon optimization to break through the bottleneck of low yield in heterologous expression of EntA. The total protein level in the Note: "+" refers to positive antimicrobial activity (inhibition zone > 6 mm); "-" refers to negative antimicrobial activity (inhibition zone ≤ 6 mm). [fig_ref] Figure 2: Expression and purification of rEntA [/fig_ref] after the gene was optimized. Although the yield of target peptide was still low, and even lower than 280 mg/l as the highest result of expression in case of enterocin L50 in P. pastoris [bib_ref] Use of the yeast Pichia pastoris as an expression host for secretion..., Basanta [/bib_ref] , it was much higher than that of Pediocin PA-1 (0.4 mg/l), Enterocin P (0.006 mg/l), Divercin V41 (23 mg/l) and EntA (0.027 mg/l) expressed in E. coli and L. lactis [bib_ref] Cloning, production and functional expression of enterocin P, a sec-dependent bacteriocin produced..., Gutiérrez [/bib_ref] [bib_ref] Heterologous coproduction of enterocin A and pediocin PA-1 by Lactococcus lactis: detection..., Martínez [/bib_ref] [bib_ref] Heterologous expression and purification of active Divercin V41, a Class IIa bacteriocin..., Richard [/bib_ref]. Furthermore, the production of rEntA increased 2.99-times compared with its Effects of pH, temperature and proteolytic enzymes on the rEntA activity. A, pH stability of rEntA. Purified rEntA was incubated in buffers with a pH range from 2 to 10 at 37°C for 12 h. The initial activity of the sample in a buffer with a pH of 6 was described as 100% activity. B, Thermal stability of EntA. Purified rEntA was incubated in buffers with a pH range from 2 to 10 at temperatures of 37, 60, 80, and 100°C for 1 h. The initial activity of the sample in a buffer with a pH of 6 was described as 100% activity. C, Proteolysis resistance of rEntA. Purified rEntA was incubated with pepsin, papain and trypsin at 37°C for 4 h. The residual antimicrobial activity of samples was tested after the pH was readjusted to pH 6.0 with sodium phosphate buffer.
native sequence expressed in P. pastoris (45.1 mg/l), which indicated codon optimization is a good tool to enhance expression efficiency and level in P. pastoris, and at the same time, it also left a large room to improve in future work at the similar aim and technical scheme. However, the maximal activity of rEntA in the supernatant was reached at an early stage (24 h) of induction [fig_ref] Figure 2: Expression and purification of rEntA [/fig_ref]. This is similar to previous results in which the highest level of rEntA was reached at 36 h. An even earlier peak of rEntA at 6 h was observed in other yeasts such as Kluyveromyces lactis and Hansenula polymorpha [bib_ref] Cloning, production, and functional expression of the bacteriocin enterocin A, produced by..., Borrero [/bib_ref]. Obviously, its final successful application suffered from this strong decomposition in the supernatant at an earlier period of expression related to the possible disruption of rEntA to host cells and the proteolysis of the target protein. The latter situation was reported in "collagen-like" bacteriocin with a high cleavage by collagenase [bib_ref] Production of pediocin PA-1 in the methylotrophic yeast Pichia pastoris reveals unexpected..., Beaulieu [/bib_ref]. However, the exact mechanism of the above described early degradation and its solution should be further studied.
A series of methods, such as ion exchange chromatography (SP and CM FF), hydrophobic exchange chromatography (Phenyl HP), and gel filtration (Superose 12), were attempted to purify rEntA in this study. Only gel filtration could purify rEntA with a yield of 3.02 mg/l [fig_ref] Figure 2: Expression and purification of rEntA [/fig_ref] after attempts with SP FF, CM FF, and phenyl HP in which almost all rEntA was lost in the penetration peak (data not shown) due to unknown reasons.
Although there are different antibacterial spectrums between class IIa bacteriocins, they consistently have particularly high antibacterial activity against Listeria -the most common pathogen in foodat nanomolar concentrations [bib_ref] Development of Class IIa bacteriocins as therapeutic agents, Lohans [/bib_ref]. The MICs of purified native EntA from E. faecium T136 against Listerias ranged from 40 to 120 ng/ml. Similarly, rEntA also showed a narrow antibacterial spectrum [fig_ref] Table 1: Antimicrobial spectrum of rEntA [/fig_ref] including L. ivanovii ATCC19119, and with a low MIC value of 20 ng/ml, it is approximately 20-fold lower than that of ampicillin (390 ng/ml). The re-growth after MVL achievement was a common phenomenon when the Listeria was treated with bacteriocins such as EntA, pediocin, sakacin A and enterococcin EFS2 in relatively low concentrations (1× or 2 × MIC) [bib_ref] Anti-Listeria effect of enterocin A, produced by cheese-isolated Enterococcus faecium EFM01, relative..., Ennahar [/bib_ref] , but we found no re-growth after MVL within 10 h when 4 × MIC rEnA was used with the Listeria [fig_ref] Figure 3: Time-kill curves of rEntA [/fig_ref] , indicating that higher concentrations of rEnA are essential to inhibit the multiplication of Listeria.
The bactericidal activity and overall structure of Pediocin PA-1 and piscicolin 126 were well maintained at higher temperatures [bib_ref] Dynamic relationships among type IIa bacteriocins: temperature effects on antimicrobial activity and..., Kaur [/bib_ref] [bib_ref] Characterization of the chemical and antimicrobial properties of piscicolin 126, a bacteriocin..., Jack [/bib_ref]. The native EntA was stable at 100°C and acidic pH conditions [bib_ref] Enhancement of enterocin A production by Enterococcus faecium MMRA and determination of..., Rehaiem [/bib_ref]. We found that rEntA also exhibited high stability under a wide range of temperatures (37-80°C) and pH levels (2-8) . These properties were potentially due to the higher cysteine content of the antimicrobial peptides [bib_ref] Cloning and heterologous expression of the antibiotic peptide (ABP) genes from Rhizopus..., Yamada [/bib_ref] , similar to the EntA containing four cysteine residues. In addition, the antimicrobial activity of some bacteriocins (nisin, sakacin P and curvacin A) was significantly enhanced with the addition of NaCl from 0 to 1.17 M [bib_ref] Effect of ecological factors on the inhibitory spectrum and activity of bacteriocins, Gänzle [/bib_ref]. However, the activity of rEntA against Listeria was enhanced only at low NaCl concentrations (25 and 50 mM). Despite the unknown mechanisms of the above differential effects, the high stability of rEntA over wide ranges of temperature, pH, and NaCl concentration supports its use as a food preservative and drug candidate.
Due to the high content of basic and aromatic amino acids in class IIa bacteriocins, pediocin PA-1, enterocin B, plantaricin 423 and native EntA were very sensitive to the digestive proteases trypsin and pepsin [bib_ref] Enterococcus faecium P21: a strain occurring naturally in dry-fermented sausages producing the..., Herranz [/bib_ref] [bib_ref] Pediocin PA-1, a wide-spectrum bacteriocin from lactic acid bacteria, Rodriguez [/bib_ref]. Similarly, the purified rEntA, with 12.76% basic amino acids and 10.63% aromatic amino acids, was inactivated with trypsin and pepsin . This high sensitivity to digestive proteases of rEntA contributes to its safety in foods and drugs, during and after oral administration.
# Conclusion
In conclusion, rEntA, as an antimicrobial agent with merit, could selectively kill important and pathogenic Listeria and retain bio-activity over a wide range of pH values, temperature and NaCl concentrations. These excellent antibacterial properties make it a potential candidate as a food preservative and therapeutic antimicrobial agent. rEntA was successfully expressed in P. pastoris X-33 at the highest level of 51,200 AU/ml and was purified through a gel filtration column. This yeast system may be a feasible technological approach to produce rEntA as a potent anti-Listeria agent after further optimization.
# Methods
## Strains and vectors
Escherichia coli DH5α, Pichia pastoris X-33 and pPIC-ZαA were purchased from Invitrogen (Beijing, China). Target strains for the antimicrobial activity assays are listed in [fig_ref] Table 2: Strains used in the antimicrobial activity assays Pseudomonas aeruginosa CVCC 2087 CVCC... [/fig_ref]. Restriction enzymes were purchased from New England Biolabs (NEB, Beijing, China). The kits for plasmid extraction and DNA purification were purchased from Tiangen (Beijing, China). Other chemical reagents used in this research were all of analytical grade.
## Construction of the expression vector and transformation
The optimized EntA gene (GenBank accession No. KJ155693) was generated by the 'ReverseTranslateTool' (http://www.bioinformatics.org/sms2/rev_trans.html) according to the codon usage of P. pastoris (http://www. kazusa.or.jp/codon/). To express the target protein with a native N-terminus, the Kex2 signal cleavage site was fused to the EntA sequence. The designed sequence was synthesized by Sangon Biotech (Shanghai, China) and digested using XhoI and XbaI. Resulting DNA fragments were ligated into pPICZαA to generate the recombinant vector pPICZαA-EntA. The latter was transformed into E. coli DH5α, and positive transformants were confirmed by DNA sequencing. The recombinant plasmid was linearized with PmeI and transformed into P. pastoris X-33 competent cells by electroporation [bib_ref] Expression of plectasin in Pichia pastoris and its characterization as a new..., Zhang [/bib_ref]. Positive transformants were screened on YPDS medium containing 100 μg/ml of zeocin and further confirmed by colony-PCR.
## Expression of renta at the shake-flask level
The positive transformants were grown in BMGY medium until the cultures reached an OD600 nm of 5.0-6.0 at 30°C. Cells were harvested by centrifugation at 4000 rpm for 10 min and resuspended in BMMY medium to an OD600 nm of 1.0. Methanol was added daily to a final concentration of approximately 0.5%. Samples were taken at 0, 12, 24, 36, 48, 60 and 72 h for analysis.
## Expression of renta at the fermenter level
A single colony of P. pastoris X-33 (pPICZαA-EntA) was grown in 10 ml of YPD medium at 30°C overnight. The culture was inoculated into 200 ml fresh YPD medium and cultivated at 29°C to an OD600 nm of approximately 6.0. The 200-ml seed culture was transferred into a 5-L fermenter (Sartorius StedimBiotech) containing 1.8 L of basal salt medium with 45 g/L of NH 4 H 2 PO 4 , 20 g/L K 2 SO 4 , 0.4 g/L CaSO 4 , 15 g/L MgSO 4 7H 2 O, 6 g/L KH 2 PO 4 , 1.5 g/L KOH, and 200 ml 45% w/v glucose. The initial fermentation was a glucose batch phase (approximately 18 h). After exhaustion of the glucose, 50% w/v glucose was added for 6 h at a feed rate of 36 ml/h. After the glucose was exhausted, methanol was supplied from 2 to 12 ml/h. The whole fermentation period was performed at 29°C. During the glucose batch and glucose-fed phases, the pH was kept at 5.0 and increased to 5.5 at the methanol induction phase [bib_ref] Contribution of bovine lactoferrin inter-lobe region to iron binding stability and antimicrobial..., Bai [/bib_ref]. The protein in the supernatant was determined by the Bradford protein assay (Tiangen, Beijing, China) and Tricine-SDS-PAGE.
## Purification of renta
The supernatant with rEntA from P. pastoris X-33 (pPICZαA-EntA) X-33 was desalted by a gel filtration column (Sephadex G-25) with a flow rate of 2 ml/min and then freeze-dried and dissolved in 100 mM of ammonium acetate buffer. The sample was passed through a gel filtration column (Superose 12) and eluted with the same buffer at a flow rate of 0.5 ml/min. Purified rEntA was further lyophilized to remove ammonium acetate.
## Antimicrobial activity assay
Tested strains including L. ivanovii, E. faecalis, and E. faecium were grown in Mueller-Hinton (MH) broth containing 3% fetal bovine serum (FBS). S. epidermidis, B. subtilis, L. lactis, B. bifidum, B. licheniformis, B. coagulans and S. aureus were grown in MH broth. P. aeruginosa, E. coli and S. enteritidis were grown in LB medium. All tested strains were grown to 0.4 of OD600 nm at 37°C. One hundred microliters of the cell suspension was inoculated into 50 ml of preheated medium containing 1.5% agar. This was rapidly mixed and poured into a Petri dish. Sterile Oxford cups were put on the surface of the solidified media. Each cup was filled with 50 μl of samples [bib_ref] Expression of plectasin in Pichia pastoris and its characterization as a new..., Zhang [/bib_ref]. Titer assays were used to quantify the antimicrobial activity of rEntA according to the method of Liu [bib_ref] Controlling Listeria monocytogenes in cottage cheese through heterologous production of enterocin A..., Liu [/bib_ref]. The titer was expressed as arbitrary units (AU/ml). One arbitrary unit (AU) was defined as the reciprocal of the highest dilution showing a clear zone of inhibition to the indicator strain. When a clear inhibition zone was followed by a turbid one, the critical dilution was taken to be the average of the final two dilutions.
Minimal inhibitory concentrations (MICs) and Minimum bactericidal concentrations (MBCs) assays were determined using the microtiter broth dilution method [bib_ref] Expression of plectasin in Pichia pastoris and its characterization as a new..., Zhang [/bib_ref]. Ampicillin was also tested with the same concentration gradient as a positive control. All tests were performed in triplicate.
## In-vitro killing curve assay
To evaluate the antibacterial activity of rEntA against L. ivanovii ATCC19119, a time-kill assay was performed as described by the methods of Mao [bib_ref] Design, expression, and characterization of a novel targeted plectasin against methicillin-resistant Staphylococcus..., Mao [/bib_ref]. In addition, tubes with only bacterial inoculum were used as growth controls. All experiments were performed in triplicate.
Effects of pH, temperature, proteases and NaCl on the activity of rEntA
The effects of pH, temperature and proteases on rEntA activity were determined as described previously [bib_ref] Expression of plectasin in Pichia pastoris and its characterization as a new..., Zhang [/bib_ref] [bib_ref] Characterization and structure analysis of a novel bacteriocin, lacticin Z, produced by..., Iwatani [/bib_ref] with the following modifications: 1) The titer of purified rEntA used in the test was 12,800 AU/ml; 2) The initial activity of sample in the buffer with a pH of 6 was taken as 100% activity for pH and thermal stability assays; and 3) The residual antimicrobial activity of samples was tested after the pH was readjusted to 6.0 with sodium phosphate buffer (pH 6.0) for the proteolytic sensibility assay.
To evaluate the effect of NaCl concentration on the activity of rEntA, an overnight culture of L. ivanovii ATCC19119 was diluted to 10 5-6 CFU/ml in fresh MHB medium (3% FBS). Ten microliters of purified rEntA and 10 μl of NaCl solution were added to 80 μl of diluted cell culture. The final rEntA concentration was 4 × MIC, and the final NaCl concentrations were 0, 25, 50, 100, 200, and 400 mM. Samples without rEntA were used as controls. All samples were incubated at 37°C for 10 h. The CFU of tested strains was determined. All tests were performed in triplicate.
[fig] Figure 1: Construction of the expression plasmid pPICZαA-EntA. A, The nucleotide sequence of EntA and its corresponding amino acid sequence. The upper line indicates the wild-type EntA gene sequence. The middle line is the codon-optimized EntA gene sequence. Optimized codons are underlined with boldface type. The lower line represents the amino acid sequence of EntA. The termination codon is marked by an asterisk. B, Map of the recombinant plasmid pPICZαA-EntA. C, Electrophoretic analysis of the recombinant vector containing the EntA gene. Lane 1, DNA marker; lane 2, pPICZαA-EntA digested by XhoI and XbaI. [/fig]
[fig] Figure 2: Expression and purification of rEntA. A, Total secreted protein level and antimicrobial titer of the fermentation supernatants of recombinant P. pastoris at the shake-flask level (bars represent the standard error of the mean). B, Antimicrobial activity of the fermentation supernatants of recombinant P. pastoris at the fermenter level. 1-9, 50 μl supernatant taken at 0, 12, 24, 36, 48, 60, 72, 84, and 90 h of induction, respectively; 10, 1 μg ampicillin. C, The total secreted protein level and antimicrobial titer in the fermenter level (bars represent the standard error of the mean). D, Tricine-SDS-PAGE analysis of rEntA secreted in the fermentation supernatant of P. pastoris cultures at the fermenter level. Lane M, 5 μl molecular mass standards (from top to bottom: 40, 25, 15, 10, 4.6 and 1.7 kDa); Lanes 1-9, 20 μl supernatant taken at 0, 12, 24, 36, 48, 60, 72, 84 and 90 h of induction, respectively. E, MALDI-TOF map of rEntA. F, Purification and identification of rEntA. Lane 1, purified rEntA (0.1 μg); Lane M, 5 μl molecular mass standards (from top to bottom: 40, 25, 15, 10, 4.6 and 1.7 kDa). Lane 2, 10 μl of rEntA supernatant taken at 24 h of induction. [/fig]
[fig] Figure 3: Time-kill curves of rEntA. L. ivanovii ATCC19119 was incubated in the presence of medium alone or in the presence of 1×, 2×, or 4× MIC of rEntA. Ampicillin of 2 × MIC was used as a positive control. Three duplicate observations were made; bars represent the standard error of the mean. supernatant reached 180 mg/l with the activity of 51,200 AU/ml at 24 h of induction in 5-L fermenter level ( [/fig]
[fig] Figure 5: Effect of NaCl concentration on the activity of rEntA. Control: L. ivanovii ATCC19119 was incubated in the absence of rEntA. 4 × MIC: L. ivanovii ATCC19119 was incubated in the presence of rEntA at 4 × MIC. [/fig]
[table] Table 1: Antimicrobial spectrum of rEntA [/table]
[table] Table 2: Strains used in the antimicrobial activity assays Pseudomonas aeruginosa CVCC 2087 CVCC Salmonella enteritidis CVCC3377 CVCC Note: a China Center of Industrial Culture Collection, b China General Microbiological Culture Collection, c China Veterinary Culture Collection, d China Center for Medical Culture Collection. [/table]
|
Depression Predicts All-Cause Mortality
OBJECTIVEdDepression affects up to 20-25% of adults with type 2 diabetes and may increase all-cause mortality, but few well-designed studies have examined the effects of depression on the full range of cardiovascular disease outcomes in type 2 diabetes.RESEARCH DESIGN AND METHODSdA total of 2,053 participants in the ACCORD (Action to Control Cardiovascular Risk in Diabetes) Health-Related Quality of Life substudy completed the Patient Health Questionnaire (PHQ)-9 measure of depression symptoms at baseline and 12, 36, and 48 months. Cox proportional hazards regression models were used to estimate hazard ratios (HRs) (95% CI) for the time-varying impact of depression on protocoldefined clinical outcomes with and without adjustment for demographic, trial-related, clinical, and behavioral variables.
RESULTSdIn fully adjusted models, depression was not significantly related to the ACCORD primary composite outcome (cardiovascular death, nonfatal heart attack, or stroke) (HR 1.53 [95% CI 0.85-2.73]) or to the ACCORD microvascular composite outcome (0.93 [0.53-1.62]), but all-cause mortality was significantly increased both in those with PHQ-assessed probable major depression (2.24 .06]) and PHQ score of $10 (1.84 . The effect of depression on all-cause mortality was not related to previous cardiovascular events or to assignment to intensive or standard glycemia control. Probable major depression (by PHQ-9) had a borderline impact on the ACCORD macrovascular end point (1.42 [0.99-2.04]).
CONCLUSIONSdDepression increases the risk of all-cause mortality and may increase the risk of macrovascular events among adults with type 2 diabetes at high risk for cardiovascular events.
Diabetes Care 35:1708-1715, 2012 P atients with diabetes are approximately twice as likely to meet DSM-IV criteria for major depression than the general medical population, with depressive symptoms affecting up to 20-25% of these patients [bib_ref] The prevalence of comorbid depression in adults with diabetes: a metaanalysis, Anderson [/bib_ref]. Patients with diabetes and depression have younger age of diabetes onset; poor adherence to diet, exercise, and disease control medications; poorer glycemic control; and an increased risk of macrovascular and microvascular complications [bib_ref] Relationship of depression and diabetes self-care, medication adherence, and preventive care, Lin [/bib_ref] [bib_ref] Association of depression and diabetes complications: a meta-analysis, De Groot [/bib_ref]. Six prospective epidemiologic studies have shown that after controlling for sociodemographic factors and clinical severity of illness, comorbid depression in patients with diabetes compared with diabetes alone was associated with a 33-52% increased risk of all-cause mortality [bib_ref] Depressive symptoms and mortality among persons with and without diabetes, Zhang [/bib_ref] [bib_ref] The association of comorbid depression with mortality in patients with type 2..., Katon [/bib_ref] [bib_ref] Depression predicts increased incidence of adverse health outcomes in older Mexican Americans..., Black [/bib_ref] [bib_ref] Depression and all-cause and coronary heart disease mortality among adults with and..., Egede [/bib_ref] [bib_ref] Depression and diabetes: a potentially lethal combination, Katon [/bib_ref]. One recent study of .4,000 patients with diabetes found that probable major depression assessed by Patient Health Questionnaire (PHQ)-9 was associated with a more than twofold risk of noncancer and non-atherosclerotic associated mortality [bib_ref] Depression and increased mortality in diabetes: unexpected causes of death, Lin [/bib_ref]. However, few studies have examined specific effects of depression on macrovascular and microvascular complications. Studies have also not examined the moderating role of preexisting cardiovascular disease (CVD) or intensity of glucose control [bib_ref] Increased mortality risk in women with depression and diabetes mellitus, Pan [/bib_ref].
The ACCORD (Action to Control Cardiovascular Risk in Diabetes) trial offers a unique opportunity to examine the relationship between depression, mortality, and cardiovascular events in a sample receiving standardized diabetes care. The ACCORD trial also includes rigorous criteria for defining macrovascular and microvascular complications and cause of death. It also allowed us to examine how the effect of depression on mortality and CVD differs between those with and without previous CVD and examine whether any effect of depression on mortality or CVD is modified by randomization to intensive versus standard glucose control. Specific hypotheses from the ACCORD Health-Related Quality of Life (HRQL) substudy examined in this epidemiological analysis are as follows [bib_ref] ACCORD Study Group. Health-related quality of life and cost-effectiveness components of the..., Sullivan [/bib_ref]. After controlling for baseline factors, depression will be associated with the following: 1) primary composite outcome (death from cardiovascular causes, nonfatal myocardial infarction, and nonfatal stroke); 2) all-cause mortality; 3) composite macrovascular outcome (primary outcome plus any revascularization plus hospitalization for congestive heart failure [CHF]); and 4) composite microvascular outcome (fatal or nonfatal renal failure or retinal photocoagulation or vitrectomy for diabetic retinopathy). We further hypothesize that these associations will not be affected by preexisting CVD or by randomization to intensive versus standard glucose control. We have previously reported that randomization to intensive glycemic control did not lead to benefits in HRQL in ACCORD but was associated with modest improvement in diabetes treatment satisfaction [bib_ref] Action to Control Cardiovascular Risk in Diabetes (ACCORD) Investigators. Effect of intensive..., Anderson [/bib_ref]. trial have been described elsewhere and the randomized trial results published [bib_ref] Depression predicts increased incidence of adverse health outcomes in older Mexican Americans..., Black [/bib_ref] [bib_ref] Depression and all-cause and coronary heart disease mortality among adults with and..., Egede [/bib_ref]. In brief, this was a multicenter randomized controlled treatment trial testing independent effects of two strategies of control of blood glucose, blood pressure, and lipids on CVD in patients with type 2 diabetes. The glycemia trial randomized 10,251 participants with type 2 diabetes to intensive (goal HbA 1c ,6%) or standard (goal HbA 1c 7.0-7.9%) glucose control. All participants were also randomized within either the blood pressure or the lipid trial arms, resulting in assignment to one of eight treatment cells as follows: 1) intensive glucose/intensive blood pressure, 2) intensive glucose/standard blood pressure, 3) standard glucose/ intensive blood pressure, 4) standard glucose/standard blood pressure, 5) intensive glucose/fibrate, 6) intensive glucose/ placebo, 7) standard glucose/fibrate, and 8) standard glucose/placebo.
The primary outcome of the ACCORD trial is a composite of death from CVD, nonfatal myocardial infarction, or nonfatal stroke. Secondary outcomes include the following: 1) all-cause mortality, 2) composite macrovascular outcome (major coronary artery disease events, specifically fatal events, nonfatal myocardial infarction, and unstable angina), and 3) composite microvascular outcome (fatal or nonfatal renal failure, retinal photocoagulation, or vitrectomy for diabetic retinopathy). These outcomes are not exclusive of each other.
The goal of the ACCORD HRQL investigation is to assess the overall effect of the ACCORD interventions from the patient's point of view in 2,053 participants randomly sampled from the eight ACCORD treatment groups. Measurements were taken at baseline and at 12, 36, and 48 months. Mean follow-up time was 4.67 6 1.45 years.
Because of the documented relationship between depression and cardiovascular events and glycemic control (3-7), depressive symptoms were measured in the ACCORD study using the nine-item Patient Health Questionnaire (PHQ-9). The PHQ-9 is the self-report version of the Primary Care Evaluation of Mental Disorders (PRIME-MD), a well-validated psychiatric diagnostic interview for use in primary care settings [bib_ref] Depression predicts increased incidence of adverse health outcomes in older Mexican Americans..., Black [/bib_ref]. The PHQ-9 depression measure provides diagnostic and severity information and is used serially to assess responsiveness to depression treatment. Since the PHQ-9 items mirror those of the major depression and minor depression diagnostic criteria in the DSM-IV, it is also possible to derive provisional diagnostic categories from PHQ-9 responses. Major depression requires five symptoms scored $2, at least one of which is depressed mood or lack of pleasure. A score of $10 on the PHQ-9 has been shown to have 77% sensitivity and 94% specificity to the diagnosis of major depression by structured psychiatric interview [bib_ref] The PHQ-9: validity of a brief depression severity measure, Kroenke [/bib_ref]. In patients with type 2 diabetes, a PHQ-9 score of $10 has been associated with higher risk of mortality and dementia, as well as macrovascular and microvascular complications [bib_ref] Depression and diabetes: unhealthy bedfellows, Katon [/bib_ref]. A recent review of the reliability and validity of depression screening tools in patients with diabetes gave the PHQ-9 generally high rates of sensitivity (66-100%) but lower rates of specificity (52-85%) [bib_ref] Screening tools used for measuring depression among people with Type 1 and..., Roy [/bib_ref]. Minor depression is listed as a provisional diagnosis for further study in DSM-IV and requires three to four symptoms scored $2, at least one of which is depressed mood or lack of pleasure.
# Statistical methods
For each of the four depression measures (PHQ-9 $10, probable major depression, probable minor depression, and continuous PHQ-9 score), we ran a series of separate proportional hazards regressions models where the outcome is the time until first occurrence of each event and the predictor of interest is the measure of depression. In each model, depression was a dichotomous, time-varying indicator measured at baseline, year 1, year 3, and year 4. We report rates at which patients met any depression criteria at any time point ("ever depressed") versus not meeting any depression criteria at any time point ("never depressed"). The first series of models contained the depression measure as well as variables indicating the randomization assignments for the main glycemia trial, with stratification for the blood pressure and lipid trial arms, and primary versus secondary prevention status. As specified in the ACCORD protocol, these models also included adjustments for the following baseline factors: demographics (age, sex, race/ ethnicity, BMI, weight, waist circumference, and duration of diabetes), blood pressure (systolic and diastolic), laboratory values (triglycerides, LDL and HDL cholesterol, serum creatinine, HbA 1c , and fasting glucose), presence of microvascular complications, and blood pressure and lipid medications. A second series of models contained all of the above adjustments, plus factors known to be related to both depression and mortality in patients with diabetes: education, smoking, alcohol, and living alone. We also divided PHQ-9-assessed depression symptoms into somatic and psychological subsets and assessed whether each was related to all-cause mortality. All models used Cox proportional hazards regression model analyses to obtain hazard ratios (HRs) (95% CI) of the measure of depression. These analyses combined participants from the randomized groups and are thus epidemiologic in nature. There were no adjustments made for postrandomization events or measures. Participants with missing data were omitted from the models, resulting in~4% loss in the most complex models. Since we conducted 12 statistical tests of hypotheses related to secondary end points and subgroups, there was a 46% chance (i.e., 1 2 [1 2 0.05] [bib_ref] Increased mortality risk in women with depression and diabetes mellitus, Pan [/bib_ref] that at least one of these tests would be statistically significant at an a level of 0.05, assuming independence between tests.
RESULTSdTable 1 compares the demographic and clinical characteristics of ACCORD HRQL substudy participants compared with the full ACCORD sample. As expected, because of randomization, there are no significant differences between the groups. displays the rates at which ACCORD HRQL participants screened positive for depression (PHQ score $10) or met criteria for probable major depression or minor depression based on responses to the PHQ. Nearly 20% of participants screened positive for clinically significant depressive symptoms (PHQ-9 $10) at baseline, with 8% meeting DSM-IV probable major depression criteria and another 7% meeting DSM-IV probable minor depression criteria. Approximately 31% of participants screened positive for depression at one of the four assessments, with 15% meeting major depression and 18% meeting minor depression criteria at least at one assessment. displays the baseline characteristics of the subjects based on whether they ever met any of the depression criteria. As would be expected from the epidemiology of depression, ever depressed patients were more likely to be younger, female, less educated, cigarette smokers, obese, and have higher HbA 1c , higher pulse, lower HDL cholesterol, and higher total cholesterol and triglycerides. They were also more likely to be treated with insulin or sulfonylureas.
The primary composite outcome was reached by 2.1% per year (6,000 personyears observed) in the never-depressed group and 1.9% per year (3,239 personyears observed) in the ever depressed group. All-cause mortality was 1.4% per year in both the never-depressed (6,313 person-years observed) and everdepressed (3,438 person-years observed) groups. The macrovascular composite outcome was reached by 5.2% per year (5,552 person-years observed) in the never-depressed group and 5.9% per year (2,908 person-years observed) in the ever-depressed group. The microvascular composite outcome was achieved by 2.6% per year (5,573 person-years observed) in the never-depressed group and 3.2% per year (2,999 person-years observed) in the ever-depressed group. Outcome rates are similar between the ever-versus neverdepressed groups, but these results do not account for other clinical differences between these groups. displays the HRs for different measures of PHQ-assessed depression as a time-dependent covariate for time to the ACCORD primary and secondary end points. The first set of models was adjusted for the demographic, trial, and clinical variables described as follows. 1) Probable major depression did not significantly increase the risk for the primary composite outcome of cardiovascular mortality, nonfatal myocardial infarction, or nonfatal stroke. The point estimate for major depression is HR 1.53, but the effect is not statistically significant. Neither PHQ score $10 nor the PHQ continuous score was significant. 2) Both PHQ score of $10 (HR 1.84 [95% CI 1.17-2.89]) and PHQ-assessed probable major depression (2.24 .06]) increased risk for the secondary outcome of all-cause mortality after controlling for age, sex, race/ethnicity, primary/secondary CVD prevention, HbA 1c , lipids, blood pressure, BMI, smoking, alcohol consumption, living alone, blood pressure, presence of microvascular complications, CHF, education, duration of diabetes, antidepressant medications, glucose, blood pressure and lipid medications, and assignment to one of eight study intervention arms. The continuous PHQ-9 score was also a significant predictor of all-cause mortality (1.05 [1.01-1.09]). For each point increase on the PHQ-9, all-cause mortality increased by 5%. The increase in absolute risk for all-cause mortality associated with probable major depression at any time point is estimated to be 0.92%. were adjusted for demographic, trial, clinical, and behavioral variables. The additional adjustment for these potentially confounding or mediating variables slightly attenuated the risk associated with depression. Effects of depression on all-cause mortality remained significant, though the marginal effects of depression on the macrovascular outcome became nonsignificant.
In order to verify that the effects of probable major depression on mortality and the primary outcome were similar in both glycemia arms, we fit a model for each outcome controlling for the randomization factors and including an interaction term between glycemia arm and major depression. We fit the same set of models to see if the effects were the same for participants with prior CVD at baseline. None of the interaction terms of interest in the four models were significant, indicating that effect of probable major depression was consistent across the groups. While some studies of patients post-myocardial infarction have found that only the somatic symptoms of depression (e.g., fatigue, insomnia) are associated with subsequent mortality, we found that both psychological and somatic symptoms of depression as assessed by the PHQ-9 were significantly (and nearly equally) associated with allcause mortality [bib_ref] The association of cognitive and somatic depressive symptoms with depression recognition and..., Smolderen [/bib_ref].
CONCLUSIONSdThis epidemiological analysis of data from the ACCORD trial revealed that depression, defined as PHQ-assessed probable major depression, PHQ score of $10, or continuous PHQ-9 score, was associated with increased risk of all-cause mortality regardless of whether previous CVD was present and regardless of randomization to intensive versus standard glycemia control. Depression marginally increased the risk of the combined macrovascular outcome. Depression was not significantly associated with the primary composite ACCORD outcome or the secondary composite microvascular outcome. In ACCORD, probable major depression was associated with an approximate twofold risk of all-cause mortality. This is somewhat higher than in previous studies and might be due to the increased capability in ACCORD to control for other clinical variables or due to the fact that study subjects were selected for high risk for cardiovascular events. It is notable that this risk was observed in the context of good glucose control in all groups, since poor adherence and poor glucose control have been proposed to account for the effect of depression on mortality in patients with diabetes [bib_ref] Depression and diabetes: a potentially lethal combination, Katon [/bib_ref].
The point prevalence of major depression in primary care patients is between 5 and 10% [bib_ref] Epidemiology of depression in primary care, Katon [/bib_ref] , whereas prevalence rates of major depression in patients with diabetes and coronary heart disease (CHD) have been estimated to be 12-18% (21) and 15-23% [bib_ref] The nature and course of depression following myocardial infarction, Schleifer [/bib_ref] [bib_ref] Depression following myocardial infarction: first-ever versus ongoing and recurrent episodes, Spijkerman [/bib_ref] , respectively. The relationship between major depression and diabetes and/or heart disease appears to be bidirectional. A recent meta-analysis of 13 studies found that the pooled relative risk for depressed patients subsequently developing diabetes was 1.60 (95% CI 1.37-1.88) [bib_ref] Depression and type 2 diabetes over the lifespan: a meta-analysis, Mezuk [/bib_ref]. This meta-analysis also found modest evidence that diabetes was a risk factor for subsequent major depression (relative risk 1.15 [1.02-1.30]) [bib_ref] Depression and type 2 diabetes over the lifespan: a meta-analysis, Mezuk [/bib_ref]. Major depression following myocardial infarction is also very common, occurring in up to 25% of patients [bib_ref] The nature and course of depression following myocardial infarction, Schleifer [/bib_ref] [bib_ref] Depression following myocardial infarction: first-ever versus ongoing and recurrent episodes, Spijkerman [/bib_ref]. Recent data suggest that approximately one-half of patients who developed major depression post-myocardial infarction had recurrent depressive episodes and that half had their first depressive episode postmyocardial infarction [bib_ref] Depression following myocardial infarction: first-ever versus ongoing and recurrent episodes, Spijkerman [/bib_ref].
In the general population, both diabetes and depression increase mortality. They exert a greater than additive effect when both are present. A large prospective study of an aging Hispanic population found that lifetime major depression was associated with a 1.64 (95% CI 1.17-2.28) and diabetes a 1.51 (1.23-1.86) HR for all-cause mortality, respectively, compared with those without history of depression or diabetes [bib_ref] Depression predicts increased incidence of adverse health outcomes in older Mexican Americans..., Black [/bib_ref]. When combined, depression and diabetes had an HR of 4.59 (2.12-9.93) of all-cause mortality compared with control subjects without history of diabetes or depression [bib_ref] Depression predicts increased incidence of adverse health outcomes in older Mexican Americans..., Black [/bib_ref]. Another study that followed .10,000 participants for 8 years found that subjects with significant depressive symptoms (Center for Epidemiologic Studies Depression Scale $16) but no diabetes had a 1.20 (1.03-1.40) increase in all-cause mortality, those with diabetes but no depression had a 1.88 (1.55-2.27) increase, and those with both depression and diabetes had a 2.50 (2.04-3.08) increase in all-cause mortality compared with subjects without depression or diabetes [bib_ref] Depression and all-cause and coronary heart disease mortality among adults with and..., Egede [/bib_ref].
Recent prospective studies have also examined the association of depression with subsequent development of macrovascular and microvascular complications in patients with diabetes. A 5-year prospective study of .4,000 diabetic patients found that comorbid probable major depression on PHQ-9 was associated with a 24% increased risk of macrovascular complications and a 36% increased risk of microvascular complications [bib_ref] Depression and diabetes: a potentially lethal combination, Katon [/bib_ref]. ACCORD showed a 42% increase in risk of macrovascular complications, but this was of borderline significance, possibly due to the smaller sample. In ACCORD, there was not a significant effect of depression on microvascular complications. It is possible that some previous studies may have overestimated Our study adds to the above studies in a number of ways. First, we were able to control for a wide range of baseline clinical characteristics related to cardiovascular risk including age, sex, race/ ethnicity, CHD status, HbA 1c , lipids, systolic and diastolic blood pressure, BMI, presence of microvascular complications, CHF status, and duration of diabetes. We were also able to control for a number of social and behavioral factors that might confound the depression effect, including smoking, alcohol consumption, living alone, and education. Second, all subjects had reasonably well-controlled glucose, blood pressure, and lipids. Third, we were able to examine whether the depression effect on cardiovascular events and mortality differed by intensive versus standard use of glucose, blood pressure and lipid medications, clinical center network, and assignment to one of eight study intervention arms. None of these factors interacted significantly with depression status. Several factors must be considered in interpreting the results of our study. First, study participants may have had less baseline depression and different subsequent depression trajectories based on the requirement to volunteer for the study and to provide informed consent at enrollment. Second, this analysis used a screening instrument (PHQ-9) to detect depression characterized by high sensitivity but low specificity. Up to one-half of patients with diabetes who score $10 on the PHQ-9 may not have major depression on structured interview. Patients with only minor depression tend to do as well with placebo as they do with antidepressant treatment. Third, this study does not consider depression treatment in detail, although the use of PHQ-9 scores obtained at four set time points may reflect the adequacy of depression treatment. Fourth, the composite microvascular end point we evaluated was comprised of advanced complications and we did not assess the impact of depression on early onset of microvascular complications, such as on new-onset microalbuminuria. Fifth, while we controlled for severity and duration of cardiometabolic disease, we did not control for all medical comorbidity. ACCORD did exclude patients with significant kidney or liver disease and cancer and those expected to live ,3 years.
Despite these limitations, the results of this study highlight the importance of depression detection and effective depression treatment as key elements in quality diabetes care. Patients with diabetes and PHQ-9 scores of $10 randomized to collaborative depression and diabetes treatment showed greater improvements in depression, functioning, and quality of life than those randomized to usual care [bib_ref] Collaborative care for patients with depression and chronic illnesses, Katon [/bib_ref]. However, no study has shown that depression treatment reduces mortality in patients with diabetes. Moreover, a recent study showed that the older tricyclic antidepressants may increase CVD with long-term use [bib_ref] Antidepressant medication use and future risk of cardiovascular disease: the Scottish Health..., Hamer [/bib_ref]. This effect was not found for the more commonly used selective serotonin reuptake inhibitor antidepressants.
The results we report here, in conjunction with data from other studies, support the need for a randomized controlled trial to assess the impact of depression care on mortality in adults with type 2 diabetes.
[table] Table 1dDemographic: and clinical characteristics of ACCORD HRQL substudy participants compared with full ACCORD sample [/table]
[table] Table 4dProportional: hazard models of depression predicting ACCORD outcomes Macrovascular composite outcome (major coronary artery disease events, specifically fatal events, nonfatal MI, and unstable angina) Microvascular composite outcome (fatal or nonfatal renal failure, retinal photocoagulation, or vitrectomy for diabetic retinopathy) [/table]
|
Evolution of echovirus 11 in a chronically infected immunodeficient patient
Deep sequencing was used to determine complete nucleotide sequences of echovirus 11 (EV11) strains isolated from a chronically infected patient with CVID as well as from cases of acute enterovirus infection. Phylogenetic analysis showed that EV11 strains that circulated in Israel in 1980-90s could be divided into four clades. EV11 strains isolated from a chronically infected individual belonged to one of the four clades and over a period of 4 years accumulated mutations at a relatively constant rate. Extrapolation of mutations accumulation curve into the past suggested that the individual was infected with circulating EV11 in the first half of 1990s. Genomic regions coding for individual viral proteins did not appear to be under strong selective pressure except for protease 3C that was remarkably conserved. This may suggest its important role in maintaining persistent infection.
# Introduction
Enteroviruses widely circulate in human populations and only rarely cause clinical disease. Polioviruses were the first enteroviruses to be isolated and the first to have the pattern of changes in their genome during chains of transmission characterized [bib_ref] Calibration of multiple poliovirus molecular clocks covering an extended evolutionary range, Jorba [/bib_ref] [bib_ref] Enteroviruses: Polioviruses, Coxsackieviruses, Echoviruses, and newer Enteroviruses, Pallansch [/bib_ref]. They cause acute flaccid paralysis of limbs and occasionally bulbar paralysis. Paralysis rate in naïve individuals, however, is around one clinical case per 100 to 1000 infections. Other enteroviruses can also cause paralysis but at a lower rate. Enteroviruses cause a variety of other clinical conditions including mild fever, hand-foot-and-mouth disease, herpangina, myocarditis, diabetes, etc. Human echovirus 11 is a member of the human enterovirus B species and is one of the most commonly isolated enteroviruses [bib_ref] Enterovirus surveillance-United States, Khetsuriani [/bib_ref]. EV11 viral capsid protein 1 (VP1) sequences segregate into six or more genogroups [bib_ref] Molecular epidemiology and typespecific detection of echovirus 11 isolates from the Americas, Oberste [/bib_ref] [bib_ref] Molecular comparison of echovirus 11 strains circulating in Europe during an epidemic..., Chevaliez [/bib_ref]. Data from the Annual Reports of the Central Virology Laboratories, Public Health Service of the Israeli Ministry of Health, Tel Hashomer Israel, indicate that EV11 is endemic in Israel with at least one EV11-positive case being reported on 17 of the 27 years between 1986 and 2012. There was a major peak of 90 EV11-positive stools in 1999 and smaller peaks of in Enteroviruses are also able to chronically infect individuals with several kinds of primary immunodeficiency [bib_ref] Enteroviral infections in primary immunodeficiency (PID): a survey of morbidity and mortality, Halliday [/bib_ref] [bib_ref] Genomic variations in echovirus 30 persistent isolates recovered from a chronically infected..., Bailly [/bib_ref] , and persist for months or years being regularly excreted in stool. This phenomenon is best known for poliovirus, which belongs to species C of human enteroviruses. Chronic poliovirus infection can last for over 30 years [bib_ref] Twenty-Eight Years of Poliovirus Replication in an Immunodeficient Individual: Impact on the..., Dunn [/bib_ref]. At least 110 immune deficient individuals have been identified who had persistent infections with immunodeficiencyassociated vaccine-derived polioviruses (iVDPVs) [bib_ref] Modeling the prevalence of immunodeficiencyassociated long-term vaccine-derived poliovirus excretors and the potential..., Tebbens [/bib_ref] [bib_ref] Immunodeficiency-related vaccinederived poliovirus (iVDPV) cases: a systematic review and implications for polio..., Guo [/bib_ref]. Additional cases of persistent infection from unidentified individuals have been inferred from environmental surveillance and isolation of ambiguous highly evolved vaccine-derived polioviruses (aVDPVs) [bib_ref] Characteristics of an environmentally monitored prolonged type 2 vaccine derived poliovirus shedding..., Hovi [/bib_ref]. Circulating vaccine-derived polioviruses (cVDPV) represent another kind of VDPV. They have regained virulence and the ability to transmit in human populations indistinguishable from wild polioviruses, and can serve as a model for natural evolution of enteroviruses. VDPVs can cause outbreaks of paralytic disease in susceptible immune competent individuals. The pattern of amino acid substitutions and recombination of iVDPVs and aVDPVs has been shown to differ from cVDPV [bib_ref] Circulating vaccine-derived polioviruses: current state of knowledge, Kew [/bib_ref].
Among the EV11 isolates from Israel were 10 referred to hereafter as iEV11s that were obtained from the CSF of an immune deficient individual with Common Variable Immunodeficiency (CVID) who suffered from symptoms of chronic encephalomyelitis between 1995 and 1999. VP1 sequences of these isolates were obtained using Sanger-based sequencing. These sequences were compared with VP1 sequences EV11s isolated from cerebrospinal fluid, stool, rectal and throat swabs of contemporary cases of EV11 infections of non-immune deficient patients (2 in1992, 3 in1993, 1 each in . A comparison of the patterns of differences in the VP1s of the two groups indicated that the iEV11s represented a persistent infection rather than repeated re-infections from contemporary EV11s circulating in the community.
Here we report the use of deep sequencing to determine complete genomic sequences of the 10 iEV11 isolates and 16 contemporary cEV11s to confirm that the iEV11s represented a persistent infection and to obtain information on the pattern of its evolution in humans during persistent infections of immune deficient individuals. To our knowledge this is the first analysis of evolution of HEV B during a persistent infection using complete genome sequences.
# Materials and methods
# Ethics statement
The Ethical Review Board of the Sheba Medical Center, Tel Hashomer approved this study (SMC-7685). The clinical samples from which viruses were isolated are a part of the virus strain collection of the Central Virology Laboratory in Israel. They were received between 1986 and 2014 and results were stripped of all links to personal details pertaining to, or which could be used to identify individual patients. All data were analyzed anonymously. The Ethical Review Board exempted this study from a requirement for obtaining informed consent.
## Sources of viruses
The annual incidence of laboratory confirmed EV11-positive stools received by the Central Virology Laboratory in Israel between 1986 and 2014 is presented in S1 there were 212 individuals with EV11-positive clinical samples. Among them were eleven EV11 isolates obtained between November 13, 1995 and December 21, 1999 from a male patient. These isolates are referred to as iEV11s to indicate that they were isolated form an immune deficient individual. The patient, born in 1955, was diagnosed with CVID at age 20 after recurrent bouts of pneumonia, sinusitis, and ear infections along with episodes of aseptic arthritis, as well as leukopenia and alopecia areata advancing to alopecia universalis. His younger brother had died from a bacterial complication known to be associated with CVID. The patient was treated with intravenous immunoglobulins (IViG) between 1975 and 1995, and prophylactically with antibiotics between 1985 and 1995. Following a year or more of poor adherence to IViG treatment, decreasing the infusions to once in 2 months, he presented with decreased hearing, a rash and neurological symptoms of headaches, stroke-like episodes and proximal lower limb weakness. He was diagnosed at the department of Neurology as suffering from chronic enteroviral meningoencephalomyelitis due to EV-11 that did not respond to reintroduction of frequent (bi-monthly) IViG infusions. In 1998 the patient received two one-week courses of daily Pleconaril (400 mg/day), under expanded access in February and then again in October. This treatment did not affect the course of his disease and his lower limb weakness gradually deteriorated along with mild cognitive difficulties. In December 2012 there was sudden worsening of his condition and he lapsed into a coma associated with acute hydrocephalus that did not respond to a ventriculoperitoneal shunt and he succumbed to his illness.Representative EV11 isolates collected prior to and during the isolation of these iEV11 samples that were sent for clinical analysis were chosen for comparison: eight from sporadic cases presenting between 1992 and 1998 and eight collected during a large EV11 outbreak in 1999. Clinical samples and symptoms associated with each isolate are indicated in S1 Virus from the first tissue culture passage (TC1) that had been kept at -20˚C were passaged once (TC2) on human human kidney epithelial cells (HuKi) received from Dr. Furin (Department of Health and Welfare, Ottawa, Canada) and RD cells (human rhabdomyosarcoma cell line; ATCC CCL136). Virus from four samples from the CVID patient and seven from among the representative samples were no longer viable by the time of this analysis ("Original TC" in S1 .
## Nucleic acid extraction
RNA was extracted from the virus in 1 ml of supernatant from the second passage for viable EV11 isolates, or from 1 ml of the original frozen stock for non-viable EV11 using a MagNA Pure LC2.0 Automatic extractor with MagNA Pure LC Total Nucleic Acid Isolation kit-High Performance (Roche Diagnostics, IN, USA) and eluted into 50 µl of elution buffer according to manufacturer's instructions.
## Full length genomic sequences
The RNA library was prepared from the total RNA of 27 EV 11 isolates, the NEBNext mRNA Sample Prep Master Mix Set 1 (New England BioLabs, Ipswich, MA) was used according to the manufacturer's protocol (NEB). Briefly, 0.5 µg of the total RNA was used for fragmentation by focused ultrasonicator (Covaris) to generate the fragments of optimal sizes (250-300 nt) suitable for Illumina sequencing. cDNA was synthesized using SuperScript III Reverse Transcriptase (Invitrogen) and random primers. The cDNA was converted into double stranded cDNA followed by end repair procedure (Klenow fragment, T4 polynucleotide kinase and T4 polymerase), and was ligated to Illumina paired end (PE) adaptors.
Size selection was performed using double AMPure bead selection step (Beckman Coulter), generating DNA libraries ranging from 200 to 500 bp in size. Next the DNA libraries were expanded using 15 cycles of PCR with multiplex indexed primers and purified by magnetic beads (Agencourt AMPure PCR purification system, BeckmanCoulter). Finally, the DNA fragments were analyzed for quality and size distribution by BioAnalyzer using a high sensitivity kit (Agilent Technologies, Inc., Santa Clara, CA).
Deep sequencing was performed using MiSeq (Illumina) producing paired end reads of 250 nt long, or HiSeq2000 (Illumina) producing 101 nt long paired-end reads. The sequencing reads were analyzed by in-house 'swarm' or High Performance Integrated Virtual Environment (HIVE) software [bib_ref] High-Performance Integrated Virtual Environment (HIVE) Tools and Applications for Big Data Analysis, Simonyan [/bib_ref]. Raw sequence reads were subjected to quality control and reads with phred score below 30 were removed. The reads were aligned to a curated database of 500 reference enteroviruses. Next, aligned sequences were separated into discrete sub-populations [bib_ref] Separation and assembly of deep sequencing data into discrete sub-population genomes, Karagiannis [/bib_ref] and assembled into full length EV11 consensus genomic sequences (see S1 . This resulted in complete or near-complete genomic sequences of 25 isolates. For some virus isolates the depth of coverage (S1 at the ends of genome was insufficient for reliable sequence reconstruction. To assemble individual consensuses of viruses present in mixtures a new algorithm was used [bib_ref] Separation and assembly of deep sequencing data into discrete sub-population genomes, Karagiannis [/bib_ref]. The names of virus isolates were shortened to the name of EV11 genogroup followed by an incremental number by incremental order in which they were isolated.
## Molecular analyses
Neighbor-joining phylogenetic comparisons were prepared by Clustal X in MEGA v5.22or MacVector software using the Tamura-Nei substitution model [bib_ref] Estimation of the number of nucleotide substitutions in the control region of..., Tamura [/bib_ref]. Separate analyses were prepared for near-complete sequences with 5'-ends trimmed to the longest sequence common to all sequences in the alignment, the entire open reading frame, the P1 region encoding all four viral capsid proteins, Viral Capsid protein 1 (VP1), the P2 plus P3 coding region encoding all non structural proteins, the 3D RNA-primed RNA polymerase, and the 3' untranslated region (3'UTR, also partly truncated). The initial alignments of these regions were prepared for MEGA using the Sequencher v5.2.2 program (Gencodes, Anne Arbor, MI, USAS).
Three-dimensional folding models of Loop V of the internal ribosomal entry site (IRES) in the 5'UTR were prepared using the web-based Unified Nucleic Acid Folding and hybridizing Package [bib_ref] Mfold web server for nucleic acid folding and hybridization prediction, Zuker [/bib_ref] at the University of Albany NY, USA, (http://mfold.rna.albany.edu; last accessed November 2014) and the consensus sequence for the Israeli isolates. The threedimensional folding model for the human enterovirus Z domain in the 3'UTR was taken from Merkel et. al. [bib_ref] Biological significance of a human enterovirus B-specific RNA element in the 3'..., Merkle [/bib_ref].
Amino acid differences between Israeli EV11 isolates from EV11 genogroup A and the earliest isolate from this genogroup, EV11-A01, were determined using the Sequencher program. Amino acid differences that became fixed during the persistent EV11 infection in the CVID patient were mapped onto the 5 capsomers located at the 5-fold axis of symmetry on the X-ray diffraction crystalline structure of the EV11 capsid (Accession number 1H8T) using the Mac-PyMol program (The PyMOL Molecular Graphics System, v7.2.1, Schrödinger, LLC.,.
## Accession numbers
Genomic sequences were deposited in the GenBank/EMBL/DDBJ database (accession numbers KY981557 to KY981581), and the raw Illumina data in sequence read archive (https:// hive.biochemistry.gwu.edu/review/Echovirus11).
# Results
## Determination of genomic sequences using direct rna library deep sequencing approach
To determine genomic sequences of EV11 isolates (S1 we have used direct RNA library deep sequencing approach because the amount of live virus in some samples was very low, and the quantity of isolated RNA would not be sufficient for multiple PCR reactions needed for conventional Sanger sequencing. To recover RNA from these samples carrier tRNA was added. RNA was then fragmented by focused ultrasonication, converted to double-stranded cDNA and sequenced using Illumina Miseq or Hiseq technology. This resulted in a variable level of sequencing coverage in most cases ranging from 1,000 to 250,000 with the exception of four samples in which the coverage was lower (See S1 . However, even for the sample with the lowest depth of coverage (40) we were able to assemble consensus sequence that had an uninterrupted open reading frame of 2,195 amino acids. The nature of RNA library preparation is such that there is a gradual decline of the number of molecules representing the very ends of the genome. In some cases the low coverage of the terminal regions resulted in consensus sequences that missed few nucleotides from the ends because they could not be reconstructed reliably.
The use of this protocol proved to be remarkably effective and the level of sequencing coverage per each nucleotide was sufficient not only for unambiguous reconstruction of consensus sequences but also for determining sequence heterogeneity profiles.
# Phylogenetic analysis
The comparison of complete or near-complete genomes the EV11 isolates revealed that they could be separated into four clades or genogroups labeled A through D [fig_ref] Fig 1: Phylogenetic relationship between full-length sequences of EV11 strains isolated in Israel in... [/fig_ref]. Their relationship with other EV11 was determined by comparing with a representative subset of 100 published sequences of virion protein VP1 [fig_ref] Fig 2: Phylogenetic tree based on nucleotide sequences of VP1 capsid protein showing the... [/fig_ref]. On this and subsequent dendrograms the incremental number after the genogroup indicates the temporal order in which they were isolated. Attempts were made previously to classify EV11 isolates into genogroups [bib_ref] Molecular epidemiology and typespecific detection of echovirus 11 isolates from the Americas, Oberste [/bib_ref] [bib_ref] Phylogenetic analysis of complete VP1 sequences of echoviruses 11 and 6: high..., Fares [/bib_ref]. The comparison of the newly sequenced isolates showed that the closest relative to our genogroup A was 1995 Tunisian isolate (HQ674714), which had 3.4-3.7% nucleotide differences with cEV11, and 5.8-8.3% with iEV11. This means that our genogroup A matches genogroup IV-4 of Fares. Genogroup B was related to 1991 isolate from Maine (AY121384) with 4.3-4.6% difference, placing it into genogroup IV-4 of Fares and D4 of Oberste. Genogroup C was related at 11.2% to 1994 Tunisian isolate HQ674718 and at 11.1-11.5% to 1999 isolate from Kuwait (AY121408), belonging to Fares genogroup IV-3 and Oberste genogroup D3. Finally our genogroup D was very close (1.1-1.8%) to 1998 Tunisian isolate HQ674721 (Fares genogroup IV-5, Oberste genogroup A).
Enteroviruses frequently recombine with other viruses of the same species. To determine whether the Israeli isolates are also products of recent recombination we have compared them with a reference set of 72 full-length sequences of Species B Enterovirus that contained at least one representative from each serotype, and all 12 full-length genomic sequences of EV11 available in Genbank. Figs 3A, 3B and 3C show the phylogeny of P1, P2, and P3 regions providing no evidence of recombination within the group of isolates from immunodeficient patient. In contrast, analysis of circulating strains revealed potential recombination event between P1 and P2 regions of genogroup D, as evidenced by its "jump" to another branch of the phylogenetic tree.
Previous studies found about 30% divergence between different isolates of EV11 [bib_ref] Phylogenetic analysis of complete VP1 sequences of echoviruses 11 and 6: high..., Fares [/bib_ref]. Therefore close similarity between strains described in this study lends support to the notion that all the iEV11s isolated between 1995 and 1999 were sequentially related to four group A isolates from 1992 and 1993. The hypothesis that these 1992 and 1993 isolates were closely related to the iEV11s is strengthened by the observation that cEV11 of genogroup A were closest to the earliest iEV11 isolate (5.8%), and more distant (8.3%) from the latest ones. This is illustrated by neighbor-joining phylogenetic tree shown on [fig_ref] Fig 4: Gradual accumulation of mutations in iEV11 illustrated by neighbor-joining tree [/fig_ref] It suggests that iEV11 may have evolved from a cEV11 strain close to the Israeli isolates from 1992-1993. Regression lines for the accumulation of nucleotide and amino acid substitutions relative to the closest cEV111 strain shown on [fig_ref] Fig 4: Gradual accumulation of mutations in iEV11 illustrated by neighbor-joining tree [/fig_ref] can be extrapolated to the early 1990s, a tentative time when the patient became chronically infected with EV11.
The first iEV11 strain isolated in 1995 (A5) had a total of 139 nucleotide differences from the closest 1992 cEV11 isolate A1. These mutations never reverted back, and remained in all subsequent iEV11 isolates. An additional 12 "substitutions" that emerged during the same period, reverted back in subsequent iEV11 isolates. Between February 1996 and April 1998 there were 113 additional mutations that emerged and became fixed in all subsequent isolates and 8 more that emerged but then "reverted back" by October 1999. An additional 79 substitutions emerged between April 1998 and October 1999 followed by 2 more between October and December of 1999. Therefore during the four time intervals between 1995, 1996, 1998, and two dates in 1999 the rates of mutations fixation were: 0.83 Ã 10 −2 , 1.16 Ã 10 −2 , 0.71 Ã 10 −2 , and 0.68 Ã 10 −2 mutations / site / year, a rate close to the molecular clock rate of 1.03 Ã 10 −2 reported for circulating poliovirus [bib_ref] Calibration of multiple poliovirus molecular clocks covering an extended evolutionary range, Jorba [/bib_ref] , 1.51 Ã 10 −2 for iVDPV [bib_ref] Twenty-Eight Years of Poliovirus Replication in an Immunodeficient Individual: Impact on the..., Dunn [/bib_ref] , and 0.84 Ã 10 −2 for EV30 isolated from chronically infected immunodeficient patient [bib_ref] Genomic variations in echovirus 30 persistent isolates recovered from a chronically infected..., Bailly [/bib_ref].
## Accumulation of amino acid differences
Analysis of the predicted amino acid substitutions that occurred during the iEV11 evolution showed that most (28 of 35, or 80%) of the amino acid substitutions became fixed in the evolving genome (S2 . Fixation of 10 amino acid substitutions resulted from 2 nucleotide differences in the codon, one involved 3 nucleotide differences, and one fixed amino acid was represented by a single nucleotide difference followed by a second silent mutation in the six subsequent isolates. Four amino acid differences were unique, e.g., found in only one of the 10 isolates, and one substitution appeared in isolates A06 through A10, but not A11 through A14.
The distribution of amino acid changes along the iEV11 genome is shown on [fig_ref] Fig 5: Distribution of mutations in the entire open reading frame of iEV11 [/fig_ref] It revealed that the entire 3C protein was highly conserved, along with parts of 3D (between amino acids 142 and 300), 3A (amino acids 1-128 and 140-236), 2C (amino acids 2-118), and VP1 (amino acids 157-267). Hot spots of mutations included amino acids 116-144 and 267-292 of VP1, and 145-164 of VP2.
The analysis of selective nature of mutations in individual viral proteins was performed by calculating density of nucleotide and amino acid substitutions, Dn and Da, respectively. The values represent the number of nucleotide and amino acid substitutions that have been fixed in at least one isolate, divided by the size of the coding region (nt) or the protein (aa). If the ratio Da/Dn is less than 1, then there is a negative selection against amino acid changes, while the ratio greater than 1 indicates possible positive selection of certain amino acid changes. [fig_ref] Fig 6: Amino acid and nucleotide substitutions densities [/fig_ref] that amino acid substitutions in most viral proteins were largely neutral, except for proteins 3C (protease) and 3D (RNA polymerase) (p-values 0.0366 and 0.0332), and marginally significant for 3A (p-value of 0.0910) that were under negative selective pressure. Amino acid sequence conservation was the strongest for 3C protease, in which only 1 out of 183 amino acids reverted from threonine at position 15 to alanine present in cEV11 strain A01-EV11-5789. In contrast, Dn and Da values for all genes of circulating EV11 differed significantly (p<0.001), suggesting that the selective pressure at amino acid level was higher and the conservation of proteins was more uniform across the genome [fig_ref] Fig 6: Amino acid and nucleotide substitutions densities [/fig_ref].
The predicted amino acid difference among iEV11s relative to the earliest cEV11 isolate A01 from 1992 were mapped onto the three dimensional crystallographic model of the EV11 capsid (Protein Data Bank, PDB ID 1H8T). The fixed amino acid substitutions are shown for the capsid pentamer centered around the five-fold axis of symmetry [fig_ref] Fig 7: Amino acid differences mapped on the virion pentamer outer surface [/fig_ref]. Many of the substitutions were on the outer capsid surface and the walls of the canyon, a circular depression surrounding the 5-fold axis of symmetry and important for high affinity binding to host cell receptors . When only iEV11s were compared, most amino acid substitutions accumulated on the northern rim of the canyon next to the 5-fold axis of symmetry. There were 7 amino acids differences in P1 between A01-cEV11-5824 isolated in 1992 and the earliest iEV11, A05-iEV11-7482, isolated in 1995. The same was true for another cEV11 isolate from 1992 A02-cEV11-5789. The amino acids that differed were: N18S, R106C, D143G, E217D, T220A, A223V and H697C. (Notation: amino acid in cEV11 from 1992, the number of the amino acid position in P1, amino acid in iEV11 from 1995.)
## Nucleotides differences in domain v of the ires in the 5'utr
The 5'-untranslated region of Enterovirus genome contains the internal ribosome entry site (IRES) . Analysis of iEV11 sequences revealed that between 1998 and 1999 a 13-14 nucleotide deletion became fixed in the hypervariable region of the 5'-UTR. Domain V in the IRES has been shown to be critical for enterovirus replication and virulence . Stem and loop structures Arrows indicate each position where one or more of the Israeli isolates had an alterative nucleotide, and a pentagram or rectangle indicates whether the difference would disrupt or conserve the stem structure, respectively. In the first model one pair was conserved and three disrupted, while in the second model, four pairs were conserved and three disrupted. The overall shape of domain V predicted by the second model is similar to the shape of domain V of polioviruses.
## Nucleotide differences in z domain of the 3'utr
A cloverleaf structure at the 5' end of the 3'-UTR serves as the specific signal for a number of functions including switching from transcription to replication . In HEV B viruses there is Evolution of echovirus 11 in a chronically infected immunodeficient patient an addition Z domain that is well conserved among HEV B viruses and may be important for viral growth in vivo, but not in vitro [bib_ref] Biological significance of a human enterovirus B-specific RNA element in the 3'..., Merkle [/bib_ref]. The resulting structure [fig_ref] Fig 9: Predicted secondary structures of Z-domain in the 3'-UTR for HEV-B and Coxsackie... [/fig_ref] was different than that proposed previously [bib_ref] Biological significance of a human enterovirus B-specific RNA element in the 3'..., Merkle [/bib_ref] for the HEV B consensus motif when nucleotide sequences specific for the iEV11 were mapped into this structure. Nucleotide differences found in one isolate, A13, would have further disrupted this structure in two positions and elongated a stem in another.
# Discussion
Normally enteroviruses cause acute infections that are cleared within few days or weeks. However, in rare cases, they can establish persistent infection lasting for years. Most known cases of Evolution of echovirus 11 in a chronically infected immunodeficient patient chronic infection with enteroviruses involve patients with various kinds of primary immunodeficiency affecting B cell functions. However many aspects of this phenomenon remain poorly understood, including the cell and tissue localization, the factors controlling the outcome of virus-host interactions, and the reasons why the infection cannot be resolved. This current study was initiated in attempts to understand the evolution of the virus that persists in a patient chronically infected with EV11 who experienced periodic bouts of encephalitis that eventually led to death. This investigation was conducted within the context of patterns of nucleotide and amino acid differences observed among EV11 viruses isolated from sporadic and clustered EV11 infections in the community during the same time frame. The clinical course of this case differed from persistent poliovirus infections in immune deficient people who are largely asymptomatic. Our goal was to identify the source of infection and to shed some light on the environment the virus replicates in and the selective pressures it is experiencing. We have studied 10 iEV11 strains isolated over a period of almost 4 years from November 1995 till December 1999. To identify the source of the infection we have also sequenced strains of circulating EV11 isolated in Israel between 1992 and 1999. The material that was available to us included primary isolates and a cell culture passage performed in attempt to increase virus concentration. Virus in some samples lost its viability and could not be expanded in cell culture, and therefore the quantity of RNA that was available was insufficient for traditional Sanger sequencing of the entire genome. Therefore we have tried another approach based on deep sequencing using Illumina technology, which proved to be very effective and resulted in reconstruction of complete or near-complete consensus sequences and identification of sequence heterogeneities. This report demonstrates that deep sequencing methodology requires smaller amounts of starting material and is sequence-independent, i.e. does not require specific PCR primers used in traditional Sanger sequencing. Comparison of the 25 genomic sequences generated in this study showed that four different genogroups circulated in Israel in the 1990s and that two of them co-circulated during the 1999 EV11 outbreak. Genogroup A viruses were isolated in , and genogroup D only in 1999. Co-circulation of two clusters of EV11 (C and D) in 1999 has been reported previously [bib_ref] Molecular epidemiology and typespecific detection of echovirus 11 isolates from the Americas, Oberste [/bib_ref].
All iEV11 strains isolated from the chronically infected patient clustered together with four cEV11 isolates of genogroup A. The virus that was the closest relative of the earliest (1995) iEV11 isolate was cEV11 strain isolated in 1992. They had 24 nucleotide differences in VP1-coding sequence (2.7%). Subsequent isolates gradually accumulated mutations so that the latest 1999 iEV11 isolate contained 51 differences (5.8%). The pace of this evolution (3.1% over a period of 3.8 years) is consistent with the rate of mutants accumulation established for circulating polioviruses . The time plot of the number of nucleotide and amino acid differences in VP1 region was a straight line that could be extrapolated back to the early 1990s, consistent with the isolation date of the closest relative, and suggesting that the patient was infected at around that time with a contemporary cEV11 strain. Comparison of full-length sequences revealed no evidence of recombination with other viruses, and other parts of the viral genome also evolved linearly, albeit at a different pace.
Amino acid substitutions that occurred in cEV11 capsid differed from those that accumulated in iEV11. In circulating viruses mutations mostly occurred in the canyon and its southern rim (across from the 5-fold axis of symmetry), while in iEV11 many of them concentrated on the northern rim adjacent to the 5-fold axis. The canyon is the binding site for high affinity receptors that belong to the immunoglobulin superfamily, such as CD155, which is poliovirus receptor . In contrast, EV11 and some other enteroviruses use Decay-Accelerating Factor (DAF, or CD55). DAF binds close to the 5-fold axis of symmetry on the surface and not in the canyon . Interaction with DAF is low affinity with high on-off rates compared with canyon binding receptors capable of causing conformational changes of bound viruses . For instance, EV11-DAF interaction has a dissociation constant KD = 3µM [40] whereas poliovirus-CD155 has KD = 80-700µM . EV11 infection of polarized epithelial cells uses two different entry routes, directly from apical surface as well as tight junctions, which is dependent on DAF binding . Preferential accumulation of mutations on the northern wall of the canyon and close to 5'-fold axis of symmetry during evolution of iEV11 suggests that the virus may be undergoing adaptation to DAF in specific cell types. This part of the virion surface is also known to harbor neutralizing epitopes, so the changes may also reflect immune evasion by the virus.
EV11 were the most sensitive of all enteroviruses to antiviral drug pleconaril . Approximately two years after the onset of chronic infection the immunodeficient patient was treated with pleconaril in an attempt to clear the viral infection. Sequencing results revealed that the strain isolated from the patient at the time contained two mutations in VP1 (V 117 I and V 119 I) associated with EV11 resistance to pleconaril [44], explaining why the treatment attempt was unsuccessful.
Highly conserved RNA secondary structures rather than sequences are important for replication . Changes of nucleotides in essential structures such as domain V of the 5'UTR and the Z motif in the 3'UTR were observed for the iEV11s. Some nucleotide variations between isolates conserved the stem loop structures while others altered them.
The rate of accumulation of total nucleotide substitutions (Kt) in capsid regions of iEV11 and replicating poliovirus were similar [bib_ref] Calibration of multiple poliovirus molecular clocks covering an extended evolutionary range, Jorba [/bib_ref] (0.76 Ã 10−2 and 1.03 Ã 10−2 mutations per site per year, respectively). However, the degree of amino acid sequences conservation was much lower in iEV11 than in poliovirus (Ka values of 0.72 Ã 10−2 and 0.03 Ã 10−2, respectively), suggesting that evolution of iEV11 was taking place under lower selective pressure. Respective values for cEV11 could not be calculated because of the scarcity of data and the unknown evolutionary trajectory of the few known sequences of cEV11. However, this conclusion was confirmed by comparing rates at which nucleotide and amino acid changes are fixed in different parts of viral genome during evolution. There are a total of 61 nucleotide triplets coding for 20 amino acids, therefore if the mutational process is random and there is no selective pressure restricting changes in the protein sequence, there should be 3 times more fixed nucleotide substitutions than fixed amino acid changes. Therefore comparison of densities of mutations in amino acid and nucleotide sequences (Da and Dn, the number of sites at which mutations are present divided by the length of protein and nucleotide sequences respectively) can reveal whether there is a positive or negative selective pressure. The Da/Dn ratio below 1 indicates that protein sequences are conserved, and that a disproportionately bigger number of mutations result in synonymous codons. A ratio greater than 1 may suggest that there is a positive selective pressure favoring accumulation of adaptive mutations. This analysis performed for iEV11 strains showed that in most viral proteins the ratio was close to 1, suggesting that they largely experience a random neutral evolution. The exception was viral 3C protease that was highly conserved and had only one mutation that changed isoleucine at amino acid 6 that is unique for genogroup A strains to methionine present in genogroups B, C, and D. In contrast, the pattern of selective pressure in cEV11 was different, and these viruses experienced higher selective pressure at amino acid level (the average Da/Dn ratio was 0.79 for iEV11 versus 0.22 for cEV11). Lower conservation of capsid proteins and different distribution of mutations on the capsid surface in iEV11 compared to cEV11 suggest that the virus may undergo adaptation to replication in different types of cells. Protease 3C that is highly conserved in iEV11 plays an important role in maturation of viral proteins as well as in cleaving several cellular proteins involved in innate antiviral defense . Its conservation in iEV11 isolates may suggest that this function is important for viral persistence. Therefore establishment of chronic EV11 infection involves not only impairment of host immune system (B-cell immunodeficiency), bus also may require virus adaptation while preserving its functions mediating virus-host interactions.
Supporting information S1 [fig_ref] Fig 1: Phylogenetic relationship between full-length sequences of EV11 strains isolated in Israel in... [/fig_ref]
[fig] Fig 1: Phylogenetic relationship between full-length sequences of EV11 strains isolated in Israel in 1992-1999. iEV11 strains belonging to genogroup A are shown in red, cEV11 of genogroup A are shown in orange, while genogroups B, C, D in purple, blue, and green, respectively. The tree was constructed using unweighted pair group method with arithmetic mean (UPGMA). Bootstrap values (N = 1000) are indicated near the internal nodes of the tree. https://doi.org/10.1371/journal.ppat.1006943.g001 Evolution of echovirus 11 in a chronically infected immunodeficient patient PLOS Pathogens | https://doi.org/10.1371/journal.ppat.1006943 March 19, 2018 [/fig]
[fig] Fig 2: Phylogenetic tree based on nucleotide sequences of VP1 capsid protein showing the relationship between sequences determined in this study and all published EV11 sequences. Color codes are the same as in Fig 1. The tree was constructed using UPGMA algorithm. https://doi.org/10.1371/journal.ppat.1006943.g002 Evolution of echovirus 11 in a chronically infected immunodeficient patient PLOS Pathogens | https://doi.org/10.1371/journal.ppat.1006943 March 19, 2018 [/fig]
[fig] Fig 3: Phylogenetic trees constructed based on P1, P2, and P3 regions of EV11 genomes determined in this study along with reference HEV-B sequences. The trees were constructed using neighbor-joining method of Saitou and Nei [46]. https://doi.org/10.1371/journal.ppat.1006943.g003 Evolution of echovirus 11 in a chronically infected immunodeficient patient PLOS Pathogens | https://doi.org/10.1371/journal.ppat.1006943 March 19, 2018 [/fig]
[fig] Fig 4: Gradual accumulation of mutations in iEV11 illustrated by neighbor-joining tree (A) and plots of the number of nucleotide and amino acid substitutions vs. time of isolation (B). The analysis was performed for the entire genome of these strains relative to the closest cEV11 isolate. Phylogenetic tree was constructed with neighbor-joining method and bootstrap values (N = 1000) are shown near the internal nodes of the tree. [/fig]
[fig] Fig 5: Distribution of mutations in the entire open reading frame of iEV11. Mutations were identified by comparing with the closest contemporary cEV11 isolate A01-cEV11-5789. The top line represents mutations that may have occurred prior to the infection of the patient, and all subsequent lines illustrate virus evolution in the patient. https://doi.org/10.1371/journal.ppat.1006943.g005 Evolution of echovirus 11 in a chronically infected immunodeficient patient PLOS Pathogens | https://doi.org/10.1371/journal.ppat.1006943 March 19, 2018were predicted for RNA sequences of the all Israeli EV11s equivalent to domain V of poliovirus using the m-Fold program. Two highest probability stem-loop structures are shown inFig 8. [/fig]
[fig] Fig 6: Amino acid and nucleotide substitutions densities (Da and Dn respectively) for individual parts of EV11 genome for iEV11 (A) and cEV11 (B). Error bars show the 95% confidence limits for binomial proportions. https://doi.org/10.1371/journal.ppat.1006943.g006 [/fig]
[fig] Fig 7: Amino acid differences mapped on the virion pentamer outer surface. Differences in iEV11 are shown in red, cEV11 in blue, and those that occurred in both are shown in black. Differences determined relative to isolate A01-EV11-5789 were mapped onto the three dimensional crystallographic model of the EV11 capsid (PDB ID: 1H8T).https://doi.org/10.1371/journal.ppat.1006943.g007 [/fig]
[fig] Fig 8: Predicted secondary structure of domain V of EV11 IRES element. Two best models are presented on panels A and B. Threedimensional folding models of Loop V of the internal ribosomal entry site (IRES) in the 5'UTR were prepared using the web-based Unified Nucleic Acid Folding and hybridizing Package[17,18] at the University of Albany NY, USA, (http://mfold.rna.albany.edu; last accessed November 2014) and the consensus sequence for the Israeli isolates.https://doi.org/10.1371/journal.ppat.1006943.g008Evolution of echovirus 11 in a chronically infected immunodeficient patient [/fig]
[fig] Fig 9: Predicted secondary structures of Z-domain in the 3'-UTR for HEV-B and Coxsackie B3 [24]and iEV11-5789 determined in this paper. The Israeli sequences were superimposed on the three-dimensional folding model for the human enterovirus Z domain in the 3'UTR taken from Merkel et. al.[24].https://doi.org/10.1371/journal.ppat.1006943.g009 [/fig]
|
The European Cystic Fibrosis Society Patient Registry: valuable lessons learned on how to sustain a disease registry
Background: Disease registries have the invaluable potential to provide an insight into the natural history of the disease under investigation, to provide useful information (e.g. through health indicators) for planning health care services and to identify suitable groups of patients for clinical trials enrolment. However, the establishment and maintenance of disease registries is a burdensome initiative from economical and organisational points of view and experience sharing on registries management is important to avoid waste of resources. The aim of this paper is to discuss the problems embedded in the institution and management of an international disease registry to warn against common mistakes that can derail the best of intentions: we share the experience of the European Cystic Fibrosis Society Patient Registry, which collects data on almost 30,000 patients from 23 countries. Methods: We discuss the major problems that researchers often encounter in the creation and management of disease registries: definition of the aims the registry has to reach, definition of the criteria for patients referral to the registry, definition of the information to record, set up of a data quality process, handling of missing data, maintenance of data confidentiality, regulation of data use and dissemination of research results.Results:We give examples on how many crucial aspects were solved by the European Cystic Fibrosis Society Patient Registry regarding objectives, inclusion criteria and variables definition, data management, data quality controls, missing data handling, confidentiality maintenance, data use and results dissemination. Conclusions: We suggest an extensive literature research and discussions in working groups with different stake holders, including patient representatives, on the objectives, inclusion criteria and the information to record. We propose to pilot the recording of few variables and test the applicability of their definition first. The use of a shared electronic platform for data collection that automatically computes derived variables, and automatically performs basic data quality controls is a good data management practice, that also helps in reducing missing data. We found crucial for success the collaboration with existing national and international registries, cystic fibrosis organisations and patients' associations.
# Background
A disease registry is a paper list or an electronic database containing information on the characteristics of a population affected by a given disease [bib_ref] A short history of pathology registries, with emphasis on cancer registries, Terracini [/bib_ref]. Disease registries have become essential for the investigation of chronic diseases thanks to their potential to epidemiologically describe the natural history of the disease. They are particularly useful in rare diseases, such as cystic fibrosis (CF), where important research questions cannot be answered without large multicentre studies because of the limited number of patients followed by individual CF centres.
The importance of disease registries has been acknowledged also by EUCERD, an EU Committee of experts in rare diseases that discusses policies and recommends activities in collaboration with the EU Commission and Parliament and the Council of Ministers. Many international projects, such as EPIRAREand RD-CONNECT, promoting international registries have also been funded by the EU.
The institution and long term management of disease registries are not trivial challenges, since many key hurdles need to be overcome. To avoid the garbage in, garbage out phenomenon, the first step requires the definition of patients' inclusion/exclusion criteria: it is vital to determine whether the registry records only confirmed cases, based on a set of pre-defined criteria, or whether it is open to all subjects with a suggestive set of symptoms or signs. The clinical and epidemiological questions the registry has to address and the set up of the information system depend on these criteria. The difficulty in addressing most of the hurdles is magnified in an international setting, where agreements on data definition are crucial to ensure uniformity of data collection across the participating countries.
Many registries begin on a voluntary and unfunded basis, being championed by single enthusiasts, but an effective disease registry requires sustained funding, adequate manpower and an efficient organisational structure to achieve its main purpose: to describe the clinical status of patients to foster care improvement [bib_ref] Using the national registry of HIV-infected veterans in research: lessons for the..., Rabeneck [/bib_ref] [bib_ref] Using disease registries for pharmacoepidemiological research: a case study of data from..., Strobl [/bib_ref].
The aim of this paper is to discuss the problems embedded in the institution and management of an international disease registry, to warn against common mistakes that can derail the best of intentions. The experience gained in the establishment and maintenance of the European Cystic Fibrosis Society Patient Registry (ECFSPR), that collects data from individual CF centres and national CF registries from Europe and the neighbouring countries, can be used to show the hidden dangers of disease registries. In this paper we provide examples on how several crucial aspects were solved by the ECFSPR in the areas of objectives, inclusion criteria and variables definition, data management, data quality controls, missing data handling, confidentiality maintenance, data use and results dissemination.
Prior to the establishment of the ECFSPR, a pan-European registry originating from a database set up to monitor a clinical trial on a respiratory medicine was used to collect data from approximately 10,000 patients in 9 countries. This registry was funded by F. Hoffmann-La Roche Ltd, and was named Epidemiologic Registry of Cystic Fibrosis [bib_ref] European Epidemiologic Registry of Cystic Fibrosis (ERCF): comparison of major disease manifestations..., Koch [/bib_ref]. With the termination of funding in 2002, it was decided by CF specialists across Europe, members of the European CF Society (ECFS), to set up an independent registry with clear objectives. At that time several national registries already existed, each built on separate proprietary platforms, whereas the majority of European countries did not have a CF registry in place, or had registries based on one-centre only. In 2003, under the auspices of ECFS, a working group of representatives of existing national registries was appointed to set up this new registry. A pilot study collected data from seven national registries (Belgium, Denmark, France, Ireland, Italy, Russia and Sweden), using a simple electronic spreadsheet. This provided a precious starting point for data collection, but also showed that different definitions were separately developed by each national registry, and even if specifications on data formats and coding were given, these were not always followed by data contributors. This aspect, along with the need for a tool for data entry for countries without a CF registry, led to the re-evaluation of the data collection system. A registry steering group was created, in charge of setting up the structure of the ECFSPR, defining its milestones and its roadmap.
The first turning point came with the EU funded project EuroCareCF, that in cooperation with the ECFSPR steering group set up patient consent forms and collected demographic data from 35 countries, thereby laying the foundation for expansion of the ECFSPR outside the existing national registries [bib_ref] Comparative demographics of the European cystic fibrosis population: a cross-sectional database analysis, Mccormick [/bib_ref]. The second and probably most important turning point came with the support from the ECFS Board who decided to expand the financial support allowing appointment of professional staff (coordinator, helpdesk, statisticians) and lately the development of a bespoke software building on the experience of previous data collection and software.
# Methods
In this section we introduce the major problems that researchers often encounter in the creation and management of an international disease registry. Many critical aspects that we had to face in the activation of this international registry can be found also in the implementation of national/local disease registries and they can be faced in a similar way.
## Definition of the objectives and of the population under study
A registry is not an end in itself; it is rather a tool to reach predetermined objectives, such as enhancement of knowledge of the epidemiology of the disease under study, evaluation of different diagnostic and therapeutic protocols, evaluation and planning of public health resources, creation of health indicators to evaluate the disease burden and the efficacy of the care.
The use of registry data is effective and efficient only if there is a clear definition of the objectives the registry has to reach and a tracking of the outcomes is carefully planned. Stepping beyond these agreed objectives could be problematic especially during the set up phase, because the community's expectations might become unachievable in the time frame of the available funding. In this phase, estimation of the observation time needed to obtain the results is a key factor, especially for rare diseases, due to the risk of having to wait for years before an adequate number of patients or outcomes are recorded. In a later phase, when core functions are running smoothly, objectives can be re-evaluated. Objectives should also be shareable by the patients whose data are collected in order to obtain their consent to process their data.
Ideally, a disease registry should only contain data on patients diagnosed with the disease of interest. This apparently trivial aspect is the core of data collection, because it determines the uniformity of population under study and it explicitly defines the data to be recorded. For this reason, it is crucial to set up an unequivocal definition of case, which rests upon a general agreement on the diagnostic criteria. These criteria may be subject to change over time: often, the change of diagnostic practices due to advances in knowledge of the disease and advances in scientific technology change the inclusion criteria to registry referral. The key aspect is to keep an audit trail of these changes.
## Definition of what to measure and how to do it
In the planning of a disease registry, it is necessary to decide which aspects of the disease are to be recorded and identify their appropriate indicators. Enthusiasm often leads to substantial overestimation of the amount of information necessary to record: it is easy to fall into the mistake of recording far too much information than what is actually needed to answer the questions the registry has to address. This is a highly inefficient approach in terms of time needed to retrieve the data and to register them into the database, because ultimately the information will not be used. The amount of information recorded is a trade-off between the researchers' needs and the requirement to keep the registry easy to handle. It is therefore advisable in the planning phase to clearly define objectives and to agree on which information has to be recorded. These tasks, if well conducted, make data collection more efficient and avoid frustration in people who enter the data if the information is not used and frustration in statisticians who analyse empty databases. For some rare diseases, the information to be recorded may be difficult to target because the disease is not yet fully understood. It is therefore advisable in the planning phase to select information to be recorded that is agreed in literature, being aware that new information could be added in the future.
Definitions for registries have to reflect what is obtainable during daily clinical work at various centres, and in case of international registries, across various countries. Strict definitions based on well-defined clinical and paraclinical aspects, as are often implemented in clinical trials, are often not applicable in the daily clinic, making compromises necessary. Unless the data collected are useful to the clinician or other ways of data collector reward (including economic repayment) many registries can fail at this step.
## Data management and data quality controls
To reach its aims, an international patient registry has to set up an efficient data management system, preferably automated, that accommodates both national registries' and individual centres' needs, coordinating the work of database managers, statisticians, epidemiologists, and covering many aspects, such as data collection, data quality controls, error correction, data analyses and reporting. A limitation to smooth data streaming for an international patient registry is that national data registries (or individual centres) often change their data acquisition systems and any subsequent data sending to the international registry requires an agreement between all parties involved to prevent rupture of the data stream or unwelcome changes to it.
Data quality control is one of the core data management activities of a disease registry and it is probably the most important aspect because the quality of the data, together with efficiency of data management, inevitably affects the quality of research. For this reason, it is vital for a registry to have accurate data quality controls and efficient data processing systems in place.
## Handling of missing data
Missing data are the bane of researcher's lives because they can reduce the precision of the estimates and may lead to biased results. Missing data are a widespread problem: in many registries, although there is an adequate completeness level in the demographic data, there is questionable completeness for clinical follow-up data [bib_ref] The how (and why) of disease registers, Mehta [/bib_ref]. Although there are statistical techniques to address missing data, it is always preferable to prevent information loss, therefore it is essential to understand the mechanisms that cause it and avoid missing data as much as possible.
## Maintaining confidentiality of registry data
The binding element to legally collect data is to obtain the patient consent to process their data. In order to reassure the patients to increase their willingness to give their consent, it is crucial to ensure that patient's identification and data security measures are state-of-the-art, particularly when dealing with electronic data, that such measures meet the data protection legislation, and that they are described and available to the patients. The collection of data into national registries and their potential export into an international registry must be approved by the data protection authority in the country of residence of the patient. An international registry data collection must itself be approved by the data protection authority of the country of the data controller. The key aspect is that the data sender must be legally allowed to send the data and the central registry must be legally allowed to process the data. The use of data must be described in data protection applications and in the patient consent information sheets; and it is important that the information given is detailed, but it also has to accommodate practical changes during the course of the registry life: revised applications for the data protection authorities in case of e.g. a change of data processor can be easily made, but obtaining the patient consents from all patients within a short time interval might be impossible. The registry needs to keep up to date with changing data legislation, nationally and internationally; this is an important challenge across Europe as new regulations are about to be enacted.
## Dissemination of the results and use of data
Timely dissemination of results and appropriate use of data collected are the key elements for achieving the registry primary objectives: enhance knowledge of the disease under study and promote research. The objectives of a disease registry also allow use of data for identifying patients eligible for clinical trials, e.g. patients with a rare genotype. These patients are of course not to be contacted directly by the company conducting the trial, but via their care giver, so anonymity is withheld until the patient consents to participate in the trial.
# Results
The results section presents the solutions to the hurdles described in the methods section, as implemented by the ECFSPR. The critical aspects and the solutions are summarised in [fig_ref] Table 1: Overview of critical aspects when setting up a registry and the solutions... [/fig_ref].
## Definition of the objectives and of the population under study definition of the objectives
During the set up meetings of the first working group of the ECFSPR, the objectives of the registry were thoroughly discussed and agreed, as stated in the ECFSPR website and in patient consent: "The purpose of the ECFS Patient Registry is to measure, survey and compare aspects of cystic fibrosis and its treatment in the participating countries, thereby encouraging new standards of dealing with the disease, to provide data for epidemiological research and to identify special patient groups suitable for multi-centre trials. The information will facilitate long-term planning of health expenditure allocations and developing pan European support systems". When defining the objectives it is important to include all stake-holders in the process. For a patient registry relying on the patients acceptance of their data being collected, we found it crucial to involve patient representatives very early in the project; for this reason we cooperated closely with CF Europe, the European CF patients' organisation, who appointed two of their associates as members of the steering group, one of whom is also member of the executive committee.
## Definition of the population under study
Due to the international nature of the ECFSPR, the first concern was to ensure that the registry collected data from patients meeting uniform inclusion criteria across countries. A working group was set up with the aim of defining the inclusion and exclusion criteria for CF patient referral to the ECFSPR. This was a small group, whose tasks included extensive literature research, retrieval of necessary information from CF registries representatives and database managers, and harmonisation of criteria.
The diagnostic criteria for CF are not internationally agreed and, often, their verification is not strictly performed in clinical practice, mainly due to costs. For example, sodium and chloride concentrations in an agreed sweat test protocol are considered gold standards for the diagnosis of CF [bib_ref] A test for concentration of electrolytes in sweat in cystic fibrosis of..., Gibson [/bib_ref] , but a quicker method of estimating sweat chloride (conductance) was introduced, even if it has never been recognized as equivalent to the concentration measurements [bib_ref] Clinical evaluation of the macroduct sweat collecting system and conductivity analyser in..., Hammond [/bib_ref]. However, this method is widely used as a screening tool and may be the only sweat test performed on a patient. Moreover, improvements in DNA analysis resulted in the diagnosis of both patients on whom sweat test was never performed (e.g. because two known disease causing mutations were found by DNA analysis) and patients with clinically milder forms of CF or CF-like syndromes, opening the debate in the international scientific community about the definition of CF [bib_ref] Who is reported in the Belgian, Dutch and French CF registries?, Thomas [/bib_ref].
The ECFSPR therefore adopted an operational definition that could be used as inclusion criteria for the registry purposes. The fundamental aspect in fact is that the population under study is uniform in terms of inclusion criteria, even if the official clinical definition might not be met, because this does not have an impact on research and appropriate subgroups can be selected for further study.
Ideally, the ECFSPR should record all the information necessary to assess whether the patients registered meet the inclusion criteria or not. However, the level of detail of information needed resulted in a lot of missing data from the pilot data collection, especially from the national registries, that often did not record such detailed information. The verification of the inclusion criteria was therefore delegated to the data contributors, who had the opportunity to check the clinical notes in a more efficient and timely way, and the ECFSPR data collection requires each participating country/centre to confirm, for each patient referred to the ECFSPR, that the inclusion criteria are met. The data senders therefore take responsibility on the appropriateness of referral, If the definition used is not the same across countries:
- try harmonisation by making the definition more generic
- involve stakeholders to discuss change of definitions and agree on a shared definition
- if definitions can be assimilated, report differences of definitions in the publications as caveats
## Data management and data quality controls
Data management Shared electronic platform for data collection with automatic computation of derived variables, allowing both direct data entry and remote data upload.
Use of technology (such as XML) that ensures that required data format and coding is used.
Data quality controls Automatic and immediate data quality controls on entering (plausible ranges, intra-record data coherence, and consistent information across years.)
Use of drop-down menus with fixed input possibilities (e.g. yes/no/unknown) Agreed controls with national registries in order to avoid duplication of identical data quality control processes.
Use of refined data controls based on age-and-sex-specific reference values Set up of a data error procedure that uses a software that automatically warns and points the user to the data to correct
Handling of missing data User-friendly software and useful feedback to contributors to encourage data entry
## Definition of what to measure and how to do it
Definition of what to measure
In the ECFSPR experience, the task of defining variables was particularly complex and time-consuming, due to the international nature of the registry and the fact that many countries had already established national registries using internally-chosen definitions. A small group of people was put in charge of revising the literature about the definition of critical variables and discussing the implementation and the adaptation of clinical definitions. The intention of this working group was to collect a limited set of key variables that would make it possible to achieve the ECFSPR objectives. We first undertook the pilot data collection from the existing national registries to estimate the feasibility of collection of such variables. The participants were asked to send information for 49 variables. The proportion of missing data varied according to the type of information requested: basic demographic data such as gender, year and month of birth were reported on all patients, whereas more detailed social and clinical information such as marital status or use of continuous inhaled antibiotics was missing for almost all patients. This pattern was mostly due to some national registries not recording information on the variables required by the ECFSPR. Based on this collection, the definitions group put forward a revised set of variables for the collection of data for years 2004/2005.
## Definition on how to measure
The ECFSPR could not always follow the golden rule according to which the choice of indicators should fall on those easy to observe and with unanimous definition. Often the best indicator was not the easiest one to observe: in this case, it was not included in the registry, due to the high risk of having a lot of missing data, obtunding the efforts of data collection. Particularly useful was the approach to have in the database few variables and test with a pilot data collection the applicability of their definition. Also, if the best indicator has different interpretations in the scientific community, there is the risk of putting under the same roof different quantities, making interpretation of results impossible. Some case examples are:
Medication: asking quantitative information on the dose of pancreatic enzymes taken by CF patients is clinically useful and it would be highly informative, but the intake of this medication is often variable on a meal to meal basis and estimating a "daily intake value" (or a "yearly intake value") would be extremely difficult. The definitions working group felt it was appropriate to replace this information with a more generic "use of pancreatic enzyme ever in the follow-up year"; Complications: an important prognostic factor in CF is chronic infection with Pseudomonas aeruginosa. Several definitions of chronic infection have been used in daily clinical care as well as in publications. The most commonly used is the Leeds definition [bib_ref] Evaluation of a new definition for chronic Pseudomonas aeruginosa in cystic fibrosis..., Lee [/bib_ref] or a modification of this. However, this definition requires several sputum cultures per year, and patients who were diagnosed with chronic infections for a long time and on continuous inhalation therapy may not fulfil the strict criteria, although they would still be classified as chronically infected by their care givers. Therefore, the definition on diagnosis of chronic infections for the ECFSPR purposes had to be an operational definitionthat allowed discrimination only between patients with chronic infection from those patients without infection and/or with intermittent infection.
Collecting about 85% of its data through national registries, which have already collected data from the CF centres following the national registry definitions, not necessarily the same as those of the ECFSPR, is another challenge. Whenever possible, we used a definition that would comply with most national registries. However, some of them could not be harmonised and for those the national registries have chosen either to report this variable as missing (e.g. one registry records only "Pseudomonas aeruginosa cultured this year" with no chronicity defined), or to change their definition to comply with the ECFSPR one. For example, several countries are now collecting the best value of FEV1 of the year, as required by the ECFSPR, instead of last of the year or the value registered at the annual assessment. If definitions are not the same across countries, but can be assimilated, this is reported in methods sections of manuscripts and written in notes appended to tables and graphs.
Data management and data quality controls Data management schematically shows how the ECFSPR data collection is organised. In order to provide a common platform for data collection, bespoke software was developed. It is composed of two tools: one for data upload from national registries and one for manual data-entry for countries that do not have a national database. In both cases, the data are stored on a central server, located in secure premises. The current data-entry software, which has been used for 3 years of data collection, has some limitations and new software is currently being rolled out across Europe, although the data flow structure at the core of its functioning has not changed.
Acquisition of data The data upload from national registries is performed through the Extensible Mark-up Language (XML), a strategy adopted by the ECFSPR in 2010 to overcome the inefficiency in the coding of data from national registries. In fact, this choice of language forces the variables formats and specifications to be met, saving a lot of data management time to the ECFSPR Data Management Unit (DMU). Once the national registries have prepared the data extract from their database according to the agreed definitions and coding, they upload the XML file they created, and a two-step system validates the file according to a procedure described in the next section. The use of the XML procedure has as big limitation the need of an IT-specialist to create the XML file from the data extract made by the national registry data manager. In the ECFSPR experience, this aspect was a bigger obstacle than anticipated, both practically and psychologically. This made us rethink the upload process in the development of the new data-entry software, focusing on user-friendliness and ease of use without compromising on coding requirements. The new data-entry software allows uploading data files in different formats in addition to XML (such as commaseparated values, and common formats that are easily originated from the national databases) and controls on data coding and value ranges are performed on these files.
The manual data-entry is performed as follows: the software sends the patients' data, except the identifiers, to a web server. The data are anonymised through the creation of a randomly-generated code. The identifiers, like name and full date of birth, are stored in an encrypted form on a server, but only the centre holds the key to decrypt the data. The rest of the database is also stored in an encrypted format and it is password-protected. Some data-quality controls are automatically performed by the software, which warns the user by means of flagged fields whenever discrepancies are found and when items are left blank. Further details on automatic data quality controls are described in the next section. One limitation of the current software is that the non-anonymous data are stored at the local hospital computer. This caused significant problems with installation and upgrading working with doctors and IT-technicians from many centres in many countries. In the new software, the identifying data are stored encrypted on the central server or on a national server, but only the centre holds the de-cryption key to view the data, and they are not accessible in any way by the ECFSPR. In case the user loses the identifying data, they will have to re-enter them whereas the rest of the data will always be accessible.
Raw data and derived variables Good data management practices impose that derived variables, such as BMI, are centrally computed by the DMU. The main reason for this is that, should an error in the computation occur, it would be easily traceable and recoverable. In order to minimise data-entry errors, it was decided to collect raw data (to which automatic plausibility checks are carried out) and delegate the computation task to Flow-chart of data collection and data quality controls of European Cystic Fibrosis Society Patient Registry.
the central DMU or to the software. This approach has the further advantage that the raw data can be transformed into standardised values (such as FEV1% of predicted, or height/weight standard deviation scores) according to different reference equations, as needed.
## Data quality controls
Basic data quality controls should initially check that entered values lie within plausible ranges. Intra-record data coherence must be checked, such as chronological sequence of dates (e.g. the date of diagnosis has to follow the date of birth unless pre-natal diagnosis was performed) and consistent information (e.g. if values for a test are present, then the test must be recorded as "performed") also across years (e.g. patients reported as liver transplanted one year should still be reported as transplanted throughout all subsequent years).
The ECFSPR has implemented a data quality control procedure composed of two levels: one carried out by the software for data entry and another carried out by the DMU.
First level of data quality controls (performed by the software) For the manual data entry, correct coding is obtained by dropdown menus with fixed input possibilities (e.g. yes/no/unknown). The only field allowing free text is the mutation field: with more than 1900 known CF mutations, only the most frequent ones are included in a drop-down menu, with the possibility to manually enter new ones as free text. For numeric values, the format must be correct (e.g. integer, decimal, date), and the value should be within a pre-set range or it will not be accepted. Furthermore, values outside certain agespecific ranges will be flagged as possible errors, but will be accepted. For example, any value of height within the range 35-250 cm is accepted by the software, but a one-year old boy recorded as 100 cm high would be outside the age-specific range of 67-79 cm. His height value would therefore be accepted by the software but it will be flagged as a probable error to the user.
For the users who send the data through file upload we originally implemented the same controls, with upload denied for patients with unacceptable values and an error report of the problems. This could result in national registries not being able to finish uploading the full data set until errors had been corrected and re-uploaded. For this reason, we subsequently adopted another approach, described in the next section.
Second level of data quality controls (DMU) The second step of data quality control procedure is identical for the two means of data sending and it is carried out after the annual data collection is closed. The DMU carries out more refined data plausibility and data coherence controls, for example, by using reference values for height and weight to detect potential errors by means of standard deviation scores, or by comparing values across years (e.g. decreasing values for height). The errors found are then uploaded onto the ECFSPR server as a file that points directly to the patient and the erroneous values with a short explanatory text. When opening the software, the users are led straight to the error and can either correct the value or confirm it; with a free text field available to send messages of explanation to the DMU.
When inconsistent data are found, data contributors should receive notification to revise data and send corrections within a pre-agreed short time frame. It is important that such notification is performed as quickly as possible. Mehta [bib_ref] The how (and why) of disease registers, Mehta [/bib_ref] reports that two weeks is a practical time limit by which centres should receive notification, because within this time people in charge of data-entry effectively remember the clinical notes and are able to retrieve the necessary information to correct the data. After this time, correcting the errors becomes less timely and efficient. This process becomes even more burdensome when requests of clarifications do not go directly to the centres that entered the data, but to intermediate data-management units (such as national registries).
For the centres manually entering the data, automatic data checks help avoid entry of wrong data, and since the patient's file is usually available during data entry, the user can quickly find the correct value. In the new software we have added even further automated controls, diminishing the need for further corrections by the DMU.
For the national registries, however, the two-level data control turned out to be very inefficient. For a lot of the corrections requested, they need to contact the individual centres in their countrymaybe even twice if errors were found in both levels of data control. This process could be very long and inefficient particularly if the national registries have already performed their own data quality controls and frozen the database for the analysis when the ECFSPR data collection starts. This is a specific issue for the ECFSPR, since we collect data with a delay of three years at the moment. The delay is primarily to allow the national registries to perform their data cleaning process and produce their own report before sending data to the ECFSPR. We aim to minimize the delay to one year. In order to optimize the error correction process a data quality control group, composed of national registry data managers and the ECFSPR team, has agreed on a common and exhaustive list of internal data checks to be carried out on the national databases by the national registries during their internal data cleaning process before data upload to the ECFSPR.
The national registries can now upload their checked data set without interruption and will just receive a list of any remaining questionable values.
After all the errors have been corrected by the data contributors, the data are frozen for analyses. Potential errors found by ECFSPR but not confirmed/corrected by the user, for the purposes of the annual data report, are deemed as erroneous and are set to missing.
## Handling of missing data
The amount of missing values has an important role on the interpretation of the results: if the data are missing for a non negligible portion of the database, then the estimates can be very imprecise or even biased if the underlying missing data mechanism is related to some specific values. For example, if 10% of the patients are reported to use insulin, but 20% of patients have unknown/missing information on use of insulin, the true percentage of use of insulin can be anything between 10 and 30%. For the annual ECFSPR data report we always report the number of missing data to illustrate the problem. The long experience of the ECFSPR team reveals that there are several reasons why missing data occur:
1. The lack of protected time, motivation and funding by a dedicated trained person in charge of data retrieval and data-entry is the biggest reason for the occurrence of missing data. Information not recorded at recruitment is rarely retrieved afterwards: data already entered are only occasionally revised by the CF centres, and their modification is performed only upon explicit request from the ECFSPR DMU. For example, some CF centres that send the data to the ECFSPR did not enter the information on the vital status of the patient (deceased/alive); when the ECFSPR DMU asked the centres if really the vital status of their patients was unknown, the centres answered that they did not have the time to enter this information, but that all the patients they reported were alive. The type of information requested and the time when the revision is requested are two critical points in the data correction process. There might be the need to consult the original clinical records and, sometimes, this can be problematic, especially if the time lapse is long. One way to reduce the workload for data entry is to automatically extract the information necessary for the registry from the computerised case report system that is used by a the CF centre. This has the advantage that if the data entered are used for clinical purposes, the centre operators are motivated to enter good-quality data for their own use. This solution was not applicable to the ECFSPR, due to the big heterogeneity of CF centres' IT systems. With the implementation of the new data-entry software, though, we paid attention to create a user-friendly tool that could be used in clinical practice. Although it may not replace the software for patients management within the CF centres, the availability of a tool that allows users to see graphs of patient's outcomes over time, centre data reports, and to download their centre's data, hopefully will boost the motivation to actively participate to the ECFSPR by entering good quality data as well as reducing the amount of missing data; 2. The second reason for missing data is poor variables definition or misunderstandings in their coding. 4. The fourth reason is the bad habit of answering a yes-no question only when there is an affirmative reply (sometimes aggravated by the software with only one tick box implying that an un-ticked box means no). This makes it hard to distinguish between true negatives (no), true unknowns (the user doesn't know the information, so they deliberately leave the field untouched) and the omissions (the user forgot to answer the question: the answer could be either yes or no), as described by the insulin example above. From the beginning, the ECFSPR data collection form has required an active answer to all questions, without any pre-set values; 5. The fifth reason of the occurrence of missing data is the fact that not all the national registries collect the same variables, therefore some information is missing for some countries. For example, in the 2008-2009 data collection, for five countries the information on the use of inhaled hypertonic saline is missing because such information is not collected in a systematic way. In other cases the national registry definition is so different from the ECFSPR definition that the country chose to set the whole variable to missing, such as happened for one country with the information on chronic infection by Pseudomonas aeruginosa; 6. A final reason is the lack of information for particular sub-groups of patients due to specific patients' characteristics. These are, for example, adults that do not have access to an adult CF centre and are lost to follow-up after leaving a children's clinic, or transplanted patients who move from a CF centre to a transplant centre that does not participate to the registry. In the ECFSPR experience, there have been anecdotal reporting of the last group of patients from one national registry, but a thorough investigation of the problem has still to be carried out.
## Maintaining confidentiality of registry data
Anonymity of the individual patient is fundamental both in the patient's decision to consent to their data use, and in the management of the data. This is particularly important in a rare disease where the number of patients is limited. Patient identifiers such as name, address and full date of birth are very rarely needed for the aims of any disease registry, as long as the patient can be identified for error correction by the patient's care giver, who is privy to this information. Therefore, partial anonymisation (often called pseudo-anonymisation) can be used, whereby a patient ID can be linked to the full patient data only at the hospital site.
This poses other problems that need to be overcome: local handling of consents (because the signature on the consent form would disclose the patient's name) and tailored software solutions (the local care giver must see the name of the patient while entering data to avoid confused identities, but the name cannot be sent to the central database).
The anonymity of data stored in the ECFSPR database was guaranteed by two means: access to personidentifiable data is granted only locally (CF centre or national registry) and creation of a random EU centre/ country number so that the centre is not identifiable in the database. For national registries, no person identifiable data are transferred, but for the individual centres, who needed to see the patient name in order to enter the data correctly, these data are stored separately and encrypted without means of access from the ECFSPR team. The centres or national registries have all been issued a random centre. The link between the centre number and the centre name is held by the helpdesk personnel and the centre names do not appear in the registry database, thus enhancing anonymity also for patients attending very small centres.
## Dissemination of results and use of data
Timely dissemination of results and appropriate use of data collected are the key elements for achieving the registry gold objectives: enhance knowledge of the disease under study and promote research. There are several ways through which results can be disseminated: publication of technical reports, communication at conferences and publication in peer-reviewed journals. The use of data should be governed to avoid misuse of data. For this purpose the ECFSPR developed an initial set of guidelines, a code of conduct and terms of reference documents endorsed by the participating countries.
## Annual reports
The ECFSPR annually publishes epidemiological descriptions of the data in a technical report, usually presented during the annual European Cystic Fibrosis Society Conference and subsequently available from the ECFSPR web site. Frequency tables, descriptive statistics and graphs give updates on main aspects of CF epidemiology such as demographics, diagnosis, genetics, lung function, nutrition, microbiology, complications and mortality. Results are presented at European level and separately by country, to allow comparisons.
In the latest issue of the annual data report, a special effort was made to make the report contents more patient-friendly than in the previous issues: we used technical jargon only when necessary, we commented tables and graphs, we provided instructions on how to read more complex graphs (such as box-plots), and we added a glossary of medical and statistical terms. The report also features a section dedicated to patients, containing a message from the ECFSPR team and an invitation to send comments.
Another reporting activity, fundamental in a disease registry activity, is the feedback on the data given to the data contributors. This has the double advantage of improving data collection and empowering data contributors. The ECFSPR sends to the participating CF centres a customised report summarising the centre's data and comparing them with data from other centres in the same country and with data from other countries. This gives the data contributor a report of the quality of the centre's data and, most importantly, the status of their patients compared with other centres. Data comparisons are performed in an anonymous way: the centre names are not disclosed, safeguarding confidentiality of each centre's aggregated data.
## Peer-reviewed publications
Publication in peer-reviewed journals is probably the most efficient and effective way to share with the scientific community the findings of research and submit it to its scrutiny.
The scientific activity of the ECFSPR allowed, for example, enhancing knowledge on potential risk factors of pulmonary disease in CF patients, highlighting the importance of their early identification and timely intervention [bib_ref] on behalf of the ECFS Patient Registry Steering Group: Factors associated with..., Kerem [/bib_ref].
The availability of a large database, such as the ECFSPR, offers a unique opportunity to analyse data from different populations. This opportunity was fully exploited to provide the CF specialists with a useful additional tool for patients care: Boëlle et al.published reference percentiles for FEV1 and BMI derived from a CF population. This allows CF specialist to have additional information on the patients they care for by comparing their lung function and BMI with their CF peers instead of against a healthy population alone.
Another important area of research activity is providing information for therapy development. This fundamental contribution was achieved in another publication [bib_ref] The relative frequency of CFTR mutation classes in European patients with cystic..., Boeck [/bib_ref] , where the authors describe the CFTR mutation class spectrum across Europe, highlighting which are the mutation classes to target for drug development in order to maximise the number of patients that will benefit from it and pointing out in which countries clinical trials could be performed thanks to the availability of patients carrying specific mutations.
Finally, research outcomes have the potential to urge for political decisions that have an impact on CF patients' life. The paper by McCormick et al. [bib_ref] Comparative demographics of the European cystic fibrosis population: a cross-sectional database analysis, Mccormick [/bib_ref] is an example of how comparison of simple demographic indicators highlights health care inequalities: the authors in fact showed that very different age structures of CF patients were observed between countries, despite a common genotype, according to their longevity of European Union membership, suggesting that health-care spending in new EU entrant countries would correspond to improved survival of patients.
## Access to data
Making the data available to the scientific community for research purposes is a noble and a burdensome responsibility. Granting access entails the responsibility to provide high quality data and ensuring legal and ethical use of the data by third parties.
The ECFSPR decided to grant access of the data only after some years of its activity, when crucial issues on uniformity of definitions and data completeness had been resolved. Access is strictly regulated by a Standard Operating Procedure, according to which requests on the use of data are reviewed by a scientific committee that formulates a recommendation for its approval and the request is forwarded to all the data contributors that ultimately give permission to use their data. A clear authorship, acknowledgement and publication policy has been set up, to ensure fair recognition of people's work and contribution. The data collected by the ECFSPR are at the moment being analysed in order to estimate the need for and plan the care of adult CF patients in the future (ERS/ECFS Task Force on Adult Care in Cystic Fibrosis).
# Discussion
There are some critical points in a disease registry development that are independent of resource allocation and that should be carefully considered in the planning phases of a disease registry.
The first important point is the definition of the information to record. The experience gained by the ECFSPR definitions working group proved that the most costeffective approach is to first work on the definition of a few variables, then pilot the definition for data collection on a restricted group of data contributors or for one data collection point, make the necessary amendments and then move on with the definition of other variables. This sequential approach, as opposed to the one of completing the full set of definitions in one go, has the advantages of not delaying the data collection for too long: if definitions are agreed on a core set of variables of high research interest first, data collection can start straight away, instead of being postponed by the time needed to define other, less important variables. The initial ECFSPR pilot study highlighted the importance of having common definitions for the variables to collect. If we had started data collection before already established databases from national CF registries, fewer compromises would have been necessary on definitions and on the level of detail of information collected, but such an opportunity is probably not available for most rare diseases. In order to set up definitions that would make data analyses results useful and at the same time data retrieval from clinical files feasible, we had to sacrifice some of the internationally acknowledged definitions; in some cases, to avoid large amount of missing data, we had to use proxy variables, such as use of pancreatic enzymes as a proxy for pancreatic insufficiency in order to guarantee fair comparisons across countries. In the most critical cases, national registries recorded information according to internally-agreed definitions, and where compromise was not feasible we had to content ourselves with missing data from those countries where their definitions were too far from what we intended. Fortunately, the national registries have been very collaborative, sometimes even changing their definitions to accommodate the ECFSPR ones in order to reduce the amount of missing data. Involving the national registry representatives in the definitions workgroup has been crucial for this understanding. For this reason we encourage researchers willing to set up disease registries to have early meeting with all potential data contributors, and to start a collaborative relationship. Finally, the ECFSPR definitions group advised a periodic revision of the information registered. This task is considered fundamental for an efficient perspective data collection, for three reasons: (1) definitions need often to be validated in real settings, and in some cases clarifications are necessary to people in charge of data retrieval and data recording;
(2) variables might prove to be of limited utility for research due to the way they have been defined or coded;
(3) improvements in knowledge of the disease and advances in scientific/technological discoveries make the collection of new information necessary: the registry needs to be constantly in tune with changes, to avoid that the information collected is no longer useful to researchers and clinicians. The ECFSPR is starting the second revision of its data collection form, reviewing the definition of some variables (such as diabetes) and evaluating the inclusion of new ones (e.g. computerised tomography imaging and lung clearance index). It is important to remember that such revision and the choice to modify the information collected has an impact on the data already collected: careful examination of whether there will be limitations in linking the data across the years should be performed. Similarly, when choosing to add new information to data collection, the effort needed to retrieve such information for the patients already included in the registry has to be carefully evaluated. For these reasons, the ECFSPR will have to carefully balance the advantages and disadvantages of modifying the data-entry forms.
Although adoption of common definitions, together with the use of a common data collection platform, should guarantee comparable data across countries, differences in outcomes between countries can still be observed, and there is a risk of over-or mis-interpreting these. They can be due to many factors, like different population demographics, health care systems, standards of care and national economics. These are the true differences that should be discussed and used for improvement of CF outcomes. However, differences may also be due to measurement methods and different translation of clinical findings: observational studies (such as patient registries) are more prone than other study designs (such as clinical trials) to the risk of artificial differences due to e.g. different measurement instruments or measurement practices. When such heterogeneity is observed, it is advisable that the registry validates the data to ensure that the differences seen are real. For example, in the ECFSPR 2008-2009 data collection, we found bigger differences in liver disease severity than expected from the natural history of this complication. For this reason the ECFSPR recently started a project on liver disease investigating the diagnostic work-up performed on randomly chosen patients and centres participating to the ECFSPR. Also, the ECFSPR decided to use internationally agreed reference values for pulmonary function and anthropometric measurements, but there is the awareness that such choice is not appropriate for all the CF populations registered in the ECFSPR, therefore standardised values are used only when comparison between countries are carried out, and careful comments always accompany the outcomes of such comparisons.
Another crucial aspect in a disease registry development is the careful consideration of the growing registry needs. With 23 countries, almost as many languages and the reality of working with both national registries with more than 5,000 patients and small centres with less than 50 patients, the ECFSPR had to face many challenges. The choice of an electronic data collection system eased the data-recording process, but it triggered the recruitment of additional workforce (IT experts, helpdesk staff for technical assistance) and the need of technical devices (e.g. server) that need maintenance. The software for data-entry must fulfil standards of quality and security, it should be tailored to the needs of the registry, and it should be user-friendly. Ideally, the physicians should use the data collected through the software in the daily patient management, thus rewarding them directly for the data retrieval and data-entry effort. The patients that consented to have their data collected should be given graphical feedback through the software, to see the benefits of participation to the registry. The first software adopted by the ECFSPR was quite demanding in terms of on-site installation and maintenance, and the burden on the centres as well as the technical helpdesk was considerable. This was an important lesson learned for the development of the new software. A multilingual helpdesk approach has been necessary even if the official language of the registry is English. The development of standard operating procedures, the drafting of technical documentation, and the use of document-sharing technologies and web teleconferencing have proven very effective in the daily management of the operational activities of the ECFSPR. This increased complexity in the registry structure has to be carefully considered by researchers, and an accurate cost-benefit analysis should be performed, especially when the funding opportunities are not secured. At the moment, the ECFSPR employs one full time coordinator, one full time statistician and two part-time help-desk personnel. Additional cost of software development will vary depending on the need. Further to this, the cost of running the national registries may be even higher in the larger countries, whereas some smaller countries rely almost entirely on volunteer effort by doctors or patient organisations. It is crucial for the sustainability of the registry, that these costs are compared to the possibility of funding.
One future aspect of the ECFSPR is to expand cooperation with the pharmaceutical industry and EMA in order to collect pharmacovigilance data on new drugs. This work is in progress in some of the national registries (such as UK and Germany), but until now the ECFSPR has been hampered by the three year delay in data collection, which would not allow timely identification of serious side effects. However, with the new software, which allows real-time use in the clinical setting, the possibility of pharmacovigilance data collection is open for the centres and countries reporting directly to the ECFSPR.
Finally, missing data is a well-known problem for the registries without a unique solution. Often missing data can be avoided by appropriate choice of information to be recorded, adequate level of detail and coding of information, and availability of well-trained, motivated and dedicated personnel in charge of data retrieval and data entry. The ECFSPR collects data from national registries that for some centres have funded data entry, and for others require data entry in order to be acknowledged as a CF centre. These are ways to motivate people, but at the moment, the ECFSPR does not have funding for the participating centres (and neither any authority over them). Another way to motivate people is to show them the utility of the data they have been collecting at regular feedback sessions as we do with the annual report and the centre report; and lately with the new software that offers interactive graphs and tables for use in the daily clinic management and for patient information.
# Conclusions
Setting up and maintaining a disease registry is a burdensome venture. Since the tentative beginning in 2003, the ECFSPR has evolved from a small working group of enthusiastic and knowledgeable national registry representatives, collecting data on spreadsheet files with very little funding and resources, to a professionally run patient registry with full and part time dedicated personnel that includes an executive director, an executive coordinator, two statisticians, an helpdesk service, a governing body composed by ten people (Executive Committee), a bespoke software and extensive use of data collected from 23 countries and more than 30,000 patients. The road to this has been paved with obstacles and challenges, and the journey is by no means at an end. A registry of these proportions may be initiated, but cannot be run as an amateur project by a few dedicated people; we could have probably accomplished our goals faster and have had fewer bumps on the road had the funding been in place at an earlier stage. For this reason, we recommend that objectives, structure and research outcomes are planned according to the available funding, in order to optimise resources allocation and avoid early frustration.
In the ECFSPR experience we found crucial for success the collaboration with existing national and international registries and cystic-fibrosis organisations (such as the ECFS Clinical Trial Network). Particularly helpful has been the patients' involvement in the registry activities through their representatives in governance committees in order to better meet the patients' needs and to convey the information about the registry in an effective way (through the patient-friendly report and the publication of web pages dedicated to patients in the ECFSPR website).
Another key aspect has been the networking for the recruitment of new centres to contribute the data, and the stimulation of their participation through their empowerment (participation to the ECFSPR governance bodies), through fair reward by co-authorship in peer-reviewed publications, and by publication of periodical (centrebased) data reports.
Finally, funding is a pivotal aspect in a disease registry running. The current registry sustainability cost is in staff (100,000 Euros per year) to retrieve, check, and analyse the data. But since the real cost to set up the registry exceeded 1,000,000 Euros over the last decade, this experience creates a cost-effective approach because the experience is donated as a gift to the community. A system for centres accreditation and funding according to centre's data completeness and data quality has been successfully used in many disease registries. This approach, however, is quite ambitious for most registries, especially for rare diseases. Pragmatically, where funding is limited, much can be achieved by restricting the data collection to a core set of data, usually referring to demographic, diagnosis and death information, which are easy to retrieve for most data contributors [bib_ref] Comparative demographics of the European cystic fibrosis population: a cross-sectional database analysis, Mccormick [/bib_ref]. The cost of running a national registry varies depending on the size and organisation of the national registry, and whether it is used locally for e.g. quality control. The new ECFSPR software will offer a cheaper solution for some countries by supplying free software and data availability locally and nationally.
[table] Table 1: Overview of critical aspects when setting up a registry and the solutions implemented by European Cystic Fibrosis Society Patient Registry [/table]
|
Extragenital lichen sclerosus et atrophicus within a skin graft scar
# Introduction
Lichen sclerosus et atrophicus (LSA) is rare chronic relapsing skin condition typically arising in postmenopausal women and prepubescent children in the genital area. Only 6% of cases occur in extragenital skin without concomitant anogenital involvement. [bib_ref] Lichen sclerosus et atrophicus, Wallace [/bib_ref] LSA can exhibit the Koebner phenomenon. Although various types of inciting events have been documented for the development of extragenital lichen sclerosus et atrophicus, we present the first case, to our knowledge, to occur within a skin graft donor site. Significantly, this is the fourth reported case of koebnerizing extragenital LSA without coexisting or previous genital involvement.
## Case presentation
A 78-year-old woman with a history of diabetes, chronic kidney disease, and stage 1A breast cancer status postlumpectomy presented for evaluation of a new itchy and painful rash of the right posterior thigh and subsequently the bilateral inner thighs for 6 months' duration. Symptoms initially developed within a skin graft scar on the right posterior thigh, which was taken during early childhood to repair a vaginal wall birth defect called Rokitansky-Mayer-K€ uster-Hauser syndrome. She denied ever having genital involvement.
Examination found a large, well-circumscribed ivory white atrophic scar with a thin peripheral erythematous scaly plaque on the right posterior thigh [fig_ref] Fig 1: Extragenital lichen sclerosus et atrophicus [/fig_ref]. The bilateral medial thighs were diffusely atrophic with ill-defined xerotic plaques with a cigarette paper texture [fig_ref] Fig 1: Extragenital lichen sclerosus et atrophicus [/fig_ref]. A genital examination was unremarkable.
Histologic sections from the right posterior thigh and the right medial thigh showed an atrophic epidermis with compact hyperkeratosis and focal areas of thickened granular layer. In the dermis was mild edema and homogenization of the papillary dermis with a scant perivascular lymphocytic inflammatory infiltrate [fig_ref] Fig 2: Punch biopsy of LSA from the skin graft scar on low power... [/fig_ref]. Based on the clinical and histopathologic features, extragenital LSA within a skin graft site scar was diagnosed. She was treated with clobetasol 0.05% ointment and calcipotriene 0.005% ointment twice daily. At her 2-months follow-up, the patient had significant symptomatic and cosmetic improvement [fig_ref] Fig 3: Extragenital LSA on the right posterior thigh skin graft scar site after... [/fig_ref].
# Discussion
Lichen sclerosus et atrophicus (LSA) is a chronic relapsing skin condition characterized by early inflammation followed by chronic scarring and skin atrophy. LSA can occur at any age, but there is a typical bimodal onset in prepubescent children and postmenopausal women. Etiology of LSA remains unclear, but the literature suggests a likely autoimmune phenomenon in a genetically predisposed individual. Previous infections, trauma, and occlusive moist environments are also contributing factors. The lesions of LSA are typically ivory or porcelain-white shiny papules and macules that coalesce to form sclerotic plaques, as seen in our patient. Although classically seen in the anogenital area (80%-98%), it can be seen in extragenital sites in 15% to 20% of patients. [bib_ref] Lichen sclerosus, Powell [/bib_ref] Extragenital sites, although far less common, are most often seen on the buttocks, thighs, breasts, submammary area, and neck. Only 6% of LSA are isolated to extragenital lesions, as seen in our case. [bib_ref] Lichen sclerosus et atrophicus, Wallace [/bib_ref] Extragenital LSA can exhibit the isomorphic or Koebner phenomenon in which typical lesions of the disorder occur after inciting events. Few case reports are available supporting this concept, with LSA found within postirradiation sites, cholecystectomy scars, vaccination sites, jellyfish stings, herpes zoster scars, sun burns, sites of friction, and injection sites among others. [bib_ref] Postirradiation lichen sclerosus, Nemer [/bib_ref] [bib_ref] Lichen sclerosus et atrophicus in a surgical scar treated with fractional laser, Mendieta-Eckert [/bib_ref] [bib_ref] Lichen sclerosus et atrophicus in a vaccination site, Anderton [/bib_ref] [bib_ref] Koebner phenomenon in a patient with lichen sclerosus following a jellyfish sting:..., Perez-Lopez [/bib_ref] [bib_ref] Bilateral Koebner phenomenon in lichen sclerosus et atrophicus, Ronnen [/bib_ref] [bib_ref] Lichen sclerosus et atrophicus appearing in an old burn scar, Meffert [/bib_ref] [bib_ref] Koebner phenomenon in lichen sclerosus et atrophicus, Pock [/bib_ref] Although various inciting events have been documented, we believe this patient to be the first documented case of LSA involving a skin graft donor site.
Most cases that show koebnerization have concurrent or a history of genital LSA. To the best of our knowledge, this patient represents the fourth reported case of extragenital LSA, with a known inciting event, developing without concomitant or preexisting anogenital LSA. Four cases of koebnerized extragenital LSA without anogenital involvement have been reported previously: in a vaccination site, an old burn scar, a cholecystectomy scar, and a surgical scar. [bib_ref] Lichen sclerosus et atrophicus in a surgical scar treated with fractional laser, Mendieta-Eckert [/bib_ref] [bib_ref] Lichen sclerosus et atrophicus in a vaccination site, Anderton [/bib_ref] [bib_ref] Koebner phenomenon in a patient with lichen sclerosus following a jellyfish sting:..., Perez-Lopez [/bib_ref] [bib_ref] Koebner phenomenon in lichen sclerosus et atrophicus, Pock [/bib_ref] The timeframe varies among koebnerizing cases of extragenital LSA. Some appear within months of the trigger, [bib_ref] Lichen sclerosus et atrophicus in a vaccination site, Anderton [/bib_ref] [bib_ref] Bilateral Koebner phenomenon in lichen sclerosus et atrophicus, Ronnen [/bib_ref] and others develop several years after the trigger. [bib_ref] Postirradiation lichen sclerosus, Nemer [/bib_ref] [bib_ref] Lichen sclerosus et atrophicus in a vaccination site, Anderton [/bib_ref] [bib_ref] Bilateral Koebner phenomenon in lichen sclerosus et atrophicus, Ronnen [/bib_ref] In a case series of 6 of patients with LSA in postirradiation sites, patients had lesions anywhere from 2 to 12 years after radiation. [bib_ref] Postirradiation lichen sclerosus, Nemer [/bib_ref] Meffert and Grimwood 8 reported a case in 1994 of LSA developing 50 years after a childhood burn injury. This case was the longest time lapse reported to date until our patient, who had LSA at least 70 years after the skin graft was initially taken. Unlike anogenital LSA, extragenital LSA does not carry an increased risk of squamous cell carcinoma. Little is known about the long-term complications of extragenital LSA and the risk of repeated koebnerizing lesions developing in areas of trauma. First-line treatment for extragenital LSA consists of high-potency topical corticosteroids and emollients. [bib_ref] Guidelines for the management of lichen sclerosus, Neil [/bib_ref] Although extragenital lesions are not as responsive as genital lesions, our patient has noted significant improvement with these measures in only 2 months time. Second-and third-line treatment options include topical calcineurin inhibitors, systemic retinoids, surgery, phototherapy, photodynamic therapy, and fractional ablative lasers.
[fig] Fig 1: Extragenital lichen sclerosus et atrophicus. A, On the right posterior thigh within a scar from a remote skin graft. B, Along the right medial thigh. [/fig]
[fig] Fig 2: Punch biopsy of LSA from the skin graft scar on low power view. (Hematoxylin-eosin stain.) [/fig]
[fig] Fig 3: Extragenital LSA on the right posterior thigh skin graft scar site after 4 months of treatment. [/fig]
[table] Abbreviation used: LSA: lichen sclerosus et atrophicus //doi.org/10.1016/j.jdcr.2018.09.007 [/table]
|
Primary and secondary dystonic syndromes: an update
Purpose of reviewThe dystonias are a common but complex group of disorders that show considerable variation in cause and clinical presentation. The purpose of this review is to highlight the most important discoveries and insights from across the field over the period of the past 18 months.Recent findingsFive new genes for primary dystonia (PRRT2, CIZ1, ANO3, TUBB4A and GNAL) have made their appearance in the literature. New subtypes of neuronal brain iron accumulation have been delineated and linked to mutations in C19orf12 and WDR45, while a new treatable form of dystonia with brain manganese deposition related to mutations in SLC30A10 has been described. At the same time, the phenotypes of other forms of dystonic syndromes have been expanded or linked together. Finally, there has been increasing recognition of both the extramotor phenotype in dystonia and the part played by the cerebellum in its pathophysiology.SummaryRecently, there has been unprecedented change in the scientific landscape with respect to the cause of various dystonic syndromes that is likely to make a direct impact on clinical practice in the near future. Understanding the genetic cause of these syndromes and the often wide phenotypic variation in their presentations will improve diagnosis and treatment. With time, these discoveries may also lead to much-needed progress in elucidating the underlying pathophysiology of dystonia.
# Introduction
The dystonias are a heterogenous group of hyperkinetic movement disorders, characterized by involuntary sustained muscle contractions affecting one or more sites of the body that lead to twisting and repetitive movements or abnormal postures of the affected body part. They represent the third most common movement disorder worldwide [bib_ref] Dystonia 2, Fanh [/bib_ref] [bib_ref] Do primary adult-onset focal dystonias share aetiological factors?, Defazio [/bib_ref] [bib_ref] The pathophysiological basis of dystonias, Breakefield [/bib_ref] [bib_ref] The diagnosis of dystonia, Geyer [/bib_ref]. There are currently 23 DYT loci referring to primary dystonic syndromes, as well as a number of primary dystonic syndromes not currently assigned to any DYT loci (see [fig_ref] Table 1: The current DYT loci [/fig_ref]. In addition, there are numerous heredodegenerative disorders in the context of which dystonia is commonly seen.
Recent advances in the field of dystonia have mirrored to a considerable degree advances in genetic technology. Fuelled by the increasing use of wholeexome sequencing approaches, researchers have been able to elucidate the genetic cause of various familial forms of dystonia at an unprecedented rate. In doing so, several new dystonic syndromes have come into existence, whereas other long-recognized dystonic syndromes have been linked either to novel or occasionally even well-known genes. These discoveries are addressed in the following sections. dystonia with a prevalence of about one in 150 000 individuals [bib_ref] Clinical evaluation of idiopathic paroxysmal kinesigenic dyskinesia: new diagnostic criteria, Bruno [/bib_ref]. It is characterized by frequent (up to 100 times a day) attacks of dystonic or choreiform movements lasting from a few seconds to a few minutes. In late 2011, mutations in PRRT2 (proline-rich transmembrane protein 2) were identified as the cause of PKD in nearly all cases [bib_ref] Exome sequencing identifies truncating mutations in PRRT2 that cause paroxysmal kinesigenic dyskinesia, Chen [/bib_ref] [bib_ref] Identification of PRRT2 as the causative gene of paroxysmal kinesigenic dyskinesias, Wang [/bib_ref]. Most mutations are truncating and by far the most common of these is the c.649dupC mutation, but missense variants (possibly with reduced penetrance) have also been described [bib_ref] Novel PRRT2 mutation in an African-American family with paroxysmal kinesigenic dyskinesia, Hedera [/bib_ref] [bib_ref] Mild paroxysmal kinesigenic dyskinesia caused by PRRT2 missense mutation with reduced penetrance, Friedman [/bib_ref] [bib_ref] The PRRT2 mutation c.649dupC is the so far most frequent cause of..., Steinlein [/bib_ref] [bib_ref] Identification of a novel PRRT2 mutation in patients with paroxysmal kinesigenic dyskinesias..., Cao [/bib_ref]. Interestingly, the c.649dupC mutation was also found in the family used to define DYT19, a supposedly separate form of PKD, suggesting that the initial linkage in this family was incorrect and DYT19 is in fact synonymous with DYT10 [bib_ref] PRRT2 gene mutations: from paroxysmal dyskinesia to episodic ataxia and hemiplegic migraine, Gardiner [/bib_ref].
## Key points
The past 2 years have seen considerable change in the field of dystonia driven predominantly by rapid advances in genetic sequencing technologies.
Five new genes for primary dystonia have appeared in the literature (PRRT2, CIZ1, ANO3, TUBB4A and GNAL).
Two new forms of neuronal brain iron accumulation and a potentially treatable form of generalized dystonia with brain manganese deposition are also now known.
Pathophysiological understanding continues to lag behind, but increasing emphasis recently on dystonia as a network disease, involving brain areas beyond the basal ganglia, is an important step forward. The gene encodes a protein which is highly expressed throughout the nervous system. On a cellular level, it appears that the PRRT2 protein is predominantly localized to the synapse and that it may participate in neurotransmitter release, suggesting disrupted neurotransmission as the possible pathological mechanism in PKD [bib_ref] Mutations in the gene PRRT2 cause paroxysmal kinesigenic dyskinesia with infantile convulsions, Lee [/bib_ref].
Most interesting of all, perhaps, has been the recognition that mutations in this gene also cause a number of other paroxysmal disorders. In fact, PRRT2-related disease appears to be a stunning example of the kind of phenotypic heterogeneity often encountered in genetic forms of dystonia, with clinical presentation differing not merely between individuals carrying different mutations, but also between individuals carrying the same mutation and even between affected members within the same family [bib_ref] Clinical features of childhood-onset paroxysmal kinesigenic dyskinesia with PRRT2 gene mutations, Silveira-Moriyama [/bib_ref]. Two forms of childhood epilepsy, namely, infantile convulsions with choreoathetosis (ICCA) and benign familial infantile seizures, as well as episodic ataxia, hemiplegic migraine and benign paroxysmal torticollis of infancy are all now recognized as possible manifestations of mutations in this gene [bib_ref] PRRT2 gene mutations: from paroxysmal dyskinesia to episodic ataxia and hemiplegic migraine, Gardiner [/bib_ref] [bib_ref] PRRT2 mutations cause benign familial infantile epilepsy and infantile convulsions with choreoathetosis..., Heron [/bib_ref] [bib_ref] Mutations in the novel protein PRRT2 cause paroxysmal kinesigenic dyskinesia with infantile..., Lee [/bib_ref] [bib_ref] Familial PRRT2 mutation with heterogeneous paroxysmal disorders including paroxysmal torticollis and hemiplegic..., Dale [/bib_ref].
## Expanding phenotypes in dystonia
In a similar fashion to PRRT2, the phenotype of mutations in ATP1A3 has recently been revised to include what was previously considered a separate condition. Mutations in this gene are traditionally associated with rapid onset dystonia-parkinsonism. In the last few months, however, it has become clear that ATP1A3 mutations are also responsible for a neuropaediatric condition labelled alternating hemiplegia of childhood (AHC) ]. AHC is a rare disorder characterized by the onset before the age of 18 months of paroxysmal neurological events, such as hemiplegia alternating in laterality, quadriplegia, dystonic spells, oculomotor abnormalities and autonomic dysfunction, all of which abate during sleep . Nonparoxysmal manifestations develop after a few months or years of the disease, comprising developmental delay, intellectual disability of variable degree, ataxia, dysarthria, choreoathetosis, and, in some patients, pyramidal tract signs. Two separate studies were able to demonstrate de-novo heterozygous mutations in ATP1A3 as the cause of the disorder in 75-100% of their case cohorts ]. Meanwhile, the phenotypic spectrum of mutations in SGCE, which causes myoclonus dystonia, has also been revised to reflect an increased awareness of psychiatric manifestations of the disease. One recent study systematically assessed 27 SGCE mutation carriers using a battery of standardized psychiatric interviews and questionnaires and compared them with a disability-matched control group of patients with alcohol-responsive tremor ]. Obsessive-compulsive disorder was eight times more likely and generalized anxiety disorder and alcohol dependence five times more likely in SGCE mutation carriers than in the tremor controls. This well designed study indicates that SGCE mutations are associated with a specific psychiatric phenotype consisting of compulsivity, anxiety and alcohol misuse, beyond the typical motor manifestations ]. Finally, primary dystonia itself has undergone a phenotypic expansion of sorts in light of growing evidence indicating an important nonmotor component to the condition, including abnormalities in sensory and perceptual functions as well as neuropsychiatric, cognitive and sleep domains. These additional features may help explain the fact that reduction in quality of life in dystonia does not always correlate with the extent of the motor deficit. Moreover, it may be possible to capitalize on some of these extramotor signs -impaired temporal discrimination thresholds, for instance -to detect nonmanifesting mutation carriers in familial forms of dystonia, which would greatly facilitate the identification of new genes wherein establishing segregation is key. Interested readers are directed to a recent review by Stamelou et al.
## Four new genes for primary pure dystonia
In 2012, four new genes have been reported to cause primary pure dystonia, in which dystonia is the sole manifestation with the exception of tremor (see [fig_ref] Table 2: Summary of the four new forms of primary pure dystonia identified in... [/fig_ref]. Most studies used to identify the new genes have employed a whole-exome sequencing approach combined with linkage analysis to tackle smaller kindreds than was previously possible.
The first gene to appear on the scene was CIZ1. A missense mutation in this gene was published as the likely cause of focal, adult-onset cervical dystonia, variably associated with mild tremor, which was inherited as an autosomal dominant trait with reduced penetrance [24 others have raised questions regarding the quality and coverage of the exome data, meaning it will be particularly important to await independent confirmation of mutations in this gene before it can be fully accepted as a definite cause of dystonia [bib_ref] Exome sequencing for gene discovery: time to set standard criteria, Klein [/bib_ref]. The protein encoded by CIZ1 is a p21 Cip1/Waf1interacting zinc finger protein that is expressed in the brain and involved in DNA synthesis and cell-cycle control [bib_ref] Ciz1 cooperates with cyclin-A-CDK2 to activate mammalian DNA replication in vitro, Copeland [/bib_ref].
Mutations in a second gene, ANO3, were identified as the cause of autosomal dominant craniocervical dystonia in a moderately sized white kindred ]. Subsequent screening of the whole gene in a cohort of 188 individuals with cervical dystonia revealed five further novel of variants, of which segregation could be demonstrated for two. Clinically, patients with mutations in this gene exhibited focal or segmental dystonia, variably affecting the craniocervical, laryngeal or brachial regions. There was often dystonic tremor with a jerky quality affecting the head, voice or upper limbs. The age at onset ranged from the very early childhood to 40 years of age. As with CIZ1, the dystonia was never seen to generalize in any affected individual, remaining confined to the head, neck and upper limbs. Indeed, some individuals with mutations in this gene manifested upper limb tremor alone and had been misclassified as essential tremor ]. Little is known about the function of ANO3, but it is most highly expressed in the striatum and is thought to encode a Ca 2þ -activated chloride channel . Such channels are believed to have a role to play in modulating neuronal excitability, suggesting a possible mechanism by which altered channel function could lead to aberrant neuronal excitability manifest as dystonic movements [bib_ref] Physiological roles and diseases of Tmem16/Anoctamin proteins: are they all chloride channels?, Duran [/bib_ref].
TUBB4A has recently been identified by two groups independently as the cause of DYT4 dystonia, also known as whispering dysphonia & ]. The condition was first described by forensic psychiatrist Neville Parker [bib_ref] Hereditary whispering dysphonia, Parker [/bib_ref] in a large family with third decade onset of autosomal dominantly inherited laryngeal dystonia with generalization. Over 25 affected individuals have since been reported, with some exhibiting a distinctive 'hobby horse' gait [bib_ref] Whispering dysphonia in an Australian family (DYT4): a clinical and genetic reappraisal, Wilcox [/bib_ref]. Alcohol responsiveness is not uncommon, leading to severe alcohol abuse in some DYT4 patients [bib_ref] Whispering dysphonia in an Australian family (DYT4): a clinical and genetic reappraisal, Wilcox [/bib_ref]. To date, however, no other family with this characteristic phenotype has been described, suggesting that the condition is very rare. TUBB4A encodes b-tubulin-4a, a constituent of microtubules, and the mutation in this family (p.Arg2Gly) results in an arginine to glycine aminoacid substitution in the key, highly conserved autoregulatory MREI (methionine-arginine-glutamic acid-isoleucine) domain of the protein. One further novel variant in this gene (p.Ala271Thr) was detected in an individual exhibiting spasmodic dysphonia with oromandibular dystonia and dyskinesia with an age at onset of 60 ]. However, segregation analysis was not possible, meaning the pathogenicity of the variant is uncertain.
Finally, mutations in GNAL have recently been identified as a cause of autosomal dominant primary dystonia and subsequently confirmed by a second group [34
[formula] && ,35 & ]. [/formula]
Initially, it appeared that GNAL may have been an exceptionally common cause of familial dystonia with the first group reporting that six out of only 39 families screened ($15%) haboured mutations in this gene ]. However, the second group found only three mutations in 760 individuals screened (<0.5%), which is perhaps more in line with what might be expected ]. Regardless, clinical characterization of individuals carrying mutations in GNAL showed a strong cervical predilection, with 82% having onset in the region of the neck and 93% displaying cervical involvement by the time they were examined for the study in question. However, progression to other sites had occurred in at least half of those affected and generalized dystonia was seen in about 10% of cases ]. Thus, the clinical phenotype in GNAL-related dystonia appears to be not dissimilar to that caused by mutations in THAP1. The gene is highly expressed in the olfactory epithelium and hyposmia was noted in some affected individuals, though not consistently ]. GNAL encodes Gaolf, the alpha subunit of triheteromeric G protein Golf, which is involved in dopamine (D1) signalling [bib_ref] G(olf)alpha mediates dopamine D1 receptor signaling, Zhuang [/bib_ref]. As D1 receptors have a known role in mediating locomotor activity, the link between GNAL and dystonia is biologically plausible.
## A new wilson's disease?
Recently, the first inborn error of manganese metabolism has been identified, clinically resembling Wilson's disease. Biallelic mutations in the gene SLC30A10, which encodes a manganese transporter, were shown to be the cause of a syndrome that consisted of early-onset generalized dystonia, liver cirrhosis, polycythemia, and hypermanganesaemia ]. Serum manganese levels are generally in the range 1000-6000 nmol/l (normal <320) and MRI of the brain typically shows hyperintensities in the basal ganglia as well as the subthalamic and dentate nuclei. Over 20 affected individuals from 10 different families have been described by two separate groups.
As with Wilson's disease, the disorder appears at least partially treatable. Using a combination of ethylenediaminetetraacetic acid and ferrous fumarate, clinicians were able to reduce serum manganese and induce improvement in dystonia in patients with this condition [bib_ref] Dystonia with brain manganese accumulation resulting from SLC30A10 mutations: a new treatable..., Stamelou [/bib_ref]. Again, as with Wilson's disease, early introduction of treatment is likely to be key so that evaluation of serum manganese may also take its place alongside copper and caeruloplasmin studies as one of the few obligatory investigations in young-onset dystonia of unexplained cause.
## New forms of neurodegeneration with brain iron accumulation
Neurodegeneration with brain iron accumulation (NBIA) results from excessive iron deposition in the brain, mainly the basal ganglia. Pantothenate kinaseassociated neurodegeneration (PKAN, NBIA1) and PLA2G6-associated neurodegeneration (PLAN, NBIA2) are the core syndromes, but several other less frequent genetic causes have been identified (including FA2H, ATP13A2, CP and FTL) [bib_ref] Excess iron harms the brain: the syndromes of neurodegeneration with brain iron..., Schneider [/bib_ref]. Now, two further subtypes have been identified and linked to mutations in the genes C19orf12 and WDR45.
Mutations in C19orf12, an orphan mitochondrial protein, appear to be a relatively common cause of NBIA, accounting for 23 of 161 idiopathic cases ($15%) screened by one group [bib_ref] Absence of an orphan mitochondrial protein, c19orf12, causes a distinct clinical subtype..., Hartig [/bib_ref]. The condition, dubbed mitochondrial membrane protein-associated neurodegeneration or MPAN, is characterized by cognitive decline progressing to dementia, prominent neuropsychiatric abnormalities and a motor neuronopathy [bib_ref] New NBIA subtype: genetic, clinical, pathologic, and radiographic features of MPAN, Hogarth [/bib_ref]. Dystonia is common ($75%) and frequently affects the hands and feet, although it is generalized in some. On MRI, the 'eye-of-the-tiger' sign is absent, but iron deposition is seen in the globus pallidus and the substantia nigra. As with PLAN, cortical Lewy bodies were present on neuropathological examination of brain tissue from an affected individual [bib_ref] New NBIA subtype: genetic, clinical, pathologic, and radiographic features of MPAN, Hogarth [/bib_ref].
WDR45 is located on the X chromosome and encodes a beta-propeller scaffold protein with a putative role in autophagy. In late 2012, mutations in this gene were linked to a novel form of NBIA, now known by the acronym SENDA (static encephalopathy of childhood with neurodegeneration in adulthood). SENDA is characterized by global developmental delay in early childhood and slow motor and cognitive gains until adolescence or early adulthood, at which point dystonia, parkinsonism, and dementia supervene & ]. A unique feature of this form of NBIA was reported to be T1 hyperintensity surrounding a central linear region of signal hypointensity within the substantia nigra and cerebral peduncles. From a genetic point of view, the condition is interesting for at least two reasons. First, all mutations occurred de novo such that Mendelian inheritance was never observed. Second, although WDR45 is located on the X chromosome and undergoes inactivation, the clinical features of this form of NBIA do not follow a pattern typical of an X-linked disorder. Instead, the phenotype of affected males is indistinguishable from that of females. It is thought that the most likely explanation for this is that males with germline mutations are nonviable and that postzygotic mutations are instead responsible for the condition in males ].
## Distinguishing psychogenic from nonpsychogenic dystonia
Accurately distinguishing psychogenic from nonpsychogenic, primary dystonia is essential if patients are to receive the form of treatment that is most likely to benefit them and progress is to be made with regard to understanding the underlying neurobiology. Two recent studies have sought reliable markers to distinguish these conditions. The first investigated cortical plasticity using transcranial magnetic stimulation and the paired associative stimulation protocol in patients with psychogenic or organic dystonia and compared them with healthy controls [bib_ref] Abnormal sensorimotor plasticity in organic but not in psychogenic dystonia, Quartarone [/bib_ref]. Whereas cortical plasticity was increased in the group with organic dystonia (as expected), it was normal in both patients with psychogenic dystonia and healthy controls. A second study used functional neuroimaging to investigate movement-related cerebral activation in three groups, one with psychogenic fixed dystonia, another with genetically determined organic dystonia and a third consisting of healthy controls ]. In organic dystonia, averaged regional cerebral blood flow was increased in the primary motor, premotor and parietal cortices, whereas it was decreased in the cerebellum. In contrast, patients with psychogenic dystonia displayed an opposite pattern of activation, with increased regional blood flow in the cerebellum and basal ganglia and decreased flow in the primary motor cortex. Together these studies suggest that distinct neurobiological mechanisms underpin organic and psychogenic dystonia, with the former related to aberrant motor cortical plasticity and metabolism and the latter related to abnormalities of frontosubcortical processing ].
## The rising importance of the cerebellum in dystonia
In terms of our understanding of the pathophysiology of dystonia, researchers have continued to make slow but steady progress. One notable trend in the last year has been a growing recognition of the part played by the cerebellum in the disorder. Multiple convergent strands of evidence from animal models, clinical observations, and anatomical, imaging and electrophysiological studies have highlighted the importance of the cerebellum in dystonia. At present, it remains unclear whether the cerebellar abnormalities observed in some of these studies are secondary (and perhaps compensatory) to basal ganglia dysfunction or are, in some cases at least, the driving force behind the development of dystonia. A recent review by Sadnicka et al. ] of the accumulating evidence in this area as well as its implications is recommended to the interested reader.
# Conclusion
Fuelled by the rapid improvement in genetic sequencing technologies, there has been an unprecedented advance in our understanding of the cause of several forms of primary and heredodegenerative dystonia. It is likely that the continued improvement of these technologies -in particular the introduction of inexpensive whole-genome sequencing -will lead to further advances in coming years. At present, many of the new genes identified have yet to be independently confirmed and their prevalence as a cause of dystonia is as yet unclear. Nonetheless, an understanding of the function of these new genes may help researchers identify key pathways in the development of dystonia that could form the targets of novel therapeutic agents or interventions. Indeed, there remains an urgent need to clarify the underlying pathophysiology of the disorder, which is still poorly understood despite considerable research and effort. This is likely to be a difficult task given the apparent heterogeneity of causes in dystonia, but it is probable that a shift away from the paradigm of dystonia as a disorder of the basal ganglia alone towards one of dystonia as the common manifestation of dysfunction at any one of a number of points in a complex network of brain areas will prove an important step in this process.
[table] Table 1: The current DYT loci (after removal of withdrawn, duplicate and unpublished loci) with associated phenotype, mode of inheritance and genetic cause or linkage interval, wherever known [/table]
[table] Table 2: Summary of the four new forms of primary pure dystonia identified in the last year [/table]
|
Health-related quality of life after gastric cancer treatment in Brazil: Narrative review and reflections
# Introduction
Most gastric adenocarcinomas affect men approximately 60-70 years of age. Approximately 65% of patients are over 50 years old. Gastric cancer is the third most common type of cancer among men and fifth among women in Brazil, with an estimated 13360 new cases among men and 7870 among women each year for the 2020-2022 period . According to the Brazilian National Cancer Institute, in the country's five geographic regions, without considering nonmelanoma skin tumors, gastric cancer in men is the second most frequent cancer in the north, followed by the northeast, occupying third position. In the south, southeast, and midwest, it is the fourth most frequent. For women, it is the fifth most frequent cancer in the south and in the north. The midwest and northeast occupy the sixth position, followed by the southeast, occupying the seventh position[1]. Although its incidence has declined, the registered mortality of gastric cancer is still high, reaching approximately 70% to 90% in Latin America, Asia, and Eastern Europe[2-4].
The term "quality of life" (QoL) can be used in everyday language by people from the general population and professionals in different fields. However, here, we refer specifically to its scientific research context in different fields of knowledge such as economics, sociology, education, and health specialties. The concept of QoL most frequently used in health studies is health-related QoL, encompassing the impact of a disease and its treatments on diverse aspects of life. The idea of QoL is centered on subjective assessment and reported by patients themselves, relating the influence of their health status to their ability to live fully .
The definition of "quality of life" is still being discussed in the scientific literature; its broad meaning might not be fully understood . Some authors confuse functional assessments and isolated elements of patients' lives with the broad and comprehensive definition used by the World Health Organization, which considers the subjective and multidimensional facets of QoL and defines it as the individual's perception of his/her position in life in the context of his/her culture and value system (including spiritual matters); QoL concerns one's goals, expectations, and standards . This notion incorporates components such as life experiences, well-being, satisfaction, and social and physical functions, which are influenced by physical and socio-economic factors, psychological factors, and perceived health status .
The evaluation of health-related QoL should be carried out using scientific instruments that are internationally validated in several languages and cross-cultural, reproducible, and comparable statistical tools. Such instruments, in the form of questionnaires, usually address physical, psychological, occupational, social, environmental, and spiritual relations (personal beliefs and religion), and they always maintain a multidimensional character and assess the individual's perception of his/her QoL. To avoid the researcher's influence in such an evaluation, most questionnaires have been developed in a self-administered manner, and prior healthcare team training is required. These questionnaires can be generic when applied to determine Given the increase in survival and the growing variety of cancer treatments, modern oncology is forced to confront its results and global impact, since no treatment is harmless, even when curative. It is thus necessary to develop and improve the various questionnaires used for this research, including both generic questionnaires such as the medical outcomes study 36-item short form health survey (SF-36) and specific instruments for each illness. Here, we mention some questionnaires for evaluating patients with gastric cancer, such as the functional assessment of cancer therapygastric (FACT-Ga). In exploring these instruments to better understand their scope and dynamics, please note the SF-36, a generic instrument for appraising multidimensional QoL, allows for comparison with other chronic diseases and the general population. The SF-36 has been translated and validated for the Brazilian population and consists of 36 items (questions) that encompass 8 domains: Functional capacity, pain, physical aspects, vitality, general health, emotional aspects, social aspects, and mental health. A score is assigned in the 8 domains ranging from 0 (worst) to 100 (best) for each domain for each question. The FACT-Ga is an example of a specific questionnaire to assess QoL in patients with gastric cancer and consists of 27 items divided into scales of functional, physical, emotional, social/family well-being, and additional concerns (GaCS). When added together, these scales derive a TOI (Trial Outcome Index), a FACT-G or a FACT-GASTRIC total score [3-8,17,18].
This study presents reflections and attempts to add knowledge to the theme of QoL among patients with gastric adenocarcinoma. The study describes some of the characteristics of this cancer in three regions of Brazil, with quantification of previously (supposedly) abstract facets, through the use of validated questionnaires. The study also evaluates the impacts of the disease in various dimensions of life, as reported by patients themselves (the protagonists of the information).
# Methodlogy
# Discussion and reflections
Gastric cancer is a significant disease with varying social impacts, depending on the reality of the Brazilian or worldwide region where it occurs and the affected population's access to specialized services. Its prognosis is directly related to the tumor's extension through the organ wall, lymph node involvement, and the team's expertise and qualification[1-4,24-30]. Surgical resection is the main therapeutic modality with curative potential, but the interdisciplinary approach is fundamental and leads to improved results. Nevertheless, therapy can adversely affect healthrelated QoL and is therefore undesirable, making it challenging to balance standardized treatment with the most complete and ideal response (including patients' perceptions of, and expectations about, their disease)[3,4].
In analyzing national data, we noted an average age of patients of 59.8 years. In addition, 54.9% of the patients were smokers, 51.9% were male, their average income range consisted of two minimum wages, and 43.7% were white . The predominant histopathological type observed in 62 (61.4%) patients, the intestinal type, reveals that the family relationship did not predominate in the disease's spread in our series. The frequency of 40.6% Helicobacter pylori contamination and 54.9% smokers reinforces the evidence of the correlation widely described in the literature, as well as the disease's primarily sporadic character, allowing for efficient collective strategies for prevention and for identifying the suppression risk factors most responsible for stomach carcinogenesis. This information can help to build policies on prevention and health promotion .
Data from Brazilian authors suggest that the QoL is more related to the type of treatment itself rather than anatomopathological, epidemiological, or demographic characteristics . This implies that most socioepidemiological variables do not interfere in QoL. Further, if these variables do not interfere in other aspects, they should not be considered relevant information for therapeutic decisions, even though some authors consider schooling to be an influencing factor of QoL . Such data strengthen the perception that specialized teams must address these patients from the beginning of their natural history, with multimodal and preventive interventions in the different elements of the disease and its treatment, not only the clinical or biological components. Interdisciplinary care in physiotherapy, nutrition, psychotherapy, and spiritual support (here, different from religion) is recommended for all patients. Each symptom or sign must be informed and treated in an early, specific, and individualized way .
The literature has tried to correlate initial staging with better QoL, notably, stage I[31]. Data from our group suggest that the higher the lymph node staging, the worse some scores may be in the domains of the SF-36 and the FACT-Ga questionnaires, hence reinforcing the importance of adequate lymphadenectomy, with an essential impact on the oncological outcomes of treatment and improvement of the staging (by more representative lymph node numerical sampling). Proper lymphadenectomy and best surgical decisions are directly linked to the surgeon's learning curve[32-38]. Such information contributes to the assertion that specialized treatment for gastric cancer reduces morbidity and mortality related to the treatment and improves survival by allowing better therapeutic strategies[28,33-41]. On the other hand, the argument that there is a negative impact of lymphadenectomy extension on QoL loses strength, making the oncological indication of this procedure prevail. Our finding of statistical non-significance in the correlation between QoL and the number of resected lymph nodes leaves open a discussion about nonsurgical factors (such as research and anatomopathological processing) influencing lymph node count and their possible statistical relationship with other variables and QoL outcomes .
The location of the tumor in the stomach affects QoL in Brazil . Patients who underwent partial gastrectomy (PG) performed better than patients who underwent total gastrectomy. Patients with tumors in the distal region had better scores than patients with proximal tumors . In practice, this result is compatible with literature reports and easily understandable when we realize that the tumor's location in the organ determines the extent of surgical procedure. Proximal tumors indicate the need for total gastrectomy in the search for adequate proximal surgical margins. In contrast, distal tumors can be treated with PG, which is initially less morbid[2-4,24,25,28, . The symptoms and signs that impact QoL can be attributed to nutritional status changes and the remaining gastric reservoir[3,4,45]. In Brazil, according to the FACT-Ga (GaCS items), answers confirming a negative impact were recorded for the following percentages of patients[3]: Bothered by flatulence (46.8%), loss of appetite (42.6%), avoiding going out to eat because of illness (38.5%), a feeling of fullness or heaviness in the stomach area (35.4%), diarrhea (33.3%), feeling tired (32.2%), bothered by a change in eating habits (31.9%), concerned by stomach problems (31%), discomfort or pain when eating (28.1%), losing weight (28.1%), feeling weak all over (28.1%), bothered by reflux or heartburn (26%), discomfort or pain in the stomach area (25%), swelling or cramps in the stomach area (20.1%), and trouble swallowing food (11.4%)[3]. More studies aimed at this point should be designed before the subject is exhausted.
Using different QoL tools can help suggest (or even audit) behaviors such as gastrectomy extension. In the literature, partial gastrectomy has proven to be superior to total gastrectomy in terms of QoL outcomes. Therefore, total gastrectomy (for patients who are more symptomatic in the early post-operative period) should be reserved for patients with oncological needs as long as radical treatment is maintained. In Brazilian studies, partial gastrectomy can be superior when the QoL outcome is analyzed and preferred whenever appropriate . There is also a clear correlation of improvement in QoL scores over the postoperative period . According to the consulted literature, the improvement of QoL in gastrectomy patients starts at three months post-operatively, being marked after six months. According to some authors, patients may have a complete recovery, with the resolution of symptoms resulting from sequelae of the surgery between 12 and 24 mo . In cardia tumor patients, esophagectomy seems to match total gastrectomy QoL scores starting from the sixth month. This positive, temporal correlation between post-operative time and QoL can be used for planning preventive measures for symptom control and rehabilitation; informing patients of this disease behavior pattern can also contribute to better therapy adherence .
Another surgical aspect being highlighted is the reconstruction of intestinal transit. Some authors admit reconstruction to Billroth I; others (such as most in Brazil) prefer and perform the Roux-en-Y. A consensus has not yet been reached regarding the debate on the best reconstruction technique[27, , but the preference for Rouxen-Y finds support, as it offers the best post-operative control of alkaline reflux and its sequelae . This information set leads us to reflect on the need for interdisciplinary, prophylactic intervention in the post-operative period. These strategies can mitigate difficulties, answer questions, and rule out unforeseen events caused by incomplete therapeutic planning. Once again, QoL information can be an essential treatment tool, enabling rational and preventive interference in the complex, multidimensional illness process .
Due to this evidence of QoL changes in the post-operative gastrectomy period, a paradigm shift in cancer care becomes desirable. Interdisciplinarity, specialized assistance, good surgical techniques, therapeutic planning, nutritional assistance, physiotherapy, and psychotherapy, as well as accurate information on the disease evolution pattern, can reduce expectations and increase treatment adherence and results .
The multimodal treatment of cancer confers unparalleled complexity in the interpretation of its effects on QoL. Diverse influences of other therapeutic modalities on QoL outcomes must be meticulously investigated[3,4,55-60].
# Conclusion
The evolution of QoL research allows for statistical analysis and, consequently, more precise and personalized approaches, even for aspects of the disease previously considered abstract. The QoL concept and its measurement tools bring the possibility of using this information in scientific research, therapeutic planning, and healthcare policies . We believe that the development of specialized, interdisciplinary healthcare in oncology should be a priority for improving outcomes. QoL statistical data can support decisions and consolidate or change therapy, generating even more scientific knowledge by including the patient's perceptions. |
The EROP-Moscow oligopeptide database
Natural oligopeptides may regulate nearly all vital processes. To date, the chemical structures of nearly 6000 oligopeptides have been identified from .1000 organisms representing all the biological kingdoms. We have compiled the known physical, chemical and biological properties of these oligopeptides-whether synthesized on ribosomes or by non-ribosomal enzymes-and have constructed an internet-accessible database, EROP-Moscow (Endogenous Regulatory OligoPeptides), which resides at http://erop.inbi.ras.ru. This database enables users to perform rapid searches via many key features of the oligopeptides, and to carry out statistical analysis of all the available information. The database lists only those oligopeptides whose chemical structures have been completely determined (directly or by translation from nucleotide sequences). It provides extensive links with the Swiss-Prot-TrEMBL peptide-protein database, as well as with the PubMed biomedical bibliographic database. EROP-Moscow also contains data on many oligopeptides that are absent from other convenient databases, and is designed for extended use in classifying new natural oligopeptides and for production of novel peptide pharmaceuticals.
# Introduction
For more than a century, natural oligopeptides have attracted scientific attention (1) as biochemical regulators. The very first such oligopeptide, carnosine (b-ala-his), was discovered by Gulevitch and Amiradzhibi in 1900 (2), but its chemical structure was not determined until 1918 [bib_ref] Conserning histidine and carnosine, Baumann [/bib_ref]. Since that time, thousands oligopeptide regulators have been described, and now $500 new natural oligopeptides emerge annually, out of a literature of >20 000 publications each year on oligopeptide chemistry and biology.
Regulatory oligopeptides generally do not exceed $50 amino acid residues (4), and they differ substantially from larger polypeptides (proteins) in their physicochemical and biological properties. Specifically, the smaller peptides rarely possess strong enough intramolecular attractions to form stable globules [bib_ref] Energy characteristics of the structure of protein molecules, Privalov [/bib_ref] , so they are able to shift configurations [bib_ref] X-ray analysis conformation of peptides in the crystalline state, Karle [/bib_ref] and to fit themselves into specific receptor molecules, a process which is further aided by high diffusional mobility [bib_ref] Biophysical problems of oligopeptide regulation, Zamyatnin [/bib_ref].
The terms 'natural' oligopeptide and 'regulatory' oligopeptide can be considered synonymous. Few integrated biological processes are known which are not regulated, or at least modulated, by small peptides. Such roles are especially well known in the regulatory organ systems, viz. nervous, endocrine and immune systems (1), but their functions extend well beyond the bounds of single organ systems or even of single biological species. Antimicrobial oligopeptides produced by prokaryotes themselves, for example, regulate competition for ecological niches and simultaneously function as signaling molecules for species-specific intercellular communication [bib_ref] Facilitation of horizontal transfer of antimicrobial resistance by transformation of antibiotic-induced cell-wall-deficient..., Woo [/bib_ref]. And even eukaryotic oligopeptide toxins seem to play important roles in regulating interspecies reactions [bib_ref] Physicochemical and biological features of endogenous oligopeptide toxins, Zamyatnin [/bib_ref].
It has been clear for >15 years that detailed understanding of the complex regulatory processes involving oligopeptides requires a system for classifying these molecules and for cataloguing their major properties. Our first attempt at such a system, in 1991 (4,10), yielded the MS-DOS version of EROP-Moscow, which contained structures, functions and sources of the oligopeptides then known. That database, however, was not widely accessible to the research community. In the meanwhile, a number of extensive and highly utilized peptide-protein databases have been created (e.g. Swiss-Prot-TrEMBL), but their data on small natural peptides is far from comprehensive and constitutes only a small fraction of their total information, so that retrieving relevant oligopeptide data from them can be laborious and excessively time-consuming period.
Here we present an internet version of the compact specialized database EROP-Moscow, recreated to provide a comprehensive description of all presently known natural oligopeptides. Neuropeptides, peptide hormones, antimicrobial agents and toxins represent the largest functional classes.
*To whom correspondence should be addressed. Tel: +7 095 9543066; Fax: +7 095 9542732; Email: [email protected] This database should now enable investigators to search easily for oligopeptides via a wide variety of different features, to compare their properties quickly, and to retrieve statistics about all relevant information in the database.
## Information sources
Since the objectives of an internet version of EROP-Moscow are to collect and disseminate all essential information about currently known oligopeptides, authenticity is of paramount importance. Therefore, all information in this database has been extracted directly from the primary sources, the great majority of which are publications in scientific journals. More than 100 journals in biochemistry, biophysics, physiology, genetics and general biology are being continuously screened and descriptions of the structures of newly found natural oligopeptides are being retrieved. In many of these articles, the authors compare the novel structures with known ones and provide references to publications not included in our systematic screening. Such publications then become an additional source of primary information for the EROP-Moscow. Finally, initial reports on novel oligopeptides are sometimes found in book chapters, patent descriptions and other proteinpeptide databases, and these sources are used, as well, and are appropriately documented. The total number of useful sources for basic information on natural oligopeptides is now >250.
In addition to the client-server features of EROP-Moscow, a library of publications has been created, containing the actual primary descriptions of oligopeptides, along with their pdf files.
## Selection of oligopeptides
Only those oligopeptides are entered into EROP-Moscow, whose chemical structures have been completely determined (either directly or by translation from nucleotide sequences), and can be described by the standard single-letter amino acid code. Although most peptides included in this version of EROP-Moscow are formed by ribosomal synthesis, a small number formed by non-ribosomal enzymes (11)-mostly from bacteria and fungi-are also included, provided they comprise residues fitting the standard one-letter code. Oligopeptides with still ambiguous structures, such as asparagine/aspartic acid or glutamine/glutamic acid at single residues, have been deliberately excluded from this database, as have artificially synthesized molecules that are not found in nature.
## Organization of erop-moscow
The EROP-Moscow database presents multilevel bioinformation via an HTML-based interface. This interface includes: Home page, Query page, Peptide page, Results page, Family page and Statistics pages [fig_ref] Figure 1: EROP-Moscow architecture illustrated on the Site Map page [/fig_ref]. All of these pages contain internal EROP-Moscow links (including Home page, Site map, Contact us and Help) as well as links to external databases such as Swiss-Prot, Protein Identification Resource (PIR), PDB and PubMed.
The following programming elements, freely available on the basis of the GNU License, have been used as server software, and these are updated as new versions appear.
(i) MySQL database server, version 4.0.14-max; The basic operational unit of EROP-Moscow is an entry (¼ record). Each individual entry describes the physical, chemical and biological characteristics of one unique natural oligopeptide. Each entry is tagged by a unique accession number, beginning with character 'E' (from EROP) followed by five numerical digits. Individual sequences which are found in multiple organisms, even taxonomically remote ones, are presented only once in EROP-Moscow, with the names of all known organisms possessing that oligopeptide being suspended.
On the other hand, oligopeptides existing in two or more distinct chemical modifications (usually accompanied by clear functional differences) are registered as separate entries and are assigned distinct names and accession numbers. Good examples of this are two natural chemical forms of gastrin: one having a simple tyrosine residue and the other, a sulfated tyrosine residue [bib_ref] Structures of human gastrins I and II, Bentley [/bib_ref].
## Home page
Users would normally enter EROP-Moscow via the site address http://erop.inbi.ras.ru. The Home page lists various information about the database itself, including the date of the most recent version, the current number of entries, on-going changes in the content (EROP-news), the list of database authors and some descriptive information. Home page is linked to the Query page and to Statistics pages for individual peptides. A Contact-us button facilitates ordinary E-mail messages and inqueries to the database manager.
## Query page
The Query page, entered from Home page via the 'Query page' button, provides a rapid search for the peptide records signaled by specific characteristics. These characteristics are subdivided into the following groups: general information (such as oligopeptide name or accession number), organismic classification (including multiple trivial species names), physicochemical properties (such as partial amino acid sequences), biochemical or biologic functions and literature references. Query examples (single words, phrases or numbers) are provided adjacent to each Query window, and pull-down menus are provided with most query options. An 'Abbreviations' button, located near the Query window for amino acid sequence, elicits display of the standard oneletter code for amino acid residues, along with optional abbreviations.
Because some oligopeptides, particularly the smallest ones, are chemically modified at the N-and C-termini, six more symbols augment the standard one-letter code. These are:
(i) '+' to denote + H 2 , which is the open N-terminus, (ii) 'b' for an acetyl residue or other chemical group at the N-terminus, (iii) 'À' to denote O À , which is the open C-terminus, (iv) 'z' for an amide bond at the C-terminus, (v) 'J' to denote the pyroglutaminyl linkage, formed by an N-terminal glutamine, owing to side-chain reaction with the terminal amine residue [bib_ref] Glycine-directed peptide amidation: presence in rat brain of two enzymes that convert..., Kizer [/bib_ref] [bib_ref] Nonenzymatic peptide alpha-amidation. Implications for a novel enzyme mechanism, Bateman [/bib_ref] , and finally, (vi) 'U' for the (occasional) aminoisobutyric acid residue.
# Results page
After entering a search word or phrase on the Query page, the user should click on the 'Submit query' button, which initiates the search and returns with the Results page, containing a list of oligopeptides that meet the specified characteristics. Each item in this list will contain the preferred name of oligopeptide, the trivial and taxonomic names of organisms where the peptide has been identified, and the accession number. The accession number, in turn, links to the appropriate Peptide page (record). When the query returns only a single oligopeptide, its record opens immediately.
## Peptide page
This page, reached via accession number, presents the collected data on each oligopeptide, including the number of amino acid residues, primary structure, precursors, known posttranslational modification(s), affinity to any definite structure-function family, taxon(s) of biological sources, tissue/cell localization in each organism, major known biological functions, molecular mass (Da), isoelectric point, pI (calculated and experimentally observed), literature sources and linking accession numbers (if any) in other peptideprotein databases (see above) or PubMed.
## Family page
Tentative homologous family assignments, for each oligopeptide in EROP-Moscow, have been developed by sequence alignment, and the entire family can be reached from the Peptide page via a 'View family' button. Equally located amino acid residues are highlighted in red and the attached oligopeptide name for each sequence links back to the appropriate Peptide page.
## Statistics pages
A special set of pages is devoted to the overall characteristics of data on oligopeptides listed in EROP-Moscow. The starting Statistics page is reached from Home page via the 'EROP Statistics' button, and it contains the list of statistical parameters compiled, each named parameter being a link to one of 15 additional Statistics pages (pp. 1-10.2). These in turn present graphic and tabular information on oligopeptides currently available in EROP-Moscow, information including: This statistical information demonstrates, for example, that the greatest number of natural oligopeptides have been identified in humans, especially in the human brain; to date neuropeptides represent the single largest functional class. The data also show that the most prolific journal for publication of new oligopeptides is the Journal of Biological Chemistry and that an American laboratory, that of J. M. Conlon, leads the discovery of new oligopeptides.
## Service modules
The EROP-Moscow database accesses a set of special software tools for alignment, and for calculating molecular masses and isoelectric points. These operations are executed outside the EROP-Moscow site, and results are returned to EROP-Moscow (as a part of its update capability) and displayed on the Peptide page and the Family page. This system creates dynamic web pages using cgi-scripts written in the PHP language. In response to each appropriate user query, the required HTML-pages are generated interactively. Once the user's web browser has sent the HTTP query to the web server, the required script containing the database query is executed. After the survey of needed records, the PHP script dynamically generates the results, in the form of an HTML-page sent to the user's computer.
Graphic information is also generated and plotted dynamically, e.g. in the statistical processing of data, thus permitting online display of statistical summaries for all natural oligopeptides currently available in the EROP-Moscow database.
## Comparison of erop-moscow with other peptide-protein databases
The internet now provides free access to a large number of both generalized and specialized peptide-protein databases. Best known among the generalized databases are PIR [bib_ref] Protein information resource: a community resource for expert annotation of protein data, Barker [/bib_ref] and Swiss-Prot (Swiss Protein), which is linked with the database of amino acid sequences translated from nucleotides TrEMBL (Translated, European Molecular Biology Laboratory) [bib_ref] The SWISS-PROT protein knowledgebase and its supplement TrEMBL in 2003, Boeckmann [/bib_ref]. Smaller databases, containing information about selected classes of oligopeptides, include Peptaibol (Peptide aminoisobuturic, for data on peptides possessing at least one residue of aminoisobutyric acid) [bib_ref] The peptaibol database: a database for sequences and structures of naturally occurring..., Whitmore [/bib_ref] , ANTIMIC (ANTIMI-Crobial, concerned with antimicrobial peptides) (18) and SCORPION [bib_ref] SCORPION, a molecular database of scorpion toxins, Srinivasan [/bib_ref] , especially created for the peptide-protein toxins from a single order of arachnoids, the scorpions.
About half the natural oligopeptide structures now available in EROP-Moscow, however, are absent from the abovenamed databases, for several reasons, especially (i) that little attention is paid to oligopeptides formed by means of nonribosomal or pure enzymatic synthesis, and (ii) that precursorproduct series are not handled systematically. In particular, very many oligopeptides generated as natural fragments of large precursors are not specifically indexed in the above databases and can be found there only by their amino acid sequences. Swiss-Prot contains one record (P01019), e.g. on human angiotensinogen, which includes the amino acid sequences for angiotensins I and II, but omits angiotensins V and VI [bib_ref] Angiotensin II and its heptapeptide (2-8), hexapeptide (3-8), and pentapeptide (4-8) metabolites..., Semple [/bib_ref]. EROP-Moscow contains these (records E00165 and E00166), as well as the more familiar oligopeptides.
In addition, EROP-Moscow lists a considerable number of oligopeptides with unique amino acid sequences that are simply not included in the other databases-owing either to the source journals (e.g. Biological Bulletin) being out of view of most database managers, or to failure in tracing primary sources keyed from secondary publications.
Information on very short oligopeptides is particularly deficient in the other major databases. displays a useful comparison of these (di-to hepta-) peptides in EROP-Moscow, versus Swiss-Prot.
## Concluding remarks
The database EROP-Moscow has been developed for simple and rapid retrieval of information on natural regulatory oligopeptides and their structurally homologous families. In addition to solving true informational problems, EROP-Moscow can serve as a basis for new research and for elucidating general principles of structural and functional organization for these substances. For example, the current size distribution of oligopeptides, by number of amino acid residues [fig_ref] Figure 3: Distribution of known oligopeptides by amino acid residue number [/fig_ref] , shows a numerical peak $8-10 residues, but this peak has no proper rationale at present. Study of structurally homologous families should facilitate prediction of the functional properties of newly found oligopeptide molecules, should provide bases for classifying newly discovered molecules and in turn should promote creation of novel, highly efficient pharmaceutics derived from the natural regulatory oligopeptides.
Because the discovery of new regulatory oligopeptides is a vigorous and continuing process [fig_ref] Figure 2: Increase in the number of decoded amino acid sequences of the natural... [/fig_ref] , continuous revision and upgrading of EROP-Moscow will be essential. In this cause, we would ask users of EROP-Moscow to alert us to newly discovered oligopeptides which may not yet have been entered into the EROP-Moscow database. For this purpose, the Contact-us button on Home page should be convenient.
## Citing erop-moscow
Users of the EROP-Moscow database are asked to cite this paper, in their relevant published research.
[fig] Ó: The Author 2006. Published by Oxford University Press. All rights reserved. The online version of this article has been published under an open access model. Users are entitled to use, reproduce, disseminate, or display the open access version of this article for non-commercial purposes provided that: the original authorship is properly and fully attributed; the Journal and Oxford University Press are attributed as the original place of publication with the correct citation details given;if an article is subsequently reproduced or disseminated not in its entirety but only in part or as a derivative work this must be clearly indicated. For commercial re-use, please contact [email protected] [/fig]
[fig] Figure 1: EROP-Moscow architecture illustrated on the Site Map page. [/fig]
[fig] Figure 3: Distribution of known oligopeptides by amino acid residue number (Statistics page 2). [/fig]
[fig] Figure 2: Increase in the number of decoded amino acid sequences of the natural oligopeptides (Statistics page 1). [/fig]
|
Infectious Complications in Home Parenteral Nutrition: A Systematic Review and Meta-Analysis Comparing Peripherally-Inserted Central Catheters with Other Central Catheters
## Question 1. described as randomized
Was the study described as randomized? A study does not satisfy quality criteria as randomized simply because the authors call it randomized; however, it is a first step in determining if a study is randomized
## Questions 2 and 3. treatment allocation-two interrelated pieces
Adequate randomization: Randomization is adequate if it occurred according to the play of chance (e.g., computer generated sequence in more recent studies, or random number table in older studies). Inadequate randomization: Randomization is inadequate if there is a preset plan (e.g., alternation where every other subject is assigned to treatment arm or another method of allocation is used, such as time or day of hospital admission or clinic visit, ZIP Code, phone number, etc.). In fact, this is not randomization at all-it is another method of assignment to groups. If assignment is not by the play of chance, then the answer to this question is no. There may be some tricky scenarios that will need to be read carefully and considered for the role of chance in assignment. For example, randomization may occur at the site level, where all individuals at a particular site are assigned to receive treatment or no treatment. This scenario is used for group-randomized trials, which can be truly randomized, but often are "quasi-experimental" studies with comparison groups rather than true control groups. (Few, if any, group-randomized trials are anticipated for this evidence review.)
Allocation concealment: This means that one does not know in advance, or cannot guess accurately, to what group the next person eligible for randomization will be assigned. Methods include sequentially numbered opaque sealed envelopes, numbered or coded containers, central randomization by a coordinating center, computer-generated randomization that is not revealed ahead of time, etc.
## Questions 4 and 5. blinding
Blinding means that one does not know to which group-intervention or control-the participant is assigned. It is also sometimes called "masking." The reviewer assessed whether each of the following was blinded to knowledge of treatment assignment: (1) the person assessing the primary outcome(s) for the study (e.g., taking the measurements such as blood pressure, examining health records for events such as myocardial infarction, reviewing and interpreting test results such as x ray or cardiac catheterization findings);
(2) the person receiving the intervention (e.g., the patient or other study participant); and (3) the person providing the intervention (e.g., the physician, nurse, pharmacist, dietitian, or behavioral interventionist).
Generally placebo-controlled medication studies are blinded to patient, provider, and outcome assessors; behavioral, lifestyle, and surgical studies are examples of studies that are frequently blinded only to the outcome assessors because blinding of the persons providing and receiving the interventions is difficult in these situations. Sometimes the individual providing the intervention is the same person performing the outcome assessment. This was noted when it occurred.
## Question 6. similarity of groups at baseline
This question relates to whether the intervention and control groups have similar baseline characteristics on average especially those characteristics that may affect the intervention or outcomes. The point of randomized trials is to create groups that are as similar as possible except for the intervention(s) being studied in order to compare the effects of the interventions between groups. When reviewers abstracted baseline characteristics, they noted when there was a significant difference between groups. Baseline characteristics for intervention groups are usually presented in a table in the article (often .
Groups can differ at baseline without raising red flags if: (1) the differences would not be expected to have any bearing on the interventions and outcomes; or (2) the differences are not statistically significant. When concerned about baseline difference in groups, reviewers recorded them in the comments section and considered them in their overall determination of the study quality.
## Questions 7 and 8. dropout
"Dropouts" in a clinical trial are individuals for whom there are no end point measurements, often because they dropped out of the study and were lost to followup.
Generally, an acceptable overall dropout rate is considered 20 percent or less of participants who were randomized or allocated into each group. An acceptable differential dropout rate is an absolute difference between groups of 15 percentage points at most (calculated by subtracting the dropout rate of one group minus the dropout rate of the other group). However, these are general rates. Lower overall dropout rates are expected in shorter studies, whereas higher overall dropout rates may be acceptable for studies of longer duration. For example, a 6-month study of weight loss interventions should be expected to have nearly 100 percent followup (almost no dropouts-nearly everybody gets their weight measured regardless of whether or not they actually received the intervention), whereas a 10-year study testing the effects of intensive blood pressure lowering on heart attacks may be acceptable if there is a 20-25 percent dropout rate, especially if the dropout rate between groups was similar. The panels for the NHLBI systematic reviews may set different levels of dropout caps.
Conversely, differential dropout rates are not flexible; there should be a 15 percent cap. If there is a differential dropout rate of 15 percent or higher between arms, then there is a serious potential for bias. This constitutes a fatal flaw, resulting in a poor quality rating for the study.
## Question 9. adherence
Did participants in each treatment group adhere to the protocols for assigned interventions? For example, if Group 1 was assigned to 10 mg/day of Drug A, did most of them take 10 mg/day of Drug A? Another example is a study evaluating the difference between a 30-pound weight loss and a 10-pound weight loss on specific clinical outcomes (e.g., heart attacks), but the 30-pound weight loss group did not achieve its intended weight loss target (e.g., the group only lost 14 pounds on average). A third example is whether a large percentage of participants assigned to one group "crossed over" and got the intervention provided to the other group. A final example is when one group that was assigned to receive a particular drug at a particular dose had a large percentage of participants who did not end up taking the drug or the dose as designed in the protocol.
## Question 10. avoid other interventions
Changes that occur in the study outcomes being assessed should be attributable to the interventions being compared in the study. If study participants receive interventions that are not part of the study protocol and could affect the outcomes being assessed, and they receive these interventions differentially, then there is cause for concern because these interventions could bias results. The following scenario is another example of how bias can occur. In a study comparing two different dietary interventions on serum cholesterol, one group had a significantly higher percentage of participants taking statin drugs than the other group. In this situation, it would be impossible to know if a difference in outcome was due to the dietary intervention or the drugs.
## Question 11. outcome measures assessment
What tools or methods were used to measure the outcomes in the study? Were the tools and methods accurate and reliable-for example, have they been validated, or are they objective? This is important as it indicates the confidence you can have in the reported outcomes. Perhaps even more important is ascertaining that outcomes were assessed in the same manner within and between groups. One example of differing methods is selfreport of dietary salt intake versus urine testing for sodium content (a more reliable and valid assessment method). Another example is using BP measurements taken by practitioners who use their usual methods versus using BP measurements done by individuals trained in a standard approach. Such an approach may include using the same instrument each time and taking an individual's BP multiple times. In each of these cases, the answer to this assessment question would be "no" for the former scenario and "yes" for the latter. In addition, a study in which an intervention group was seen more frequently than the control group, enabling more opportunities to report clinical events, would not be considered reliable and valid.
## Question 12. power calculation
Generally, a study's methods section will address the sample size needed to detect differences in primary outcomes. The current standard is at least 80 percent power to detect a clinically relevant difference in an outcome using a two-sided alpha of 0.05. Often, however, older studies will not report on power.
## Question 13. prespecified outcomes
Investigators should prespecify outcomes reported in a study for hypothesis testingwhich is the reason for conducting an RCT. Without prespecified outcomes, the study may be reporting ad hoc analyses, simply looking for differences supporting desired findings. Investigators also should prespecify subgroups being examined. Most RCTs conduct numerous post hoc analyses as a way of exploring findings and generating additional hypotheses. The intent of this question is to give more weight to reports that are not simply exploratory in nature.
## Question 14. intention-to-treat analysis
Intention-to-treat (ITT) means everybody who was randomized is analyzed according to the original group to which they are assigned. This is an extremely important concept because conducting an ITT analysis preserves the whole reason for doing a randomized trial; that is, to compare groups that differ only in the intervention being tested. When the ITT philosophy is not followed, groups being compared may no longer be the same. In this situation, the study would likely be rated poor. However, if an investigator used another type of analysis that could be viewed as valid, this would be explained in the "other" box on the quality assessment form. Some researchers use a completers analysis (an analysis of only the participants who completed the intervention and the study), which introduces significant potential for bias. Characteristics of participants who do not complete the study are unlikely to be the same as those who do. The likely impact of participants withdrawing from a study treatment must be considered carefully. ITT analysis provides a more conservative (potentially less biased) estimate of effectiveness.
## General guidance for determining the overall quality rating of controlled intervention studies
The questions on the assessment tool were designed to help reviewers focus on the key concepts for evaluating a study's internal validity. They are not intended to create a list that is simply tallied up to arrive at a summary judgment of quality.
Internal validity is the extent to which the results (effects) reported in a study can truly be attributed to the intervention being evaluated and not to flaws in the design or conduct of the study-in other words, the ability for the study to make causal conclusions about the effects of the intervention being tested. Such flaws can increase the risk of bias. Critical appraisal involves considering the risk of potential for allocation bias, measurement bias, or confounding (the mixture of exposures that one cannot tease out from each other). Examples of confounding include co-interventions, differences at baseline in patient characteristics, and other issues addressed in the questions above. High risk of bias translates to a rating of poor quality. Low risk of bias translates to a rating of good quality.
Fatal flaws: If a study has a "fatal flaw," then risk of bias is significant, and the study is of poor quality. Examples of fatal flaws in RCTs include high dropout rates, high differential dropout rates, no ITT analysis or other unsuitable statistical analysis (e.g., completers-only analysis).
Generally, when evaluating a study, one will not see a "fatal flaw;" however, one will find some risk of bias. During training, reviewers were instructed to look for the potential for bias in studies by focusing on the concepts underlying the questions in the tool. For any box checked "no," reviewers were told to ask: "What is the potential risk of bias that may be introduced by this flaw?" That is, does this factor cause one to doubt the results that were reported in the study?
NHLBI staff provided reviewers with background reading on critical appraisal, while emphasizing that the best approach to use is to think about the questions in the tool in determining the potential for bias in a study. The staff also emphasized that each study has specific nuances; therefore, reviewers should familiarize themselves with the key concepts.
Supplemental data
## Guidance for assessing the quality of observational cohort and cross-sectional studies
The guidance document below is organized by question number from the tool for quality assessment of observational cohort and cross-sectional studies.
## Question 1. research question
Did the authors describe their goal in conducting this research? Is it easy to understand what they were looking to find? This issue is important for any scientific paper of any type. Higher quality scientific research explicitly defines a research question.
## Questions 2 and 3. study population
Did the authors describe the group of people from which the study participants were selected or recruited, using demographics, location, and time period? If you were to conduct this study again, would you know who to recruit, from where, and from what time period? Is the cohort population free of the outcomes of interest at the time they were recruited?
An example would be men over 40 years old with type 2 diabetes who began seeking medical care at Phoenix Good Samaritan Hospital between January 1, 1990 and December 31, 1994. In this example, the population is clearly described as: (1) who (men over 40 years old with type 2 diabetes); (2) where (Phoenix Good Samaritan Hospital); and (3) when (between January 1, 1990 and December 31, 1994). Another example is women ages 34 to 59 years of age in 1980 who were in the nursing profession and had no known coronary disease, stroke, cancer, hypercholesterolemia, or diabetes, and were recruited from the 11 most populous States, with contact information obtained from State nursing boards.
In cohort studies, it is crucial that the population at baseline is free of the outcome of interest. For example, the nurses' population above would be an appropriate group in which to study incident coronary disease. This information is usually found either in descriptions of population recruitment, definitions of variables, or inclusion/exclusion criteria.
You may need to look at prior papers on methods in order to make the assessment for this question. Those papers are usually in the reference list.
If fewer than 50% of eligible persons participated in the study, then there is concern that the study population does not adequately represent the target population. This increases the risk of bias.
## Question 4. groups recruited from the same population and uniform eligibility criteria
Were the inclusion and exclusion criteria developed prior to recruitment or selection of the study population? Were the same underlying criteria used for all of the subjects involved? This issue is related to the description of the study population, above, and you may find the information for both of these questions in the same section of the paper.
Most cohort studies begin with the selection of the cohort; participants in this cohort are then measured or evaluated to determine their exposure status. However, some cohort studies may recruit or select exposed participants in a different time or place than unexposed participants, especially retrospective cohort studies-which is when data are obtained from the past (retrospectively), but the analysis examines exposures prior to outcomes. For example, one research question could be whether diabetic men with clinical depression are at higher risk for cardiovascular disease than those without clinical depression. So, diabetic men with depression might be selected from a mental health clinic, while diabetic men without depression might be selected from an internal medicine or endocrinology clinic. This study recruits groups from different clinic populations, so this example would get a "no."
However, the women nurses described in the question above were selected based on the same inclusion/exclusion criteria, so that example would get a "yes."
## Question 5. sample size justification
Did the authors present their reasons for selecting or recruiting the number of people included or analyzed? Do they note or discuss the statistical power of the study? This question is about whether or not the study had enough participants to detect an association if one truly existed.
A paragraph in the methods section of the article may explain the sample size needed to detect a hypothesized difference in outcomes. You may also find a discussion of power in the discussion section (such as the study had 85 percent power to detect a 20 percent increase in the rate of an outcome of interest, with a 2-sided alpha of 0.05). Sometimes estimates of variance and/or estimates of effect size are given, instead of sample size calculations. In any of these cases, the answer would be "yes."
However, observational cohort studies often do not report anything about power or sample sizes because the analyses are exploratory in nature. In this case, the answer would be "no." This is not a "fatal flaw." It just may indicate that attention was not paid to whether the study was sufficiently sized to answer a prespecified question-i.e., it may have been an exploratory, hypothesis-generating study.
## Question 6. exposure assessed prior to outcome measurement
This question is important because, in order to determine whether an exposure causes an outcome, the exposure must come before the outcome.
For some prospective cohort studies, the investigator enrolls the cohort and then determines the exposure status of various members of the cohort (large epidemiological studies like Framingham used this approach). However, for other cohort studies, the cohort is selected based on its exposure status, as in the example above of depressed diabetic men (the exposure being depression). Other examples include a cohort identified by its exposure to fluoridated drinking water and then compared to a cohort living in an area without fluoridated water, or a cohort of military personnel exposed to combat in the Gulf War compared to a cohort of military personnel not deployed in a combat zone.
With either of these types of cohort studies, the cohort is followed forward in time (i.e., prospectively) to assess the outcomes that occurred in the exposed members compared to nonexposed members of the cohort. Therefore, you begin the study in the present by looking at groups that were exposed (or not) to some biological or behavioral factor, intervention, etc., and then you follow them forward in time to examine outcomes. If a cohort study is conducted properly, the answer to this question should be "yes," since the exposure status of members of the cohort was determined at the beginning of the study before the outcomes occurred.
For retrospective cohort studies, the same principal applies. The difference is that, rather than identifying a cohort in the present and following them forward in time, the investigators go back in time (i.e., retrospectively) and select a cohort based on their exposure status in the past and then follow them forward to assess the outcomes that occurred in the exposed and nonexposed cohort members. Because in retrospective cohort studies the exposure and outcomes may have already occurred (it depends on how long they follow the cohort), it is important to make sure that the exposure preceded the outcome.
Sometimes cross-sectional studies are conducted (or cross-sectional analyses of cohortstudy data), where the exposures and outcomes are measured during the same timeframe. As a result, cross-sectional analyses provide weaker evidence than regular cohort studies regarding a potential causal relationship between exposures and outcomes. For crosssectional analyses, the answer to Question 6 should be "no."
## Question 7. sufficient timeframe to see an effect
Did the study allow enough time for a sufficient number of outcomes to occur or be observed, or enough time for an exposure to have a biological effect on an outcome? In the examples given above, if clinical depression has a biological effect on increasing risk for CVD, such an effect may take years. In the other example, if higher dietary sodium increases BP, a short timeframe may be sufficient to assess its association with BP, but a longer timeframe would be needed to examine its association with heart attacks.
The issue of timeframe is important to enable meaningful analysis of the relationships between exposures and outcomes to be conducted. This often requires at least several years, especially when looking at health outcomes, but it depends on the research question and outcomes being examined.
Cross-sectional analyses allow no time to see an effect, since the exposures and outcomes are assessed at the same time, so those would get a "no" response.
## Question 8. different levels of the exposure of interest
If the exposure can be defined as a range (examples: drug dosage, amount of physical activity, amount of sodium consumed), were multiple categories of that exposure assessed? (for example, for drugs: not on the medication, on a low dose, medium dose, high dose; for dietary sodium, higher than average U.S. consumption, lower than recommended consumption, between the two). Sometimes discrete categories of exposure are not used, but instead exposures are measured as continuous variables (for example, mg/day of dietary sodium or BP values).
In any case, studying different levels of exposure (where possible) enables investigators to assess trends or dose-response relationships between exposures and outcomes-e.g., the higher the exposure, the greater the rate of the health outcome. The presence of trends or dose-response relationships lends credibility to the hypothesis of causality between exposure and outcome.
For some exposures, however, this question may not be applicable (e.g., the exposure may be a dichotomous variable like living in a rural setting versus an urban setting, or vaccinated/not vaccinated with a one-time vaccine). If there are only two possible exposures (yes/no), then this question should be given an "NA," and it should not count negatively towards the quality rating.
## Question 9. exposure measures and assessment
Were the exposure measures defined in detail? Were the tools or methods used to measure exposure accurate and reliable-for example, have they been validated or are they objective? This issue is important as it influences confidence in the reported exposures. When exposures are measured with less accuracy or validity, it is harder to see an association between exposure and outcome even if one exists. Also as important is whether the exposures were assessed in the same manner within groups and between groups; if not, bias may result.
For example, retrospective self-report of dietary salt intake is not as valid and reliable as prospectively using a standardized dietary log plus testing participants' urine for sodium content. Another example is measurement of BP, where there may be quite a difference between usual care, where clinicians measure BP however it is done in their practice setting (which can vary considerably), and use of trained BP assessors using standardized equipment (e.g., the same BP device which has been tested and calibrated) and a standardized protocol (e.g., patient is seated for 5 minutes with feet flat on the floor, BP is taken twice in each arm, and all four measurements are averaged). In each of these cases, the former would get a "no" and the latter a "yes."
Here is a final example that illustrates the point about why it is important to assess exposures consistently across all groups: If people with higher BP (exposed cohort) are seen by their providers more frequently than those without elevated BP (nonexposed group), it also increases the chances of detecting and documenting changes in health outcomes, including CVD-related events. Therefore, it may lead to the conclusion that higher BP leads to more CVD events. This may be true, but it could also be due to the fact that the subjects with higher BP were seen more often; thus, more CVD-related events were detected and documented simply because they had more encounters with the health care system. Thus, it could bias the results and lead to an erroneous conclusion.
## Question 10. repeated exposure assessment
Was the exposure for each person measured more than once during the course of the study period? Multiple measurements with the same result increase our confidence that the exposure status was correctly classified. Also, multiple measurements enable investigators to look at changes in exposure over time, for example, people who ate high dietary sodium throughout the followup period, compared to those who started out high then reduced their intake, compared to those who ate low sodium throughout. Once again, this may not be applicable in all cases. In many older studies, exposure was measured only at baseline. However, multiple exposure measurements do result in a stronger study design.
## Question 11. outcome measures
Were the outcomes defined in detail? Were the tools or methods for measuring outcomes accurate and reliable-for example, have they been validated or are they objective? This issue is important because it influences confidence in the validity of study results. Also important is whether the outcomes were assessed in the same manner within groups and between groups.
An example of an outcome measure that is objective, accurate, and reliable is death-the outcome measured with more accuracy than any other. But even with a measure as objective as death, there can be differences in the accuracy and reliability of how death was assessed by the investigators. Did they base it on an autopsy report, death certificate, death registry, or report from a family member? Another example is a study of whether dietary fat intake is related to blood cholesterol level (cholesterol level being the outcome), and the cholesterol level is measured from fasting blood samples that are all sent to the same laboratory. These examples would get a "yes." An example of a "no" would be self-report by subjects that they had a heart attack, or self-report of how much they weigh (if body weight is the outcome of interest).
Similar to the example in Question 9, results may be biased if one group (e.g., people with high BP) is seen more frequently than another group (people with normal BP) because more frequent encounters with the health care system increases the chances of outcomes being detected and documented.
## Question 12. blinding of outcome assessors
Blinding means that outcome assessors did not know whether the participant was exposed or unexposed. It is also sometimes called "masking." The objective is to look for evidence in the article that the person(s) assessing the outcome(s) for the study (for example, examining medical records to determine the outcomes that occurred in the exposed and comparison groups) is masked to the exposure status of the participant. Sometimes the person measuring the exposure is the same person conducting the outcome assessment. In this case, the outcome assessor would most likely not be blinded to exposure status because they also took measurements of exposures. If so, make a note of that in the comments section.
As you assess this criterion, think about whether it is likely that the person(s) doing the outcome assessment would know (or be able to figure out) the exposure status of the study participants. If the answer is no, then blinding is adequate. An example of adequate blinding of the outcome assessors is to create a separate committee, whose members were not involved in the care of the patient and had no information about the study participants' exposure status. The committee would then be provided with copies of participants' medical records, which had been stripped of any potential exposure information or personally identifiable information. The committee would then review the records for prespecified outcomes according to the study protocol. If blinding was not possible, which is sometimes the case, mark "NA" and explain the potential for bias.
## Question 13. followup rate
Higher overall followup rates are always better than lower followup rates, even though higher rates are expected in shorter studies, whereas lower overall followup rates are often seen in studies of longer duration. Usually, an acceptable overall followup rate is considered 80 percent or more of participants whose exposures were measured at baseline. However, this is just a general guideline. For example, a 6-month cohort study examining the relationship between dietary sodium intake and BP level may have over 90 percent followup, but a 20-year cohort study examining effects of sodium intake on stroke may have only a 65 percent followup rate.
## Question 14. statistical analyses
Were key potential confounding variables measured and adjusted for, such as by statistical adjustment for baseline differences? Logistic regression or other regression methods are often used to account for the influence of variables not of interest.
This is a key issue in cohort studies, because statistical analyses need to control for potential confounders, in contrast to an RCT, where the randomization process controls for potential confounders. All key factors that may be associated both with the exposure of interest and the outcome-that are not of interest to the research question-should be controlled for in the analyses.
For example, in a study of the relationship between cardiorespiratory fitness and CVD events (heart attacks and strokes), the study should control for age, BP, blood cholesterol, and body weight, because all of these factors are associated both with low fitness and with CVD events. Well-done cohort studies control for multiple potential confounders.
## Some general guidance for determining the overall quality rating of observational cohort and cross-sectional studies
The questions on the form are designed to help you focus on the key concepts for evaluating the internal validity of a study. They are not intended to create a list that you simply tally up to arrive at a summary judgment of quality.
Internal validity for cohort studies is the extent to which the results reported in the study can truly be attributed to the exposure being evaluated and not to flaws in the design or conduct of the study-in other words, the ability of the study to draw associative conclusions about the effects of the exposures being studied on outcomes. Any such flaws can increase the risk of bias.
Critical appraisal involves considering the risk of potential for selection bias, information bias, measurement bias, or confounding (the mixture of exposures that one cannot tease out from each other). Examples of confounding include co-interventions, differences at baseline in patient characteristics, and other issues throughout the questions above. High risk of bias translates to a rating of poor quality. Low risk of bias translates to a rating of good quality. (Thus, the greater the risk of bias, the lower the quality rating of the study.)
In addition, the more attention in the study design to issues that can help determine whether there is a causal relationship between the exposure and outcome, the higher quality the study. These include exposures occurring prior to outcomes, evaluation of a dose-response gradient, accuracy of measurement of both exposure and outcome, sufficient timeframe to see an effect, and appropriate control for confounding-all concepts reflected in the tool.
Generally, when you evaluate a study, you will not see a "fatal flaw," but you will find some risk of bias. By focusing on the concepts underlying the questions in the quality assessment tool, you should ask yourself about the potential for bias in the study you are critically appraising. For any box where you check "no" you should ask, "What is the potential risk of bias resulting from this flaw in study design or execution?" That is, does this factor cause you to doubt the results that are reported in the study or doubt the ability of the study to accurately assess an association between exposure and outcome?
The best approach is to think about the questions in the tool and how each one tells you something about the potential for bias in a study. The more you familiarize yourself with the key concepts, the more comfortable you will be with critical appraisal. Examples of studies rated good, fair, and poor are useful, but each study must be assessed on its own based on the details that are reported and consideration of the concepts for minimizing bias.
Supplemental data text. References of the full-text articles screened-in but excluded for the meta-analysis, including reason for exclusion :
a. Inpatients, retrospective or pediatric patients. . Touré A, Lauverjat M, Peraldi C, Boncompain-Gerard M, Gelas P, Barnoud D, Chambrier C.
[table] Table 2: The National Institutes of Health Quality Assessment Tool for Observational Cohort and Cross-Sectional Studies. Was the research question or objective in this paper clearly stated?2. Was the study population clearly specified and defined?3. Was the participation rate of eligible persons at least 50%?4. Were all the subjects selected or recruited from the same or similar populations (including the same time period)? Were inclusion and exclusion criteria for being in the study prespecified and applied uniformly to all participants?5. Was a sample size justification, power description, or variance and effect estimates provided?6. For the analyses in this paper, were the exposure(s) of interest measured prior to the outcome(s) being measured? 7. Was the timeframe sufficient so that one could reasonably expect to see an association between exposure and outcome if it existed? 8. For exposures that can vary in amount or level, did the study examine different levels of the exposure as related to the outcome (e.g., categories of exposure, or exposure measured as continuous variable)?9. Were the exposure measures (independent variables) clearly defined, valid, reliable, and implemented consistently across all study participants?10. Was the exposure(s) assessed more than once over time?Additional Comments (If POOR, please state why):*CD, cannot determine; NA, not applicable; NR, not reported [/table]
|
Pitch and Timbre Interfere When Both Are Parametrically Varied
Pitch and timbre perception are both based on the frequency content of sound, but previous perceptual experiments have disagreed about whether these two dimensions are processed independently from each other. We tested the interaction of pitch and timbre variations using sequential comparisons of sound pairs. Listeners judged whether two sequential sounds were identical along the dimension of either pitch or timbre, while the perceptual distances along both dimensions were parametrically manipulated. Pitch and timbre variations perceptually interfered with each other and the degree of interference was modulated by the magnitude of changes along the un-attended dimension. These results show that pitch and timbre are not orthogonal to each other when both are assessed with parametrically controlled variations.
# Introduction
In everyday life, people attend simultaneously to both the pitch and timbre of sounds. In speech, the pitch contour of a sentence carries prosodic information, while timbre variations enable listeners to identify phonemes and vowels necessary for speech segmentation; both contribute to speaker gender and body size perception [bib_ref] The interaction of glottal-pulse rate and vocal-tract length in judgments of speaker..., Smith [/bib_ref]. Both are also central to music perception: pitch typically defines melodies and timbre distinguishes different instruments.
Pitch and timbre have been historically viewed as orthogonal perceptual dimensions [bib_ref] Timbre perception and auditory object identification, Handel [/bib_ref] , but they both are based on spectrotemporal frequency information. Previous studies have produced mixed results on whether these dimensions interfere with each other or not, and thus on whether they may be processed entirely independently [bib_ref] Perceptual space for musical structures, Miller [/bib_ref] [bib_ref] The skill of recognizing musical structures, Beal [/bib_ref] [bib_ref] Effect of frequency, timbre, experience, and feedback on musical tuning skills, Platt [/bib_ref] [bib_ref] Perceptual organization of complex-tone sequences: A tradeoff between pitch and timbre?, Singh [/bib_ref] [bib_ref] HARD and SOFT interacting dimensions: Differential effects of dual context on classification, Melara [/bib_ref] [bib_ref] Interaction among auditory dimensions: Timbre, pitch, and loudness, Melara [/bib_ref] [bib_ref] Perceptual primacy of dimensions: Support for a model of dimensional interaction, Melara [/bib_ref] [bib_ref] Dissociation of pitch from timbre in auditory shortterm memory, Semal [/bib_ref] [bib_ref] Further evidence for an autonomous processing of pitch in auditory short-term memory, Semal [/bib_ref] [bib_ref] Perceptual Interactions Between Musical Pitch and Timbre, Krumhansl [/bib_ref] [bib_ref] Influence of spectral locus and F0 changes on the pitch and timbre..., Singh [/bib_ref] [bib_ref] Perception of pitch and timbre by musically trained and untrained listeners, Pitt [/bib_ref] [bib_ref] Interference effects in short-term memory for timbre, Starr [/bib_ref] [bib_ref] Influence of tonal context and timbral variation on perception of pitch, Warrier [/bib_ref] [bib_ref] The dependency of timbre on fundamental frequency, Marozeau [/bib_ref] [bib_ref] An interval size illusion: The influence of timbre on the perceived size..., Russo [/bib_ref] [bib_ref] The effect of fundamental frequency on the brightness dimension of timbre, Marozeau [/bib_ref].
We hypothesized that these discrepant results might relate to unequal manipulation of the perceptual spacing of pitch and timbre cues, as well as to variations in subjects' previous experience with particular timbres, especially musical timbres. Thus, we sought to assess the interactions between pitch and timbre by parametrically varying them in the threshold-to-suprathreshold range using novel sounds. Listeners attended to one dimension at a time, and judged whether two sounds presented in sequence differed along the attended dimension. Such sequential comparisons are ecologically relevant because meaningful pitch and timbre variations in music and speech streams typically involve serial comparisons over time.
# Methods
## Stimuli
Sounds were synthesized using SIGNAL, a Digital Programming Language (version 5.04.17, Engineering Design, Berkeley, CA). We filtered white noise with a spectrum made up from the sum of 10 Gaussian filters (Equation 1) with constant width and exponentially decaying amplitude, centered at the first 10 harmonics of a fundamental frequency f 0 .
In Equation 1, S is the filtering spectrum, G is an individual Gaussian filter (with unitary amplitude, peak at the central frequency nNf 0 and width s k ) and A is the exponentially decaying weight scaling each Gaussian filter (10 filters in the spectrum, indicated by the index n).
Sounds with different timbres were characterized by differences in their filter width (s k in Equation 1), and sounds with different pitches were characterized by their fundamental frequency f 0 . Two independent experiments, a timbre-rating test (see Timbral step spacing determination below and [fig_ref] Figure 1: Intra-pair pitch and timbre distances [/fig_ref] and a pitch-matching test (Results), proved that filter width variation was related to the perceived timbre (on a continuum from tonal to noise-like) and fundamental frequency variation was related to the perceived pitch of the sounds.
The timbre rating experiment also provided a metric for selecting a set of sounds with different filter widths that were evenly spaced along the timbre dimension and easily discriminable by all subjects. The selected filter widths were s k = 10, Hz, corresponding to sounds defined, for the purposes of this experiment, as one ''timbral step'' apart. The fundamental frequencies were f 0 = 300, 318, 337, 357, 378 Hz, corresponding to sounds one semitone (one experimental ''pitch step'') apart. No rating procedure was required for pitch, because it is known that its perception depends on log-linear frequency ratios between stimuli [bib_ref] Frequency discrimination as a function of frequency and sensation level, Wier [/bib_ref]. It was also not important for the design of this experiment whether timbre steps were of the same ''perceptual size'' as pitch steps: it was only necessary to ensure that, within each dimension, a perceptually-consistent step size was used.
Two sets of sounds were prepared for the two conditions of the task (described below).
For the attend-timbre condition we chose 4 filter widths (s k = 16.9, 25.1, 32.6, 48 Hz) and 3 fundamental frequencies (f 0 = 300, 318, 378 Hz). These values yielded 4 intra-pair timbre distances: 0 timbral steps (same s k for the two sounds), 1 timbral step (pairs with adjacent values of s k in the above list, for example 16.9 and 25.1), 2 timbral steps (pairs with s k values two steps apart, for example 16.9 and 32.6), 3 timbral steps (pairs with s k values 3 steps apart, 16.9 and 48); as well as 4 intra-pair pitch distances: 0 semitones (same f 0 for the 2 sounds), 1 semitone (pairs with f 0 = 300 and f 0 = 318), 3 semitones (pairs with f 0 = 318 and f 0 = 378), and 4 semitones (pairs with f 0 = 300 and f 0 = 378).
For the attend-pitch condition, we chose 4 fundamental frequencies (f 0 = 318, 337, 357, 378 Hz) and 3 filter widths (s k = 16.9, 25.1, 48 Hz), yielding, in the same way as described in the preceding paragraph, 4 intra-pair pitch distances (0, 1, 2, 3 semitones) and 4 intra-pair timbre distances (0, 1, 2, 3 timbral steps).
Thus, for each stimulus set, the dimension relevant for the task (pitch or timbre intra-pair distance) could take 4 levels, while the irrelevant dimension could take 3 levels [fig_ref] Figure 1: Intra-pair pitch and timbre distances [/fig_ref]. All sounds were 500 ms long, including a 50 ms cosine rise and fall. The interval between sounds within a pair was 300 ms.
To obtain the same loudness, the sounds were equated for their RMS value [bib_ref] Evaluation of Objective Loudness Meters, Soloudre [/bib_ref].
## Timbral step spacing determination
A dissimilarity rating task was used to test how naïve listeners perceived the timbre of the experimental sound stimuli at a constant fundamental frequency. The goal was to adjust the width of the Gaussian filters in the sound spectrum (s k , Equation 1), in order to obtain a set of stimuli evenly spaced along the timbre dimension.
Ten sounds were synthesized as described above, with constant fundamental frequency (f 0 = 300 Hz), duration (500 ms) and loudness (equal RMS; [bib_ref] Evaluation of Objective Loudness Meters, Soloudre [/bib_ref] and ten levels of timbre (s k = 10, [bib_ref] Interference effects in short-term memory for timbre, Starr [/bib_ref] [bib_ref] Frequency discrimination as a function of frequency and sensation level, Wier [/bib_ref] [bib_ref] Analysis of covariance using the rank transformation, Conover [/bib_ref] , 50, 55, Equation 1). They were arranged in all possible pairs with 300 ms inter-sound intervals.
Eight subjects (2 men, age range 20-35 years, volunteers) gave written informed consent and reported having no hearing deficits. They were paid 8 J. The subjects first listened to all 10 sounds, presented singly for familiarization. Then they listened to each pair of sounds in random order (6 repetitions) and rated the dissimilarity of the sounds in the pair on a scale from 0 to 9, where 0 meant ''identical'' and 9 meant ''very different''. They were not instructed about, and not asked to focus on, the way in which the sounds varied. Each pair could be heard more than once, with no time limit.
These data were analyzed using custom-written programs in MATLAB R2007a (Version 7.4.0.287, The MathWorks, Natick MA). For each subject, the average over six ratings of the same pair was calculated (pairs made up of the same sounds in a different order counted as the same pair). A linear regression analysis of the average ratings on the timbre intra-pair distances was then computed. Data from one subject were excluded at this stage, because his ratings did not appear to scale with intra-pair distance, as indicated by the squared correlation coefficient (r 2 ) between this subject's rating and the parametric intra-pair distances (r 2 = 0.21). All the other subjects (N = 7) gave ratings that scaled with the parametric intra-pair distances (r 2 .0.63) and were in strong agreement with each other (the average of the intersubjects squared correlation coefficients was 0.70). The judgments of the remaining subjects (N = 7) were averaged and rearranged into a 10610 upper triangular matrix with no diagonal. Multidimensional scalingwas used to reconstruct, from this dissimilarity matrix, a map of the position of the sounds in an acoustic space that could best account for the perceived distances. Since the response scale used here only assumed an ordinal scale of measurement, we chose a non-metric MDS algorithm (Matlab function mdscale: nonmetric scaling with Kruskal's nonmetric stress criterion). We used the values of stress associated with the MDS results to evaluate how well a particular configuration reproduced the observed distance matrix. The MDS analysis suggested that the timbre of the tested sounds was best described by a 2 dimensional space, although 1 dimension was already a good approximation (the stress was 0.026 for 1 dimension and less than 0.001 for 3 dimensions, [fig_ref] Figure 1: Intra-pair pitch and timbre distances [/fig_ref].
These results provided a quantitative representation of the sounds that approximated the ranks of the dissimilarities. As all sounds were equated for duration, intensity and pitch, it was assumed that the dimensions found by the MDS algorithm were inherent to timbre. The goal of this analysis was not to try to interpret what acoustic variables contributed to timbre, so that we did not systematically attempt to determine what acoustic qualities were represented by the dimensions produce by the MDS analysis. Rather, the purpose was to have a systematic, albeit approximate, way of reorganizing the parameters of the sound-generating function in order to produce a set of stimuli that a typical human subject would perceive as evenly-spaced along the timbre dimension. Thus, the MDS solution in the three dimensional space was used to adjust the parameters of the sound-generating function, and obtain a new evenly-spaced timbre series.
In the new series, distances between adjacent parameters s k were proportional to the inverse of the distances between adjacent sounds, as reconstructed by the MDS algorithm. The new series was s k = 10, 13.5, 16.9, 21.0, 25.1, 28.5, 32.6, 39.5, 48.0, 55 Hz. To have adjacent sounds perceptually distinguishable, only five levels were used in the main experiment: s k = 10, 16.9, 25.1, 32.6, 48.0 Hz.
## Procedures
Subjects heard pairs of sounds (binaurally through Sennheiser HD 580 headphones at a constant level of their choice) and indicated whether the pairs were the same or different along the specified dimension -timbre for the first condition (attend-timbre) or pitch for the second condition (attend-pitch)-by pressing a key on the keyboard with no time limit. The sounds in each pair could vary along both dimensions [fig_ref] Figure 1: Intra-pair pitch and timbre distances [/fig_ref].
Both conditions included an initial training phase to familiarize subjects with the task and the terminology, and to select a homogeneous group of participants: those who reached a minimum performance criterion (described below) were admitted to the test phase.
During training, subjects received feedback after judging each pair, and repeated trials in which they made the wrong judgment. After five consecutive correct answers, the subjects took a small exam with no feedback: they judged 10 pairs, from among the most difficult pairs (having close or equal values along the relevant dimension). The criterion to be admitted to the test phase was a score of 7 or more correct answers in this exam.
In the test phase all the possible pairs were presented 6 times in random order, with no feedback.
## Subjects
All procedures were approved by the SISSA Ethics Committee. All participants (75 total, 24 men, age range 18-35 years) gave written informed consent and reported having no hearing deficits. They were paid 8J per hour.
All 75 subjects were tested in the attend-timbre condition in the initial session of these experiments; 57 subjects (21 men) successfully passed the training phase and completed the task. The attend-pitch condition was tested during a second, later session. A subset of 21 subjects (6 men) who had successfully completed the attend-timbre condition was selected blindly with respect to their performance in the attend-timbre condition. They all successfully completed the training session and the task. A minimum of 2 days intervened between the two conditions for each subject.
# Analysis
All analyses were performed using custom-written programs in MATLAB R2007a (Version 7.4.0.287, The MathWorks, Natick MA).
For each participant, the hit and false alarm rates were computed (proportions of correct and incorrect ''different'' responses, respectively) for each intra-pair distance along the attended and un-attended dimension in the test phase (6 repetitions in each condition). Average values are shown in [fig_ref] Table 1: Mean hit rate and false alarm rates for all timbre and pitch... [/fig_ref].
We computed the d' scores of individual participants from the hit and false alarm rates. d' scores measure discriminability, the separation between the means of the signal and the noise distributions, in units of the standard deviation of the noise distribution. A d' score equal to zero indicates that subjects are unable to discriminate between change and no change; a d' score significantly greater than zero means that subjects can detect the difference between change and no change. Once d' scores are significantly different from zero, higher d' scores indicate that a difference is more readily perceived. Group d' scores are shown in [fig_ref] Figure 2: Average sensitivity as a function of pitch and timbre variations [/fig_ref].
Non-parametric statistical analyses were carried out on the d' scores, as the normality assumption was violated for most combinations of timbre and pitch distances (Lilliefors tests, [fig_ref] Table 2: Lilliefors normality test p-values for the d' scores in all conditions [/fig_ref].
Wilcoxon signed-rank testswere used for assessing whether the d' scores were significantly different from zero in all combinations of timbre and pitch distances.
The interaction between pitch and timbre was assessed separately in the attend-timbre and attend-pitch conditions by the Scheirer-Ray-Hare extension of the Kruskal-Wallis test, a non-parametric two-way ANOVA, factors: distance along the attended and un-attended dimension).
Nonparametric Spearman rank correlationswere used for calculating correlation coefficients between the d' scores and the variations along the un-attended dimension [fig_ref] Figure 2: Average sensitivity as a function of pitch and timbre variations [/fig_ref] and a nonparametric ANCOVA based on ranks [bib_ref] Analysis of covariance using the rank transformation, Conover [/bib_ref] was used for testing the hypotheses of equal slopes and intercepts. [fig_ref] Figure 2: Average sensitivity as a function of pitch and timbre variations [/fig_ref] shows the average d' (mean6SE) across subjects (57 subjects in the attend-timbre condition and 21 subjects in the attend-pitch conditions), obtained from the answers given to identical and different stimulus pairs along the attended dimension. In all conditions, the discriminability significantly decreased as the variation in the un-attended dimension grew larger, indicating interference between pitch and timbre (Spearman rank correlations were all negative and significantly different from zero, indicated in [fig_ref] Figure 2: Average sensitivity as a function of pitch and timbre variations [/fig_ref]. Nevertheless, a degree of distinction between timbre and pitch was retained under all interference conditions: the ability of subjects to detect change at significantlygreater-than chance levels was never abolished by variations along the irrelevant dimension (the d' scores in all conditions remained significantly different from zero -Wilcoxon signed-rank tests, all p,0.001). In other words, the variations along the irrelevant dimension never fully determined the subjects' response to the pairs. Significant interference between pitch and timbre, which did not disrupt the subjects' ability to perform the task, was confirmed by nonparametric two-way analysis of varianceand rank analysis of covariance [bib_ref] Analysis of covariance using the rank transformation, Conover [/bib_ref] ; both analyses indicated that the interference was similar for different distances along the attended dimensions.
# Results
The Scheirer-Ray-Hare extension of the Kruskal-Wallis test, a nonparametric two-way analysis of variance, yielded a main effect for the variations along the attended dimension of timbre (attend-timbre: H(2) = 275.64, p,0.001) but not along the attended dimension of pitch (attend-pitch: H(2) = 5.30, p = 0.07) indicating that in the attend-timbre condition, an increase in the timbre distance yielded a substantial increase in discriminability, while in the attend-pitch condition, discriminability was roughly equivalent at all 3 pitch distances (a ''ceiling'' effect due to the fact that a semitone is already an easily discriminable pitch interval). The main effect of the un-attended dimension was significant in both conditions (attend-timbre: H(3) = 82.89, p,0.001; attendpitch: H(3) = 33.11, p,0.001), indicating a clear interference between pitch and timbre both when subjects judged timbre and when they judged pitch. The interaction term was not significant (attend-timbre: H(6) = 3.31, p = 0.77; attend-pitch: H(6) = 0.14, p = 0.99), indicating that the effect of variations along the unattended dimension were consistent across variations along the attended dimension.
The rank analysis of covariance [bib_ref] Analysis of covariance using the rank transformation, Conover [/bib_ref] confirmed these results in both attend-timbre and attend-pitch conditions. We considered the ranked d' values, and tested the linear regression model of 3 lines with equal slopes and different intercepts against the model with different slopes and different intercepts (test one). Since the model with equal slopes and different intercepts was not rejected in both conditions (see below), we tested it against the model with equal slopes and equal intercepts (test two).
In the attend-timbre condition, the first test showed no indication of statistically different slopes (F(2,678) = 1.82, p = 0.163), while the second test showed statistically different intercepts (F(2,680) = 76.42, p,0.001). In other words, 3 different but parallel lines fit the ranked d' values at variable timbre distances. This confirms that an increase in the timbre distance yielded a substantial increase in discriminability, and that variations along the unattended dimension of pitch affected discriminability similarly for all timbre distances.
In the attend-pitch condition, the results for the two tests showed no indication of statistically different slopes (F(2,246) = 1.04, p = 0.353) or intercepts (F(2,248) = 1.36, p = 0.259). In other words, the same line fits the ranked d' values at variable pitch distances. This confirms that performance for different pitch distances was at ''ceiling'', but nevertheless, variations along the unattended dimension of timbre affected pitch discriminability.
The ''step sizes'' for pitch and timbre used in this experiment were perceptually different (the smallest timbre distance was less discriminable than the smallest pitch distance, according Wilcoxon rank sum tests, all of which yielded p,0.001). Indeed, all discriminations in the attend-pitch condition were easier for subjects to perform than in the attend-timbre condition [fig_ref] Figure 2: Average sensitivity as a function of pitch and timbre variations [/fig_ref]. On average, in the attend-pitch condition subjects also completed the training phase faster (26.33 vs. 94.71 average trials; Wilcoxon matched pair sign rank test: z = 24.02, p,0.0001).
To ascertain whether the timbre manipulation used here may have inadvertently produced small but robust directional changes in pitch percepts, a frequency-matching control experiment was conducted. The experimental set-up was the same as the main experiment. Ten subjects (9 women, age range: 20-35 years) were selected from a pool of 15 new volunteers using the same performance criteria as the main experiment. They listened to each sound from the experimental set (5 repetitions, random order) and adjusted the frequency of a pure tone to match it. No time limit was set, and no feedback was given to the subjects about their performance. Four levels of pitch (f 0 = 300, 318, 357, 378) and four levels of timbre (s k = 16.9, 25.1, 32.6 48) were used. For each individual and for each pitch level, the matched tone frequencies were regressed against the timbre levels. There were no consistent directional effects of timbre variations on pitch perception (the slopes of the regression lines were not significantly different from zero, [fig_ref] Table 3: No consistent directional effects of timbre on pure tone matching [/fig_ref]. The variability of the matched tone frequencies (measured as the standard deviation of the matched values for each subject and each target sound) were similarly analyzed using regression [fig_ref] Table 4: No consistent effect of timbre on the variability of pure tone matching [/fig_ref]. Again, there were no consistent effects of variability in the timbre dimension on the variability of pitch percepts. This failure of pitch percepts to appear more difficult as filter width increased was also confirmed by a similar analysis on the number of adjustments taken for each stimulus which also did not reveal any significant variation that depended on timbre variation (data not shown). These results are consistent with previous studies [bib_ref] Effect of frequency, timbre, experience, and feedback on musical tuning skills, Platt [/bib_ref] , [bib_ref] An interval size illusion: The influence of timbre on the perceived size..., Russo [/bib_ref].
# Discussion
A sequential comparison task revealed a consistent interference between pitch and timbre variations for a set of novel sounds, where pitch and timbre were both appropriately quantified and parametrically varied. A group of listeners judged whether two sounds presented in sequence differed along one attended dimension (pitch or timbre) while the second dimension varied in an uncorrelated way. The listeners were always able to perform the task at above-chance levels, but had their judgments consistently biased by variations along the unattended dimension. The possibility that the timbre manipulation procedure may have inadvertently produced pitch changes was empirically ruled out by a frequency-matching control experiment. Alternatively, interactions between pitch and timbre may also be caused by poor selective attention or by a misunderstanding of pitch and timbre concepts on the part of listeners. We believe this was not the case in the present experiments, because of the performance criterion that participants had to meet in order to be included. Indeed, the irrelevant variations never completely dominated the response patterns of subjects: even large pitch variations when the timbre was kept constant did not lead to a complete deterioration of timbre discrimination, indicating that throughout the experiment, the subjects were following instructions and the conceptual distinction between timbre and pitch was retained even under strong interference conditions.
Our results both explain and extend the inconsistent pattern of findings from previous studies using salient timbre manipulations that were not parametrically or perceptually scaled [bib_ref] Perceptual space for musical structures, Miller [/bib_ref] , [bib_ref] The skill of recognizing musical structures, Beal [/bib_ref] , [bib_ref] Influence of spectral locus and F0 changes on the pitch and timbre..., Singh [/bib_ref] , [bib_ref] Perception of pitch and timbre by musically trained and untrained listeners, Pitt [/bib_ref] , [bib_ref] Influence of tonal context and timbral variation on perception of pitch, Warrier [/bib_ref] [bib_ref] The dependency of timbre on fundamental frequency, Marozeau [/bib_ref] [bib_ref] An interval size illusion: The influence of timbre on the perceived size..., Russo [/bib_ref] [bib_ref] The effect of fundamental frequency on the brightness dimension of timbre, Marozeau [/bib_ref] (but see [bib_ref] Interference effects in short-term memory for timbre, Starr [/bib_ref]. These studies mostly focused on musical timbre (recorded or synthesized instrumental sounds) or on brightness, which is one of the major aspects of musical timbre. In some of these studies, salient variations along the irrelevant dimension (instrument identity for timbre or different octaves for pitch) consistently interfered with small variations in the relevant dimension for both pitch and timbre, e.g. [bib_ref] Perceptual space for musical structures, Miller [/bib_ref] , [bib_ref] Influence of spectral locus and F0 changes on the pitch and timbre..., Singh [/bib_ref] , [bib_ref] Influence of tonal context and timbral variation on perception of pitch, Warrier [/bib_ref] but the interaction was inconsistent for larger variations along the relevant dimension [bib_ref] Influence of spectral locus and F0 changes on the pitch and timbre..., Singh [/bib_ref] , [bib_ref] Perception of pitch and timbre by musically trained and untrained listeners, Pitt [/bib_ref] , [bib_ref] An interval size illusion: The influence of timbre on the perceived size..., Russo [/bib_ref] , often depending on the musical training of the subjects (which in turn correlated with perceptual acuity). Unfortunately, because previous studies did not parametrically measure or control the amount of timbre variation they imposed on their stimuli, it is very difficult to meaningfully compare the relative effect sizes seen in previous work with what was found here.
The present study found a symmetrical interaction between pitch and timbre for controlled perceptual step sizes. The timbre variations were supra-threshold but were relatively small in comparison to the typical variations adopted in previous studies. These sounds were reported by listeners to vary along the continuum between noisy and tonal, rather than being categorically different (such as a trumpet vs. a piano). A symmetrical interaction was also previously observed in long sound sequences [bib_ref] Perceptual organization of complex-tone sequences: A tradeoff between pitch and timbre?, Singh [/bib_ref] , [bib_ref] Influence of tonal context and timbral variation on perception of pitch, Warrier [/bib_ref] and in single-sound speeded classification tasks (but only for the speed, and not the accuracy of judgements [bib_ref] HARD and SOFT interacting dimensions: Differential effects of dual context on classification, Melara [/bib_ref] [bib_ref] Interaction among auditory dimensions: Timbre, pitch, and loudness, Melara [/bib_ref] [bib_ref] Perceptual primacy of dimensions: Support for a model of dimensional interaction, Melara [/bib_ref] , [bib_ref] Perceptual Interactions Between Musical Pitch and Timbre, Krumhansl [/bib_ref] , [bib_ref] Perception of pitch and timbre by musically trained and untrained listeners, Pitt [/bib_ref].
The sounds created for this study share similarities with iterated rippled noise (IRN). IRN is constructed by adding a random noise to a copy of itself that is delayed (with a delay d) and attenuated (with a scale factor g) for a number of iterations (n). It corresponds to the environmental case of a sound mixed with its multiple reflections from regularly spaced flat surfaces [bib_ref] A time domain description for the pitch strength of iterated rippled noise, Yost [/bib_ref]. Since IRN has temporal regularities, it possesses a pitch at the reciprocal of the delay d, and a timbre (described variously as tone/noise ratio, pitch strength, or pitch saliency) related to the number of iterations n (more iterations produce a more tonal sound, i.e. a more salient pitch) and to the attenuation g (a bigger attenuation generates a stronger noise percept, and a less salient pitch [bib_ref] The relative strength of the tone and noise components in iterated rippled..., Patterson [/bib_ref] , [bib_ref] Pitch strength and Stevens's power law, Shofner [/bib_ref]. These pitch and timbre attributes can be extracted from the sound spectrum or from the autocorrelation function (which is the Fourier transform of the power spectrum [bib_ref] Pitch strength and Stevens's power law, Shofner [/bib_ref] , [bib_ref] The role of the envelope in processing iterated rippled noise, Yost [/bib_ref].
While the sounds in this study share some of the perceptual attributes of IRN, in that both have a pitch that varies in strength from a noisy to a tonal quality, and both have similar spectra and autocorrelation functions, the sounds used in this study differ from IRN in that the amplitude of their spectral peaks decays exponentially from low frequencies to high frequencies. This minimizes any conflict between spectral and periodicity pitch cues. While the pitch and pitch saliency ( = timbre) of IRN stimuli have been considered as independent percepts [bib_ref] A time domain description for the pitch strength of iterated rippled noise, Yost [/bib_ref] , to our knowledge there is no previous study that has examined the perceptual interaction between pitch and timbre in these sounds. Such testing could reveal a similar effect to the one found here, where, in spite of a clear pitch and identifiable timbral quality of particular sounds, variations along one of the dimensions may become less identifiable when the second dimension also varies.
Interference between pitch and timbre has been interpreted as a conflict based on common mechanisms used for ''spectral'' contributions to pitch and timbre determination [bib_ref] Effect of frequency, timbre, experience, and feedback on musical tuning skills, Platt [/bib_ref] , [bib_ref] Perceptual organization of complex-tone sequences: A tradeoff between pitch and timbre?, Singh [/bib_ref] , [bib_ref] Perceptual primacy of dimensions: Support for a model of dimensional interaction, Melara [/bib_ref] , [bib_ref] Influence of spectral locus and F0 changes on the pitch and timbre..., Singh [/bib_ref] , [bib_ref] An interval size illusion: The influence of timbre on the perceived size..., Russo [/bib_ref] , [bib_ref] The effect of fundamental frequency on the brightness dimension of timbre, Marozeau [/bib_ref]. In pitch perception, spectral information is presumably combined with periodicity information. The timbre manipulation used in this study minimizes the conflict between spectral and periodicity pitch, since the fundamental always has the maximum amplitude. Future experiments examining pitchtimbre interference at small perceptual distances for sounds where spectral and temporal mechanisms make unequal contributions to pitch perception could provide a fuller understanding of the joint contributions of spectral and temporal information to both pitch and timbre processing.
In summary, a consistent interference between pitch and timbre can be identified when timbre and pitch are appropriately quantified and parametrically varied. The interference does not abolish the distinction between pitch and timbre, but variations along the un-attended dimension make listeners less certain about whether they heard variation in the attended dimension. Supporting Information Author Contributions
[fig] Figure 1: Intra-pair pitch and timbre distances. (A) attend-timbre condition, (B) attend-pitch condition. For each pair of sounds the timbre and pitch variations are indicated in the graphs. For each single sound in the pairs, pitch is related to the fundamental frequency, f 0 , and timbre is related to the width of the Gaussian filter in the spectrum, s k (Equation1). The intra-pair distances in (A), attend-timbre condition, are obtained with sounds parameters: f 0 = 300, 318, 378 Hz, s k = 16.9, 25.1, 32.6, 48 Hz. The intra-pair distances in (B), attend-pitch condition, are obtained with sounds parameters: f 0 = 318, 337, 357, 378 Hz, s k = 16.9, 25.1, 48 Hz. doi:10.1371/journal.pone.0087065.g001 [/fig]
[fig] Figure 2: Average sensitivity as a function of pitch and timbre variations. Values represent average d' score (6SE) across subjects, for each intra-pair distance along the attended dimension (1step circle, 2 steps diamond or 3 steps triangle) and un-attended dimension (on the x-axis). (A) attend-timbre condition, n = 57 for each data point; (B) attend-pitch condition, n = 21 for each data point. r indicates the Spearman rank correlations coefficient between the d' scores and the variations along the un-attended dimension, p indicates the p-values for the significance test. doi:10.1371/journal.pone.0087065.g002 [/fig]
[fig] Figure S1: Results of the multidimensional scaling (MDS). (a) Stress as a function of the dimensionality of the MDS solution; (b) three-dimensional spatial representation based on the dissimilarity ratings for 10 sounds. (EPS) [/fig]
[table] Table 1: Mean hit rate and false alarm rates for all timbre and pitch distances in the two conditions. [/table]
[table] Table 2: Lilliefors normality test p-values for the d' scores in all conditions. [/table]
[table] Table 3: No consistent directional effects of timbre on pure tone matching. [/table]
[table] Table 4: No consistent effect of timbre on the variability of pure tone matching. Note: The table shows uncorrected p-values for the slopes of the individual regressions of the standard deviation of the matched tones on timbre levels for each fundamental frequency f 0 . Uncorrected p-values are shown in this instance because these are the more conservative alternative when making the argument of no consistent directional effects. p-values .0.05 mean that the hypothesis of equal slopes cannot be rejected. S1-10 indicates subjects from 1 to 10. *uncorrected p-value ,0.05. doi:10.1371/journal.pone.0087065.t004 [/table]
|
Promoter- and cell-specific epigenetic regulation of CD44, Cyclin D2, GLIPR1 and PTEN by Methyl-CpG binding proteins and histone modifications
Background :The aim of the current study was to analyze the involvement of methyl-CpG binding proteins (MBDs) and histone modifications on the regulation of CD44, Cyclin D2, GLIPR1 and PTEN in different cellular contexts such as the prostate cancer cells DU145 and LNCaP, and the breast cancer cells MCF-7. Since global chromatin changes have been shown to occur in tumours and regions of tumour-associated genes are affected by epigenetic modifications, these may constitute important regulatory mechanisms for the pathogenesis of malignant transformation.Methods :In DU145, LNCaP and MCF-7 cells mRNA expression levels of CD44, Cyclin D2, GLIPR1 and PTEN were determined by quantitative RT-PCR at the basal status as well as after treatment with demethylating agent 5-aza-2'deoxycytidine and/or histone deacetylase inhibitor Trichostatin A. Furthermore, genomic DNA was bisulfite-converted and sequenced. Chromatin immunoprecipitation was performed with the stimulated and unstimulated cells using antibodies for MBD1, MBD2 and MeCP2 as well as 17 different histone antibodies.Results :Comparison of the different promoters showed that MeCP2 and MBD2a repressed promoter-specifically Cyclin D2 in all cell lines, whereas in MCF-7 cells MeCP2 repressed cell-specifically all methylated promoters. Chromatin immunoprecipitation showed that all methylated promoters associated with at least one MBD. Treatment of the cells by the demethylating agent 5-aza-2'-deoxycytidine (5-aza-CdR) caused dissociation of the MBDs from the promoters. Only MBD1v1 bound and repressed methylation-independently all promoters. Real-time amplification of DNA immunoprecipitated by 17 different antibodies showed a preferential enrichment for methylated lysine of histone H3 (H3K4me1, H3K4me2 and H3K4me3) at the particular promoters. Notably, the silent promoters were associated with unmodified histones which were acetylated following treatment by 5-aza-CdR.Conclusions :This study is one of the first to reveal the histone code and MBD profile at the promoters of CD44, Cyclin D2, GLIPR1 and PTEN in different tumour cells and associated changes after stimulation with methylation inhibitor 5aza-CdR.
# Background
Global chromatin changes have been shown to occur in tumours. In chromosomal regions of tumour-associated genes epigenetic modifications may constitute important regulatory mechanisms for the pathogenesis of malignant transformation [bib_ref] Genomic DNA methylation: the mark and its mediators, Klose [/bib_ref]. Inactivation of tumour suppressor genes by promoter hypermethylation has been reported for diverse tumours and is thought to play a crucial role in carcinogenesis [bib_ref] Cancer as an epigenetic disease: DNA methylation and chromatin alterations in human..., Esteller [/bib_ref]. DNA methylation affects mainly the cytosine base in a CpG dinucleotide, which is found isolated or clustered in so called CpG islands, and may induce gene repression by inhibiting the access of transcription factors to their binding sites, and by recruiting methyl-CpG binding proteins (MBDs) to methylated DNA together with histone modifications [bib_ref] The methyl-CpG binding domain and the evolving role of DNA methylation in..., Hendrich [/bib_ref].
To date, five MBDs have been identified: MBD1, MBD2, MBD3, MBD4 and MeCP2. These proteins are implicated in the transcriptional repression of methylated DNA [bib_ref] Identification and characterization of a family of mammalian methyl-CpG binding proteins, Hendrich [/bib_ref] [bib_ref] The thymine DNA glycosylase MBD4 represses transcription and is associated with methylated..., Kondo [/bib_ref]. With the exception of MBD4, belonging to the uracil DNA glycosylase superfamily [bib_ref] The thymine DNA glycosylase MBD4 represses transcription and is associated with methylated..., Kondo [/bib_ref] , the members of the family associate with histone deacetylases (HDACs). MBD1 is alternatively spliced to produce five protein isoforms (PCM1, MBD1v1, MBD1v2, MBD1v3 and MBD1v4) which differ in the number of cysteine-rich (CXXC) domains and the carboxyl-terminal sequence. Although repression of unmethylated genes has been reported to depend on the third CXXC domain [bib_ref] Mbd1 is recruited to both methylated and nonmethylated CpGs via distinct DNA..., Jorgensen [/bib_ref] , recent findings indicate that the two other CXXC domains may also contribute to the repression of unmethylated promoters, however, with a weaker affinity [bib_ref] Methyl-CpG binding domain proteins and their involvement in the regulation of the..., Wischnewski [/bib_ref]. Two isoforms of MBD2 are known: MBD2a and MBD2b. The shorter form, MBD2b, starting at the second methionine lacks the N-terminal sequence of MBD2a [bib_ref] Methyl-CpG-binding protein 2 represses LINE-1 expression and retrotransposition but not Alu transcription, Yu [/bib_ref]. MBD2a may act either as an activator or a repressor of transcription [bib_ref] Methyl-CpG binding domain proteins and their involvement in the regulation of the..., Wischnewski [/bib_ref] [bib_ref] Methyl-CpG-binding protein 2 represses LINE-1 expression and retrotransposition but not Alu transcription, Yu [/bib_ref] [bib_ref] Antithetic effects of MBD2a on gene regulation, Fujita [/bib_ref] [bib_ref] MBD2 is a transcriptional repressor belonging to the MeCP1 histone deacetylase complex, Ng [/bib_ref].
Epigenetic modifications include not only methylation of DNA but also configurational changes in chromatin which are implicated in transcriptional regulation, as well. The N-terminal tails of histones are subject to posttranslational modifications, such as acetylation, phosphorylation, ubiquitination and methylation. Histone acetylation may be a predominant mark in active chromatin regions, and acetyl groups are removed by HDACs. Methylation of the lysine residue 4 of histone H3 (H3K4) is highly conserved and associated with transcriptionally active genes. Methylation of the lysine residue 9 of histone H3 (H3K9) recruits the heterochromatin protein HP-1, which condenses the chromatin into an inactive conformation. Both, DNA methylation and histone modifications may be linked by MBDs. Nearly all members of the family can interact with histone methyltransferases and deacetylases [bib_ref] Chemical regulation of epigenetic modifications: Opportunities for new cancer therapy, Zheng [/bib_ref].
Tumour invasion is accompanied by migration of malignant cells into the surrounding connective tissue [bib_ref] Dissecting the metastatic cascade, Pantel [/bib_ref]. Alterations in cell-cell and cell-matrix interactions are involved in this process. CD44 is a glycoprotein and main receptor for hyaluronic acid, collagen, fibronectin and osteopontin, and regulates the cytoskeleton by transduction of signals from the extracellular matrix. Moreover, CD44 is involved in leukocyte binding to vascular endothelium at sites of inflammation [bib_ref] The hyaluronan receptor, CD44, Isacke [/bib_ref]. Numerous isoforms of CD44 exist, and some of them are overexpressed on breast tumour cells which seems to be correlated with the metastatic potential [bib_ref] CD44 crosslinking-mediated matrix metalloproteinase-9 relocation in breast tumor cells leads to enhanced..., Peng [/bib_ref]. Furthermore, the phenotype of breast tumour cells showed that CD44 may distinguish tumour-initiating from non-tumourigenic cells [bib_ref] Prospective identification of tumorigenic breast cancer cells, Al-Hajj [/bib_ref]. Recent experimental and clinical investigations showed that CD44 together with heparanase and hyaluronan regulates tumour cell proliferation, migra-tion, invasion and angiogenesis and associates with breast cancer patient survival [bib_ref] Heparanase, hyaluronan, and CD44 in cancers: a breast carcinoma perspective, Gotte [/bib_ref]. In respect to the down-regulation of CD44 during progression and metastasis of prostate cancer, CD44 is a metastasis suppressor for this tumour type [bib_ref] Hypermethylation of the CD44 gene is associated with progression and metastasis of..., Kito [/bib_ref]. Aberrant promoter hypermethylation has been described for CD44 gene silencing [bib_ref] Silencing of CD44 expression in prostate cancer by hypermethylation of the CD44..., Verkaik [/bib_ref].
The D-type cyclins D1, D2 and D3 and their associated cyclin-dependent kinases are critical components for cell proliferation. They are expressed during the cell cycle at G0/G1-S-transition. Cyclin D2, implicated in cell differentiation and malignant transformation, is inactivated by promoter hypermethylation in several human cancers. High DNA methylation levels of Cyclin D2 cause deregulation of the G1/S checkpoint, and correlate with clinicopathologic features of tumour aggressiveness in breast and prostate cancer [bib_ref] Promoter hypermethylation of p16INK4A, p14ARF, CyclinD2 and Slit2 in serum and tumor..., Sharma [/bib_ref] [bib_ref] Hypermethylation of Cyclin D2 is associated with loss of mRNA expression and..., Henrique [/bib_ref]. GLIPR1 (glioma pathogenesis-related protein 1) is a novel p53-target gene [bib_ref] RTVP-1, a novel human gene with sequence similarity to genes of diverse..., Rich [/bib_ref] cloned from human glioblastoma cell lines and its expression in astrocytic tumours correlated with tumour grade [bib_ref] Related to testes-specific, vespid, and pathogenesis protein-1 (RTVP-1) is overexpressed in gliomas..., Rosenzweig [/bib_ref]. In contrast to its oncogenic effect in glioma, where GLIPR1 regulates proliferation, migration and survival of glioma cells, it acts as a tumour suppressor in prostate cancer. Down-regulation in this context appears to be caused by epigenetic rather than genetic changes [bib_ref] RTVP-1, a tumor suppressor inactivated by methylation in prostate cancer, Ren [/bib_ref].
PTEN (phosphatase and tensin homologue) is a wellknown tumour suppressor that inhibits cell proliferation and migration by antagonizing the phosphatidylinositol 3-kinase (PI3K) signaling pathway [bib_ref] PTEN, more than the AKT pathway, Blanco-Aparicio [/bib_ref]. In many primary and metastatic human tumours PTEN is inactivated by mutations, deletions or promoter hypermethylation [bib_ref] The phosphatidyl inositol 3-kinase signaling network: implications for human breast cancer, Dillon [/bib_ref] [bib_ref] Akt-regulated pathways in prostate cancer, Majumder [/bib_ref].
In the present study, the promoters of the four abovedescribed tumour-associated genes (CD44, Cyclin D2, GLIPR1 and PTEN) were examined for methylationdependent gene regulation, the participation of MBDs in gene silencing and the histone modifications associated with the respective promoter areas. Comparison of the settings at the promoters among each other and between different cellular contexts show a MBD-mediated promoter-and cell-specific repression of the four genes. Our data provide new insights on the histone signature at the promoters of these genes, and deliver valuable information on their epigenetic regulatory mechanism.
# Methods
## Cell culture
All cell lines were obtained from ATCC. Prostate carcinoma cells DU145 and LNCaP, and breast adenocarcinoma cells MCF-7 were maintained in Dulbecco's modified Eagle's medium (DMEM, Invitrogen, Karlsruhe, Germany), supplemented with 10% fetal calf serum (FCS) and 2 mM L-glutamine (Invitrogen) and cultured under standard conditions (37°C, 5% CO 2 , humidified atmo-sphere). Cell viability was determined by trypan blue staining. Each cell line was stimulated by 5-aza-2'-deoxycytidine (5-aza-CdR, f.c. 1 μM, Sigma-Aldrich, Taufkirchen, Germany) for 72 h. 5-aza-CdR-treated cells or a mock control were stimulated by Trichostatin A (TSA, f.c. 500 nM, Sigma-Aldrich) for the last 24 h of the 72 h incubation.
# Mrna expression analysis
To determine the mRNA expression of CD44, Cyclin D2, GLIPR1 and PTEN, total RNA was extracted from DU145, LNCaP and MCF-7 cells using the RNeasy ® Mini Kit (Qiagen, Hilden, Germany) according to the manufacturer's description. Synthesis of cDNA was carried out using the SuperScript First strand System with random hexamer primers (Invitrogen). PCR amplification of cDNA was performed with primers specific for CD44: 5' (forward) GTGATCAACAGTGGCAATGGA and 3' (reverse) TCACCAAATGCACCATTTCCT (PCR product 94 bp), Cyclin D2: 5' TGGGGAAGTTGAAGTG-GAAC and 3' ATCATCGACGGTGGGTACAT (175 bp), GLIPR1: 5' TGCCAGTTTTCACATAATACAC and 3' GGATTTCGTCATACCAGTTT (142 bp), PTEN: 5' TTGAAGACCATAACCCACCACAG and 3' GGCA-GACCACAAACTGAGGATTG (387 bp), β-Actin: 5' GGCGGCACCAGCATGTACCCT and 3' AGGGGC-CGGACTGGTCATACT (202 bp). The reaction was performed in a final volume of 20 μl containing PCR Buffer (Qiagen), 200 μM of each dNTP (Roche Applied Science, Mannheim, Germany), 0.5 μM of each primer and 2.5 units of Taq polymerase (Qiagen). After a PCR run for 30 cycles on a Peltier Thermal Cycler (PTC-200, Biozym, Oldendorf, Germany), the PCR products were electrophoretically separated on a 1% agarose gel.
## Bisulfit genomic sequencing
Genomic DNA was isolated from the cultured cells using the QIAamp DNA Mini Kit (Qiagen) according to the manufacturer's description. Approximately 0.5-1 μg genomic DNA was bisulfite-converted and purified according to the recommended protocol of the EpiTect Bisulfite Kit (Qiagen). One μl of converted DNA was amplified and sequenced by the following primers: CD44 5' (forward) TGTGAAATTTAGAGATTTTGTTTTAG and 3' (reverse) AAATTTTAAAAAATAACAACCCTC CC, Cyclin D2 5' GGGTTAGTTGTTGTTTTTTTTAA TAA and 3' AAAAAAATTTTTCTATTTTTATTTTT, GLIPR1 5' TTATTATGTGTTGATATG ATTTTAAA AAG, and 3' AACCCACAACTTTACAAACCTAACC, PTEN 5' GTTTTTTTTGAAAGGGAAGGTG and 3' CAAACCCCCTCCCTAAAACTA. Sequencing amplification was run using BigDye reagent and buffer (Amersham Biosciences, Freiburg, Germany). After ethanol precipitation of the PCR products the pellets were resus-pended by HiDi formamide (Applied Biosystems) and sequenced on a Genetic Analyzer 3130 (Applied Biosystems).
## Mbd protein expression analysis
Protein levels of MBD1, MBD2 and MeCP2 in basal and stimulated DU145, LNCaP and MCF-7 cells as well as MBD1 knock out (MBD1 -/-) mouse embryonic fibroblasts (MEF) were investigated by Western blot analysis as recently described using nuclear extracts and antibodies recognizing these epigenetic factors [bib_ref] Methyl-CpG binding domain proteins and their involvement in the regulation of the..., Wischnewski [/bib_ref].
To test the specificity of the antibodies against modified histones, acidic protein extraction was performed according to a special protocol. Briefly, cells were resuspended in TEB buffer (PBS containing 0.5% v/v Triton X100 and 2 mM PMSF) and incubated for 10 min on ice. Acidic extraction was carried out in 0.2 N HCl in a rotation shaker at 4°C over night. The histone protein content in the supernatant was measured according to Bradford.
Twenty-five μg of protein extracts were separated by a 12% SDS polyacrylamide gel and transferred onto the nitrocellulose membrane Hybond-C extra (Amersham). After blocking, the membrane was probed with a 1:500 or 1:1000 dilution of antibodies which are directed against the following human proteins: MBD1, MeCP2 (Abcam, Cambridge, UK), MBD2, monomethylated lysine 9 of histone H3 (H3K9me), dimethylated lysine 9 of histone H3 (H3K9me2), trimethylated lysine 9 of histone H3 (H3K9 me3), monomethylated lysine 4 of histone H3 (H3K4me), dimethylated lysine 4 of histone H3 (H3K4me2), trimethylated lysine 4 of histone H3 (H3K4me3), monomethylated lysine 20 of histone H4 (H4K20me), dimethylated lysine 20 of histone H4 (H4K20me2), trimethylated lysine 20 of histone H4 (H4K20me3) (Millipore, Schwalbach, Germany).
Detection of the proteins was carried out using a peroxidase-conjugated secondary antibody (Sigma-Aldrich) and the chemiluminescence ECL detection Kit (Amersham).
## Construction of plasmids
Reporter plasmids were constructed by cloning CD44 (-908/-118), Cyclin D2 (-507/-30), GLIPR1 (-521/-142) and PTEN (-737/-41) promoter fragments into the XhoI and HindIII sites of the pGL2-Luciferase reporter plasmid (Promega, Mannheim, Germany). For the assay targeting the Gal4-linked transcriptional repressor domain (TRD) of the MBDs to the Gal4 DNA-recognition motif, five Gal4 sequences were inserted into the MluI and XhoI sites directly upstream of the cloned promoter fragments. All clones were verified by enzymatic digestion and DNA sequencing.
The construction of the expression plasmids encoding the full length proteins MBD1 (isoforms MBD1v1 and MBD1v3), MBD2 (isoforms MBD2a and MBD2b) or MeCP2, and the expression plasmids containing sequences encoding a fusion protein consisting of the Gal4 DNA binding domain and the TRD of MBD1 (amino acids 383-605, MBD1-TRD), MBD2 (45-262, MBD2-TRD) or MeCP2 (196-486, MeCP2-TRD) has been previously described [bib_ref] Methyl-CpG binding domain proteins and their involvement in the regulation of the..., Wischnewski [/bib_ref] [bib_ref] Methyl-CpG-binding protein 2 represses LINE-1 expression and retrotransposition but not Alu transcription, Yu [/bib_ref].
## In vitro methylation of plasmid dna
Twenty μg reporter plasmid containing CD44, Cyclin D2, GLIPR1 and PTEN promoter fragments were methylated by the HpaII or SssI methylase (New England Biolabs, NEB) for 4 h at 37°C using the methyl donor SAM (S-Adenosyl methionine, NEB). Efficient and complete methylation of the plasmid DNA was confirmed by its resistance to digestion with the methylation-sensitive restriction enzyme HpaII. A control digestion with the isoschizomer MspI was performed.
## Transfection and luciferase reporter assay
The DU145, LNCaP and MCF-7 cells as well as the MBD1 -/mouse embryonic fibroblasts [bib_ref] Methyl-CpG binding domain proteins and their involvement in the regulation of the..., Wischnewski [/bib_ref] were transiently transfected with 0.5 μg of reporter plasmids and expression plasmids using the FuGENE Reagent (Roche Applied Science). For efficiency control 0.2 μg of a vector encoding for the Renilla Luciferase (Promega, Mannheim, Germany) was co-transfected. After 48 h incubation, the transfected cells were lysed using the Dual-Luciferase Reporter Assay System (Promega). Promoterdriven luciferase activity was measured on a 20/20 n Luminometer (Turner Biosystems, Sunnyvale, USA) and normalized by the Renilla Luciferase activity. Each transfection experiment was carried out in triplicate wells and repeated at least twice.
## Chromatin immunoprecipitation assay (chip)
DU145, LNCaP and MCF-7 cells at 80% confluence were fixed with 1% formaldehyde in minimal medium for 10 min at room temperature. The DNA/protein cross linking reaction was stopped by adding a glycine stop-fix solution. The cells were washed with ice cold PBS, scraped and pelleted by centrifugation for 10 min, 4°C, at 720 g. Cells were lysed with a hypotonic lysis buffer, and the nuclei were pelleted by centrifugation for 10 min, 4°C, at 2400 g. The nuclei pellet was sheared in 1 ml shearing buffer by sonication at 25% power for 4 min on ice (Sonicator UP50H, Dr. Hielscher GmbH, Teltow, Germany) to chromatin fragment lengths of 200 to 1000 bp. The stopfix solution, hypotonic lysis buffer and shearing buffer were obtained from the ChIP-IT kit (Active Motif, Rixensart, Belgium). The chromatin extract was pre-cleared with protein G beads. 170 μl aliquots of the supernatant were immunoprecipitated using 3 μg of the antibodies specific for IgG (Active Motif ), the antibodies against acetylated histones H2A (K5), H2B (K12), H3 (K9), H4 (K8) and non-modified histones H2A, H2B, H3, H4 (Cell Signaling, Danvers, USA), and the antibodies as described above, overnight at 4°C. The DNA/protein/ antibody complexes were incubated with protein G beads for 2 h at 4°C. After washing the beads, the immunoprecipitated DNA was eluted from the beads by 100 μl elution buffer containing 1% SDS and 50 mM NaHCO 3 for 15 min, and protein-DNA crosslinks were reversed with 200 mM NaCl by incubation at 65°C for 4 h. Digestion of the proteins was performed with 0.1 mM EDTA, 20 mM Tris-HCl pH 6.5 and 2 μl Proteinase K solution (Active Motif ) for 2 h at 42°C. The DNA was purified by minicolumns (Active Motif ).
## Quantitative real-time pcr
The immunoprecipitated DNA fragments were amplified by the following primer pairs: To quantify the mRNA expression in the cell lines the following primers were used: CD44 5' CCCAGATGGA-GAAAGCTCTG and 3' GTTGTTTGCTGCACAGAT GG (113 bp), Cyclin D2 5' TTCCGCAGTGC TCCT ACTTC and 3' CGCACTTCTGTTCCTCACAG (105 bp), GLIPR1 5' CTGTGGCCACTACACTCAGG and 3' AGAGCGTCAAAGCCAGAAAC (95 bp), PTEN 5' CCCAGACATGACAGCCATC and 3' TCTGCAGGAA ATCCCATAGC (126 bp), RPLP0 5' ACCCAGCTCTG-GAGAAACTGC and 3' TGAGGTCCTC CTTGGT-GAACA (72 bp). The PCR reaction contained 2 μl template, 7.5 μl SYBRGreen Mastermix (Qiagen), and 4 pmol primer sets (Sigma-Aldrich, München, Germany) in a final volume of 15 μl, and was carried out at a melting temperature of 60°C and in 45 cycles on a Realplex System (Mastercycler epgradient S, Eppendorf, Hamburg, Germany). A dilution series of 10, 2.5, 1.25, 0.3125 and 0.078 ng/μl template DNA served as internal standard for quantification. All experiments were done in triplicate and each PCR was repeated at least twice. Evaluation of the data was performed by the Realplex software.
# Statistical analysis
The statistical analyses were performed using the SPSS software package, version 13.0 (SPSS Inc. Chicago, IL). Student's t-test and Fisher's exact test were used to identify possible statistical differences in activation and repression of gene expression, binding affinities of MBDs and histone modifications between basal and stimulated cell lines. Analysis of variance (ANOVA) was performed to determine if the means of several groups are likely to be equal.
The analyses were explorative and generated hypotheses that have to be validated in further studies. Therefore, no adjustment for multiple testing like a Bonferroni correction was performed.
The diagrams are based on the mean values of measured values. The error bars represent the standard deviation (STDEV).
# Results
## Mrna expression of cd44, cyclin d2, glipr1 and pten in du145, lncap and mcf-7 cells prior to and after treatment with 5-aza-2'-deoxycytidine and/or trichostatin a
Basal mRNA expression of CD44, Cyclin D2, GLIPR1 and PTEN was determined by reverse transcription-PCR and quantitative real-time PCR using gene-specific primers.
To measure the influence of DNA methylation and histone deacetylation the cells were also incubated with the demethylating agent 5-aza-2'-deoxycytidine (5-aza-CdR) and/or the histone deacetylase inhibitor Trichostatin A (TSA). In [fig_ref] Figure 1: mRNA expression of CD44, Cyclin D2, GLIPR1 and PTEN in DU145, LNCaP... [/fig_ref] , the basal mRNA levels in the cell lines are depicted in comparison to the cells treated with 5aza-CdR and/or TSA.
CD44 was constitutively expressed in LNCaP cells, and 5-aza-CdR and TSA had no further influence on the expression level. In contrast, in DU145 and MCF-7 cells the transcription of CD44 could be significantly up-regulated by 5-aza-CdR, whereas TSA had no activating effect. These findings suggest that in basal LNCaP cells the CD44 promoter is unmethylated, whereas in basal DU145 and MCF-7 cells CD44 promoter activity may be repressed by DNA methylation.
In LNCaP and MCF-7 cells Cyclin D2 was constitutively expressed. Amazingly, TSA had a significantly stronger effect than 5-aza-CdR on the expression of Cyclin D2 in DU145 cells indicating potential inactive histone modifications at the promoter.
GLIPR1 was also constitutively expressed in LNCaP and MCF-7 cells, while 5-aza-CdR stimulation was significantly sufficient for a high re-expression of GLIPR1 in DU145 cells.
The basal transcriptional activity of PTEN could not be further up-regulated by the agents in all cell lines tested.
## Methylation status of the cd44, cyclin d2, glipr1 and pten promoters
To define the methylation status of the four promoters in the three cell lines bisulfite genomic sequencing was performed. The promoters of the highly expressed genes, such as PTEN in all cell lines, CD44 in LNCaP and Cyclin D2 in LNCaP and MCF-7, were unmethylated. Sixty percent of the CpG sites of the GLIPR1 promoter were methylated in basal DU145 cells, and treatment of these cells by 5-aza-CdR led to a decrease in methylation and activation of gene expression. The Cyclin D2 promoter was hardly methylated (6%) in basal DU145 cells, which parallels with the resistance to 5-aza-CdR. Immunoblot analysis using antibodies specific for MBD1, MBD2 and MeCP2 documented efficient protein expression in the transfected cells (data not shown).
## Repression
Evaluation of promoter-driven luciferase activities shows that DNA methylation of the reporter plasmids by SssI caused a stronger decrease of the CD44, Cyclin D2, GLIPR1 and PTEN promoter activity than that by HpaII due to the higher number of SssI sites than HpaII sites in the promoters (data not shown).
The Gal4 domain-mediated binding of MBD1-TRD to the Gal4 motif upstream of the promoters led to a significant repression of luciferase activity with all promoters (CD44, Cyclin D2, GLIPR1 and PTEN) in nearly all transfected cells [fig_ref] Figure 2: Luciferase activities of the co-transfected reporter plasmids containing the promoters of CD44,... [/fig_ref]. These findings indicate that MBD1 may have a general repressive effect on these tumour-associated genes. However, MBD2-TRD linked by the fusion part of the Gal4 DNA binding domain to the Gal4 sequence had a variable effect on the particular promoters and in the different tumour cells. While MBD2-TRD did not suppress the promoter of CD44 in all cells used, it was able to significantly repress Cyclin D2 in LNCaP and MCF-7 cells, and GLIPR1 in DU145 cells. Moreover, MBD2-TRD had only a slightly repressive effect on the basal expression of PTEN in LNCaP cells. With exception of the promoters of CD44 and GLIPR1 in LNCaP cells, MeCP2-TRD could repress to a different extent all investigated genes in all cells [fig_ref] Figure 2: Luciferase activities of the co-transfected reporter plasmids containing the promoters of CD44,... [/fig_ref].
In order to cover more informative aspects of the regulation of the four tumour-associated genes, co-transfections using expression plasmids, which encode for the full length MBD proteins, and unmethylated and HpaIImethylated reporter plasmids were accomplished without using the artificial link by the Gal4 system. Taken together, the results of these transient co-transfections largely support the data of the co-transfections based on the Gal4 system, with the exception of the promoters of PTEN and Cyclin D2. Here, in contrast to the Gal4-mediated binding, MBD2a and MeCP2 had a repressive effect on the methylated PTEN promoter in LNCaP and MCF-7 cells, and MBD2a suppressed the Cyclin D2 promoter in all cell lines tested [fig_ref] Figure 2: Luciferase activities of the co-transfected reporter plasmids containing the promoters of CD44,... [/fig_ref].
Usually, MBD1v1 and MBD2a had a stronger influence on the promoter activity than MBD1v3 and MBD2b, respectively [fig_ref] Table 1: MBD-mediated repression of CD44, Cyclin D2, GLIPR1 and PTEN [/fig_ref]. Moreover, co-transfections using the unmethylated reporter plasmids show that in contrast to the other members of the MBD family only MBD1v1 was able to repress the activity of the unmethylated promoters (data not shown and [fig_ref] Table 1: MBD-mediated repression of CD44, Cyclin D2, GLIPR1 and PTEN [/fig_ref]. The ability of MBD1v1 to bind unmethylated DNA depends on its third CXXC domain [bib_ref] Mbd1 is recruited to both methylated and nonmethylated CpGs via distinct DNA..., Jorgensen [/bib_ref]. Although the isoform MBD1v3 has only two of these domains, it could slightly repress the unmethylated promoters [fig_ref] Table 1: MBD-mediated repression of CD44, Cyclin D2, GLIPR1 and PTEN [/fig_ref] , which was also observed for other promoters [bib_ref] Methyl-CpG binding domain proteins and their involvement in the regulation of the..., Wischnewski [/bib_ref].
To exclude that the observed repressive effect of MBD1 is owing to endogenous MBD1 and to emphasize its role as a putative and general repressor, MBD1 -/mouse embryonic fibroblasts, which do not express MBD1 [fig_ref] Figure 3: Protein expression of MBD1, MBD2 and MeCP2 in basal and stimulated MCF-7... [/fig_ref] , were subsequently co-transfected with the reporter and expression plasmids. Transfected MBD1 -/mouse embryonic fibroblasts showed comparable data to the other cell lines used and a similar repressive effect of the transfected full length MBD1 on the promoter-driven luciferase activity (Additional File 1).
## In vivo binding of mbd1, mbd2 and mecp2 to the promoters of cd44, cyclin d2, glipr1 and pten
In [fig_ref] Figure 4: Chromatin immunoprecipitation using antibodies specific for MBD1, MBD2 and MeCP2 [/fig_ref] five examples of the evaluation of the realtime PCR products are shown, which were representa-tively chosen from the data of the immunoprecipitation. To pursue the changes of the in vivo DNA binding of MBD1, MBD2 and MeCP2 to the promoters, the bar charts show besides the precipitated, amplified DNA derived from basal DU145, LNCaP and MCF-7 cells, also DNA from the 5-aza-CdR-stimulated cells. Due to the characteristics of a housekeeping gene, which is unmethylated and constitutively expressed, the amplified, precipitated RPLP0 (ribosomal protein, large protein 0) gene served as negative and internal control of the real-time PCR. The values of the immunoprecipitation of the RPLP0 gene were approx. 5% and used as background level [fig_ref] Figure 4: Chromatin immunoprecipitation using antibodies specific for MBD1, MBD2 and MeCP2 [/fig_ref].
The examples of [fig_ref] Figure 4: Chromatin immunoprecipitation using antibodies specific for MBD1, MBD2 and MeCP2 [/fig_ref] to 4D were chosen because of the stimulatory effect of 5-aza-CdR and TSA on the expression of these genes [fig_ref] Figure 1: mRNA expression of CD44, Cyclin D2, GLIPR1 and PTEN in DU145, LNCaP... [/fig_ref]. As shown in [fig_ref] Figure 4: Chromatin immunoprecipitation using antibodies specific for MBD1, MBD2 and MeCP2 [/fig_ref] , the DNA of CD44 in DU145 cells was fairly enriched by the antibody specific for MBD1 (11-13%), whereas the immunoprecipitation by MBD2 and MeCP2 was at background level. In these cells 5-aza-CdR treatment caused a significant decrease of DNA enriched by MBD1 to the background level [fig_ref] Figure 4: Chromatin immunoprecipitation using antibodies specific for MBD1, MBD2 and MeCP2 [/fig_ref]. These findings are supported by the expression analyses demonstrating that 5aza-CdR could activate the expression of CD44 in these cells [fig_ref] Figure 1: mRNA expression of CD44, Cyclin D2, GLIPR1 and PTEN in DU145, LNCaP... [/fig_ref]. Furthermore, the transfection assays sustained these data and showed that most notably MBD1v1 (80%) had a strong repressive effect on CD44, whereas MBD2a and MeCP2 had almost no influence on the methylated promoter [fig_ref] Figure 2: Luciferase activities of the co-transfected reporter plasmids containing the promoters of CD44,... [/fig_ref] , [fig_ref] Table 1: MBD-mediated repression of CD44, Cyclin D2, GLIPR1 and PTEN [/fig_ref]. Similar data were obtained for CD44 in MCF-7 cells (data not shown).
As shown in [fig_ref] Figure 4: Chromatin immunoprecipitation using antibodies specific for MBD1, MBD2 and MeCP2 [/fig_ref] , in DU145 cells, where Cyclin D2 was not expressed [fig_ref] Figure 1: mRNA expression of CD44, Cyclin D2, GLIPR1 and PTEN in DU145, LNCaP... [/fig_ref] , an enrichment of Cyclin D2 could be observed using antibodies for MBD1 (22-30%), MBD2 (16-22%) and MeCP2 (22-27%). The stimulation of DU145 cells by 5-aza-CdR resulted in the expression of Cyclin D2 [fig_ref] Figure 1: mRNA expression of CD44, Cyclin D2, GLIPR1 and PTEN in DU145, LNCaP... [/fig_ref] and entailed a significant decrease in DNA yields of Cyclin D2 to the back-ground level [fig_ref] Figure 4: Chromatin immunoprecipitation using antibodies specific for MBD1, MBD2 and MeCP2 [/fig_ref]. These data indicate that MBD1, MBD2 and MeCP2 are able to bind to the methylated promoter of Cyclin D2 in basal DU145 cells, and leave the promoter following administration of 5-aza-CdR to the cells. In accordance with these findings, the transfection experiments showed that MBD1v1 (65%), MBD2a (5%) and MeCP2 (50%) were able to suppress the methylated promoter of Cyclin D2 in DU145 cells [fig_ref] Figure 2: Luciferase activities of the co-transfected reporter plasmids containing the promoters of CD44,... [/fig_ref] , [fig_ref] Table 1: MBD-mediated repression of CD44, Cyclin D2, GLIPR1 and PTEN [/fig_ref].
Another example in [fig_ref] Figure 4: Chromatin immunoprecipitation using antibodies specific for MBD1, MBD2 and MeCP2 [/fig_ref] demonstrates the enrichment of GLIPR1 DNA by the antibodies MBD1 (40-65%), MBD2 (15-25%) and MeCP2 (25-40%) in basal DU145 cells, and the significant decrease to the background level in 5-aza-CdR-treated cells. In DU145 cells GLIPR1 was scarcely expressed, and the stimulation of the cells by 5-aza-CdR led to a highly elevated transcriptional level [fig_ref] Figure 1: mRNA expression of CD44, Cyclin D2, GLIPR1 and PTEN in DU145, LNCaP... [/fig_ref]. These findings show that the in vivo binding of MBD1, MBD2 and MeCP2 to the methylated promoter of GLIPR1 in basal DU145 cells is abrogated by 5-aza-CdR, and are supported by the transfection assays demonstrating the ability of MBD1v1 (35%), MBD2a (60%) and MeCP2 (65%) to repress the methylated promoter of GLIPR1 in DU145 cells [fig_ref] Figure 2: Luciferase activities of the co-transfected reporter plasmids containing the promoters of CD44,... [/fig_ref] , [fig_ref] Table 1: MBD-mediated repression of CD44, Cyclin D2, GLIPR1 and PTEN [/fig_ref].
The DNA immunoprecipitated from MCF-7 [fig_ref] Figure 4: Chromatin immunoprecipitation using antibodies specific for MBD1, MBD2 and MeCP2 [/fig_ref] , DU145 and LNCaP cells (data not shown) by the antibodies for MBD1, MBD2 and MeCP2 did not enrich the PTEN sequence and did not exceed the background range of 5%. These findings agree with those of the expression analyses where PTEN was constitutively expressed and could not be further up-regulated by 5aza-CdR [fig_ref] Figure 1: mRNA expression of CD44, Cyclin D2, GLIPR1 and PTEN in DU145, LNCaP... [/fig_ref] suggesting that the PTEN promoter is unmethylated in these cells. The lacking occupancy of MBDs to the promoter was also observed for constitutively expressed CD44 in LNCaP cells, as well as for Cyclin D2 and GLIPR1 in LNCaP and MCF-7 cells (data not shown and [fig_ref] Figure 1: mRNA expression of CD44, Cyclin D2, GLIPR1 and PTEN in DU145, LNCaP... [/fig_ref].
## Protein expression of mbd1, mbd2 and mecp2 in du145, lncap and mcf-7 cells
The distribution of the different endogenous MBDs in each cell line was scrutinized by immunoblot analyses, which show similar expression levels of MBD1, MBD2 and MeCP2 in untreated and stimulated DU145, LNCaP (data not shown) and MCF-7 cells [fig_ref] Figure 3: Protein expression of MBD1, MBD2 and MeCP2 in basal and stimulated MCF-7... [/fig_ref]. These findings show that the different binding affinities and repressive effects of the MBDs were not caused by the different expression levels of these proteins in the various cell lines and by the stimulation of these cells. Moreover, the loss of expression of MBD1 in the MBD1 -/mouse embryonic fibroblasts is demonstrated in [fig_ref] Figure 3: Protein expression of MBD1, MBD2 and MeCP2 in basal and stimulated MCF-7... [/fig_ref].
## Histone signature at the promoters of cd44, cyclin d2, glipr1 and pten
Promoter activity may also be regulated by numerous modifications of the histones associated with the promoter. In general, acetylation of the N-terminal histone tails is a dominant signal for active chromatin facilitating the binding of the components of the basal transcription machinery. Histone methylation can be either an active or repressive signal. Mono-, di-and trimethylation of H3K4 are involved in gene expression. In contrast, mono-, di-and trimethylation of H3K9 and H4K20 correlate with stably transcriptional repressive regions of the genome.
In order to characterize the signature of the histones, which are bound to the active and repressive promoters, ChIP assays using antibodies specific for the methylated histones H3K4, H3K9 and H4K20 as well as for unmodified and acetylated histones were accomplished. To pursue the changes in the histone signature of the promoters of CD44, Cyclin D2, GLIPR1 and PTEN in basal DU145, LNCaP and MCF-7 cells, the code was compared with that in 5-aza-CdR-or TSA-stimulated cells. DNA enriched by the antibody IgG served as negative control and background level of the respective assay. Performing the analyses of methylated histones, a specific enrichment of the DNA above background level was only detected using the antibodies for di-and tri-methylated H3K4 in the basal cells. The stimulation of the cells by 5aza-CdR caused an increase in the immunoprecipitated dimethylated H3K4 [fig_ref] Figure 5: Chromatin immunoprecipitation using antibodies specific for methylated, unmodified and acetylated histones [/fig_ref]. Additionally, the analyses of the unmodified and acetylated histones showed a strengthened immunoprecipitation of the unmodified histones H2A, H2B, H3 and H4 in basal cells, which could not be observed in 5-aza-CdR-stimulated cells [fig_ref] Figure 5: Chromatin immunoprecipitation using antibodies specific for methylated, unmodified and acetylated histones [/fig_ref]. This histone signature in DU145 cells was similar to the code of PTEN in LNCaP and MCF-7 cells, as well as to that of CD44 in LNCaP cells and Cyclin D2 and GLIPR1 in LNCaP and MCF-7 cells (data not shown).
Moreover, it might reflect the respective constitutive expression of these genes in the appropriate cell lines [fig_ref] Figure 1: mRNA expression of CD44, Cyclin D2, GLIPR1 and PTEN in DU145, LNCaP... [/fig_ref].
As shown in [fig_ref] Figure 5: Chromatin immunoprecipitation using antibodies specific for methylated, unmodified and acetylated histones [/fig_ref] , only dimethylated H3K4 was enriched at the promoter of CD44 in basal MCF-7 cells and to a less extent in 5-aza-CdR-stimulated cells. In addition, an enrichment of the unmodified histones existed in the basal cells. However, in the 5-aza-CdRstimulated cells this enrichment was reduced in favour of the elevated levels of the immunoprecipitated, acetylated histones H3 and H4 [fig_ref] Figure 5: Chromatin immunoprecipitation using antibodies specific for methylated, unmodified and acetylated histones [/fig_ref]. The acetylation of H3 and H4 might correlate with the up-regulation of the gene expression of CD44 by 5-aza-CdR in MCF-7 cells [fig_ref] Figure 1: mRNA expression of CD44, Cyclin D2, GLIPR1 and PTEN in DU145, LNCaP... [/fig_ref].
In [fig_ref] Figure 5: Chromatin immunoprecipitation using antibodies specific for methylated, unmodified and acetylated histones [/fig_ref] a slight enrichment of dimethylated H3K4 could be observed for GLIPR1 in basal DU145 cells, whereas a considerable immunoprecipitation of monoand dimethylated H3K4 occurred in the 5-aza-CdR-stimulated cells. It was remarkable that in 5-aza-CdR-stimulated DU145 cells more acetylated than unmodified H2B was associated with the promoter of GLIPR1. On the other hand, more unmodified than acetylated H2A was bound to the gene in the basal cells [fig_ref] Figure 5: Chromatin immunoprecipitation using antibodies specific for methylated, unmodified and acetylated histones [/fig_ref]. The increase in mono-and dimethylation of H3K4 and acetylation of H2B in the stimulated cells may be concordant with the high transcript level mediated by 5-aza-CdR [fig_ref] Figure 1: mRNA expression of CD44, Cyclin D2, GLIPR1 and PTEN in DU145, LNCaP... [/fig_ref]. However, a high amount of acetylated H4 at the promoter of GLIPR1 was also observed in the basal cells.
As demonstrated in [fig_ref] Figure 5: Chromatin immunoprecipitation using antibodies specific for methylated, unmodified and acetylated histones [/fig_ref] , a similar enrichment of dimethylated H3K4 could be found for Cyclin D2 in basal and treated DU145 cells. Since the promoter of Cyclin D2 could be much stronger activated by TSA than by 5-aza-CdR in DU145 cells [fig_ref] Figure 1: mRNA expression of CD44, Cyclin D2, GLIPR1 and PTEN in DU145, LNCaP... [/fig_ref] and 5-aza-CdR did not largely affect the histone signature (data not shown), an additional ChIP assay for Cyclin D2 was performed using TSA-stimulated DU145 cells. The administration of TSA to the cells led to a strong acetylation of histones H2A, H2B and H4. In contrast, unmodified H2A and H4 were enriched in basal DU145 cells [fig_ref] Figure 5: Chromatin immunoprecipitation using antibodies specific for methylated, unmodified and acetylated histones [/fig_ref].
# Discussion
In the current study, the participation of MBDs in the transcriptional repression of the four selected tumourassociated genes CD44, Cyclin D2, GLIPR1 and PTEN was examined. The modifications of histones binding at the respective promoters in basal and 5-aza-CdR-stimulated prostate cancer cells DU145 and LNCaP and breast tumour cells MCF-7 were investigated, as well. Comparison of the events at the promoters of the individual genes in the different cell lines aimed to clarify whether their regulation is promoter-and/or cell-specific.
Highly constitutively expressed genes, such as PTEN in all three cell lines, Cyclin D2 and GLIPR1 in LNCaP and MCF-7 cells or CD44 in LNCaP cells are unmethylated as shown by bisulfite sequencing and could therefore not be further up-regulated by the demethylating agent 5-aza-CdR. These genes also showed no in vivo binding of MBDs to their promoters. However, transient transfection assays demonstrated a repressive potential of MBDs, when these promoters were methylated in vitro. In contrast, CD44 in DU145 and MCF-7 cells as well as Cyclin D2 and GLIPR1 in DU145 cells may be inactivated by DNA methylation. ChIP assays showed that at least one member of the MBD family bound to these promoters, and its binding affinity to the promoters correlated with its ability to repress the promoter activity in transient cotransfections. 5-aza-CdR may cause demethylation of the DNA and the release of the MBDs from the promoters. These findings show that these tumour-associated genes may be targets of therapeutic drugs, such as demethylating agents.
As described for patients with myelodysplastic syndrome (MDS) of the bone marrow, 5-aza-CdR has already been introduced as a therapeutic drug and was found to prolong survival of these patients. Besides, TSA is already used as a therapeutic drug for acute myeloid leukaemia. It can induce cell differentiation and apoptosis, has antiproliferative effects and leads to cell-cycle arrest [bib_ref] Histone deacetylase inhibitors such as sodium butyrate and trichostatin A induce apoptosis..., Sawa [/bib_ref]. Combined epigenetic therapies of a demethylating agents with a histone deacetylase inhibitor indicate that these agents have significant activity in patients with MDS/ acute myelogenous leukaemia [bib_ref] Demethylating agents in myeloid malignancies, Garcia-Manero [/bib_ref]. Due to the findings of our study, these treatments should also be considered for patients with solid tumours.
As shown in ChIP and transient co-transfection assays, MBD1v1 bound and repressed the methylated promoters in all cell lines used. In transfection assays MBD1v1 showed additionally a repressive effect on unmethylated promoters. Based on the repressive effect of exogenous MBD1 in MBD1 -/mouse embryonic fibroblasts, MBD1v1 may act as a general, epigenetic factor for these methylated promoters. The ability of MBD1v1 to bind also demethylated DNA is owing to its third CXXC domain [bib_ref] Mbd1 is recruited to both methylated and nonmethylated CpGs via distinct DNA..., Jorgensen [/bib_ref]. MBD1v3 does not possess this domain, but had a weak repressive effect on the unmethylated promoters [bib_ref] Methyl-CpG binding domain proteins and their involvement in the regulation of the..., Wischnewski [/bib_ref]. Furthermore, the ChIP assays revealed the in vivo binding of MBD1, MBD2 and MeCP2 to the promoter of Cyclin D2 and GLIPR1 in DU145 cells which exhibited a very low level of the respective transcripts. In contrast, CD44 was only bound by MBD1 in DU145 and MCF-7 cells in which the transcription was slightly higher than the basal expression of Cyclin D2 and GLIPR1 in DU145 cells suggesting that the promoter of CD44 might be partly demethylated. In transient cotransfection assays MeCP2 and MBD2a had a minor and rather heterogeneous influence on the methylated promoters than MBD1v1 and could inhibit gene expression in 75% and 50% of the cases, respectively. The repressive effect of MeCP2 and MBD2a on the promoter of Cyclin D2 was promoter-specific in the three cell lines, albeit the influence of MBD2a was weak in DU145 cells. In MCF-7 cells MeCP2 repressed cell-specifically the methylated promoters. Moreover, the promoters of Cyclin D2 and PTEN could be suppressed by all MBDs and their isoforms studied in LNCaP and MCF-7 cells. Our findings show that at least one member of the MBD family was always involved in repression of the methylated promoters in each cell line. These findings suggest the potential impact of therapeutical intervention on cancer patients by means of increasing expression and tumour-suppressive function of genes, which are epigenetically silenced by MBD protein occupancy.
As far as we know, only two publications have reported on such an epigenetic comparison of MBDs regarding different genes in different cell lines. In a large-scale study Lopez-Serra et al. described the binding affinity of MBDs to 22 tumour suppressor genes in 10 cell lines, among others in MCF-7 cells, in which, however, none of the four presented genes was considered [bib_ref] A profile of methyl-CpG binding domain protein occupancy of hypermethylated promoter CpG..., Lopez-Serra [/bib_ref]. These authors also referred to the binding affinity of MBD1 to unmethylated promoters. Furthermore, they showed that MBD2 and MeCP2, but not MBD1, are promoter-specific factors of the 22 genes and MBD2, but not MBD1 or MeCP2, is a cell-specific factor [bib_ref] A profile of methyl-CpG binding domain protein occupancy of hypermethylated promoter CpG..., Lopez-Serra [/bib_ref]. The promoter-specificity of MeCP2 and MBD2a in these ChIP analyses is consistent to our data based on transient co-transfection assays. The second study of Ballestar et al. investigated the binding affinity of MBDs to the promoters of 6 tumour suppressor genes in two different cell lines and normal lymphocytes using the ChIP assay. Contrary to the study of Lopez-Serra et al. and our findings this laboratory could not observe any binding of MBD1 to unmethylated as well as methylated promoters. However, they could show a gene-specific binding pattern of MBDs at the methylated promoters. Whereas MeCP2 bound to all methylated promoters, MBD2 had only binding affinity to one of the promoters studied [bib_ref] Methyl-CpG binding proteins identify novel sites of epigenetic inactivation in human cancer, Ballestar [/bib_ref].
To investigate the signature of histones binding to the promoters of CD44, Cyclin D2, GLIPR1 and PTEN in basal and 5-aza-CdR-stimulated DU145, LNCaP and MCF-7 cells, 17 antibodies specific for mono-, di-and trimethylated, acetylated and unmodified histones were applied. In agreement to the association of mono-, diand trimethylated histone H3K4 with active genes [bib_ref] Q2ChIP, a quick and quantitative chromatin immunoprecipitation assay, unravels epigenetic dynamics of..., Dahl [/bib_ref] the analysis of the methylation status of the histones show mainly associations with these modifications. All three methylation grades of H3K4 were reported to be localized in the region of the transcriptional start site of known genes. An increased binding of H3K4me3 could be observed at highly expressed genes. However, H3K4me1 and H3K4me2 associated at intermediately active promoters and were found rather downstream of the transcriptional start site [bib_ref] Q2ChIP, a quick and quantitative chromatin immunoprecipitation assay, unravels epigenetic dynamics of..., Dahl [/bib_ref] [bib_ref] High-resolution profiling of histone methylations in the human genome, Barski [/bib_ref]. The findings of this study showed that high levels of H3K4me2 and H3K4me3 or H3K4me2 alone could be confined to highly expressed genes. The intermediately active promoters were bound by H3K4me2 and in one case by H3K4me1 and H3K4me2.
Furthermore, generally no mono-, di-and trimethylated histones H3K9 and H4K20 at the promoters were detected. This may be explained by their preferred occurrence in heterochromatin and implication in gene silencing. It was reported that H3K9me3 and H4K20me3 associated with constitutive heterochromatin and stable gene repression, and H3K9me2 and H4K20me1 associated with facultative heterochromatin and temporarily inactive genes [bib_ref] Q2ChIP, a quick and quantitative chromatin immunoprecipitation assay, unravels epigenetic dynamics of..., Dahl [/bib_ref] [bib_ref] High-resolution profiling of histone methylations in the human genome, Barski [/bib_ref]. Moreover, confirmed the presence of H3K9me2 and H3K9me3 at inactive genes while they also perceived H3K9me1 at active genes [bib_ref] Q2ChIP, a quick and quantitative chromatin immunoprecipitation assay, unravels epigenetic dynamics of..., Dahl [/bib_ref] [bib_ref] High-resolution profiling of histone methylations in the human genome, Barski [/bib_ref]. These data support our analysis showing the exclusive binding of H3K9me1 to the active promoter of Cyclin D2 in basal and 5-aza-CdR-stimulated MCF-7 cells. However, a recent publication showed that H3K9me3 may also be enriched in numerous active promoters [bib_ref] Suz12 binds to silenced regions of the genome in a cell-type-specific manner, Squazzo [/bib_ref]. To sum up, the genes of CD44, Cyclin D2, GLIPR1 and PTEN are obviously located neither in constitutive nor facultative heterochromatin.
In addition, the ChIP assays presented in this work showed that, in general, unmodified histones at repressive as well as active promoters were acetylated following stimulation of the cells by 5-aza-CdR. The activation of the repressive promoters mediated by this demethylating agent sustains the fact that unmodified and acetylated histones bind preferentially within areas of repressive and active genes, respectively. Moreover, the recruitment of HDACs by MBDs to methylated DNA leads to histone deacetylation, whereas DNA demethylation promotes the release of MBDs together with HDACs and consequently histone acetylation. A high level of acetylation after 5aza-CdR-mediated stimulation was observed for histone H3 at the promoter of CD44 in MCF-7 cells. Comparative analyses of the methylation status and chromatin structure of the p14(ARF)/p16(INK4A) promoters showed generally the presence of higher levels of acetylated H3 at unmethylated than methylated CpG dinucleotides in a series of normal and cancer cells [bib_ref] Altered chromatin structure associated with methylation-induced gene silencing in cancer cells: correlation..., Nguyen [/bib_ref]. As a result, acetylated H3 is particularly associated with transcription and involved in histone deposition and chromatin assembly [bib_ref] The language of covalent histone modifications, Strahl [/bib_ref]. In contrast, it was reported that TSA treatment induced significant acetylation of H3 at the multidrug resistance gene 1 (MDR1) but did not activate transcription of this gene [bib_ref] Precipitous release of methyl-CpG binding protein 2 and histone deacetylase 1 from..., El-Osta [/bib_ref]. Furthermore, in 5-aza-CdR-treated cells an increase of the amount of acetylated H3 at the cytomegalovirus promoter was observed. However, hypoacetylation played only a moderate role in the inactivation of this promoter [bib_ref] Cell type-specific activation of the cytomegalovirus promoter by dimethylsulfoxide and 5-Aza-2'-deoxycytidine, Radhakrishnan [/bib_ref]. This observation is similar to the present results showing that unmodified histones, which are frequently detected at inactive genes, were associated with the constitutively expressed PTEN gene. The lacking acetylation of the histones might be compensated by the high di-and trimethylation of H3K4. Although 5-aza-CdR had no effect on the constitutive expression of this gene, it affected indirectly the acetylation of the surrounding histones. In one case of Cyclin D2 in DU145 cells TSA had a stronger effect on gene expression than 5-aza-CdR. The stimulation of the cells by TSA confirmed its function as a histone deacetylase inhibitor and caused a high degree of acetylation of the unmodified histones.
However, TSA treatment has not only been shown to open the chromosomal structure increasing the accessibility of transcription factor complexes to their binding sites on the promoter and to up-regulate gene transcription. TSA, as well as 5-aza-CdR, is also involved in the post-transcriptional regulation by changing mRNA stability. TSA can reduce the half life of mRNAs and acetylate cytoplasmic proteins [bib_ref] Trichostatin A and 5 Aza-2' deoxycytidine decrease estrogen receptor mRNA stability in..., Pryzbylkowski [/bib_ref] [bib_ref] HDAC inhibitors regulate claudin-1 expression in colon cancer cells through modulation of..., Krishnan [/bib_ref]. 5-aza-CdR and TSA could affect expression levels by indirect modulation of downstream gene regulatory mechanisms and alter the subcellular distribution of proteins that mediate posttranscriptional regulation. They caused the interaction between an RNA binding protein and the estrogen receptor mRNA leading to increased mRNA stability [bib_ref] Trichostatin A and 5 Aza-2' deoxycytidine decrease estrogen receptor mRNA stability in..., Pryzbylkowski [/bib_ref]. Moreover, the modulation of mRNA stability of claudin-1, a tight junction protein, by its 3'-UTR has been revealed as the major mechanism underlying HDACdependent claudin-1 expression [bib_ref] HDAC inhibitors regulate claudin-1 expression in colon cancer cells through modulation of..., Krishnan [/bib_ref]. To sum up, the reports show that TSA may influence the mRNA stability of tumour-associated genes in different manners, it is able to destabilize the claudin-1 mRNA [bib_ref] HDAC inhibitors regulate claudin-1 expression in colon cancer cells through modulation of..., Krishnan [/bib_ref] while in case of the cell-cycle control gene p21 WAF1 it stabilized the mRNA [bib_ref] Histone deacetylase inhibitors regulate p21WAF1 gene expression at the post-transcriptional level in..., Hirsch [/bib_ref]. Since we only considered the impact of 5aza-CdR and TSA on the promoter settings, we do not know the contribution of the RNA stability in the regulation of our gene set. This requires further investigation.
# Conclusions
The presented combined investigations on biologically relevant tumour-associated genes in different tumour cell types show diverse and characteristic profiles of MBD patterns and histone signatures at the promoters. These results contribute to a more comprehensive understanding of the epigenetic interplay in tumourigenesis and the role of MBDs and histone modifications in the regulation of tumour-associated genes, and may constitute valuable information in therapeutical approaches for re-expression of tumour suppressor genes as part of individual cancer treatments.
[fig] Figure 1: mRNA expression of CD44, Cyclin D2, GLIPR1 and PTEN in DU145, LNCaP and MCF-7 cells. The data assessed by RT-PCR (gel electrophoretically separated on a representative agarose gel, left panel) and quantitative real-time PCR (bar charts, right panels) show the relative levels and changes in the mRNA expression of the tumour-associated genes in unstimulated basal (b), 5-aza-CdR-(A), TSA-(T) and 5-aza-CdR+TSA-(A+T) stimulated DU145, LNCaP and MCF-7 cells. The housekeeping gene β-Actin was selected as internal control. * Statistical significance of p < 0.05 according to the Fisher's exact test in respect of changes in stimulated samples compared to the basal status. [/fig]
[fig] Figure 2: Luciferase activities of the co-transfected reporter plasmids containing the promoters of CD44, Cyclin D2, GLIPR1 or PTEN in MCF-7, DU145 and LNCaP cells. MCF-7, DU145 and LNCaP cells were transiently transfected with pGL2-Luciferase reporter plasmids containing promoter fragments of CD44 (A), Cyclin D2 (B), GLIPR1 (C) or PTEN (D) immediately downstream of five Gal4 binding motifs (left diagrams), or with methylated reporter plasmids containing the same promoters without the Gal4 motifs (right diagrams). The co-transfected expression plasmids encoded either for Gal4-fused TRDs of MBD1, MBD2, or MeCP2 (left diagrams) or for full length proteins of MBD1v1, MBD1v3, MBD2a, MBD2b or MeCP2 proteins (right diagrams). The activities derived from the reference Renilla Luciferase were used for normalization of the data. The relative luciferase activities of the reporter constructs co-transfected with the empty expression plasmids (-) were arbitrarily set to 100%. * Statistical significance of p < 0.05 according to the Fisher's exact test in respect of repression of promoters by MBDs compared to the basal level activity (-). [/fig]
[fig] Figure 3: Protein expression of MBD1, MBD2 and MeCP2 in basal and stimulated MCF-7 cells and in MBD1 -/mouse embryonic fibroblasts. The MBD protein levels in basal, 5-aza-CdR-, TSA-and 5-aza-CdR&TSA-stimulated MCF-7 cells were evaluated by Western Blot analysis using antibodies specific for MBD1 (60 kDa), MBD2 (49 kDa), MeCP2 (70 kDa) and HSC70 (70 kDa, loading control) (A). The MBD1 protein level in MBD1 -/mouse embryonic fibroblasts was evaluated in comparison to basal MCF-7 cells with antibodies specific for MBD1 (60 kDa) and HSC70 (70 kDa, loading control) (B). [/fig]
[fig] Figures: 5A and 5B show the modifications of the histones binding to the promoter of PTEN in DU145 cells. [/fig]
[fig] Figure 4: Chromatin immunoprecipitation using antibodies specific for MBD1, MBD2 and MeCP2. Quantitative real-time PCR analysis of the immunoprecipitated DNA (IP) derived from unstimulated (basal) and 5-aza-CdR-stimulated DU145, LNCaP and MCF-7 cells was performed using primer pairs specific for the promoter fragments of RPLP0 (A), CD44 (B), Cyclin D2 (C), GLIPR1 (D), and PTEN (E). All values obtained were normalized and referred to 100% of the input DNA. * Statistical significance of p < 0.05 according to the Fisher's exact test. ** Statistical significance of p < 0.05 according to the analysis of variance (ANOVA). [/fig]
[fig] Figure 5: Chromatin immunoprecipitation using antibodies specific for methylated, unmodified and acetylated histones. Representative results of quantified DNA derived from unstimulated (basal) and 5-aza-CdR-or TSA-stimulated DU145 and MCF-7 cells immunoprecipitated by antibodies specific for methylated histones (left diagrams) as well as for unmodified and acetylated histones (right diagrams). Examples of PTEN in DU145 cells (A, B), CD44 in MCF-7 cells (C, D), GLIPR1 in DU145 cells (E, F) and Cyclin D2 in DU145 cells (G, H) are shown. All values obtained were normalized and referred to 100% of the input DNA. IgG, negative control; H3K9, Lysine 9 of histone H3; H3K4, Lysine 4 of histone H3; H4K20, Lysine 20 of histone H4; H2A, histone H2A; H2B, histone H2B; me, mono-methylated; me2, dimethylated; me3, trimethylated; Ac, acetylated. [/fig]
[table] Table 1: MBD-mediated repression of CD44, Cyclin D2, GLIPR1 and PTEN [/table]
|
Direct Visualization of Antigen-specific CD8+T Cells during the Primary Immune Response to Epstein-Barr Virus In Vivo
Primary infection with virus can stimulate a vigorous cytotoxic T cell response. The magnitude of the antigen-specific component versus the bystander component of a primary T cell response remains controversial. In this study, we have used tetrameric major histocompatibility complex-peptide complexes to directly visualize antigen-specific cluster of differentration (CD)8 ϩ T cells during the primary immune response to Epstein-Barr virus (EBV) infection in humans. We show that massive expansion of activated, antigen-specific T cells occurs during the primary response to this virus. In one individual, T cells specific for a single EBV epitope comprised 44% of the total CD8 ϩ T cells within peripheral blood. The majority of the antigen-specific cells had an activated/memory phenotype, with expression of human histocompatibility leukocyte antigen (HLA) DR, CD38, and CD45RO, downregulation of CD62 leukocyte (CD62L), and low levels of expression of CD45RA. After recovery from AIM, the frequency of antigen-specific T cells fell in most donors studied, although populations of antigen-specific cells continued to be easily detectable for at least 3 yr.
P rimary infection with virus can stimulate a vigorous T cell response, with activation and proliferation of lymphocytes at the site of infection (1), within draining lymph nodes, and sometimes also within peripheral blood. The extent to which these lymphocytes have been specifically activated by antigen or nonspecifically activated, perhaps by cytokines, (bystander activation) is still a matter of debate. Recent work has suggested that the antigen-specific component of the primary T cell response to infection may be greater than was originally suggested. However, progress has been limited by the difficulties of identifying antigen-specific T cells in vivo. Little is known about the phenotypic differences between different functional subsets of T cells and it is not yet possible to distinguish between antigen-stimulated effector T cells, antigen specific memory T cells, and bystander-activated T cells on the basis of cell phenotype. Antibody and PCR techniques may be used to identify T cells expressing receptors known to be selected by certain MHC-peptide complexes, but antigen-specific T cells expressing other receptors will inevitably be missed using these types of approach.
Recently, a novel method of identifying antigen-specific T lymphocytes has been described. Tetrameric MHCpeptide complexes have been shown to bind stably and specifically to appropriate MHC-peptide-specific T cells. This technique may be used both to quantify and to characterize antigen-specific T cells directly. We have exploited this technique to study antigen-specific T cells during the primary and early persistent phases of EBV infection in humans. Infection with this highly immunogenic virus provides an ideal natural situation in which to study the development of a cluster of differentiation (CD)8 ϩ immune response in vivo in humans. Primary EBV infection in adolescence or young adulthood may be clinically manifest as acute infectious mononucleosis (AIM), a disease characterized by a striking expansion of CD8 ϩ T cells in peripheral blood. Resolution of the symptoms is associated with the return to a normal blood picture and with the establishment of a life-long virus carrier state under the control of memory T cell surveillance.
EBV is a gamma herpes virus that can establish both latent and fully productive (lytic) infections, involving the expression of different sets of viral genes. The B cell reservoir is the site of virus latency, whereas lytic infections 1396 Visualization of Antigen-specific T Cells appear to be focused in epithelial and/or locally infiltrating B cells in the oropharynx. Most work on CTL control of EBV infection in humans has concentrated on the response to latent cycle antigens; here, the evidence from limiting dilution assays in vitro suggests that the frequency of circulating CD8 ϩ T cells with specificity for given target epitopes ranges from 1:100 to 1:500 during primary infection and from 1:500 to 1:2,500 once the long-term virus carrier state has been established. Recently, CTL responses to lytic cycle antigens have been identifiedand, certainly during the primary response, their frequencies appear to be at least as high as the latent antigen-specific reactivities. Even so, if such values from in vitro outgrowth are accurate, then the virus-specific response as a whole would account for just a minor fraction of the total CD8 ϩ T cell population in AIM. Somewhat in contrast, a study of the TCR repertoire used in AIM indicates the presence of unusually large T cell expansions that can constitute up to 25% of the total CD8 ϩ pool and whose monoclonal or oligoclonal pattern of TCR usage implies that they are antigen driven.
In an effort to reconcile these data, we have made tetramers of MHC molecules complexed to defined EBV peptide epitopes, and used these to identify directly and to characterize epitope-specific T cells within the primary T cell response in AIM patients. In selected individuals we have also used the tetramers to search for epitope-specific T cells in the circulation between 6 and 37 mo later, after establishment of the long-term virus carrier state. We chose to study the response to two immunodominant epitopes from EBV lytic cycle proteins (an HLA A2-restricted epitope in BMLF1 and an HLA B8-restricted epitope in BZLF1) and to one immunodominant epitope from an EBV latent cycle protein (an HLA B8-restricted epitope in Epstein-Barr nuclear antigen (EBNA) 3A). We went on to analyze the phenotype of the antigen-specific cells, looking at expression of known markers of activation (HLA DR, CD38, downregulation of CD62 leukocyte [CD62L]), memory (CD45RO), and late differentiation (CD57) in the T cell lineage. We demonstrate massive expansion of activated, antigen-specific CD8 ϩ T cells during the primary immune response to EBV and show that antigen-specific CD8 ϩ T cells persist at relatively high frequencies for at least 3 yr after primary infection.
# Materials and methods
Patients. 10 HLA A2-positive individuals and three HLA B8-positive individuals with recent onset of AIM and a positive EBV-specific latex agglutination test (Oxoid, Basingstoke, UK) were studied. Samples of peripheral blood were taken as soon as the diagnosis of AIM was confirmed. Further samples were taken from selected donors 6-37 mo later, long after full recovery from the clinical illness.
Normal Individuals. Samples of peripheral blood were also taken from HLA A2 ϩ and HLA B8 ϩ normal healthy adult individuals of known serological status with respect to anti-EBV antibodies.
Isolation and Fractionation of Lymphocyte Preparations. Peripheral blood was diluted with an equal volume of RPMI. PBMCs were separated using Ficoll-Hypaque density gradient centrifugation and cryopreserved within 3 h of venesection. Tissue Typing. Class I tissue typing was performed on lymphocytes using classical serological methods.
Synthesis of MHC-Peptide Tetrameric Complexes. Soluble peptide-MHC tetramers were produced using a similar method to that described by Altman et al.. Recombinant class I heavy chain (HLA A2 and HLA B8) or  2 microglobulin was produced in Escherichia coli cells transformed with the relevant expression plasmids. Expression of class I heavy chain was limited to the extracellular domain only and the sequence of this domain was modified by the addition of a substrate sequence for BirA biotinylation at the COOH terminus. HLA A2 and HLA B8 complexes were folded in vitro using 30 mg of heavy chain protein, 25 mg of  2 microglobulin, and 10 mg of synthetic peptide. The HLA A2 peptide ligand was GLCTLVAML from the EBV lytic protein BMLF1), and the HLA B8 peptide ligands were RAKFKQLL; from the EBV lytic protein BZLF1) and FLRGRAYGL (16; from the latent protein EBNA 3A). Protease inhibitors (2 g/ml pepstatin, 2 g/ml leupeptin, and 0.2 mM phenylmethylsulphonyl fluoride) were used to preserve the protein. The MHC complexes were biotinylated using purified BirA enzyme at a concentration of 5 g/ml, 0.5 mM biotin, and 5 mM ATP. The reaction was incubated at room temperature for 16 h. Typically, levels of biotinylation of 80% were achieved. The biotinylated MHC-peptide complexes were recovered by FPLC purification (using buffer containing 20 mM Tris, pH 8.0, and 50 mM NaCl) and ion exchange chromatography (0-0.5 M NaCl gradient). Tetramers were made by mixing biotinylated protein complex with streptavidin-PE (Sigma Chemical Co.) at a molar ratio of 4:1.
Cell Staining.
# Results
Specificity of the HLA-Peptide Tetrameric Complexes. We made phycoerythrin-labeled tetrameric complexes of HLA A2 folded with the GLCTLVAML peptide from the EBV lytic protein BMLF1and of HLA B8 folded with the RAKFKQLL epitope from the EBV lytic protein BZLF1and with the FLRGRAYGL epitope from the EBV latent protein EBNA 3A.
The HLA A2-GLCTLVAML tetrameric complex stained HLA A2-restricted GLCTLVAML-specific T cell clones, but did not stain other HLA A2-restricted clones specific for the HIV gag epitope (SLYNTVATL;b ) or for the influenza matrix protein epitope (GILGFVFTL; data not shown). Additional controls confirmed the specificity of the tetramer staining. Thus the complex did not stain CD8 ϩ T cells within PBL preparations from healthy HLA A2-negative individualswithin parallel preparations from HLA A2-negative AIM patientsor within preparations from healthy HLA A2-positive individuals who were EBV seronegative.
The specificities of the HLA B8-FLRGRAYGL and the HLA B8-RAKFKQLL tetrameric complexes were tested in a similar way. The tetramers stained T cell clones specific for the appropriate ligand, but not T cell clones specific for other HLA B8-restricted peptides (data not shown). Neither of the HLA B8 tetrameric complexes stained T cells within PBLs taken from HLA B8-negative donors (data not shown), nor did they stain T cells within PBLs taken from HLA B8-positive individuals who were EBV seronegative (data not shown).
Frequency of Circulating EBV Epitope-specific T Cells during Primary EBV Infection. We used the three tetramers to analyze the frequency of circulating CD8 ϩ T cells specific for immunodominant EBV lytic (GLCTLVAML and RAK-FKQLL) and latent (FLRGRAYGL) cycle epitopes during primary EBV infection.
PBLs taken from 10 HLA A2-positive individuals suffering from recent onset AIM were stained with the phycoerythrin-conjugated HLA A2-GLCTLVAML tetrameric complex and with an antibody to CD8 conjugated to tricolor. In all HLA A2-positive individuals, a population of CD8 ϩ T cells clearly stained with the HLA A2-GLCTL-VAML tetrameric complex. The frequency of tetramerreactive T cells in PBL ranged from 0.5 to 6.6% of the CD8 ϩ T cells in blood in the 10 donors with AIM. In 4 of The frequency of CD8 ϩ T cells specific for three epitopes from EBV within PBLs taken from donors suffering from AIM (a) PBLs from the two HLA A2-positive donors, IM74 and IM83, suffering from AIM, were stained with phycoerythrin-conjugated HLA A2-GLCTL-VAML tetrameric complex and with an antibody to CD8 conjugated to tricolor. (b) PBLs from HLA B8-positive donors IM70 and IM59, suffering from AIM, were stained with phycoerythrin-conjugated HLA B8-RAKFKQLL tetrameric complex and with an antibody to CD8 conjugated to tricolor. (c) PBLs from HLA B8-positive donors IM63 and IM59, suffering from AIM, were stained with phycoerythrin-conjugated HLA B8-FLRGRAYGL tetrameric complex and with an antibody to CD8 conjugated to tricolor. In each figure, the number of CD8 ϩ T cells that stain with the tetrameric complex is shown as percent frequency. the 10 donors, the frequency was Ͼ 5%.a illustrates the data obtained from patients infectious mononucleosis (IM)74 and IM83 in whom 5.6 and 6.6% of circulating CD8 ϩ T cells stained with the HLA A2-GLCTLVAML complex.
In a second series, PBLs taken from three HLA B8-positive individuals suffering from recent onset AIM were stained with the phycoerythrin-conjugated HLA B8-RAKFKQLL or HLA B8-FLRGRAYGL tetrameric complex and with an anti-CD8 antibody conjugated to Tricolor. The frequency of HLA B8-RAKFKQLL-reactive CD8 ϩ T cells was very high in these donors.shows that 44 and 40% of CD8 ϩ T cells in PBLs from patients IM70 and IM59, respectively, stained with this tetrameric complex. In the third patient, IM63, the level of staining was 29%. In contrast, the frequency of CD8 ϩ T cells reactive with the HLA B8-FLRGRAYGL tetrameric complex was relatively low..2% of CD8 ϩ T cells in PBLs from patients IM63 and IM59, respectively, stained with this complex. Only 0.3% of CD8 ϩ T cells from patient IM70 were identified in this way. FLRGRAYGL is known to be a dominant EBV latent cycle epitope and in donor IM63 we have previously shown that the T cell response to this epitope dominates the T cell response to EBV latent cycle antigens. Nevertheless in all three individuals in this study, the frequency of CD8 ϩ T cells specific for the HLA B8-restricted EBV lytic cycle epitope (RAKFKQLL) was at least 10-fold greater than the frequency of cells specific for the HLA B8-restricted latent cycle epitope (FLRGRAYGL).
At the same time, we again conducted control experiments showing that PBL from HLA-mismatched individuals suffering from AIM did not stain with the tetrameric MHC-peptide complexes.
Frequency of Circulating EBV Epitope-specific T Cells in Postconvalescent AIM Patients. We had access to samples of peripheral blood taken from two of the HLA A2-positive patients and all three of the HLA B8-positive patients between 6 and 37 mo after primary infection with EBV. At this time all donors had fully recovered from the clinical ill-ness. The frequency of CD8 ϩ T cells reactive with the EBV lytic cycle epitopes (GLCTLVAML and RAK-FKQLL) fell in all the donors studied, although populations of tetramer-reactive cells remained easily detectable . Interestingly, the frequency of CD8 ϩ T cells reactive with the EBV latent cycle epitope (FLRGRAYGL) was similar in the acute and follow-up samples . In light of these findings, we used the tetrameric complexes to study the frequency of EBV epitope-specific T cells in healthy adult donors who were seropositive for EBV, but gave no history of AIM. Such individuals are likely to have carried the viral infection for many years. Staining of PBLs taken from these donors revealed a low but still significant frequency (usually Ͻ 1:200) of tetramer-reactive CD8 ϩ lymphocytes in some individuals with the relevant HLA class I allele (data not shown). Importantly, in control experiments of the kind illustrated in, c and e , we again showed that PBLs taken from HLA-matched EBV-seronegative donors or from HLA-mismatched EBV-seropositive donors gave no staining with the tetrameric complexes, c and e , and data not shown).
Phenotype of EBV Epitope-specific T Cells during the Primary T Cell Response In Vivo. The phenotype of the tetramer reactive cells was analyzed in six individuals suffering from AIM using three-color FACS ® analysis. The results of such an analysis, in this case involving the HLA A2-GLCTL-VAML tetramer-reactive cells in patient IM73, are shown in. The epitope-specific cells had an activated/memory phenotype; thus the majority expressed the activation markers HLA DR, and CD38 (17;, but showed downregulation of the lymph node homing receptor CD62L (18;, consistent with the finding that the molecule is downregulated after lymphocyte activation. Note also that the majority of the tetramer-reactive cells expressed the CD45RO isoform of CD45 and not the CD45RA isoform (19;, e and f ); in this context, CD45RO has been proposed as a marker both for recently activated and for "memory" T cells. CD28, the ligand for B7, is an important costimulatory molecule that is normally . The frequency of CD8 ϩ T cells specific for three epitopes from EBV within PBLs taken from donors suffering from AIM and from the same donors at least 6 mo later. PBLs from HLA A2-positive donors (a) and HLA B8-positive donors (b and c) suffering from AIM were stained with the phycoerythrinconjugated HLA A2-GLCTL-VAML tetrameric complex (a), the phycoerythrin-conjugated HLA B8-RAKFKQLL tetrameric complex (3), or the phycoerythrin-conjugated HLA B8-FLR-GRAYGL tetrameric complex (c). The frequency of tetramer-positive cells within CD8 ϩ PBLs taken from donors suffering from AIM is shown in solid bars, and the frequency of tetramer-positive cells within CD8 ϩ PBLs taken from the same donors at least 6 mo later is shown in shaded bars. expressed on the majority of CD4 ϩ T cells, but on only %05ف of CD8 ϩ T cells. This molecule was expressed on 48% of the epitope-specific cells in patient IM73, whereas CD57, which has been postulated to be a marker of terminal T cell differentiation, was expressed on only 2% of that population. Complete results of the phenotypic analysis of the tetramer-reactive T cells taken from four HLA A2-positive and two HLA B8-positive patients with AIM are shown in. Although there was some variation between patients, the tetramer-reactive cells were generally CD38 ϩ , HLA DR ϩ , CD45RO ϩ , and CD45RA Ϫ ; levels of CD28 positivity were the most variable.
Phenotype of EBV Epitope-specific T Cells in Postconvalescent AIM Patients. We were able to compare the phenotype of the HLA B8-restricted FLRGRAYGL-and RAK-FKQLL-specific T cells present within peripheral blood taken from donor IM63 during primary infection and 37 mo later. During primary infection 44% of CD8 ϩ T cells reacted with the HLA B8-RAKFKQLL tetrameric complex. At follow up, still 18% of CD8 ϩ T cells reacted with this complex, though by this time the patient was completely well and the circulating blood picture was normal. Compared with the phenotype of HLA B8-RAK- FKQLL tetramer-reactive cells during the primary infection, those cells present in the follow-up bleed showed lower levels of expression of the activation markers CD38and HLA DR (data not shown) and higher levels of the lymph node homing receptor CD62L (data not shown). There was also an increase in the percentage of tetramer-reactive cells expressing CD57 (data not shown).
Significantly, there was a clear reduction in the frequency of expression of CD45ROby the antigenspecific cells, and an increase in the frequency of expression of CD45RA. Analysis of the HLA B8restricted FLRGRAYGL-specific T cells in the same patient showed a similar fall in the proportion of cells expressing CD38, HLA DR, and CD45RO and in increase in the proportion of cells expressing CD62L, CD57, and CD45RA over the 3 yr since primary infection.
# Discussion
Here we have used primary EBV infection in humans as the model in which to ask what fraction of a primary virusinduced T cell response is actually made up of virus-specific (as opposed to coincidentally activated bystander) cells. Our earlier attempts to address this question using two different approaches have given somewhat conflicting results. Thus, in vitro cloning experiments suggested that the frequency of T cells specific for a single EBV epitope was no more than 1:100 (12), whereas a study of TCR usage suggested that the frequency of antigen-specific cells might be as high as . In this study, we have used a novel technique to directly quantify and characterize epitope-specific T cells during the immune response to EBV. We made tetrameric complexes of HLA A2 folded with an EBV lytic cycle epitope, and of HLA B8 folded with an EBV lytic cycle epitope or an EBV latent cycle epitope. Preliminary studies showed that the tetrameric complexes stained T cell clones of the appropriate specificity, but not T cell clones with different specificities. There were also no nonspecific reactivities detected on T cells either from EBV-seropositive but HLA-mismatched donors or from EBV-seronegative HLA matched donors tested as controls. We then went on to use the tetramers to study the circulating T cells in HLA A2-or HLA B8-positive AIM patients.
A very high frequency of EBV epitope-specific T cells was present in peripheral blood in such individuals. Thus peripheral blood T cells reactive with the HLA A2-restricted GLCTLVAML peptide from the EBV lytic cycle antigen BMLF1 were detected in all 10 HLA A2-positive AIM donors tested at levels between 0.5 and 6.6% of CD8 ϩ T cells. Even higher frequencies of T cells reactive with the HLA B8-restricted RAKFKQLL peptide from the EBV lytic cycle antigen BZLF1 were observed in all three HLA B8 ϩ donors studied; in one individual, 44% of circulating CD8 ϩ T cells reacted with the HLA B8-RAKFKQLL tetrameric complex. This represents, in peripheral blood alone, a population of 3ف ϫ 10 9 T cells. In these same three patients, the frequency of T cells specific for the HLA B8-restricted FLRGRAYGL peptide from the EBV latent protein EBNA 3A was lower. The relative immunodominance of the RAKFKQLL (BZLF1) epitope over the FLRGRAYGL (EBNA 3A) epitope during the primary immune response to EBV in these donors is consistent with the results of previously published functional studies.
These numbers of EBV antigen-specific cells, measured directly using tetrameric MHC-peptide complexes, are much greater than the previously reported precursor frequencies, estimated using limiting dilution analysis, in a similar group of patients. In donor 59, we can compare the fre-. Comparison of the frequency and phenotype of antigen-specific CD8 ϩ T cells within PBLs taken from a donor suffering from AIM and 37 mo later. PBLs taken from donor IM63 while he was suffering from AIM (IM63.1; a-e) and 37 mo later (IM63.3; f-j) were stained with the phycoerythrin-conjugated HLA B8-RAKFKQLL tetrameric complex, with a tricolor-conjugated anti-CD8 antibody, and with one of a panel of antibodies conjugated to FITC. Only CD8 ϩ T cells were included in the phenotypic analysis. Staining with the tetrameric complex is shown on the y-axis and staining with normal mouse serum (negative control; a and f), and with an antibody specific for CD38 (b and g), CD45RO (c and h), CD45RA (d and i), or CD28 (e and j) is shown on the x-axis. In each figure, the number of tetramer-reactive T cells that stain with the particular phenotypic marker is shown as percent frequency. quency of FLRGRAYGL-specific CTLs previously estimatedby in vitro outgrowth in a limiting dilution assay (frequency of 1:358) with that estimated in this study by tetramer staining (frequency of 1:50). The comparison shows that in vitro outgrowth underestimates the true magnitude of this latent antigen-induced primary response. We would anticipate a similar situation for lytic cycle epitope-specific T cells, although this has not yet been tested rigorously. To be detectable in a limiting dilution analysis, T cells must be able to survive and proliferate in culture. T cells in a late differentiation compartment (perhaps the majority of the "effector" cells in AIM) are therefore unlikely to be detected using this technique. The proportion of epitopespecific T cells detectable in a limiting dilution assay may depend on the "phenotype" of these cells in vivo.
In individuals with AIM, there was some phenotypic heterogeneity within cell populations, both within an individual patient and between patients, perhaps reflecting differences in the duration of antigen exposure. However, certain common themes emerge. The majority of tetramerreactive cells had an activated/memory phenotype with high levels of expression of CD38, HLA DR, and CD45RO and relatively low levels of expression of CD62L and of CD45RA. By comparison, expression of CD28, the ligand for B7, was extremely variable. In patients IM59 and IM63, we were able to compare CD28 expression on the relatively small population of HLA B8-FLRGRAYGL tetramer-reactive cells with that of the greatly expanded population of HLA B8-RAKFKQLL tetramer-reactive cells. In both donors, the proportion of the FLRGRAYGL-specific T cells expressing CD28 was much higher than the proportion of the RAKFKQLL-specific T cells expressing this molecule. This would be consistent with the idea that CD28 is expressed by naive T cellsand T cells in an early differentiation compartment, but that expression is lost after multiple rounds of cell division in vivo. It has been suggested that CD57 expression may also be a marker for late or terminal differentiation. Consistent with this, analysis of the HLA B8-RAKFKQLL and HLA B8-FLRGRAYGL tetramer-reactive cells in donor IM59 showed that expression of CD57 was more frequent on the greatly expanded population of HLA B8-RAKFKQLLspecific cells.
Although our results clearly show that high numbers of activated, epitope-specific cells are present in peripheral blood during the primary response to EBV, the role of these cells in the control of EBV infection requires further study. We are currently analyzing the expression of fas ligandand perforin (24) by tetramer-reactive cells in order to investigate whether the cells are able to kill target cells and by what mechanisms of cytotoxicity.
After recovery from AIM, the numbers of epitope-specific T cells, as estimated by staining with the HLA-peptide tetrameric complexes, fell in most individuals studied. Here it is interesting to note that the highly expanded lytic cycle epitope-specific CTLs were more heavily culled with transition to the virus-carrier state than the less expanded latent cycle epitope-specific CTLs; this concept of heavier culling of the more abundant components of the primary response was also suggested by the work with functional assays. Nevertheless, populations of both lytic cycle and latent cycle epitope-specific CTLs were still detectable in all the individuals studied at up to 37 mo after primary EBV infection. Comparison of the phenotype of the tetramer-reactive T cells in samples of blood taken from donor IM63 during primary infection and 37 mo later revealed interesting changes in cell surface marker expression. The epitopespecific T cells in the follow-up samples were unexpectedly heterogenous with respect to expression of markers of activation and differentiation. Most significantly, CD45RA was expressed on a substantial subset of tetramer-reactive CD8 ϩ T cells in the follow-up samples, clearly showing that this CD45 isoform is not a specific marker for "naive" T cells within the CD8 ϩ population. This finding is consistent with the results of a recently published functional study of the characteristics of a population of CD27 Ϫ CD8 ϩ T cells that expressed CD45RA.
In conclusion, this study is the first to demonstrate directly the magnitude of the specific primary T cell response induced in vivo by a natural virus infection in humans. The results show that massive expansion of activated, epitope-specific CD8 ϩ T cells can occur during the primary immune response to EBV and that a population of epitope-specific CD8 ϩ T cells persists, in peripheral blood, at a relatively high frequency, for at least 3 yr after primary infection. This work was supported by grants from the Medical Research Council. |
Functional analysis of human olfactory receptors with a high basal activity using LNCaP cell line
Humans use a family of more than 400 olfactory receptors (ORs) to detect odorants. However, deorphanization of ORs is a critical issue because the functional properties of more than 80% of ORs remain unknown, thus, hampering our understanding of the relationship between receptor function and perception. HEK293 cells are the most commonly used heterologous expression system to determine the function of a given OR; however, they cannot functionally express a majority of ORs probably due to a lack of factor(s) required in cells in which ORs function endogenously. Interestingly, ORs have been known to be expressed in a variety of cells outside the nose and play critical physiological roles. These findings prompted us to test the capacity of cells to functionally express a specific repertoire of ORs. In this study, we selected three cell lines that endogenously express functional ORs. We demonstrated that human prostate carcinoma (LNCaP) cell lines successfully identified novel ligands for ORs that were not recognized when expressed in HEK293 cells. Further experiments suggested that the LNCaP cell line was effective for functional expression of ORs, especially with a high basal activity, which impeded the sensitive detection of ligandmediated activity of ORs. This report provides an efficient functional assay system for a specific repertoire of ORs that cannot be characterized in current cell systems.OPEN ACCESSCitation: Ieki T, Yamanaka Y, Yoshikawa K (2022) Functional analysis of human olfactory receptors with a high basal activity using LNCaP cell line. PLoS ONE 17(4): e0267356. https://doi.org/ 10.
# Introduction
Olfaction, the sense of smell, plays an important role in recognizing odors and inspiring preferences for products. The discovery of olfactory receptors (ORs) in olfactory sensory neurons (OSNs) has initiated the molecular era of olfactory research [bib_ref] A novel multigene family may encode odorant receptors: A molecular basis for..., Buck [/bib_ref]. Humans use a family of approximately 400 ORs, each of which recognizes a unique subset of odorants with distinct affinities [bib_ref] Extreme expansion of the olfactory receptor gene repertoire in African elephants and..., Niimura [/bib_ref]. To date, 12% of human ORs have been functionally characterized and matched with their cognate ligands [bib_ref] Human olfactory receptor responses to odorants, Mainland [/bib_ref] [bib_ref] Odor coding by a mammalian receptor repertoire, Saito [/bib_ref] [bib_ref] Identification of agonists for a group of human odorant receptors, Gonzalez-Kristeller [/bib_ref] [bib_ref] The missense of smell: Functional variability in the human odorant receptor repertoire, Mainland [/bib_ref] [bib_ref] Genetic variation across the human olfactory receptor repertoire alters odor perception, Trimmer [/bib_ref] [bib_ref] Rapid deorphanization of human olfactory receptors in yeast, Yasi [/bib_ref]. This progress has not only furthered the understanding of olfactory coding but also provided opportunities for industrial applications. The number of patents on ORs has drastically increased over the past three decades. However, recent advances in understanding olfaction and technological development seem to have reached a plateau, probably due to a large number of uncharacterized ORs. The understanding of the functions of OR has progressed slowly, and a well-recognized reason is the difficulty in expressing ORs on the surface of heterologous cells. Most exogenous OR proteins are retained in the endoplasmic reticulum, following which they are degraded [bib_ref] Endoplasmic reticulum retention, degradation, and aggregation of olfactory G-protein coupled receptors, Lu [/bib_ref]. Several approaches have been successfully established to promote cell surface expression of ORs in heterologous cell systems. Fusing an N-terminal epitope tag from bovine rhodopsin improves the membrane trafficking of OR proteins [bib_ref] RTP family members induce functional expression of mammalian odorant receptors, Saito [/bib_ref]. Co-expression of chaperones, such as receptor transporting protein 1 (RTP1) and other G protein-coupled receptors (GPCRs), in OSNs promotes membrane trafficking of ORs in heterologous cells [bib_ref] RTP family members induce functional expression of mammalian odorant receptors, Saito [/bib_ref] [bib_ref] Receptor-transporting protein 1 short (RTP1S) mediates translocation and activation of odorant receptors..., Wu [/bib_ref] [bib_ref] Activation state of the M3 muscarinic acetylcholine receptor modulates mammalian odorant receptor..., Li [/bib_ref]. Signal amplifiers are also effective in detecting weak levels of signal transduction derived from a small number of cell surface ORs [bib_ref] Ric-8B promotes functional expression of odorant receptors, Von Dannecker [/bib_ref] [bib_ref] Myr-Ric-8A enhances Gα15-mediated Ca2+ response of vertebrate olfactory receptors, Yoshikawa [/bib_ref]. Although these approaches resulted in the deorphanization of a significant number of ORs, more than 80% of ORs are still orphan [bib_ref] Genetic variation across the human olfactory receptor repertoire alters odor perception, Trimmer [/bib_ref] , indicating that heterologous cells lack additional factor(s) required for the functional expression of ORs.
A promising approach to overcome this difficulty is the use of homologous systems that naturally equip the cell with factors required for cell surface expression and signal transduction of the entire repertoire of ORs [bib_ref] Functional expression of a mammalian odorant receptor, Zhao [/bib_ref] [bib_ref] Functional identification and reconstitution of an odorant receptor in single olfactory neurons, Touhara [/bib_ref]. However, the success achieved through this approach has been limited, especially for human ORs. Genetic manipulation of animals, in which defined OR-expressing neurons are labeled with fluorescent probes or in which an activitydependent fluorescent reporter is expressed in OSNs, has identified pairs of ORs and their ligands [bib_ref] In vivo identification of eugenol-responsive and muscone-responsive mouse odorant receptors, Mcclintock [/bib_ref] [bib_ref] Ultrasensitive detection of amines by a trace amine-associated receptor, Zhang [/bib_ref] [bib_ref] Odorant receptor map in the mouse olfactory bulb: In vivo sensitivity and..., Oka [/bib_ref] [bib_ref] Olfactory receptor and neural pathway responsible for highly selective sensing of musk..., Shirasu [/bib_ref] [bib_ref] Mammalian-specific OR37 receptors are differentially activated by distinct odorous fatty aldehydes, Bautze [/bib_ref]. However, these approaches are low throughput, and they cannot be easily applied to the characterization of many human ORs. Application of recent approaches for the in vivo identification of OR repertoires has been restricted to mouse ORs [bib_ref] Large-scale transcriptional profiling of chemosensory neurons identifies receptor-ligand pairs in vivo, Von Der Weid [/bib_ref] [bib_ref] Molecular profiling of activated olfactory neurons identifies odorant receptors for odors in..., Jiang [/bib_ref]. An immortalized cell line derived from the OSN lineage appears to be a powerful tool to functionally express a given OR; however, it has not been widely utilized probably due to genetic instability and difficulty in preventing differentiation and maintaining the molecular characteristics of normal OSN [bib_ref] An olfactory sensory neuron line, odora, properly targets olfactory proteins and responds..., Murrell [/bib_ref].
Interestingly, evidence suggests that the expression of ORs is not only restricted to OSNs but is also apparent in non-olfactory tissues. This expression has been observed in a variety of tissues, such as the heart, lung, sperm, skin, and cancerous tissues, and cell lines [bib_ref] Therapeutic potential of ectopic olfactory and taste receptors, Lee [/bib_ref] [bib_ref] Expression profile of ectopic olfactory receptors determined by deep sequencing, Flegel [/bib_ref] [bib_ref] Identification of a testicular odorant receptor mediating human sperm chemotaxis, Spehr [/bib_ref] [bib_ref] Functional characterization of a mouse testicular olfactory receptor and its role in..., Fukuda [/bib_ref] [bib_ref] Oxygen regulation of breathing through an olfactory receptor activated by lactate, Chang [/bib_ref] [bib_ref] Enterochromaffin cells are gut chemosensors that couple to sensory neural pathways, Bellono [/bib_ref] [bib_ref] Olfactory receptor responding to gut microbiotaderived signals plays a role in renin..., Pluznick [/bib_ref] [bib_ref] MOR23 promotes muscle regeneration and regulates cell adhesion and migration, Griffin [/bib_ref]. The repertoire of non-nasally expressed ORs appears to vary between tissues [bib_ref] Expression profile of ectopic olfactory receptors determined by deep sequencing, Flegel [/bib_ref]. At least a part of the non-nasal ORs have been proven to be functional and play a physiological role, indicating that the non-nasal cells have a molecular machinery that sufficiently promotes the functional expression of a distinct repertoire of ORs [bib_ref] Identification of a testicular odorant receptor mediating human sperm chemotaxis, Spehr [/bib_ref] [bib_ref] Functional characterization of a mouse testicular olfactory receptor and its role in..., Fukuda [/bib_ref] [bib_ref] Oxygen regulation of breathing through an olfactory receptor activated by lactate, Chang [/bib_ref] [bib_ref] Enterochromaffin cells are gut chemosensors that couple to sensory neural pathways, Bellono [/bib_ref] [bib_ref] Olfactory receptor responding to gut microbiotaderived signals plays a role in renin..., Pluznick [/bib_ref] [bib_ref] MOR23 promotes muscle regeneration and regulates cell adhesion and migration, Griffin [/bib_ref] [bib_ref] Olfactory receptor OR2AT4 regulates human hair growth, Chéret [/bib_ref]. Therefore, we conducted this study to test a series of potential cell lines that can endogenously and functionally express ORs that can be used as a novel functional expression system for human ORs.
# Materials and methods
## Expression vector
Human ORs and other genes were cloned as described earlier [bib_ref] Involvement of the olfactory system in the induction of anti-fatigue effects by..., Saito [/bib_ref] [bib_ref] Ligand specificity and evolution of mammalian musk odor receptors: Effect of single..., Sato-Akuhara [/bib_ref]. Briefly, all human OR genes and trace amine-associated receptor (TAAR) genes were amplified from human genomic DNA (Promega, Madison, WI, USA). The identified single nucleotide polymorphisms (SNPs) that were different from the reference sequences in GenBank (http://www.ncbi.nlm. nih.gov/genbank/) or HORDE (http://genome.weizmann.ac.il/horde/) but were found in the NCBI dbSNP database (http://www.ncbi.nlm.nih.gov/SNP/) were not modified (S1 . When we amplified an OR gene with unknown missense mutations, we modified them to reference sequences. Each of the amplified receptor genes was inserted into the pME18S vector to generate OR proteins fused with N-terminal epitope tags of FLAG and 20 N-terminal amino acids of bovine rhodopsin. For the primary screening, which was performed on 378 of the 412 ORs, we screened HEK293 cells against each of the 6 odorants (S2 , and the remaining 34 ORs were tested during later screening (S3 [fig_ref] Table 1: ORs for each odorant a [/fig_ref] , and inserted into pME18S without any N-terminal epitope tag. The human phosphodiesterase 1C (PDE1C, NM_001191059) gene was amplified from cDNA derived from the human lung (BioChain Institute, Newark, CA, USA). cDNA of M3AchR was purchased from Thermo Fisher Scientific (Waltham, MA, USA), and the ORF was inserted into pME18S containing both the N-terminal epitope tags of FLAG and 20 amino acids of bovine rhodopsin.
## Cells
HEK293 cells were provided by Prof. Touhara Kazushige at the University of Tokyo; HepG2, HuH7, and LNCaP cells were purchased from ECACC, JCRB cell bank (JCRB0403), and ATCC (LNCaP clone FGC, ATCC CRL-1740), respectively. HEK293 and HepG2 cells were grown in Dulbecco's modified Eagle's medium (DMEM, 4.5 g/L glucose, Nacalai Tesque, Kyoto, Japan) supplemented with 10% fetal bovine serum (FBS). Antibiotic-antimycotic (1%, 100X, Thermo Fisher Scientific, Waltham, MA, USA) was applied for HepG2 cells. HuH7 cells were grown in DMEM (low glucose, pyruvate, Thermo Fisher Scientific. Waltham, MA, USA) supplemented with 10% FBS, and LNCaP cells were grown in RPMI1640 medium (ATCC modification, Thermo Fisher Scientific, Waltham, MA, USA) supplemented with 10% FBS. All cells were cultured at 37˚C in a humidified atmosphere containing 5% CO 2 . Cell cultures were split using 0.25% trypsin-EDTA (Invitrogen, Waltham, MA, USA) every two to six days before reaching confluence.
## Camp response element (cre)-regulated luciferase reporter gene assay
Transfection and luciferase assays were performed as previously described with some modifications [bib_ref] Involvement of the olfactory system in the induction of anti-fatigue effects by..., Saito [/bib_ref] [bib_ref] Ligand specificity and evolution of mammalian musk odor receptors: Effect of single..., Sato-Akuhara [/bib_ref]. For the 96-well plate assay for identifying ORs for six odorants in HEK293 cells, 0.075 μg of a FLAG-Rho-tagged OR pME18S vector, 0.03 μg of CRE/luc2PpGL4.29 (CRE-dependent firefly luciferase), 0.03 μg of pRL-CMV (constitutively expressed Renilla luciferase), and 0.03 μg of RTP1S pME18S vector per well were transfected in HEK293 cells. To compare between HEK293 cells and other three cell lines, 0.075 μg of a FLAG-Rho-tagged OR pME18S vector, 0.015 μg of CRE/luc2PpGL4.29, 0.015 μg of pRL-CMV, 0.03 μg of RTP1S pME18S, and 0.03 μg of Gαolf pME18S vector per well were transfected into each cell line with the transfection reagent in poly D-lysine-coated 96-well plates (Corning, NY, USA). When cotransfecting with PDE1C, the volume of pRL-CMV was reduced by the same amount, to maintain a constant total volume of the transfected plasmid. The transfection reagent differed depending on the cell type. For HEK293 cells, 0.41 μL/well of 0.1% polyethylenimine Max (PEI-MAX, Polysciences, Warrington, PA, USA) was applied. PEI-MAX was diluted with distilled water, and its pH was adjusted to 7.4. Further, 0.2 μL/well and 0.35 μL/well of Lipofectamine 2000 (Thermo Fisher Scientific, Waltham, MA, USA) were used for HuH7 and HepG2 cells, respectively, and 0.15 μL/well of Lipofectamine 3000 (Thermo Fisher Scientific, Waltham, MA, USA) was used for LNCaP cells. We added 2 μL of P3000 reagent per 1 μg plasmid when we used Lipofectamine 3000. As a buffer for transfection reagents and plasmids, DMEM was used for HEK293 and HuH7 cells, and Opti-MEM was used for HepG2 and LNCaP cells. After incubation for 20 min, 90 μL of the cell suspension (2 × 10 5 cells/cm 2 in each growth medium) was added to 10 μL of the above transfection solution per well, and the plate was incubated at 37˚C in 5% CO 2 for 24 h.
For the 384-well plate assay, 0.029 μg of a FLAG-Rho-tagged OR pME18S vector, 0.022 μg of CRE/luc2PpGL4.29, 0.0011 μg of pRL-CMV, and 0.012 μg of RTP1S pME18S vectors per well were transfected into HEK293 cells with the transfection reagent in poly D lysine-coated 384-well black plates (Corning). Following this, 0.029 μg of a FLAG-Rho-tagged OR pME18S, 0.011 μg of CRE/luc2PpGL4.29, 0.0011 μg of pRL-CMV, 0.012 μg of RTP1S pME18S, and 0.010 μg of Gαolf pME18S vectors per well were transfected into the other three cell lines. The presence or absence of Gαolf was determined based on our previous results that Gαolf did not improve detection of OR-mediated cAMP response in HEK293 cells but did so in other types of cells [bib_ref] Myr-Ric-8A enhances Gα15-mediated Ca2+ response of vertebrate olfactory receptors, Yoshikawa [/bib_ref] [bib_ref] An unsaturated aliphatic alcohol as a natural ligand for a mouse odorant..., Yoshikawa [/bib_ref] [bib_ref] Identification of specific ligands for orphan olfactory receptors: G protein-dependent agonism and..., Shirokova [/bib_ref]. Thus, cell lines other than HEK293 cells were tested with Gαolf. Regarding transfection reagents, 0.16 μL/well of PEI-MAX (0.1%, pH 7.5) was added to HEK293 cells, 0.14 μL/well or 0.077 μL/well of Lipofectamine 2000 was applied to HepG2 or HuH7 cells, respectively, and Lipofectamine 3000 (0.058 μL/well) was added to LNCaP cells. Identical buffers, as described above, were used to dilute the transfection reagents and plasmids. After incubation for 20 min, 40 μL of cell suspension (2 × 10 5 cells/cm 2 in each growth medium) was added to 4.4 μL of the above transfection solution per well, and the solution was incubated at 37˚C in 5% CO 2 for 24 h. The 384-well plate assay was performed using the BiomekFX laboratory automation system (Beckman Coulter, Brea, CA, USA).
After 24 h of transfection, the medium was removed, and the transfected cells were treated with an odorant solution diluted in the growth medium without FBS. The odorant solution diluted in Ringer's solution (140 mM NaCl, 5 mM KCl, 1 mM MgCl 2, 2 mM CaCl 2, 10 mM HEPES, 5 mM glucose, pH 7.4) was also used for HEK293 cells to reduce the background signal from transfected ORs. The 96-or 384-well plates were sealed, and they were incubated at 37˚C for 4 h. The luciferase reporter gene activities were measured using a Dual-Glo Luciferase Assay System (Promega, Madison, USA), and luminescence was measured with Mithras LB940 (Berthold Technologies, Bad Wildbad, Germany) for the 96-well plate and with Ensight multimode plate reader (PerkinElmer, Waltham, MA, USA) for the 384-well plate.
Odorants l-Carvone, anis aldehyde, methyl β-naphthyl ketone, d-carvone, and linalool were purchased from FUJIFILM Wako Pure Chemicals (Osaka, Japan), and cis-3-hexenol was purchased from Merck (Darmstadt, Germany). These six odorants were used in both single and mixed solutions. For primary screening of ORs, the tested concentrations were as follows: l-carvone (300 μM), cis-3-hexenol (1 mM), anis aldehyde (3 mM), methyl β-naphthyl ketone (100 μM), d-carvone (300 μM), and linalool (3 mM). The mixture comprised 100 mM l-carvone, cis-3-hexenol, methyl β-naphthyl ketone, d-carvone, 300 mM anis aldehyde, and linalool in EtOH, and it was diluted with the assay medium for stimulation. The final concentrations of odorant mixture applied to OR screening were as follows: l-carvone (100 μM), cis-3-hexenol (100 μM), anis aldehyde (100 μM), methyl β-naphthyl ketone (100 μM), d-carvone (300 μM), and linalool (300 μM). For the dose-response analyses of odorants, a stock solution (1 M) of each compound was prepared in EtOH and subsequently diluted with the growth medium without FBS. Control stimulation without an odorant employed the highest concentration of EtOH used for each odorant (i.e., the solution for stimulation with 0 μM l-carvone contained 0.3% EtOH).
## Immunocytochemistry (live-cell staining)
A 35-mm dish (IWAKI, Shizuoka, Japan) was coated with 0.01% poly D-lysine PBS solution for 15 min. After incubation, the dish was washed with distilled water thrice and then used. 1.5 μg of a FLAG-Rho-tagged OR pME18S vector and 0.6 μg of RTP1S pME18S plasmids were transfected. Following this, 8.4 μL of 0.1% PEI-MAX solution was added to HEK293 cells, and 3.04 μL of Lipofectamine 3000 and 2 μL of P3000 reagent per 1 μg plasmid were used for LNCaP cells. The plasmids and transfection solutions were mixed in 200 μL Opti-MEM (Invitrogen, Waltham, MA, USA), and incubated for 20 min. After incubation, 1 mL of cell suspension (2 × 10 5 cells/cm 2 in each growth medium) was added to 200 μL of the aforementioned transfection solution per dish, and the solution was incubated at 37˚C in 5% CO 2 for 24 h.
After 24 h of transfection, the dish was placed on ice until PBS was added. First, the medium was removed, and the dish was incubated with a mouse anti-FLAG antibody (1:1000, KO602-S, Trans genic Inc. Fukuoka, Japan) for 1 h. The medium used for HEK293 cells was DMEM with 10% FBS and 10 mM HEPES (pH 7.5) and that for LNCaP cells was RPMI1640 with 10% FBS. After washing with Ringer's solution thrice, the dish was incubated with a goat anti-mouse IgG-Alexa Fluor 488 antibody (1:500, Invitrogen, Waltham, MA, USA) for 1 h. The dish was then washed again with Ringer's solution thrice. After washing, the cells in the dish were fixed with 1% PFA/PB for 15 min. Finally, the dish was washed with PBS three times, and PBS was allowed to remain in the dish. Fluorescent images were obtained using a microscope (BZ-X700, Keyence, Osaka, Japan).
# Data analysis
Data analysis was performed using Microsoft Excel or GraphPad Prism software. Fold increase was calculated as Luc(S) divided by Luc(NS), where Luc(S) was the luminescence intensity of firefly luciferase divided by the luminescence intensity of Renilla luciferase of a certain odorant-stimulated well and Luc(NS) was the luminescence intensity of firefly luciferase divided by the luminescence intensity of Renilla luciferase of a certain non-stimulated well. We excluded the data with higher concentrations of odorants which induced cell toxicity, defined as statistically significant reduction of the luminescence intensity of Renilla luciferase compared with that of non-stimulated control cells (P<0.05, Student's t-test). Data were fitted to a sigmoidal curve using the GraphPad Prism software.
# Results
Six odorants commonly used as perfumery raw materials, namely l-carvone for mint, d-carvone for caraway seeds, cis-3-hexenol for green, anis aldehyde for spicy/anisic, methyl βnaphthyl ketone for rose, and linalool for citrus, were selected. First, we used HEK293 cells, the most common heterologous cell system, to investigate ORs that were sensitive to odorants [bib_ref] Human olfactory receptor responses to odorants, Mainland [/bib_ref] [bib_ref] Odor coding by a mammalian receptor repertoire, Saito [/bib_ref] [bib_ref] Identification of agonists for a group of human odorant receptors, Gonzalez-Kristeller [/bib_ref] [bib_ref] The missense of smell: Functional variability in the human odorant receptor repertoire, Mainland [/bib_ref] [bib_ref] Identification of ligands for olfactory receptors by functional expression of a receptor..., Krautwurst [/bib_ref]. As an assay platform for ORs, we selected a CRE-regulated luciferase reporter gene assay, which is the most commonly used for detecting cAMP signals mediated by activated ORs in HEK293 cells [bib_ref] Human olfactory receptor responses to odorants, Mainland [/bib_ref] [bib_ref] Odor coding by a mammalian receptor repertoire, Saito [/bib_ref] [bib_ref] The missense of smell: Functional variability in the human odorant receptor repertoire, Mainland [/bib_ref] [bib_ref] Involvement of the olfactory system in the induction of anti-fatigue effects by..., Saito [/bib_ref] [bib_ref] Ligand specificity and evolution of mammalian musk odor receptors: Effect of single..., Sato-Akuhara [/bib_ref] [bib_ref] Odorant response assays for a heterologously expressed olfactory receptor, Katada [/bib_ref]. For the primary screening, HEK293 cells were transfected with 378 ORs and stimulated with a single concentration of each of the six odorants [fig_ref] Fig 1: HEK293 cell-based screening of human ORs for six perfumery raw materials [/fig_ref]. The response of cells expressing each OR was presented as fold increase, where a signal from stimulated cells was divided by that from non-stimulated cells expressing the same OR. We then selected ORs that showed response amplitudes (fold increase) above the mean+SD among all the examined ORs in the primary screening and conducted follow-up dose-response analysis [fig_ref] Fig 1: HEK293 cell-based screening of human ORs for six perfumery raw materials [/fig_ref]. We concluded that an OR is a receptor for the odorant if at least one concentration of the odorant produced a statistically significant response in ORexpressing cells as compared to both the OR-expressing cells stimulated with medium alone and vector-transfected control cells stimulated with the concentration of the odorant (Sidak-Bonferroni method with alpha = 5.0%). The identified ORs were as follows: OR2W1, OR5K1, OR5P3, OR8B3, OR8H1, and OR10A6 for l-carvone; OR1A1, OR5P3, and OR10A6 for d-carvone; OR1A1, OR2J3, OR2W1, OR4S2, OR5P3, and OR10A6 for cis-3-hexenol; OR1A1, OR2J2, OR2J3, OR5K1 OR5P3, OR8B3, and OR10A6 for anis aldehyde; OR1A1, OR1D2, OR2J2, OR2W1, OR5P3, OR8B3, OR8H1, and OR10A6 for methyl β-naphthyl ketone; and OR1A1 and OR1C1 for linalool. Negative ORs that were tested in this follow-up dose-response analysis but did not show reproducible responses were as follows: OR51A7 for l-carvone and d-carvone; OR2J2 for cis-3-hexenol; OR13C2, OR10G4, and OR1L6 for anis aldehyde; and OR11H6, OR10Q1, and OR2S2 for methyl β-naphthyl ketone (S2 . These data not only provided 25 novel pairs of ORs and their ligands (white circles in [fig_ref] Table 1: ORs for each odorant a [/fig_ref] , contributing to future studies to understand the relationship between receptor function and perception, but also provide a basis for evaluating the capability of novel heterologous cell systems other than HEK293 cells.
We then examined whether the three cell lines, HepG2 (human hepatocellular carcinoma cells), HuH7 (human hepatocellular carcinoma cells), and LNCaP (human prostate carcinoma cells), all of which express functional ORs, were effective as an alternative expression system in comparison with HEK293 cells. OR1A1 is expressed in HepG2 cells, and it responds to l- A. Primary screening of human ORs using 384-well plates. 378 ORs were expressed in HEK293 cells, and they were stimulated with a single concentration of each odorant. HEK293 cells were co-transfected with each OR, CRE/ luc2PpGL4.29, pRL-CMV, and RTP1S. Odorant solutions were diluted in the growth medium without FBS. X-axis lists each of the 378 ORs, and color bars represent OR families. Y-axis indicates the response of cells expressing each OR as fold increase where a signal from stimulated cells was divided by that from nonstimulated cells expressing the same OR (mean values from two screening replicates). B. Verification of dose-dependent responsiveness of cell-expressing candidate ORs using 96-well plates. Each OR was co-transfected with CRE/luc2PpGL4.29, pRL-CMV, and RTP1S. 32 pairs of ORs and their agonists that meet our statistical criteria for dose-dependent responses are shown. The grey-shaded ORs do not fit a sigmoidal curve because their responses were not saturated in the range of concentrations tested due to less sensitivity. Y-axis indicates the fold increase value, and data are shown as mean ± SE from three independent experiments. https://doi.org/10.1371/journal.pone.0267356.g001 carvone [bib_ref] Activation of OR1A1 suppresses PPAR-γ expression by inducing HES-1 in cultured hepatocytes, Wu [/bib_ref]. OR1A2 is expressed in HuH7 cells, and it is sensitive to (-)-citronellal and citronellol [bib_ref] Monoterpene (-)-citronellal affects hepatocarcinoma cell signaling via an olfactory receptor, Maßberg [/bib_ref]. LNCaP cells express OR51E1 and OR51E2 (also named prostate-specific G-protein-coupled receptor [PSGR]), which recognize β-ionone, propionic acid, butyric acid, and nonanoic acid [bib_ref] Activation of an olfactory receptor inhibits proliferation of prostate cancer cells, Neuhaus [/bib_ref] [bib_ref] Ectopically expressed olfactory receptors OR51E1 and OR51E2 suppress proliferation and promote cell..., Pronin [/bib_ref]. As HepG2, HuH7, and LNCaP have not been widely utilized as functional expression systems for GPCRs, we first screened several experimental conditions, such as transfection reagents and culture buffers, based on the responsiveness of OR1A1 (S1 .
Under optimized conditions, we tested whether the three cell lines could functionally express a unique subset of ORs, including those that could not be expressed in HEK293 cells. The cell lines were transfected with 412 ORs, including TAARs (see Materials and Methods), and they were stimulated with a mixture of the 6 odorants. Measuring activation against the odorant mixture suggested that each cell line was capable of functionally expressing a distinct set of ORs [fig_ref] Fig 2: Capacity of three cell lines as a functional expression cell system for... [/fig_ref]. We tentatively identified activated ORs based on the criterion that the response amplitude (fold increase) should be above mean + 2SD of all the examined ORs except for those validated in HEK293 cells [fig_ref] Fig 1: HEK293 cell-based screening of human ORs for six perfumery raw materials [/fig_ref] ORs sensitive to each component of the mixture in HEK293 cells). Then we conducted follow-up experiments testing higher and lower concentrations of the odorant mixture (S3 As a result, ORs were identified as positive if at least one concentration of the odorant produced a statistically significant response of an OR-expressing cell as compared to that of cells stimulated with medium alone and a vectortransfected control cell (Sidak-Bonferroni method with alpha = 5.0%), labeled in [fig_ref] Fig 2: Capacity of three cell lines as a functional expression cell system for... [/fig_ref] We identified OR1A1, OR1D2, OR2J2, OR2J3, OR2W1, OR5P3, and OR8B3 as ORs sensitive to the mixture using HEK293 cells, which included 7 out of 12 ORs identified in [fig_ref] Fig 1: HEK293 cell-based screening of human ORs for six perfumery raw materials [/fig_ref] Some of the ORs that were activated against each component of the mixture [fig_ref] Fig 1: HEK293 cell-based screening of human ORs for six perfumery raw materials [/fig_ref] were undetectable in this experiment likely due to the low concentration of the tested component in the mixture (see Materials and Methods) and/or antagonistic effects [bib_ref] Identification of a testicular odorant receptor mediating human sperm chemotaxis, Spehr [/bib_ref] [bib_ref] Widespread inhibition, antagonism, and synergy in mouse olfactory sensory neurons in vivo, Inagaki [/bib_ref] [bib_ref] Odorant receptor inhibition is fundamental to odor encoding, Pfister [/bib_ref] [bib_ref] Olfactory receptor antagonism between odorants, Oka [/bib_ref]. The same experiments using three other cell lines resulted in the identification of OR1A1, OR2J2, and OR10A6 from HepG2; OR1A1, OR5P3, and 9Q1 from HuH7; and OR1A1, OR2W1, OR5P3, 11L1, and OR51T1 from LNCaP [fig_ref] Fig 2: Capacity of three cell lines as a functional expression cell system for... [/fig_ref]. The response amplitudes and errors of the identified ORs in three cell lines were smaller than those in HEK293 cells, suggesting a difference in capacity of cAMP metabolism between cell lines. These data provide evidence of a unique pattern of ORs functionally expressed in each cell line. Notably, activation of OR9Q1 and OR51T1, which are orphan receptors, was specifically detected in HuH7 cells or LNCaP cells but not in HEK293 cells [fig_ref] Fig 1: HEK293 cell-based screening of human ORs for six perfumery raw materials [/fig_ref]. This result suggests that these cell lines are candidate functional expression systems for ORs non-functional in HEK293 cells. Here, we focused on OR51T1 and LNCaP cell pairs with the most robust response among those of the orphan ORs.
When the odorant mixture was tested at high concentrations, OR51T1 activation was still not monitored in HEK293 cells; however, it was robustly detected in LNCaP cells [fig_ref] Fig 3: Activations of specific ORs were selectively monitored in LNCaP cells [/fig_ref]. This result was in contrast with that for OR1A1, where HEK293 cells as well as LNCaP cells showed an OR1A1-mediated response in a dose-dependent manner. A potential reason for LNCaPselective detection of OR51T1 activation was co-expression of Gαolf. We expressed Gαolf in three cell lines including LNCaP cells but did not in HEK293 cells because previous studies indicated that co-expression of Gαolf improved detection of OR activation from specific cell types, except for HEK293 cells [bib_ref] Myr-Ric-8A enhances Gα15-mediated Ca2+ response of vertebrate olfactory receptors, Yoshikawa [/bib_ref] [bib_ref] An unsaturated aliphatic alcohol as a natural ligand for a mouse odorant..., Yoshikawa [/bib_ref] [bib_ref] Identification of specific ligands for orphan olfactory receptors: G protein-dependent agonism and..., Shirokova [/bib_ref]. However, co-expression of Gαolf was not required for sensitive detection of OR activation in LNCaP cells, indicating that the observed difference in activation of ORs expressed in HEK293 cells and LNCaP cells is due to an inherent capacity of the cells rather than the presence or absence of exogenous Gαolf (S7 . Following this, we tested six components of the odorant mixture. OR51T1-expressing LNCaP cells showed a significant dose-dependent response to all components of the mixture, leading to the identification of novel pairs of odorant-ORs that could not be achieved using HEK293 cells [fig_ref] Fig 4: Identification of novel pairs of ligands and ORs in LNCaP cells [/fig_ref]. demonstrates the deorphanization of OR51T1, and it supports the idea that LNCaP is an efficient cell line for functional expression of at least one OR. Next, to understand the range of applicable ORs and factors that are effective for functional expression in LNCaP cells, we addressed the mechanism by which LNCaP cells allowed the detection of OR51T1 activation. One potential reason was that LNCaP cells efficiently promoted the cell surface expression of OR51T1, but HEK293 cells did not. However, OR51T1 was detected on the surface of HEK293 as well as LNCaP cells in live-cell staining experiments to detect the N-terminal FLAG-tag of OR51T1 proteins . We also found that the basal signal level from OR51T1 under no ligand stimulation was significantly higher in HEK293 cells than in LNCaP cells . This difference in OR51T1-derived basal signal between HEK293 and LNCaP cells seemed to be attributed to the difference in the capacity of cAMP production in each cell because forskolin, a potent adenylate cyclase activator, induced higher cAMP-mediated response in HEK293 cells than in LNCaP cells . However, the difference in OR51T1-mediated basal signal from each cell line was not proportional to the capacity of cAMP production, and they were still clear when the basal signals (CRE-Luc ratio, in response to 10 μM forskolin were normalized to 100% . Thus, HEK293 cells produced large basal signals from a specific repertoire of ORs as reported previously [bib_ref] Agonist-independent GPCR activity regulates anterior-posterior targeting of olfactory sensory neurons, Nakashima [/bib_ref] , and a high basal signal may impede the sensitive detection of ligand-mediated response of OR51T1. A reverse experiment supported this notion. Treatment with 3-isobutyl-1-methylxanthine (IBMX), an inhibitor for phosphodiesterase (PDE), allowed LNCaP cells to produce larger OR51T1-mediated basal signal and enhanced forskolin-mediated response of mock-transfected LNCaP cells more robustly than HEK293 cells . Notably, IBMXtreated LNCaP cells reconstituted the HEK293 cell-like pattern of OR activations to the odorant mixture: responses of LNCaP-specific ORs with high basal activity (i.e. OR51T1)
[formula] - 3,4) O - - 3) O O O d-carvone - 3,4) - 4) O ▲ 3) O O cis-3-hexenol O ▲ 3) ▲ 3) - 3) - 3) O O O ▲ 3) O anis aldehyde O O O O O ▲ 3) O O O m-b-n-ketone O O O ▲ 38) O O O O O O linalool O - 3)▲ 3 [/formula]
## Fig 5. capacity of hek293 and lncap cells for membrane trafficking of nascent or proteins and for controlling background intracellular signal from non-stimulated ors.
A. Immunohistochemical analysis (live-cell staining) of HEK293 and LNCaP cells expressing OR51T1 or TAAR1. M3AChR was tested as a positive control of GPCRs for efficient cell-surface expression. All three tested receptors were expressed as fused proteins with an N-terminal FLAG-tag, and they were detected using an anti-FLAG antibody. Scale, 100 μm. B, C. Difference in background signal levels from constitutive activity of ORs in HEK293 (green) and LNCaP (purple) cells. Each type of cell was transfected with CRE/luc2PpGL4.29, pRL-CMV, RTP1S and Gαolf. X-axis indicates transfected ORs. After 24 h of transfection, cells were stimulated with growth medium without odorants, and luciferase activity was measured. Data are presented as CRE-Luc ratio (luminescence intensity of firefly luciferase divided by the luminescence intensity of Renilla luciferase) (B) or %forskolin (C) where CRE-Luc ratio in each transfection condition is presented as normalized values and CRE-Luc ratio in response to 10 μM forskolin was set to 100%. D. Dose-response of were reduced, whereas those of other ORs activated strongly in HEK293 cell were enhanced (i.e. OR2J2, S3 . We analyzed the correlation of response amplitudes of ORs to the odorant mixture in HEK293 cells and LNCaP cells with or without IBMX treatment . For this analysis, we selected the 40 ORs which showed response amplitudes above mean + 2SD from all examined ORs in each cell line of the primary screening [fig_ref] Fig 2: Capacity of three cell lines as a functional expression cell system for... [/fig_ref]. The pattern of OR responses in HEK293 cells showed a weak correlation with that in untreated LNCaP cells, but showed a stronger correlation when LNCaP cells were treated with IBMX (r = 0.50, P = 0.0010 v.s. r = 0.89, P<0.0001). Thus, we reasoned that depleting the background signal allowed HEK293 cells to detect activation of OR51T1 and tested experimental conditions to manipulate the basal level of signals. We tested the effect of coexpression of PDE1C, which was expressed in the cilia of OSNs. We observed that the coexpression of PDE1C induced decreased levels of OR51T1-mediated basal signal in HEK293 cells, suggesting a reduction in the background noise for detection of an agonistmediated OR51T1 activation (P< 0.0001, student's t-test; . PDE1C expressed in LNCaP cells without OR51T1 did not reduce sensitive detection of stimulant-induced cAMP signal of LNCaP cells (P = 0.198, student's t-test; . Under this condition, we tested whether activation of OR51T1 against l-carvone was detectable in HEK293 cells. However, no significant response was measured even when different concentrations of PDE1C were tested . In addition, we also tested the reduction of the amount of transfected plasmid for OR51T1. This approach successfully reduced background signal and led to sensitive detection of ligand-mediated activation of two ORs with high basal activity (OR2W1 and OR51E1), but it was not effective to OR51T1 (S9 These observations highlight the advantage of LNCaP cells, which are effective for the functional characterization of ORs, especially those with a high basal activity in HEK293 cells.
Thus, we searched for another case in which OR with a high basal activity in HEK293 cells was functionally expressed and successfully assayed in LNCaP cells. Our data on background signals from HEK293 cells transfected with each of the 412 ORs confirmed that OR51T1 elicited a relatively high background signal and identified other ORs with a high basal signal [fig_ref] Fig 4: Identification of novel pairs of ligands and ORs in LNCaP cells [/fig_ref]. These data should be interpreted with caution because the majority of OR species were not recruited on the cell surface of HEK293 cells, and the present data missed a substantial number of ORs with high basal activity. Among the identified receptors with a high basal activity, we were particularly interested in TAAR1, a receptor for organic amine compounds, though the human olfactory epithelium does not express TAAR1 at all or only slightly expresses it [bib_ref] The human olfactory transcriptome, Olender [/bib_ref]. We focused on TAAR1 because TAAR1-expressing LNCaP cells in the initial screening experiment seemed to be responsive to the odorant mixture, although the response amplitude did not meet the criteria described above. The response values showed an average fold increase of 2.8 in TAAR1-expressing LNCaP cells as compared to 1.5 in the vector-transfected control [fig_ref] Fig 2: Capacity of three cell lines as a functional expression cell system for... [/fig_ref]. Thus, the same series of experiments on OR51T1 were conducted using TAAR1. CRE-regulated luciferase reporter gene assays indicated that TAAR1 activation against the mixture and its components was detectable only in LNCaP cells, and it revealed novel types of ligands, l-carvone, anis aldehyde, and cis-3-hexenol [fig_ref] Fig 3: Activations of specific ORs were selectively monitored in LNCaP cells [/fig_ref]. The observed agonistic activity of these odorants was significant but weak because higher concentrations were required for detecting activation of TAAR1 than those of known agonists of organic amines (S10 [fig_ref] Fig 3: Activations of specific ORs were selectively monitored in LNCaP cells [/fig_ref]. Both HEK293 and LNCaP cells were capable of recruiting TAAR1 protein on the cell surface , whereas HEK293 cells showed higher background signals when they were expressed with TAAR1 than did LNCaP cells . Taken together, our results suggest that the LNCaP cell line is an effective heterologous cell system for functional characterization of a specific repertoire of ORs with high basal activity.
# Discussion
Since the discovery of the OR gene family in 1991 [bib_ref] A novel multigene family may encode odorant receptors: A molecular basis for..., Buck [/bib_ref] , many studies have highlighted the importance of the functions of individual ORs. Genetically manipulated animals lacking a specific OR showed abnormal behavioral olfactory responses in terms of sensitivity, preference, and social communication [bib_ref] Ligand specificity and evolution of mammalian musk odor receptors: Effect of single..., Sato-Akuhara [/bib_ref] [bib_ref] The male mouse pheromone ESP1 enhances female sexual receptive behaviour through a..., Haga [/bib_ref] [bib_ref] Contribution of individual olfactory receptors to odor-induced attractive or aversive behavior in..., Horio [/bib_ref] [bib_ref] Identification of an Intraand Inter-specific Tear Protein Signal in Rodents, Tsunoda [/bib_ref] [bib_ref] Single olfactory receptors set odor detection thresholds, Dewan [/bib_ref] [bib_ref] An adenosine receptor for olfaction in fish, Wakisaka [/bib_ref] [bib_ref] Olfactory receptor for prostaglandin F2α mediates male fish courtship behavior, Yabuki [/bib_ref]. In humans, taking advantage of the natural knockout of ORs due to genetic polymorphism, a significant function of perception and potential ligands of respective ORs have been suggested [bib_ref] The missense of smell: Functional variability in the human odorant receptor repertoire, Mainland [/bib_ref] [bib_ref] Genetic variation across the human olfactory receptor repertoire alters odor perception, Trimmer [/bib_ref] [bib_ref] Sequence variants in TAAR5 and other loci affect human odor perception and..., Gisladottir [/bib_ref] [bib_ref] A mendelian trait for olfactory sensitivity affects odor experience and food selection, Jaeger [/bib_ref] [bib_ref] Genetic elucidation of human hyperosmia to isovaleric acid, Menashe [/bib_ref] [bib_ref] Genetic variation in a human odorant receptor alters odour perception, Keller [/bib_ref]. However, pairing ORs with cognate ligands has progressed slowly due to their inefficient functional expression in heterologous cells. The significance of the present study is as follows: 1) it provides a novel methodology for functional assay of ORs by utilizing cell lines that endogenously express functional ORs and are effective as heterologous expression systems, and 2) it identifies 31 novel pairs of ORs and their ligands that will potentially contribute to our understanding of how the sense of smell is constructed and technologically controlled in the future.
Prior to investigating the capability of the three cell lines, we used HEK293 cells to perform control experiments to validate the OR and ligand pairs. Previous reports on identification of OR-ligand pairs are useful for evaluating the consistency of our results [bib_ref] Human olfactory receptor responses to odorants, Mainland [/bib_ref] [bib_ref] Odor coding by a mammalian receptor repertoire, Saito [/bib_ref] [bib_ref] Involvement of the olfactory system in the induction of anti-fatigue effects by..., Saito [/bib_ref]. In the present study, we consistently detected the activation of OR2W1, OR5P3, and OR8B3 by l-carvone, OR1A1 by d-carvone, OR2J3 and OR2W1 by cis-3-hexenol, and OR1C1 by linalool (black circles in [fig_ref] Table 1: ORs for each odorant a [/fig_ref]. The present study successfully identified novel OR-ligand pairs, including OR5K1, OR8H1, and OR10A6 for l-carvone; OR5P3 and OR10A6 for d-carvone; OR1A1, OR4S2, OR5P3, and OR10A6 for cis-3-hexenol; OR1A1, OR2J2, OR2J3, OR5K1, OR5P3, OR8B3, and OR10A6 for anis aldehyde; OR1A1, OR1D2, OR2J2, OR2W1, OR5P3, OR8B3, OR8H1, and OR10A6 for methyl β-naphthyl ketone; and OR1A1 for linalool (white circles in [fig_ref] Table 1: ORs for each odorant a [/fig_ref]. However, our primary screening missed OR-ligand pairs identified in previous studies: OR1A1 for l-carvone; OR2W1 and OR8B3 for d-carvone; OR2J1, OR2J2, and OR14J1 for cis-3-hexenol; OR6P1 for anis aldehyde; OR2J3 for methyl β-naphthyl ketone; and OR1N2 for linalool (black triangles in [fig_ref] Table 1: ORs for each odorant a [/fig_ref]. The inconsistency may be attributed to differences in the amino acid sequences of OR variants. As genetic variations in human ORs are abundant, a part of ORs used in this study were variants with different amino acids from their reference sequences. Similarly, previous studies tested a series of ORs which included OR variants with different amino acid residues from those in this study, likely providing different assay outcomes to the same odorants. S5 Table shows a comparison with responsiveness of OR variants reported in previous studies. In addition, the inconsistency with previous studies is likely explained by a difference in tested concentrations. A previous study also conducted an OR screen for five odorants tested in this study: l-carvone, d-carvone, anis aldehyde, cis-3-hexenol, and linalool in a 100 μM concentration for primary screening, whereas our current study used higher concentrations [bib_ref] Human olfactory receptor responses to odorants, Mainland [/bib_ref]. This difference may explain why a large number of ORs were identified in our study. Potential false negative identifications of ORs in our primary screen using 384-well plates may also cause some discrepancy with previous findings. OR14J1 and 6P1 were not included in our plasmid library at the time of primary screening, and therefore, they were not tested. The activated ORs that were missed may be detected in our assay system after testing them at higher concentrations.
The use of LNCaP cells allowed the deorphanization and functional characterization of OR51T1. OR51T1 was found to be sensitive to structurally dissimilar odorants, and therefore, appeared to be classified into broadly tuned ORs [bib_ref] A large-scale analysis of odor coding in the olfactory epithelium, Nara [/bib_ref] [bib_ref] Responsiveness of G protein-coupled odorant receptors is partially attributed to the activation..., Yu [/bib_ref] [bib_ref] How broadly tuned olfactory receptors equally recognize their agonists. Human OR1G1 as..., Charlier [/bib_ref]. The observed high basal activity of OR51T1-expressing cells was also consistent with the reported characteristics of broadly tuned ORs [bib_ref] Responsiveness of G protein-coupled odorant receptors is partially attributed to the activation..., Yu [/bib_ref]. One potential olfactory function of this type of OR is to act as an intensity analyzer by providing an easy readout of odor concentrations regardless of the identity of odors in the olfactory system [bib_ref] Responsiveness of G protein-coupled odorant receptors is partially attributed to the activation..., Yu [/bib_ref]. Future identification of a ligand that specifically activates OR51T1 will allow us to pinpoint its cognate perception and examine its potential function.
Next, agonists for TAAR1 were identified from non-amine compounds, such as l-carvone, anis aldehyde, and cis-3-hexenol. These three odorants showed an extremely weak affinity to TAAR1 because the response of cells did not reach saturation up to a concentration of 3000 μM; therefore, their EC50 values could not be determined. This is inconsistent with known monoamine agonists, which show a higher affinity and efficacy with EC50 values of 0.1-100 μM (S10 [fig_ref] Fig 2: Capacity of three cell lines as a functional expression cell system for... [/fig_ref]. The structural basis underlying the sensitive binding to amines was predicted to involve an essential salt bridge between the amino group of agonists and an aspartic acid on the third transmembrane domain, which is highly conserved in many TAARs, including TAAR5 [bib_ref] Convergent olfactory trace amine-associated receptors detect biogenic polyamines with distinct motifs via..., Jia [/bib_ref]. The observed lower affinities of identified agonists were probably attributed to the lack of a salt bridge in the interaction. However, the present data suggest a statistically significant activation of TAAR1 against non-amine odorants, which was consistent with the binding capacity of a non-amine odorant toward TAAR5 [bib_ref] Timberol® inhibits TAAR5-mediated responses to trimethylamine and influences the olfactory threshold in..., Wallrabenstein [/bib_ref]. Recent studies have further suggested that human TAAR1 is expressed in several brain regions but not in the olfactory epithelium; TAAR1 was proposed as a promising therapeutic target for the treatment of schizophrenia, psychosis in Parkinson's disease, substance abuse, metabolic syndrome, and obesity [bib_ref] Olfactory expression of trace amine-associated receptors requires cooperative cis-acting enhancers, Shah [/bib_ref] [bib_ref] Coordination of two enhancers drives expression of olfactory trace amine-associated receptors, Fei [/bib_ref]. Thus, TAAR1 may not be classified as an OR. Taken together, our results suggest that LNCaP cells have the capacity to functionally express a wide range of receptors with high basal activity, including ORs and other GPCRs outside the nose.
We argue that prostate epithelial cells as well as LNCaP cells may possess appropriate signal to noise ratio of cAMP signal for utilizing ORs with high basal activity, including endogenous OR51E1 and OR51E2 as well as exogenous OR51T1 and TAAR1. On the other hand, HEK293 cells appear to be efficient for analyzing ORs with relatively low basal activity. It is worth noting that assay outcomes in each cell line also depend on the potency of available agonists for target ORs. While the detection of OR51T1 activation was specific in LNCaP cells, OR51E1 with high basal activity could also be analyzed in HEK293 cells (S9 This apparent discrepancy is likely reconciled by the different agonistic potency between nonanoic acid for fatty acid-selective OR51E1 and the identified agonists for broadly tuned OR51T1, which may recognize structurally diverse molecules with weak affinity. Our assumption is that an approach for depleting basal cAMP signal will lead to the optimization of HEK293 cells capable of detecting activation of ORs with a high basal activity. This hypothesis is partially supported by the converse experiment. Up-regulation of basal cAMP signal in LNCaP cells by IBMX treatment provided a similar assay outcome to HEK293 cells. However, our attempts to reduce basal signal in HEK293 cells did not allow the detection of OR51T1 activation potentially because they also reduced the signal from activated ORs. This result highlights the advantage of LNCaP cells for analyzing a specific repertoire of ORs with a high basal activity. Although a previous study and the current study did not detect response of LNCaP cells to agonistic odorants for endogenous OR51E1 and OR51E2 likely due to their low expression levels (S11 [fig_ref] Fig 4: Identification of novel pairs of ligands and ORs in LNCaP cells [/fig_ref] , the potential responsiveness of the endogenous receptors to a subset of odorants may pose a risk in the use of the cell line as a functional assay system for a given exogenous OR. However, a notable advantage of homologous expression systems for characterizing ORs has been proven in this study as well as by a number of previous studies using olfactory sensory neurons [bib_ref] Functional expression of a mammalian odorant receptor, Zhao [/bib_ref] [bib_ref] Functional identification and reconstitution of an odorant receptor in single olfactory neurons, Touhara [/bib_ref] [bib_ref] The importance of odorant conformation to the binding and activation of a..., Peterlin [/bib_ref] [bib_ref] MouSensor: A versatile genetic platform to create super sniffer mice for studying..., D'hulst [/bib_ref]. Future studies will establish a more sensitive condition in LNCaP cells via long-term trial and error optimization of the current assay condition using HEK293 cells.
Several heterologous cell systems, other than HEK293 cells, have been reported previously. For example, vertebrate and insect ORs were successfully analyzed in Xenopus oocytes using electrophysiological methods, and a variety of common mammalian cell lines have also been utilized for the successful characterization of ORs [bib_ref] Odorant receptor map in the mouse olfactory bulb: In vivo sensitivity and..., Oka [/bib_ref] [bib_ref] An unsaturated aliphatic alcohol as a natural ligand for a mouse odorant..., Yoshikawa [/bib_ref] [bib_ref] Identification of specific ligands for orphan olfactory receptors: G protein-dependent agonism and..., Shirokova [/bib_ref] [bib_ref] Odorant response assays for a heterologously expressed olfactory receptor, Katada [/bib_ref] [bib_ref] Insect olfactory receptors are heteromeric ligand-gated ion channels, Sato [/bib_ref]. These attempts demonstrated the high consistency of the observed functionality of ORs, regardless of the type of heterologous cells and assay methodologies. However, the limited number of tested ORs failed to provide an understanding of the range of applicable ORs in each type of heterologous cell. The current report tested the highest number of ORs for comparison of multiple cell lines, and it identified cases of cell line-dependent detection of OR activation. Although we have focused only on the LNCaP cell line and ORs with a high constitutive activity, the present data also include potential ORs functionally expressed only in a distinct cell line (S3 Further investigation using a different set of odorants against ORs expressed in the three cell lines tested here and others will increase the usefulness of cell lines with endogenous ORs as an additional heterologous cell system. Accumulating information on odorant-OR pairs will provide an understanding of how the brain computes odor inputs, constructs the sense of smell, and develops an approach to efficiently control olfactory sense. [fig_ref] Table: Comparison with genetic variants used in previous studies [/fig_ref] OR sequences used in this study. (XLSX) S2 [fig_ref] Table: Comparison with genetic variants used in previous studies [/fig_ref] OR screening using HEK293 cells against six odorants [fig_ref] Fig 1: HEK293 cell-based screening of human ORs for six perfumery raw materials [/fig_ref]. [fig_ref] Table: Comparison with genetic variants used in previous studies [/fig_ref] OR screening using four cell lines against an odorant mixture [fig_ref] Fig 2: Capacity of three cell lines as a functional expression cell system for... [/fig_ref]. (XLSX) S4 [fig_ref] Table: Comparison with genetic variants used in previous studies [/fig_ref] Background basal activity of 412 ORs expressed in HEK293 cells (related to .
## Supporting information
[formula] (XLSX) S3(XLSX) S5 [/formula]
[fig] Fig 1: HEK293 cell-based screening of human ORs for six perfumery raw materials. [/fig]
[fig] Fig 2: Capacity of three cell lines as a functional expression cell system for ORs. A total of 412 ORs were expressed in different cell lines. Each OR was co-transfected with CRE/luc2PpGL4.29, pRL-CMV, and RTP1S in HEK293 cells whereas Gαolf was also co-transfected in the other three cell lines. They were stimulated with a mixture of six odorants in the growth medium without FBS. The X-axis lists each of the 412 ORs, and color bars represent OR families. Y-axis indicates the response of cells expressing each OR as a fold increase. Data are presented as the mean values from two screening replicates. The dashed line denotes mean + 2SD of all the examined ORs except for those validated in HEK293 cells. The name of the receptor is described if the response value in the follow-up dose-response analysis meets our criteria (S3-S6 Figs).https://doi.org/10.1371/journal.pone.0267356.g002 [/fig]
[fig] Fig 3: Activations of specific ORs were selectively monitored in LNCaP cells. Dose-response curves of OR1A1 (top, black), OR51T1 (middle, magenta), and TAAR1 (bottom, blue) expressed in HEK293 (left) or LNCaP cells (right) against increasing concentrations of the odorant mixture. Each OR was co-transfected with CRE/luc2PpGL4.29, pRL-CMV, RTP1S and Gαolf. X-axis indicates final concentrations of each of the six components of the mixture. Ringer's solution was used to dissolve the odorant mixture. Assays were also conducted on mock-transfected cells transfected with empty vector, CRE/luc2PpGL4.29, pRL-CMV, RTP1S, and Gαolf. (white). Yaxis indicates the fold increase in response. Data are shown as mean ± SE of three independent experiments. https://doi.org/10.1371/journal.pone.0267356.g003 [/fig]
[fig] Fig 4: Identification of novel pairs of ligands and ORs in LNCaP cells. Dose-response curves of OR51T1 (A, magenta) and TAAR1 (B, blue) expressed in HEK293 (left) or LNCaP cells (right) against increasing concentrations of each odorant. Each OR in both cell lines was co-transfected with CRE/luc2PpGL4.29, pRL-CMV, RTP1S and Gαolf. Ringer's solution was used to dissolve the odorant mixture. Y-axis indicates the fold increase in response. Data are shown as mean ± SE of three independent experiments. https://doi.org/10.1371/journal.pone.0267356.g004 [/fig]
[fig] S1: Fig. Optimization of transfection conditions for the three cell lines. A-C. Six patterns of transfection conditions, including (i) 0.41 μL of PEI-MAX (condition for HEK293 cells), (ii) 0.2 μL of Lipofectamine 2000, (iii) 0.35 μL of Lipofectamine 2000, (iv) 0.5 μL of Lipofectamine 2000, (v) 0.15 μL of Lipofectamine 3000, and (vi) 0.3 μL of Lipofectamine 3000, were tested.DMEM was used as a transfection buffer for HEK293 cells, and Opti-MEM was used as a transfection buffer for the other three cell lines. Based on the result, the condition adopted for each cell line is highlighted in red. Data are shown as mean ± SE of three independent experiments. D. Buffers for transfection reagent and plasmid for HepG2, HuH7, and LNCaP cells. This figure shows the fold increase values of OR1A1-expressing cells stimulated by l-carvone. Each cell line was tested with Opti-MEM or a growth medium for each cell (DMEM for HepG2 and HuH7 cells and RPMI1640 for LNCaP cells). These data show the average of duplicated wells from one experiment. Based on the result of this experiment, DMEM was used for HEK293 and HuH7, and Opti-MEM was used for HepG2 and LNCaP(Fig 2).(EPS) S2 Fig. Dose-response analysis of candidate ORs for each odorant (related to Fig 1). HEK293 cells were transfected with each ORs, as well as CRE/luc2PpGL4.29, pRL-CMV, and RTP1S and stimulated with odorants. The response of cells expressing each OR is presented as fold increase. Data are shown as mean ± SE of three independent experiments. (EPS) S3 Fig. Dose-response of OR-expressing HEK293 cells to the odorant mixture. HEK293 cells were transfected with each OR, CRE/luc2PpGL4.29, pRL-CMV, Gαolf, and RTP1S and stimulated with multiple concentrations of the odorant mixture. X-axis indicates concentration of the odorant mixture. For an example, 1000 μM of odorant mixture was composed of lcarvone (100 μM), cis-3-hexenol (100 μM), anis aldehyde (100 μM), methyl β-naphthyl ketone (100 μM), d-carvone (300 μM), and linalool (300 μM). Data are shown as mean ± SE of three or six replicates from one or two independent experiments. � At least one concentration of the odorant produced a statistically significant response of an OR-expressing cell as compared to both that of cells stimulated with medium alone and a vector-transfected control cell (Sidak--Bonferroni method with alpha = 5.0%). (EPS) S4 Fig. Dose-response of OR-expressing HepG2 cells to the odorant mixture. HepG2 cells were transfected with the indicated ORs, CRE/luc2PpGL4.29, pRL-CMV, Gαolf, and RTP1S and stimulated with multiple concentrations of the odorant mixture. Error bars, S.E. from three replicates. � At least one concentration of the odorant produced a statistically significant response of an OR-expressing cell as compared to both that of cells stimulated with medium alone and a vector-transfected control cell (Sidak--Bonferroni method with alpha = 5.0%). (EPS) S5 Fig. Dose-response of OR-expressing HuH7 cells to the odorant mixture. HuH7 cells were transfected with the indicated ORs, CRE/luc2PpGL4.29, pRL-CMV, Gαolf, and RTP1S and stimulated with multiple concentrations of the odorant mixture. Error bars, S.E. over three or six replicates from one or two independent experiments. � At least one concentration of the odorant produced a statistically significant response of an OR-expressing cell as compared to both that of cells stimulated with medium alone and a vector-transfected control cell (Sidak--Bonferroni method with alpha = 5.0%). (EPS) S6 Fig. Dose-response of OR-expressing LNCaP cells to the odorant mixture. LNCaP cells were transfected with the indicated ORs, CRE/luc2PpGL4.29, PRL-CMV, Gαolf, and RTP1S and stimulated with multiple concentrations of the odorant mixture. Error bars, S.E. over three replicates. � At least one concentration of the odorant produced a statistically significant response of an OR-expressing cell as compared to both that of cells stimulated with medium alone and a vector-transfected control cell (Sidak--Bonferroni method with alpha = 5.0%). (EPS) S7 Fig. The effect of co-expression of Gαolf on the assay sensitivity of OR-expressing LNCaP cells. LNCaP cells were transfected with the indicated ORs, CRE/luc2PpGL4.29, and pRL-CMV RTP1S with or without Gαolf. They were then stimulated with multiple concentrations of the odorant mixture. Data are shown as mean ± SE of three independent experiments. (EPS) S8 Fig. No responsiveness of OR51T1-expressing LNCaP cells to odorants. LNCaP cells were transfected with each OR, CRE/luc2PpGL4.29, pRL-CMV, Gαolf, and RTP1S and stimulated with the indicated concentration of ambrettolide, cedramber, and galaxolide. (EPS) S9 Fig. Experimental conditions to reduce the basal level of signals failed to detect OR51T1 activation by l-carvone in HEK293 cells. A-C. Reduction of the amount of transfected plasmid improved response of HEK293 cells expressing OR2W1 or OR51E1; examples of previously characterized ORs with a high basal activity that did not allow the detection of OR51T1 activation. HEK293 cells were transfected with each OR, CRE/luc2PpGL4.29, pRL-CMV, and RTP1S. (EPS) 0 Fig. Response of TAAR1-expressing LNCaP cells and HEK293 cells to known ligands. LNCaP cells and HEK293 cells were transfected with TAAR1, CRE/luc2PpGL4.29, pRL-CMV, Gαolf and RTP1S and stimulated with the indicated concentration of amines. Data are shown as mean ± SE of three independent experiments. (EPS) 1 Fig. No responsiveness of LNCaP cells to agonists for endogenous OR51E1 and OR51E2. Mock-transfected LNCaP cells were stimulated with the indicated concentration of odorants. LNCaP cells were transfected with CRE/luc2PpGL4.29, pRL-CMV, Gαolf, and RTP1S. Error bars, S.E. from three replicates. (EPS) [/fig]
[table] PLOS ONE: | https://doi.org/10.1371/journal.pone. [/table]
[table] Table 1: ORs for each odorant a . [/table]
[table] Table: Comparison with genetic variants used in previous studies. (XLSX) 28. Raka RN, Wu H, Xiao J, Hossen I, Cao Y, Huang M, et al. Human ectopic olfactory receptors and their food originated ligands: a review. Crit Rev Food Sci Nutr. 2021: 1-20. https://doi.org/10.1080/ 10408398.2021.1885007 PMID: 33605814 29. Feldmesser E, Olender T, Khen M, Yanai I, Ophir R, Lancet D. Widespread ectopic expression of olfactory receptor genes. BMC Genomics. 2006; 7: 121. https://doi.org/10.1186/1471-2164-7-121 PMID: 16716209 30. Malik B, Elkaddi N, Turkistani J, Spielman AI, Ozdener MH. Mammalian taste cells express functional olfactory receptors. Chem Senses. 2019; 44(5): 289-301. https://doi.org/10.1093/chemse/bjz019 PMID: 31140574 [/table]
|
A meta-analysis comparing stereotactic body radiotherapy vs conventional radiotherapy in inoperable stage I non-small cell lung cancer
Background: Stereotactic body radiotherapy (SBRT) superseded conventional radiotherapy (CRT) for the treatment of patients with inoperable early stage non-small cell lung cancer (NSCLC) over a decade ago. However, the direct comparisons of the outcomes of SBRT and CRT remain controversial. This meta-analysis was performed to compare the survival and safety of SBRT and CRT in patients with inoperable stage I NSCLC.Methods: We systematically searched the Cochrane Library, Embase, PubMed, Web of Science, Ovid MEDLINE, ScienceDirect, Scopus and Google Scholar for relevant articles. Overall survival (OS), progression-free survival (PFS), lung cancer-specific survival (LCSS), local control rate (LCR) and adverse effects (AEs) were the primary outcomes. Results: We identified 11,110 articles, 17 of which were eventually included in this study; these 17 articles had 17,973 patients (SBRT: 7395; CRT: 10,578). Compared to CRT for the treatment of inoperable stage I NSCLC, SBRT had superior survival in terms of OS (hazard ratio [HR]: 0.66, 95% confidence interval [CI]: 0.62-0.70, P < .00001), LCSS (HR: 0.42 [0.35-0.50], P < .00001), and PFS (HR: 0.34 [0.25-0.48], P < .00001). The 4-year OS rate (OSR); 4-year LCSS rate (LCSSR); 3-year local control rate (LCR); 5-year PFS rate (PFSR) with SBRT were all higher than those with CRT. With regard to all-grade AEs, the SBRT group had a significantly lower rate of dyspnea, esophagitis and radiation pneumonitis; no significant difference was found in grade 3-5 AEs (risk ratio [RR]: 0.68 [0.30-1.53], P = .35).Conclusions: With better survival and a lower rate of dyspnea, esophagitis and radiation pneumonitis than CRT, SBRT appears to be more suitable for patients with inoperable stage I NSCLC.Abbreviations: 3DCRT = 3D conformal radiotherapy, AEs = adverse effects, AHRT = accelerated hypofractionated radiotherapy, BED = biologically effective dose, CFRT = conventional fractionated radiotherapy, CI = confidence interval, CRT = conventional radiotherapy, GRADE = the grading of recommendations assessment, development and evaluation system, HR = hazard ratio, LCR = local control rate, LCSS = lung cancer-specific survival, LCSSR = lung cancer-specific survival rate, NOS = Newcastle-Ottawa Scale, NSCLC = non-small cell lung cancer, OS = overall survival, OSR = overall survival rate, PFS = progression-free survival, PFSR = progression-free survival rate, PRISMA = preferred reporting items for systematic review and meta-analysis, RR = risk ratio, RCT = randomized controlled trial, SBRT = stereotactic body radiotherapy.Editor: Jianxun Ding.All analyses were based on previously published studies, and hence no ethical approval and patient consent were required. meta-analysis comparing stereotactic body radiotherapy vs conventional radiotherapy in inoperable stage I non-small cell lung cancer. Medicine 2020;99:34(e21715).
# Introduction
Lung cancer remains a major contributor to cancer-related mortality and cancer incidence worldwide; its annual incidence is predicted to increase continually for at least the next few decades.Recent advances in screening have made substantial progress with regard to detecting early-stage lung cancer.As far as stage I non-small cell lung cancer (NSCLC) patients are concerned, the current standard therapy is surgery.However, multiple comorbidities and poor physical performance status result in nearly 20% of stage I NSCLC patients being unable to tolerate surgery.Conventional radiotherapy (CRT) has been used as the standard noninvasive strategy for more than a decade.However, given the low 5-year survival of 10% to 22%,CRT has long puzzled clinicians.
The past decade has seen great advances in stereotactic body radiotherapy (SBRT), and SBRT has gradually superseded CRT in clinical practice for the treatment of patients with inoperable early-stage NSCLC. The CHISEL trial reported that the rate of 2year local control for inoperable Stage I NSCLC patients was 89% with SBRT compared with 65% with CRT.However, Borst et aldemonstrated that SBRT had a significant doseresponse relationship to radiation pneumonitis. In addition, the SPACE trial revealed that no apparent difference was observed in overall survival or local control between patients treated with SBRT and CRT.It is controversial whether the SBRT can replace CRT.
This meta-analysis aimed to directly compare the efficacy and safety of CRT and SBRT for inoperable stage I NSCLC.
# Materials and methods
The PRISMA (Preferred Reporting Items for Systematic Review and Meta-Analysis) guidelines guided the performance of this meta-analysis (Supplementary Digital Content,, http:// links.lww.com/MD/E693).2.1. Search strategy
The following internet sources were used to retrieve the relevant literature:
(1) PubMed;
(2) Web of Science; (3) Embase; (4) Cochrane Library; (5) Ovid MEDLINE; (6) ScienceDirect; (7) Scopus;Google Scholar.
We updated the last search on September 3, 2019. "Stereotactic body radiotherapy" and "lung neoplasms" were the key terms used in the search. The complete search is outlined in Appendix 1 (http://links.lww.com/MD/E698). For the further identification of eligible articles, we identified relevant references of the retrieved literature as well. We set no limitations on language.
## Selection criteria
We performed this search pursuant to the PICOS guidelines:
(1) participants: stage I NSCLC patients who were inoperable, high-risk operable (the factors were: a. heart disease; b. advanced age; c. chronic obstructive pulmonary disease; d. poor performance status)or refused surgical intervention; (2) intervention and comparison: SBRT versus CRT (including conventional radiotherapy,conventional fractionated radiotherapy,3D conformal radiotherapy,and accelerated hypofractionated radiotherapy; (3) outcomes: OS (overall survival), OS rate (OSR), local control rate (LCR), lung cancer-specific survival (LCSS), LCSS rate (LCSSR), progression-free survival (PFS), PFS rate (PFSR) and adverse effects (AEs); and (4) study design: RCTs and cohort studies.
Protocols, abstracts, meta-analyses, animal experiments, articles with duplicated data and studies without original data were excluded.
## Data extraction
The following data were extracted independently by 2 investigators: first author, country, publication year, number of participants, participant characteristics (age, sex, tumor size, TNM stage, histology, performance status (PS), medically inoperable rate), treatment characteristics (type, dose, time), median follow-up, study design, antitumor efficacy indices (OS, OSR, LCSS, LCSSR, LCR, PFS, and PFSR) and the types and quantity of AEs (all-grade AEs and grade 3-5 AEs). Disagreements on any terms were resolved by a third investigator.
## Quality assessment
The 5-point Jadad scale was adopted to evaluate the quality of the RCTs; it consisted of three sections, namely, randomization, masking and accountability. Studies with scores ≥3 were assessed as being of high quality.The Newcastle-Ottawa Scale (NOS, 9 points) was adopted to determine the cohort study quality; it contained three sections, namely, selection, comparability and exposure. Studies with scores between 8 and 9 were assessed as being of high quality; studies with scores between 6-7 were assessed as being of medium quality.The grading of recommendations assessment, development and evaluation (GRADE) system was use to explore the quality of the data; it consisted of five sections, namely, inconsistency, indirectness, imprecision, risk of bias and publication bias. Data were assessed as being of high, medium, low or very low quality.2.5. Statistical analysis Stata 14.0 (Stata Corp) and Review Manager 5.3 (The Nordic Cochrane Center) were utilized to perform this meta-analysis. To analyze the OS, LCSS and PFS, hazard ratios (HRs, HRs > 1 favored the CRT arm) and 95% confidence interval (CIs) were applied. HR, OSR, LCSSR, LCR, PFSR were extracted directly from the included articles and indirectly from Kaplan-Meier curves, pursuant to the methods proposed by Tierney et al.Risk ratios (RRs) and their 95% CIs were used for AEs (RRs > 1 favored the CRT arm) as well as the OSR, LCSSR, LCR, and PFSR (RR > 1 favored the SBRT arm). To clarify whether the We utilized the x 2 test and I 2 statistic to assess heterogeneity. A P < .1 (on the x 2 test) or I 2 > 50% reflected evident heterogeneity and a random-effects model was used; otherwise the fixed-effects model was used. Begg rank correlationand Egger linear regression testswere utilized to assess publication bias; P < .05 indicated statistical significance.
# Results
## Search results and study quality assessment
On the last search on September 3, 2019, 11,110 potentially qualified studies were initially identified. After strict screening, 17 studies (2 RCTs,2 prospective cohortsand 13 retrospective cohortsinvolving 17,973 patients (SBRT, 7395; CRT, 10,578) were included in this analysis. The subgroup analyses revealed that the
[formula] T1-T2N0M0 SBRT 77 77 - - - - 30.4 RC CRT 193 77 - - - - 2015 [/formula]
KoshyUSA75
[formula] - - - - 21 RC CFRT 5375 75 - 60 Gy in 1.8-2 Gy/F - - 2015 [/formula]
ValleUSA
[formula] 2007-2011 T1-T2aN0M0 SBRT 184 - - - - - - RC CRT 64 - - - - - 2016 [/formula]
Nyman , P = .1368) were higher in the SBRT group than in the CRT group. Five articles evaluated LCSS (heterogeneity: P = .58, I 2 = 0). SBRT showed significant superiority in LCSS (HR: 0.42 [0.35-0.50], P < .00001) compared with CRT.
## Adverse effects.
Six articles compared all-grade AEs , http://links.lww.com/MD/E696). The 10 most commonly reported AEs in all grades were pneumonitis, dyspnea, cough, fibrosis, acute kidney disease, radiation pneumonitis, skin reactions, fatigue, and chest wall painFive articles referred to grade 3-5 AEs,in which SBRT was associated with no obvious difference (RR: 0.68 [0.30-1.53], P = .35) when compared with CRT (heterogeneity: P = .08, I 2 = 52%,(Supplementary Digital Content, , http://links.lww.com/MD/E697). Dyspnea, cough, radiation pneumonitis, fatigue, chest wall pain, lung infection, pain, cataracts, hypoxia and weight loss were the 10 most commonly reported grade 3-5 AEs induced by SBRT and CRT, among which we observed no significant differences. Subgroup analysis for overall survival, lung cancer-specific survival and progression-free survival. .
## Os
# Discussion
Early-stage NSCLC accounts for approximately 10% to 20% of new NSCLC diagnoses, and the figure is continually rising due to the implementation of the new NSCLC screening guidelines.SBRT has improved substantially during the past decade, and it has supplanted CRT and gained popularity among early-stage inoperable NSCLC patients.However, several debates still exist concerning whether SBRT can completely replace CRT. This meta-analysis is the first to directly compare SBRT with CRT in patients with inoperable stage I NSCLC. The outcomes revealed that the patients who were administered SBRT had superior OS, LCSS, LCR and PFS when compared with the patients who received CRT. In addition, the subgroup results were also statistically significant and in favor of better survival with SBRT. The analyses also demonstrated that the SBRT group had a significantly lower risk of dyspnea, radiation pneumonitis and esophagitis among allgrade AEs, though there was no apparent difference in the rate of grade 3-5 AEs between the SBRT and CRT groups. The main benefit of SBRT treatment is the significantly better survival (OS, LCSS, LCR and PFS). Compared with CRT, SBRT had obvious advantages in OSR-4y, LCSSR-4y, LCR-3y and PFSR-5y in our study. Navarro-Martin et alfound that the 3-year OSR was 66%. In addition, Timmerman et alshowed the 5-year OSR and 5-year PFSR with SBRT were 40.0% and 25.5%, respectively, in the RTOG 0236 Trial. However, the SPACE trialindicated no apparent difference in local control or OS after directly comparing SBRT and CRT. We speculated that the superior survival with SBRT could be due to many reasons. First, SBRT delivers high doses (e.g., 3 fractions of 15-22 Gy) of radiation that precisely and directly target tumors for ablation, while CRT is based on a protracted treatment with minimal 1.8 to 2.0 dose-per-fraction sizes.Additionally, the most common SBRT dose-fractionation schedules provide a BED of at least 100 Gy to achieve antitumor efficacy, while the BED of CRT usually reaches 80 Gy or less, which is not high enough to completely kill all the tumor cells in the gross target volume, resulting in a lower rate of long-term local control.Finally, with highly accurate doses delivered during minimal courses of therapy, SBRT prevents tissues within the irradiation volume from radiation injury to some extent. CRT involves 6 to 7 weeks of daily radiation; as these doses are above the radiation tolerance, the accumulated dose injuries inevitably lead to some degree of lung fibrosis.Our subgroup analyses showed that SBRT led to better survival than CRT, conventional fractionated radiotherapy (CFRT) and 3D conformal radiotherapy (3DCRT), though not in the subgroup of patients who received accelerated hypofractionated radiotherapy (AHRT). We also found that the patients with a performance status score of 0-1, patients aged 75 years, males, patients with squamous histology and patients who received an SBRT fraction dose ≥20 Gy were predicted to have relatively better survival. Additionally, residence in Asian or European regions, age between 75 and 80, and a fraction dose of 10 to 15 Gy were also likely to influence the outcomes. In terms of AHRT subgroup results, we surmised that several schedules with an adequately high BED of fractionated and accelerated CRT could achieve tumor control rates similar to those achieved with SBRT. The differences in the SPACE trial concerning survival may be attributed to the lower BED in the SBRT group with the 3 prescribed fractions of 45 Gy. However, a worse performance status (24% PS > 2) of the patients treated in the SPACE trial cannot be ignored. As van Baardwijk et al stated, the high BED in SBRT could also be considered "overkill."Overall, SBRT appears to lead to better survival in patients with inoperable stage I NSCLC.
Our meta-analysis demonstrated that the risk of toxicity was low in both the SBRT and CRT groups. Our subgroup analyses indicated that SBRT induced significantly lower rates of dyspnea, radiation pneumonitis and esophagitis in all-grade AEs while there were no significant difference in grade 3-5. For the statistical difference, we suspected the difference might because the special courses and doses of SBRT and CRT treatment. Both BED are rather "safe", there were rarely treatment-related deaths and the majority of the AEs were mostly grade 1-2. This may illustrate the low correlation in grade 3-5 AEs. For SBRT, higher fraction dose and less fractions meant more of the ray was precisely concentrated on the tumor regions with less harm to normal structures and tissues, which effectively promoted a declined probability and quantity of AEs. Additionally, the suitable BED effectively meant that latent radiotherapy damage and lung fibrosis could be avoided during a shorter treatment time. While CRT with lower BED struggles to get enough efficacy to kill tumor cells in the gross target volume via a longer course treatment involving 6 to 7 weeks of daily radiation. Accumulated dose injuries inevitably lead to some grade 1-2 AEs. In the published literature, it has been reported that tumors located in the central lung may enhance the risks associated with SBRT due to the potential damage to the mediastinal tissues; for instance, there may be an increased risk of pleural effusion, pneumonitis and lung capacity reduction. In addition, related rare AEs are tracheobronchial injury, esophageal ulcer and myeleterosis.When the tumors are near the brachial plexus nerve, severe damage could result in neuropathic pain and brachial palsy.For tumors near the chest wall, the latent complications are rib fractures and chest wall pain; therefore, the SBRT dose should be limited to 30 to 35 Gy. SBRT seems to be acceptable for patients with tumors located in the peripheral lung but away from chest wall; however, radiation pneumonitis is the most common complication in these patients.It is accepted that, in SBRT, 4 to 10 fractions are safe and effective, yet 3 fractions with 54 to 60 Gy pose a risk that needs to be avoided.The prospective RTOG 0813 study reported that no serious toxicities occurred at99:www.md-journal.com the maximum tolerated dose (50 Gy) in 5-fraction regimens.However, particular attention should be paid to tumors abutting the bronchial tree and esophagus to avoid severe toxicity. In summary, compared with CRT, SBRT is a safer treatment strategy. Several limitations exist in the present meta-analysis. First and foremost, the articles we included were all in the English language; therefore, there may have been language bias. Additionally, while most of the 17 articles were of mediumand high-quality, the 2 RCTs included in the analysis could have weakened the study conclusions. Moreover, attention should be paid to the significant heterogeneity that existed in the analyses of all-grade and grades 3-5 AEs, which might have reduced the reliability of the results. Furthermore, we obtained the data pertaining to grades 3-5 AE mainly from 2 RCTs, which may have led to representativeness bias. Finally, the CRT types and the treatment doses were diverse, which perhaps increased the heterogeneity and decreased the quality, although subgroup analyses were conducted. Accordingly, we suggest that determining the suitable dose for tumors in each anatomical region should be considered in future studies.
# Conclusions
Our study illustrated that SBRT tends to lead to better survival (OS, LCSS, LCR and PFS) and carries lower risks of dyspnea, radiation pneumonitis and esophagitis than CRT for patients with inoperable stage I NSCLC. The subgroup analyses indicated that a 0-1 patient performance status and suitable BED predicted an improved prognosis. Given the inherent limitations of the present study, the conclusions need to be confirmed in more largescale high-quality RCTs. |
Immunosenescence and Inflamm-Aging As Two Sides of the Same Coin: Friends or Foes?
[bib_ref] Aging and health -a systems biology perspective. introduction, Jazwinski [/bib_ref] [bib_ref] Complex systems dynamics in aging: new evidence, continuing questions, Cohen [/bib_ref] [bib_ref] Does the human immune system ever really become "senescent, Pawelec [/bib_ref] [bib_ref] Immunobiography and the heterogeneity of immune responses in the elderly: a focus..., Franceschi [/bib_ref] [bib_ref] Markers of T cell senescence in humans, Xu [/bib_ref] [bib_ref] From inflamm-aging to immune-paralysis: a slippery slope during aging for immune-adaptation, Fulop [/bib_ref]
## Immune changes with aging: immunosenescence and inflamm-aging, as the two sides of the same coin
The prevailing current opinion is that the most marked changes that occur with aging in the adaptive immune system determine the state of immunosenescence [bib_ref] Hallmarks of human "immunosenescence": adaptation or dysregulation?, Pawelec [/bib_ref] [bib_ref] Lymphocyte generation and population homeostasis throughout life, Yanes [/bib_ref]. It is of note that since the 1980s it has been recognized that the innate system is influenced by aging but perhaps not always in the same direction [bib_ref] Innate immunosenescence: effect of aging on cells and receptors of the innate..., Solana [/bib_ref] [bib_ref] Paradoxical changes in innate immunity in aging: recent progress and new directions, Montgomery [/bib_ref] [bib_ref] Age-dependent alterations of Fc gamma receptor-mediated effector functions of human polymorphonuclear leucocytes, Fülöp [/bib_ref]. At the turn of the century, a new concept has been put forward by Claudio Franceschi. This concept suggested that aging was associated with a chronic, sterile, low-grade inflammation called inflamm-aging [bib_ref] Inflamm-aging. An evolutionary perspective on immunosenes cence, Franceschi [/bib_ref]. The first related question that arises is what are the characteristic aging-associated changes in the various compartments of the immune system? Furthermore, are these changes faithful indicators of senescence (and progressive incapacitation of the system) or an adaptation, as well as ultimately whether they have any clinical significance. The second question is what the relationship between immunosenescence and inflamm-aging is and how it can be integrated into the broader mechanism of the aging process? Finally, the third question is whether we should attempt to intervene to modulate it.
## Innate immune changes
The innate immune response is the most phylogenetically conserved protection in the animal kingdom that allows the organism to efficiently defend against an impressive number of aggressive pathogens [bib_ref] Immunosenescence in vertebrates and invertebrates, Müller [/bib_ref]. This compartment is meant to recognize and react to the conserved pathogen-associated molecular patterns (external threats) and danger-associated molecular patterns (internal threats) by way of specific receptors that play a key role in elimination of the aggressors [bib_ref] Innate cell communication kick-starts pathogen-specific immunity, Rivera [/bib_ref] [bib_ref] Molecular determinants in phagocyte-bacteria interactions, Kaufmann [/bib_ref]. There are three classes of pattern-recognition receptors (PRRs), each one having distinct roles although all of them elicit some form of inflammation.
The first class of PRR comprises the Toll-like receptors which, using various intracellular signaling pathways, results in NF-κB activation and production of various mediators such as cytokines and chemokines [bib_ref] Tolllike receptors: significance, ligands, signaling pathways, and functions in mammals, Vidya [/bib_ref]. The second class comprises the NOD-like receptors, which are able to stimulate the inflammasome complex and that results in the production of IL-1, IL-18, and IL-33 [bib_ref] Multifaceted functions of NOD-like receptor proteins in myeloid cells at the intersection..., Kufer [/bib_ref]. The third class includes the Rig-like receptors that act through the interferon response elements [bib_ref] What really rigs up RIG-I?, Barik [/bib_ref]. There are also other receptors which are crucial for the functionality of innate immune cells, such as various Fcγ receptors, chemokine receptors such as fMLP receptors, and complement receptors [bib_ref] Age-dependent alterations of Fc gamma receptor-mediated effector functions of human polymorphonuclear leucocytes, Fülöp [/bib_ref]. One of the most important observations of these last years, besides the discovery of PRR, is the fact that the innate immune system possesses some sort of memory which has been termed trained innate immune memory [bib_ref] Bacille Calmette-Guerin induces NOD2-dependent nonspecific protection from reinfection via epigenetic reprogramming of..., Kleinnijenhuis [/bib_ref] [bib_ref] Trained immunity: an ancient way of remembering, Netea [/bib_ref]. As described by Franceschi et al., the innate immune system may be viewed as possessing the "bow tie" architecture where many signals converge to a few sensors but result in many effectors [bib_ref] Network, degeneracy and bow tie. Integrating paradigms and architectures to grasp the..., Tieri [/bib_ref]. These coordinated events help to elicit a precise and efficient response.
Whereas many immune changes have been described with aging, we will not describe in details all the aging-associated changes for each immune cell type, as this topic has been comprehensively reviewed recently [bib_ref] Human T cell immunosenescence and inflammation in aging, Bektas [/bib_ref] [bib_ref] Reduced dynamic range of antiviral innate immune responses in aging, Molony [/bib_ref]. Collectively, the main characteristics of aging with respect to the innate system are immune stimulation at the basal level on the one hand and, immune paralysis when specific functions such as free radical production are needed, on the other hand [bib_ref] From inflamm-aging to immune-paralysis: a slippery slope during aging for immune-adaptation, Fulop [/bib_ref]. This dichotomy was initially proposed to be at the basis of the inflamm-aging concept, which stated that the relatively maintained innate response overrode the more altered adaptive immune response, resulting in higher production of pro-inflammatory mediators. Since that original observation, it became obvious that other processes may contribute to inflamm-aging, such as cell senescence, mitochondrial dysfunction, and microbiota changes (dysbiosis) [bib_ref] From cell senescence to age-related diseases: differential mechanisms of action of senescence-associated..., Byun [/bib_ref] [bib_ref] Aging of the human metaorganism: the microbial counterpart, Biagi [/bib_ref]. Furthermore, under some circumstances the effects of oxidative stress were included as part of the process (oxy-inflamm-aging), emphasizing the role of the oxidative stress in the complex mechanisms of aging [bib_ref] The role of oxidative and inflammatory stress and persistent viral infections in..., Bauer [/bib_ref]. Whatever the nature and the involvement of the components of inflammaging are, it is a fact that the phenomenon results in subclinical low-grade inflammation. For instance, life-long antigenic stimulation by pathogens would maintain this quiescent state of innate immune system activation. The innate immune system can also be stimulated by the so-called internal GARBage system [bib_ref] Inflammaging and 'Garb-aging', Franceschi [/bib_ref]. Thus, a heightened inflamm-aging state is produced as a consequence of (1) dysfunctional mitochondria, (2) defective autophagy/mitophagy (disposal of dysfunctional organelles), (3) endoplasmic reticulum stress, (4) activation of inflammasome by cell debris and misplaced self molecules, (5) defective ubiquitin/proteasome system (misfolded/oxidized proteins), (6) activation of DNA damage response, (7) senescent T cells and their senescence-associated secretory phenotype (SASP), and (8) age-related changes in the composition of gut microbiota (dysbiosis) [bib_ref] From cell senescence to age-related diseases: differential mechanisms of action of senescence-associated..., Byun [/bib_ref] [bib_ref] Aging of the human metaorganism: the microbial counterpart, Biagi [/bib_ref] [bib_ref] The role of oxidative and inflammatory stress and persistent viral infections in..., Bauer [/bib_ref].
The demonstration of trained immune memory may explain, at least partially, some of the immune aspects of aging [bib_ref] Bacille Calmette-Guerin induces NOD2-dependent nonspecific protection from reinfection via epigenetic reprogramming of..., Kleinnijenhuis [/bib_ref] [bib_ref] Trained immunity: an ancient way of remembering, Netea [/bib_ref]. Following their response, innate immune cells return to a Frontiers in Immunology | www.frontiersin.org January 2018 | Volume 8 | Article 1960 quiescent state due to epigenetic changes and modulation of cell metabolism alternating between (aerobic) oxidative phosphorylation (OX-PHOS) and anaerobic glycolysis (Warburg effect) (31). A subsequent stimulation (e.g., by the same or different type of pathogen) elicits a faster and higher response than the first one due to trained innate memory [bib_ref] In vitro experimental model of trained innate immunity in human primary monocytes, Bekkering [/bib_ref]. The hypothesis of trained innate memory may, at least in part, explain why aging innate immune cells are in a state of sustained activation [bib_ref] Age-dependent alterations of Fc gamma receptor-mediated effector functions of human polymorphonuclear leucocytes, Fülöp [/bib_ref]. This concept is relatively novel and was first observed after a specific stimulation, such as the Calmette-Guérin bacillus (BCG). Even after 3 months following challenge, innate cells (monocyte/macrophages) were still able to sustain a certain "memory" of the initial infection and to react in the absence of BCG to any other stimulation [bib_ref] Bacille Calmette-Guerin induces NOD2-dependent nonspecific protection from reinfection via epigenetic reprogramming of..., Kleinnijenhuis [/bib_ref]. This observation has led to the concept that the innate immune system has a certain "memory, " which was not foreseen from the previous paradigm of the immune system function.
Within the context of the aging innate immune system, it can also be suggested that the sustained trained immune memory is (or may be) leading to a sustained state of activation even in the absence of a specific challenge [bib_ref] Immunobiography and the heterogeneity of immune responses in the elderly: a focus..., Franceschi [/bib_ref]. This memory is likely due to a shift in the epigenetic landscape (epigenome) of the innate cells and fueled by an energetic shift of these cells. As yet, there is no formal proof for the contribution of these phenomena to the basic activation of innate immune cells, but this seems strongly probable. If this was the case, the epigenome and pathways involved in energy production, use, and conservation by immune cells could be targets of choice for immune modulation in the elderly. This possibility may explain the suggestions that macrophages are at center stage of inflamm-aging [bib_ref] Inflamm-aging. An evolutionary perspective on immunosenes cence, Franceschi [/bib_ref]. Macrophages are able to modify their phenotype, produce pro-and anti-inflammatory mediators, and orchestrate many vital functions. Moreover, this phenomenon may to some extent resemble hormesis, providing a possibility to better react after each repeated stimulation [bib_ref] What is hormesis and its relevance to healthy aging and longevity?, Calabrese [/bib_ref] [bib_ref] Mediterranean diet and inflammaging within the hormesis paradigm, Martucci [/bib_ref]. Finally, it has recently been reported that not only does a high low-grade controlled inflammation was present in aged individuals (centenarians) but also that inflammation showed a better correlation with longevity than any other parameters, according to two longitudinal studies.
The epigenetic clock notion in aging whole-organism has been proposed recently [bib_ref] Epigenetic clock analyses of cellular senescence and ageing, Lowe [/bib_ref]. ELOVL2 (elongase of omega 3 and 6 fatty acids) was found to be the most powerful single epigenetic biomarker of aging [bib_ref] Methylation of ELOVL2 gene as a new epigenetic marker of age, Garagnani [/bib_ref]. Furthermore, Franceschi's group has shown that centenarians and their offsprings are epigenetically younger than one could deduct from their chronological age. According to this study [bib_ref] Decreased epigenetic age of PBMCs from Italian semi-supercentenarians and their offspring, Horvath [/bib_ref] , semi-supercentenarians are on average 8.7 years younger than expected based on chronological age, and offsprings of aged greater than 105 years are 5.2 years younger than age-matched controls where DNA methylation age and chronological age overlap. These findings reinforce the idea that epigenome control of the innate trained memory and its possible dysregulation with aging lead to DAMAge by inflammaging. Can this be reconciled with the heightened inflamm-aging in centenarians? Perhaps trained immune memory is the key regulated by epigenetic changes. The methylation age observed in centenarians suggests that the heightened inflamm-aging is either not connected to it or, paradoxically, methylation age should be younger. These observations may represent a trade-off between a potentially harmful process which, when under tight control, may remain beneficial. It is tempting to suggest that inflammaging may be considered the essence of life and the real "Fountain of Youth. " In this context, centenarians may be considered as the standard and not the exception and may serve as model for the better understanding the role of inflammation and epigenetics in aging.
Chronic challenges during aging are paralleled by intracellular changes such as mitochondrial dysfunction, altered autophagy and changes in DNA repair mechanisms. However; immune cells are constantly maintained in an alert state due to chronic low-grade inflammation. However, this state may be counterbalanced by anti-inflammatory molecules as shown in the case of centenarians [bib_ref] Inflammaging and anti-inflammaging: a systemic perspective on aging and longevity emerged from..., Franceschi [/bib_ref] [bib_ref] Inflamm-aging does not simply reflect increases in pro-inflammatory markers, Morrisette-Thomas [/bib_ref] [bib_ref] Inflammaging and human longevity in the omics era, Monti [/bib_ref]. Chronic low-grade inflammation (inflamm-aging) is a physiological response to the life-long antigenic stress and represents an efficient defense mechanism as long as it is under control. Without the essential counter-regulation by anti-inflammatory molecules as seen in aging, it is now clear how damaging this physiological state may be to the whole organism [bib_ref] The role of immunosenescence in the development of age-related diseases, Fülöp [/bib_ref]. Centenarians seem to represent an exception to the inflammaging effects on physiological aging or, perhaps, they present the physiological dynamics of aging. Thus, we may suggest that the norm (successful) aging is exemplified by centenarians whereas individuals that do not reach this age are the biological exceptions that lack the individual epigenetic history and machinery to reach that pinnacle.
The corollary of chronic low-grade inflammation is the downregulation of the innate immune functions or immune paralysis or eventually a sort of innate immune tolerance [bib_ref] From inflamm-aging to immune-paralysis: a slippery slope during aging for immune-adaptation, Fulop [/bib_ref]. This physiological condition protects the organism against further self-induced damage even if it is at the expense of the defense from pathogens or from GARBAge. However, it is a reductionist view to assume that this immune paralysis is equal to a non-functional state. Although there are functional alterations in elderly individuals when compared to young subjects, a straightforward assumption that immune cells of elderly humans lose their protective functions. For instance, most elderly humans are able to defend against many types of infections even if the adaptive immune response is somewhat less functional. However, there are relatively few longitudinal studies concerning innate immune function changes with age. Therefore, we do not know whether this is a continuous phenomenon or whether they remain stable during aging. We can speculate that incongruent results of cross-sectional studies suggest that there is not a uniform decrease either related to cell type or to immune cell functions.
Could there be any advantage associated with innate immune paralysis? It can easily be conceptualized that maintenance of identical intensities in the innate cell functions in elderly subjects to the level of young individuals would be energetically very difficult. For instance, this is illustrated by the situation of the large amounts of energy needed for maintaining a M1 (pro-inflammatory and anti-cancer) phenotype in the case of macrophages [bib_ref] Macrophage metabolism as therapeutic target for cancer, atherosclerosis, and obesity, Geeraerts [/bib_ref]. On the other hand, the M2 phenotype (healing, promoting angiogenesis and cancer growth) consumes much less energy and, therefore, is not as much affected. A decreased production of free radicals can be considered harmful for the eradication of pathogens. However, low free radical production may protect the whole body against further age-and oxidative stress-related damages. Decreased chemotactic activity may also be suggested to be harmful for pathogen destruction, although a sterile inflammatory process may protect against excessive tissue damage. Thus, the question is within the context of evolution, what is most rewarding for the body in an aged organism? On the one hand, is it to destroy pathogens at any cost? On the other hand, is it to maintain physiological integrity by way of chronic inflammation? This dichotomy can mirror recent findings on mitochondria where a certain degree of dysfunction was linked to successful aging and longevity, in contrast to normal or excessively altered functioning in unsuccessful aging [bib_ref] Mitochondria and mitochondria-induced signalling molecules as longevity determinants, Rose [/bib_ref]. This mild impairment may work as a hormetic signal [bib_ref] Aging and Parkinson's disease: inflammaging, neuroinflammation and biological remodeling as key factors..., Calabrese [/bib_ref]. Although there is at present no definitive answer to this question, it illustrates the fact that a broader perspective is needed to understand changes in the innate immune system with aging. The innate immune system influences the adaptive immune response in many ways. One of these cases is antigen presentation by dendritic cells (DCs). There are conflicting results in this domain and it seems that DCs are less able to prime CD4 + T cells in the elderly [bib_ref] Role of dendritic cells in inflammation and loss of tolerance in the..., Agrawal [/bib_ref]. It is not clear whether the problem is related to antigen presentation or to reaction to antigen presentation. In likelihood; both aspects may be affected by aging. The processing of antigenic peptides by the immune proteasome may not be so efficient and perhaps either the T cell receptor (TCR) itself or TCR-dependent-signaling could be altered [bib_ref] The immunoproteasome in oxidative stress, aging, and disease, Johnston-Carey [/bib_ref] [bib_ref] Cellular signaling in the aging immune system, Fulop [/bib_ref]. The interaction may also occur through an interplay with the cytokines that are secreted by the innate immune system cells. Increased levels of pro-inflammatory cytokine production by the innate cells during aging may also influence the reactivity of the CD4 + T cells; e.g., the increased amounts of TNFα may downregulate the expression of CD28 which will negatively affect clonal expansion [bib_ref] Decreased proliferative capability of CD4 + cells of elderly people is associated..., Bryl [/bib_ref]. Moreover, these cytokines may elicit increased free radical production in T cells, which will paralyze their function by increasing inhibitory events of signaling [bib_ref] Oxidative inactivation of CD45 protein tyrosine phosphatase may contribute to T lymphocyte..., Rider [/bib_ref]. In sum, alterations in the innate immune system may also impact adaptive immune changes with aging.
## Adaptive immune system
The adaptive immune system is composed of the cellular and the humoral immune response. T cells are orchestrating the cellular immune responses. These cells are basically divided into CD4 + and CD8 + T cell populations, which possess very clearly defined functions. CD4 + T cells are helper cells that regulate the functions of all the other immune cells. They also possess effector functions [bib_ref] Effector/memory CD4 T cells making either Th1 or Th2 cytokines commonly co-express..., Das [/bib_ref]. CD8 + T cells are effector and memory T cells responsible for clearing the aggressors [bib_ref] Heterogeneity in the differentiation and function of CD8? T cells, Mittrücker [/bib_ref]. The CD4 + compartment may be subdivided, taking into account functionalities, in Th1, Th2, Th17, and regulatory T cell (Treg) subpopulations [bib_ref] Different subsets of T cells, memory, effector functions, and CAR-T Immunotherapy, Golubovskaya [/bib_ref]. Phenotypically, CD4 + and CD8 + T cell compartments are subdivided into four functionally distinct subpopulations, which are naïve, central memory, effector memory, and T effector memory cells re-expressing CD45RA (TEMRA) [bib_ref] From "truly naïve" to "exhausted senescent" T cells: when markers predict functionality, Larbi [/bib_ref].
Many alterations in the adaptive immune system have been described in aging [bib_ref] Lymphocyte generation and population homeostasis throughout life, Yanes [/bib_ref] [bib_ref] The life cycle of a T cell after vaccination -where does immune..., Kim [/bib_ref] [bib_ref] Mechanisms underlying T cell immunosenescence: aging and Cytomegalovirus infection, Tu [/bib_ref]. With respect to T cell subpopulations, aging is characterized by two main changes: a decrease in naïve T cells that leads to the shrinking of the TCR repertoire and an increase in memory T cells that is primed by different aggressors. Recent thymic emigrants of new naïve cells are vanishingly rare in the elderly because of thymic involution at puberty and acute and chronic antigenic stress over the lifetime and, age-associated hematopoietic stem cell insufficiency [bib_ref] Hallmarks of human "immunosenescence": adaptation or dysregulation?, Pawelec [/bib_ref] , This phenomenon is considered as one of the most basic changes in the adaptive immune system with aging. How does this situation happen? This seems to be the main explanation for the increased incidence of infections, cancers and the failure of vaccination in elderly [bib_ref] Naive T cells: the crux of cellular immune aging, Appay [/bib_ref] [bib_ref] Perturbed CD8+ T cell immunity across universal influenza epitopes in the elderly, Nguyen [/bib_ref] [bib_ref] Immunosenescence and cancer, Pawelec [/bib_ref]. These observations mean that elderly individuals are less able to respond to neoantigens than young individuals. However, this idea has been seriously challenged in recent years, mainly on the basis that there may not be a dramatic shrinkage of the TCR repertoire involving the remaining and slowly produced new emigrants, as supposed for decades [bib_ref] Immunosenescence and cancer, Pawelec [/bib_ref]. Moreover, the newly reconsidered homeostatic proliferation of naïve T cells under IL-7 stimulation may replace the failing thymus, at least partially. The recently discovered Stem Cell-like Memory T cells may also participate in incomplete replenishment of the naïve T cell compartment [bib_ref] T memory stem cells in health and disease, Gattinoni [/bib_ref]. Overall, the alterations may not be so dramatic and even the T cell repertoire may be relatively sufficient to supply the demand. Indeed, centenarians do not present more cancer, as its prevalence plateaus after the age of 90. Furthermore, there is no tendency for these individuals to suffer from unknown new pathogen-induced infections. Findings of two longitudinal studies led to the conclusion that having more CD8 + naïve T cells was not considered a survival advantage [bib_ref] Immunosenenescence: role of cytomegalovirus, Pawelec [/bib_ref] [bib_ref] Lower proportion of naïve peripheral CD8+ T cells and an unopposed pro-inflammatory..., Derhovanessian [/bib_ref]. These new data shed serious doubts on the present concept of "immunosenescence, " at least in the adaptive compartment. However, the debate between immunologists and gerontologists is far from being settled.
There could be some evolutionary reasons for thymic involution. First, the maintenance of an organ so metabolically active may be very resource-demanding in the situation where the whole organism tends to reduce energy consumption during aging. This phenomenon can parallel the other two very energy-demanding organ shrinkage situations seen in aging, namely those of muscles and bone marrow. Second, during life, the organism has already encountered most of the pathogens typically active in the temporal and spatial region of its dwelling. Thus, resources must be allocated preferentially to combat these "usual, " cognate pathogens by the memory part of the immune system rather than spend energy for a useless fight, which may be terminated in any event by destruction of the invading organism.
Thymic involution is a double-edged sword. On the one hand, it may indirectly be responsible for the death of the organism which would then lack the right TCR to mount an effective response against neo-antigen(s). On the other hand, it results in lower energy consumption which becomes available for other survival-supportive functions and activities of the organism. Given the relative rarity of direct infectious causes of death in the elderly, it would appear that downregulation of capacity to respond to novel pathogens during aging does not come at an excessive cost. The increase in the number of memory T cells may be very rewarding for the aging organism as this will continuously assure survival against attacks by cognate pathogens that may threaten the survival of the organism. T cells are all directed against specific internal and external aggressors. The body hosts many latent infections which can re-activate from time to time under specific conditions [bib_ref] From "truly naïve" to "exhausted senescent" T cells: when markers predict functionality, Larbi [/bib_ref]. One well characterized pathogen of this type is the cytomegalovirus (CMV) [bib_ref] Immunosenenescence: role of cytomegalovirus, Pawelec [/bib_ref]. CMV was once considered the main cause of age-related immune changes in the elderly. Although accumulating data are still quite contradictory, the current belief is that the presence of CMV infection does not seem to be only detrimental [bib_ref] Lower proportion of naïve peripheral CD8+ T cells and an unopposed pro-inflammatory..., Derhovanessian [/bib_ref] [bib_ref] CMV and immunosenescence: from basics to clinics, Solana [/bib_ref] [bib_ref] The impact of CMV infection on survival in older humans, Pawelec [/bib_ref] [bib_ref] Cytomegalovirus driven immunosenescence-an immune phenotype with or without clinical impact?, Söderberg-Nauclér [/bib_ref]. On the contrary, CMV infection may be considered a recurrent stimulation that maintains a sustained immunological alertness that favors a better immune response, e.g., to vaccination [bib_ref] Predictors of the antibody response to influenza vaccination in older adults with..., Mcelhaney [/bib_ref]. The global response to the many various CMV antigens has even been linked to better survival [bib_ref] CMV-specific T-cell responses at older ages: broad responses with a large central..., Bajwa [/bib_ref]. Thus, the increased number of committed memory T cells may not be considered unequivocally as detrimental or related only to aging.
One of the most important features of aging is the notion of senescent cells [bib_ref] Cell senescence: role in aging and age-related diseases, Campisi [/bib_ref]. This idea has re-gained popularity in recent years as a way to explain the decreased functionality of the immune system with aging [bib_ref] Replicative senescence: the final stage of memory T cell differen tiation?, Effros [/bib_ref]. Senescent cells conform to the model of Hayflick replicative senescence as they are not proliferating but remain metabolically active and secrete several pro-inflammatory substances (SASP) [bib_ref] Senescence-associated secretory phenotypes reveal cell-nonautonomous functions of oncogenic RAS and the p53..., Coppé [/bib_ref]. Formerly, accumulated memory T cells were considered "senescent" [bib_ref] Replicative senescence: the final stage of memory T cell differen tiation?, Effros [/bib_ref]. However, experimental evidence suggests that these cells are still able to function when pathogens such as CMV are re-activated [bib_ref] CMV and immunosenescence: from basics to clinics, Solana [/bib_ref] [bib_ref] The impact of CMV infection on survival in older humans, Pawelec [/bib_ref]. Furthermore, there are no universally accepted markers of cell senescence [bib_ref] Markers of T cell senescence in humans, Xu [/bib_ref]. Finally, cell senescence is also a double-edged sword as these cells are needed in the case of some physiological functions, for instance repair and fight against cancerous transformation, whereas they are detrimental-to other cell functions [bib_ref] Cancer secretome and inflammation: the bright and the dark sides of NF-κB, Capece [/bib_ref].
There is also a large confusion in the field of aging with respect to the number and distinction between senescent and exhausted cells [bib_ref] Molecular and cellular insights into T cell exhaustion, Wherry [/bib_ref] [bib_ref] Senescence of T lymphocytes: implications for enhancing human immunity, Akbar [/bib_ref] [bib_ref] Blockade of PD-1 or p38 MAP kinase signaling enhances senescent human CD8..., Henson [/bib_ref]. Senescent and exhausted immune cells are to be distinguished as the former may be functionally inert, whereas the latter may be functionally "dormant. " This distinction is crucial when considering immune functions in relationship to aging. Exhausted T cells can be awakened by modulation of some surface receptors called the immune checkpoint inhibitors and, they can then resume function [bib_ref] Reversing T-cell dysfunction and exhaustion in cancer, Zarour [/bib_ref]. The most important of these receptors are PD-1, CTLA-4, LAG-3, TIM-3, and TIMIN. This distinction has gained considerable importance since it has been reported that a number of cancers in some elderly subjects could be successfully modulated and T cells engineered to be immunotoxic toward such cancers, namely melanoma and NSCLC [bib_ref] Immunotherapy comes of age: immune aging & checkpoint inhibitors, Elias [/bib_ref] [bib_ref] Immune checkpoint inhibitors and elderly people: a review, Daste [/bib_ref] [bib_ref] Considerations for successful cancer immunotherapy in aged hosts, Hurez [/bib_ref].
One additional aspect where the literature has not, in our view, paid enough attention is the age-associated impairment of metabolic regulation of immune cell functions which is of vital importance for an adequate immune response. Quiescent cells compared to activated cells require different metabolic responses [bib_ref] Metabolic regulation of T lymphocytes, Maciver [/bib_ref] [bib_ref] The metabolic life and times of a T-cell, Michalek [/bib_ref] [bib_ref] Impact of metabolism on T-cell differentiation and function and cross talk with..., Kouidhi [/bib_ref] [bib_ref] Cellular metabolism modulation in T lymphocyte immunity, Liu [/bib_ref] [bib_ref] Aerobic glycolysis promotes T helper 1 cell differentiation through an epigenetic mechanism, Peng [/bib_ref]. Whereas quiescent cells use the OX-PHOS pathway for their functions that generates 36 ATP per metabolized glucose, activated cells use anaerobic glycolysis, which generate two ATP. Why is it so? When cells are activated, they need energy very quickly which cannot be provided by the OX-PHOS pathway, but only by aerobic glycolysis (Warburg effect). This is another example where the body is trading efficiency for rapidity, as in many circumstances in the aging immune system. Not only do the quiescent and activated states have different metabolic requirements but also the differentiation of the various subtypes of T cells is dependent on the specific metabolic pathways used [bib_ref] Metabolic instruction of immunity, Buck [/bib_ref]. The master of cellular metabolism is the mTOR pathway that regulates clonal expansion, whereas its inhibition drives (via autophagy) the reconstruction of damaged cells [bib_ref] p38 signaling inhibits mTORC1-independent autophagy in senescent human CD8 + T cells, Henson [/bib_ref]. Very few studies in aging have addressed the metabolic changes that occur in immune cells with aging [bib_ref] The convergence of senescence and nutrient sensing during lymphocyte ageing, Akbar [/bib_ref]. However, one of these studies has emphasized the metabolic differences in T cells of young and elderly subjects. T cells of elderly individuals suffer from insufficient substrate to feed mitochondrial respiration and, consequently, are energy deprived. Instead of breaking down glucose, they shunt it into the pentose phosphate pathway, promoting an anabolic state. One of the metabolic consequences is the accumulation of reductive elements, particularly NADPH and reduced glutathione, and the scavenging of reactive oxygen species. Energy-deprived T cells upregulate activation of the energy sensor 5′-AMP-activated protein kinase (AMPK). A downstream target of inappropriately activated AMPK in aging T cells is the dual-specificity protein phosphatase 4 (DUSP4) [bib_ref] Signal inhibition by the dual-specific phosphatase 4 impairs T cell-dependent B-cell responses..., Yu [/bib_ref] , which negatively regulates members of the MAPK superfamily, in particular ERK1, ERK2, and JNK. ERK is also subject to increased negative regulation by another dual-specificity protein phosphatase, DUSP6 [bib_ref] Decline in miR-181a expression with age impairs T cell receptor sensitivity by..., Li [/bib_ref]. ERK is a key regulator of the T-cell receptor signaling cascade, and its dephosphorylation by DUSP4 and DUSP6 functions as a suppressive mechanism, weakening the TCR-induced signal and dampening T-cell function. Thus, it may be very important to take into account age-related changes to determine how the extrinsic nutritional availability of glucose, amino-acids, and lipids will modulate age-related changes in immune system functioning.
Once immune cells are stimulated as a result of recognition of cognate antigen presentation, they initiate signaling pathways that result in the transcription of the appropriate molecules required for the expected functions [bib_ref] Insights into the initiation of TCR signaling, Chakraborty [/bib_ref] [bib_ref] CD28 costimulatory signals in T lymphocyte activation: emerging functions beyond a qualitative..., Porciello [/bib_ref] [bib_ref] New insights into the T cell synapse from single molecule techniques, Dustin [/bib_ref]. In elderly subjects, this cascade of events has been found to be altered, from impaired immune synapse formation to defects associated with translocation of transcription factors [bib_ref] Cellular signaling in the aging immune system, Fulop [/bib_ref]. Interestingly, detailed investigations of these alterations have revealed that they were mostly related to factors which could be considered modulable. One of the key factors that are involved in these changes are the composition and the organization of components of the plasma membrane, which orchestrates assembly of signaling molecules in cholesterol/ ganglioside-containing nanoclusters [bib_ref] Differential role of lipid rafts in the functions of CD4+ and CD8+..., Larbi [/bib_ref]. Thus, changes that were once considered a part of the aging process could be viewed as only the manifestation of some environmental interference and modulated by lifestyle factors such as exercise and nutrition [bib_ref] Does regular exercise counter T cell immunosenescence reducing the risk of developing..., Turner [/bib_ref]. wHAT iS THe ReLATiONSHiP BeTweeN iMMUNOSeNeSCeNCe AND iNFLAMM-AGiNG?
According to the original concept of inflamm-aging, a consequence of immunosenescence, the relatively conserved innate immune system overtakes the more altered adaptive immune system in aging. However, recent data are more in line with the interpretation that this is not a unidirectional relationship, but a mutually maintained state where immunosenescence is induced by inflamm-aging and vice versa. The main changes in the aging adaptive immune system occur in the T cell compartment [bib_ref] Mechanisms underlying T cell immunosenescence: aging and Cytomegalovirus infection, Tu [/bib_ref]. There is an increase in the number of memory CD8 + T cells, which were originally considered relatively nonfunctional [bib_ref] Cytomegalovirus persistence and T-cell immunosenescence in people aged fifty and older: a..., Weltevrede [/bib_ref]. These cells are characterized by the loss of naïve T cell surface markers, such as CD28, CD27, and the emergence of new senescent markers such as KLRG1. It has been found that the increase in the number of memory T cells and, later on, that of B cells may be due to a continuous chronic antigenic stimulation similar to the phenomenon of inflamm-aging. Infection by CMV emerged above all the large variety of potential stimulating agents [bib_ref] Immunosenenescence: role of cytomegalovirus, Pawelec [/bib_ref]. However, some data indicate that CMV infection could not be differentiated from inflamm-aging between seropositive and seronegative individuals [bib_ref] The age-related increase in low-grade systemic inflammation (inflammaging) is not driven by..., Bartlett [/bib_ref]. It is conceivable that the body devotes a huge part of its immune resources to contain this specific infection throughout life. Consequently, the immune space becomes filled with CMV-specific memory CD8 + T cells. These cells have been previously considered to be inactive but recent data have shown that they are metabolically active and their senescent phenotype (SASP) can participate in the development of inflamm-aging [bib_ref] Senescence of T lymphocytes: implications for enhancing human immunity, Akbar [/bib_ref]. Thus, chronic antigenic stimulation leads both to the phenomenon of inflamm-aging and the increase of the number of senescent T cells. One additional consequence of chronic stimulation is the phenomenon of exhaustion, characterized by the emergence of inhibitory receptors, such as PD-1, CTLA-4, and many others [bib_ref] Molecular and cellular insights into T cell exhaustion, Wherry [/bib_ref]. Other cell types of the adaptive immune system are also affected by aging but to various extents. For instance, the CD4 + T cell population also undergoes similar changes to CD8 + T cells but to a different extent [bib_ref] From "truly naïve" to "exhausted senescent" T cells: when markers predict functionality, Larbi [/bib_ref]. The Treg population also increases with aging as well as the pro-inflammatory Th17 subpopulation (101). Finally, the B cell compartment is also altered with aging (102). The functional consequences of these overall changes result collectively in the decreased ability to fight new challenges. Thus, clonal expansion, cytokine production, and specific antibody production are compromised. This situation leads to increased infections, cancer, and chronic diseases in the elderly [bib_ref] The role of immunosenescence in the development of age-related diseases, Fülöp [/bib_ref]. It appears that inflamm-aging and immunosenescence progress in parallel and form a vicious cycle. Increased production of inflammatory mediators characteristic of inflamm-aging contributes to the decrease of the adaptive immune response and, eventually, to immunosenescence. In contrast, the decrease of the adaptive immune response reinforces the stimulation of the innate immune response (as the means to protect organism from infections in the circumstances when adaptive immunity fails) leading to inflamm-aging. Both processes are important not only as causes of immune changes in the elderly but also (or even mainly) because of their consequences in the aging organism.
iMMUNOSeNeSCeNCe/iNFLAMM-AGiNG; wHY DOeS iT MATTeR?
One can ask why all these changes in the immune system with aging do matter. The paradigm for many years has been that immunosenescence and inflamm-aging are the fertile soil for the development of diseases mostly considered as age-related, either acute such as infections, or chronic such as cancer, frailty, Alzheimer's disease (AD), and cardiovascular diseases (CVD) [bib_ref] The role of immunosenescence in the development of age-related diseases, Fülöp [/bib_ref]. The bulk of these observations has led to the field of geroscience [bib_ref] GeroScience: understanding the interaction of processes of aging and chronic diseases, Sonntag [/bib_ref] [bib_ref] Suggestions from geroscience for the genetics of age-related diseases, Franceschi [/bib_ref] [bib_ref] The emergence of geroscience as an interdisciplinary approach to the enhancement of..., Sierra [/bib_ref]. The consequence of this new field leads to a novel approach that consists in targeting the aging process as the single most important risk factor instead of treating each disease separately. This notion should be nuanced by individual aging, suggesting that all individuals do not age in the same way and perhaps the underlying mechanisms may be different [bib_ref] Immunobiography and the heterogeneity of immune responses in the elderly: a focus..., Franceschi [/bib_ref]. Alterations in T cell functions, more precisely the decrease in the number of naïve T cells and the increase in number of memory T cells, has been considered the main explanation for increased incidence of infections and cancers in the elderly [bib_ref] The role of immunosenescence in the development of age-related diseases, Fülöp [/bib_ref]. However, there is still no direct evidence from experimental observations or longitudinal studies, which could really support this hypothesis. It is of note to mention that the overall incidence of malignancies decreases after the age of 90 [bib_ref] Aging and cancer mortality: dynamics of change and sex differences, Yang [/bib_ref]. However, would the incidence of infections, which is claimed to increase; still be true if one would only take into account elderly individuals with healthy aging? It is also of note that relatively few elderly subjects die of infections, even if severe and requiring hospital treatment. For example, in Canada in 2013, only 4.1% of deaths in individuals aged more than 65 years could be directly attributable to infectious causes (International Classification of Diseases A00-A99, B00-B99, G00-G03, J09-J21) (Statistics Canada: http://www5.statcan.gc.ca/cansim/pick-choisir?lang=eng&sear chTypeByValue=1&id=1020561, accessed .
It has also been commonly believed for decades that elderly individuals responded poorly to vaccination, which sometimes led to generalized doubts about the efficacy of vaccination in old age in general. Among the many vaccines which were considered less efficient (107), influenza vaccination was generally cited as the gold standard [bib_ref] Impact of aging and Cytomegalovirus on immunological response to influenza vaccination and..., Merani [/bib_ref] [bib_ref] Immunosenescence: influenza vaccination and the elderly, Haq [/bib_ref]. It is now established that there are many factors besides immunosenescence, which influence effectiveness of this vaccine and others. In fact, knowledge and better understanding of these factors has already led to enormous enhancement of efficacy of vaccination (e.g., against herpesviruses) [bib_ref] Efficacy of an adjuvanted herpes zoster subunit vaccine in older adults, Lal [/bib_ref].
Inflamm-aging could play a role in the late manifestation of diseases such as AD, frailty syndrome (FS), and CVD. It is of note that, besides FS, the onset of these diseases start in young or middle age when the immune system is still efficient, that is before signs of immunosenescence or inflamm-aging are detectable. Frailty can be considered a manifestation of (unsuccessful) aging and may represent the clinical sign of biological age, in contrast to chronological age [bib_ref] Aging, frailty and age-related diseases, Fulop [/bib_ref]. Furthermore, in a longitudinal study, inflammation was found to be the most important factor to account for longevity, especially in semi-super centenarians. In conclusion, there is no doubt that immunosenescence and inflamm-aging contribute to the increased incidence of agerelated diseases. However, their exact role is not yet well defined, and in some case, it may be even doubtful. There has been little exploration of the possibility that there are optimum levels of immunosenescence and inflamm-aging and that too much or too little could exacerbate the risks of various diseases in the elderly.
## How can successes in vaccination and immunotherapy be interpreted in light of immunosenescence?
There has been recent reports of therapeutic successes in domains which were strongly considered to have the potential to overcome immunosenescence, namely the decreased immune response of the elderly to vaccination and the failures in treatment of some cancers. For example, in one study a new vaccine was tested for herpes zoster (shingles) prevention. This study showed that the administration of the antigen in combination with an adjuvant induced a strong immune response even in subjects older than 80 years of age [bib_ref] Efficacy of an adjuvanted herpes zoster subunit vaccine in older adults, Lal [/bib_ref]. Furthermore, protection by this vaccine was high even after the 3.2 years of follow-up. This report should serve as a lead to question whether reported lack of vaccine efficacy in the elderly is due to immunosenescence or to improperly designed vaccines. Alternatively, even if a decrease in the immune response occurs it may be overcome by a well targeted vaccine [bib_ref] Efficacy of an adjuvanted herpes zoster subunit vaccine in older adults, Lal [/bib_ref]. Recently, there has been a rise of immune checkpoint inhibitors as effective therapeutics in cancer treatment. While still sparse, the observations of effectiveness in elderly subjects are very encouraging. For example, in the case of metastatic melanoma, the use of Nivolumab ® and Ipilimumab ® either alone or in combination had survival effects in elderly subjects similar to young patients [bib_ref] Immunotherapy comes of age: immune aging & checkpoint inhibitors, Elias [/bib_ref] [bib_ref] Immune checkpoint inhibitors and elderly people: a review, Daste [/bib_ref] [bib_ref] Considerations for successful cancer immunotherapy in aged hosts, Hurez [/bib_ref]. This is a remarkable result that suggests that the exhausted T cells of the elderly are still able to respond to inhibition of their inhibitory receptors with a recovery of cytotoxic activity. It is to be mentioned that immunotherapy often works well in the elderly, but the treatment has to be adapted to the patients and be different than for young people [bib_ref] Considerations for successful cancer immunotherapy in aged hosts, Hurez [/bib_ref]. What works for one does not necessarily work for the other, but this does not mean that the elderly will not respond. The longevity advantage in some cases of CMV infection may also militate against the exclusively detrimental effect of the immunosenescence and inflamm-aging [bib_ref] CMV-specific T-cell responses at older ages: broad responses with a large central..., Bajwa [/bib_ref].
## How should we interpret immunosenescence and inflamm-aging?
We suggest that there is a need for a complete reconsideration of immune changes with aging to gain access to a better understanding of their mechanisms and to ensure that eventual interventions do more good than harm. The question does follow: which immune changes in the elderly may be beneficial? The answer to this question would require careful reinterpretation of current data, but it could also be highly useful to cope with an extended perspective of aging. There are several potential changes in the adaptive immune system with aging. First, increased proportions of adaptive memory cells may be beneficial to fight cognate pathogens more efficiently. Thymic involution may be considered as needed for reduction of energy consumption by an organ which is not absolutely necessary for survival and the maintenance of which is energy-costly. Finally, the increased proportions of Tregs observed in the elderly may prevent autoimmune onslaught.
There are some alterations that may not be solely detrimental in the innate immune system. For instance, increased proportions of innate ("trained") memory cells may help to efficiently fight cognate-and some not so cognate-pathogens. Increased asymptomatic pro-inflammatory state (conceptualized as increased "readiness" of innate immunity to pathogen challenge) may have some evolutionary advantages and could even be considered necessary if it is well regulated, i.e., not excessive. However, too much of a good thing could ultimately lead to disastrous consequences at any age, e.g., free radicals.
Accordingly, we can propose a new paradigm for dynamic immune changes with aging [fig_ref] FiGURe 1 |: The new paradigm for the role of inflammaging and immunoadaptation/remodeling in the... [/fig_ref]. We suggest that aging leads to modified/modulated responses of the immune system, making it more adapted to cope with challenges (pathogens) in a given (local) environment, and not just to an eventually terminal deterioration of the immune system. From an evolutionary perspective, this is a simple optimization of the resources of the aging body, even if it ultimately leads to pathologies and death. From this perspective, many or most age-related changes in the immune system may be desirable adaptations to the aging process, and thus no need for rejuvenation seems to be necessary.
## Two important recent approaches to better understand immune changes with aging, supporting the need for a change of the current paradigm
There are two new approaches which can be adopted to re-conceptualize immune changes with aging. These integrate almost all aspects mentioned above. Immune systems of the elderly are remodeled with fewer naive cells and dysfunctional (exhausted vs. senescent) memory cells, due to chronic antigenic stimulation (including, but not limited to, CMV and neo-antigens from emerging malignant cells) and thymic involution, with altered innate immune response resulting in inflamm-aging eventually contributing to some age-related disease development [fig_ref] TABLe 1 |: Summary of some immune changes associated with aging in innate and adaptive... [/fig_ref].
## How do inflamm-aging and immunosenescence stand from an evolutionary perspective?
Considering all the alterations in the immune system with aging, the question arises whether and how inflamm-aging and immunosenescence can be the cause of these numerous agerelated alterations and pathologies attributed to them. Recent [bib_ref] Efficacy of an adjuvanted herpes zoster subunit vaccine in older adults, Lal [/bib_ref]. An additional recent report has further suggested that inflammation is a driving force for longevity in super semi-centenarians. In another study [bib_ref] New advances in CMV and immunosenescence, Sansoni [/bib_ref] , Franceschi's group determined HCMV prevalence in 132 centenarians, 245 centenarian offspring, and 101 offspring of non-long-lived parents. These authors found that infection did not impact on the longevity of these elderly individuals. Finally, the studies concerning the diversity of the microbiota in centenarians may also support this changing paradigm as dysbiosis is not always the equivalent of dysregulated inflamm-aging, particularly in centenarians and semi-super centenarians. The changes in the composition of the gut microbiota with age in subjects ranging from 22 to 109 years can be mentioned as one of the best example of remodeling, with possible large influence on inflamm-aging and immunosenescence, taking into account how important is the gut microbiota for the immune system. An increase in sub-dominant species and among them, species which are considered very "good, " was observed in Italian, Japanese, and Chinese centenarians despite the differences of diet and genetics in aged subjects, particularly in semi-supercentenarians [bib_ref] Gut microbiota changes in the extreme decades of human life: a focus..., Santoro [/bib_ref] [bib_ref] Gut microbiota and extreme longevity, Biagi [/bib_ref]. This situation did not involve a chronic uncontrolled inflammation, but the well-balanced inflammatory and antiinflammatory equilibrium. Immune changes observed during aging may thus only represent an adaptation to a challenging environment (containing mostly the cognate pathogens, with the exception of cancer cells generated by semi-random mutations) that results in maintenance of homeostasis via hormesis. However, under conditions of not previously encountered pressure (i.e., contact with a novel, previously unknown pathogen), the aging immune system either can adapt by using the available reserves. Conversely, if it is unable to do so, that leads to a maladaptation manifested by age-related diseases or a lethal outcome.
From this perspective, it is interesting to consider the putative role of inflamm-aging in frailty. The definition of frailty is already controversial as two main designations exist, namely the phenotypic definition [bib_ref] Frailty in older adults evidence for a phenotype, Fried [/bib_ref] and the multiple composite burden (deficit accumulation) [bib_ref] A comparison of two approaches to measuring frailty in elderly people, Rockwood [/bib_ref]. Frailty in fact can be conceptualized from an evolutionary point of view as the decrease of the physiological/biological/molecular reserves of the aging organs/ organism, leading to less efficient responses to stresses and, Frontiers in Immunology | www.frontiersin.org January 2018 | Volume 8 | Article 1960 therefore, producing deleterious effects, even death [bib_ref] Aging, frailty and age-related diseases, Fulop [/bib_ref]. This could be considered normal (usual) aging, in contrast to successful aging on the one hand or pathological aging on the other hand. Independently of its definition, one of the most accepted causes of frailty is inflamm-aging [bib_ref] Emerging roles of frailty and inflammaging in risk assessment of age-related chronic..., Wu [/bib_ref] that represents the biological threshold between successful and pathological aging. This event suggests a dynamic process that could still be reversed if the underlying causes such as inflamm-aging were contained. However, this situation may also progress to death through diseases when it becomes uncontrolled and hyperinflammatory, as would be predicted by the trained innate memory process [bib_ref] From inflamm-aging to immune-paralysis: a slippery slope during aging for immune-adaptation, Fulop [/bib_ref]. From this perspective, the extent of symptoms included in the clinical picture of frailty may be considered as a surrogate measure for biological age independently of chronological age, indicating whether inflamm-aging tends toward health or disease/vulnerability. Then an "optimal inflamm-aging" may be defined for longevity and health (optimal aging). Eventually, as is the case with all biological processes, an equilibrium is needed with functional checkpoint gatekeepers. If, for any reason, this equilibrium is perturbed the pathological pathway may prevail. The aim of optimization efforts and approaches would be to help maintain this very complex equilibrium to achieve an adequate functional longevity for optimal aging.
## A dysregulatory approach integrating the immune system and other systems
Clearly, the immune system does not exist in isolation but is influenced by and, in turn, influences many other systems such as the central and the peripheral nervous system, the endocrine system and others [bib_ref] Detection of a novel, integrative aging process suggests complex physiological integration, Cohen [/bib_ref]. This fact fits perfectly with the new approach of the study of aging which states that aging is the sum of results of the dysregulation of different system(s) from a normal regulatory level (i.e., a homeostatic state). This state is not necessarily identical between young and old subjects and may well reflect adaptations to intrinsic (GARBAge) and extrinsic changes (mainly pathogens and adverse environmental influences) related to aging. Thus, dysregulations are neither obligatorily detrimental nor beneficial but indicate a state of dyshomeostasis. This statement leads to the important insight that the pro-inflammatory state cannot be considered separately from the anti-inflammatory state [bib_ref] Inflamm-aging does not simply reflect increases in pro-inflammatory markers, Morrisette-Thomas [/bib_ref]. In this respect, this possibility will likely be expanded and nuanced by inclusion of other systems and cellular subsystems, e.g., mitochondrial metabolism. More broadly, this approach suggests that there are clear limits to the relatively linear, pathway-based, reductionist approaches to understanding physiology in general and immunology in particular. "Optimal" levels of various cell types, surface markers, and cytokines are unlikely to be very high or very low, but often intermediate, suggesting non-linear associations with risk. These optimal levels are also likely to vary depending on many other factors, such that it will be quite tricky to identify generally healthy or unhealthy states. Because of the possibility that changes with age or health state represent adaptations rather than aspects of pathology, substantial care must be exercised when interpreting changes in the system. This way of thinking is a completely innovative approach which can be studied by various statistical analyses using smaller or larger databases obtained in the studies of elderly subjects. This approach can also lead to the discovery of new biomarkers and their use in clinical settings. Such a broad approach would allow the integration of the omics, single cell assessment and systems biology. Different statistical approaches may be warranted both in cases when the dysregulation(s) follow a single unidirectional pathway (as appears to be the case for inflamm-aging and likely metabolic syndrome) [bib_ref] Detection of a novel, integrative aging process suggests complex physiological integration, Cohen [/bib_ref] [bib_ref] Validating metabolic syndrome through principal component analysis in a medically diverse, realistic..., Dusseault-Belanger [/bib_ref] , as well as in cases where dysregulation can produce a wide array of phenotypes sharing little beyond their departure from a homeostatic state [bib_ref] Homeostatic dysregulation proceeds in parallel in multiple physiological systems, Li [/bib_ref] [bib_ref] Trajectories of physiological dysregulation predicts mortality and health outcomes in a consistent..., Milot [/bib_ref]. An appealing hypothesis is that canalized dysregulations occur as a result of adaptations to the aging process and thus reflect an optimized response to an imperfect situation. In contrast, non-canalized dysregulations reflect a true loss of homeostatic control and may themselves be the imperfect situation causing the canalized responses.
Identification of cell types in both the innate and adaptive immune systems that are affected by age-related changes would be an additional application of these kinds of integrative statistical approaches. Immunology has generally considered individual cells to belong to discrete types that can be distinguished based on their surface markers. Certainly, this is a valid paradigm for many of the major classes of immune cells (T-cells vs., B-cells, CD4 + vs. CD8 + , etc.), but many examples of less distinct, partially overlapping classes are starting to emerge: classes based on the levels of surface receptors rather than just their presence or absence, or classes confounded by some subpopulations that are simultaneously expressing two markers that were supposed to distinguish populations (e.g., CD45RA + vs. CD45RO + ). It would thus appear that variation in cell surface receptors can happen in different ways. Discrete variation produces populations of cells with distinct functions and properties ("classes" or "types"), whereas continuous variation produces cells with functions and properties that vary along a gradient. This distinction is important because if one wrongly considers cells varying along a gradient as being from discrete classes, one is likely to (a) misidentify many cells with intermediate phenotypes and (b) to fail to understand the true biological processes driving cell diversity, thus wrongly interpreting the functional consequences of this diversity. For example, one can suppose that a population of cells varies along a gradient that determines their affinity for two types of pathogens (say, A and B). The cells with high affinity for A would have low affinity for B and vice versa. If there were a true gradient, a reasonable strategy would be to have a large population of cells with an intermediate affinity for both, maximizing the flexibility of the response. This may be a particularly good strategy during aging when the total cell population declines and the ability to maintain large numbers of cells at both extremes of the gradient is compromised. If one wrongly divides cells into A-affinitive and B-affinitive types, one may obtain uninterpretable or confusing results, depending on where along the gradient the threshold is set. It would then be not possible to understand any strategies that involve the use of intermediate values along the gradient. Immunosenescence and Inflamm-Aging: Friends or Foes?
Frontiers in Immunology | www.frontiersin.org January 2018 | Volume 8 | Article 1960
Should we intervene and How?
If we consider immune changes related to aging as an adaptation/ remodeling, interventions are presently very difficult to foresee. In particular, a single, generalized immune intervention does not appear to be likely. Anti-inflammatory interventions may depend on the state (level) of inflammation and its duration and, on interactions of the innate immune system with other systems, as well as on the appropriate inflammatory state of the individual given age and disease status. If one assumes that the immune/ inflammatory system in the elderly/aged organism is adapted/ remodeled in order to provide the best possible anti-pathogen protection when the adaptive immune system fails, the rejuvenation approach as currently proposed (e.g., IL-7/IL-15) seems likely to cause potential long-term harm in the aged organism. Perhaps, more general, however, purposeful interventions may be necessary, such as lifestyle interventions with personalized exercise and nutrition. Specific epigenetic diets may have their role in this modulation [bib_ref] Vanden Berghe W. From inflammaging to healthy aging by dietary lifestyle choices:..., Szarc Vel Szic [/bib_ref] [bib_ref] Epigenetic clock analysis of diet, exercise, education, and lifestyle factors, Quach [/bib_ref]. Some drugs with global action have been suggested to decrease the (over)activation of the immune/inflammatory system. One such drug is metformin, suspected for a long time to be a powerful modulator of aging [bib_ref] Metformin reduces all-cause mortality and diseases of ageing independent of its effect..., Campbell [/bib_ref]. Still, its effects in aging immune/inflammatory system are not yet clear. In any case, any interventions will need to be personalized and the immune history (immunobiography) of the individual will have to be taken into account.
## Conclusion and future perspectives
Aging is a highly complex process but an increased understanding of the process should lead to the efficient treatment of the many age-related diseases. The immune system interacts with many other systems in the organism (mainly the neural, metabolic, and the endocrine systems) and is, therefore, one of the most ubiquitous master systems of the organism. As such, it orchestrates health when it functions well but, when maladapted, it leads to diseases in the aging organism. Many changes in the immune system with age have been described and most of them have been considered deleterious and causes of many age-related diseases. Changes occur in both the innate and the adaptive immune arms of the immune system, but perhaps not to the same extent or with the same consequences. There is an intricate interrelationship between inflamm-aging and immunosenescence, which are nearly identical in some ways but very different in other aspects and, occurring in concert, mutually influencing each other. Future studies are obviously necessary to elucidate these interactions and raise targets for new interventions to decrease the deleterious effects of aging and use the beneficial effects for a better health and functionspan in the elderly. Therefore, the phenomenon traditionally termed "immunosenescence" may be considered an immunoremodeling/adaptation as a result of chronic aggressions and time. Immunosenescence may be necessary for an adequate response to known antigens, but detrimental for responses to new antigens in most circumstances. The discovery of new processes, new immune cell subtypes, and the integration of genetic/epigenetic/metabolic and environmental factors (nutrition) will nuance our «evil» and apparently mistaken perception that aging-associated immune changes are only detrimental. Immunosenescence/inflamm-aging may contribute to diseases such as cancer but its role during aging is still controversial. Elderly in clinical settings are doing much better than predicted from experiments thus, human studies in particular are badly needed.
In view of the successes of cancer immunotherapy and vaccination in the elderly, no such intervention should be refused to an elderly subject based on a dogmatic assumption that agingrelated immune changes are detrimental. Thus, time is of the essence; and the future is already now for the elderly.
# Author contributions
TF has written and conceptualized the article; AL, GD, AP, EF, AC, JW, and CF have contributed to the writing and critically read it.
# Funding
This work was partly supported by grants from the Canadian Institutes of Health Research (No. 106634 and No. 106701) to TF, the Université de Sherbrooke, and its Research Center on Aging; the Polish Ministry of Science and Higher Education (statutory grant 02-0058/07/262) to JW, the Agency for Science Technology and Research (A*STAR) to AL. AC is supported by a New Investigator Salary Award from the Canadian Institutes of Health Research and is a member of the Fonds de Recherche du Québec-Santé (FRQS)-supported Centre de Recherche sur le Vieillissement and Centre de Recherche Clinique du Centre Hospitalier Universitaire de Sherbrooke (CHUS). CF is supported by progetto CARIPLO rif. 2015-0564.
## References
[fig] FiGURe 1 |: The new paradigm for the role of inflammaging and immunoadaptation/remodeling in the aging process. *Optimization: all three processes increase in concert, balancing each other. **Deterioration: inflammaging increases, and is not balanced by opposite processes of antiinflammaging and immuneadaptation/ remodeling, which are decreasing. We mean by antiinflammaging all compensatory mechanisms which emerged to compensate the chronic inflammaging. The most important diseases that could have an inflammaging component are cancers, cardiovascular diseases, and neurodegenerative diseases. to challenge the established view of the role of immune changes with aging. The publication of Lal et al. cited above that reported a positive response to a new herpes zoster vaccine even in the very old raises the question of the role of immunosenescence and inflamm-aging in the decreased response to vaccination [/fig]
[table] TABLe 1 |: Summary of some immune changes associated with aging in innate and adaptive immune systems. [/table]
|
The Evolution of Polycentric Governance in the Galapagos Small-Scale Fishing Sector
Addressing the multiple anthropogenic and non-anthropogenic factors affecting small-scale fisheries requires collaboration from diverse regions, geographical scales, and administrative levels in order to prevent a potential misfit between governance systems and the socio-ecological problems they address. While connecting actors and stakeholders is challenging, as they often hold opposing perceptions and goals, unveiling the network configurations of governance systems remains one effective way to explore collaborative alliances in light of the diverse drivers of change present in small-scale fishery systems. This study employed descriptive statistics, exponential random graph models (ERGMs), and qualitative data analysis to explore preferential attachments of new nodes to well-positioned nodes within the Galapagos small-scale fishery governance system network and the propensity of cross-sectoral reciprocity and cross-sectoral open triads formation in the network. Our findings identified significant players and network configurations that might be essential in the collaboration diffusion and robustness of the Galapagos small-scale fishery sector governance system.
# Introduction
Today, small-scale fishing governance systems face different challenges in formulating strategies capable of addressing multiple problems. We live in an increasingly interconnected world, where problems in a social-ecological system originate in numerous and simultaneous interactions and exposures from local and global scales [bib_ref] Polycentric systems: multilevel governance involving a diversity of organizations, Ostrom [/bib_ref] ; [bib_ref] The social structural foundations of adaptation and transformation in social-ecological systems, Barnes [/bib_ref] [bib_ref] Environmental governance for the anthropocene? Social-ecological systems, resilience, and collaborative learning, Berkes [/bib_ref]. The effects of global and local dynamics at present-characterized by high uncertainty, complexity and unexpected changes-make social-ecological transformations and uncertainty an inevitable occurrence in small-scale fisheries systems. Consequently, aligning small-scale fishery governance systems with the social-ecological dimensions they are meant to address also becomes challenging.
Small-scale fishery governance systems should consider spatial scale (i.e., capacity to match a social-ecological system's geographical extent), temporal scale (i.e., capacity to act on time), and functional scale (i.e., capacity to match a socialecological system's functional dynamics and interactions); recognizing this is crucial when dealing with complex socialecological systems [bib_ref] The problem of fit among biophysical systems, environmental and resource regimes, and..., Galaz [/bib_ref] [bib_ref] Disentangling intangible social-ecological systems, Bodin [/bib_ref] [bib_ref] Conservation success as a function of good alignment of social and ecological..., Bodin [/bib_ref] [bib_ref] Institutional fit and the sustainability of social-ecological systems, Epstein [/bib_ref]. However, it is necessary to recognize that the management scale of those governance systems encompasses multiple types of fit simultaneously to span a socio-ecological system's scope in the face of change [bib_ref] How does network governance affect social-ecological fit across the land-sea interface? An..., Pittman [/bib_ref] [bib_ref] Identifying governance gaps among interlinked sustainability challenges, Bergsten [/bib_ref] [bib_ref] Achieving multiple socioecological institutional fits: the case of spiny lobster comanagement in..., Ishihara [/bib_ref]. Today, the management capacity of governance systems depends not only on its ability to fit with environmental and ecological concerns but also on its ability to fit with various societal problems and stakeholders' expectations [bib_ref] Does polycentricity fit? Linking social fit with polycentric governance in a large-scale..., Acton [/bib_ref] [bib_ref] Achieving multiple socioecological institutional fits: the case of spiny lobster comanagement in..., Ishihara [/bib_ref]. Global sustainability challengespressure governance systems to align as much as possible with the spatial, temporal, and functional dimensions of the system (e.g., with the interactions between marine species and fishers' actions). Moreover, unexpected socio-economic and environmental changes and needs can emerge in socio-ecological systems (e.g., due to the adverse impacts of novel pandemics such as 2019 novel coronavirus or COVID-19, climate change, or unreported fishing); this has broadened the management scope of social-ecological governance systems and the need to address "the problem of fit" more closely [bib_ref] The problem of fit among biophysical systems, environmental and resource regimes, and..., Galaz [/bib_ref] [bib_ref] Closing integrative gaps in complex environmental governance systems, Fried [/bib_ref].
How society responds to the evolving conditions through collaborative approaches is an essential component of addressing the problem of fit in small-scale fishery governance systems. By recognizing this, this paper aims to improve the Galapagos small-scale fishery collaboration network and the notion of governance fit within the Galapagos smallscale fishery sector by considering attributes stemming from institutional fit, adaptive co-management, polycentrism and subsidiarity (summarized in [fig_ref] Figure 1: Governance fitFig [/fig_ref]. Here, we offer a methodological approach that draws on social network analysis and qualitative data analysis. This novel research approach enables analysts to represent, capture and unveil relationships and interdependencies in social and ecological environments; we thus employ it to examine specific network patterns and configurations that may strengthen the collaborative links of Galapagos small-scale fishery governance system.
We explore in this study the preferential attachment of nodes (i.e., the likelihood of adding collaborative ties to well-positioned nodes) in the Galapagos small-scale fishery governance network. In doing so, we employ descriptive statistics (centrality measures), estimate the propensity toward reciprocity and open triad formation (as explained in across sectors using exponential random graph models (ERGMs), and analyze interviews. Throughout the paper, we use nodes, referring to those organizations and agencies connected to the Galapagos' small-scale fishery governance network and those organizations that may be part of the collaborative network of Galapagos small-scale fishery governance in the future. At the same time, we refer to links/ties to the organizations' connectivity in terms of other organizations and agencies.
In this paper, we argue that institutions and agencies may be able to more wisely discern how to choose and create collaboration partners based on the nodes' positions, features and needs, rather than leaving it to chance or to policies and laws to define collaboration ties. This implies that organizational ties in a governance system network can become more dynamic, moving from delegated organizational links to organizational ties where actors can make choices regarding the partners with whom they collaborate. Our theoretical framework might guide practitioners as to the spread and allocation of elements needed in socialecological governance systems networks, such as governmental and international support, including financial aid, economic incentives and subsidies, technology, data exchange, and co-production of knowledge, along with other determinants and instruments deemed significant in building robust collaborative networks.
The theoretical approach of this paper provides stakeholders with a broader image of where collaboration links might have a more extensive influence on collaboration diffusion in a network. It offers stakeholders a platform for evaluating whether collaboration alliances need to be created, enhanced or reformulated. Stakeholders may analyze whether or not it is necessary to create mutual links (A ↔ B) or include a third collaboration party C into an existing A-B collaboration. If so, the inclusion of a new collaboration partner would lead to an open triadic or a triadic closed organizational network configuration , where organizations involved can benefit and strengthen each other by sharing organizational goals and resources. : if there are mutual interactions between organizations and agencies (A ↔ B) in a governance structure, such organizations and agencies are likely share efforts, such as financial resources, technicians, knowledge, and data. They also serve as baseline for the formation of open or closed triadic network configurations, implying further diffusion and propagation of collective efforts. This often starts when organizations and agencies create an initial organizational link and reciprocate organizational ties (A ↔ B). : if collaboration connections A-B and B-C exist, it is likely that a new organizational link A-C would be formed (red line in , giving rise to a triadic closure configuration. The analysis of open triads enables us to indicate the likelihood of partners of partners to become collaboration partners, which implies that the A-B and B-C collaboration ties might be transmitted to A-C in a governance system structure [bib_ref] How to close a hole: exploring alternative closure mechanisms in interorganizational networks, Lomi [/bib_ref] [bib_ref] How does network governance affect social-ecological fit across the land-sea interface? An..., Pittman [/bib_ref].
## Paths toward collaboration and polycentric links in galapagos
Although the conservation of Galapagos marine resources was not a priority at the time, institutional ties to protect these resources date back to the 1960s, when the Charles Darwin Research Station (CDRS), the first international research organization in the Islands, and the Galapagos National Park (GNPS), the first governmental organization for conservation purposes in Galapagos, were created and signed the first collaboration agreement to foster research and conservation in the Galapagos. This agreement marked a turning point for the development of Galapagos fishery science by giving rise to a series of institutional links in the Galapagos, which were initiated when the GNPS and the CDRS requested that US Peace Corps volunteer Jerry Wellington explore coastal intertidal and subtidal ecosystems of Galapagos in the 1970s. Wellington's outcomes highlighting the marine biodiversity and endemism of the Galapagos served to consolidate the first official inter-institutional cooperation agreement between the CDRS and the National Fisheries Institute (Spanish acronym: INP) in 1976, joined a year later by the University of Guayaquil, in order to explore the Galapagos fishery resources state in terms of abundance and distribution. This effort gave rise to the first triadic network configuration involving reciprocated ties in the Galapagos small-scale fishery sector network (A ↔ B; B ↔ C; C ↔ A), represented by the red links in (i.e., network configurations that narrow and facilitate collaboration in networks).
Galapagos marine resources have been subjected to fishery exploitation since the late eighteenth century. British and North American whalers and sealers pioneered commercial exploitation in the archipelago. Sperm whales (Physeter macrocephalus), fur seals (Arctocephalus galapagoensis), and Galapagos sea lions (Zalophus wollebaeki) were the primary targets species. Notably, the demands of the Asian market for shark fins, together with the sea cucumber (Isostichopus fuscus) capture in the 1980s and later collapse in the 1990s in close collaboration between Asian intermediaries with Galapagos local fishers and fishers from coastal provinces of Ecuador; prompted great interest in the management, conservation and commercialization of. As a result, between the 1980s and 1990s, the number of immigrants from Ecuador's mainland, small-scale fishing fleets, and tourists on the islands increased significantly, giving rise to the establishment and interests of diverse scientific institutions, governmental and non-governmental bodies, and various local fishing cooperatives, as well as diverse legal provisions, institutional arrangements and strategies, shaping changes from a topdown command control form of governance to one with more polycentric links in the Galapagos.
The completion of the management plan of the Galapagos Marine Reserve (Spanish acronym: PMRMG) by the so-called Grupo Nucleo in 1994, the preparation process and later adoption of the so-called Galapagos Special Law (GSL) in 1998, that led to the Marine Reserve (GMR) establishment and the Galapagos co-management system (GCM) implementation, as well as the 2007 inclusion of the Galapagos Islands into the list of endangered World Heritage Sites by UNESCO [bib_ref] Political dynamics and governance of World Heritage ecosystems, Morrison [/bib_ref] , marked significant milestones in constructing governance environments with more polycentric links by prompting diverse ties between national public and private international and local organizations and agencies.
Significantly, the GCM, administered mainly from the governmental side, gave rise to delegated institutional ties under two management bodies: the so-called Participative Management Board (PMB), formed by representatives from the GNPS, the small-scale fishery (elected among the Galapagos Fishing Cooperatives), the Galapagos Chamber of Tourism, the CDRS and naturalist guides to represent the local level; and the so-called Inter-institutional Management Authority (IMA), formed by representatives from three ministries based on Ecuador's mainland (Ministry of Environment, Ministry of Defense and Ministry of Foreign Trade, Industrialization, Fisheries and Tourism), representatives of local sectors (the small-scale fishery sector and the Galapagos Chamber of Tourism) and the Ecuadorian Committee for the Defense of Nature and the Environment (Spanish acronym: CEDENMA) to represent a higher level of the decision-making process and decide if there was no consensus among the representatives of the PMB at the local level. Under the IMA structure, CDRS acted as a technical advisor and the GNPS as a technical secretariat for the Ministry of Environment.
Since the reform of the GSL in 2015, the Galapagos cogovernance has been changing its original governance structure. With GSL reforms, the PMB and the IMA were repealed, giving rise to new delegated organizational links -formed by, and run primarily from, the governmental side -to lead decision-making processes. Today, the GSL is being amended, giving rise to discussions to consolidate a new consultative governance scheme, whose operational legal framework remains unclear and inactive. Therefore, collaboration and organizational ties in the Galapagos small-scale fishery sector, involving actors from diverse administrative levels and scales, continue to change due to changes on the governance structure and the creation of new management tools, including the management plans of the Galapagos National Park Directorate (Spanish acronym: DPNG) and the Galapagos Special Regime Governing Council (Spanish acronym: CGREG).
Nodes indicate organizations and agencies. Ties represent the organizational connections between organizations and agencies. The red ties of show the first triadic network configuration involving reciprocated ties in the Galapagos small-scale fishery network, as described above. Note: Despite the vital role that Asian intermediaries played in the development of the Galapagos sea cucumber fishery, they were not recognized as a sector or actor in the Galápagos fishery system; therefore, they were not members of the PMB. This gave rise to two parallel management systems: the GCM, and the system formed by Asian intermediaries who, in partnership with local fishers, set up clandestine camps to catch and process sea cucumbers. See also the discussions regarding evolving polycentric governance of the Great Barrier Reef in who initially coined the term delegated polycentricity, and the role of Asian intermediaries in the exploitation of Galapagos sea cucumbers in.
## Moving beyond co-management
Although co-management has undoubtedly been a significant approach in response to the limitations of centralized, top-down governance, as well as the increasing demands of natural resource users and local communities to be part of the decision-making processes that affect their livelihoods, it is essential to note that the social-ecological interactions that span the small-scale fisheries systems are more complex and dynamic than the way that comanagement literature initially considered them. The human and ecological environments of small-scale fisheries change day to day; this is due to significant problems that create multiple socio-ecological interactions beyond the comanagement scope as a category of institutional arrangements to share power and responsibility between the government and local resource users. Today, we have witnessed closely that we live in a new era of the Anthropocene [bib_ref] Coral reefs in the Anthropocene, Hughes [/bib_ref] [bib_ref] Advancing coral reef governance into the Anthropocene, Morrison [/bib_ref]. The uncertain behavior of complex social-ecological interactions has broadened the small-scale fisheries governance scale. The incomplete transition toward the new Galapagos governance system established by the new GSL, in combination with the adverse impacts of the COVID-19, climate change and illegal, undeclared and unregulated fishing by national and international fleets, makes evident the need to explore other governance forms and further organizational links at diverse geographical and administrative levels, from local to international beyond the Galapagos Marine Reserve protected area and the DPNG jurisdiction that enable to align the Galapagos small-scale fishery governance system as much as possible with the extent, timing and functional diversity of social-ecological systems interactions and prevent a misfit.
Achieving an approximation to such socio-ecological fit requires strategic approaches that support the cooperation and interaction of diverse public and private actors from various jurisdictional levels and geographical scales in order to ensure more sustainable outcomes [bib_ref] Resilience and sustainable development: building adaptive capacity in a world of transformations, Folke [/bib_ref] [bib_ref] Enhancing the fit through adaptive co-management: creating and maintaining bridging functions for..., Olsson [/bib_ref] [bib_ref] Local sustainability initiatives in English National Parks: what role for adaptive governance?, Clark [/bib_ref]. Adaptive co-management (AC) is an emerging approach for common-pool resources management that enables the delivery of responses to socialecological changes operating on multiple scales and levels, guided by subsidiarity principles and polycentricism [bib_ref] Adaptive governance of social-ecological systems, Folke [/bib_ref] [bib_ref] Polycentric systems for coping with collective action and global environmental change, Ostrom [/bib_ref] [bib_ref] Diagnosing adaptive comanagement across multiple cases, Plummer [/bib_ref] [bib_ref] Polycentric systems of governance: a theoretical model for the commons, Carlisle [/bib_ref]. The subsidiarity principle implies that actions should be taken at the lowest practical level of governance, which in complex social-ecological systems ensures that decisions are made as near as possible to those whose livelihoods might be affected by decision-making structures. Significantly, the subsidiarity principlesometimes referred to as "good governance"-provides an important platform for taking into account the proper stakeholders and local priorities; disregarding these considerations could reinforce the current status quo, which often reflects political economic inequalities and vested interests [bib_ref] Environmental governance and its implications for conservation practice, Armitage [/bib_ref]. Different from monocentric forms of governance characterized by hierarchical governance structures (e.g., driven by a governmental authority or private monopoly) [fig_ref] Figure 4 a: Monocentric forms of governance, b polycentric forms of governance [/fig_ref] that span broad geographical scales and administrative levels in order to act as close as possible to social-ecological interactions and the underlying causes of vulnerability [bib_ref] Adaptive governance of social-ecological systems, Folke [/bib_ref] [bib_ref] Polycentric systems for coping with collective action and global environmental change, Ostrom [/bib_ref] [bib_ref] Diagnosing adaptive comanagement across multiple cases, Plummer [/bib_ref].
Nodes indicate organizations and agencies in a governance structure. Ties indicate the organizational links between organizations and agencies. The gray nodes in 4a represent either governmental organizations (e.g., in a common-pool resource governance system with a strong presence of the government over decisions) or private organizations (e.g., in a monopoly). Node color in 4b indicates the economic sector (blue nodes = private sector organizations and agencies; gray nodes = public sector organizations and agencies). This implies cross-sectoral interaction between different organizations regardless of their economic sector and administrative level. Furthermore, it is important to point out that while the existence of multiple semi-autonomous decision centers might be enough to deem a governance arrangement as polycentric, it does not mean that there will be enough coordination among such centers to ensure that a system acts as a polycentric governance system, see discussion in [bib_ref] Polycentric systems of governance: a theoretical model for the commons, Carlisle [/bib_ref] and [bib_ref] Polycentric to monocentric governance: power dynamics in Lake Victoria's fisheries, Mudaliar [/bib_ref]. The latter consideration is particularly important in terms of the influence of power in the management of common-pool resources. Power dynamics are pivotal to defining polycentric systems and coordination among decision-making centers [bib_ref] Polycentric to monocentric governance: power dynamics in Lake Victoria's fisheries, Mudaliar [/bib_ref]. Without the actual intention to share power, crosssectoral and cross-level interactions are challenging to achieve, keeping a system from functioning as one polycentric governance system [bib_ref] The black box of power in polycentric environmental governance, Morrison [/bib_ref] [bib_ref] Polycentric to monocentric governance: power dynamics in Lake Victoria's fisheries, Mudaliar [/bib_ref]. At the same time, it is essential to bear in mind that an unclear distribution of responsibilities among decisionmaking centers in polycentric systems may give rise to confusion and functional and geographical overlaps between higher and lower administrative levels, also hindering the polycentric governance system [bib_ref] Cross-scale linkages in connectivity conservation: adaptive governance challenges in spatially distributed networks, Wyborn [/bib_ref] [bib_ref] Crowding-out lower-level authorities: Interactions and transformations of higher and lower-level authorities in..., Mudliar [/bib_ref].
Much of the criticism placed on common-pool resources governance systems has emerged because, among some reasons and deficiencies, they tend to suggest panacea/ blueprint solutions for all types of problems (i.e., fixed standard universal solutions for various issues, see discussion in [bib_ref] A diagnostic approach for going beyond panaceas, Ostrom [/bib_ref] and [bib_ref] Moving beyond panaceas: a multi-tiered diagnostic approach for social-ecological analysis, Ostrom [/bib_ref]. The complexity that embraces small-scale fishery socialecological systems' interactions demands management strategies and policies should be viewed as place-specific experiments that can be revised, adapted and changed as different social-ecological circumstances demand [bib_ref] Resilience and sustainable development: building adaptive capacity in a world of transformations, Folke [/bib_ref] [bib_ref] Adaptive governance of social-ecological systems, Folke [/bib_ref] [bib_ref] Adaptive comanagement and the paradox of learning, Armitage [/bib_ref]. AC is an evolving framework that provides elements to be learned via experimentation and learning from joint actions on broad geographical scales and administrative levels (i.e., learn by doing) [bib_ref] Adaptive comanagement for social-ecological complexity, Armitage [/bib_ref] [bib_ref] Polycentric systems for coping with collective action and global environmental change, Ostrom [/bib_ref]. AC provides platforms that allow the participation of various stakeholders from local to broader non-local organizations and actors-possessing different sorts of resources such as social memory, financial resources, knowledge and data, among other adaptive capacity determinants, which can be activated when needed to navigate the dynamic nature (nonlinear relationship) of interconnected socio-ecological dimensions (complex systems thinking) to deal more appropriately with uncertainty and rapid changes of smallscale fishery social-ecological systems [bib_ref] Resource and environmental management, second. Routledge, London Morrison TH (2017) Evolving polycentric..., Mitchell [/bib_ref].
## Case study
Our case study focuses on the Galapagos small-scale fishery sector, a crucial socio-economic sector in the biodiversity hotspot that inspired Darwin's theory of evolution, located 1200 km off the Ecuadorian coastline [fig_ref] Figure 5: Location map [/fig_ref]. We focus our study on this sector considering that it plays a significant role in providing seafood to~30,000 residents and 271,000 tourists who arrive annually in the Galapagos (in pre-COVID-19 conditions), making it a crucial sector for the food security of the archipelago. The case study of the Galapagos small-scale fishery sector serves to highlights today's need for governance systems to deal with the unforeseen trans-boundary social-ecological interactions (e.g., due to the effects of COVID-19) present in complex socio-ecological systems. These have affected diverse fishing communities in the islands due to the linkage of the fishery sector with the tourism sector. Fishing communities are seafood suppliers assisting the development of tourism, the main livelihood and source of income in the Galapagos. In this context, an approximate reduction of 73% in visitors to the Galapagos as a result of measures designed to reduce the spread of the COVID-19 virus and the number of people infected directly affected the socio-economic situation of the Galapagos small-scale fishing sector. The measures, which included the prohibition of all national and international tourist arrivals in the archipelago during the early months of the pandemic, and a subsequent mandatory negative polymerase chain reaction test for entry into Ecuador and the Galapagos, led to the number of visitors to the Galapagos dropping from 271,238 visitors in 2019 to 72,519 in 2020 (DPNG 2021).
# Methods
## Data collection
The study used various methods to collect data, since data collection coincided with the COVID-19 pandemic, which limited human contact. We explored the history and institutional interactions in the Galapagos through a review of previous studies on marine and conservation science development in the Galapagos (n = 41), including peerreviewed journal articles, policy documents, organizational records and institutional publications from the government and the private sector. To this end, we used Google and Google Scholar to search the following keywords: Galapagos governance, Galapagos small-scale fishery, Galapagos collaborative arrangements and governance, and GCM. We also used the reference list of relevant peer-reviewed papers about the development of marine and conservation science in the Galapagos as a guide to decide which articles to read. The review enabled us to examine organizations from different geographical scales and administrative levels and create a list of nodes that traditionally do not possess significant links within the Galapagos small-scale fishery governance system network (n = 28). However, they operate directly and indirectly in the Galapagos conservation and marine development areas (in normal conditions-pre-COVID-19). We used this list and the Galapagos smallscale fishery collaboration network of the work by Representatives and officials noted in (c) were presented with a series of open and closed questions. They were asked (1) how the respondent's organization might collaborate in the Galapagos small-scale sector if there were institutional arrangements in place (e.g., financial resources, technical and scientific knowledge, local knowledge acquired over time, data and information, equipment and technology, infrastructure, or monitoring of illegal fishing or research projects), and (2) about the administrative level (local, national or international) and economic sector (public or private) of their organizations. A Qualtrics software, Version 6.2020 of Qualtrics (Copyright © [2020] Qualtrics) was used to create our study questions (in Spanish), send personalized links to the individuals' institutional email address and store respondent's answers. Informed consent was obtained via an initial question in the Qualtrics survey. We collected the data of the study between June 2020 and December 2020. We kept the survey open from September 10 to December 18, 2020. This study received ethics clearance (ORE #41927) from our university's research ethics system.
# Data analysis
Representatives' and officials' answers [noted in (1) in the data collection section] were translated from Spanish to English, and transcribed and coded using the qualitative data analysis software NVivo (released March 2020) (QSR International Pty Ltd. 2020). The coding procedure, undertaken by the study's corresponding author, was both deductive and inductive. The codes were developed using categories from the question [noted in (1) in the data collection section]. We used Gephi network visualization 0.9.2 software [bib_ref] Gephi: an open source software for exploring and manipulating networks Berardo R,..., Bastian [/bib_ref] to suggest the preferential attachment of nodes [indicated in a) in the data collection section] using centrality measures, degree centrality, eigenvector centrality, and closeness centrality. We used PNet softwareto examine the propensity of reciprocity cross-sectoral formation and crosssectoral open triads formation in the network [indicated in b) in the data collection section]. For this purpose, we developed a series of hypotheses using a building blocks (motifs) approach (i.e., network configurations representing specific network patterns in an observed network), representing basic network configurations we deem significant preconditions to facilitate network collaboration within governance systems (see more regarding "building blocks," in [bib_ref] Network motifs: simple building blocks of complex networks, Milo [/bib_ref] and their application in various studies in Berardo and Scholz (2010) [bib_ref] Innovation, cooperation, and the structure of three regional sustainable agriculture networks in..., Levy [/bib_ref] and [bib_ref] External exposure, boundary-spanning, and opinion leadership in remote communities: a network experiment, Matous [/bib_ref].
To capture the propensity toward the network configurations/building blocks shown in , we used one asymmetric adjacency matrix (i.e., a value assignation of zeros and ones according to the presence or not of ties between nodes in the network) and two attribute matrices (i.e., a value assignation of zeros and ones according to the presence or not of nodes' attributes). In the adjacency matrix, organizational links in the Galapagos small-scale governance system network were set as 1, and the absence of the organizational relations was set as 0. In the first attribute matrix, public sector nodes were assigned as 1, and private sector nodes were assigned as 0. In the second attribute matrix, private sector nodes were established as 1, and public sector nodes were established as 0. We used these matrices and the parameters presented in to run two models on PNet software (see also. We tested whether the parameters converged at t-statistic <0.1 and had a good fit at goodness-of-fit < 0.1 [bib_ref] Illustrations: simulation, estimation, and goodness of fit, Robins [/bib_ref].
# Results
# Descriptive statistics results
Our descriptive statistical analysis identified actors whose position and centrality values within the network can contribute to and influence collaboration diffusion in the Galapagos small-scale fishery sector (e.g., CGREG, DPNG, fishing cooperatives, Charles Darwin Foundation (CDF). Our centrality analysis indicated that various actors with high centrality (i.e., nodes' that sent and received more collaboration ties compared to others in the network) were present in the network and [fig_ref] Table 1: Descriptive statistics of the Galapagos small-scale fishery sector [/fig_ref]. Specifically, these were: the governmental organizations GO01 and GO02, the fishing cooperative FC02, the governmental organization GO05, the fishing cooperative FC01, the international non-governmental organization NGO01, the governmental organization GO04, the municipal government MG01, the governmental organizations GO03 and GO06 and the international non-governmental organization NGO05, respectively (see and [fig_ref] Table 1: Descriptive statistics of the Galapagos small-scale fishery sector [/fig_ref].
Our analysis showed various actors with higher eigenvector centrality values compared to others in the network (i.e., nodes' importance based on their connections to influential nodes in the Galapagos small-scale fishery governance system, in other words the value of well-connected friends) and [fig_ref] Table 1: Descriptive statistics of the Galapagos small-scale fishery sector [/fig_ref]. Specifically, these were: the governmental organizations GO01 and GO02, the fishing cooperatives FC04, FC03, FC01 and FC02, the governmental organizations GO03, GO05, GO06, GO08, GO09 and GO07, respectively (see and [fig_ref] Table 1: Descriptive statistics of the Galapagos small-scale fishery sector [/fig_ref].
Our closeness centrality analysis indicated various actors with higher closeness centrality values than other organizations and agencies in the Galapagos small-scale fishing governance system network (i.e., nodes' importance based on their closeness to all nodes in the network) and [fig_ref] Table 1: Descriptive statistics of the Galapagos small-scale fishery sector [/fig_ref]. Specifically, these were: the governmental organizations GO01 and GO05, the international nongovernmental organizations NGO01 and NGO05, the governmental organization GO04, the fishing cooperative FC02, the municipal government MG01, the fishing cooperative FC01, the governmental organization GO06 and the international non-governmental organization NGO02, respectively (see and [fig_ref] Table 1: Descriptive statistics of the Galapagos small-scale fishery sector [/fig_ref]. Building blocks/hypotheses used when estimating cross-sectoral reciprocity and open triad network configurations of the Galapagos small-scale fishing governance system. ERGMs are a class of statistical models that enable capturing the presence or absence of specific network configurations in a social network. ERGMs provide a platform to statistically examine the propensity of building blocks in a more extensive network [bib_ref] Disentangling intangible social-ecological systems, Bodin [/bib_ref]. Pink nodes represent organizations and agencies from the private sector, and yellow nodes represent organizations and agencies from the public sector in the network
# Ergms results
In terms of cross-sectoral reciprocity configuration formations, we found no strong evidence that nodes from the private sector tended to reciprocate organizational links between them (Hypothesis 2,. However, we found a positive and significant propensity of nodes from the public sector to reciprocate organizational links between them (Hypothesis 1,. Significantly, we found a positive and significant propensity of nodes from the private sector and the public to return ties (Hypothesis 3,; this was notable considering the value of multi-sectoral links in decision-making structures of common-pool resource governance systems. Estimates on cross-sectoral open triad formation were positive and significant. Our results showed a positive and significant effect based on [Attr]-in-2-star (Hypothesis 4,, [Attr]-out-2-star (Hypothesis 5), and [Attr]-2-path (Hypothesis 6,parameters. We believe this signified cross-sectoral collaboration diffusion and likely diffusion of diverse determinants of adaptive capacity in the network (such as knowledge, technology, data, and expertise) needed to address diverse multidimensional internal and external factors of change that might affect the present and future stability of the sector.
# Qualitative data analysis
Effective collaborative responses to diverse simultaneous drivers of change necessitate embracing a socialecological perspective that involves different sorts of information, skills, and stakeholders at different geographical scales and administrative levels. Recognizing this is significant if we aim to improve the governance capacity to anticipate and adjust to simultaneous drivers of change, particularly in this era of constant change and evolution [bib_ref] Adaptation, adaptive capacity and vulnerability, Smit [/bib_ref]. Our results show that diverse organizations and agencies from various geographical and administrative levels, with no significant collaboration ties within the Galapagos small-scale fishery governance system, might collaborate with the Galapagos small-scale fishery system through diverse forms:
Our collaboration in the management of the artisanal fishing sector could be carried out through technical support and donation of equipment to strengthen the infrastructure they have and improve marketing strategies for their products. PG02 node of Figs. 7-9, Level: Local, Sector: Public. Degree centrality of the Galapagos small-scale fishery governance network. Nodes indicate the organizations and agencies (GO governmental organization, PO private organization, FA fishery association, NGO nongovernmental organization, MG municipal government, PG parish government, ARO academic and research organization). Ties indicate the organizational links between organizations and agencies. Green nodes indicate nodes connected to the Galapagos small-scale fishery network. Pink nodes indicate organizations and agencies that traditionally do not have significant organizational links with the Galapagos small-scale fishery governance system network. Node size indicates degree centrality, meaning that as the size increases, they send and receive more organizational links than others in the network, making them important players in this fishery governance network, as most of the links pass through them We might stimulate the consumption of local fishery products in the tourism sector and to report incidents or non-regulated vessels within the Galapagos Marine Reserve. GO15 node of Figs. 7-9, Level: National, Sector: Public.
We can deliver specific projects that can provide information for decision-making. NGO15 node of Figs. 7-9, Level: International, Sector: Private.
We constantly make reports of the guided visits, and we can provide information about the management of the fishing sector in the places of visit. P005 node of Figs. 7-9, Level: Local, Sector: Private.
Using the language employed by [bib_ref] The struggle to govern the commons, Dietz [/bib_ref] , it is important to note that governance systems should be viewed as a co-evolutionary race. While the existing Galapagos small-scale fishery sector collaborative network provides a significant umbrella to deal with multiple drivers of change, incorporating new actors at different geographical scales and administrative levels into the current Galapagos small-scale fishery collaborative network might lead to exploring further external cooperation links. The following quotes from interviewees are significant in that regard:
We have projects related to fisheries in other parts of the world whose experience and information could be made available to local actors. NGO07 node of Figs. 7-9, Level: International, Sector: Private. Eigenvector centrality of the Galapagos small-scale fishery governance network. Nodes indicate the organizations and agencies (GO governmental organization, PO private organization, FA fishery association, NGO non-governmental organization, MG municipal government, PG parish government, ARO academic and research organization). Ties indicate the organizational links between organizations and agencies. Green nodes indicate those nodes connected to the Galapagos small-scale fishery network. Pink nodes indicate nodes that traditionally do not possess significant organizational links in the Galapagos small-scale fishery governance system network. Node size indicates eigenvector centrality, which signifies that, as a node's size increases, it is deemed more important in the network based on its connections to important players in the fishery governance system. Larger nodes are thus influential players in the network, able to reach important organizations and agencies and diffuse critical information and knowledge in the Galapagos small-scale fishery network We have research groups at both the University of Malaga and the Spanish Institute of Oceanography based in Fuengirola (Malaga) with experience in fisheries. ARO10 node of Figs. 7-9, Level: International, Sector: Private.
We are a multidisciplinary research center that brings together researchers from different universities in Ecuador and the world. Our alliances with academia are very important in developing knowledge, information gathering, and training that contribute to sustainability. ARO04 node of Figs. 7-9, Level: International, Sector: Private.
We could sign an Inter-institutional Cooperation Agreement with the fishing sector to finance projects of interest. PG01 node of Figs. 7-9, Level: Local, Sector: Public.
In adaptive co-management, continuous learning is crucial in approximating a governance system as close as possible to one desired functional state. From a governance perspective, learning refers to the process of detecting and correcting errors to achieve better outcomes over time. In this context, the literature of socialecological systems often differentiates between different types of learning include single-loop learning (i.e., correcting mistakes by adjusting resource management strategies and actions), double-loop learning (i.e., correcting errors by adjusting behaviors and attitudes) and triple-loop learning (i.e., addressing conflicts by designing or revising governance norms and protocols to produce significant changes in governance) [bib_ref] Adaptive comanagement and the paradox of learning, Armitage [/bib_ref]. Managing complex social-ecological systems largely depends on Closeness centrality of the Galapagos small-scale fishery governance network. Nodes indicate the organizations and agencies (GO governmental organization, PO private organization, FA fishery association, NGO non-governmental organization, MG municipal government, PG parish government, ARO academic and research organization). Ties indicate the organizational links between organizations and agencies. Green nodes indicate those nodes connected to the Galapagos small-scale fishery network. Pink nodes indicate nodes that traditionally do not possess significant organizational links in the Galapagos small-scale fishery governance system network. Node size indicates closeness centrality, which signifies that as a node's size increases, it is deemed important based on its closeness to all nodes in the network. This makes more central nodes significant players in the network for dispersing knowledge or information faster than others, due to their closeness to all nodes in the Galapagos small-scale fishery governance system network moving from scattered and individual learning processes to collective learning, transitioning from single-loop learning to double-loop and triple-loop learning. The following quotes from interviewees are significant in that regard:
We are an educational entity; our collaboration would be clearly linked to education. We have previously linked the children of fishers in educational programs such as the Sea Turtle Monitoring Program. NGO08 node of Figs. 7-9, Level: International, Sector: Private.
As has been done in previous years, our collaboration would be oriented to training and workshops for the socio-organizational consolidation of the fishing cooperatives and the organization and strengthening their legal scope. NGO15 node of Figs. 7-9, Level: International, Sector: Private.
With the above in mind, a crucial development in socialecological systems lies in the question of who is learning and from whom [bib_ref] Adaptive comanagement and the paradox of learning, Armitage [/bib_ref]. It is important to recognize that the scientific community and rigid governance structures have often viewed scientific production as the only way of solving problems. However, learning at the local scale is crucial to addressing uncertainty and the changing local conditions that generate vulnerability. Local actors possess particular knowledge and experience acquired over the years, which if it is aligned to the right actors, might potentially strengthen the Galapagos smallscale fishery collaborative network. The following quotes from interviewees are significant in that regard:
Marketing in conjunction with the fishing sector as part of a macro project to collect food products that involve the rural sector. We could contribute with local knowledge acquired overtime to motivate youth to get involved in the fishing sector. PG04 node of Figs. 7-9, Level: Local, Sector: Public.
They could count on our group of local volunteers to be part of the participatory processes. P006 node of Figs. 7-9, Level: Local, Sector: Private.
# Discussion
Our research suggests that understanding the structures of governance systems is a significant contributor to creating synergies among stakeholders to achieve collective outcomes that lead to more robust social-ecological systems in light of multiple adverse drivers of change. Governance systems often represent the different structures by which societies shape collective actions [bib_ref] Water governance: some critical issues, Tortajada [/bib_ref] [bib_ref] Governance principles for natural resource management, Lockwood [/bib_ref]. Bearing this in mind, our research indicates that addressing the extent of the effects of unprecedented and simultaneous drivers of change (such as climate change, novel pandemics, illegal marine fishing, invasive species, among other wicked problems) demands a deeper understanding by those involved in governance systems [bib_ref] Advancing coral reef governance into the Anthropocene, Morrison [/bib_ref]. These must have a clear grasp of the governance actors, with their interactions and network configurations between different sectors, geographical scales and administrative [bib_ref] Collaborative governance for climate change adaptation in Canada: experimenting with adaptive co-management, Baird [/bib_ref] [bib_ref] An institutional analysis of the Kaipara Harbour Governance Network in New Zealand, Kanwar [/bib_ref] [bib_ref] Identifying governance gaps among interlinked sustainability challenges, Bergsten [/bib_ref]. In this context, we argue that actors within the Galapagos small-scale fishing governance system network may create strategic alliances to deal with external and internal drivers of change and enhance the governance system fit. This will be possible if they explore further organizational ties and network configurations across sectors and geographical scales, and keep track of the organizations' positions and features in the existing small-scale fishing governance network. Approximating as closely as possible the governance scale of the Galapagos small-scale fishing sector with the extension of the multiple socialecological interactions in the Galapagos (fit) by including a few delegated organizations and organizational links designated by law is challenging, if not impossible to achieve [bib_ref] Collaborative environmental governance: achieving collective action in social-ecological systems, Bodin [/bib_ref] [bib_ref] Closing integrative gaps in complex environmental governance systems, Fried [/bib_ref]. Managing and controlling wicked problems spanning the Galapagos smallscale sector, such as climate change or the introduction of rapid mitigation measures to address novel pandemics, requires the collective effort of diverse organizations and agencies beyond state and national boundaries.
Our results show that understanding certain degrees of network distribution can provide valuable information for strengthening the Galapagos small-scale fishery collaborative network. It could provide additional pathways for the diffusion of determinants of adaptive capacity, along with better coordination and collaboration among actors within the fishing governance network. Our descriptive statistics suggest that various organizations and agencies occupy important positions within this network. Our centrality analysis indicates that certain organizations and agencies send and receive more organizational links than others in the network and [fig_ref] Table 1: Descriptive statistics of the Galapagos small-scale fishery sector [/fig_ref]. We deem it important to unveil these nodes in the governance network considering that these organizations and agencies probably control decisions in this network. Therefore, if we aim to incorporate new collaboration links into the existing network, it is necessary to recognize the organizations and agencies possessing the authority and power to make changes to approximate the management of governance systems with socio-ecological interactions and operationalize transitions to adaptive co-management forms of governance.
Our results also point to various organizations and agencies have higher eigenvector centrality values than others in the Galapagos small-scale fishing governance system network and [fig_ref] Table 1: Descriptive statistics of the Galapagos small-scale fishery sector [/fig_ref]. We consider this an important feature to recognize if we aim at aligning new actors with diverse technological, behavioral, financial, institutional, and informational resources, among other determinants of adaptive capacity, with said network. We argue that these organizations and agencies are influential and well-positioned, not so much for the number of organizational links that they send and receive, but because of their connections to organizations and agencies with higher centrality values than others in the Galapagos small-scale fishing governance system. This means that these organizations and agencies may serve as channels of communication to reach other organizations and agencies often in charge of the decision-making structures of the governance network, facilitating the creation of links between external stakeholders and decision-making actors. We claim that this access might lead to governance arrangements and the formation of new organizational links that facilitate the connection between local priorities and international, regional and national levels of management.
Our outcomes also indicate that diverse organizations and agencies within the Galapagos small-scale fishing governance system network have higher closeness values than others and [fig_ref] Table 1: Descriptive statistics of the Galapagos small-scale fishery sector [/fig_ref]. We argue that this is a good sign for collaboration and the diffusion and incorporation of adaptive capacity determinants into the network, as these organizations and agencies are closer to any others in the network. From a governance perspective, reaching all other actors more rapidly implies that the incorporation and diffusion of ideas, financial resources and technical solutions might occur more quickly and more efficiently in the network. This is significant considering that approximating a governance fit partially depends on the capacity of governance systems to act in time [bib_ref] Institutional fit and the sustainability of social-ecological systems, Epstein [/bib_ref]. Recognizing that the capacity of governance systems to achieve such a fit has been gradually reduced due to the growing human and ecological interactions spanning governance systems is needed in managing common-pool resources like the Galapagos small-scale fisheries. Recently this has been evidenced more explicitly as governance systems have been struggling with measures and strategies to limit the spread of the COVID-19 virus and cope with the associated socioeconomic and public health fallout. Therefore, evaluating organizations and agencies closer to all in the network might signify acting faster in crisis and delivering rapid responses in the Galapagos small-scale fishing governance system network.
Although we found no strong evidence of mutual interaction between organizations and agencies from the private sector (hypothesis 2), our ERGM outcomes suggest a propensity toward a cross-sectoral interaction network among various organizations and agencies in the Galapagos smallscale fishery system. We found evidence of this propensity toward mutual interaction among nodes of the public sector (hypothesis 1), and a significant, positive propensity of nodes from the private and public sectors to form organizational links in the Galapagos small-scale fishing governance system network (hypothesis 3). The latter, from our perspective, may be seen as a significant feature of analysis, bearing in mind the need for cross-sectoral interactions to deliver adequate policy-making solutions in the sector. We further noted positive and significant effects toward crosssectoral open triadic network configurations (hypotheses 4, 5, and 6). We argue that the prevalence of these configurations in the network can be interpreted as a good sign for the evolution of cross-sectoral collaboration relationships within the Galapagos small-scale fishing governance system network. It is likely that the prevalence of a reciprocal relationship (A ↔ B) might further be developed into either an open triadic or a closed triadic network configuration.
The propensity toward open triadic configurations might potentially lead to closed triad configurations if organizations and agencies deem that the participation of a third party (C) could contribute to the achievement of common institutional goals and collective actions for governing a shared natural resource, through more densely clustered relations of collaboration.
According to our qualitative data analysis, diverse organizations, and agencies with no strong presence in the Galapagos small-scale fishing governance system may collaborate within the network through various means. This includes contributions in the form of technical assistance, equipment, information and training capabilities, and local fisheries knowledge. The willingness to collaborate is speculative, and our observations in this regard ignore the role of power and the level of trust required among organizations and agencies to cement collaborative partnerships. However, we argue that these results demonstrate that additional alliances and collaboration may emerge in the network, transforming it into a more densely clustered collaboration network. Well-positioned organizations and agencies in the network, such as CGREG, DPNG, fishing cooperatives, and CDF, can play an important role in creating a more collaborative network because they possess ties to other governmental organizations, NGOs, funding, academic and research institutions, and local resource users.
## Limitations and future directions
Our study coincided with the Coronavirus disease 2019 (COVID-19) pandemic, restricting nearly all in-person interactions; as a result, reaching organizations' representatives and officials to be included in the study was a challenging endeavor. Therefore, in the future there remains room for this paper's outcomes to be expanded in scope. This can be accomplished by integrating other organizations and agencies at diverse geographical scales into our analysis, as well as administrative levels and organizational links that this study may have missed. Furthermore, this paper may serve as a guide for future theoretical frameworks geared toward exploring further network configurations of the Galapagos small-scale fishery governance system. For example, there is clearly a need to examine the propensity toward triadic network configurations (i.e., interactions and links between the three nodes A, B and C) and investigate further hypotheses considering actors' attributes (e.g., hypotheses regarding trust between nodes, a central feature that drives stakeholders to engage in collaboration and choose collaboration partners) [bib_ref] Trust, confidence, and equity affect the legitimacy of natural resource governance, Turner [/bib_ref] [bib_ref] Collective action in a polycentric water governance system, Baldwin [/bib_ref] [bib_ref] The impacts of trust, cost and risk on collaboration in environmental governance, Bodin [/bib_ref]. Further, while we deem polycentric governance arrangements attractive to create and deliver solutions to the various socio-ecological problems affecting the Galapagos small-scale fishing sector, we also recognize that understanding the manifestations of power and its influences is critical to fostering collaboration among multiple actors within the Galapagos small-scale fishing governance system. Conflicts usually emerge in polycentric governance arrangements because of conflict of interest and resource access inequality, increasing polarization among stakeholders and obstacles to forming collaborative partnerships between higher and lower administrative levels (Mudaliar 2020; [bib_ref] Crowding-out lower-level authorities: Interactions and transformations of higher and lower-level authorities in..., Mudliar [/bib_ref]. Therefore, we suggest that future investigations evaluate the role of power dynamics in the governance of the Galapagos small-scale fishery system, which is an aspect that our research does not address. By no means the inclusion of multiple organizations and agencies across various administrative levels and geographical scales will be sufficient to enhance collaboration and functionality within the Galapagos small-scale fishing governance system [bib_ref] Breaking bad: when does polycentricity lead to maladaptation rather than adaptation?, Biddle [/bib_ref]. Additional research efforts are needed to unveil the power dimensions of the Galapagos small-scale fishing governance system. The transition to a new Galapagos governance system regime, which is currently being amended, will most likely redistribute responsibilities and decision-making power, potentially leading to recentralization pathways and monocentric governance arrangements. Thus, we suggest exploring the power dynamics of the Galapagos small-scale fishing governance system based on the typology of power proposed by [bib_ref] The black box of power in polycentric environmental governance, Morrison [/bib_ref]. These authors define three dimensions of power: power by design, pragmatic power and framing power. Based on this research approach, it will be possible to elucidate the concentration of power within the Galapagos small-scale fishing governance system network. Such knowledge is fundamental for improving the collaborative ties in the Galapagos small-scale fishing governance system, marking a critical step in addressing the complex socio-ecological problems that hinder the sustainable development of the Galapagos small-scale fisheries.
# Conclusions
Since Elinor Ostrom's publications, there has been a significant rise of scientific interest in polycentrism in the literature on complex social-ecological systems. However, to our knowledge, the number of studies in the Galapagos Islands aimed at improving marine resource management of complex social-ecological systems, considering social network approaches and polycentric governance arrangements, is still limited. Addressing simultaneous wicked problems, such as public health, socio-economic, environmental, institutional and climate issues, requires a multi-level approach across different scales. This study, therefore, proposes that the Galapagos small-scale fishing governance system should explore more polycentric approaches to governance, including linkages (partnerships) spanning multiple scales and levels, from global to local, relying on formal and informal networks. More polycentric ties in the sector might contribute to creating the correct links at the right time in light of multiple drivers of change [bib_ref] Enhancing the fit through adaptive co-management: creating and maintaining bridging functions for..., Olsson [/bib_ref] [bib_ref] Polycentric systems of governance: a theoretical model for the commons, Carlisle [/bib_ref]. Complex social-ecological systems, like the Galapagos small-scale fishing sector, need to embrace a socialecological perspective involving different sorts of information, skills, and stakeholders, at different scales and levels. This would enable the sector to approximate as closely as possible the governance scope required to handle the multiple social-ecological dynamics in the archipelago and prevent a misfit. By no means are we suggesting that the state should cede control over marine resources in the Galapagos. We do, however, consider that the multiple social-ecological interactions that comprise the sector require the cooperation and collaboration of multi-scale and multi-level organizations to deal with the multiple drivers of change, particularly in these current times of constant change und uncertainty. Without question, the adverse effects of the COVID-19 pandemic on the social-economic situation of the Galapagos population, together with the difficulties controlling illegal international fishing within the Galapagos Marine Reserve protected area, highlight the need to create an adaptive capacity based on a polycentric governance network. Systems with high adaptive capacity are those most capable of reconfiguring themselves when subjected to shocks [bib_ref] Adaptive governance of social-ecological systems, Folke [/bib_ref]. Therefore, this paper might guide practitioners and decision-makers to explore further organizational links and network configurations, allowing for the development of collaboration strategies to cope with the various multidimensional problems faced by the Galapagos small-scale fishing system.
We contend that the gauging of nodes' positions, features, and needs can enable actors within governance systems to better discern among collaborative partnerships from which to choose, rather than relying on chance or policies and laws to define collaboration ties. In our view, this argument contributes to the discussion analyzing polycentric arrangements by implying that, rather than being arbitrarily forced to adjust to polycentric structures, actors can do so voluntarily because it helps them to consolidate strategic alliances considering mutual goals and concerns. Notably, we argue that the insights presented in this study contribute to elucidating the notion of institutional fit, initially explored by. It is significant to consolidate the idea that the concept of fit in common-pool resources depends on governance systems' ability to fit in with environmental and ecological concerns, but also on their ability to fit in with various global sustainability challenges and stakeholder expectations [bib_ref] Does polycentricity fit? Linking social fit with polycentric governance in a large-scale..., Acton [/bib_ref] [bib_ref] Achieving multiple socioecological institutional fits: the case of spiny lobster comanagement in..., Ishihara [/bib_ref]. Finally, we see our research as a timely study that might open discussions in the ongoing reformulation of the GSL-bearing in mind that the distribution of functions and power in the Galapagos Islands centers around the guidelines and policy decisions established under the GSL. COVID-19 is a new driver of change in the Galapagos that has led to the archipelago's worst-ever socio-economic scenario and the need to explore new ways to address various issues beyond environmental and ecological concerns. In this context, we consider the insights presented in our study to have usefully introduced governance-related insights hardly explored among the related public and political discussions in the Galapagos.
[fig] Figure 1: Governance fitFig. 2 a Reciprocity, b Open triads, c Closed triads [/fig]
[fig] Figure 4 a: Monocentric forms of governance, b polycentric forms of governance [/fig]
[fig] Figure 5: Location map [/fig]
[table] Table 1: Descriptive statistics of the Galapagos small-scale fishery sector [/table]
|
Structural and functional neural adaptations in obstructive sleep apnea: An activation likelihood estimation meta-analysis
a b s t r a c tObstructive sleep apnea (OSA) is a common multisystem chronic disorder. Functional and structural neuroimaging has been widely applied in patients with OSA, but these studies have often yielded diverse results. The present quantitative meta-analysis aims to identify consistent patterns of abnormal activation and grey matter loss in OSA across studies. We used PubMed to retrieve task/resting-state functional magnetic resonance imaging and voxel-based morphometry studies. Stereotactic data were extracted from fifteen studies, and subsequently tested for convergence using activation likelihood estimation. We found convergent evidence for structural atrophy and functional disturbances in the right basolateral amygdala/hippocampus and the right central insula. Functional characterization of these regions using the BrainMap database suggested associated dysfunction of emotional, sensory, and limbic processes. Assessment of task-based co-activation patterns furthermore indicated that the two regions obtained from the meta-analysis are part of a joint network comprising the anterior insula, posterior-medial frontal cortex and thalamus. Taken together, our findings highlight the role of right amygdala, hippocampus and insula in the abnormal emotional and sensory processing in OSA.
# Introduction
Obstructive sleep apnea (OSA) is a chronic disorder that arises from recurrent partial or complete pharyngeal obstruction during sleep [bib_ref] Prevalence of symptoms and risk of sleep apnea in the US population:..., Hiestand [/bib_ref] [bib_ref] Adult obstructive sleep apnoea, Jordan [/bib_ref] [bib_ref] Obstructive sleep apnea: diagnosis, epidemiology, and economics, Kapur [/bib_ref] [bib_ref] Sleep in primary care international study, G. Prevalence of symptoms and risk..., Netzer [/bib_ref] [bib_ref] Sleep apnoea and the brain: a complex relationship, Rosenzweig [/bib_ref]. In patients with OSA, this leads to nocturnal apneas and hypopneas, intermittent hypoxia, reoxygenation and hyper-/hypocapnia events, along with sleep fragmentation, and changes in cerebral blood flow [bib_ref] Effect of continuous positive airway pressure on regional cerebral blood flow during..., Shiota [/bib_ref] [bib_ref] Regional cerebral blood flow alterations in obstructive sleep apnea, Yadav [/bib_ref]. The prevalence of OSA is noticeable in general population and around 50% in patients with cardiovascular or metabolic disorders [bib_ref] Prevalence of symptoms and risk of obstructive sleep apnea syndrome in the..., Khazaie [/bib_ref] [bib_ref] Obstructive sleep apnoea syndrome, Lévy [/bib_ref] [bib_ref] Obstructive sleep apnea in adults: epidemiology, clinical presentation, and treatment options, Lurie [/bib_ref].
Several recent studies highlighted that OSA contributes to emotional and cognitive decline, and it is increasingly considered as one of the rare modifiable risk factors for neurodegenerative dementia [bib_ref] Alzheimer's disease neuroimaging, I. Sleep-disordered breathing advances cognitive decline in the elderly, Osorio [/bib_ref] [bib_ref] Sleep apnoea and the brain: a complex relationship, Rosenzweig [/bib_ref] [bib_ref] Connections between sleep and cognition in older adults, Yaffe [/bib_ref] [bib_ref] Sleep-disordered breathing, hypoxia, and risk of mild cognitive impairment and dementia in..., Yaffe [/bib_ref]. If untreated, OSA can result in varying degrees of cognitive deficits such as difficulties with attention, memory, executive functioning, and quality of life [bib_ref] Sleep apnoea and the brain: a complex relationship, Rosenzweig [/bib_ref]. In addition, excessive daytime sleepiness, labile interpersonal relationships, and decreased work and school efficiency have all been documented in OSA patients [bib_ref] Sleep apnoea and the brain: a complex relationship, Rosenzweig [/bib_ref] [bib_ref] Obstructive sleep apnea syndrome is associated with deficits in verbal but not..., Twigg [/bib_ref]. Moreover, it is recognized that OSA patients are two to thirteen times more likely to experience a driving-related traffic accident [bib_ref] Systematic review of motor vehicle crash risk in persons with sleep apnea, Ellen [/bib_ref] [bib_ref] Sleep apnea related risk of motor vehicle accidents is reduced by continuous..., Karimi [/bib_ref] [bib_ref] Attention deficits detected in cognitive tests differentiate between sleep apnea patients with..., Karimi [/bib_ref] [bib_ref] Obstructive sleep apnea syndrome; a neglected cause of traffic collision among Iranian..., Khazaie [/bib_ref]. Such accidents are more likely to occur in those who manifest greater daytime sleepiness, but are not necessarily related to sleepiness alone [bib_ref] Systematic review of motor vehicle crash risk in persons with sleep apnea, Ellen [/bib_ref]. In addition to cognitive and emotional deficits, increased prevalence of OSA in several psychiatric disorders has been reported, of which major depressive disorder (MDD), anxiety, and posttraumatic stress disorder (PTSD) appear best documented [bib_ref] Obstructive sleep apnea and psychiatric disorders: a systematic review, Gupta [/bib_ref] [bib_ref] Association of psychiatric disorders and sleep apnea in a large cohort, Sforza [/bib_ref].
It has been suggested that adaptive and maladaptive processes both occur in patients with OSA in response to hypoxemia [bib_ref] Sleep apnoea and the brain: a complex relationship, Rosenzweig [/bib_ref]. The fine balance of these processes, and its eventual impact on neurocognitive and emotional performance, will depend on the stage of this dynamic process, effects on other organ systems, cognitive reserve, and idiosyncratic susceptibility [bib_ref] Oxidative stress in obstructive sleep apnea and intermittent hypoxia-revisited-the bad ugly and..., Lavie [/bib_ref] [bib_ref] Association of psychiatric disorders and sleep apnea in a large cohort, Sforza [/bib_ref]. Although these deficits are not always reversed with treatment [bib_ref] Continuous positive airway pressure devices for the treatment of obstructive sleep apnoea-hypopnoea..., Mcdaid [/bib_ref] , a meta-analysis and a meta-review [bib_ref] Neurocognitive function in obstructive sleep apnoea: a meta-review, Bucks [/bib_ref] suggest beneficial effects of treatment (e.g. continuous positive airway pressure (CPAP)) on cognitive performance, sleepiness and neural injury in patients with OSA.
Over the last three decades, numerous structural and functional neuroimaging studies, including voxel-based morphometry (VBM), task functional magnetic resonance imaging (fMRI), and resting-state fMRI (rs-fMRI) have been conducted on patients with OSA. Structural and functional MRI imaging studies, however, often point to diverse results in OSA [bib_ref] Desperately seeking grey matter volume changes in sleep apnea: a methodological review..., Celle [/bib_ref] [bib_ref] The brain in sleep-disordered breathing: a vote for the chicken?, Morrell [/bib_ref]. The variability of the findings has been suggested to be due to relatively small sample sizes, with heterogeneous patient groups that differed in several key respects (e.g. diagnostic criteria, IQ, age, gender, and the imaging acquisition, preprocessing, and analysis methods); for more detailed discussion, please refer to [bib_ref] CrossTalk proposal: the intermittent hypoxia attending severe obstructive sleep apnoea does lead..., Gozal [/bib_ref] [bib_ref] Is brain injury in obstructive sleep apnea reversible?, Macey [/bib_ref] [bib_ref] The brain in sleep-disordered breathing: a vote for the chicken?, Morrell [/bib_ref]. To date, OSA structural studies have used a spatially unbiased analytical approach, such as commonly used mass-univariate approaches that rely on conservative statistical thresholds mandated by the large number of voxels compared between-groups [bib_ref] Voxel-based morphometry-the methods, Ashburner [/bib_ref]. On the other hand, task fMRI studies used a variety of paradigms to study functional disturbances in particular disease. Recently, the activation likelihood estimation (ALE) method has been proposed as a useful methodology that, using coordinatebased meta-analyses (CBMA), provides a powerful tool to attain a synoptic view of distributed neuroimaging findings and different neuroimaging methods (e.g. structural and functional) in an objective and quantitative fashion [bib_ref] Coordinate-based activation likelihood estimation meta-analysis of neuroimaging data: a random-effects approach based..., Eickhoff [/bib_ref] [bib_ref] Meta-analysis of the functional neuroanatomy of single-word reading: method and validation, Turkeltaub [/bib_ref]. More specifically, CBMA method searches for "where" in the brain the amount of convergence between reported foci is more than expected by chance, which yields to statistical inference on the integration of previous findings [bib_ref] Activation likelihood estimation meta-analysis revisited, Eickhoff [/bib_ref] [bib_ref] Investigating the functional heterogeneity of the default mode network using coordinate-based meta-analytic..., Laird [/bib_ref] [bib_ref] Meta-analysis of the functional neuroanatomy of single-word reading: method and validation, Turkeltaub [/bib_ref].
Only structural OSA studies were hitherto analysed using the ALE method [bib_ref] Mapping gray matter reductions in obstructive sleep apnea: an activation likelihood estimation..., Weng [/bib_ref] , and in order to fully address some of the previously raised concerns in the field about the diversity of these findings, we undertook the ALE meta-analysis of both functional and structural abnormalities recorded in patients with OSA. Our aim was to elucidate converging findings and to emphasize important brain nodes as highlighted via different neuroimaging modalities. We then functionally characterized the obtained regions that showed neurobiological aberrations in OSA patients by means of the BrainMap database, and also assessed their brain wide co-activation patterns to reveal networks that are (conjointly) connected to these obtained areas.
# Methods
## Search strategies and study selection
In accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement (Moher Obstructive"[Mesh]) OR sleep apnea AND (("voxel-based morphometry") OR "VBM") that resulted in 21 studies [fig_ref] Figure 1: Paper selection strategy flow chart [/fig_ref]. No positron emission tomography (PET) studies met our criteria. Of note is that here; "study" reflects an individual scientific paper and "experiment" represents a single analysis or contrast of interest in a given study yielding localization information (i.e. OSA>controls; OSA<controls).
Only peer-reviewed cross-sectional studies that were published in English language and that compared groups of human adult OSA patients (above 18 years old) to healthy controls were included. The exclusion criteria were as follows:
- Case-reports, letter to editors, meta-analysis or review studies reporting no original data. - Studies that did not report whole-brain analysis.
- Studies which did not report standard space coordinates.
- Studies that reported coordinates only in sub-sample.
- Intervention studies (pre/post treatment contrasts such as CPAP).
- Studies without a "control group" i.e. those focused only on a group of OSA patients. - Studies where less than 7 patients were included in each group.
## Data extraction
Two investigators independently extracted the information (M.T and A.A.S). Recorded data included the first author's name, year of publication, age, gender and number of patients and controls, the imaging modality (resting-state/task fMRI or VBM) and type of task in the task fMRI studies. Moreover, we recorded the peak coordinates (x,y,z) in Talairach [bib_ref] Co-planar Stereotaxic Atlas of the Human Brain: 3-dimensional Proportional System: An Approach..., Talairach [/bib_ref] or Montreal Neurological Institute (MNI) [bib_ref] 3D statistical neuroanatomical models from 305 MRI volumes, Evans [/bib_ref] stereotactic space from all experiments and transformed all data into MNI coordinates for analysis [bib_ref] Bias between MNI and talairach coordinates analyzed using the ICBM-152 brain template, Lancaster [/bib_ref]. We used the extracted stereotactic coordinates to conduct the ALE metaanalysis.
If a publication did not report the coordinates of activation maxima, we contacted the authors. Two studies were excluded [bib_ref] Altered brain activation during response inhibition in obstructive sleep apnea, Ayalon [/bib_ref] [bib_ref] Altered intrinsic regional brain activity in male patients with severe obstructive sleep..., Peng [/bib_ref] because the authors published two papers based on the same samples i.e. [bib_ref] Relationship between obstructive sleep apnea severity and brain activation during a sustained..., Ayalon [/bib_ref] [bib_ref] Altered brain activation during response inhibition in obstructive sleep apnea, Ayalon [/bib_ref] [bib_ref] Aberrant spontaneous low-frequency brain activity in male patients with severe obstructive sleep..., Li [/bib_ref] [bib_ref] Altered intrinsic regional brain activity in male patients with severe obstructive sleep..., Peng [/bib_ref]. It is worthy to note that one study performed both VBM and rs-fMRI measurements [bib_ref] Altered resting-state brain activity in obstructive sleep apnea, Zhang [/bib_ref] and another one applied both VBM and task fMRI [bib_ref] Functional and structural changes in the brain associated with the increase in..., Fatouleh [/bib_ref] [fig_ref] Figure 1: Paper selection strategy flow chart [/fig_ref].
## Activation likelihood estimation
The statistical analyses were performed using the revised version of the activation likelihood estimation (ALE) [bib_ref] Activation likelihood estimation meta-analysis revisited, Eickhoff [/bib_ref] based on coordinate-based meta-analyses (CBMA) [bib_ref] Coordinate-based activation likelihood estimation meta-analysis of neuroimaging data: a random-effects approach based..., Eickhoff [/bib_ref] [bib_ref] Investigating the functional heterogeneity of the default mode network using coordinate-based meta-analytic..., Laird [/bib_ref] [bib_ref] Meta-analysis of the functional neuroanatomy of single-word reading: method and validation, Turkeltaub [/bib_ref]. ALE assessed the significant convergence between activation foci from different experiments (e.g. OSA>controls, OSA<controls) for a given study in comparison with a random distribution of foci. More specifically, as the first step, ALE algorithm models the reported foci as center peaks of 3D Gaussian probability distributions that acknowledge the spatial uncertainty associated with each focus. The uncertainty is mainly due to between-subject variations (neuroanatomical variability and small sample sizes) and between-laboratory differences (various brain templates, normalization, and analysis strategies). The number of participants per experiment determines the width of the spatial uncertainty of any focus [bib_ref] Activation likelihood estimation meta-analysis revisited, Eickhoff [/bib_ref]. As the second step, the probability distributions of all activation foci in a particular experiment are combined for each voxel, which creates a modeled activation map (MA map). Thus, these MA maps summarize localization probabilities of studies and the final ALE map results from interpolation of these MA maps describing the convergence of results across all experiments. During the third step, an analytical approach based on a non-linear histogram integration is applied to test against the null hypothesis of randomly distributed foci and subsequently significant statistical threshold set at p < 0.05 family-wise error in cluster level . The recent analysis approach tested for convergence by experiments (random effects) rather than foci (fixed effects) [bib_ref] Activation likelihood estimation meta-analysis revisited, Eickhoff [/bib_ref] ; for further details and a summary of the ALE method please refer to .
## Functional decoding using the brainmap database
To assess the functional roles of the abnormal brain regions in OSA, behavioral decoding using the BrainMap database was consequently performed. More specifically, we tested which types of tasks were more likely than chance to activate for each of the meta-analysis common regional gray matter loss and functional abnormalities (or seeds) [bib_ref] Differentiated parietal connectivity of frontal regions for what and where memory, Rottschy [/bib_ref]. The behavioral domain and paradigm class meta-data categories from the BrainMap database were used for functional characterization of the clusters. At the time of analysis, the database included coordinates of reported activation foci and associated meta-data of more than 10,000 neuroimaging experiments [bib_ref] The cognitive paradigm ontology: design and application, Turner [/bib_ref]. Behavioral domains of the BrainMap database consist of several main categories including cognition, action, perception, emotion, and interoception, and their related sub-categories, which define the neural processes isolated by the respective contrast. However, paradigm classes specify the particular applied task [bib_ref] BrainMap taxonomy of experimental design: description and evaluation, Fox [/bib_ref] (see http:// www.brainmap.org/scribe). For the functional characterization of common regional atrophy and functional abnormalities, all experiments in the BrainMap database that featured at least one focus of activation within the seed regions were determined based on reported activation coordinates. Subsequently, for each behavioral domain and paradigm class category, we identified regional functional profile by discerning taxonomic labels for which the probability of finding activation in the respective cluster is significantly higher than the overall chance across the entire BrainMap database. Significant level was identified as p < 0.05 using a binomial test [bib_ref] Dysregulated left inferior parietal activity in schizophrenia and depression: functional connectivity and..., Muller [/bib_ref] [bib_ref] Differentiated parietal connectivity of frontal regions for what and where memory, Rottschy [/bib_ref].
## Whole-brain co-activation profiles
In order to map brain regions that feature significant coactivation with the regions identified in the OSA structural and functional meta-analysis, we performed meta-analytic coactivation modeling . More specifically, we tested how likely it is that the experiments activating the particular region also activate other brain voxels above chance [bib_ref] Metaanalytic connectivity modeling: delineating the functional connectivity of the human amygdala, Robinson [/bib_ref]. In order to perform MACM, first we identified all experiments in the BrainMap database that activate the convergent seeds. Then, quantitative meta-analysis was applied to test for convergence across the foci reported in the experiments. Inevitably, the highest convergence will be observed in the seed regions because experiments are already selected by activation in those seeds. Significant convergence of reported foci in other brain areas represents consistent co-activation or functional connectivity of other voxels with the seeds. More specifically, MACM provides information on the functional interactions of cortical modules based on their whole-brain co-activation pattern across the BrainMap database [bib_ref] Networks of task co-activations, Laird [/bib_ref].
# Results
Fifteen publications that recruited 290 OSA patients and 290 healthy controls were included [bib_ref] Relationship between obstructive sleep apnea severity and brain activation during a sustained..., Ayalon [/bib_ref] [bib_ref] Increased brain activation during verbal learning in obstructive sleep apnea, Ayalon [/bib_ref] [bib_ref] Functional and structural changes in the brain associated with the increase in..., Fatouleh [/bib_ref] [bib_ref] FMRI responses to cold pressor challenges in control and obstructive sleep apnea..., Harper [/bib_ref] [bib_ref] Neural responses during Valsalva maneuvers in obstructive sleep apnea syndrome, Henderson [/bib_ref] [bib_ref] Reduced brain gray matter concentration in patients with obstructive sleep apnea syndrome, Joo [/bib_ref] [bib_ref] Aberrant spontaneous low-frequency brain activity in male patients with severe obstructive sleep..., Li [/bib_ref] [bib_ref] Inspiratory loading elicits aberrant fMRI signal changes in obstructive sleep apnea, Macey [/bib_ref] [bib_ref] Functional magnetic resonance imaging responses to expiratory loading in obstructive sleep apnea, Macey [/bib_ref] [bib_ref] Changes in brain morphology in patients with obstructive sleep apnoea, Morrell [/bib_ref] [bib_ref] Task positive and default mode networks during a parametric working memory task..., Prilipko [/bib_ref] [bib_ref] Altered cortical and subcortical local coherence in obstructive sleep apnea: a functional..., Santarnecchi [/bib_ref] [bib_ref] Cognitive profile and brain morphological changes in obstructive sleep apnea, Torelli [/bib_ref] [bib_ref] A combined neuropsychological and brain imaging study of obstructive sleep apnea, Yaouhi [/bib_ref] [bib_ref] Altered resting-state brain activity in obstructive sleep apnea, Zhang [/bib_ref] [fig_ref] 1: Introduction [/fig_ref] , [fig_ref] Figure 1: Paper selection strategy flow chart [/fig_ref]. These publications collectively reported results from 30 experiments; of which 17 experiments were reported as the "OSA<controls" contrasts and 13 experiments for the "OSA>controls" contrasts.
## Convergence of neuroimaging findings in osa
Testing for significant convergence across all eligible neuroimaging experiments comparing subjects with OSA to healthy controls yielded two significant clusters, one located in the right amygdala/hippocampus, the other in the right central insula (p < 0.05 FWE corrected in cluster level) [fig_ref] Figure 2: Fig [/fig_ref].
The cluster of convergence in the right amygdala/hippocampus was driven by a 62.5% contribution from task-fMRI studies (cognitive stimulation paradigms contributed 35% and sensory stimulation paradigms contributed 27.5%). Further 37.5% of contribution was driven by VBM studies. No contribution from resting-state fMRI was noted. The identified shared cluster of reduced grey matter volume and functional hypo-activation in OSA patients (local maximum: 22/-10/-22 in MNI space) was then anatomically allocated to the internal subdivision of the human right amygdala and hippocampus, as defined by histological and functional [bib_ref] Definition and characterization of an extended social-affective default network, Amft [/bib_ref] [bib_ref] An investigation of the structural, connectional, and functional subspecialization in the human..., Bzdok [/bib_ref] [bib_ref] Metaanalytic connectivity modeling: delineating the functional connectivity of the human amygdala, Robinson [/bib_ref] criteria using the SPM Anatomy Toolbox [bib_ref] A new SPM toolbox for combining probabilistic cytoarchitectonic maps and functional imaging..., Eickhoff [/bib_ref]. Majority of the cluster volume (31%) appeared assigned to the basolateral nucleus of the amygdala, with smaller portions extending into the CA1 (14%), CA3 (13%), and subiculum (9%) [fig_ref] Figure 2: Fig [/fig_ref]. A second cluster of significant convergence was located in the right central insula (local maximum: 40/2/-4 in MNI space) [bib_ref] Meta-analytic clustering of the insular cortex characterizing the meta-analytic connectivity of the..., Cauda [/bib_ref] [bib_ref] A link between the systems: functional differentiation and integration within the human..., Kurth [/bib_ref]. This result was almost completely driven by the hypo-activations in task-fMRI studies (99.9%). Among those, cognitive stimulation paradigms contributed 55.4% and sensory stimulation paradigms contributed 44.5% to the right central insula [fig_ref] Figure 2: Fig [/fig_ref].
Supplementary analyses targeting convergence among specific aspects confirmed these key findings. We then tested for significant convergence across the functional MRI findings, pooling across task-and resting-state fMRI analyses in order to provide a global assessment of aberrant functional patterns in OSA patients. Accounting for multiple comparisons across the entire brain, a significant convergence was again identified in the right basolateral amygdala/hippocampus and right central insula. Subsequently, we sub-analysed only task-based fMRI studies, identifying the same regions as significant. Due to the low number of available experiments, resting-state functional imaging findings could not be reliable assessed in isolation at this point. Convergence of structural and functional difference in obstructive sleep apnea compared with healthy controls. Location of the significant convergence of gray matter reduction and functional disturbance in the right basolateral nucleus of the human amygdala/hippocampus (A) and in the right central insula (B). Results are from the Activation Likelihood Estimation for sleep apnea meta-analyses. All activations are significant at P < 0.05 corrected for multiple comparisons using the family-wise error rate in cluster level (cFWE).
In summary, the performed series of quantitative meta-analyses on structural and functional neuroimaging findings in OSA patients revealed consistent evidence for primarily structural atrophy and functional disturbances (task-related hypo-activations) in the right amygdala/hippocampus and right central insula.
## Functional decoding using the brainmap database
In order to assess the functional roles of these two brain regions, which feature the most consistent evidence for OSA-related aberrations, we performed behavioral decoding using the BrainMap database. More precisely, we tested which types of tasks were more likely than by chance to activate the right basolateral amygdala/hippocampus and right central insula regions identified in the main ALE analysis. We found a significant (p < 0.05, corrected for multiple comparisons) association of the right basolateral amygdala/hippocampus region with affective and emotional processing, perception/interoception, memory-related processes, somatosensory functions [fig_ref] Figure 3: Behavioral characterization of the significant cluster in the right amygdala/hippocampus [/fig_ref]. The cluster located on the right central insular cortex was reported as associated with somatosensory processing, and in particular with pain processing (p < 0.05 corrected for multiple comparisons) [fig_ref] Figure 3: Behavioral characterization of the significant cluster in the right amygdala/hippocampus [/fig_ref].
In summary, the behavioral decoding of the two identified regions via BrainMap functional database indicates their previous significant association with sensory and phylogenetic older limbic processes.
## Whole-brain co-activation profiles
As our next step and in order to map brain regions that feature significant co-activation with the two identified regions, we performed MACM. In first instance, those experiments in BrainMap that feature activation in the region of the right basolateral amygdala/hippocampus and the right central insula were identified. Those regions that were more likely than by chance to be corecruited with these seeds and may hence be considered as a part of the functionally connected networks, were then further investigated by an ALE meta-analysis across identified experiments.
The MACM analysis indicated significant (p < 0.05 FWE corrected in cluster level) co-activation of the right amygdala/hippocampus region with several bilateral brain regions. In particular, coactivations were found with the left amygdala/hippocampus [bib_ref] Metaanalytic connectivity modeling: delineating the functional connectivity of the human amygdala, Robinson [/bib_ref] , medial prefrontal (area FP2 [bib_ref] Cytoarchitecture, probability maps and functions of the human frontal pole, Bludau [/bib_ref] , anterior cingulate cortex (areas s24 and s32 (Palomero-Gallagher et al., 2015)), posterior-medial frontal The results of meta-analytic connectivity modeling analysis. Task-based co-activation pattern of the right basolateral amygdala/hippocampus (A) and of the right central insula (B). All activations are significant at P < 0.05 corrected for multiple comparisons using the family-wise error rate in cluster level (cFWE).
cortex [bib_ref] Functional role of the supplementary and pre-supplementary motor areas, Nachev [/bib_ref] , and bilateral mid-fusiform gyrus (presumably corresponding to the fusiform face region) . We also found significant co-activation with the bilateral thalamus (e.g. mediodorsal and anterior nuclei projecting to the prefrontal cortex and the ventral lateral and ventral anterior nuclei shown to project to motor and premotor cortices [bib_ref] Non-invasive mapping of connections between human thalamus and cortex using diffusion imaging, Behrens [/bib_ref]. In addition the anterior insula and adjacent ventro-lateral prefrontal cortex (vlPFC) [bib_ref] A link between the systems: functional differentiation and integration within the human..., Kurth [/bib_ref] , as well as precuneus and posterior cingulate cortex (PCC) were identified bilaterally [fig_ref] Figure 4: Fig [/fig_ref].
Significant (p < 0.05 cFWE corrected) co-activation of the right central insula was largely symmetrically across both hemispheres. Co-activated regions comprised the bilateral parietal operculum (areas OP 1 and OP 4) [bib_ref] The human parietal operculum: II. Stereotaxic maps and correlation with functional imaging..., Eickhoff [/bib_ref] [bib_ref] The human parietal operculum. I. Cytoarchitectonic mapping of subdivisions, Eickhoff [/bib_ref] and adjacent inferior parietal cortex (areas PFop and PFcm [bib_ref] The human inferior parietal lobule in stereotaxic space, Caspers [/bib_ref] , bilateral thalamus [bib_ref] Non-invasive mapping of connections between human thalamus and cortex using diffusion imaging, Behrens [/bib_ref] and posteriormedial frontal cortex . We also found significant co-activation with the bilateral ventral striatum and the basal fore-brain. Finally, co-activated regions comprised the entire insula and perisylvian cortex including the adjacent ventro-lateral prefrontal cortex [bib_ref] Cytoarchitecture, probability maps and functions of the human frontal pole, Bludau [/bib_ref] [bib_ref] Meta-analytic clustering of the insular cortex characterizing the meta-analytic connectivity of the..., Cauda [/bib_ref] [bib_ref] A link between the systems: functional differentiation and integration within the human..., Kurth [/bib_ref] [fig_ref] Figure 4: Fig [/fig_ref].
As the last step, we analysed the concurrence activations across both MACM analyses in order to identify regions that feature significant task-based co-activation with both seeds (right amygdala/hippocampus and right central insula). This analysis revealed the bilateral anterior insula and left opercular area OP 1, bilateral thalamus, as well as posterior-medial frontal cortex [fig_ref] Figure 5: Conjunction analysis demonstrated regions significantly co-activated with both seeds [/fig_ref]. Assessing the functions significantly associated with this entire network through behavioral characterization using the BrainMap database similarly suggested a strong connection with perception and somatosensory processing [fig_ref] Figure 5: Conjunction analysis demonstrated regions significantly co-activated with both seeds [/fig_ref].
# Discussion
It has been suggested that the diverse and often conflicting findings regarding brain structure and function in a variety of disorders may be ameliorated by a more finely tuned understanding of which structures in networks are most implicated [bib_ref] Desperately seeking grey matter volume changes in sleep apnea: a methodological review..., Celle [/bib_ref]. In this vein, in our study, we have undertaken to gain greater understanding of core features of regional volume and activity alteration in patients with OSA across the published literature by using the ALE meta-analysis of currently available functional and structural imaging studies. This has been done with view to address the impact of diffuse changes in important areas on emotions, sensory and cognition impairments in patients with OSA.
In our study, the combined ALE quantitative analyses on group contrasts between patients with OSA and healthy controls highlighted structural atrophy and functional disturbances (task-related hypo-activations) in the clusters corresponding to the right basolateral amygdala/hippocampus and the right central insula [fig_ref] Figure 2: Fig [/fig_ref]. These highlighted regions, the amygdala/hippocampus and the insular cortex, have so far been relatively neglected nodes in the OSA neurocircuitry fingerprint. The likely contribution of these structures to recognized deficits and disabilities in OSA is hence explored here further. Behavioral decoding analyses of these two regions demonstrated the possible dysfunction of emotional, sensory, and cognitive processes [fig_ref] Figure 3: Behavioral characterization of the significant cluster in the right amygdala/hippocampus [/fig_ref]. The results of MACM analysis inferred that the right basolateral amygdala/hippocampus and central insula also comprise a network with the bilateral anterior insula, posterior-medial frontal cortex and thalamus (Figs. 4 and 5).
## Amygdala/hippocampus 'node' and osa
Convergent functional and structural alteration in the right basolateral amygdala and the CA1/CA3 regions of hippocampus and [fig_ref] 1: Introduction [/fig_ref] Studies entered into the meta-analysis are listed based on the year of publication and further alphabetically for each year. OSA = Obstructive sleep apnea; VBM = Voxel-based morphometry; fMRI = Functional magnetic resonance imaging; rs-fMRI = Resting-state fMRI; BMI = Body Mass Index. * Age was matched between groups and the authors reported only one value for two groups; ** Standard errors or confident intervals have been transformed to standard deviations (SD the subiculum [fig_ref] Figure 2: Fig [/fig_ref] in our results appears to have been driven by contribution of both structural and functional neuroimaging.
## Basolateral nucleus of amygdala
It has been posited that one of the functions of the amygdala is to link sensory input to emotional responses that then guide behavior. Aberrant facial emotions processing, emotional blunting and even aberrant sexual behaviors, dysfunctional memory and olfactory processing [bib_ref] The amygdala: vigilance and emotion, Davis [/bib_ref] , have all been reported with damage to amygdala [bib_ref] The amygdala: vigilance and emotion, Davis [/bib_ref]. Apart from the emotional processing, it has also been shown that basolateral nucleus of amygdala is involved in spatial and motor learning [bib_ref] The amygdala: vigilance and emotion, Davis [/bib_ref]. The majority of listed deficits have also been reported to a smaller or larger degree in patients with OSA [bib_ref] Sleep apnoea and the brain: a complex relationship, Rosenzweig [/bib_ref].
The importance of amygdala as one of structures affected by the chronic process of OSA has been previously intimated by several studies [bib_ref] Functional and structural changes in the brain associated with the increase in..., Fatouleh [/bib_ref] [bib_ref] Preliminary functional MRI neural correlates of executive functioning and empathy in children..., Kheirandish-Gozal [/bib_ref] [bib_ref] Preferential suppression of limbic Fos expression by intermittent hypoxia in obese diabetic..., Mukai [/bib_ref]. For instance, in children with OSA, during watching empathy-eliciting scenarios, the severity of OSA predicted less sensitivity to harm in the left amygdala (Kheirandish- [bib_ref] Preliminary functional MRI neural correlates of executive functioning and empathy in children..., Kheirandish-Gozal [/bib_ref]. In general, a greater neural recruitment of regions implicated in cognitive control, conflict monitoring, and attentional allocation has been required in those children with OSA in order to perform at the same level as children without OSA (Kheirandish- [bib_ref] Preliminary functional MRI neural correlates of executive functioning and empathy in children..., Kheirandish-Gozal [/bib_ref]. The results of our meta-analysis may go some way to explain an aberrant facial cues processing previously noted in children. Namely, our findings implicated the co-activation of nominally important regions and circuitry for this processing, e.g. the right amygdala region co-activation with the mid-fusiform gyrus, corresponding to the face region, and the anterior hippocampus, ventral striatum, contralateral amygdala, and the contralateral prefrontal cortex was found [fig_ref] Figure 4: Fig [/fig_ref]. Of all nuclei, the basolateral nucleus of amygdala appeared most affected according to the pooled results of the meta-analysis [fig_ref] Figure 2: Fig [/fig_ref].
Preclinical investigations of the amygdala connectivity suggest that the basolateral amygdala receives sensory information from the thalamus, hippocampus and cortex and then activates or modulates synaptic transmission in target areas appropriate for the reinforcement signal with which the sensory information has been associated. In animals, amygdala has been shown as vital for learning procedures and stress-induced conditioning that involves pairings of potent and arbitrary chemosensory stimuli [bib_ref] Temporary basolateral amygdala lesions disrupt acquisition of socially transmitted food preferences in..., Wang [/bib_ref]. For example, an animal study demonstrated that corticotropin-releasing factor receptors within the basolateral amygdala are involved in regulating fear-conditioned alterations in sleep [bib_ref] Basolateral amygdala and the regulation of fear-conditioned changes in sleep: role of..., Wellman [/bib_ref].
Similarly, and in keeping with these studies behavioral decoding of the regions highlighted by our meta-analysis suggested a significant association with affective and emotional processing, memory-related processes, and chemo-sensory processing [fig_ref] Figure 3: Behavioral characterization of the significant cluster in the right amygdala/hippocampus [/fig_ref].
Moreover, it has been suggested that the role of the amygdala in modulating momentary levels of vigilance in response to uncertainty underscores its likely importance for the etiology of anxiety disorders, MDD and PTSD [bib_ref] The amygdala: vigilance and emotion, Davis [/bib_ref] [bib_ref] Disrupted amygdalar subregion functional connectivity and evidence of a compensatory network in..., Etkin [/bib_ref] [bib_ref] Aberrant intrinsic connectivity of hippocampus and amygdala overlap in the fronto-insular and..., Tahmasian [/bib_ref]. The increased prevalence of these psychopathologies has been recognized in patients with OSA, possibly also suggesting a dual relationship. A meta-analysis of functional neuroimaging studies in PTSD and anxiety patients has suggested the basal portion of the amygdala as the major focus of hyperactivity [bib_ref] Functional neuroimaging of anxiety: a meta-analysis of emotional processing in PTSD, social..., Etkin [/bib_ref] , the same part of amygdala suggested by findings of our meta-analysis [fig_ref] Figure 2: Fig [/fig_ref].
Recent evidence further suggests that the amygdala is part of a complex network that mediates the formation of a larger repertoire of positive and negative emotions and it's dysfunctions may lead to various psychiatric disorders [bib_ref] The amygdala as a target for behavior surgery, Langevin [/bib_ref]. The pivotal role for the basolateral amygdala in differentiation of stimuli, and subsequent prediction of either positive or negative outcomes, has been suggested [bib_ref] A circuit mechanism for differentiating positive and negative associations, Namburi [/bib_ref]. Specifically, it has been shown that synaptic plasticity in the basolateral amygdala mediates the acquisition of associative memories of both end of emotional valences, and that different populations of neurons of that complex may encode fearful or rewarding associations [bib_ref] A circuit mechanism for differentiating positive and negative associations, Namburi [/bib_ref]. For example, it appears that basolateral nucleus neurons that project to the nucleus accumbens undergo synaptic changes following reward conditioning, whilst those that project to centromedial nucleus of amygdala undergo synaptic changes following fear conditioning [bib_ref] A circuit mechanism for differentiating positive and negative associations, Namburi [/bib_ref]. Hence, it follows that the importance of hypotrophy of such pivotal region such is basolateral nucleus of amygdala in patients with OSA should not be ignored. It is easy to postulate that this deficit must have reverberating impact on the patients' ability to decipher the very valence of ongoing life experiences and that in others lead to emotional blunting such can be seen in Alzheimer's disease (AD) or behavioural variant frontotemporal dementia [bib_ref] Emotion comprehension in the temporal variant of frontotemporal dementia, Rosen [/bib_ref] [bib_ref] The lower hippocampus global connectivity, the higher its local metabolism in Alzheimer..., Tahmasian [/bib_ref] [bib_ref] Based on the network degeneration hypothesis: separating individual patients with different neurodegenerative..., Tahmasian [/bib_ref].
Bzdok and colleagues highlighted the role of basolateral amygdala for processing and integrating environmental information and coordinating high-level sensory input, while the centromedial area is associated with mediating attentional, vegetative, and motor responses. Furthermore, they showed that the right basolateral amygdala is coactivated with the following regions: the left amygdala, dorsomedial prefrontal cortex, temporal pole, precuneus, inferior parietal cortex bilaterally, ventromedial prefrontal cortex, superior temporal gyrus/associative auditory cortex, middle frontal gyrus/frontal eye field, hippocampus, and posterior superior temporal sulcus on the left side . Our results are largely in agreement with the above-mentioned findings and also with the results of a previous study on amygdala that applied MACM [bib_ref] Metaanalytic connectivity modeling: delineating the functional connectivity of the human amygdala, Robinson [/bib_ref] [fig_ref] Figure 4: Fig [/fig_ref]. It also follows that dysfunction of emotional, sensory, and limbic processes reported in some patients with OSA might be partly explained with structural and functional alteration of right amygdala/hippocampus region [fig_ref] Figure 3: Behavioral characterization of the significant cluster in the right amygdala/hippocampus [/fig_ref].
## Hippocampus
The very possibility of a tripartite link between amygdala, memory impairment and OSA also posits itself as worthy of further exploration. The notion that activation of the amygdala during emotional arousal enhances memory in part by modulating plasticity in the hippocampus is not a new one [bib_ref] Memory-enhancing amygdala stimulation elicits gamma synchrony in the hippocampus, Bass [/bib_ref] [bib_ref] Psychic blindness and other symptoms following bilateral temporal lobectomy in rhesus monkeys, Kluver [/bib_ref] [bib_ref] Amygdala modulation of memory consolidation: interaction with other brain systems, Mcgaugh [/bib_ref] [bib_ref] Role of the basolateral amygdala in memory consolidation, Mcintyre [/bib_ref] [bib_ref] Role of the basolateral amygdala in memory consolidation, Pare [/bib_ref]. In addition, aberrant functional connectivity of amygdala and hippocampus may interact with dysfunctional intrinsic network activity in MDD , which might be related to emotional memory disturbances in OSA as well. Moreover, an aberrant connectivity between the hippocampus and the cerebellum has been reported in OSA patients, with view that this may lead to alterations in a distributed memory system for associative learning [bib_ref] CrossTalk opposing view: the intermittent hypoxia attending severe obstructive sleep apnoea does..., Rosenzweig [/bib_ref]. In addition, bilateral enlargement of hippocampus in OSA patients has been reported previously [bib_ref] The impact of sleep and hypoxia on the brain: potential mechanisms for..., Rosenzweig [/bib_ref] Of some note is that our results suggest higher damage to the anterior/ventral part of the hippocampal formation. In primates amygdala-projecting neurons are focally restricted to the most anterior (uncal) CA1 and pro-subiculum [bib_ref] Functional organization of the hippocampal longitudinal axis, Strange [/bib_ref] , and fMRI activations associated with emotional memory in humans have been found to be primarily in anterior regions [bib_ref] Functional organization of the hippocampal longitudinal axis, Strange [/bib_ref]. There is also evidence of double dissociation between semantic processing in the anterior hippocampus and non-semantic processing in the posterior hippocampus. One example of semantic processing that requires flexible expression of memory is transitive inference [bib_ref] Functional organization of the hippocampal longitudinal axis, Strange [/bib_ref] , or a form of inferential and deductive reasoning, another previously reported deficit in OSA patients [bib_ref] Neurocognitive impairment in obstructive sleep apnea, Lal [/bib_ref]. Similar to our findings, Weng and colleagues found significant grey matter atrophy mainly on the right in the parahippocampus in their meta-analysis [bib_ref] Mapping gray matter reductions in obstructive sleep apnea: an activation likelihood estimation..., Weng [/bib_ref]. Several preclinical studies also highlighted parahippocampus and hippocampal formation as important node contributing to the structural and functional abnormalities in OSA. For example, it has been demonstrated that the hippocampus is the region with high vulnerability to intermittent hypoxia, which may underlie the high frequency of neurobehavioral deficits observed frequently in OSA [bib_ref] Developmental differences in cortical and hippocampal vulnerability to intermittent hypoxia in the..., Gozal [/bib_ref]. Moreover, chronic episodic hypoxia during sleep impairs substantial region-selective neuronal loss within the CA1 region and leads to spatial learning impairments [bib_ref] Behavioral and anatomical correlates of chronic episodic hypoxia during sleep in the..., Gozal [/bib_ref]. The co-activation results of our study for the right hippocampal formation, (inclusive of anterior CA1 region), are also in agreement with the previously reported hippocampal co-activated networks (e.g. mPFC, bilateral superior frontal gyri, left hippocampus and parahippocampal gyrus, amygdala, cingulate cortex, thalamus, fusiform gyrus) and as such they might go some way towards explaining deficits in patients with OSA in perceptual, cognitive and affective processing domains. In the same vein, the functional characterization of the regions highlighted by our study suggests deficits in the perceptual, cognitive and affective processing domains [fig_ref] Figure 3: Behavioral characterization of the significant cluster in the right amygdala/hippocampus [/fig_ref]. Also, one can argue that recently shown association of OSA with AD and cognitive deficits noted in patients with OSA could be in part explained with hippocampal dysfunction, as previously demonstrated in AD [bib_ref] Alzheimer's disease neuroimaging, I. Sleep-disordered breathing advances cognitive decline in the elderly, Osorio [/bib_ref] [bib_ref] Link between hippocampus' raised local and eased global intrinsic connectivity in AD, Pasquini [/bib_ref] [bib_ref] The lower hippocampus global connectivity, the higher its local metabolism in Alzheimer..., Tahmasian [/bib_ref]. In addition, another meta-analysis recently suggested that patients with AD could have 5 times higher rate of presenting OSA symptoms than healthy individuals [bib_ref] The association between obstructive sleep apnea and Alzheimer's disease: a meta-analysis perspective...., Emamian [/bib_ref].
Of note is also that a majority of highlighted aberrant findings in our study were found to be non-dominant-sided. It has been suggested that there is a general hemispheric lateralization of perceptual processing, with lateralization of serial or local processing (on the left) versus parallel, global or holistic processing (on the right) [bib_ref] Lateralized human hippocampal activity predicts navigation based on sequence or place memory, Igloi [/bib_ref]. In a similar manner, some authors argue that rather than providing a single common function, the two hippocampi provide complementary representations for navigation (e.g. places on the right and temporal sequences on the left), both of which likely contribute to different aspects of episodic memory [bib_ref] Lateralized human hippocampal activity predicts navigation based on sequence or place memory, Igloi [/bib_ref]. The right hippocampus appears specifically involved in memory tasks requiring allocentric processing of spatial locations and hence its impairment may have clinical reverberations in patients' accurate large-scale navigation [bib_ref] Lateralized human hippocampal activity predicts navigation based on sequence or place memory, Igloi [/bib_ref] , and it may possibly also negatively affect the driving ability in OSA patients. By contrast, the left hippocampus appears to be involved in episodic/autobiographical memory [bib_ref] Lateralized human hippocampal activity predicts navigation based on sequence or place memory, Igloi [/bib_ref]. Similarly, it has been suggested that during response monitoring, the right amygdala and rostral anterior cingulate cortex may mediate aversive conditioning to errors, whereas the left amygdala may underpin detrimental negative affect concerning performance [bib_ref] Hemispheric differences in amygdala contributions to response monitoring, Polli [/bib_ref]. The possible clinical implication for our OSA patients would be that combined activity in these structures, which may serve to help us learn from our mistakes without becoming overly upset about them, is not optimal. Malfunction within this system could conceivably also contribute to the genesis of previously reported neuropsychiatric deficits prevalent in OSA, such as depression, emotional lability, anxiety or even aggravation of the PTSD.
Taken together, our study posits the amygdala and the hippocampal formation as important affected nodes in OSA neuropathology.
## Insula and osa
The right insular cortex has been highlighted as another cortical structure of importance by our results [fig_ref] Figure 2: Fig [/fig_ref]. So far, insular cortex has been somewhat ignored in theoretical constructs of likely neuropathology that underscores the affective and cognitive deficits in patients with OSA. Neuroanatomically, insula presents a nexus at the confluence of several neural pathways. It is densely interconnected with itself and with almost all cortical association regions [bib_ref] Meta-analytic clustering of the insular cortex characterizing the meta-analytic connectivity of the..., Cauda [/bib_ref] [bib_ref] A convergent functional architecture of the insula emerges across imaging modalities, Kelly [/bib_ref] [bib_ref] A link between the systems: functional differentiation and integration within the human..., Kurth [/bib_ref]. It has been shown to receive sensory, somesthetic and interoceptive inputs from cortical areas and via the thalamus [bib_ref] A convergent functional architecture of the insula emerges across imaging modalities, Kelly [/bib_ref] [bib_ref] Salience processing and insular cortical function and dysfunction, Uddin [/bib_ref]. It is also interconnected with the medial temporal lobe, amygdala, and basal ganglia [bib_ref] A convergent functional architecture of the insula emerges across imaging modalities, Kelly [/bib_ref]. It has been long proposed that the integration of all these inputs might present the neural substrate for human phenomenological experience, i.e. our own idiosyncratic conscious perception and understanding of a particular situation or phenomenon, further underscoring the importance of this structure (for detailed overview please refer to [bib_ref] A convergent functional architecture of the insula emerges across imaging modalities, Kelly [/bib_ref].
Our data suggests that OSA patients have convergent hypoactivation of the right central insula compared to healthy controls [fig_ref] Figure 2: Fig [/fig_ref]. Novel insights into the functional organization and specialization of the insula suggest that activity in the insula correlates with the degree of subjective salience, whether it is influenced by homeostatic, emotional or cognitive factors [bib_ref] A convergent functional architecture of the insula emerges across imaging modalities, Kelly [/bib_ref] [bib_ref] Salience processing and insular cortical function and dysfunction, Uddin [/bib_ref]. Moreover, the subdivisions of insular cortex have been shown to be co-dependent, and each subdivision participates to varying degrees in nearly every task domain that has been investigated, including those involving language, memory, sensory (from gustation and olfaction, to music perception), interoception, somesthesis and emotional processing [bib_ref] Salience processing and insular cortical function and dysfunction, Uddin [/bib_ref]. Again, it can be argued that the majority of listed functions have been recognized as abnormal or lacking to a varied degree in some patients with OSA.
The relative salience of the multitude of informational inputs during wakefulness determines those inputs which deserve attention [bib_ref] Salience processing and insular cortical function and dysfunction, Uddin [/bib_ref]. An intrinsic brain system known as the 'salience network', with key nodes in the insular cortices, has a central role in the detection of behaviorally relevant stimuli and the coordination of neural resources [bib_ref] Salience processing and insular cortical function and dysfunction, Uddin [/bib_ref]. Emerging evidence suggests that atypical engagement of specific subdivisions of the insula within the SN is a feature of many neuropsychiatric disorders, including of AD and MDD [bib_ref] Regional atrophy of the insular cortex is associated with neuropsychiatric symptoms in..., Moon [/bib_ref] [bib_ref] Insular cortex and neuropsychiatric disorders: a review of recent literature, Nagai [/bib_ref]. In OSA, localized cortical thinning has been reported in the region of insular cortices [bib_ref] Localized cortical thinning in patients with obstructive sleep apnea syndrome, Joo [/bib_ref] and bilaterally anterior insular neuronal damage and increased glial activation has also been shown [bib_ref] Insular cortex metabolite changes in obstructive sleep apnea, Yadav [/bib_ref]. Similarly, selective functional disconnection between the right anterior insula and the medial prefrontal cortex was correlated with the severity of the OSA in a very recent study, whilst the functional disconnection between the insula and the posterior cingulate cortex was correlated with depressive scores and working memory performance of patients with OSA . Of other sleep disorders, patients with insomnia have also been shown to have a greater involvement of the anterior insula, as well as insula BOLD correlation with EEG gamma frequency power during rest . Moreover, this increased involvement of the anterior/ventral insula was associated with negative affect. For instance, it has been suggested that here aberrant activation of the insula in arousal networks may underlie the misperception of sleep quality and subjective distress in insomnia . Given that insomnia and sleep apnea frequently co-exist [bib_ref] Comorbid insomnia and obstructive sleep apnea: challenges for clinical practice and research, Luyster [/bib_ref] , it is tempting to postulate that recently highlighted divergent results of subjective versus objective complaints in OSA patients may also be a reflection of similar misperception and/or interoception of variety of bodily and cognitive functions in a subgroup of patients with OSA [bib_ref] Sleep apnoea and the brain: a complex relationship, Rosenzweig [/bib_ref].
Cauda and colleagues demonstrated that anterior insula, mainly on the right side, plays an important role in saliency detection and cognition [bib_ref] Meta-analytic clustering of the insular cortex characterizing the meta-analytic connectivity of the..., Cauda [/bib_ref]. Moreover, the insula can be parcellated to the anterior part, which is characterized by an attentional pattern of connectivity with frontal, cingulate, parietal, cerebellar regions, whereas the posterior part is characterized by a more local connectivity pattern with connections to sensorimotor, temporal and posterior cingulate areas [bib_ref] Meta-analytic clustering of the insular cortex characterizing the meta-analytic connectivity of the..., Cauda [/bib_ref]. It has also been shown that posterior insula is mainly activated by interoception, perception and emotion [bib_ref] Meta-analytic clustering of the insular cortex characterizing the meta-analytic connectivity of the..., Cauda [/bib_ref]. In broad agreement with previous functional implications for insular cortices, the results of our meta-analyses implicate right central insular region as well as point to the deficits in perceptual, somatosensory and affective processing [fig_ref] Figure 3: Behavioral characterization of the significant cluster in the right amygdala/hippocampus [/fig_ref]. Arguably our findings also suggest that sensory and cognitive task-related modulations of the altered wider neurocircuitry in sleep apnea patients lead to a weaker central insular cortex's functional connectivity and activation during a given task, by comparison to healthy volunteers. However, the possibility that, as of yet unclear maladaptive structural process in sleep apnea patients, leads to plastic alterations and to the "functional" deactivation of this very region, should not be overlooked.
In that vein and regarding hemispheric lateralization of insular changes highlighted by our study, an interesting study suggests insight into the potential mechanisms behind the pathological process at hand [bib_ref] Insular cortex metabolite changes in obstructive sleep apnea, Yadav [/bib_ref]. Namely, Yadav and colleagues recently showed bilateral insular neuronal damage in OSA patients with higher glial activation and neuroinflammation on the left [bib_ref] Insular cortex metabolite changes in obstructive sleep apnea, Yadav [/bib_ref]. The authors speculated that asymmetrical outcome could stem from the larger cerebral blood flow on the right side with the consequences of hypoxemic periods during apnea having a relatively reduced effect on the right over the left side [bib_ref] Insular cortex metabolite changes in obstructive sleep apnea, Yadav [/bib_ref]. In view of our results, however, another interesting possibility presents itself. In contrast to their finding, our results suggest higher impact on the right and conceivably may also suggest an adaptive role for the microglial activation early on, or at certain stages, during the chronic process of OSA in some patients. How, when and which of the neuroinflammation processes may be protective/adaptive, and which detrimental/maladaptive, has become one of the crucial questions in fields of neurodegenerative, psychiatric, traumatic and ischemic diseases, and it is of further interest that many of diseases (etc. MDD, AD, stroke) have been found co-morbid or even prevalent in patients with OSA [bib_ref] Sleep apnoea and the brain: a complex relationship, Rosenzweig [/bib_ref].
## The convergent role for amygdala and insula?
In order to perform a particular neural function, to form a thought or an emotion, a set of brain networks or systems needs to transiently interact [bib_ref] Brain connectivity during resting state and subsequent working memory task predicts behavioural..., Sala-Llonch [/bib_ref]. The function and structure of such networks is of particular interest given that abnormalities in the interactions of network components can play a critical role in neuropsychiatric disorders, with damage to specific functional connectivity nodes and networks giving rise to distinct neurological and psychiatric syndromes [bib_ref] Human brain networks in health and disease, Bassett [/bib_ref] [bib_ref] Neurodegenerative diseases target large-scale human brain networks, Seeley [/bib_ref] [bib_ref] Based on the network degeneration hypothesis: separating individual patients with different neurodegenerative..., Tahmasian [/bib_ref] [bib_ref] A systematic review on the applications of resting-state fMRI in Parkinson's disease:..., Tahmasian [/bib_ref] [bib_ref] Evaluation of sleep problems in preeclamptic, healthy pregnant and non-pregnant women, Khazaie [/bib_ref]. The results of our conjunctional meta-analysis point to functionally important co-activation between our seeds (in the right amygdala/hippocampus and right central insula) and the bilateral anterior insula, bilateral thalamus, as well as the posterior-medial frontal cortex [fig_ref] Figure 5: Conjunction analysis demonstrated regions significantly co-activated with both seeds [/fig_ref]. Assessing the behavioral characterization of both seeds using the BrainMap database demonstrated significant involvement of somatosensory and affective processing [fig_ref] Figure 5: Conjunction analysis demonstrated regions significantly co-activated with both seeds [/fig_ref]. These findings are in agreement with previous MACM and behavioral decoding studies on the right amygdala/hippocampus and right insula [bib_ref] Meta-analytic clustering of the insular cortex characterizing the meta-analytic connectivity of the..., Cauda [/bib_ref] [bib_ref] A link between the systems: functional differentiation and integration within the human..., Kurth [/bib_ref] [bib_ref] Metaanalytic connectivity modeling: delineating the functional connectivity of the human amygdala, Robinson [/bib_ref].
The regions highlighted by our meta-analyses: the right amygdala/hippocampus and right central insula all exist, or contribute to, the flow of information as connection hubs on one of the three important intrinsic connectivity networks: the default mode network (DMN), the central executive network (CEN) and the salience network (SN). Of these, CEN (including the dorsolateral prefrontal cortex and posterior parietal cortex) and DMN (including the medial prefrontal cortex (mPFC)) and PCC are two important anticorrelated cognitive-related networks [bib_ref] Large-scale brain networks and psychopathology: a unifying triple network model, Menon [/bib_ref] [bib_ref] Brain connectivity during resting state and subsequent working memory task predicts behavioural..., Sala-Llonch [/bib_ref]. The DMN plays a competitive role with the majority of task-related networks and the activation of a cognitive task-related network is commonly accompanied by deactivation of the DMN [bib_ref] The brain's default network: anatomy, function, and relevance to disease, Buckner [/bib_ref] [bib_ref] Brain connectivity during resting state and subsequent working memory task predicts behavioural..., Sala-Llonch [/bib_ref]. It has been shown that the connectivity within DMN regions in task-fMRI studies contributes to the facilitation or monitoring of cognitive performance, and that the differences in functional coupling within DMN regions can predict differences in cognitive performance [bib_ref] Brain connectivity during resting state and subsequent working memory task predicts behavioural..., Sala-Llonch [/bib_ref].
Over the years, numerous neuroimaging studies have been performed to identify structural and functional brain impairments in OSA patients in order to explain for noted emotional and cognitive deficits, including altered brain activation and deactivation in the CEN, DMN and SN [bib_ref] Altered resting-state brain activity in obstructive sleep apnea, Zhang [/bib_ref]. Defective deactivation of DMN in patients with OSA during task has been suggested by findings of one study, which also proposed the intermittent hypoxia damage as a more likely culprit behind observed aberrant connectivity [bib_ref] Task positive and default mode networks during a parametric working memory task..., Prilipko [/bib_ref]. In addition, differential compensatory spatial recruitment of the task positive network and DMN has been demonstrated, with a different pattern of spatial recruitment and deactivation noted in comparison to healthy controls [bib_ref] Task positive and default mode networks during a parametric working memory task..., Prilipko [/bib_ref]. An altered activation of the CEN and deactivation of the DMN during working memory tasks in OSA patients has also been shown . Similarly, in rs-fMRI studies functional disconnection in brain areas of the CEN and DMN in OSA patients has been reported [bib_ref] Altered resting-state brain activity in obstructive sleep apnea, Zhang [/bib_ref]. Taken together, these findings might be taken to suggest a role for the functional impairment of the CEN and DMN in cognitive deficits in patients with OSA. Of note, some of those alterations have been reported as reversible and CPAP treatment has been shown to increase the connectivity of the DMN in elderly patients with OSA and to attenuate cortical thinning.
The functional disconnection of the insular cortex with the CEN and DMN networks has been reported in normal aging and many neuropsychiatric, neurodegenerative disorders [bib_ref] Insular dysfunction within the salience network is associated with severity of symptoms..., Manoliu [/bib_ref]. In keeping with the findings of our meta-analysis that highlighted ventral insular cortex as an important node in somatosensory, neurocognitive, perceptual and affective deficits in patients with OSA, a recent study has demonstrated decreased functional connectivity of the right insular cortex with the main nodes of the DMN. This has been taken to indicate the functional disconnection between the SN and DMN in OSA patients . In addition, the decrease in functional connectivity between the right AI and the mPFC has been significantly correlated with the apnea and hypopnea index (AHI value). This correlation has been taken to suggest the injury by the intermittent hypoxia as the most likely culprit underscoring the aberrant disconnection . It has been suggested that the functional disconnection between the insular cortex and the DMN may in itself be sufficient to lead to aberrant cognitive control sig-nals and to result in cognitive impairment in OSA patients . In addition, in another study, the abnormal insular cortex metabolites in adults with OSA showed significant correlations with disease severity and neuropsychological status, suggesting not just cognitive but also emotional/affective impact of this disconnection [bib_ref] Insular cortex metabolite changes in obstructive sleep apnea, Yadav [/bib_ref].
Finally, we found significant connectivity between the seeds (the right amygdala/hippocampus and right central insula) and bilateral anterior insula and left opercular area, bilateral thalamus, as well as posterior-medial frontal cortex [fig_ref] Figure 5: Conjunction analysis demonstrated regions significantly co-activated with both seeds [/fig_ref]. This pattern closely resembles the canonical frontoparietal executive control network (e.g. CEN) identified in many studies of cognitive control over emotional and non-emotional material [bib_ref] Disrupted amygdalar subregion functional connectivity and evidence of a compensatory network in..., Etkin [/bib_ref]. This coordinated network has also been observed using rs-fMRI and as previously noted; it has been reliably dissociated from the SN [bib_ref] Disrupted amygdalar subregion functional connectivity and evidence of a compensatory network in..., Etkin [/bib_ref]. The executive control network in healthy controls does not normally include the amygdala or insula, and thus coupling of these two structures with this network and impaired dissociation with SN, and possibly also impaired deactivation of DMN in patients, likely further reflects a chronic disorder driven network-level neural adaptation.
In summary, the noted aberrant connectivity could be argued to underlie previously reported dysmetric deficits of affect, attention, information processing and visuo-motor control [bib_ref] Sleep apnoea and the brain: a complex relationship, Rosenzweig [/bib_ref] [bib_ref] The impact of sleep and hypoxia on the brain: potential mechanisms for..., Rosenzweig [/bib_ref]. Of further note, similar aberrant connectivity with some of these networks has also been recently reported in patients with generalised anxiety disorder, MDD, and PTSD [bib_ref] Altered resting-state functional connectivity of basolateral and centromedial amygdala complexes in posttraumatic..., Brown [/bib_ref] [bib_ref] Disrupted amygdalar subregion functional connectivity and evidence of a compensatory network in..., Etkin [/bib_ref] [bib_ref] Insular dysfunction within the salience network is associated with severity of symptoms..., Manoliu [/bib_ref] [bib_ref] Aberrant intrinsic connectivity of hippocampus and amygdala overlap in the fronto-insular and..., Tahmasian [/bib_ref].
## Research in context
Recently, an ALE meta-analysis on eight VBM studies found significant reductions in gray matter of the bilateral parahippocampal (more robust on the right side) and frontotemporal regions (less-pronounced) in patients with OSA [bib_ref] Mapping gray matter reductions in obstructive sleep apnea: an activation likelihood estimation..., Weng [/bib_ref]. Similarly, our meta-analysis suggests convergent grey matter atrophy and functional hypo-activation in the basolateral amygdala and the hippocampal formation. Unlike their study, our ALE analysis included both structural and functional (e.g. task-fMRI and rs-fMRI) neuroimaging studies in order to comprehensively assess both abnormalities in OSA. Due to our stringent inclusion criteria we have excluded several structural studies that were otherwise incorporated by them [bib_ref] Mapping gray matter reductions in obstructive sleep apnea: an activation likelihood estimation..., Weng [/bib_ref] , For example, we excluded all those previous studies that did not report standard space coordinates [bib_ref] Brain morphology associated with obstructive sleep apnea, Macey [/bib_ref] [bib_ref] Neural consequences of sleep disordered breathing: the role of intermittent hypoxia, Morrell [/bib_ref] , studies that did not find group difference between OSA patients and controls [bib_ref] Cerebral structural changes in severe obstructive sleep apnea, O'donoghue [/bib_ref] , and one interventional study [bib_ref] Obstructive sleep apnea: brain structural changes and neurocognitive function before and after..., Canessa [/bib_ref]. In addition, in our meta-analysis we used as a statistically significant cut off point P < 0.05 corrected for multiple comparisons using the FWE in cluster level. This cut off point is more conservative than False Discovery Rate (FDR) used in that meta-analyses [bib_ref] Mapping gray matter reductions in obstructive sleep apnea: an activation likelihood estimation..., Weng [/bib_ref] and as such has likely had an impact on our results. Moreover, we indentified behavioral characterization of the seed regions using the BrainMap database and also task-based co-activation patterns of functionally connected areas to these regions, which have not been previously done.
## Potential limitations of the present study
Coherent summary of the neuroimaging literature about the impact of OSA on the brain that is growing in scope and complexity requires increasingly sophisticated tools for synthesizing findings across studies. The meta-analysis has been accepted as an important tool to develop new hypotheses on structure-function correspondence and to establish consensus on the locations of func-tional regions in diseases such as OSA across previously published studies [bib_ref] Activation likelihood estimation meta-analysis revisited, Eickhoff [/bib_ref] [bib_ref] Evaluating the consistency and specificity of neuroimaging data using meta-analysis, Wager [/bib_ref]. Nonetheless, the meta-analysis itself is not fully immune of limitations that arise from characteristics of the primary studies under review . It is hence of importance to recognize that despite best efforts to circumvent many limitations connected to single studies and the stringent exclusion and inclusion criteria used, the studies included in our meta-analysis differed regarding design, methodology, and the study population. Unfortunately, none of the above mentioned analysis, controlled for age or gender covariates across studies due to current methodological limitations. In addition, the majority of the included studies in our meta-analysis investigated the brain activation or changes in a sample of men without consideration of possible gender differences and thus, we were not able to identify the activation patterns separately for both genders. Similarly, it is impossible to account for separate activation patterns at various stages of OSA in different study populations, and to discern the temporal vector of the noted changes. Additionally, it should be also noted that although we included all available neuroimaging studies that satisfied our predetermined criteria, it is still possible that the sample size for our meta-analysis has underpowered our findings. For example, in this study our exclusion criteria stipulated exclusion of interventional studies and it is possible that their inclusion might have strengthened our findings. Future studies can apply various cognitive and emotional task fMRI to understand different angels of the associated neuropsychiatric symptoms of OSA (e.g. memory loss, depression). Moreover, we did not find any results from rs-fMRI experiments maybe because of low number of available rs-fMRI studies. rs-fMRI as a promising non-invasive tool is, however, currently widely employed to measure functional connectivity alteration in different neuropsychiatric disorders and maybe future studies can implement it to assess the intrinsic functional abnormalities in OSA.
Beside those primary source-related issues, the very determination of the consistent brain structures differences remains the problem in various meta-analysis methods, including in the ALE method, which was utilized in this study. The CBMA method uses precise coordinates (rather than general regional labels) as its input. The issue arises from the fact that these peak coordinate foci are limited indicators of the location of a significant anatomical difference [bib_ref] Evaluating the consistency and specificity of neuroimaging data using meta-analysis, Wager [/bib_ref]. Finally, during the ALE analysis, overlapping clusters of difference are commonly found by averaging across different peak coordinates, increasing the risk that with two or more relatively nearby peak foci, ALE will find an average, 'significant' cluster somewhere between these foci in a brain region which was actually not reported in the source studies . On the other hand, one study recommended using image-based rather than CBMA method in ALE meta-analysis, which was applied in this study. The reason is the CBMA method provide less information from each study [bib_ref] Meta-analysis of neuroimaging data: a comparison of image-based and coordinate-based pooling of..., Salimi-Khorshidi [/bib_ref]. However, the original data is often very difficult to obtain from previous studies than reported peak coordinates. Hence, CBMA is still the standard ALE approach to detect convergent regional abnormalities in neuropsychiatric disorders .
# Conclusions
The impact of oxidative and neuroinflammatory effects of OSA on the right amygdala/hippocampus along with the right insular cortex and other subcortical and cortical structures plays an important role, which has been previously suggested to underscore several of subjective and objective cognitive and emotional complaints of adult OSA patients [bib_ref] Sleep apnoea and the brain: a complex relationship, Rosenzweig [/bib_ref]. Here, we present findings of the meta-analysis on neuroimaging studies that also implicates the right amygdala/hippocampus complex and the insular cortex as important nodes on the affected cognitive and affective circuitry in OSA patients. Moreover, in accordance with this and previous studies, the behavioral characterization of the entire highlighted network using the BrainMap database suggested implications for the emotional and memory related functions, as well as somatosensory processing in the affected patients. A MACM analysis demonstrated that the right amygdala/hippocampus and insula are part of a joint network comprising the anterior insula, posterior-medial frontal cortex and thalamus. Further, our study strongly suggested non-dominant lateralization of noted chronic deficits in OSA. It was outside the scope of this paper to provide any detailed mechanistic insights behind this phenomenon, however, its significance should not be ignored and should be further explored in future studies.
Taken collectively, neuroimaging and neurophysiological studies in patients with OSA have delineated a putative regional "fingerprint" of OSA-induced brain injury. They purport a disconnection of the fronto-parietal regions and a disruption of the thalamocortical oscillator, with involvement of the hippocampal formation .
One of the challenges for future research will be to establish and differentiate the nuanced task fMRI profiles and patterns of functional connectivity of particular subdivisions of the insula and amygdala with the intrinsic brain networks in OSA patients. Given the burgeoning body of research into aberrant connectivity of intrinsic brain networks and their implication in disorders such as AD and other neuropsychiatric and neurodegenerative disorders, the ability to decipher correct or convergent biomarkers for each of these disorders can not be overstated. Ideally, any such research would also make it possible to delineate specific contribution of several of neuropathological facets of OSA injury on any changes noted, e.g. sleep fragmentation versus the impact of intermittent hypoxia versus any other confounding factors such as obesity. This has so far been difficult to implement, but future careful experimental designs might help with this issue. In particular targeted cognitive and emotional tasks fMRI studies could be well positioned to explore some of the previously reported neurocognitive and neuropsychiatric symptoms associated with OSA. Finally, it is hoped that the findings presented here may offer a tentative first step towards this task, as well as to provide an initial theoretical framework for interpreting the aberrant activity within these network nodes.
## Conflict of interest
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
[fig] Figure 1: Paper selection strategy flow chart. fMRI = functional magnetic resonance imaging; VBM = voxel-based morphometry. et al., 2009), references for this meta-analysis were collected by a search of the PubMed database in April 2015, and by reference tracing of retrieved articles. Keywords for the search were as follows: (i) (("Sleep Apnea Syndromes"[Mesh] OR "Sleep Apnea; Central"[Mesh] OR "Sleep Apnea; Obstructive"[Mesh]) OR sleep apnea) AND (("functional magnetic resonance imaging") OR "fMRI"); which resulted in 45 studies; (ii) ("Sleep Apnea Syndromes"[Mesh] OR "Sleep Apnea; Central"[Mesh] OR "Sleep Apnea; [/fig]
[fig] Figure 2: Fig. 2. Convergence of structural and functional difference in obstructive sleep apnea compared with healthy controls. Location of the significant convergence of gray matter reduction and functional disturbance in the right basolateral nucleus of the human amygdala/hippocampus (A) and in the right central insula (B). Results are from the Activation Likelihood Estimation for sleep apnea meta-analyses. All activations are significant at P < 0.05 corrected for multiple comparisons using the family-wise error rate in cluster level (cFWE). [/fig]
[fig] Figure 3: Behavioral characterization of the significant cluster in the right amygdala/hippocampus (A) and in the right central insula (B). Only behavioral domains significantly associated with the respective clusters at p < 0.05 (corrected for multiple comparisons) are illustrated. [/fig]
[fig] Figure 4: Fig. 4. The results of meta-analytic connectivity modeling analysis. Task-based co-activation pattern of the right basolateral amygdala/hippocampus (A) and of the right central insula (B). All activations are significant at P < 0.05 corrected for multiple comparisons using the family-wise error rate in cluster level (cFWE). [/fig]
[fig] Figure 5: Conjunction analysis demonstrated regions significantly co-activated with both seeds (right basolateral amygdala/hippocampus and right central insula) (A). (B) Functional characterization of network shown in figure 5A (p < 0.05 corrected). [/fig]
[table] 1: Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143 2. Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143 2.1. Search strategies and study selection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143 2.2. Data extraction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144 2.3. Activation likelihood estimation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144 2.4. Functional decoding using the BrainMap database . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145 2.5. Whole-brain co-activation profiles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145 [/table]
|
Alkaline ceramidase 2 is a novel direct target of p53 and induces autophagy and apoptosis through ROS generation
ACER2 is a critical sphingolipid metabolizing enzyme, and has been shown to be remarkably upregulated following various stimuli such as DNA damage. However, the transcriptional regulatory mechanism of ACER2 gene and its potential role in the regulation of autophagy remain unknown. In this study, we have for the first time identified the human ACER2 gene promoter, and found that human ACER2 transcription is directly regulated by p53 and ACER2 is implicated in the induction of autophagy as well as apoptosis. A series of luciferase reporter assay demonstrated that ACER2 major promoter is located within its first intron where the consensus p53-binding sites exist. Consistently, forced expression of p53 significantly stimulated ACER2 transcription. Notably, p53-mediated autophagy and apoptosis were markedly enhanced by ACER2. Depletion of the essential autophagy gene ATG5 revealed that ACER2-induced autophagy facilitates its effect on apoptosis. Further studies clearly showed that ACER2-mediated autophagy and apoptosis are accompanied by ROS generation. In summary, our present study strongly suggests that ACER2 plays a pivotal role in p53-induced autophagy and apoptosis, and thus might serve as a novel and attractive molecular target for cancer treatment.Tumor suppressor p53 plays a crucial role in the regulation of stress response and tumor development 1-5 . p53 acts as nuclear transcription factor to regulate a variety of its target genes implicated in diverse cellular processes including DNA repair, cell cycle arrest, apoptosis, autophagy, senescence, angiogenesis and migration 6-10 . A growing body of recent evidence suggests that, in addition to stress response, p53 also modulates cellular metabolism such as glycolysis, oxidative phosphorylation, fatty acid metabolism, amino acid metabolism and reactive oxygen species through transactivating metabolic enzyme-related genes 11,12 . Indeed, it has been shown that several genes encoding important metabolic enzymes are the direct transcriptional targets of p53 including TP53-induced glycolysis and apoptosis regulator (TIGAR), glutaminase 2 (GLS2) and cytochrome oxidase 2 (Sco2) 13-15 . As expected, p53 has an ability to impair the oncogenic metabolic reprogramming required for cancer cell growth and survival 11,12 .Sphingolipids are membrane lipids that are ubiquitously expressed in all eukaryotic cells. In addition to their structural roles in membrane biology, sphingolipids act as bioactive signaling molecules within cells. For example, metabolites of sphingolipid such as ceramide, sphingosine and sphingosine-1-phosphate (S1P) have been shown to participate in various cellular processes including proliferation, differentiation, adhesion, migration, apoptosis and autophagy[16][17][18][19][20][21][22]. Ceramidases represent a family of sphingolipid-metabolizing enzymes, which catalyze the hydrolysis of ceramides to generate sphingosine. To date, five distinct human ceramidases have been identified including acid ceramidase (AC), neutral ceramidase (NC), alkaline ceramidase 1 (ACER1), alkaline ceramidase 2 (ACER2) and alkaline ceramidase 3 (ACER3) 23 . Among these five ceramidases, ACER2 is ubiquitously expressed at low level in various normal tissues except placenta, and remarkably up-regulated in response to various cellular stimuli such as DNA damage and serum starvation 23-28 . Recently, it has been shown that DNA damage-mediated 4,6-diamidino-2-phenylindole (DAPI), and cover slips were mounted onto glass slides by using ProLong Gold Antifade Reagent. Cells were observed under a confocal microscope.Staining autophagosomes with GFP-LC3. Cells seeded on glass cover slips coated with poly-lysine in a 24-well culture plate were co-transfected with 100 ng of GFP-LC3 together with 200 ng of pcDNA3.0-Flag empty vector, pcDNA3.0-Flag-ACER2 or with pcDNA3.0-Flag-p53 using Lipofectamine 2000 (Invitrogen). Forty-eight hours after transfection, the fluorescence of GFP-LC3 was observed under a confocal microscope. Cell nuclei and the exogenous Flag-ACER2/Flag-p53 were stained with DAPI and anti-Flag antibody, respectively. Flow cytometric analysis for apoptosis. Annexin V-FITC apoptosis detection kit (BD Pharmingen) was used to detect apoptotic activity according to the manufacturer's instructions with slight modifications. Briefly, cells were harvested, washed twice in ice-cold 1 × PBS and resuspended in binding buffer. Annexin V-FITC and propidium iodide (PI) were then added to the reaction mixture, and incubated for 15 min at room temperature in the dark. Cells were then analyzed for apoptosis by flow cytometry.Flow cytometric analysis for ROS. Reactive oxygen species was detected by cell-permeable fluorogenic probe 2ʹ , 7ʹ -dichlorodihydrofluorescin diacetate (DCFH-DA) according to the manufacturer's instructions. Briefly, cells were washed in 1 × PBS and incubated with DCFH-DA-containing medium at 37 °C for 30 min. Cells were then washed in 1 × PBS and subjected to DCF fluorescent quantification using flow cytometry.Transmission electron microscopy (TEM). Cells were pre-fixed in 4% glutaraldehyde for 1.5 h at 4 °C and fixed in 1% OsO4. After dehydration and embedding, ultrathin sections were prepared with a Reichert-Jung Ultracut Ultramicrotome (Leica, Vienna, Austria). Images were observed under a H7500 transmission electron microscope (Hitachi Ltd, Tokyo, Japan).DNA sequence alignment and database analysis. ACER2 mRNA and genomic sequences were extracted from GeneBank and UCSC database (https://genome.ucsc.edu/). The 5′ -RACE sequences were aligned with ACER2 genomic sequences to verify gene identity, exon usage and location of the transcription start sites.
Scientific RepoRts | 7:44573 | DOI: [bib_ref] TFIIS.h, a new target of p53, regulates transcription efficiency of pro-apoptotic bax..., Liao [/bib_ref].1038/srep44573 up-regulation of ACER2 promotes apoptosis [bib_ref] Alkaline ceramidase 2 and its bioactive product sphingosine are novel regulators of..., Xu [/bib_ref]. Of note, dysregulation of ACER2 has been observed in several types of cancers. However, the transcriptional regulatory mechanism of ACER2 gene following cellular stimuli and its potential role in the regulation of certain important biological processes such as autophogy and tumorigenesis remain largely unknown [bib_ref] Golgi alkaline ceramidase regulates cell proliferation and survival by controlling levels of..., Xu [/bib_ref] [bib_ref] Expression profiling and identification of novel genes in hepatocellular carcinomas, Graveel [/bib_ref].
In the present study, we have identified ACER2 promoter and also found that ACER2 is a direct transcriptional target gene of p53. Furthermore, we have demonstrated that ACER2 is required for the induction of autophagy and apoptosis through the enhancement of ROS generation. Together, our present findings strongly suggest that ACER2 which is tightly implicated in sphingolipid metabolism plays an important role in the regulation of p53-dependent autophagy and apoptosis.
# Results
Gene organization and chromatin state of ACER2 gene locus. To analyze the genomic organization and chromatin state of human ACER2 gene, we have employed UCSC genome browser (https://genome.ucsc.edu/). As shown in [fig_ref] Figure 1: Chromatin state and 5′-RACE analysis of Human ACER2 gene [/fig_ref] , ACER2 gene is mapped at chromosome 9p22 and composed of six exons and five introns. According to ENCODE histone modification data, the transcription elongation hallmark of H3K36me3 accumulated within ACER2 gene body. Consistent with this finding, ChromHMM chromatin state segmentation data indicated that ACER2 gene might be actively transcribed. Of note, the first entire exon and the 5′ -region of the first intron of ACER2 gene contained a classic CpG island with high DNase I hypersensitivity and H3K4me3 (a hallmark of transcription initiation), suggesting that ACER2 gene promoter might be present within the first exon and/or 5′ -part of the first intron of ACER2 gene.
## Identification of the transcription start site(s) of acer2 gene.
To identify the transcription initiation site(s) of ACER2 gene, we have employed SMART RACE assay system using two reverse primers (GSP1 and GSP2) and adapter primers (NUP and AP) [fig_ref] Figure 1: Chromatin state and 5′-RACE analysis of Human ACER2 gene [/fig_ref]. As shown in [fig_ref] Figure 1: Chromatin state and 5′-RACE analysis of Human ACER2 gene [/fig_ref] , the primary PCR (AP/GSP2) failed to amplify any products, whereas the nested PCR (NUP/GSP1) generated two distinct products. Based on the subsequent cloning and sequencing of the above two PCR products, the shorter fragment (191 bp in length without The genomic region of ACER2 (chr9:19405000-19455000, human species genomic assembly version, GRCh37/hg19) was retrieved and is schematically represented with the indicated tracks (https://genome.ucsc.edu/). The chromatin state segmentation is displayed as active promoter (bright red), strong enhancer (orange), insulator (blue), transcriptional elongation (dark green) and weakly transcribed region (light green). (B) Schematic representation of human ACER2 gene organization and 5′ -RACE primer design. Exons and introns are indicated by filled boxes and thin lines, respectively. The translation start codon (ATG) and the translation stop codon are indicated by arrows. (C) 5′ -RACE analysis. 5′ -RACE analysis was conducted using human testis total RNA. Primary and nested PCR were performed using AP/GSP2 and NUP/ GSP1, respectively. PCR products were analyzed on 1.0% agarose gel electrophoresis, purified for cloning and subsequently verified by sequencing. Lane M: DL2000 size markers; Lane 1: Primary PCR products; Lane 2: Nested PCR products.
Scientific RepoRts | 7:44573 | DOI: 10.1038/srep44573 RACE adapter primer) was specific to ACER2 gene. The first base of this PCR product (G, guanine) was mapped at 76 bp upstream of ACER2 start codon, which might be the transcription initiation site of ACER2 gene (+ 1).
## Identification of the acer2 promoter region.
To identify the potential proximal promoter region(s) of ACER2 gene, we generated three luciferase reporter constructs containing the indicated genomic fragments of ACER2 gene such as P1285 (− 470/+ 815), P800 (+ 15/+ 815) and P676 (+ 140/+ 815) [fig_ref] Figure 2: Identification of ACER2 promoter region [/fig_ref]. The results obtained from the chromatin state and transcription start site analyses [fig_ref] Figure 1: Chromatin state and 5′-RACE analysis of Human ACER2 gene [/fig_ref] prompted us to generate P1285 The positions relative to the major ACER2 transcription start site (+ 1) are indicated. Exons and introns of ACER2 are shown by open boxes and thin lines, respectively. For luciferase reporter assay, H1299 (left) and HEK293 (right) cells were transiently co-transfected in triplicate in 12-well plates with the indicated luciferase reporter constructs together with Renilla luciferase reporter plasmid (pRL-TK) by using Lipofectamine 2000 transfection reagent. Forty-eight hours after transfection, firefly and Renilla luciferase activities were measured by Dual Luciferase Assay System (Promega). Data obtained from a representative of at least three independent experiments were shown as fold induction compared to the activity of cells transfected with the empty pGL3basic vector (lower). (B) The major ACER2 promoter exists within its first intron (+ 419/+ 815). A schematic diagram of ACER2 gene promoter reporter constructs (upper). For luciferase reporter assay, H1299 and HeLa cells were transiently co-transfected with the indicated luciferase reporter constructs along with pRL-TK. Fortyeight hours after transfection, firefly and Renilla luciferase activities were measured and analyzed as in [fig_ref] Figure 2: Identification of ACER2 promoter region [/fig_ref] (lower).
(− 470/+ 815) carrying the 5′ -upstream region, first exon and 5′ -part of the first intron of ACER2 gene. The indicated luciferase reporter constructs were introduced into H1299 and HEK293 cells and their luciferase activities were measured. As shown in [fig_ref] Figure 2: Identification of ACER2 promoter region [/fig_ref] , a significant increase in luciferase activities arising from each of three constructs [P1285 (− 470/+ 815), P800 (+ 15/+ 815) and P676 (+ 140/+ 815)] was detectable as compared to that of the empty pGL3-basic plasmid, suggesting that a genomic region from + 140 to + 815 of ACER2 gene has a strong promoter activity.
To further delineate ACER2 gene promoter, we have generated the additional four luciferase reporter constructs including P397 (+ 419/+ 815), P304 (+ 140/+ 443), P223 (+ 140/+ 362) and P130 (+ 140/+ 275), based on their parental P676 (+ 140/+ 815) [fig_ref] Figure 2: Identification of ACER2 promoter region [/fig_ref]. Luciferase reporter assay demonstrated that luciferase activity driven from P397 (+ 419/+ 815) is comparable to that from the parental P676 (+ 140/+ 815), whereas P304 (+ 140/+ 443) fails to drive luciferase reporter gene expression, indicating a genomic region from + 419 to + 815 of ACER2 gene contains an obvious promoter activity. Intriguingly, P223 (+ 140/+ 362) had a lower but considerable promoter activity relative to pGL3-basic vector, suggesting that the genomic region from + 140 to + 362 contains an alternative minor promoter for ACER2 gene.
Taken together, we concluded that the major ACER2 promoter exists within the first intron spanning from + 419 to + 815.
Sequence and homology analysis of the ACER2 promoter. To gain an insight into understanding the transcriptional regulatory mechanisms of ACER2 gene, the promoter region of ACER2 gene was subjected to transcription factor binding site analysis using MatInspector professional and TFSEARCH software. As shown in [fig_ref] Figure 3: ACER2 gene promoter region contains consensus p53-binding sites [/fig_ref] , ACER2 gene promoter lacks the classical TATA box, but contains the classical GC box and the other putative binding sites of transcription factors such as p53, AP-1 and GATA-1. In addition, homology analysis using ClustalW2 algorithm revealed that ACER2 promoter sequence is evolutionarily conserved across various species such as human, mouse and rat [fig_ref] Figure 3: ACER2 gene promoter region contains consensus p53-binding sites [/fig_ref]. This high degree of homology among ACER2 promoter regions was comparable to that among the coding regions of ACER2. Since two potential p53-binding sites were evolutionarily well conserved among species, it is highly suggestive that p53 might directly regulate ACER2 gene transcription.
p53 directly transactivates ACER2. To ask whether p53 could directly transactivate ACER2 gene, p53-deficient H1299 cells were co-transfected with p53 expression plasmid together with the indicated ACER2 luciferase reporter plasmids. Luciferase reporters bearing p53-target p21 and Bax promoters were employed as positive controls. The luciferase reporter assay revealed that ectopic expression of wild-type p53 enhances luciferase activities driven by the indicated ACER2 promoters including P1285, P800, P676 and P397 [fig_ref] Figure 4: ACER2 is a direct transcriptional target of p53 [/fig_ref]. Under our experimental conditions, p21 and Bax promoters responded to the exogenous p53. In a sharp contrast to wild-type p53, mutant p53 (p53R175H) had an undetectable effect on P397 ACER2 luciferase reporter [fig_ref] Figure 4: ACER2 is a direct transcriptional target of p53 [/fig_ref].
To verify whether p53-mediated ACER2 transactivation could be dependent on the putative p53-binding sites of ACER2 promoter, we have introduced point mutation(s) at the two putative p53-binding site(s) of ACER2 promoter and generated P397 p53m1, P397 p53m2 and P397 p53m1 + 2 luciferase reporter plasmids [fig_ref] Figure 4: ACER2 is a direct transcriptional target of p53 [/fig_ref]. H1299 cells were then co-transfected with p53 expression plasmid together with the indicated luciferase reporter plasmids. As shown in [fig_ref] Figure 4: ACER2 is a direct transcriptional target of p53 [/fig_ref] , disruption of the first p53-binding site (p53CBS1) but not the second one (p53CBS2) completely impaired p53-dependent ACER2 transactivation, suggesting that the first p53-binding site is required for p53-mediated ACER2 transcriptional activation. Consistent with these results, ChIP assay demonstrated that, like p21 promoter, p53 is efficiently recruited onto the first p53-binding site within ACER2 promoter region in cells [fig_ref] Figure 4: ACER2 is a direct transcriptional target of p53 [/fig_ref].
Next, we tested whether p53 could transactivate the endogenous ACER2 expression. As shown in [fig_ref] Figure 4: ACER2 is a direct transcriptional target of p53 [/fig_ref] , ectopic expression of wild-type p53 resulted in up-regulation of the endogenous ACER2 in H1299 cells. Under our experimental conditions, p53-dependent endogenous transcriptional induction of p21, Bax and MDM2 was detectable. Then, we assessed whether ACER2 could be induced following DNA damage in a p53-dependent manner. To this end, p53-deficient H1299 and p53-proficient A549 cells were treated with 1 μ M of adriamycin (ADR). Twenty-four hours after treatment, total RNA was prepared and subjected to qRT-PCR. As seen in [fig_ref] Figure 4: ACER2 is a direct transcriptional target of p53 [/fig_ref] , a massive increase in the endogenous ACER2 transcription was detected in A549 cells but not in H1299 cells. Moreover, siRNA-mediated silencing of p53 in A549 cells significantly prohibited ADR-dependent transcriptional induction of ACER2 [fig_ref] Figure 4: ACER2 is a direct transcriptional target of p53 [/fig_ref]. Collectively, our results strongly suggest that ACER2 is a novel direct target gene of p53.
ACER2 stimulates autophagy and apoptosis. Since p53 plays an important role in the regulation of both autophagy and apoptosis, we next asked whether ACER2 could induce autophagy and/or apoptosis in a p53-dependent manner. For this purpose, H1299 cells were transfected with the indicated combinations of the expression plasmids. Forty-eight hours after transfection, cell lysates were prepared and analyzed by p53CBS2). Exons of ACER2 are indicated as filled boxes (upper) and the potential p53-binding sites are shown as open boxes (lower). (D) p53CBS1 is required for p53-dependent activation ACER2 promoter. The indicated luciferase reporter constructs were introduced into H1299 cells together with empty plasmid or with wild-type p53 expression plasmid. Forty-eight hours after transfection, their luciferase activities were determined as in (A). (E) Direct recruitment of p53 onto ACER2 promoter region in cells. H1299 cells were transfected with p53 expression plasmid or with the empty plasmid. Forty-eight hours after transfection, chromatin fragments were prepared and immunoprecipitated with p53 antibody or with control IgG. The co-precipitated DNA fragments were purified and subjected to qPCR analysis. p53-target p21 promoter was amplified as a positive control. (F) p53 activates the endogenous ACER2 transcription. H1299 cells were transiently transfected with the empty plasmid or with p53 expression plasmid. Forty-eight hours after transfection, total RNA was extracted and subjected to qRT-PCR with the indicated primer sets. (G) Adriamycin-dependent induction of the endogenous ACER2. p53-deficient H1299 cells and p53-proficient A549 cells were treated with or without 1 μ M of adriamycin (ADR). Twenty-four hours after treatment, qRT-PCR was conducted. (H) p53 is required for DNA damage-induced ACER2 up-regulation. A549 cells were transfected with control siRNA or with p53specific siRNA. Twenty-four hours after transfection, cells were treated with 1 μ M of ADR. Twenty-four hours after treatment, qRT-PCR was conducted. immunoblotting. As shown in [fig_ref] Figure 5: ACER2 induces apoptosis and autophagy [/fig_ref] , forced expression of either ACER2 or p53 alone caused a remarkable induction of autophagy marker LC3-II. It has been well known that mTOR/Akt pathway plays an inhibitory role in autophagy [bib_ref] Sphingolipids: regulators of crosstalk between apoptosis and autophagy, Young [/bib_ref] [bib_ref] Concomitant reduction of c-Myc expression and PI3K/AKT/mTOR signaling by quercetin induces a..., Granato [/bib_ref]. Consistently, the amounts of phosphor-mTOR and phosphor-Akt, which suppress autophagy, were decreased in H1299 cells expressing either ACER2 or p53, strongly suggesting that ACER2 has an ability to promote autophagy at least in part through the inhibition of mTOR-Akt pathway. Under the same experimental conditions, transfected cells were analyzed by flow cytometry for apoptosis. As shown in [fig_ref] Figure 5: ACER2 induces apoptosis and autophagy [/fig_ref] , overexpression of either ACER2 or p53 alone caused a remarkable increase in PI + /AnexinV + apoptotic population, suggesting that, like p53, ACER2 has an ability to induce apoptosis. Notably, co-expression of ACER2 with p53 resulted in a much more larger PI + /AnexinV + apoptotic population than ACER2 or p53 alone, indicating that ACER2 augments p53-dependent apoptosis.
In support of the results obtained from immunoblotting, electron microscopic analysis demonstrated a significant accumulation of the double-membraned autophagic vesicles (autophagosomes) in cells transfected with either ACER2 or p53 expression plasmid [fig_ref] Figure 5: ACER2 induces apoptosis and autophagy [/fig_ref]. We then examined the cellular distribution of LC3. Once autophagosomes appear, the diffused cytoplasmic form of LC3, LC3-I, is conjugated with lipid to form LC3-II, which is associated with autophagosome membranes, and thus can be visualized in small puncta. Indirect immunofluorescence staining revealed that overexpression of either ACER2 or p53 causes a marked increase in number of LC3 puncta [fig_ref] Figure 1: Chromatin state and 5′-RACE analysis of Human ACER2 gene [/fig_ref]. Notably again, co-expression of ACER2 with p53 resulted in much . ACER2 and ATG5 are required for DNA damage-induced apoptosis and autophagy. A549 cells were transiently transfected with control siRNA (NCsi), siRNA against ACER2 (ACER2si) or with ATG5 siRNA (ATG5si). Twenty-four hours after transfection, cells were exposed to 2 μ M of ADR or left untreated. Thirty-six hours after treatment, cells were subjected to qPCR analysis (A), immunoblot analysis (B), and flow cytometric analysis (C).
Scientific RepoRts | 7:44573 | DOI: 10.1038/srep44573 more double-membraned autophagic vesicles [fig_ref] Figure 5: ACER2 induces apoptosis and autophagy [/fig_ref] and LC3 puncta [fig_ref] Figure 1: Chromatin state and 5′-RACE analysis of Human ACER2 gene [/fig_ref] , and much less phosphorylated mTOR and Akt [fig_ref] Figure 5: ACER2 induces apoptosis and autophagy [/fig_ref] , than p53 or ACER2 alone, suggesting that ACER2 augments p53-dependent autophagy.
ACER2 contributes to DNA damage-induced autophagy and apoptosis. Next, we examined whether ACER2 induction could be required for p53-mediated autophogy and apoptosis following DNA damage. To this end, we have employed siRNA-mediated knockdown of ACER2 in A549 cells bearing wild-type p53. Twenty-four hours after siRNA transfection, cells were treated with DMSO or with ADR (2 μ M). Thirty-six hours after treatment, total RNA was isolated and analyzed by qRT-PCR. As shown in , silencing of ACER2 was successful. Under the same experimental conditions, cell lysates were prepared and analyzed for LC3-II by immunoblotting. As shown in , ADR-mediated up-regulation of LC3-II was attenuated by ACER2 depletion. FACS analysis revealed that silencing of ACER2 reduces number of PI + /AnnexinV + apoptotic cells in response to ADR as compared to cells transfected with control siRNA . Consistent with these observations, enforced expression of ACER2 in H1299 cells led to an increase in number of autophagic and apoptotic cells in response to ADR as compared with that of control cells [fig_ref] Figure 7: ACER2 enhances DNA damage-induced apoptosis and autophagy [/fig_ref]. Taken together, these results strongly suggest that ACER2 plays an important role in the regulation of p53-induced autophagy and apoptosis in response to DNA damage. We then examined the effects of ACER2-induced autophagy on apoptosis under our experimental conditions. To this end, we have employed specific siRNA to knockdown the essential autophagy gene ATG5 in A549 cells . The results revealed that siRNAs-mediated ATG5 silencing reduces ADR-mediated up-regulation of LC3-II and number of PI + /AnnexinV + apoptotic cells in response to ADR as compared to cells transfected with control siRNA . Of note, ATG5 depletion did not affect ADR-induced ACER2 up-regulation as compared with the cells transfected with control siRNA . Thus, these data suggest that ACER2-mediated autophagy facilitates the subsequent apoptotic cell death following DNA damage.
## Acer2 induces apoptosis and autophagy by inducing ros generation.
Previous studies demonstrated that overexpression of ACER2 induces the generation of its bioactive product, sphingosine, which leads to the production of reactive oxygen species (ROS) [bib_ref] Alkaline ceramidase 2 and its bioactive product sphingosine are novel regulators of..., Xu [/bib_ref] [bib_ref] Expression profiling and identification of novel genes in hepatocellular carcinomas, Graveel [/bib_ref] [bib_ref] Differential signaling of sphingosine derivatives in U937 human monocytes depends on the..., Kim [/bib_ref] [bib_ref] Synthesis of sphingosine is essential for oxidative stress-induced apoptosis of photoreceptors, Abrahan [/bib_ref] [bib_ref] Phytosphingosine in combination with ionizing radiation enhances apoptotic cell death in radiation-resistant..., Park [/bib_ref]. In addition, it has been well known that ROS potentiates apoptosis and autophagy [bib_ref] Free radicals in cross talk between autophagy and apoptosis, Kaminskyy [/bib_ref] [bib_ref] Superoxide is the major reactive oxygen species regulating autophagy, Chen [/bib_ref]. Therefore, we sought to examine whether ACER2 overexpression could augment apoptosis and autophagy by increasing ROS levels. H1299 cells were treated with NAC or left untreated. Two hours after treatment, cells were transfected with the empty plasmid or with the expression plasmid for ACER2. Forty-eight hours after transfection, cells were analyzed by immunoblotting or by flow cytometry. As shown in [fig_ref] Figure 8: ACER2 induces apoptosis and autophagy though enhancing ROS generation [/fig_ref] -C, ACER2 overexpression resulted in a dramatic increase in ROS level accompanied by the induction of apoptosis and autophogy. However, treatment of cells with ROS scavenger, N-acetyl-cysteine (NAC), ACER2 overexpression-mediated apoptosis and autophagy were significantly impaired. In contrast to the addition of NAC, the exogenous H 2 O 2 enhanced ACER2 overexpression-mediated apoptosis and autophagy [fig_ref] Figure 8: ACER2 induces apoptosis and autophagy though enhancing ROS generation [/fig_ref]. Together, these results suggest that ACER2 induces apoptosis and autophagy through the production of ROS.
# Discussion
In addition to its well-established role in carbohydrate and amino acid metabolisms, it has recently been shown that tumor suppressor p53 is implicated also in the regulation of sphingosine metabolism [bib_ref] Sphingolipids in the DNA damage response, Carroll [/bib_ref] [bib_ref] p53 and regulation of bioactive sphingolipids, Heffernan-Stroud [/bib_ref]. Numerous studies revealed that cellular stresses such as DNA damage significantly increase the expression levels of bioactive sphingolipids including ceramides 20,37-44 , sphingosine 45 and sphingosine-1-phosphate [bib_ref] Apoptosis induces expression of sphingosine kinase 1 to release sphingosine-1-phosphate as a..., Gude [/bib_ref]. In mammalian cells, there are at least 33 distinct enzymes involved in sphingolipid metabolism [bib_ref] Alkaline ceramidase 2 and its bioactive product sphingosine are novel regulators of..., Xu [/bib_ref]. Of note, systemic analysis on DNA damage-mediated expression of major sphingolipid-metabolizing enzymes demonstrated that ADR treatment causes a marked increase in ACER2 expression level, whereas the expression levels of the other sphingolipid-metabolizing enzymes remain unchanged or moderately up-regulated in response to ADR [bib_ref] Alkaline ceramidase 2 and its bioactive product sphingosine are novel regulators of..., Xu [/bib_ref]. Additional analysis clearly showed that ACER2 promotes ADR-dependent apoptosis through the generation of sphingosine and ROS [bib_ref] Alkaline ceramidase 2 and its bioactive product sphingosine are novel regulators of..., Xu [/bib_ref]. These observations strongly suggest that, among sphingolipid metabolizing enzymes, ACER2 contributes to the proper DNA damage response. Intriguingly, our present studies have demonstrated that p53 directly transactivates ACER2, which subsequently induces autophagy and apoptosis in response to DNA damage. Therefore, it is highly likely that ACER2 is transactivated by p53 in response to DNA damage, and contributes to DNA damage-mediated induction of autophagy and apoptosis. With these in mind, ACER2 might be a novel molecular target for cancer therapy.
It has been shown that sphingosine level is elevated in response to DNA damage such as ADR in cancer cells carrying wild-type p53 including MCF7, HCT116 and A549 cells [bib_ref] Alkaline ceramidase 2 and its bioactive product sphingosine are novel regulators of..., Xu [/bib_ref] [bib_ref] Sphingosine generation, cytochrome c release, and activation of caspase-7 in doxorubicin-induced apoptosis..., Cuvillier [/bib_ref]. From our present findings showing that p53 directly up-regulates ACER2 which catalyzes the hydrolysis of ceramides to generate sphingosines, it is conceivable that p53 might affect the intracellular concentration and/or distribution of sphingosine through the transactivation of ACER2. To confirm this issue, further experiments should be required.
Our present results have also demonstrated that ACER2 is sufficient to induce apoptosis. Consistent with our findings, Xu et al. described that DNA damage-dependent up-regulation of ACER2 contributes to the induction of apoptosis through the production of sphingosine and ROS [bib_ref] Alkaline ceramidase 2 and its bioactive product sphingosine are novel regulators of..., Xu [/bib_ref]. Although several lines of evidence indicate that sphingosine and its analogs stimulate the production of ROS and thereby inducing apoptosis [bib_ref] Differential signaling of sphingosine derivatives in U937 human monocytes depends on the..., Kim [/bib_ref] [bib_ref] Synthesis of sphingosine is essential for oxidative stress-induced apoptosis of photoreceptors, Abrahan [/bib_ref] [bib_ref] Phytosphingosine in combination with ionizing radiation enhances apoptotic cell death in radiation-resistant..., Park [/bib_ref] [bib_ref] Loss of sphingosine kinase-1 in carcinoma cells increases formation of reactive oxygen..., Huwiler [/bib_ref] , it remains elusive how ACER2-mediated generation of sphingosine could enhance the cellular ROS level. Notably, Zigdon et al. found that an accumulation of dihrdrosphingosine, a saturated analog of sphingosine, arising from a deficiency of ceramide synthase 2 (CerS2) leads to ROS generation through the inhibition of the mitochondrial respiratory chain [bib_ref] Ablation of ceramide synthase 2 causes chronic oxidative stress due to disruption..., Zigdon [/bib_ref]. In vitro experiments demonstrated that sphingosine and sphinganine directly inhibit complex IV activity of the isolated mitochondria [bib_ref] Ablation of ceramide synthase 2 causes chronic oxidative stress due to disruption..., Zigdon [/bib_ref]. Therefore, further studies should be required to adequately address whether this could be the case.
To our knowledge, we have demonstrated for the first time that ACER2 also regulates autophagy through the generation of ROS. Autophagy and apoptosis have been generally accepted to be evolutionarily conserved cellular processes to determine cell fate in response to cellular stresses including DNA damage [bib_ref] Sphingolipids: regulators of crosstalk between apoptosis and autophagy, Young [/bib_ref] [bib_ref] Regulation of autophagy by sphingolipids, Bedia [/bib_ref] [bib_ref] Autophagosomal Membrane Serves as Platform for Intracellular Death-inducing Signaling Complex (iDISC)-mediated Caspase-8..., Young [/bib_ref]. Based on our present results, DNA damage-mediated apoptosis was significantly impaired in ATG5-or ACER2-depleted cells. Thus, it is possible that ACER2 contributes to the induction of not only apoptosis but also autophagy. Previously, Young et al. described the sphingolipids-mediated crosstalk between apoptosis and autophagy 18 . According to their results, SKI-I, a pansphingosine kinase inhibitor, potentiated apoptosis accompanied by the induction of autophagy [bib_ref] Sphingolipids: regulators of crosstalk between apoptosis and autophagy, Young [/bib_ref]. Since sphingosine kinase inhibitor resulted in an accumulation of sphingosine, it is necessary to further investigate whether ACER2 could participate in sphingosine-induced autophagy and apoptosis.
Notably, the induction of autophagy by sphingosine has not yet been demonstrated [bib_ref] Sphingolipids: regulators of crosstalk between apoptosis and autophagy, Young [/bib_ref]. Based on our present findings, it is conceivable that sphingosine, a product of ACER2, could induce autophagy through increasing ROS generation. Further study should be required to adequately address this issue. In addition, it has been shown that ceramide, a substrate of ACER2, regulates both apoptosis and autophagy, and also directly inhibits mitochondrial complex III to generate ROS 18 . However, the functional importance of ROS in ceramide-mediated autophagy and apoptosis has not yet been fully clarified and thus merits further study. In summary, our present study not only identified the critical sphingolipid metabolizing enzyme ACER2 as a novel direct target of p53, but also demonstrated a novel role of ACER2 in the regulation of autophogy as well as apoptosis though the generation of sphingosine and ROS. Based on our current findings, further extensive studies on the function of ACER2 and sphingosine might provide a clue to develop a promising strategy for cancer therapy.
# Methods
Reagents and cell lines. Hydrogen peroxide (H 2 O 2 ), adriamycin hydrochloride (ADR), N-acetyl-cysteine (NAC), dimethyl sulfoxide (DMSO) and 4′ ,6-Diamidino-2-phenylindole dihydrochloride (DAPI) were purchased from Sigma-Aldrich (China). Human lung cancer H1299 and A549 cells were maintained in a humidified atmosphere containing 5% CO 2 at 37 °C in RPMI 1640 medium (H1299) or in DMEM (A549) supplemented with 100 units/ml of penicillin, 100 mg/ml of streptomycin and 10% (vol/vol) heat-inactivated FBS (Invitrogen).
Rapid amplification of cDNA ends (RACE). 5′ -RACE analysis was carried out using SMART TM RACE cDNA amplification kit (Clontech) as described previously [bib_ref] The gene pair PRR11 and SKA2 shares a NF-Y-regulated bidirectional promoter and..., Wang [/bib_ref]. Briefly, human normal testis total RNA (Ambion) was reverse-transcribed to generate first-strand cDNA with the incorporation of an adaptor onto its 5′ end. Two distinct ACER2 gene-specific primers (GSP1 and GSP2) corresponding to its second and fourth exons were designed and synthesized. The primer sequences are listed in . Primary and nested PCR were conducted by using AP/GSP2 and NUP/GSP1, respectively. The resultant nested PCR products were cloned and sequenced.
Luciferase reporter constructs and reporter assays. The putative promoter region of ACER2 was obtained by PCR-based amplification and cloned into pGL3-basic vector to generate ACER2-P1285. A series of deletion mutants were constructed using Site-Directed Mutagenesis kit (TOYOBO, Japan) according to the manufacturer's instructions. Luciferase reporter constructs containing the indicated point mutation(s) at the potential p53-binding sites were also generated by using Site-Directed Mutagenesis kit. The sequences of primers used are listed in . All of the reporter constructs were validated by direct sequencing.
For luciferase reporter analysis, cells were seeded in triplicate into 12-well plates and co-transfected with the indicated reporter vectors, pRL-TK vector (Promega) encoding Renilla luciferase together with the empty vector (pcDNA3) or with pcDNA3-Flag-p53 expression vector using Lipofectamine 2000 reagent (Invitrogen). Forty-eight hours after transfection, cells were lysed and their luciferase activities were measured using Dual-Luciferase assay system (Promega) as described previously [bib_ref] The gene pair PRR11 and SKA2 shares a NF-Y-regulated bidirectional promoter and..., Wang [/bib_ref].
## Construction of the expression plasmids and transient transfection.
The entire coding region of human ACER2 was amplified by PCR and inserted into Eco R I/Xho l sites of the mammalian expression plasmid pcDNA3.0-Flag. The resultant expression plasmid was verified by sequencing and named as pcDNA-Flag-ACER2. The mutant p53 expression plasmid, pcDNA-p53 (R175H), was kindly provided by Dr. Giovanni Blandino. For transient transfection, cells were seeded at a density of 1 × 10 5 cells/12-well tissue culture plate or 3.5 × 10 5 cells/6-well tissue culture plate and incubated overnight. Cells were then transiently transfected with the indicated plasmids using Lipofectamine 2000 reagent (Invitrogen) according to the manufacturer's protocols.
siRNA synthesis and transfection. The negative control siRNA and siRNAs against ACER2 and ATG5 were chemically synthesized by Shanghai GenePharma (Shanghai, China). The sequences of siRNAs used are listed in . siRNAs were transfected into the indicated cells using Lipofectamine RNAiMAX reagent (invitrogen) according to the manufacturer's instructions.
## Rna isolation and rt-pcr.
Total RNA isolation and quantitative RT-PCR were conducted as described previously [bib_ref] PRR11 is a novel gene implicated in cell cycle progression and lung..., Ji [/bib_ref]. The sequences of the primers for PCR are shown in . Immunoblot analysis. Immunoblot analysis was performed as described previously with slight modifications [bib_ref] PRR11 is a novel gene implicated in cell cycle progression and lung..., Ji [/bib_ref]. In brief, cells were lysed in 2 × SDS sample buffer (for the detection of ACER2 protein in Golgi membranes) or in RIPA buffer (for the detection of the other proteins) supplemented with protease inhibitor PMSF (Beyotime, Jiangsu, China). Total proteins were quantified using the bicinchoninic acid (BCA) protein assay kit (Thermo Scientific, Beijing, China). The antibodies used in the present study were shown in . The blots were visualized by the enhanced chemiluminescence (ECL, Bio-Rad Laboratories, USA).
## Chromatin immunoprecipitation (chip). chip was performed using ez chip tm chromatin
Immunoprecipitation kit (Upstate, Lake Placid, NY) as described previously [bib_ref] The gene pair PRR11 and SKA2 shares a NF-Y-regulated bidirectional promoter and..., Wang [/bib_ref]. The sequences of the primers and the antibodies used were provided in Supplementary Tables S1 and S3, respectively. qPCR data for ChIP were calculated and shown as % of recovered immunoprecipitated DNA relative to the Input DNA, i.e. % Input = 100 × 2 −∆Ct , where ∆ Ct = (Ct [ChIP] -(Ct [Input] -Log 2 100)).
Indirect immunofluorescence staining. Cells were seeded on glass cover slips coated with poly-lysine in a 24-well culture plate, and then transfected with the indicated plasmids. Forty-eight hours after transfection, cells were fixed in 4% paraformaldehyde in 1 × PBS, permeabilized with 0.05% Triton X-100 at room temperature for 15 min, and blocked with 0.5% BSA and 10% goat serum for 1 h at 37 °C. After blocking, cells were incubated with anti-LC3 antibody for overnight at 4 °C followed by the incubation with secondary antibody conjugated with fluorescein isothiocyanate (FITC) for 2 h. After washing in 1 × PBS, cells were incubated with
[fig] Figure 1: Chromatin state and 5′-RACE analysis of Human ACER2 gene. (A) Chromatin state annotation of human ACER2 gene locus. [/fig]
[fig] Figure 2: Identification of ACER2 promoter region. (A) ACER2 genomic region between + 140 and + 815 contains a strong promoter activity. A schematic diagram of ACER2 gene promoter reporter constructs (upper). [/fig]
[fig] Figure 3: ACER2 gene promoter region contains consensus p53-binding sites. (A) Nucleotide sequence of human ACER2 promoter region. The potential transcription factor binding sites are underlined. Nucleotides are numbered relative to the major transcription start site of ACER2 (+ 1). (B) Sequence alignment of human, mouse and rat ACER2 gene promoters. The potential p53-binding sites are boxed. [/fig]
[fig] Figure 4: ACER2 is a direct transcriptional target of p53. (A) Overexpression of wild-type p53 enhances ACER2 promoter activity. p53-deficient H1299 cells were transiently co-transfected with the indicated luciferase reporter constructs and pRL-TK together with the expression plasmid for wild-type p53. Forty-eight hours after transfection, their luciferase activities were examined as inFig. 2.(B) Mutant p53 has an undetectable effect on ACER2 promoter activity. H1299 cells were co-transfected with ACER2-P397 reporter construct and pRL-TK along with mutant p53 (R175H) expression plasmid. Forty-eight hours after transfection, luciferase activities were examined as in (A). (C) Schematic diagram of the luciferase reporter plasmids containing ACER2 promoter region with the indicated point mutations at the potential p53-binding sites (p53CBS1 and Scientific RepoRts | 7:44573 | DOI: 10.1038/srep44573 [/fig]
[fig] Figure 5: ACER2 induces apoptosis and autophagy. (A) Forced expression of ACER2 induces autophagy accompanied by the inhibition of mTOR-Akt pathway. H1299 cells were transiently transfected with the indicated combinations of the expression plasmids. Forty-eight hours after transfection, cell lysates were prepared and subjected to immunobloting with the indicated antibodies. (B) ACER2 induces apoptosis. H1299 cells were transfected as in (A). Forty-eight hours after transfection, cells were subjected to the flow cytometric analysis. (C) and (D) ACER2 induces autophagy. H1299 cells were transfected as in (A). Fortyeight hours after transfection, cells were analyzed by the transmission electron microscopy (C) and the indirect immunofluorescence staining with LC3 antibody (D). [/fig]
[fig] Figure 7: ACER2 enhances DNA damage-induced apoptosis and autophagy. A549 cells were transiently transfected with the empty plasmid or with ACER2 expression plasmid. Twenty-four hours after transfection, cells were incubated in the presence or absence of 2 μ M of ADR. Thirty-six hours after treatment, cells were subjected to immunoblot analysis (A) and flow cytometric analysis (B).Scientific RepoRts | 7:44573 | DOI: 10.1038/srep44573 [/fig]
[fig] Figure 8: ACER2 induces apoptosis and autophagy though enhancing ROS generation. (A-C) ACER2mediated apoptosis and autophagy is prohibited by NAC. H1299 cells were pre-treated with 2 mM N-acetylcysteine (NAC) for 2 h, and then transiently transfected with the empty plasmid or with ACER2 expression plasmid. Forty-eight hours after transfection, cells were processed for immunoblot assay (A), the flow cytometric analysis for ROS (B) and the flow cytometric analysis for apoptosis (C). (D,E) H 2 O 2 treatment augments ACER2-mediated apoptosis and autophagy. H1299 cells were transiently transfected with the empty plasmid or with ACER2 expression plasmid. Twenty-four hours after transfection, cells were treated with the indicated concentrations of H 2 O 2 or left untreated. Forty-eight hours after H 2 O 2 treatment, cells were subjected to immunoblot analysis (D) and the flow cytometric analysis for ROS (E). (F) Schematic representation of p53/ACER2-dependent apoptosis and autophagy. DNA damage causes p53 activation, which subsequently transactivates ACER2. ACER2 enhances sphingosine level and also stimulates the generation of ROS, leading to the induction of autophagy and apoptosis. Scientific RepoRts | 7:44573 | DOI: 10.1038/srep44573 [/fig]
|
Actual Anatomical and Dosimetric Changes of Parotid Glands in Nasopharyngeal Carcinoma Patients during Intensity Modulated Radiation Therapy
The goal of this study was to evaluate the actual anatomical and dosimetric changes of parotid glands in nasopharyngeal carcinoma patients during intensity modulated radiation therapy. With helical tomotherapy, its planning system, and adaptive software, weekly anatomical and dosimetric changes of parotid glands in 35 NPC patients were evaluated. Interweekly parotid volume varied significantly ( < 0.03). The rate of volume change reached the highest level at the 16th fraction. The average 1 increased by 32.2 (left) and 28.6 (right), and the average 50 increased by 33.9 (left) and 24.93 (right), respectively. Repeat data comparison indicated that the 1 and 50 varied significantly among different fractions (both with = 0.000). The variation of parotid volume was inversely correlated with that of the 1 and 50 (both with = 0.000). In conclusion, parotid volume and actual dose vary significantly in NPC patients during IMRT. Replanning at the end of the fourth week of IMRT may have clinical benefits.
# Introduction
Due to the anatomical and biological specificity of nasopharyngeal carcinoma (NPC), radiation therapy or chemoradiotherapy has been recognized as a definitive treatment [bib_ref] Clinical outcomes and prognostic factors of 695 nasopharyngeal carcinoma patients treated with..., Wang [/bib_ref]. Studies have shown that the higher the radiation dose delivered to the target volume, the better the local disease control ratio [bib_ref] Dose-response relationship of nasopharyngeal carcinoma above conventional tumoricidal level: a study by..., Teo [/bib_ref]. The escalation of the delivered dose, however, often leads to severe and related side effects. Xerostomia is one of the most frequent side effects and the amount of radiation that is delivered to the parotid glands, which assume a major role in stimulating salivary flow, affects NPC patients' quality of life. Therefore, it is crucial to minimize the dose to the parotid gland while assuring adequate dose distribution to the target volume in the treatment of NPC. Unlike twodimensional conventional radiation therapy (2DCRT) and three-dimensional conformal radiation therapy (3DCRT), intensity modulated radiation therapy (IMRT) can deliver a highly conformal dose to targets while effectively sparing critical normal organs, potentially improving the local control rate and reducing radiation-related toxicities [bib_ref] Intensity-modulated radiation therapy with or without chemotherapy for nasopharyngeal carcinoma: radiation therapy..., Lee [/bib_ref] [bib_ref] The treatment outcome and radiation-induced toxicity for patients with head and neck..., Kouloulias [/bib_ref].
Patients with head and neck cancer may be subjected to significant anatomical changes during radiation therapy, changes which can cause volume shrinkage near the facial surface. And parotid gland variations may result in an unanticipated overdose. A hybrid IMRT plan, generated by applying the beam configurations of the first plan to the anatomical structures of the second simulation CT images, has been used to evaluate possible volumetric and dosimetric variations [bib_ref] Clinical study of the necessity of replanning before the 25th fraction during..., Wang [/bib_ref] [bib_ref] Assessment of anatomical and dosimetric changes by a deformable registration method during..., Lu [/bib_ref]. However, it is inevitable that this approach will develop bias.
Helical tomotherapy (HT) is a unique IMRT modality that combines elements of diagnostic radiology and radiation therapy in a single unit. In addition to its ability to deliver a highly conformal dose distribution, HT is equipped with xenon detectors that have been designed to obtain the megavoltage computed tomography (MVCT) images that are used for pretreatment setup verification [bib_ref] Megavoltage computed tomography: an emerging tool for image-guided radiotherapy, Hong [/bib_ref]. Meanwhile, HT is equipped with adaptive planning software which can calculate actual dose distribution in each treatment fraction [bib_ref] The use of megavoltage CT (MVCT) images for dose recomputations, Langen [/bib_ref]. To evaluate the actual anatomical and dosimetric changes in NPC patients during IMRT, we performed this study.
# Materials and methods
Between March 2009 and August 2010, 35 histologically proven and locoregionally advanced NPC patients were treated with HT in our center. Informed consent was obtained from all patients before receiving treatment. Patient characteristics are summarized in [fig_ref] Table 1: Patients' clinical characteristics [/fig_ref]. All patients were treated with HT, but 15 underwent concurrent cisplatinbased chemotherapy, 2 received concurrent cisplatin-based chemotherapy with anti-EGFR monoclonal antibody (Mab), and 5 underwent concurrent anti-EGFR Mab treatment. Patients' weight was noted before and at the end of treatment. All patients underwent planning kilovoltage CT (KVCT) scanning with a slice thickness of 3 mm. The patients were immobilized with a thermoplastic head-and-shoulder mask and a head-and-shoulder immobilization board. Each patient underwent scanning through the head and neck region (from the head to below the clavicles). Enhancement CT and plain CT images were transmitted to a Pinnacle 3 8.0 workstation and fused. Enhanced CT, MRI, or PET-CT images were used to guide the contours of the target volumes. Each patient received a total of 33 fractions of radiation, resulting in 70 Gy to the gross tumor volume and positive lymph nodes (pGTVnx and pGTVnd were obtained by expanding the corresponding gross tumor volume and metastatic nodes with a margin of 3-5 mm), 60 Gy to the high-risk planning target volume, and 50-56 Gy to the low-risk planning target volume. Treatment planning was made on a TomoTherapy Hi-Art 2.2.4.1 workstation. The physician and physicist simultaneously decided whether treatment planning would be executed. No more than 5% of PTV volume received more than 110% of the prescribed dose. Dose-volume constraints for OARs were utilized similarly to the previous published paper [bib_ref] Clinical observation of 73 nasopharyngeal carcinoma patients treated by helical tomotherapy: the..., Ren [/bib_ref]. The primary dosimetric parameters of main target volumes and organs of risk are shown in [fig_ref] Table 2: Planning dosimetric parameters of targets and organs at risk [/fig_ref]. During HT therapy, patients underwent MVCT guidance at least once every week. To minimize unnecessary irradiation and to reduce in-room time, the range of MVCT scans included the entire length of parotid glands and the gross tumor (slice thickness was 6 mm). The requisite time depended on the selected range and pitch and was generally about 3 minutes. The patient setup verifications were completed through the automatic and manual coregistration of the on-set MVCT images with the planning CT images based on bony and tissue anatomy.
HT's adaptive software calculated the volume and actual dose distribution according to the pretreatment MVCT scanning. The MVCT images of the first fraction were collected, followed by additional 7 fractions (fractions numbers 6, 11, 16, 21, 26, 31, and 33) for a total of 8 series of images. According to previously noted setup errors, each patient's MVCT images were merged with each patient's corresponding KVCT images using the adaptive software. The same physician manually contoured the parotid glands of each patient on the MVCT images. According to the contoured images, the actual singlefraction dose-volume histograms of the parotid gland were gained in the adaptive software. The volume and dosimetric parameters were recorded on the basis of the dose-volume histograms. The volumes of the left and right parotids were calculated 8 times and the ratios to their volumes before the first fraction were calculated for comparison. The inside and outside target volumes were also obtained. The actual doses of each single fraction including the 1 (the relative volume of the parotid gland that received 1 Gy) and 50 (half parotid gland receiving dose) were recorded. The distances between the outside borders of the bilateral parotid glands and the facial transverse diameter at the level of the odontoid process and the root of the C2 vertebral body were measured.
Spearman's correlation analysis was used to study the correlation between the two series of parameters. The parotid volume before the first fraction of radiation therapy, as measured using MVCT images, and that from the primary KVCT planning images were compared using the paired Wilcoxon rank sum test. The interfractional parameters affecting these variations were studied using repeated measures and linear regression analyses.
# Results
## Variations of parotid volume.
Each patient had 8 series of MVCT fusion images, and a total of 280 series of images were gathered for the 35 patients. There was no significant difference in parotid volume between the MVCT images before the first fraction and the KVCT images of the initial plans ( = −0.961, = 0.337). The parotid volume gradually decreased during radiation therapy [fig_ref] Figure 1: MVCT images of one patient showing parotid volume variations during radiation therapy [/fig_ref]. Before the first fraction, the volumes of the left and right parotid glands were 29.43 ± 11.6 cm 3 (12.98-65.19 cm 3 ) and 29.03 ± 10.55 cm 3 (12.80-53.11 cm 3 ), respectively. Before the last fraction, the volumes of the left and right parotid glands were 21.02 ± 11.07 cm 3 (8.70-63.77 cm 3 ) and 22.28 ± 9.67 cm 3 (7.08-51.87 cm 3 ), respectively. When measured as a percentage of the initial volume at the end of radiation therapy, the average volume reduction was 29.47% and 24.47% for the left and right parotid glands, respectively. Repeat data comparison indicated that parotid volumes varied significantly every week ( < 0.03, [fig_ref] Table 3: Variations of parotid volumes among different fractions [/fig_ref]. The rate of volume variation changed during radiation therapy, reaching its peak at the 16th fraction and later decreasing. The left and right parotid volumes had an average reduction of 0.26 cm 3 (0.92%)/treatment day and 0.22 cm 3 (0.76%)/treatment day, respectively. At the end of radiation therapy, the patients' weight lost 11.5± 5.75% (−2.94-27.59%). Body weight changes correlated with
## Displacement of parotid glands.
The left and right parotid glands shifted medially during radiation therapy. The distance between the bilateral parotid external borders was 14.60 ± 1.14 cm (12.28-17.24 cm) and 13.52 ± 1.31 cm (10.93-17.24 cm) before the first and last fractions, respectively; the average variation was −7.5 ± 3.85%. The average ratio of the intratarget volume to the extratarget volume of the left parotid gland increased from 0.28±0.19 (0.03-0.87) to 0.53 ± 0.42 (0.07-2.2). The average ratio of the intratarget volume to the extratarget volume of the right parotid gland increased from 0.26 ± 0.16 (0.01-0.63) to 0.44 ± 0.34 (0.05-1.81). The average ratio of the intratarget volume to the extratarget volume of the left and right parotid glands increased by 102.3 ± 80.05% and 86.43 ± 122.1%, respectively. [bib_ref] Clinical outcomes and prognostic factors of 695 nasopharyngeal carcinoma patients treated with..., Wang [/bib_ref]. The 1 of both parotid glands increased gradually during treatment process. The 1 of the left parotid gland was 38.19 ± 10.56% (21.26-64.08%) and 49.21 ± 12.48% (24.36-80.47%) before the first and final fractions, respectively. The 1 of the right parotid gland was 35.46 ± 9.37% (11.44-55.52%) and 44.5 ± 12.08% (23.03-69.26%) before the first and final fraction, respectively. When measured as a percentage of the initial volume at the end of treatment, the average 1 increased by 32.2% and 28.6% in the left and right parotid glands, respectively. The volume had an average increase of 0.35% (1.0%)/treatment day and 0.28% (1.06%)/treatment day for the left and the right parotid glands, respectively. Repeat data comparisons indicated that the 1 varied significantly among different fractions ( = 0.000). . The 50 of both parotid glands increased gradually during treatment process. The 50 of the left parotid gland was 0.791 ± 0.253 Gy (0.54-1.55 Gy) and 1.04 ± 0.348 Gy (0.607-2.0 Gy) before the first and final fractions, respectively. The 50 of the right parotid gland was 0.733 ± 0.143 Gy (0.509-1.13 Gy) and 0.928 ± 0.331 Gy (0.569-2.02 Gy) before the first and final fractions, respectively. When measured as a percentage of the initial volume at the end of treatment, the average 50 increased by 33.9% and 24.93% in the left and right parotid glands, respectively. The volume had an average increase of 0.77 cGy (1.0%)/treatment day and 0.6 cGy (0.78%)/treatment day for the left and right parotid glands, respectively. Repeat data comparison indicated that the 50 varied significantly among different fractions ( = 0.000).
## Variations of parotid
## Variations of parotid
## Correlation between parotid volume and dose.
During radiation therapy, there was a negative correlation between the parotid volume and the 1 ( = −0.982, = 0.000) and between the parotid volume and the 50 ( = −0.987, = 0.000).
# Discussion
As the most important large salivary glands, parotid glands secrete 60-65% of total saliva volume. After exposure to a high dose of irradiation, the secretary function of the parotid gland is impaired and saliva secretion decreases. Xerostomia thus becomes the main complication in head and neck cancer patients who have received radiation therapy [bib_ref] Parotid gland function during and following radiotherapy of malignancies in the head..., Franzen [/bib_ref]. IMRT represents a new generation of technology and, as compared with 2DCRT and 3DCRT, has better dosimetric advantages, improved conformity and uniformity of treatment targets, and better-protected OARs [bib_ref] The treatment outcome and radiation-induced toxicity for patients with head and neck..., Kouloulias [/bib_ref]. Phase III clinical trials showed that IMRT reduces the incidence of xerostomia and improves the quality of life for patients with head and neck cancer [bib_ref] Parotidsparing intensity modulated versus conventional radiotherapy in head and neck cancer (PARSPORT):..., Nutting [/bib_ref] , but severe dry mouth symptoms still sometimes occur [bib_ref] Intensity-modulated radiation therapy with or without chemotherapy for nasopharyngeal carcinoma: radiation therapy..., Lee [/bib_ref]. Xerostomia can be caused by the actual parotid dose increases that result from changes in organ anatomy, tumor size, and body weight that take place during radiation therapy, even when image-guided techniques are used. Not only can these changes cause target underdose, but also overdose to OARs can result in additional complications. To compensate for these changes, replanning can be performed during radiation therapy. Compared to the volume changes in other salivary glands, those in the parotid glands are critical, and studying the pattern of their volume changes may be helpful to decide the replanning timing. At the end of fractionated radiation therapy, parotid volume decreases, with an average volume reduction of 21.3-42% and an average reduction rate of 0.4-1.4%/day [bib_ref] Actual dose variation of parotid glands and spinal cord for nasopharyngeal cancer..., Han [/bib_ref] [bib_ref] Radiation-induced volume changes in parotid and submandibular glands in patients with head..., Wang [/bib_ref] [bib_ref] A two-variable linear model of parotid shrinkage during IMRT for head and..., Broggi [/bib_ref]. However, some studies showed that parotid volumes had small changes during the first 3-4 weeks [bib_ref] A two-variable linear model of parotid shrinkage during IMRT for head and..., Broggi [/bib_ref] [bib_ref] Comparison of treatment plans involving intensity-modulated radiotherapy for nasopharyngeal carcinoma, Xia [/bib_ref] and stabilized after the 5th week [bib_ref] Intensity-modulated radiation therapy with or without chemotherapy for nasopharyngeal carcinoma: radiation therapy..., Lee [/bib_ref]. Wang et al. [bib_ref] Radiation-induced volume changes in parotid and submandibular glands in patients with head..., Wang [/bib_ref] reported on a group of head and neck cancer (mainly consisting of oral cavity cancer) patients with postoperative radiation therapy and found that their parotid volume changes were more evident in the first 3 weeks than in the last 3 weeks. The average reduction was 20.01% and 8.57%, respectively, and the average parotid volume had no significant changes 2 and 6 months after radiation therapy as compared with those at the end of treatment. Our study found that parotid volume variation presented a linear pattern throughout radiation therapy, and the rate of volume variation reached its peak at the 16th fraction and then decreased gradually. This can be explained by the use of HT technology and different treatment protocols which combined radiation therapy with chemotherapy or anti-EGFR Mab in locoregional advanced diseases.
In addition to volume reduction, parotid glands also move to the body midline during radiation therapy. Wang et al. [bib_ref] Anatomic and dosimetric changes during the treatment course of intensity-modulated radiotherapy for..., Wang [/bib_ref] studied the parotid displacement in 15 NPC patients at the 18th fraction and found that the center of the left and right parotid glands moved to the body midline with a median motion distance of 4.8 mm and 4.3 mm, respectively. The distance between the outside boundaries of the bilateral parotid glands varied significantly at the end of radiation therapy ( < 0.001), with an average reduction of 9.2 mm (0.4-15.2 mm), while the inner boundary distance did not reach significance ( = 0.555). Robar et al. [bib_ref] Spatial and dosimetric variability of organs at risk in head-and-neck intensitymodulated radiotherapy, Robar [/bib_ref] studied the parotid anatomical changes every week during radiation therapy in 15 head and neck cancer patients and found that despite the movement of the outside boundaries to the midline (with an average of 2.6 mm and 1.9 mm for the left and right parotid glands, resp.), the parotid centers remained unchanged. Vásquez Osorio et al. [bib_ref] Local anatomic changes in parotid and submandibular glands during radiotherapy for oropharynx..., Osorio [/bib_ref] measured the parotid displacement in three-dimensional directions in 10 oropharyngeal cancer patients treated with a nonrigid registration technique. CT scanning was executed in the 23rd fraction and they found that changes in the central region of parotid glands were minimal, while changes in the peripheral region were the largest, with an average value of 1 ± 3 mm and 3 ± 3 mm, respectively. In this study, at the end of radiation therapy, the external diameters of the parotid glands obviously shrunk, and the ratios of the intratarget volume to the extratarget volume of both parotid glands increased, results which were similar to those of the previously mentioned studies that indicate that volume reduction was the main cause of parotid displacement.
During radiation therapy, parotid glands, the tumor, and surrounding tissue shrink, deform, and shift to the body midline, leading to variations of the actual parotid dose [bib_ref] Actual dose variation of parotid glands and spinal cord for nasopharyngeal cancer..., Han [/bib_ref] [bib_ref] Quantification of volumetric and geometric changes occurring during fractionated radiotherapy for head-and-neck..., Barker [/bib_ref]. Robar et al. [bib_ref] Spatial and dosimetric variability of organs at risk in head-and-neck intensitymodulated radiotherapy, Robar [/bib_ref] sketched parotid glands using weekly KVCT scanning. The initial planning parameters were transplanted to new KVCT images to form new plans. They found that the mean dose ( mean ) of the left and right parotid glands increased by 2.6 ± 4.3% and 0.2 ± 4.0% and 26 increased by 3.5% ± 5.2% and 0.3% ± 4.7%, respectively, as compared with the initial plan. In this study, the left and right parotid glands received different actual doses, suggesting that multiple factors affected parotid dose variations, such as beam distribution and tumor and metastatic lymph node locations. Using HT with MVCT scanning and adaptive software, Han et al. [bib_ref] Actual dose variation of parotid glands and spinal cord for nasopharyngeal cancer..., Han [/bib_ref] detected a single actual dose of the parotid gland in 5 NPC patients and found that parotid 50 was 83.0 ± 28.3 (53.6-151.1) cGy in the first fraction and increased to 142.6 ± 47.3 (72.2-207.9) cGy in the last treatment, which was equivalent to 177 ± 49% of that of the initial plan (97-249%, = 0.0005) and an average increase of 1.7 cGy/fraction. With similar methods, You et al. [bib_ref] Is there a clinical benefit to adaptive planning during tomotherapy in patients..., You [/bib_ref] assessed the single dose in the last week of radiation therapy in 31 head and neck cancer patients and showed that the relative volume receiving 0.75 Gy was increased by 23.6%.
We assessed the actual parotid 1 and 50 with weekly MVCT scanning. And the dose variation was less than that reported by Han et al., a difference that was probably due to this study's large number of patients, heterogeneous tumor staging, and different treatment protocols. Replanning during radiation therapy can reduce the dose to OARs and improve the dose distribution to the tumor volume. Wang et al. [bib_ref] Clinical study of the necessity of replanning before the 25th fraction during..., Wang [/bib_ref] practiced replanning before the 25th fraction in 28 NPC patients and found that, compared with the initial plan, CTV dose increased by 4.91 ± 10.89% ( = 0.024), while the max of the spinal cord and mean and 30 of the parotid gland decreased by 5.00 ± 9.23 Gy ( = 0.008), 4.23 ± 10.03 Gy ( = 0.034), and 11.47% ± 18.89% ( = 0.003), respectively. Yan et al. [bib_ref] Predictors for replanning in loco-regionally advanced nasopharyngeal carcinoma patients undergoing intensity-modulated radiation..., Yan [/bib_ref] recommended replanning after 20 fractions of treatment.
In summary, during the IMRT of NPC, some patients' parotid volumes and locations varied significantly, generally causing an increase of the actual delivered dose. It is thus necessary to identify relevant factors that affect these changes.
Our study suggests that replanning is appropriate in the fourth week of IMRT.
[fig] Figure 1: MVCT images of one patient showing parotid volume variations during radiation therapy (the number on each image is the fraction number). that of parotid volume ( = 0.418, = 0.012). At the last fraction of treatment, patients' facial diameter was decreased by 9.49±3.94% (1.66-17.1%), without correlation with parotid volume changes ( = 0.236, = 0.172). [/fig]
[table] Table 1: Patients' clinical characteristics. [/table]
[table] Table 2: Planning dosimetric parameters of targets and organs at risk. 30 : the relative volume of the organ receiving 30 Gy; 40 : the relative volume of the organ receiving 40 Gy. [/table]
[table] Table 3: Variations of parotid volumes among different fractions (pairwise comparisons). [/table]
|
Clinical Outcomes of In-office Sutureless Amniotic Membrane Transplantation in Persistent Epithelial Defect
Purpose: To investigate the efficacy of outpatient clinic-based sutureless amniotic membrane transplantation (AMT) along with therapeutic contact lens (T-lens) application in eyes with persistent epithelial defects (PED).Methods: Nine eyes of nine patients (mean age, 71.7 ± 5.2 years) diagnosed with PED and treated with in-office sutureless AMT combined with T-lens application were retrospectively reviewed. Demographics, systemic diseases, PED etiology, corneal epithelial defect size, visual acuity, corneal scraping culture results, and clinical course were evaluated.Results: Among nine eyes with PED, three had neurotrophic keratopathy, four had infectious keratitis (three with fungal keratitis and one with bacterial keratitis), one had limbal deficiency, and one had marginal keratitis. The mean epithelial defect size (calculated as an average of the horizontal and vertical diameters) was 3.13 ± 1.42 mm, and the mean duration from AMT to epithelial healing was 30.1 ± 10.5 days (range, 14-51 days) in successful trials. The success rates were 77.8% (7/9) per patient and 66.7% (8/12) per trial. The causes of failure in two patients were AMT displacement and uncontrolled infection.Conclusions:Our results demonstrate that in-office sutureless AMT combined with T-lens application can be used in patients with PED who are refractory to medications. It will be especially helpful for elderly patients because of its easy-to-use method. To achieve successful outcomes with AMT, an appropriate periocular environment as well as infection control need to be considered.
mechanical injuries including exposure keratopathy, and inflammatory conditions such as keratoconjunctivitis sicca, rosacea, infectious keratitis, peripheral ulcerative keratitis, and autoimmune diseases [bib_ref] Persistent corneal epithelial defects: a review article, Vaidyanathan [/bib_ref]. Regardless of the cause, when damaged corneal epithelium is not managed properly, it can progress to corneal ulcer, conjunctivalization, and neovascularization, ultimately resulting in corneal opacity and permanent visual loss [bib_ref] Woundhealing studies in cornea and skin: parallels, differences and opportunities, Bukowiecki [/bib_ref]. Even with prompt diagnosis and supportive medical treatment, the visual prognosis for PED is poor due to slow recovery [bib_ref] Autologous serum 50% eyedrops in the treatment of persistent corneal epithelial defects, Jeng [/bib_ref]. Therefore, additional surgical treatments such as tarsorrhaphy, conjunctival flap, and amniotic membrane transplantation (AMT) may be considered for improved results with regard to promoting regeneration of corneal epithelium in PED from various corneal diseases.
AMT has been used in various corneal diseases since its first usage in ophthalmology by de Rotth [bib_ref] Plastic repair of conjucntival defects with fetal membrane, De Rotth [/bib_ref] in 1940 for conjunctival loss repair; evidence of its efficacy in treating corneal ulcers has been further provided by Kim and Tseng [bib_ref] The effects on inhibition of corneal neovascularization after human amniotic membrane transplantation..., Kim [/bib_ref]. Briefly, the amniotic membrane derived from the inner layer of the placenta is composed of an avascular stromal layer and a basement membrane. The amniotic membrane serves as a frame for the migration of corneal epithelium and contains many proteins and anti-inflammatory cytokines that prevent epithelial apoptosis [bib_ref] Amniotic membrane transplantation in infectious corneal ulcer, Kim [/bib_ref] [bib_ref] Amniotic membrane transplantation for conjunctival surface reconstruction, Tseng [/bib_ref] [bib_ref] The human amnion is a site of MHC class Ib expression: evidence..., Houlihan [/bib_ref] [bib_ref] Surgical reconstruction of the ocular surface in advanced ocular cicatricial pemphigoid and..., Tsubota [/bib_ref] [bib_ref] Amniotic membrane transplantation with or without limbal allografts for corneal surface reconstruction..., Tseng [/bib_ref] ; furthermore, it does not trigger immunologic rejection [bib_ref] Amniotic membrane transplantation for conjunctival surface reconstruction, Tseng [/bib_ref] [bib_ref] The human amnion is a site of MHC class Ib expression: evidence..., Houlihan [/bib_ref]. Conventional amniotic membrane patches (temporary AMT) or grafts (permanent AMT) combined with traditional suture techniques have been proven to be effective in treating various corneal diseases [bib_ref] Single and multilayer amniotic membrane transplantation for persistent corneal epithelial defect with..., Prabhasawat [/bib_ref] [bib_ref] Evaluation of amniotic membrane transplantation as an adjunct to medical therapy as..., Tamhane [/bib_ref] [bib_ref] Amniotic membrane transplantation for acute chemical or thermal burns, Meller [/bib_ref]. However, in contrast to sutureless AMT, these techniques often necessitate the use of an operating room. Sutureless AMT has recently been proposed as a possible substitute for conventional AMT [bib_ref] Sutureless amniotic membrane transplantation for severe bacterial keratitis, Sheha [/bib_ref] [bib_ref] Sutureless amniotic membrane transplantation for partial limbal stem cell deficiency, Kheirkhah [/bib_ref] [bib_ref] Temporary sutureless amniotic membrane patch for acute alkaline burns, Kheirkhah [/bib_ref] [bib_ref] Sutureless amniotic membrane ProKera for ocular surface disorders: short-term results, Suri [/bib_ref] [bib_ref] Sutureless dehydrated amniotic membrane for persistent epithelial defects, Mimouni [/bib_ref] ; this procedure can be easily performed on an outpatient basis and is thus advantageous for elderly patients who have difficulty with complicated processes and preoperative preparations. To the best of our knowledge, only a few studies have evaluated the outcomes of in-office sutureless AMT in PED, and research has been especially sparse with regard to dehydrated amniotic membranes (as compared with cryopreserved amniotic membranes). In the current study, we investigated the potential role of in-office sutureless AMT combined with therapeutic contact lens (T-lens) application for treating PED in various corneal diseases by evaluating the factors affecting treatment success and failure.
# Materials and methods
This study was approved by the institutional review board of Seoul National Bundang Hospital (No. B-2105-682-104) and was performed in accordance with the tenets of the Declaration of Helsinki and its later amendments. Informed consent was exempted for the current study with the approval of the institutional review board of Seoul National Bundang Hospital because of the investigation's minimal risk to participants.
We retrospectively reviewed the medical records of consecutively presenting patients with PED in various corneal diseases who underwent sutureless AMT with T-lens application in an outpatient clinic at Seoul National Bundang Hospital between March 2018 to March 2020. PED diagnoses were confirmed by a corneal specialist (HSJ) using slit lamp microscopy with fluorescein staining. As an inclusion criterion, the definition of PED was as follows: the size of the epithelial defect (ED) either decreased minimally or did not decrease at all despite appropriate conventional treatment for over a week, the margins were heaped up, and the underlying subepithelium or anterior stroma showed opacification. To effectively analyze the factors associated with treatment failure for AMT, we did not exclude any failed cases. Conventional medical treatments prior to AMT for the patients enrolled in our study mainly comprised preservative-free artificial tears and empirical or targeted antimicrobial medications according to antibiotic susceptibility, if causative microbial culture results were confirmed [fig_ref] Table 1: Demographics, underlying ocular diseases, clinical features and outcomes of patients with persistent... [/fig_ref].
The amniotic membrane used in this study was the Am-bioDisk (MiMedx Group, Marietta, GA, USA), which is a dehydrated amniotic membrane graft designed for ophthalmic in-office application. The thickness of the Ambio-Disk is approximately 35 µm and it has a round shape that is retained even after pinching with mircroforceps. This morphologic feature differs from that of conventional amniotic membranes (which are non-self-retaining and are paper attached) and other cryopreserved sutureless amniotic membranes (i.e., those adhering to a polymethyl methacrylate ring), enabling AmbioDisk to be used more conveniently and effectively in outpatient clinic settings .
In the current study, after gently applying the Ambio-Disk to the cornea with McPherson forceps, a T-lens (Acuve Oasys; Johnson & Johnson Vision, Jacksonville, FL, USA) was additionally applied over the AmbioDisk in order to stabilize it. Care was taken to dehydrate the ocular surface prior to the removal of the speculum and patients were instructed to apply only one drop of each eye drop solution in order to prevent dislodgment of the amniotic membrane from the ocular surface. Although the T-lens (Acuve Oasys) is generally recommended for up to 2 weeks of continuous use, we maintained the T-lens in order to prevent the amniotic membrane from dislodging until we removed the T-lens after epithelial healing. Representative serial anterior segment photographs taken from the time of AMT until epithelial healing are presented in . After AMT treatment, the use of artificial tears and topical prophylactic antibiotics is recommended three or four times a day in noninfectious cases. If infectious causes could not be ruled out, we maintained the use of pre-AMT medications. Before confirming microbial culture results, frequent application of empirical topical antibiotics (ranging from four times a day to every hour) was recommended. However, if the frequent use of medication was considered harmful for maintaining amniotic membranes, less frequent use of topical treatment was recommended in order to maintain the stability of the amniotic membrane.
The following information was collected from medical records: age, sex, systemic diseases, past ocular history (including PED etiology, corneal ED size, visual acuity before and after AMT, corneal scraping culture results, dura-tion from PED detection to treatment, and duration of healing from PED), and clinical outcomes. The success and failure of each trial of AMT with T-lens application was evaluated according to the following outcome criteria: AMT was regarded as successful when complete epithelial healing was achieved without worsening signs and was regarded as a failed treatment when epithelial healing was inadequate. In these cases, either additional therapy was recommended, or AMT with T-lens application was deemed intolerable due to various causes (and follow-up treatment was hence not recommended), regardless of corneal epithelial status.
# Results
The demographics, underlying ocular diseases, clinical [fig_ref] Table 1: Demographics, underlying ocular diseases, clinical features and outcomes of patients with persistent... [/fig_ref]. A total of 12 trials in nine eyes (among nine patients) were reviewed in this study. Four male patients and five female patients with a mean age of 71.7 ± 5.2 years (range, 61-84 years) at the time of AMT with T-lens application were included. The PED etiologies of the enrolled patients included neurotrophic keratopathy (three eyes), infectious keratitis (fungal, three eyes; bacterial, one eye), limbal deficiency related to contact lenses (one eye), and marginal keratitis (one eye). Among the nine eyes, two (cases 1 and 9) underwent more than one trial of AMT with T-lens therapy due to recurrence of PED (case 1) and AMT displacement (case 9).
The mean size of corneal ED, which was calculated as the average of the horizontal and vertical diameters measured at the time of AMT with T-lens application, was 3.13 ± 1.42 mm (range, 0.70-8.25 mm). The mean duration from the detection of ED to AMT with T-lens application was 17.5 ± 9.7 days (range, 7-37 days). Three trials of AMT with T-lens application were performed in case 9 due to AMT displacement; the follow-up period between AMT application and displacement was 3 days in the first trial and 4 days in the second trial, and both trials occurred over the course of <7 days. Hence, the duration of pre-AMT conventional therapy was calculated as a single trial. Anterior segment photographs before and after sutureless AMT with T-lens application in both successful and failed treatment cases are presented in [fig_ref] Figure 3: Anterior segment photographs from patients with persistent epithelial defect [/fig_ref] and [fig_ref] Figure 4: Anterior segment photographs from patients with persistent epithelial defect [/fig_ref] -4F, respectively.
The success rates of AMT with T-lens application within PED were 66.7% (8/12) per trial and 77.8% (7/9) per patient. The mean maintenance duration of AMT with T-lens application was 27.1 ± 13.0 days (range, 12-51 days) and the mean epithelial healing time following AMT was 30.1 ± 10.5 days (range, 14-51 days) in successful trials.
In case 7, a patient with neurotrophic keratopathy related to the long-standing use of diclofenac sodium, topical steroids, and glaucoma medications (due to uncontrolled macular edema in uvetic glaucoma) complained of pain after AMT, and AMT with T-lens application was ineffective after 4 days. We hence judged that AMT with T-lens application was ineffective in this case. In case 8, although the corneal epithelium was healed 10 days after AMT treatment, conjunctival injection continued, indicating an uncontrolled infection. This case was thus regarded as a failed treatment according to the pre-specified outcome criteria, and the use of the T-lens was discontinued thereafter. Case 9 presented with exposure keratopathy related to a history of lid reconstruction with reverse Hughes flap surgery due to sebaceous gland carcinoma. Although we recognized that this patient's infectious keratitis was not controlled, we tried AMT with T-lens application with the hope that the PED would improve; however, AMT could not be maintained until the third trial of AMT with temporary tarsorrhaphy.
# Discussion
AMT has long been recognized as an effective treatment option for PED. AMT with conventional suture technique has already been demonstrated to be effective in treating various corneal diseases (especially PED), with success rates ranging from 64% to 91% [bib_ref] Amniotic membrane transplantation for persistent epithelial defects with ulceration, Lee [/bib_ref] [bib_ref] Amniotic membrane inlay and overlay grafting for corneal epithelial defects and stromal..., Letko [/bib_ref] [bib_ref] Contact lens-related infectious keratitis, Palmer [/bib_ref]. However, the conventional suture technique is an invasive and time-consuming procedure, thus necessitating the development of a new, less invasive, and simpler method. With the development and understanding of AMT processing, in-office sutureless AMT has been proposed as a substitute for the conventional suture technique. This treatment modality is divided into cryopreserved and dehydrated AMT.
In 2013, Suri et al. [bib_ref] Sutureless amniotic membrane ProKera for ocular surface disorders: short-term results, Suri [/bib_ref] confirmed success rates of 44% in nonhealing corneal ulcers (nine eyes) and 64% in neurotrophic keratitis (11 eyes) using cryopreserved in-office sutureless AMT (ProKera; Bio-Tissue, Miami, FL, USA). Recently, Mimouni et al. [bib_ref] Sutureless dehydrated amniotic membrane for persistent epithelial defects, Mimouni [/bib_ref] reported favorable outcomes for in-office sutureless dehydrated AMT (BioDOPTIX; Labtician Ophthalmics, Oakville, ON, Canada) in nine eyes with PED of various etiologies, with an overall success rate of 89% (8/9 eyes). In our study, the success rates of in-office sutureless AMT with T-lens application (per trial) were 75.0% for neurotrophic keratitis and 50% for nonhealing corneal ulcers in infectious keratitis; our results hence varied according to PED etiology and were similar to the favorable outcomes observed in a study of neurotrophic keratopathy conducted by Suri et al. [bib_ref] Sutureless amniotic membrane ProKera for ocular surface disorders: short-term results, Suri [/bib_ref]. A patient with limbal deficiency related to contact lenses and marginal keratitis who had underlying autoimmune hepatitis and thyroid disease also demonstrated successful results at their first AMT trial. The lower overall success rate observed in our study as compared to the study by Mimouni et al. [bib_ref] Sutureless dehydrated amniotic membrane for persistent epithelial defects, Mimouni [/bib_ref] could be related to differing causes of PED among the included cases; for example, nearly half of the PED patients included in our study presented with infectious keratitis. Although sutureless AMT with T-lens can be applied easily and efficiently in elderly patients with PED (regardless of etiology) on an outpatient basis, we suggest that this treatment may be more effective for treating PED with noninfectious causes, such as neurotrophic keratopathy.
The mean PED healing duration in our study was 30.1 ± 10.5 days (range, 14-51 days) within successful trials. A previous study by Suri et al. [bib_ref] Sutureless amniotic membrane ProKera for ocular surface disorders: short-term results, Suri [/bib_ref] reported that the mean [bib_ref] Sutureless dehydrated amniotic membrane for persistent epithelial defects, Mimouni [/bib_ref] reported a mean PED resolution time of 17.8 ± 9.6 days (range 7-35 days). The longer healing duration observed in our study could be attributed to the higher number of cases with infectious causes as well as the different ED sizes included in the current investigation.
We evaluated characteristic causes of AMT failure in failed trials. The causes of treatment failure in our study were AMT displacement and uncontrolled infection. T-lenses are used indispensably in in-office sutureless AMT in order to stabilize the amniotic membrane, and this could aggravate corneal infection by serving as a bed for microbial infection [bib_ref] Contact lens-related infectious keratitis, Palmer [/bib_ref]. Based on our results, we believe that more careful observation of the signs and symptoms of ocular infection, as well as an additional close follow-up with strict usage of antibiotics or antifungals, is essential to proper treatment. If signs of infection are observed, such as conjunctival injection or aggravation of eye pain, AMT and T-lens application should be discontinued promptly.
Early displacement of the amniotic membrane is a major problem even with the conventional suture method and the use of protective contact lenses [bib_ref] Amniotic membrane inlay and overlay grafting for corneal epithelial defects and stromal..., Letko [/bib_ref]. The displacement rate is an important factor as well, especially in sutureless AMTs with T-lenses, and is expected to be inferior to that observed within the conventional suture method. In our cases, the overall displacement rate was 25% (3 out of 12 trials). Considering other complications related to sutures as well as the accessibility of the operation, this displacement rate seems tolerable. Case 9 had a history of lid reconstruction surgery and PED due to uncontrolled bacterial keratitis; the patient finally succumbed after temporary tarsorrhaphy in the third trial. Since the periocular environment could be an important factor for sutureless AMT stabilization, a careful assessment of the periocular environment as well as appropriate pre-AMT intervention should be performed prior to sutureless AMT with T-lens application if necessary.
Mimouni et al. [bib_ref] Sutureless dehydrated amniotic membrane for persistent epithelial defects, Mimouni [/bib_ref] reported on patients treated with de-hydrated sutureless AMT; specifically, BioDOPTIX for PED demonstrated complete absorption of the dehydrated amniotic membrane within less than 2 weeks in that study. The researchers explained that the shorter absorption time compared to that of ProKera (which has been reported to take up to 125 days to absorb) could be related to the reduced thickness of the dehydrated BioDOPTIX (40 µm) as compared to the cryopreserved ProKera (100-200 μm) [bib_ref] Sutureless dehydrated amniotic membrane for persistent epithelial defects, Mimouni [/bib_ref] [bib_ref] Evaluation of the role of ProKera in the management of ocular surface..., Pachigolla [/bib_ref]. In our study, all cases except two cases with spontaneous displacement of AMT with T-lens application had their AMT with T-lens application removed by a study author (HSJ) after epithelial healing was detected; the amniotic membrane was maintained until the T-lens was removed in all these cases. We considered that the absorption time may be longer than the duration of T-lens treatment (27.1 ± 13.0 days; range, 12-51 days) within our case series. We present representative anterior segment photographs of two cases before, during, and after AMT in . As the amniotic membrane (AmbioDisk) is thin and translucent, the configuration of the membrane under the T-lens cannot be fully observed accurately. We expect similar absorption time between AmbioDisk and BioDOPTIX as the thickness of the amniotic membrane is approximately 40 µm for both treatment modalities. Features such as thickness and absorption time may be important considerations for choosing amniotic membrane types according to the ocular environment. Further investigation regarding absorption time for the AmbioDisk within a larger number of cases is necessary in future research. This study has some limitations in addition to its substantial strengths. First, the number of cases enrolled in this study was small, our study enrolled PED cases with heterogenous causes, and there was no control group. These limitations are inherent to all retrospective studies. Since PED is a rare condition, it is diff icult to plan well-controlled studies evaluating PED treatments. However, our case series, along with other prior studies on sutureless AMT will be informative to clinicians encountering patients with refractory PED. Additional case series will aid in the establishment of indications for the treatment of PED, including based on size, thus allowing this expensive treatment to be performed more successfully. Second, in-office sutureless dehydrated AMT inevitably requires a T-lens in order to stabilize the amniotic membrane, which makes it difficult to distinguish whether T-lenses or AMTs with T-lenses are affecting epithelial healing. Epithelial healing may be a protective effect of T-lenses; however, since we included PED cases that were refractory to conventional treatment for over a week, our findings suggest that the amniotic membrane may have additive anti-inflammatory and epithelial healing promotive effects. To make study results more accurate and distinguishable, a further comparative case-control study with a T-lens control group (versus AMT with a T-lens group) will be necessary in future research efforts. Third, this study did not include accurate information regarding amniotic membrane melting time. Further, we did not exchange T-lenses until the amniotic membrane melted for the purpose of stabilizing the amniotic membrane. Long-standing use of a single T-lens could increase the chance of infection by providing a microbial bed and could thus be harmful for patients if not applied with adequate consideration of the amniotic membrane melting time. Studies evaluating more accurate timing with regard to amniotic membrane melting time may be helpful for reducing infection risk by providing adequate follow-up time as well as minimizing T-lens continuation.
We believe that in-office sutureless AMT is particularly helpful for elderly patients who often have difficulty with the preoperative general examination to enter the operating room while aiding in economic and efficiency optimization. Moreover, neurotrophic keratopathy (which is a main cause of PED) is prevalent in elderly patients and is a good indication for sutureless AMT. However, we conclude that this treatment modality cannot replace conventional amniotic membrane patches and grafts as it may be difficult to maintain due to the influence of the surrounding ocular environment (eyelids or conjunctiva). There may also be a possibility that the sutureless membrane will move with blinking. According to our experience, we recommend that in-office sutureless AMT be used in cases where the defect size is less than half of the cornea, there is no eyelid defect, and infection is under control.
In conclusion, the results of our case series demonstrate that in-office sutureless AMT combined with T-lens application may be a simple and effective treatment in PED patients refractory to medications, particularly among elderly patients. An appropriate periocular environment (including consideration of the eyelids) and infection control are important to achieving a successful in-office sutureless AMT with T-lens application in refractory cases of PED.
[fig] Figure 1, Figure 2: Pictures of an amniotic membrane used in treating corneal diseases. This figure presents (A) cryopreserved conventional Ami-noGraft (Bio-Tissue, Miami, FL, USA), (B) cryopreserved sutureless ProKera (Bio-Tissue), and (C) dehydrated sutureless AmbioDisk (MiMedx Group, Marietta, GA, USA) amniotic membranes. In treatment with ProKera, an amniotic membrane is adhered to a polymethyl methacrylate ring, whereas the Ambiodisk has a self-retaining form. The images are provided courtesy of (A,B) Bio-Tissue and (C) Katena (Parsippany, NJ, USA), respectively. Representative serial anterior segment photographs from two patients (A-E, case 6; F-J, case 1) with persistent epithelial defect (A,B,F,G) before, (C,H) 5 days, and (D,I) 1 month after sutureless amniotic membrane transplantation with therapeutic contact lens application. (E,J) Remnant amniotic membranes and therapeutic contact lenses were removed 1 month after amniotic membrane transplantation. and outcomes of the patients with PED are summarized in [/fig]
[fig] Figure 3: Anterior segment photographs from patients with persistent epithelial defect (A-G) before and (H-N) after in-office sutureless amniotic membrane transplantation treatment in successful cases; (A,H) case 1-1, (B,I) case 1-2, (C,J) case 2, (D,K) case 3, (E,L) case 4, (F,M) case 5, and (G,N) case 6 are presented here. [/fig]
[table] Table 1: Demographics, underlying ocular diseases, clinical features and outcomes of patients with persistent ED treated with AMT combined with T-lens application [/table]
|
Review of Laser Raman Spectroscopy for Surgical Breast Cancer Detection: Stochastic Backpropagation Neural Networks
Laser Raman spectroscopy (LRS) is a highly specific biomolecular technique which has been shown to have the ability to distinguish malignant and normal breast tissue. This paper discusses significant advancements in the use of LRS in surgical breast cancer diagnosis, with an emphasis on statistical and machine learning strategies employed for precise, transparent and real-time analysis of Raman spectra. When combined with a variety of "machine learning" techniques LRS has been increasingly employed in oncogenic diagnostics. This paper proposes that the majority of these algorithms fail to provide the two most critical pieces of information required by the practicing surgeon: a probability that the classification of a tissue is correct, and, more importantly, the expected error in that probability. Stochastic backpropagation artificial neural networks inherently provide both pieces of information for each and every tissue site examined by LRS. If the networks are trained using both human experts and an unsupervised classification algorithm as gold standards, rapid progress can be made understanding what additional contextual data is needed to improve network classification performance. Our patients expect us to not simply have an opinion about their tumor, but to know how certain we are that we are correct. Stochastic networks can provide that information.
# Introduction
Breast cancer is the second most commonly diagnosed cancer among American women after skin cancer, with approximately 1 in 8 women (12.8%) receiving a diagnosis of invasive breast cancer within their lifetime. Early detection of tumors has become possible due to technological advancement of screening techniques as well as public education and awareness. The standard of care for early stage invasive breast cancer includes breast irradiation and breast conserving surgery (BCS), during which the surgeon attempts to excise all tumors with a negative margin, that is margins of the resected tissue have no evidence of tumor. Current gold standard definition of a negative surgical margin is defined as no malignant cells at the surface of resected tissue specimen. Large studies have been conducted to compare outcomes between BCS and traditional mastectomy procedures, showing equivalent overall survival rates. Mastectomy procedures include undesirable cosmetic outcomes for the patient, increased psychological burden and increased wound infection risk. Unfortunately, the biggest risk with BCS is the chance of local reoccurrence (LR) with 15% to 35% of patients who opt for BCS requiring a second surgery to obtain negative margins. Current standard of care after BCS includes a pathological investigation of tumor margins using hematoxylin and eosin (H&E) stains. During this process the excised tissue is fixed, processed, sectioned and stained with H&E and studied under a microscope. This process is time and labor intensive, with results only becoming available days after initial surgery. If the surgical pathologist finds positive or close to positive margins, the patient has to undergo a second surgery. Hence, the accurate detection of tumor margins during surgery has become an area of intense investigation and technology development. Some intraoperative modalities that are under investigation to tackle this problem include radiofrequency spectroscopy, bioimpedance spectroscopy, photoacoustic tomography, spatial frequency domain imaging (SFDI), fluorescence imaging, elastic scattering spectroscopy, microscopy with ultraviolet surface excitation (MUSE), light-sheet microscopy, nonlinear microscopy, optical coherence tomographyand laser Raman spectroscopy (LRS). Maloney et al. review all of these intraoperative techniques as well as pre-operative imaging technologies for margin detection during BCS. Their recommendation is a hybrid system of imaging modalities and optical scanning techniques to achieve desired sensitivity and specificity required for margin detection. In recent years, many investigators have pointed out the advantages of using a highly specific biomolecular probe, laser Raman spectroscopy (LRS), to distinguish breast cancer tumors from healthy breast tissue and benign masses.
This article focuses on LRS as a real-time tool for breast tumor detection during surgery. LRS is a non-destructive and label free technique that harnesses the biochemical specificity of the Raman effect. A tiny fraction of incident light on an object is absorbed by its molecules which are set into vibrational motion and the subsequent frequency change of the scattered light is equal to the vibrational frequency of the these molecules. The metric for Raman spectroscopy is known as the Stokes shift, named after Irish physicist George Gabriel Stokes. In the mid-19th century, Stokes noted that while most light incident on matter was either absorbed or reflected back at the same wavelength as the illuminating light (Rayleigh scatter) some light came back from the interaction at a longer wavelength; he coined the term "fluorescence" for this lower energy light. This shift in wavelength to lower energy state occurs in both fluorescence and Raman scattering and was ultimately referred to as a Stokes shift in his honor. The origin of the Stokes shift is commonly represented in a Perrin-Jablonski diagram. The energy transfer occurs between the incident photons and the electrons of the target atoms. If the electrons are in the ground state a Stokes shift occurs with a loss of photon energy. If the electrons are in an excited state, the photons can gain energy in what is termed an anti-Stokes shift. This scattered light is collected by a spectrometer, creating a spectrum, which showcases a series of peaks, also referred to as bands, corresponding to the characteristic vibrational frequencies of the scattering molecules. This creates a unique biomolecular 'fingerprint' of the sample. When applied to cancer diagnostics, this unique fingerprint can be a direct indicator of the inherent biochemical differences between malignant and healthy tissue. Such biochemical specificity obtained from the surgical field can be instrumental in helping a surgeon achieve negative margins during BCS.
This paper discusses the most significant research advancements in the use of LRS as a real-time in vivo tool for surgical breast cancer detection, with an emphasis on the statistical and machine learning methods employed for spectral classification. The choice and implementation of machine learning diagnostic algorithms is critical for precise, real-time and transparent spectral classification. We start by describing various statistical and machine language algorithms commonly used for LRS data processing; followed by a brief review of multiple efforts using LRS to diagnose breast cancer using these algorithms. We will point out the variety of statistical analysis and classification techniques employed in these studies with emphasis on the absence of a standard methodology for (1) spectral data collection, (2) signal pre-processing, and (3) final classification. We posit that agreement on those three steps will be critical to the creation of national database development for the use of LRS in biomolecular breast cancer diagnostics. We suspect that stochastic neural networks will become the classification strategy of choice in oncogenic diagnostics due to their ability to (1) identify complex, non-linear boundaries between classes; (2) produce probabilistic estimates of the likelihood their predictions are correct; and, (3) estimate the expected error in their own prediction. We posit that (1) the majority of so called "machine learning" algorithms are, in fact simply statistical tools; and, (2) that stochastic backpropagation algorithms are the only true learning algorithm inherently providing the two most critical pieces of information: the Bayesian probability of correct classification and an estimate of the certainty of each classification probability for each target.
## Statistical and machine learning methods
The preparation and use of LRS data proceeds in most studies through several steps generally summarized as (1) pre-process cleaning of the raw incoming signal; (2) identification and extraction of the portion(s) of the data containing the lion's share of the information (dimensionality reduction); (3) the pure unsupervised preliminary clustering of the LRS target sites driven only by the extracted features without human bias; and (4) a formal classification of all targets using a gold standard for training. The gold standard can be either the output of an unsupervised clustering algorithm, or the opinion of a human expert after evaluating other data, e.g., H&E photomicrographs of the LRS targets, or, preferably, both.
Until recently, a fifth stage has often been omitted or minimized: rigorously estimating the likelihood that the putative classification decision is correct and then evaluating the replicability of that decision. We will briefly describe several algorithms commonly employed in each of the first four tasks and then describe an easily implemented stochastic neural network method to generate both a Bayesian probabilistic classification as well as error estimation for classification replicability.includes a flowchart of these five data processing steps as well as commonly used algorithms for each step. The data processing steps required for robust Raman spectral analysis with the various algorithms that can be used for each step.
## Data pre-processing
Raman scattering is an example of an inelastic scattering of a photon after it interacts with the electron cloud generated by a wide variety of molecular bonds, e.g., C-H, C-H 2 , C-H 3 , -OH, etc. The change in energy is recorded in the Raman spectral data as a Stokes shift in the photon's wavelength towards the red end of the spectrum (lower frequency, lower energy) if the electron cloud is in the ground state, and toward the blue end (higher frequency, higher energy) if the molecular bond has been excited to a higher energy state by some previous event. The molecular specificity of an LRS spectrum comes with a price. Very few laser photons striking a target will carry the vibrational energy signature of one of the molecular bonds in the target. The vast majority will be either scattered elastically without carrying vibrational information (Rayleigh scatter) or absorbed to be subsequently dissipated as heat or fluorescence.
Since the Raman effect is such a rare event, to minimize data acquisition time highly sensitive CCD and CMOS chips and thermal cooling systems are now the core sensor components for collecting Raman photons present at just above sensor thermal background levels. The sensitivities are such that an LRS spectrometer makes an excellent cosmic ray detector. Signal pre-processing starts in modern systems with a series of embedded software routines that look for spike events involving only one or two spectral bins which are characteristic of cosmic ray events. Once a spike is detected, the embedded algorithm then uses any one of a number of nearest neighbor routines to interpolate and replace the affected pixels. The next step in most pre-processing protocols is usually called noise reduction or smoothing. The most common choice of algorithm for the past three decades has been that devised by Savitsky and Golay. The algorithm uses a moving window based local polynomial fitting routine requiring operator decisions about parameters such the size of the moving window and polynomial order. The inherent danger in the smoothing is that as the moving window size increases, some Raman bands of interest with broad FWHM (full width half maximum) and low S/N characteristics may be lost. The third stage of preprocessing is the removal of the red-shifted fluorescence and other continuum distortions arising from variables such as changes in sample absorption characteristics as a function of wavelength, laser bleed-through, or CCD inhomogeneities. The baseline corrections that must be accomplished without distorting the information in the raw data, are quite critical, but relatively simple to automate. A wide variety of techniques have been investigated including wavelets, FFTs, first and second derivatives, and polynomial fitting. At present the most commonly implemented baseline correction techniques involve some form of polynomial fitting. An easily implemented method uses an asymmetrically reweighted penalized least squares smoothing algorithm (arPLS) developed by Baek et al.. This method employs an iterative process to estimate a baseline. The calculated baseline is then subtracted from the intensity values of the raw spectrum to remove the continuum.
The final pre-processing step, normalization, attempts to minimize the impact of variables in the data collection process that are independent and extraneous to the hypothesis driving the LRS investigation. Common confounding variables include changes in sample temperature, ambient lighting, laser power drift, and CCD/CMOS chip temperature. The two most common normalization procedures are peak normalization and vector normalization. In peak normalization a single peak, one predicted to be unchanging in the experimental conditions, is chosen as a gold standard. All peaks, including the normalizing peak, are then divided by the intensity of that normalizing peak. For a given set of experiments, evaluation of the entire data set using principal component analysis (PCA, defined below) can provide confirmation that the normalizing peak does not contribute to the information content of the LRS spectra. But that confirmation does not generalize beyond the bounds of the original experiment. In vector normalization, the square root of the sum of the squared intensities of the spectrum is calculated. Then, each of the Raman spectral intensities is simply divided by this 'norm' to obtain a normalized spectrum. Vector normalization is becoming increasingly popular and may be more useful when trying to interpret the results of disparate LRS measurements performed under a variety of experimental conditions.
## Data optimization and dimension reduction
At present it has become common practice for data analysis to begin with optimization of the raw input data using principle component analysis (PCA). PCA, also known as the Karhunen-Loeve or hotelling transform, is a classical unsupervised multivariate analysis technique. Often incorrectly thought of as a classification algorithm, PCA is actually the premier method for reducing the dimensionality of the raw data. In the case of LRS, the raw spectral data may contain~2000 bands. PCA will often be able to reduce that enormous input vector to as few as 3-10 principal components or factors. PCA identifies linear combinations of raw parameters that account for the maximum variance. It is important to note that in machine learning and artificial intelligence, variance is often characterized as the information content in a data set. Analysis of variance (ANOVA) can be used to confirm PCA detection of high information content in specific regions of LRS spectra. The data dimensionality reduction significantly increases the computational efficiency of classification algorithms and makes it more likely that sophisticated, complex algorithms such as neural networks will not "memorize" the training set data, but will instead extract more robust correlations that can be applied to classify previously unseen data sets. Finally, PCA is unaffected by modest amounts of noise in the data since the covariance matrix is an average over many data points while the noise is uncorrelated from point to point.
Partial least squares (PLS) bears considerable resemblance to PCA. However, while PCA concentrates only on modeling the y-axis of a data set, PLS simultaneously models the structure of both the x and y axes. PLS in a form known as PLS-regression (PLSR) is used extensively in chemometrics. PLSR usefulness is its ability to analyze data containing noisy, collinear, and even incomplete variables in both x and y.
## Unsupervised, autonomous data exploration
Machine learning algorithms generally fall into one of two categories: unsupervised and supervised. Unsupervised algorithms identify patterns in a data set without prior knowledge of independent data that might provide classification criteria easily interpreted by a human expert. Supervised algorithms are provided with the predicted classification of either a human expert or the results of an unsupervised algorithm. The supervised algorithm then must predict the likelihood the clustering or classification model proposed by either the unsupervised or human expert is correct.
When initially exploring any unknown environment, it is certainly desirable to identify self-organizing patterns in the raw data. The exercise can quickly point out the existence of anomalies in the data such as outliers, measurement results unexpected by the experiment and distant from the spectral patterns generated by all other samples. The human or algorithmic analysis of these readings is critical for proceeding with data processing since there are only two primary reasons for the existence of an outlier. The common assumption is that this is a measurement error and the offending data should be thrown out. The other possibility, of course, is that this is a serendipitous detection of an unexpected phenomenon and may be the most important finding in the experiment. Both unsupervised artificial neural networks and factor analysis techniques have been employed successfully for such an initial qualitative exploration. In the data flow illustrated in, PCA factors extracted from raw data are often used as inputs for two unsupervised clustering algorithms: hierarchical cluster analysis (HCA) and k-means. HCA is a completely autonomous cluster detection algorithm used extensively in molecular biology, medicine, and biochemistry to explore genome data sets. Dendritic tree HCA initially considers each member of the data set as a separate cluster and then combines the clusters by merging nearest neighbors. Repeated with the resulting combined data points the binary combination continues, reducing the number of clusters at each step, until only one cluster remains. The HCA data partitioning is then represented by a dendrogram, or a tree where each step in the clustering process is illustrated by a node or branching of the tree. The distance between each level is calculated, and significance levels can be set deciding if the data best described by 1, 2, 3,..., N clusters. HCA can be used as the very first data exploration algorithm since it does not require any help from a human expert, not even a priori selection of the probable number of classes. The algorithm can also generate a corresponding tree diagram showing contribution of each member of the input vector to the final clustering partitions. Such a dendritic tree produced the canonical genomic classification system devised by Woese et al., a central tool for understanding the diversity and evolution of life.
The second unsupervised algorithm, k-means, has been shown to be able to generate the optimal clustering model for any input data. However, it is not quite as autonomous as HCA. K-means requires an initial estimate of the likely number of clusters if it is going to solve the clustering task in a reasonable amount of time. The estimate of the likely number of clusters can be provided automatically as one of the outputs for HCA, or it can be provided by a human expert. After choosing the initial data points as temporary cluster centers, k-means assigns data points to a cluster and moves the center of the cluster such that the sum of the squared distance between the data points and the arithmetic mean of all the cluster's data points is minimized. The algorithm does not generate a probabilistic estimate of the accuracy of cluster assignment.
## Supervised data classification
Linear discriminant analysis (LDA) and support vector machines (SVM) are supervised machine learning classifiers. Often referred to as Fisher's linear discriminant, LDA, like PCA, searches for a linear combination of variables that optimally characterizes a data set. Unlike PCA, LDA has access to not only the uncontrolled variables (in the case of LRS the spectral bands), but also to the predicted classification of, usually, a human expert. Given that a priori classification, LDA attempts to model the differences between the predicted classes.
SVM is a robust, easily implemented and reliable linear binary classifier. SVM can rapidly identify class boundaries between two linearly separable clusters. Vectors of the input data are non-linearly mapped to a high-dimensional feature space where a linear decision surface is formed to create two clusters. Kernel trick is often employed when using SVM so that one can operate in input space instead of a highly dimensional one. SVM boundary decisions are relatively easy to understand. However, the final classification is purely binary. Neither LDA or SVM produce a probability that a classification is correct and do not inherently estimate the error in their prediction for each target.
## Bayesian probabilities of correct classification-stochastic neural networks
Once unsupervised algorithms assign cluster memberships to each data point for all samples, the correctness of each assignment must be estimated. Bayesian probability theory is a powerful formalized method to generate the probability that each data point should be included in each of the data clusters proposed by either a human expert or an unsupervised algorithm. Stochastic backpropagation neural networks (NNs) are simple optimization algorithms modeled on the neuronal signal processing characteristics of the brain. These stochastic algorithms have been shown to be robust estimators of the Bayesian a posteriori probability of correct classification and easily generate an error analysis for each probability. Unlike LDA and SVM, NNs can model non-linear boundaries, have been shown to be universal approximation machines, and can generate a Bayesian probability that each target has been correctly classified. NNs are not confined to binary classification tasks, but can easily assign a Bayesian probability that a target is a member of each class under consideration. Breast cancer tissue is often heterogenous, and at a given time the Raman laser spot light might be incident on a boundary region and thus capturing signal from a mixture of malignant and healthy cells. Hence a probabilistic classification model can more accurately depict the heterogenous nature of cancer and leave the final decision making to the surgeon.
Backpropagation networks consist of two or more "layers of nodes". A node (i) is just one of the numbers in a vector describing a data point. In NN terminology, a complete vector is called a layer (j). Each NN will have at least one input layer of nodes, an output layer, and a set of weights (w ij ) that fully connect the nodes of the two layers. The output nodes for the NN are the Bayesian probability that a target has been correctly classified either by HCA, k-means, or a human expert. If the classes are linearly separable, only the input and output layers and their connecting weights are required. If a nonlinear separation is expected, one or more layers are inserted between the input and output layers to extract higher order terms.illustrates a simple data flow we have used in previous work beginning with the initial analysis of variance and choice of spectral regions of interest used as qualitative input exploration with PCA, HCA, and k-means, through to a stochastic backpropagation NN using a single intermediary layer between the input and output layers (known as the hidden layer).
For this strawman exampledepicts the variance in 20 LRS spectra acquired while evaluating ex vivo tissue excised during breast conserving lumpectomy, 10 from sites identified by subsequent histopathology as tumor sites and 10 from sites identified as healthy tissue. The spectra have been corrected for cosmic rays, smoothed using Savitsky-Golay algorithm, continuum removed by arPLS, and vector normalized. The first frame depicts the variance present in all 20 spectra without attention to probable classification. Multiple spectral regions appear to contain significant information. The second frame depicting the 3-sigma error for spectra from each of the two putative classes, tumor and healthy, show considerable overlap for several of the possible bands of interest for classification. Such a qualitative assessment can then be verified using PCA to look for spectral regions that can contribute to classification. HCA and k-means unsupervised classifications using multiple combinations of possible bands of interest can also be used to explore for the optimal data dimension reduction and clustering partitions. In the mock-up depicted in, six regions of interest show high variance and are free from overlap at the 3-sigma level. The Raman shift values and the molecular targets exhibiting Raman shifts at these wavelengths appear in . . Biomolecular assignments of six regions of interest in spectra displayed in. The intensities of each of the bands representing the spectral regions of interest enter the NN at the input layer. Each node in the hidden layer is connected to each node in both the input and output layers by a weight that is initialized at the start of training with a random number between 1 and 0. Training is a simple matter of iteratively adjusting each of these weights to generate values at the output layer that match values presented as the predicted classification. For example, a node at layer j initiates a calculation of the sum of the inputs from the previous layer each multiplied by the specific w ij for that connection. The node constrains its output by transforming the sum using a nonlinear threshold function of sigmoid form across the interval [0, 1]. The weights are modified by minimizing least squares (see Storrie-Lombardi et al.for a detailed description of NN training). The novel aspect of stochastic backpropagation NNs is the manner in which the optimization is performed using the chain rule (also known as gradient descent), which was independently discovered by several investigators. During training, the algorithm measures the differences between its output and the putative class assignment and then propagates error backward by modifying the weights in each layer according to their contribution to the total error. Note that the NNs represented inare non-linear, unlike the two unsupervised clustering algorithms. These NNs are capable of drawing more complex class boundaries than can be accomplished by the two clustering algorithms.
Interestingly, if a target has been incorrectly assigned by one of the linear unsupervised algorithms or a human expert, these stochastic non-linear NNs will "refuse" the putative assignment of the target and generate a low probability that the assignment is accurate. In analyzing the output of these NNs, if the model has been correctly configured, the probabilistic predictions for all classes considered will sum tõ 1, the required hallmark of a Bayesian analytical engine. For data sets of limited size, training and testing of all samples can be easily accomplished by two different techniques: leave-one-out cross-validation and jackknife training/testing sequence. Both have their advantages and disadvantages. Leave-one-out cross-validation removes one target before training begins, trains an NN on the full data set less one, tests the hold out target, and then repeats the entire process for the next datum until all targets have been tested. The process is excellent for optimal utilization of small data sets, for the proper training of the NN, and for the precise evaluation of each target. But for large data sets the technique is computationally resource-intensive. A variety of jackknife procedures have been developed that generally involve selection of random subsets of the data for training, testing, and validation. It is the stochastic reset of the NN's weights for multiple runs that makes it possible to implement a Monte Carlo test-retest sequence and produce statistics that the NNs use to generate a true error estimate to assess their own performance. It should be noted that larger data sets (about an order of magnitude more spectra than the number of weights in the NN) are preferred since their size inhibits the NNs from "memorizing" the "correct" answers given by the unsupervised algorithms or the human expert. Instead the algorithms will learn the more robust patterns in the correlations and anti-correlations driving the unsupervised clustering algorithms.
Finally, it cannot be overemphasized that since these are stochastic NNs, the entire training and testing process can be automatically run as many times as necessary with a complete random reset of the starting weights occurring at the beginning of each run. The output from these runs, that each begin at unique initial states, provides clear statistics characterizing the inherent error in the Bayesian probability generated by the algorithm. This capability separates stochastic backpropagation algorithms from all other classifiers discussed here.summarizes all LRS breast cancer papers reviewed in this article.
## Review of major research advancements
In 2002, Shafer-Peltier et al.presented the first comprehensive and accurate morphological model of breast tissue using a Raman confocal micro imaging system by using 'basis spectra' of components found in breast tissue as building blocks for macroscopic samples. They collected Raman micro-images of breast tissue which were used to identify nine major components as the basis spectra: cell cytoplasm, cell nucleus, collagen, fat, cholesterol-like, β-carotene, calcium hydroxyapatite, calcium oxalate dihydrate and water. The Raman images were correlated with phase contrast images as well as H&E stains of the same tissue section and the images were overlapped for comparison. Using linear combinations of the nine basis spectra the investigators were able to model significant Raman features of a range of breast tissues from normal to cancerous and got consistent results with the known morphology of the tissue determined from phase contrast and H&E staining. They employed the use of least squares fitting for fitting basis spectra to the Raman spectroscopic images and used PCA to identify the independently varying components. The same year Haka et al.published a LRS investigation of microcalcifications found in benign and malignant breast lesions. LSR probing of these microcalcifications found two distinct categories, type I, calcium oxalate dihydrate which are benign and type II, calcium hydroxyapatite, which can be either malignant or benign. Using PCA on the acquired spectra the investigators were able to determine that type II microcalcifications in benign ducts have more calcium carbonate and less protein than type II microcalcifications found in malignant ducts. They were able to discriminate microcalcifications present in benign and malignant breast ducts with sensitivity of 88% and a specificity of 93%, which is marked improvement over traditional techniques like X-ray mammograms. In a subsequent paper published in 2005, Haka et al.were able to use the a linear combination of the same basis spectra introduced by Shafer-Peltierto determine the contribution of each basis spectra to breast tissue specimens of normal tissue, fibrocystic lesions, fibroadenoma and infiltrating carcinoma by normalizing the fit coefficients so that they sum to 1. These were validated using H&E stains of the same tissue specimen. Further, these fit coefficients were used to make the first spectral based diagnostic algorithm to identify specific pathologies in breast tissue. This algorithm used the fit coefficients of fat and collagen to divide all data points into two categories, one group contained infiltrating carcinomas and fibroadenomas and the second group contained normal tissue and fibrocystic lesions. Logistic regression was used to further subdivide the two groups resulting in sensitivity of 94% and specificity of 96% for infiltrating carcinoma when compared to H&E analysis by the pathologist. In 2006, the same group published their first LRS in vivo study of margins during a partial mastectomy. The investigators collected a total of 31 Raman spectra from nine patients, and were able to collect the spectra in 1 s, highlighting the potential real-time benefits of this technique. The data was analyzed and fit in real-time using the same basis spectraand modeldescribed in previous publications by the same group. The diagnostic algorithm produced 100% sensitivity and specificity for detecting carcinoma when compared to the standard H&E pathological review; although it is important to note that there was only one malignant spectral sample in this dataset. This malignancy was grossly invisible during surgery, and after H&E analysis the margin was deemed positive and the patient had to undergo a second surgery. Had the malignant spectrum been taken into account during initial surgery, a second surgery could have been avoided altogether, hence highlighting very promising results with this first in vivo study.
Mohs et al.published an in vivo study of tumor detection in mouse models in 2010. They investigated the use of a hand-held Raman spectroscopic device operating at 785 nm, called SpectroPen to detect intraoperative contrast agents. They discussed the design and performance of the SpectroPen, which can resolve NIR fluorescence from the Raman signal using optical filtering fitted into the hand-held device and compared the results from the SpectroPen to a classic 785 nm Raman spectrometer. They found that the SpectroPen was able to detect two contrast agents which adhere to malignant cells, fluorescent contrast agent (indocyanine green, ICG) and a surface-enhanced Raman scattering (SERS) contrast agent (pegylated colloidal gold). Furthermore, they studied the performance of the SpectroPen in determining tumor margins of mice injected with 4T1 tumor cells and the ICG contrast agent. They collected 14 spectra using SpectroPen and reliably found ICG signals which were correlated to the bioluminescent and bright-field images of the mouse, thus accurately detecting tumor borders.
Keller et al.propose the use of spatially offset Raman spectroscopy (SORS) for tumor detection. SORS is a technique in which Raman spectra is collected from regions spatially offset by varying amounts from the point of original incidence, which facilitates signal collection from deeper layers, since Raman photons from deeper within the sample are shifted laterally before emission from sample surface. This can be instrumental in accurately determining tumor margins since clear margin is 2 mm for infiltrating carcinoma. Keller et al.discuss the development of a SORS probe including orientation of source and collection fibers and collected in vitro breast tissue signals. The tumor signals were characterized by a strong phenylalanine band at 1006 cm −1 and a wider amide I band at 1656 cm −1 , the healthy signals exhibited a more intense CH stretch at 1445 cm −1 and the presence of a carbonyl stretch band at 1750 cm −1 and both types of signals had significantly different ratios of 1303 to 1265 cm −1 bands which is indicative of the ratio of lipid to protein content. They achieved a sensitivity of 95% and specificity of 100% using sparse multinomial logistic regression when compared to H&E histology.
Brozek-Pluska et al.present another study of LRS data from ex vivo breast tissue with 321 spectra from 44 patients. They found that normal breast tissue spectra exhibited C-C, and C=C stretching bands of carotenoids as well as the C-H symmetric and asymmetric bands of lipids, which were absent in malignant and benign tissue spectra. Additionally, their results showed more autofluorescence in malignant tissue than in normal and benign tissue and showcased a sensitivity of 72% for identifying malignant tissue using PCA analysis (using PC1). A subsequent paper published by Abramczyk et al.from the same laboratory with 1100 spectra from 99 patients confirmed all of Brozek-Ploskas' results; also obtaining a sensitivity of 72% for identifying malignant tissue with respect to the first PC in PCA analysis.
Kong et al.published a seminal study combining auto fluorescence (AF) images to highlight important sampling points for Raman spectroscopy to diagnose basal cell carcinoma during Mohs surgery, and thereby reducing acquisition time for Raman spectra. They found that this methodology was faster than frozen section histology and studies that use only infrared or Raman microscopy. PCA and K-means clustering were employed to achieve 100% sensitivity and 92% specificity. Shipp et al.recently extended this methodology to BCS. The investigators used a diagnosis model and an initial independent test based on smaller mastectomy samples with sensitivity = 91% and specificity = 83% and validated data using k-means and LDA. They analyzed 51 fresh BCS specimens with multi-modal spectral histopathology (AF and Raman spectra) and could detect residual and small tumors on the surface of whole BCS samples with 100% sensitivity and at least 80% specificity.
Garcia-Flores et al.investigated breast tissue of rats using high frequency LRS on in vivo and ex vivo samples. Using PCA and LDA resulted in discrimination accuracy of 77.2%, 82.3% and 100% for in vivo transcutaneous, in vivo skin-removed and ex vivo spectra respectively. Zúñiga et al.also employed PCA and LDA on Raman spectra of ex vivo breast tissue using a 785 nm system to get sensitivity of 90% and specificity of 86% when compared to H&E expert analysis.
SVM and LRS have been used to identify normal tissue, fibrocystic change (FCC), fibroadenoma (FA) and breast cancer, in the absence and presence of microcalcifications during stereotactic core needle biopsy. Barman et al.were able to achieve a sensitivity of 62.5% and specificity of 100% using SVM with leave-one-out cross validation. Recently LRS, PCA, LDA, and SVM have been employed to discriminate between benign lesions, fibrocystic disease, fibroadenoma, intraductal papilloma, invasive ductal carcinoma and lobular carcinoma in formalin fixed paraffin preserved tissue.
In the last year, a few groups have evaluated the use of neural networks for LRS breast cancer diagnosis. Shang et al.used back propagation NNs on breast cancer Raman spectra. They ran two separate NNs on spectra mostly comprised of collagen and those having mostly lipid content to get a discrimination accuracy of 95.33% and 98.67%, respectively. Koya et al.use convolutional neural networks (CNNs) for their classification of breast tissue spectra to achieve a sensitivity of 88% and specificity of 90.8%.
It is important to note that menstrual status and hormonal variation can affect Raman spectra. Kanter et al.demonstrated that stratifying Raman data by a patient's hormonal status (point in menstrual cycle and menopausal state) can increase the classification accuracy of cervical precancer from 88% to 94%. In another publication, the same group found that stratifying the data by menstrual state increased the classification of low-grade squamous intraepithelial lesion (LGSIL) to 97% from 74%. This finding can be extended to breast tissue as well. Shah et al.investigated the use of diffuse optical spectroscopy (DOS) to study different menopausal states in women. One of their findings included higher tissue absorption coefficient before the conset of menses than before ovulation. Cubeddu et al.also found that optical properties of breast tissue follow the physiological changes that occur during the menstrual cycle and these differences should be taken into account when using optical techniques on breast tissue.
In summary, we find that many investigators are able to produce high prediction statistics for discriminating between normal and cancerous breast tissue, but the sensitivity and specificty fall for studies that have a larger sample set. Possible compouding factors are variability in (1) menstrual state with its effect on optical properties of breast tissue, (2) percentage of lipid to fibrous tissue in healthy breast tissue, (3) tumor cell density within a healthy matrix, and, a simple item often overlooked, (4) the prescence of surgical ink used at the time of lumpectomy. Five of the six commonly used dyes that we have tested have strong contaminating Raman signals. We postulate that collecting reliable Raman spectra from all breast tissue types as well as all menstrual states that can be encountered during surgery is one way to advance the use of LRS in clinical settings. Additionally, the creation of an international database of Raman spectra of breast tissue will allow validation of diagnostic algorithms on large datasets and thus creating a more reliable, robust and accurate classification. Furthermore, augmenting Raman spectra with other imaging modalities can increase the classification accuracy and reduce data acquisition time in the operating room (as demonstrated by Shipp et al.. in using AF and Raman spectra and also proposed by Maloney et al..
# Conclusions
LRS efforts will continue to advance in real time surgical settings. Two series of LRS studies on brain cancer are particularly encouraging. Hollon et al.have recently reported on the utility of an NN trained on over 2.5 million simulated Raman histology images to predict brain tumor diagnosis in an operating room under 150 s. They report an accuracy of 94.6% compared to pathologist interpretation which had an accuracy of 93.9%. The work is an excellent example of the computational challenges amenable to current NN algorithms. Desroches et al.have pioneered the operational use of a Raman spectroscopy guidance system tightly integrated with a brain biopsy needle. The work points the way for LRS molecular diagnostic guidance in surgical intervention in breast, prostate, and hepatic cancer.
A fundamental LRS investigation is currently underway to understand tumor progression within a broader metabolic context. A metabolic syndrome that includes hyperlipidemia, hypertension, and diabetes is highly correlated with a decrease survival time for breast cancer patients. Patients with the syndrome also exhibit a clinically covert white adipose tissue inflammation (WAT). Histologic examination of WAT reveals dead or dying adipocytes that are surrounded by macrophages. Recently, LRS has been shown to be the first non-invasive technique to successfully identify the presence of WAT in both human and murine model systems. The authors posit that the LRS fingerprint signal for WAT arises from fatty acid saturation that occurs in association with an adipocyte hypertrophy. The investigation opens the doorway for LRS to explore biochemical shifts in lipid pathophysiology that may be either correlated with or directly contributory to the unfolding of breast cancer tumor progression and outcome.
Clearly, over the past two decades, multiple investigators have successfully employed both PCA and analysis of variance heuristic feature selection strategies along with unsupervised and supervised clustering and classification algorithms. We would propose that it is not feature selection or choice of easily implemented binary classification tools that will decide the evolution of LRS in medical diagnostics. Instead, it will be on the LRS community to raise the bar for the use of so called "machine learning" algorithms in cancer diagnostics. We posit that the majority of machine learning algorithms employed over the last decade fail to provide the two most critical pieces of information required by the practicing surgeon: a probability that the classification of a tissue is correct, and, more importantly, the expected error in that probability. As such, they should be considered simply statistical tools to explore data sets; tools that in some ways do not qualify as learning machines. Stochastic backpropagation artificial neural networks inherently provide both pieces of information for every tissue site examined by LRS: probability of correct classification and an estimate of the reliability of that probability. If the networks are trained using both human experts and an unsupervised classification algorithm as gold standards, rapid progress can be made understanding what additional contextual data from the human expert is needed to improve classification performance.
Colleagues and patients expect us to not simply have an opinion about a putative tumor, but to know how certain we are that we our decision is correct. Only algorithms capable of providing the same two pieces of information required of the human expert should be deployed in modern LRS oncogenic diagnostics. Serendipitously, the recent wealth of work on deep neural networks (NNs with two or more hidden layers) has paved the way for mining the large, complex, multivariable data sets critical for cancer diagnostics. These deep learning algorithms need to employ stochastic weight resets and standardized automatic controls for under-and over-learning during training, techniques easily implemented in even the more complex nets. We expect the coming decade to be an exciting time in the evolution of LRS and stochastic neural network technologies. |
High Rates of Three Common GJB2 Mutations c.516G>C, c.-23+1G>A, c.235delC in Deaf Patients from Southern Siberia Are Due to the Founder Effect
The mutations in the GJB2 gene (13q12.11, MIM 121011) encoding transmembrane protein connexin 26 (Cx26) account for a significant portion of hereditary hearing loss worldwide. Earlier we found a high prevalence of recessive GJB2 mutations c.516G>C, c.-23+1G>A, c.235delC in indigenous Turkic-speaking Siberian peoples (Tuvinians and Altaians) from the Tyva Republic and Altai Republic (Southern Siberia, Russia) and proposed the founder effect as a cause for their high rates in these populations. To reconstruct the haplotypes associated with each of these mutations, the genotyping of polymorphic genetic markers both within and flanking the GJB2 gene was performed in 28 unrelated individuals homozygous for c.516G>C (n = 18), c.-23+1G>A (n = 6), or c.235delC (n = 4) as well as in the ethnically matched controls (62 Tuvinians and 55 Altaians) without these mutations. The common haplotypes specific for mutations c.516G>C, c.-23+1G>A, or c.235delC were revealed implying a single origin of each of these mutations. The age of mutations estimated by the DMLE+ v2.3 software and the single marker method is discussed in relation to ethnic history of Tuvinians and Altaians. The data obtained in this study support a crucial role of the founder effect in the high prevalence of GJB2 mutations c.516G>C, c.-23+1G>A, c.235delC in indigenous populations of Southern Siberia.
# Introduction
Mutations in the GJB2 gene (gap junction protein, beta-2, 13q12.11, MIM 121011) encoding transmembrane protein connexin 26 (Cx26) lead to nonsyndromic autosomal recessive deafness 1A (DFNB1A, MIM 220290) which is the most common form of hereditary hearing loss in many populations. High prevalence of the GJB2-associated deafness makes the GJB2 gene testing essential for the establishment of genetic diagnosis of hearing loss.
Over 400 deafness-associated variations in GJB2 have been reported in the Human Gene Mutation Database (http://www.hgmd.cf.ac.uk). Specific ethno-geographic prevalence patterns were found for many of them. For instance, variant c.35delG (p.Gly12Valfs*2) is prevalent in deaf patients of Caucasian origin; c.235delC (p.Leu79Cysfs*3) is common in some Asian populations; c.167delT (p.Leu56Argfs*26) is frequent in Ashkenazi Jews; c.427C>T (p.Arg143Trp) is specific for population of Ghana (West Africa) and Peru (South America); c.71G>A (p.Trp24*) is widely spread in Indians and European Gypsies; c.109G>A (p.Val37Ile) prevails in populations of Southeast Asia; the splice donor variant c.-23+1G>A was found in many populations worldwide but extremely high prevalence of c.-23+1G>A was detected among Yakuts (Eastern Siberia, Russia); c.131G>A (p.Trp44*) was found with high frequency among descendants of ancestral Mayan population in Guatemala.
High prevalence of some major GJB2 mutations in certain populations was explained by the founder effect as evidenced by conservation of haplotypes with closely linked markers. In some cases, analysis of genetic background of these mutations allowed to elucidate their approximate age and a presumable region of origin. The key role of the founder effect in prevalence of mutation c.35delG was established in numerous studies by analysis of the c.35delG-bearing haplotypes: this mutation first appeared approximately 10000-14000 years ago in the Middle East and/or the Mediterranean and then spread by human migrations throughout Europe and worldwide. The conservation of haplotype bearing mutation c.167delT found in Ashkenazi Jews suggests a single origin of this mutation which began to spread since a presumed Ashkenazi population bottleneck. Haplotype analysis of genetic markers flanking the GJB2 gene showed that a high rate of mutation c.71G>A (p.Trp24*) common for Indians is most probably due to the founder effect, and the age of this mutation was calculated as 7880 years. Contribution of the founder effect in extremely high rate of mutation c.-23+1G>A among Yakuts (Eastern Siberia, Russia) was evidenced by the c.-23+1G>A haplotype analysis, and the age of this mutation was estimated at approximately 800 years. Common haplotype was established for specific mutation c.131G>A (p.Trp44*) found in individuals from Guatemala suggesting a single founder from ancestral Mayan population. The founder effect was also suggested in high prevalence of mutation c.235delC in East Asians (China, Japan, Korea), Mongolians (Mongolia), and Altaians (Southern Siberia, Russia) but there were only a few studies of the c.235delC-bearing haplotypes to support this hypothesis. Additionally,proposed that c.235delC has probably derived from a founder mutation approximately 11500 years ago in the Lake Baikal region and spread to some Asian regions through subsequent migrations. A haplotype block specific to East Asians with the c.109G>A (p.Val37Ile) mutation was found among deaf patients of Chinese, Japanese, Vietnamese, and Philippines ancestry and the age of p.Val37Ile in this Asian cohort was estimated at approximately 300 generations. confirmed the founder effect in origin of six GJB2 mutations frequently observed in Japanese hearing loss patients (c.235delC, p.Val37Ile, p.[Gly45Glu;Tyr136*], p.Arg143Trp, c.176_191del, and c.299_300delAT) and estimated the year at which each mutation occurred: c.235delC-around 6500 years ago, p.[Gly45Glu;Tyr136*]-around 6000 years ago, p.Arg143Trp-around 6500 years ago, c.176_191del-around 4000 years ago, c.299_300delAT-around 7700 years ago, and p.Val37Ile -around 14500 or 5000 years ago.
In our recent study, we evaluated the spectrum and frequency of the GJB2 gene variants in a large cohort of deaf Tuvinian patients and the ethnically matched controls from the Tyva Republic (Southern Siberia, Russia). A striking finding was a high prevalence of rare specific variant c.516G>C (p.Trp172Cys) in the GJB2 gene accounting for 62.9% of all mutant GJB2 alleles found in Tuvinian patients and having carrier frequency of 3.8% in controls. Other frequent GJB2 mutations found in Tuvinian patients were c.-23+1G>A (27.6%) and c.235delC (5.2%). The c.235delC was previously found as a major GJB2 mutation in Altaians living in the Altai Republic (Southern Siberia, Russia) neighboring the Tyva Republic. In our recent study on enlarged cohort of Altaian deaf patients, the proportion of c.235delC, c.516G>C, and c.-23+1G>A among all mutant GJB2 alleles found in Altaian patients was estimated as 51.9%, 29.6%, and 14.8%, respectively.
High rate of three GJB2 mutations c.516G>C, c.-23+1G>A, and c.235delC in Tuvinians and Altaians implies a crucial role of the founder effect in their prevalence in indigenous populations of Southern Siberia. In this study we test a presumable common origin of each of these GJB2 mutations by analysis of haplotypes bearing c.516G>C, c.-23+1G>A, and c.235delC.
# Materials and methods
## Subjects
The pathogenic contribution of the GJB2 mutations to deafness and their carrier frequencies were evaluated in our preliminary studies in indigenous populations of Southern Siberia (Tuvinians and Altaians) and three GJB2 mutations (c.516G>C, c.-23+1G>A, c.235delC) were found to be common. For the analysis of haplotypes bearing these mutations, we recruited in total 28 unrelated deaf patients who were homozygous for c.516G>C (seventeen Tuvinians and one Altaian), for c.-23+1G>A (six Tuvinians) or for c.235delC (four Altaians). The ethnically matched control samples were represented by 117 unrelated healthy individuals without mutations c.516G>C, c.-23+1G>A, and c.235delC (62 Tuvinians and 55 Altaians).
The study was conducted in accordance with the Declaration of Helsinki, and the protocol was approved by the Bioethics Commission at the Institute of Cytology and Genetics SB RAS, Novosibirsk, Russia (Protocol No. 9, 24 April 2012).
## Strs and snps genotyping
To determine common haplotypes for each of three major GJB2 mutations c.516G>C, c.-23+1G>A, c.235delC, we performed genotyping of seven Short Tandem Repeats (D13S1316, D13S141, D13S175, D13S1853, D13S143, D13S1275, D13S292) flanking the GJB2 gene and nine Single Nucleotide Polymorphisms (rs747931, rs5030700, rs3751385, rs2274083, rs2274084, rs1411911768, rs9552101, rs117685390, rs877098) intragenic and flanking the GJB2 gene both in 28 unrelated deaf patients homozygous for c.516G>C, c.-23+1G>A, or c.235delC and in 117 unrelated healthy individuals (62 Tuvinians and 55 Altaians) who were negative for these mutations. The location of analyzed genetic markers on chromosome 13 is presented in. Two additional SNPs (rs11147592, rs9509086) were genotyped in homozygous patients only. The total length of the region flanked by distal markers D13S1316 (centromeric) and D13S292 (telomeric) was approximately 3.5 Mb. All primers and genotyping methods are summarized in . Fragment analysis and Sanger sequencing were performed in the SB RAS Genomics Core Facility (Institute of Chemical Biology and Fundamental Medicine SB RAS, Novosibirsk, Russia).
## Figure 1.
Schematic presentation of the GJB2 gene structure and localization of genetic markers (seven STRs and nine SNPs) which were used for the reconstruction of haplotypes for GJB2 mutations c.516G>C, c.-23+1G>A, and c.235delC. These mutations are marked by red color. *-basal (core) promoter (128 bp). Positions of genetic markers (shown in brackets) were defined according to GRCh37.p13 Genome Assembly (https://www.ncbi.nlm.nih.gov/assembly/GCA_000001405.14).
## Estimation of mutations age
Estimation of a mutation age is based on the expected decay of linkage disequilibrium between the mutation and alleles of surrounding genetic markers due to recombination ("genetic clock" concept). We applied two approaches for estimating the age of mutations c.516G>C, c.-23+1G>A, and c.235delC. The first was the DMLE+ v2.3 software method (Disequilibrium Mapping using maximum-Likelihood Estimation, DMLE+: http://dmle.org/)which is based on multiple linked marker loci and uses the Markov Chain Monte Carlo algorithm for Bayesian estimation of the mutation age. The second, used when appropriate, was the single marker method based on intraallelic variation of a single marker. For calculation the mutation age by the DMLE+ software, the demographic parameters (population size, population growth rate, and proportion of population sampled) are required in addition to the haplotype data and the map distances among marker loci and mutations. Since population growth rates for Tuvinian and Altaian populations could not be reliably estimated because of very limited knowledge of demographic variation of these populations Positions of genetic markers (shown in brackets) were defined according to GRCh37.p13 Genome Assembly (https://www.ncbi.nlm.nih.gov/assembly/GCA_000001405.14).
## Reconstruction of str and snp haplotypes
The reconstruction of the founder haplotypes from STRs and SNPs genotyping data and analysis of their frequencies were performed using Expectation-Maximization (EM) algorithm of the Arlequin 3.5.2.2 software. The boundaries of haplotypes for each of three GJB2 mutations were determined by observed linkage disequilibrium between the marker alleles and each mutation according to equation δ = (Pd−Pn)/(1−Pn), where δ is the measure of linkage disequilibrium, Pd is the marker allele frequency among mutant chromosomes, Pn is the frequency of the same allele among normal chromosomes.
## Estimation of mutations age
Estimation of a mutation age is based on the expected decay of linkage disequilibrium between the mutation and alleles of surrounding genetic markers due to recombination ("genetic clock" concept).
We applied two approaches for estimating the age of mutations c.516G>C, c.-23+1G>A, and c.235delC. The first was the DMLE+ v2.3 software method (Disequilibrium Mapping using maximum-Likelihood Estimation, DMLE+: http://dmle.org/)which is based on multiple linked marker loci and uses the Markov Chain Monte Carlo algorithm for Bayesian estimation of the mutation age. The second, used when appropriate, was the single marker method based on intra-allelic variation of a single marker. For calculation the mutation age by the DMLE+ software, the demographic parameters (population size, population growth rate, and proportion of population sampled) are required in addition to the haplotype data and the map distances among marker loci and mutations. Since population growth rates for Tuvinian and Altaian populations could not be reliably estimated because of very limited knowledge of demographic variation of these populations along their history, we analyzed the haplotype data using several plausible growth rates: 0.05, 0.1, and 0.2. The parameter "proportion of population sampled" for each of three mutations (c.516G>C, c.-23+1G>A, c.235delC) was calculated on the basis of our previous data. The contemporary population sizes for Tuvinians and Altaians according to the 2010 census were 249299 and 68814 peoples, respectively.
The estimation of the mutation age by the single marker method was performed using algorithm proposed by:
[formula] g = log[1 − Q/(1 − Pn)]/log(1 − )(1) [/formula]
where g is the number of generations passed from the moment of the mutation appearance to the present; Q is the share of mutant chromosomes unlinked with the founder haplotype; Pn is the population frequency of allele included in the founder haplotype, and is the recombinant fraction calculated from physical distance between marker and mutation (under the assumption that 1 cM = 1000 kb). To avoid possible underestimation of a mutation age as suggested by, we also applied the Luria-Delbrűck correctionwhich takes into account the demographic parameters:
[formula] g c = g + g 0 (2) g 0 = −(1/d) ln( f d )(3) [/formula]
where d is population growth rate, also assuming f d = e d /(e d −1) and f d ≈ 1/d at small d values. The duration of one generation (g) was considered to be 25 years.
# Statistical analysis
Two-tailed Fisher's exact test with significance level of p < 0.05 was applied to compare allele frequencies between patients and controls.
# Results
We assumed that the high prevalence of GJB2 mutations c.516G>C (p.Trp172Cys), c.-23+1G>A, c.235delC in Tuvinians and Altaians is a consequence of the founder effect. To test whether all carriers of each of these mutations share a common haplotype, we performed genotyping of polymorphic genetic markers both intragenic and flanking GJB2 gene (nine SNPs and seven STRs) in 28 unrelated individuals homozygous for c.516G>C (n = 18), c.-23+1G>A (n = 6), or c.235delC (n = 4) as well as in ethnically matched controls (62 Tuvinians and 55 Altaians). The choice of analyzed genetic markers was based on their physical location, their variability in Asian populations, and the availability of previously published data for other populations. Results of the STRs and the SNPs genotyping are summarized in.
## Str haplotypes
Data on genotyping of seven STR markers (D13S1316, D13S141, D13S175, D13S1853, D13S143, D13S1275, D13S292) flanking the GJB2 gene and encompassing approximately 3.5 Mbwere used to reconstruct STR haplotypes both in deaf patients homozygous for each GJB2 mutations (c.516G>C, c.-23+1G>A or c.235delC) and in the ethnically matched controls. The boundaries of the shared STR haplotypes were determined by observed linkage disequilibrium between STR alleles and each mutation.
Three different haplotypes formed by specific alleles of five STRs (D13S1316, D13S141, D13S175, D13S1853, D13S143) with a length of approximately 1.6 Mb were found to be associated with mutation c.516G>C in Tuvinian patients and 39 STR haplotypes were reconstructed in Tuvinian control sample (data not shown). The 269-124-105-204-125 haplotype was the most common (67.9%) among mutant chromosomes bearing c.516G>C, while the frequency of this haplotype in normal chromosomes (1.6%) was significantly lower (p < 10 −14 ) . and in Tuvinian controls (data not shown), respectively. Significant differences (p < 10 −8 ) were observed between frequency of the 124-105-204-125-208-209 haplotype predominantly found among all mutant chromosomes with c.-23+1G>A (83.3%) and its frequency among normal chromosomes in Tuvinian controls (5.4%) .
The only haplotype found in all mutant chromosomes with c.235delC (Altaian patients) was 267-124-105-204-125-210 (D13S1316-D13S141-D13S175-D13S1853-D13S143-D13S1275) flanked by markers D13S1316 and D13S1275 (~1.7 Mb), whereas this haplotype was not detected on normal chromosomes in Altaian control sample (p < 10 −11 ) .
## Snp haplotypes
To thoroughly analyze the structure of haplotypes associated with specific GJB2 mutations, we have genotyped nine SNPs: four SNPs flanking GJB2 gene (rs747931, rs9552101, rs117685390, rs877098) and five intragenic SNPs (rs5030700, rs3751385, rs2274083, rs2274084, rs1411911768)in patients homozygous for c.516G>C, c.-23+1G>A, or c.235delC and in the ethnically matched controls.
Significant linkage disequilibrium was observed between each of three GJB2 mutations and certain alleles of all analyzed SNPs.
The only haplotype T-C-C-A-G-T-G-T-C (rs747931-rs5030700-rs3751385-rs2274083-rs2274084-rs141 1911768-rs9552101-rs117685390-rs877098) was found on all (100%) mutant chromosomes with c.516G>C in Tuvinian patients in contrast with normal chromosomes in Tuvinian controls where 24 different SNP haplotypes were reconstructed (data not shown) and frequency of haplotype T-C-C-A-G-T-G-T-C was estimated to be 2.17% (p < 10 −26 ). Haplotypes for c.516G>C: Two SNP haplotypes were present in Tuvinian patients homozygous for c.-23+1G>A, while 26 different SNP haplotypes were reconstructed in the Tuvinian controls (data not shown). The C-C-C-A-G-C-G-T-C haplotype was predominant in Tuvinian patients (91.7%), while its frequency in the Tuvinian controls was 5.3% (p < 10 −10 ).
[formula] rs747931-rs5030700-rs3751385-rs2274083-rs2274084-rs1411911768-rs9552101-rs117685390-rs877098 T-C-C-A-G-T-G-T-C 1 0.0217 120 <10 −26 other haplotypes 0 0.9783 - - Haplotypes for c.-23+1G>A: rs747931-rs5030700-rs3751385-rs2274083-rs2274084-rs1411911768-rs9552101-rs117685390-rs877098 C-C-C-A-G-C-G-T-C 0.9167 0.0532 64 <10 −10 C-C-C-A-G-C-G-T-T 0.0833 0.1540 0.047 0.4488 other haplotypes 0 0.7928 - - Haplotypes for c.235delC: rs747931-rs5030700-rs3751385-rs2274083-rs2274084-rs1411911768-rs9552101-rs117685390-rs877098 T-C-C-A-G-C-G-T-T 1 0. [/formula]
Only one SNP haplotype T-C-C-A-G-C-G-T-T was found in Altaian patients homozygous for c.235delC, while 22 different SNP haplotypes were identified in Altaian controls (data not shown). This haplotype was the second by frequency in the Altaian control sample and differences found between its frequency in Altaian patients (100%) and controls (15.9%) were insignificant.
Comparative analysis of SNP haplotypes associated with each of three GJB2 mutations (c.516G>C, c.-23+1G>A, or c.235delC) revealed three SNPs (rs747931, rs1411911768, and rs877098), whose allelic compositions clearly define the specificity of each of these haplotypes. Two of these SNPs, rs747931 and rs877098, are located distantly from the GJB2 gene, while rs1411911768 is located in basal (core) promoter region (128 bp) of the GJB2 gene. Allele T of rs1411911768 included in the common haplotype associated with c.516G>C in Tuvinian patients was present in all corresponding mutant chromosomes, while it was absent in common haplotypes for mutations c.-23+1G>A and c.235delC. Allele C of rs747931 was detected in both c.-23+1G>A-associated haplotypes found in Tuvinian patients but it was absent in haplotypes associated with c.516G>C or c.235delC. Variant T of rs877098 was only found in c.235delC haplotype in Altaian patients and in more rare c.-23+1G>A haplotype in Tuvinian patients, and it was absent in c.516G>C haplotype.
Thus, the unique allelic combination of three SNPs (rs747931-// -rs1411911768-// -rs877098) was found for each of the three most frequent SNP haplotypes bearing GJB2 mutations: T-T-C for c.516G>C, C-C-C-for c.-23+1G>A, and T-C-T-for c.235delC. The common haplotypes found for each of the mutations c.516G>C, c.-23+1G>A, or c.235delC prevailing in indigenous peoples of Southern Siberia imply that each of them had descended from a single ancestor. We estimated the numbers of generations (g) and years (in assumption that g = 25 years) passed from the common ancestral mutation event for each of these mutations assuming several population growth rates (0.05, 0.1, and 0.2) by the DMLE+ v2.3 program, which is sensitive to demographic parameters, and based on an analysis of multiple linked marker loci included in appropriate haplotype. The single marker method for the estimation of the mutation age is based on the linkage disequilibrium and the recombination fraction observed for the alleles of surrounding genetic markers. This approach implies analysis of alleles of the most distal markers which manifest significant linkage disequilibrium, while marker alleles with complete linkage disequilibrium (all disease chromosomes carried the same allele) are considered to be uninformative.
The . For c.516G>C age estimation by the single marker method we used allele (125) of the distal STR marker D13S143 found in high linkage disequilibrium with c.516G>C (Supplementarythat resulted in 27 generations passed from the origin of c.516G>C (675 years). After the Luria-Delbrűck correction allowing to avoid possible underestimation of a mutation age due to demographic parameters, the age of c. We were not able to estimate the age of c.235delC using the single marker method because of the lack of recombination in all markers included in STR and SNP haplotypes observed for c. .
# Discussion
We found three GJB2 mutations, c.516G>C, c.-23+1G>A, and c.235delC to be predominant in deaf Tuvinian and Altaian patients. Tuvinians and Altaians are the indigenous Turkic-speaking populations of two neighboring federal subjects of the Russian Federation, the Tyva Republic (Tuva) and the Altai Republic, respectively, which are located in Southern Siberia. The Tyva Republic is bordered by Mongolia in the south and the east, whereas the Republic of Altai is bordered by Mongolia in the southeast, China in the south, and Kazakhstan in the southwest.
## Ethnic history of tuvinians and altaians
Tuvinians (Tuvans) live mainly in the Tyva Republic in Russia (249299 people in total according to the 2010 census), though relatively small groups of Tuvinians also live in the northern part of Mongolia and in the Xinjiang Uygur Autonomous Region of China. Tuvinians are one of the most ancient Turkic-speaking peoples inhabiting Central Asia and the Sayan-Altai region. The name "Tuva" probably originates from a Samoyedic tribe (referred to the VII century Chinese sources as "Dubo" or "Tupo") that populated the upper Yenisei river region. The location of Tuva in the geographical center of the Asian continent had a significant impact on the formation of its population because of the relations with residents of neighboring regions. At different times, Tuva was at the periphery of a powerful state of Huns (II century BC-I century AD) or was incorporated in the Ancient Turkic Khaganate (VI-VIII centuries), in the Uyghur Khaganate (VIII-IX centuries), in the Yenisei Kyrgyz Khaganate (IX-XII centuries), and also in the Mongol Empire (XIII-XIV centuries), which played an outstanding role in the history of the nomadic civilization and the ethno-political development of Central Asia and the Sayan-Altai region. These historical events had a certain impact on the consolidation of ancestral Tuvinian tribes and, ultimately, on their formation into a single ethnic group. At the end of the XIII-XIV centuries, the ethnic composition of Tuva population already included those groups that took part in the formation of the Tuvinian people: descendants of different Turkic-, Mongolic-, Ket-, and Samoyedic-speaking tribes.
The Altaians, indigenous inhabitants of the Altai Republic (68814 people in total according to the 2010 census), belong to two main ethnic groups originated from several ancient Turkic-speaking tribes: Southern Altaians (Altai-kizhi, Teleut, and Telengit) and Northern Altaians (Chelkan, Kumandin, and Tubalar). Southern Altaian language belongs to the Kipchak branch of Turkic language family whereas the Northern Altai languages are greater influenced by Samoyedic, Yeniseian, and Ugric languages. In the past, the Altai region, as well as Tuva, was conquered or influenced by powerful Turkic Khaganates as well as the Mongol Empire.
Thus, archaeological, linguistic, anthropological, and historical evidences indicate similarities in the ethnogenesis of Turkic-speaking Tuvinian and Altaians.. Moreover, we found the unique allelic combinations of three SNPs (rs747931-// -rs1411911768-// -rs877098) that were highly specific for each of the most frequent haplotypes bearing GJB2 mutations (T-T-C for c.516G>C, C-C-C for c.-23+1G>A, and T-C-T for c.235delC). These combinations were absent or sufficiently less common in the control samples that allows to use them as additional markers for identification of major GJB2 mutations in indigenous populations of Siberia.
## The c.516g>c mutation
The GJB2 variant c.516G>C (p.Trp172Cys, rs1302739538) accounts for 62.9% and 29.6% of all mutant GJB2 alleles detected in deaf Tuvinian and Altaian patients, respectively, and the carrier frequencies of c.516G>C are 3.8% and 0.5% in the corresponding ethnically matched controls. The c.516G>C substitution leads to a replacement of an aromatic non-polar tryptophan with a small polar cysteine at conservative amino acid position 172 (p.Trp172Cys) in the second extracellular loop of protein connexin 26 (Cx26). The c.516G>C meets the main criteria to be classified as pathogenic for autosomal recessive hearing loss based on the ACMG/AMP criteriaas specified by the Hearing Loss Expert Panel. In our recent study, we suggested that this very rare GJB2 mutation is endemic for Tuvinians living in the Republic of Tuva, since besides them c.516G>C was only found in Altaians from neighboring the Altai Republic (with less frequency) and in one deaf patient from Mongolia, and nowhere else in the world.
In this study, we obtained convincing evidence supporting the origin of mutation c.516G>C from a single ancestor. The common STR haplotype spanning about 1.6 Mb as well as the common internal SNP haplotype were identified in most of GJB2 alleles carrying c.516G>C, and their frequencies in patients homozygous for c.516G>C were significantly different from controls. Interesting finding was a strong (100%) association of c.516G>C mutation with very rare allele T (A) of intragenic rs1411911768 (dbSNP: MAF A = 0.00002/3 TOPMED), which was found in Tuvinian and Altaian controls with sufficiently lower frequency (0.0565 and 0.0182, respectively) (Supplementary. We speculate that c.516G>C mutation could initially have arisen on the chromosome bearing rare allele of rs1411911768 in ancestors of these indigenous peoples (rather in Tuvinians, among whom c.516G>C is more prevalent) and reached current high prevalence as a result of the founder effect. The age of c.516G>C based on the single marker method was estimated to be 675 years or 1000-1275 years after the Luria-Delbrűck correction, whereas the dating of this event by the DMLE+ program led to wide time ranges (2275-4500, 1425-2650, or 775-1375 years ago) with different population growth rates (d = 0.05, 0.1, or 0.2, respectively). We tend to think that c.516G>C is rather a relatively "young" mutation since a fast population growth was probably intrinsic to Tuvinians in the past because of a traditionally large family size observed in contemporary Tuvinians. In addition, the
## The c.516g>c mutation
The GJB2 variant c.516G>C (p.Trp172Cys, rs1302739538) accounts for 62.9% and 29.6% of all mutant GJB2 alleles detected in deaf Tuvinian and Altaian patients, respectively, and the carrier frequencies of c.516G>C are 3.8% and 0.5% in the corresponding ethnically matched controls. The c.516G>C substitution leads to a replacement of an aromatic non-polar tryptophan with a small polar cysteine at conservative amino acid position 172 (p.Trp172Cys) in the second extracellular loop of protein connexin 26 (Cx26). The c.516G>C meets the main criteria to be classified as pathogenic for autosomal recessive hearing loss based on the ACMG/AMP criteriaas specified by the Hearing Loss Expert Panel. In our recent study, we suggested that this very rare GJB2 mutation is endemic for Tuvinians living in the Republic of Tuva, since besides them c.516G>C was only found in Altaians from neighboring the Altai Republic (with less frequency) and in one deaf patient from Mongolia, and nowhere else in the world.
In this study, we obtained convincing evidence supporting the origin of mutation c.516G>C from a single ancestor. The common STR haplotype spanning about 1.6 Mb as well as the common internal SNP haplotype were identified in most of GJB2 alleles carrying c.516G>C, and their frequencies in patients homozygous for c.516G>C were significantly different from controls. Interesting finding was a strong (100%) association of c.516G>C mutation with very rare allele T (A) of intragenic rs1411911768 (dbSNP: MAF A = 0.00002/3 TOPMED), which was found in Tuvinian and Altaian controls with sufficiently lower frequency (0.0565 and 0.0182, respectively) (Supplementary. We speculate that c.516G>C mutation could initially have arisen on the chromosome bearing rare allele of rs1411911768 in ancestors of these indigenous peoples (rather in Tuvinians, among whom c.516G>C is more prevalent) and reached current high prevalence as a result of the founder effect. The age of c.516G>C based on the single marker method was estimated to be 675 years or 1000-1275 years after the Luria-Delbrűck correction, whereas the dating of this event by the DMLE+ program led to wide time ranges (2275-4500, 1425-2650, or 775-1375 years ago) with different population growth rates (d = 0.05, 0.1, or 0.2, respectively). We tend to think that c.516G>C is rather a relatively "young" mutation since a fast population growth was probably intrinsic to Tuvinians in the past because of a traditionally large family size observed in contemporary Tuvinians. In addition, the prevalence of this mutation is very restricted. The most plausible scenario suggests that c.516G>C has arisen in the territory of Tuva as the result of a unique event after main formation of the Tuvinian ethnic group (which took place at the end of the XIII-XIV centuries) and then spread into the neighboring territory of Altai. Taking into account the complexity of ethnic history of Tuvinians, it remains unclear, who actually were the c.516G>C founders-different ancient Turkic-or Mongolic-speaking groups or other aboriginal peoples who lived there. The introduction of c.516G>C into Tuva territory with migration flows of ancient Mongolic-speaking groups is not consistent with the finding of c.516G>C in only one deaf patient from Mongoliaas well as with its absence in Mongolian patients living in China. It is known that several nomadic Tuvinian groups roamed in the past across the territories of Tuva and Mongolia had remained in Mongolia when Tuva was separated from Mongolia to become under Russian protectorate after the breakup of the Qing Empire in 1911-1912. Since the ethnicity of examined deaf patients was not reported in the study by Tekin et al., the question about the origin of c.516G>C in Mongolia remains open.
## The c.-23+1g>a mutation
The proportion of the splice donor site mutation c.-23+1G>A reaches 27.6% of all mutant GJB2 alleles in Tuvinian deaf patientsand 14.8% in Altaian patients. Splice donor site GJB2 variant c.-23+1G>A has been detected among deaf patients of different origin around the world. The extremely high prevalence of c.-23+1G>A (up to 92.2% of all mutant GJB2 alleles found in patients and carrier frequency reaching of 10.2%) observed in Yakuts, indigenous Turkic-speaking people living in the subarctic region of Russia (the Sakha Republic, Eastern Siberia), was explained by the founder effect in an isolated population and a probable selective advantage for the c.-23+1G>A heterozygotes in severe subarctic climate. The c.-23+1G>A is also the most common mutation in deaf Mongolian patients from Mongolia.
To our knowledge, the haplotypes bearing c.-23+1G>A were analyzed only in a few studies.suggested diverse origins of c.-23+1G>A based on multiple c.-23+1G>A-associated haplotypes found in comparative analysis of seven Mongolian and three Anatolian Turkish c.-23+1G>A homozygous patients. However, despite the fact that several different haplotypes were found to be associated with c.-23+1G>A in Mongolians, a single conserved haplotype (which appears to be a common haplotype in Mongolia) was identified in Turkish homozygous patients suggesting a single common ancestor with an intervening population bottleneck in the Turkish branch.revealed the common origin of c.-23+1G>A in Yakuts (Eastern Siberia) by the reconstruction of 140 haplotypes bearing this mutation using eight polymorphic microsatellite markers flanking the GJB2 gene and two intragenic SNP markers. These findings are consistent with the founder effect hypothesis and support a common Central Asian origin of c.-23+1G>A since the Turkic-speaking ancestors of Yakuts migrated to the Eastern part of Siberia from their initial settlement in the Baikal Lake area under pressure of the Mongol expansion in XI -XIII centuries AD.analyzed the c.-23+1G>A haplotypes in the sample of Yakut, Evenk, Russian, and Tuvinian deaf patients homozygous for c.-23+1G>A by using the same panel of SNPs (rs2313477, rs11841024, rs4769974, rs7994748, rs7987144, rs5030702, and rs1932429) as reported in the study by Tekin et al.and revealed the reduced c.-23+1G>A haplotype diversity in the analyzed sample when compared with the haplotypes in Mongolians. Interesting, that almost all examined patients (except one Yakut patient) in this study were homozygous for the allele T of intronic rs7994748 (GJB2) that is consistent with the studies byand by in which the association of this rs7994748 allele with hearing loss was presumed. In the study bywhere five SNPs (rs747931, rs3751385, rs11147592, rs9509086, and rs9552102) were used for the c.-23+1G>A haplotype analysis in six Mongolian deaf patients, two c.-23+1G>A haplotypes were identified: major haplotype G-G-C-T-A (9/12 chromosomes) and a minor haplotype A-G-C-T-A (3/12 chromosomes).
Our data on the common STR and SNP haplotypes for c.-23+1G>A found in Tuvinians evidence a single origin of this mutation and suggest the founder effect in its high prevalence in the Tyva Republic and neighboring territory of the Altai Republic. Based on the ethnic history of Tuvinians who experienced repeated influence of Mongolians at various stages of their ethnic formation, we speculate that c.-23+1G>A mutation can be introduced into Tuva by ancient Mongolic-speaking groups which were subsequently assimilated by the indigenous population of this region and then spread in Siberia by the migration flows. Our estimation of the c.-23+1G>A age yielded a wide range of 725-4100 years ago. This uncertainty could be probably attributed to a small size of the examined sample and an unclear population growth rate of Tuvinians in the past. Nevertheless, this estimation is consistent with previously reported age of c.-23+1G>A in the Sakha Republic (Yakutia) presumably introduced by Turkic-speaking ancestors of Yakuts approximately 800 years agofurther confirmed by the observed similarity of allelic composition of the common STR haplotypes in Tuvinian and Yakut patients homozygous for c.-23+1G>A (data not shown). Thus, our data support a proposed common Central Asian origin of mutation c.-23+1G>A and its further expansion defined by a specific population bottleneck at least throughout Siberia though further extensive studies in many populations are required to clarify this issue.
## The c.235delc mutation
High prevalence of c.235delC mutation was found in our previous study in the Altai Republicand was later confirmed in an extended cohort of Altaian deaf patients since alleles with c.235delC accounted for 51.9% of all mutant GJB2 alleles found in patients and the carrier frequency of c.235delC reached 3.7% in Altaian population sample.
According to numerous studies, c.235delC mutation prevails in patients with hearing loss in Asian populations (China, Japan, Mongolia, Korea). The founder effect, implying the origin of c.235delC from a common ancestor, was suggested for the explanation of high prevalence of c.235delC in Asia. Several studies focusing on the analysis of the haplotypes bearing c.235delC confirmed this hypothesis despite the certain differences between the sets of used genetic markers. Based on the STR and SNP analysis, we found only one haplotype associated with c.235delC in Altaian homozygous patients. It is worth noting that some SNPs included in c.235delC-associated haplotype in Altaians overlap with the SNP markers analyzed in other studies and alleles observed coincide with the ones found in Asian patients having c.235delC. We suggest that these findings are in favor of a common c.235delC-associated haplotype at least among Altaians, Mongolians, Chinese, and Japanese and accordingly, in favor of the origin of c.235delC from one ancestor. Additional studies using a unified panel of markers are needed to clarify the question.
As for the age of c.235delC, as far as we know, this issue was elucidated in only two studies. In the study by Yan et al. seven SNPs flanking this mutation were analyzed in deaf patients (in a total of 26 homozygotes and 19 heterozygotes for c.235delC) from various regions of Asia (China, Japan, Korea, Mongolia) and association of c.235delC with one core haplotype A-G-A-C (SNP2-V27I-E114G-SNP1), with a length of approximately 2.6 kb, was discovered. The allele T of the most distant marker SNP6 (rs747931) located at~63 kb from c.235delC was used to evaluate the mutation age resulting in 460 generations or approximately 11500 years (assuming 25 years per generation). Yan et al. speculated that c.235delC might have arisen in the Baikal area and then spread to Mongolia, China, Korea, and Japan through subsequent migration. In recent study by Shinagawa et al. the c.235delC-associated haplotypes were analyzed in total of 20 Japanese patients homozygous for c.235delC. Based on observed linkage disequilibrium for 5'SNP6 (rs4769920) located at~265 kb from c.235delC, the occurrence of c.235delC mutation was estimated around 6500 years ago. Notably, the single marker method was applied for c.235delC age estimation in both studies, while we could not estimate the age of c.235delC by this method due to the lack of recombination in c.235delC haplotype in Altaian patients. Our estimation by the DMLE+ led to the lower values of the age of c.235delC (22-126 generations or 550-3150 years at the different population growth rates). Although we do not exclude that c.235delC is really "younger" in Altaians in comparison with the data from these studies, these differences are more likely due to different methods of the age estimation, the panels of used genetic markers, the sample sizes, as well as uncertainty in growth rates of Altaian population along their history. Additionally, our data are based on the limited population of Southern Siberia (Altaians) whereas, for example, in the study bythe samples from various countries (Mongolia, China, Japan, and Korea) were analyzed.
# Conclusions
The common haplotypes specific for GJB2 mutations c.516G>C, c.-23+1G>A, and c.235delC imply a single origin for each of them. A crucial role of the founder effect in high prevalence of these mutations in indigenous populations of Southern Siberia was established.
Supplementary Materials: The following are available online at http://www.mdpi.com/2073-4425/11/7/833/s1, : Primer sequences and methods for STRs and SNPs genotyping.: The allelic frequencies of STRs (D13S1316, D13S141, D13S175, D13S1853, D13S143, D13S1275, D13S292) and SNPs (rs747931, rs5030700, rs3751385, rs2274083, rs2274084, rs1411911768, rs9552101, rs117685390, rs877098) in deaf patients homozygous for mutations c.516G>C, c.-23+1G>A or c.235delC and in the control samples (Tuvinians and Altaians). |
Managing bevacizumab-induced intraocular inflammation
The outcome of four cases of sterile endophthalmitis that developed after intravitreal injections of bevacizumab has been reported here. All four eyes received 1.25 mg/0.05 ml intravitreal bevacizumab from 0.2-ml aliquots for different etiologies. The inflammation predominantly involved the anterior chamber with mild vitreous reaction. All patients were culture negative and regained preinjection visual acuity and were culture negative following intravitreal antibiotic administration. This report highlights that intravitreal bevacizumab can cause sterile endophthalmitis and this has to be kept in mind, and clinical judgment should be used to differentiate it from infective endophthalmitis.Bevacizumab (Avastin; Basel, Switzerland) is a recombinant humanized monoclonal antibody that is directed against all isoforms of the vascular endothelial growth factor (VEGF). It is approved for intravenous administration for the treatment of metastatic colon cancer. In ophthalmology, bevacizumab has been used widely as an off-label treatment for ocular disorders like age-related macular degeneration (wet variety),[1]retinal vein occlusion, proliferative diabetic retinopathy,[2]diabetic macular edema, and retinopathy of prematurity.
Intravitreal bevacizumab (IVBe) can cause complications, including traumatic cataract, retinal detachment, and endophthalmitis, with the reported incidence of endophthalmitis ranging from 0.014% to 0.082%. [bib_ref] Complications in patients after intravitreal injection of bevacizumab, Shima [/bib_ref] [bib_ref] Infectious and noninfectious endophthalmitis after intravitreal bevacizumab, Jonas [/bib_ref] Recently, there have been a few reports of toxic anterior segment syndrome (TASS)-like culture-negative sterile endophthalmitis after IVBe injection for diverse etiologies. [bib_ref] Severe Intraocular Inflammation after intravitreal Injection of Bevacizumab, Sato [/bib_ref] [bib_ref] Sterile endophthalmitis after intravitreal injection of bevacizumab obtained from a single batch, Yamashiro [/bib_ref] [bib_ref] Acute intraocular inflammation after intravitreous injections of bevacizumab for treatment of neovascular..., Wickremasinghe [/bib_ref] We report similar cases occurring in four patients at our center. According to our knowledge, this is the first of its kind report from our country.
In our center, we routinely inject over 100 IVBe injections every month. [bib_ref] Intravitreal bevacizumab for subfoveal idiopathic choroidal neovascularization, Mandal [/bib_ref] During the last 3 months we have encountered four such cases of ocular inflammation within 3-5 days of administering 1.25 mg/0.05 ml IVBe injection. The purpose of this report is to emphasize the role of intravitreal antibiotics alone in patients with suspected post-IVBe endophthalmitis. This is important because recent reports in the literature seem to suggest that these patients need vitrectomy. [bib_ref] Severe Intraocular Inflammation after intravitreal Injection of Bevacizumab, Sato [/bib_ref] Routinely, bevacizumab was obtained from the manufacturer and was prepared by the authors' hospital pharmacy by a qualified pharmacist under sterile conditions. From the commercially available 4-ml vial containing 100 mg bevacizumab (Avastin Genentech, Inc., San Francisco, CA, USA) ,0.2-ml fractions were transferred under strict aseptic conditions (class 10 environment) into 2-ml glass ampoules which subsequently were flame sealed. During the entire formulation process, cold chain (2-8°C) was maintained. [bib_ref] Safety and cost-effectiveness of single dose dispensing of bevacizumab for various retinal..., Velpandian [/bib_ref] Bevacizumab (1.25 mg/0.05 ml) was injected into the vitreous cavity in a surgical room maintaining standard aseptic precautions. After the injection, all patients received topical antibiotic treatment with gatifloxacin and were instructed for routine examination the next day as per the protocol followed in our hospital. On day 1 after IVBe none of the patients had any evidence of ocular inflammation.
All the four cases presented 3-5 days after IVBe with a drop in visual acuity, and mild ciliary injection with severe anterior chamber reaction accompanied by mild vitreitis (grade 1 haze).
## Case report
A 65-year-old man with recurrent choroidal neovascular membrane who had previously received IVBe injection twice [] presented to us 3 days after injection with a vision hand movements associated with severe anterior chamber inflammation with mild vitreitis []. Before injection, his visual acuity was 5/200. The same day we performed vitreous tap with simultaneous intravitreal ceftazidime (2.25 mg/0.1 ml) and vancomycin (1 mg/0.1 ml) injections and started 0.5% moxifloxacin (Vigamox; Alcon Laboratories Inc.) prednisolone acetate 1% solution and homatropine 2% for cycloplegia. Gram stain and culture results were negative. One week later, he regained preinjection visual acuity and 4 weeks later, his visual acuity improved to 20/200 [].
All the other three patients were managed similarly. The clinical charts of these four patients were reviewed and the details are shown in [fig_ref] Table 1: Clinical findings of sterile endophthalmitis after intravitreal injection of bevacizumabIVBe [/fig_ref]. Three of the four patients had received previous IVBe, but none had developed intraocular inflammation before.
# Discussion
There are two views regarding the cause of sterile endophthalmitis that have been described by some of the authors. The first view is that it is caused by some endotoxin or breakdown product due to the faulty storage of bevacizumab. The second view is that it may be an immune-mediated response to bevacizumab, following repeated injections. [bib_ref] Acute intraocular inflammation after intravitreous injections of bevacizumab for treatment of neovascular..., Wickremasinghe [/bib_ref] The reports published documenting this type of ocular inflammation after IVBe have traced the drug to the same lot of vials of bevacizumab. [bib_ref] Severe Intraocular Inflammation after intravitreal Injection of Bevacizumab, Sato [/bib_ref] [bib_ref] Sterile endophthalmitis after intravitreal injection of bevacizumab obtained from a single batch, Yamashiro [/bib_ref] In our series of patients who developed severe intraocular inflammation after an IVBe injection, the cause remains to be determined. We believe that the cause seems to be sterile endophthalmitis similar to what is widely known to develop following intravitreal triamcinolone in some cases. Infective etiology was less likely as the culture was negative in all the cases and the reaction was more concentrated in the anterior chamber with only mild vitreitis in the form of grade 1-2 haze. Also all the four cases regained the preinjection visual acuity after intravitreal antibiotics and topical treatment.
In our center, multiple 4-ml vials of bevacizumab are usually dispensed at a time into 0.2-ml aliquoted glass ampoules containing 5 mg of the drug. [bib_ref] Safety and cost-effectiveness of single dose dispensing of bevacizumab for various retinal..., Velpandian [/bib_ref] So it was not possible for us to attribute the causative glass ampoule to the same lot of bevacizumab. But clustering of these four cases within a short span of time can point to the same lot.
Till date the causes of this type of sterile endophthalmitis occurring soon after IVBe remain unsolved. As discussed previously, it may be either of the two causes. Regarding management options also, there are differing views. Whereas Sato et al, have performed vitrectomy in all their cases of sterile endophthalmitis, [bib_ref] Severe Intraocular Inflammation after intravitreal Injection of Bevacizumab, Sato [/bib_ref] Yamashiro et al, have done vitrectomy in the most severely affected eyes and managed conservatively the less severely affected eyes with topical and systemic medications. [bib_ref] Sterile endophthalmitis after intravitreal injection of bevacizumab obtained from a single batch, Yamashiro [/bib_ref] It is difficult to differentiate sterile from infectious endophthalmitis, especially in the cases with severe anterior inflammation and hypopyon. There lies the importance of vitreous tap and culture for ruling out infective etiology.
We believe that the administration of intravitreal antibiotics along with topical antibiotics and steroids may be appropriate in dealing with this TASS-like picture following IVBe injection. With the rising graph of IVBe injections, the number of such cases is sure to rise, thus necessitating the need to find a prudent approach to deal with such situations.
[fig] Figure 1: (a) Fundus photograph showing the recurrent choroidal neovascular membrane with an activity at the edge in the form of retinal hemorrhage. (b) Fundus photograph 3 days after intravitreal bevacizumab injection showing vitreous haze obscuring fundus details. (c) Fundus photograph 4 weeks after the intravitreal injection showing clearing of the vitreous cavity with scarring of the neovascular membrane a b c [/fig]
[table] Table 1: Clinical findings of sterile endophthalmitis after intravitreal injection of bevacizumabIVBe: Intravitreal bevacizumab, AC: Anterior chamber, AMD: Age-related macular degeneration, CRVO: Central retinal vein occlusion, Neg.: Negative [/table]
|
Comparative efficacy and tolerability of 32 oral antipsychotics for the acute treatment of adults with multi-episode schizophrenia: a systematic review and network meta-analysis
This appendix formed part of the original submission and has been peer reviewed. We post it as supplied by the authors.Supplement to: Huhn M, Nikolakopoulou A, Schneider-Thoma J, et al. Comparative efficacy and tolerability of 32 oral antipsychotics for the acute treatment of adults with multi-episode schizophrenia: a systematic review and network meta-analysis. Lancet 2019; published online July 11. http://dx.
# Introduction
## Rationale
3 Describe the rationale for the review in the context of what is already known. Provide an explicit statement of questions being addressed, with reference to participants, interventions, comparisons, outcomes, and study design (PICOS).
## 3, evidence in context
3ff, Appendix 2 METHODS Protocol and registration Indicate whether a review protocol exists and if and where it can be accessed (e.g., Web address); and, if available, provide registration information, including registration number.
Appendix 2, PROSPERO (CRD4201401 4919) Eligibility criteria Specify study characteristics (e.g., PICOS, length of followup) and report characteristics (e.g., years considered, language, publication status) used as criteria for eligibility, giving rationale. State the process for selecting studies (i.e., screening, eligibility, included in systematic review, and, if applicable, included in the meta-analysis).
5ff, Appendix 2 Data collection process Describe method of data extraction from reports (e.g., piloted forms, independently, in duplicate) and any processes for obtaining and confirming data from investigators.
5ff, Appendix 2 Data items List and define all variables for which data were sought (e.g., PICOS, funding sources) and any assumptions and simplifications made.
5ff, Appendix 2, Appendix 18
Geometry of the network S1
Describe methods used to explore the geometry of the treatment network under study and potential biases related to it. This should include how the evidence base has been graphically summarized for presentation, and what characteristics were compiled and used to describe the evidence base to readers.
5ff, Appendix 4
Risk of bias within individual studies Describe methods used for assessing risk of bias of individual studies (including specification of whether this was done at the study or outcome level), and how this information is to be used in any data synthesis.
5ff, Appendix 2, 19
Summary measures State the principal summary measures (e.g., risk ratio, difference in means).
5ff, Appendix 4 Planned methods of analysis Describe the methods of handling data and combining results of studies for each network meta-analysis.
6ff, Appendix 4 Assessment of Inconsistency
## S2
Describe the statistical methods used to evaluate the agreement of direct and indirect evidence in the treatment network(s) studied. Describe efforts taken to address its presence when found.
6ff, Appendix 4, 13
Risk of bias across studies Specify any assessment of risk of bias that may affect the cumulative evidence (e.g., publication bias, selective reporting within studies).
6ff, Appendix 15, 19 Additional analyses Describe methods of additional analyses if done, indicating which were pre-specified. This may include, but not be limited to, the following:
- Sensitivity or subgroup analyses;
- Meta-regression analyses; 6ff, Appendix 9, 10, 16,17 Provide a brief overview of characteristics of the treatment network. This may include commentary on the abundance of trials and randomized patients for the different interventions and pairwise comparisons in the network, gaps of evidence in the treatment network, and potential biases reflected by the network structure.
# Appendix 14
Study characterist ics For each study, present characteristics for which data were extracted (e.g., study size, PICOS, follow-up period) and provide the citations.
List of included studies appendix 7 Risk of bias within studiesPresent data on risk of bias of each study and, if available, any outcome level assessment.
# Appendix 8
Results of individual studies [bib_ref] Olanzapine versus placebo: Results of a double-blind, fixed-dose olanzapine trial, Beasley [/bib_ref] For all outcomes considered (benefits or harms), present, for each study: 1) simple summary data for each intervention group, and 2) effect estimates and confidence intervals. Synthesis of results [bib_ref] Olanzapine versus placebo and haloperidol: Acute phase results of the north american..., Beasley [/bib_ref] Present results of each meta-analysis done, including confidence/credible intervals. If additional summary measures were explored (such as treatment rankings), these should also be presented. 10f, appendix 13
Risk of bias across studies [bib_ref] Olanzapine versus haloperidol: Acute phase results of the international double-blind olanzapine trial, Beasley [/bib_ref] Present results of any assessment of risk of bias across studies for the evidence base being studied.
10, appendix 15 Results of additional analyses [bib_ref] A double-blind controlled trial of pipotiazine, haloperidol and placebo in recently-hospitalized acute..., Bechelli [/bib_ref] Give results of additional analyses, if done 11, appendix 9 and 10 DISCUSSION Summary of evidence [bib_ref] Plasma levels and clinical effects of thioridazine and thiothixene, Bergling [/bib_ref] Summarize the main findings, including the strength of evidence for each main outcome; consider their relevance to key groups (e.g., healthcare providers, users, and policy-makers).
12ff Discuss limitations at study and outcome level (e.g., risk of bias), and at review level (e.g., incomplete retrieval of identified research, reporting bias).
13ff
# Conclusions 26
Provide a general interpretation of the results in the context of other evidence, and implications for future research.
13f FUNDING Funding 27
Describe sources of funding for the systematic review and other support (e.g., supply of data); role of funders for the systematic review. This should also include information regarding whether funding has been received from manufacturers of treatments in the network and/or whether some of the authors are content experts with professional conflicts of interest that could affect use of treatments in the network.
2, 15 PICOS = population, intervention, comparators, outcomes, study design.
# Appendix 2: protocol
## Protocol as published in prospero (registration number: crd42014014919)
Review title and timescale Review title
Give the working title of the review. This must be in English. Ideally it should state succinctly the interventions or exposures being reviewed and the associated health or social problem being addressed in the review.
Informing clinicians, patients and guidelines: network meta-analysis on 24 antipsychotic drugs and a broad ragne of important outcomes for schizophrenia.
## Original language title
For reviews in languages other than English, this field should be used to enter the title in the language of the review. This will be displayed together with the English language title. Review team details Named contact
The named contact acts as the guarantor for the accuracy of the information presented in the register record. To examine the comparative efficacy, acceptability, and tolerability of twelve second-and twelve first-generation antipsychotic drugs in schizophrenia by applying a network meta-analysis approach. Searches
Give details of the sources to be searched, and any restrictions (e.g. language or publication period). The full search strategy is not required, but may be supplied as a link or attachment.
1. We will search the Cochrane Schizophrenia Group Controlled Trials Register (compiled by regular systematic hand searches and searches of more than Register'. There will be no language restriction applied concerning the literature. As an exception we will exclude Chinese studies which often do not use appropriate randomization procedures and do not report their methods so that it is impossible to check on these issues.
## Url to search strategy
If you have one, give the link to your search strategy here. Alternatively you can e-mail this to PROSPERO and we will store and link to it.
I give permission for this file to be made publicly available Yes 18 Condition or domain being studied Give a short description of the disease, condition or healthcare domain being studied. This could include health and wellbeing outcomes.
## Schizophrenia 19 participants/population
Give summary criteria for the participants or populations being studied by the review. The preferred format includes details of both inclusion and exclusion criteria.
We will include adult people (age >= 18, no upper age limit, no restriction in setting, gender, ethnicity) with schizophrenia or related disorders (such as schizophreniform, or schizoaffective disorders). There is no clear evidence that the latter schizophrenia-like psychoses are caused by fundamentally different disease processes or require different treatment approaches. However, as in our previous report we will exclude studies in treatment resistant patients, in patients with predominant negative symptoms, in patients with concomitant medical illness, and studies in stable patients (mainly relapse prevention studies), because these are different patient populations and it is an important requirement of network meta-analysis to have reasonably homogeneous samples. Studies in which less than 20% of the participants were suffering from other psychiatric disorders than schizophrenia (e.g. depression or mental retardation) will be acceptable. We will include the trials irrespective of the diagnostic criteria used. It is a general strategy of the Cochrane Schizophrenia Group (CSG) to include not only studies that used specific diagnostic criteria such as ICD-10 or DSM-IV, because these criteria are not meticulously used in clinical routine either. This decision should increase generalizability.
20 Intervention(s), exposure(s)
Give full and clear descriptions of the nature of the interventions or the exposures to be reviewed
We will include all second-generation ("atypical") antipsychotic drugs available in Europe or the US (amisulpride, aripiprazole, asenapine, brexpiprazole, cariprazine, clozapine, iloperidone, lurasidone, olanzapine, paliperidone, quetiapine, risperidone, sertindole, ziprasidone, zotepine), placebo and a selection of first-generation ("typical", "conventional") antipsychotics (benperidol, chlopromazine, clopenthixol, flupenthixol, fluphenazine, fluspirilene, haloperidol, levomepromazine, loxapine, methotrimeprazine, molindone, penfluridol, perazine, perphenazine, pimozide, sulpiride, thioridazine, thiothixene, trifluoperazine, zuclopenthixol). Second-generation antipsychotics are in some countries such as the US or Germany nowadays the most frequently prescribed compounds, they are overall more costly (especially the most recent ones which still have patent protection) and they are thus obvious choices. One novel aspect of the network meta-analysis is the inclusion of first-generation antipsychotics. In addition to the reasons already mentioned Collaborating Centre for Drug Statistics" which they find most important for such a meta-analysis (http://www.whocc.no/atc_ddd_methodology/who_collaborating_centre/, detailed results can be found on our website www.cfdm.de/media/doc/Antipsychotic Survey.doc). We added benperidol and perazine which are frequently used FGAs in Germany. The selection includes FGAs from various classes, there are high-potency (e.g. haloperidol), mid-potency (zuclopenthixol, perphenazine) and low-potency (e.g. thioridazine) drugs, butyrophenones (e.g. benperidol, haloperidol), phenothiazines (e.g. fluphenazine), thioxantenes (zuclopenthixol) and a substituted benzamide (sulpiride). We will include all these compounds in any oral forms of administration (for example tablets or liquid). In fixed-dose studies we will only included target to maximum doses according to the International Consensus Study on Antipsychotic dose. We will include all flexible-dose studies, because these allow the investigators to titrate to the adequate dose for the individual patient. We will exclude depot formulations which are mainly used for long-term relapse prevention which is not the focus of this review.
## Comparator(s)/control
Where relevant, give details of the alternatives against which the main subject/topic of the review will be compared (e.g.
another intervention or a non-exposed control group).
In a network meta-analysis each treatment is compared with each other, therefore any treatment can be the comparator.
But placebo is a natural reference that will be used in our occasion. [bib_ref] Olanzapine versus haloperidol: Acute phase results of the international double-blind olanzapine trial, Beasley [/bib_ref] Types of study to be included Give details of the study designs to be included in the review. If there are no restrictions on the types of study design eligible for inclusion, this should be stated.
We will included open and blinded randomized controlled trials (RCTs) comparing one antipsychotic drug with another antipsychotic agent or placebo. Results from non double-blinded trials will be considered only for objective outcomes.
Trials in which antipsychotic drugs were used as an augmentation-or combination strategy will be excluded. In the case of cross-over studies we will use only the first cross-over phase to avoid the problem of carry-over effects which are very likely in schizophrenia. We will exclude cluster randomized trials due to the unit-of-analysis-problems associated with this design (it is anyhow unlikely that such studies on our question exist). We will also only include double-blinded studies for subjective outcomes because we recently showed that a lack of blinding can exaggerate differences between treatments in this area. For objective outcomes (e.g. weight gain) blinding is less of a problem. The minimum duration of follow-up will be 3 weeks as shorter trials are unlikely to find significant differences. In addition, short-term results (3 weeks-3 months weeks) and longer term results (>3 months) will be analysed in separate publications.
## Context
Give summary details of the setting and other relevant characteristics which help define the inclusion or exclusion criteria.
We will include adult people (age >= 18, no upper age limit, no restriction in setting, gender, ethnicity) with schizophrenia, schizophreniform or schizoaffective disorders with an acute exacerbation, primarily irrespective of the diagnostic criteria used. There is no clear evidence that the latter schizophrenia-like psychoses are caused by fundamentally different disease processes or require different treatment approaches. It is also a general strategy of the Cochrane Schizophrenia Group to include not only studies that used specific diagnostic criteria such as ICD-10 or DSM-IV, because these criteria are not meticulously used in clinical routine either. Studies in which less than 20% of the participants were suffering from other psychiatric disorders (e.g. depression or mental retardation) will be included. We will exclude studies in participants with no or only subclinical symptoms at baseline that are usually conducted to address the relapse preventing effects of antipsychotics, studies in patients with predominant negative symptoms and studies including exclusively participants with major concomitant somatic illness or psychiatric disorders (e.g. substance abuse). [bib_ref] Plasma levels and clinical effects of thioridazine and thiothixene, Bergling [/bib_ref] Primary outcome(s)
Give the most important outcomes.
Mean reduction in overall symptoms of schizophrenia Give information on timing and effect measures, as appropriate.
The primary outcome will be overall symptoms of schizophrenia as measured by rating scales such as the Positive and Negative Syndrome Scale (PANSS), the Brief Psychiatric Rating Scale (BPRS) or of any other validated scale (e.g. the Manchester Scale) for the assessment of overall schizophrenic symptomatology. Overall symptoms of schizophrenia as measured by such scales was the primary outcome in numerous previous systematic reviews. As not all studies will have used the same scale, we will apply the following hierarchy: mean change of the PANSS total score from baseline to endpoint, if not available mean change of the BPRS, or if again not available the mean values at endpoint of the PANSS/ BPRS. The results of other rating scales will only be used if the instrument has been published in a peer-reviewed journal, because it has been shown that unvalidated schizophrenia scales exaggerate differences. The minimum duration of followup will be 3 weeks as shorter trials are unlikely to find significant differences. Outcomes will be classified into short-term results (3 weeks-3 months) where the primary time point will be six weeks, if available, and longer term results (>3 months).
## Secondary outcomes
List any additional outcomes that will be addressed. If there are no secondary outcomes enter None. Give information on timing and effect measures, as appropriate.
The minimum duration of follow-up will be 3 weeks as shorter trials are unlikely to find significant differences. Outcomes will be classified into short-term results (3 weeks-3 months) where the primary time point will be six weeks, if available, and longer term results (>3 months). [bib_ref] Amisulpride versus olanzapine in the treatment of schizophrenia in indian patients: Randomized..., Bhowmick [/bib_ref] Data extraction (selection and coding) Give the procedure for selecting studies for the review and extracting data, including the number of researchers involved and how discrepancies will be resolved. List the data to be extracted.
1. Selection of trials: Two reviewers will independently inspect all abstracts identified in the searches. Disagreement will be resolved by discussion, and where doubt still remains, we will acquire the full article for further inspection. Once the full articles are obtained, at least two reviewers will independently decide whether the studies meet the review criteria. If disagreement can not be resolved by discussion, we will resolve it with a third reviewer or seek further information from the study authors. 2. Data extraction: Two reviewers will independently extract data from all selected trials on electronic forms. When disagreement arises we will resolve it by discussion with a third reviewer. Where this is not possible we will contact the study authors.
## Risk of bias (quality) assessment
State whether and how risk of bias will be assessed, how the quality of individual studies will be assessed, and whether and how this will influence the planned synthesis.
Study quality in terms of sequence generation, allocation concealment, blinding, the completeness of outcome data, selective reporting and other biases will be assessed with the Cochrane Collaboration risk of bias tool.
28 Strategy for data synthesis Give the planned general approach to be used, for example whether the data to be used will be aggregate or at the level of individual participants, and whether a quantitative or narrative (descriptive) synthesis is planned. Where appropriate a brief outline of analytic approach should be given.
1. Two-step procedure In a first step we will perform series of conventional pair-wise meta-analyses by combining studies that compared the same interventions. In a second step we will then perform a network meta-analysis within a Bayesian framework. 2. Continuous outcomes: The effect size measure for continuous outcomes will be the standardized mean difference (SMD) because we expect that the studies use different rating scales of overall schizophrenia symptomatology, especially the Positive and Negative Syndrome Scale (PANSS) or the Brief Psychiatric Rating Scale (BPRS) (see outcomes, above). Intention-to-treat (ITT) data will be used whenever available. Missing standard deviations: When standard errors instead of standard deviations (SD) are presented, the former will be converted to standard deviations (SDs). If both are missing we will estimate SDs from confidence intervals, t-values, or p-values as described in Section 7.7.3 of the Cochrane Handbook for Systematic Reviews. If none of these options is viable we will contact the original authors. When no information can be obtained we will derive SDs from those of the other studies using a validated imputation technique. 3. Dichotomous outcomes: The effect size for dichotomous outcomes will be the odds ratio (OR) and its 95% confidence intervals (CIs). The main reason to prefer odds ratios to relative risks is that this measure has mathematical properties that make it more appropriate for network meta-analysis (e.g. the odds ratio is symmetrical).
Another reason justifying the calculation of the odds ratios is the expectation that different definitions of 'response to treatment' will be used in the original trials and in such a situation the odds ratio has been shown to yield the most consistent results which are largely independent from the response cut-off used. Therefore, although the relative risk is more intuitive for clinician, the odds ratio has clear advantages for the purpose of our review. Analyses will be carried out in accordance to the 'intention-to-treat' principal when possible ('once randomized always analyze'). Everyone allocated to the intervention will be counted whether they completed the follow up or not. If the authors applied such a strategy, we will use their results. If the original authors presented only the results of the per-protocol or completer population, we will assume that those participants lost to follow-up would not have changed in a given outcome. In terms of efficacy this means that they would be conservatively considered to have not responded to treatment. In terms of tolerability it would mean that participants would not have developed a side-effect which we feel is appropriate, because otherwise side-effects, many of which are rare, would be overestimated. Applying this approach led to meaningful results in our previous reports on which we build. 4. Assessment of heterogeneity The heterogeneity (variability in relative treatment effects within the same treatment comparison) will be measured with the tau-squared (the variance of the random effects distribution). The heterogeneity variance will be assumed common across the various treatment comparisons and the empirical distributions will be used to characterise the amount of heterogeneity as low, moderate or high using the first and third quantiles at least compared to placebo, are the degree of placebo response (which has increased over the years) and industry sponsorship. Less robust factors include severity of illness at baseline, gender, chronicity and publication year. We will investigate if these variables are similarly distributed across studies grouped by comparison, whereas it is clear a priori that publication year, a composite of various factors, will differ between older and more recent antipsychotics. We will consider that placebo response in schizophrenia has increased over the years and that there could be differences between placebo-controlled trials and head-to-head trials as it is known from antidepressant trials in major depressive disorder. 6.
Network meta-analyses We assume that patients who fulfill the inclusion criteria outlined above are equally likely to be randomised to any of the antipsychotic that we plan to compare. If the collected studies appear to be sufficiently similar with respect to the distribution of effect modifiers (refer to "assessment of transitivity assumption" section), we will conduct a random effects NMA to synthesize all evidence for each outcome, and obtain a comprehensive ranking of all treatments. We will use arm-level data and the binomial likelihood for dichotomous outcomes. We will account for the correlations induced by multi-arm studies by employing multivariate distributions. We will assume a single heterogeneity parameter for each network. We will present the summary ORs or SMD for all pairwise comparisons in a league table. We will also estimate the prediction intervals to assess how much the common heterogeneity affects the relative effect with respect to the extra uncertainty anticipated in a future study. To rank the various treatments for each outcome, we will use the surface under the cumulative ranking curve (SUCRA) and the mean ranks. 7. Assessment of inconsistency The strategical and conceptual evaluation of transitivity will be supplemented with a statistical evaluation of consistency, the agreement between direct and indirect evidence. We will employ local as well as global methods to evaluate consistency.
Local methods detect 'hot spots' of inconsistency, evidence loops that are inconsistent or comparisons for which direct and indirect evidence disagree. We will employ the loop-specific approach to evaluate inconsistency within each loop of evidence, and a method that separates direct evidence from indirect evidence provided by the entire network. We will also evaluate consistency in the entire network by calculating the I2 for network heterogeneity, inconsistency, and for both.
Tests for inconsistency are known to have low power, and empirical evidence has suggested that 10% of evidence loops published in the medical literature are expected to be inconsistent. Therefore, interpretation of the statistical inference about inconsistency will be carried out with caution and possible sources of inconsistency will be explored even in the absence of evidence for inconsistency. 8. Exploring heterogeneity and inconsistency and sensitivity analyses We expect small amounts of heterogeneity and inconsistency to be present given the variety of study settings we plan to include. We will explore whether treatment effect for the primary outcome is robust in subgroup analyses and network meta-regression using the characteristics presented under 'analysis of subgroups or subsets'. 9. Publication bias We will explore the association between study size and effect size with a comparison-adjusted funnel plot that has been adapted to network meta-analysis. . Statistical software The analysis and presentation of results will be performed using the Stata packages network and network_graphs, the R package netmeta and self-programmed codes in OpenBUGS.
## Analysis of subgroups or subsets
Give any planned exploration of subgroups or subsets within the review. 'None planned' is a valid response if no subgroup analyses are planned.
The following potential effect moderators of the primary outcome will be explored by subgroup or meta-regression analysis: 1. Dose of the antipsychotics in olanzapine-equivalents according to Give brief details of plans for communicating essential messages from the review to the appropriate audiences.
The results will be published in major psychiatric journals and presented at major international and German psychiatric conferences. Our findings will be rapidly implemented in national and international treatment guidelines, for some of which Stefan Leucht is a co-author. 25. Secondary outcomes: we shortened the section and removed some outcomes 28. Strategy for data synthesis: we described our statistical approach in more detail [bib_ref] Loxapine: A controlled evaluation in chronic schizophrenic patients, Bishop [/bib_ref]. Analysis of subsets: we removed some planned subgroup analysis, because we realized that some of the proposed analyses are not meaningful.
## Differences between protocol and review
Due to requests by peer-reviewers of the manuscript, a number of changes and additions to the original protocol had to be made which are summarised in the following text:
- One reviewer requested risk ratios instead of odd´s ratios for dichotomous outcomes and weighted mean differences instead of standardized mean differences for continous outcomes. We followed these suggestions and updated the analyses accordingly.
- One reviewer requested a sensitivity analysis exluding trials published before 1990
- One reviewer requested a sensitivty analysis exluding "failed trials"
- The following outcomes were requested by the reviewers:
[formula] ---------------------------------------------------------------------------------------------------------------------------------------------------------------- [/formula]
## Who ictrp 10-01-19
Amisulpride and schizo* and random* = 1 Aripiprazole and schizo* and random* = 5 Asenapine and schizo* and random* = 1 Benperidol and schizo* and random* = 0 Brexpiprazole and schizo* and random* = 2 Cariprazine and schizo* and random* = 2 Chlorpromazine and schizo* and random* = 1 Clopenthixol and schizo* and random* = 0 Clozapine and schizo* and random* = 1 Flupenthixol and schizo* and random* = 0 Fluphenazine and schizo* and random* = 0 Fluspirilene and schizo* and random* = 0 Haloperidol and schizo* and random* = 1 Iloperidone and schizo* and random* = 0 Levomepromazine and schizo* and random* = 0 Loxapine and schizo* and random* = 0 Lurasidone and schizo* and random* = 6 Molindone and schizo* and random* = 0 Olanzapine and schizo* and random* = 6 Paliperidone and schizo* and random* = 10 Quetiapine and schizo* and random* = 4 Penfluridol and schizo* and random* = 0 Perazine and schizo* and random* = 0 Perphenazine and schizo* and random* = 0 Pimozide and schizo* and random* = 0 Risperidone and schizo* and random* = 8 Sertindole and schizo* and random* = 0 Sulpiride and schizo* and random* = 0 Thioridazine and schizo* and random* = 0 Thiothixene and schizo* and random* = 0 Trifluoperazine and schizo* and random* = 7 Ziprasidone and schizo* and random* = 1 Zotepine and schizo* and random* = 0 Zuclopenthixol and schizo* and random* = 0 Amisulpride and psycho* and random* = 1 Aripiprazole and psycho* and random* = 2 Asenapine and psycho* and random* = 0 Benperidol and psycho* and random* = 0 Brexpiprazole and psycho* and random*= 0 Cariprazine and psycho* and random* = 0 Chlorpromazine and psycho* and random* = 2 Clopenthixol and psycho* and random* = 0 Clozapine and psycho* and random* = 0 Flupenthixol and psycho* and random* = 0 Fluphenazine and psycho* and random* = 0 Fluspirilene and psycho* and random* = 0 Haloperidol and psycho* and random* = 2 Iloperidone and psycho* and random* = 0 Levomepromazine and psycho* and random* = 0 Loxapine and psycho* and random* = 0 Lurasidone and psycho* and random* = 2 Molindone and psycho* and random* = 0 Olanzapine and psycho* and random* = 3 Paliperidone and psycho* and random* = 1 Quetiapine and psycho* and random* = 2 Penfluridol and psycho* and random* = 0 Perazine and psycho* and random* = 0 Perphenazine and psycho* and random* = 0 Pimozide and psycho* and random* = 0 Risperidone and psycho* and random* = 1 Sertindole and psycho* and random* = 0 Sulpiride and psycho* and random* = 0 Thioridazine and psycho* and random* = 0 Thiothixene and psycho* and random* = 0 Trifluoperazine and psycho* and random* = 16 Ziprasidone and psycho* and random* = 0 Zotepine and psycho* and random* = 0 Zuclopenthixol and psycho* and random* = 0 Total = 88
## Appendix 4: statistical methods in detail
## Measures of treatment effect
The effect size for dichotomous outcomes was the odds ratio (OR) and its 95% confidence intervals (CIs).
A reviewer requested risk ratios (RR) as effect size for dichotomous outcomes, as these are easier to understand for clinicians. We followed the reviewer´s suggestion and changed the effect size for dichotomous otucomes to RRs. Everyone allocated to the intervention was counted whether they completed the follow up or not. If the authors applied such a strategy, we used their results. If the original authors presented only the results of the per-protocol or completer population, we assumed that those participants lost to follow-up would not have changed in a given outcome. In terms of efficacy this means that they were conservatively considered to have not responded to treatment. In terms of tolerability it would meant that participants would not have developed a side-effect which we think is appropriate, because otherwise side-effects, many of which are rare, would have been overestimated.
The effect size measure for continuous outcomes was the standardized mean difference (SMD) because we expected that the studies use different rating scales of overall schizophrenia symptomatology, especially the Positive and Negative Syndrome Scale (PANSS) or the Brief Psychiatric Rating Scale (BPRS). Missing standard deviations: When standard errors instead of standard deviations (SD) were presented, the former was converted to standard deviations (SDs). If both were missing we estimated SDs from confidence intervals, t-values, or p-values as described in Section 7.7.3 of the Cochrane Handbook for Systematic Reviews. When no information could be obtained we derived SDs from those of the other studies using a validated imputation technique. 2
## Statistical details for network meta-analysis and meta-regression models
We used standard network meta-analysis and meta-regression models fitted in JAGS using the rjags package in R. We accounted for the correlations induced by multi-arm studies by employing multivariate distributions. We assumed a single heterogeneity parameter for each network. We presented the summary RRs or SMD for all pairwise comparisons in a league table. We also estimated the prediction intervals to assess how much the common heterogeneity affects the relative effect with respect to the extra uncertainty anticipated in a future study. To rank the various treatments for each outcome, we used the surface under the cumulative ranking curve (SUCRA) and the mean ranks. If data was too sparse for a network metaanalytic approach, we did a simple pairwise meta-analysis using random effects model.
The heterogeneity (variability in relative treatment effects within the same treatment comparison) was measured with the tau-squared (the variance of the random effects distribution). The heterogeneity variance was assumed common across the various treatment comparisons and the empirical distributions were to characterise the amount of heterogeneity as low, moderate or high using the first and third quantiles 3 . Potential reasons for heterogeneity were explored by subgroup analysis and meta-regressions.
For every study i = 1, … , N we denote with , the symptoms score (either final or change from baseline) in arm k = 1, … , , , the standard deviation and , is the sample size in arm . Then the normal likelihood is employed to y i,k~N (φ i,k , (sd i,k ) 2 /n i,k ) in each arm and then we parametrize to the study-
[formula] specific standardised mean difference , * logit(φ i,1 ) = / logit(φ i,k ) = (u i + , * ) / for ≥ 2, [/formula]
where is the pooled standard deviation in the study arms.
# Network meta-analysis:
For the model without covariates we set , * = , and ,2 , … , ,~( 2 , … , , )
The matrix involves the heterogeneity standard deviation τ.
Each 2 , … , corresponds into a treatment comparison, e.g. ≡ ; then we assume consistency by putting = − , with = 0 for an arbitrarily selected reference treatment.
## Network meta-regression:
In the network meta-regression models we set , * = , + 1, × ( − )
In the model with independent and consistent coefficients we define 1, = if treatment 1 is placebo; otherwise 1, ≡ = − for any treatments , .
In the model with exchangeable coefficients we set ~( , 2 ).
## Network meta-regression for placebo response:
In placebo-controlled studies the covariate is defined as the improvement in symptoms on the PANSS scale in the placebo arm. This variable is unobserved in head-to-head studies. Previous research has shown a strong association between year of study ( ) and placebo response, hence the values were stochastically imputed for head-to-head studies using the following process, nested within the metaregression model.
[formula] ~( , 2 ) = + × [/formula]
Note: in all meta-regression analyses we excluded drugs that have been studied in less than 100 participants to ensure convergence of the models.
Priors:
We assume vague normal priors (0,10 4 ) for the paraemters and , , , ….
and (if assumed independent) or (if are assumed exchangeable), and .
For variance parameters we employed: ~(0,5), ~(0,1), 1/ 2~( 0.001,0.001)
All models were run using 100 000 iterations after an initial burn-in of 10 000; a thinning of 10 was used.
Convergence was evaluated by monitoring the mixing of several chains with different initial values.
## Assessment of the transitivity assumption
We assumed that patients who fulfilled the inclusion criteria outlined above are equally likely to be randomised to any of the antipsychotic that we planned to compare. Nevertheless we inspected the distribution of potential effect modifiers with boxplots (appendix 5). We investigateed the distribution of clinical and methodological variables that can act as effect modifiers across treatment comparisons. The main features, which have been demonstrated to date to moderate efficacy of antipsychotics, at least compared to placebo, were the degree of placebo response (which has increased over the years) and
industry sponsorship. Less robust factors included severity of illness at baseline, gender, chronicity and publication year. We investigated if these variables were similarly distributed across studies grouped by comparison, whereas it was clear a priori that publication year, a composite of various factors, will differ between older and more recent antipsychotics. We considered that placebo response in schizophrenia has increased over the years and that there could be differences between placebo-controlled trials and headto-head trials as it is known from antidepressant trials in major depressive disorder.
## Assessment of inconsistency
The strategical and conceptual evaluation of transitivity was supplemented with a statistical evaluation of consistency, the agreement between direct and indirect evidence. We employed local as well as global methods to evaluate consistency. Local methods detected 'hot spots' of inconsistency, evidence loops that were inconsistent or comparisons for which direct and indirect evidence disagreed. We employed the loop-specific approach to evaluate inconsistency within each loop of evidence, and a method that separated direct evidence from indirect evidence provided by the entire network.
## Appendix 5: assessment of transitivity
Before conducting the statistical analysis we assessed whether the trials included in the NMA were on average similar in terms of characteristics that might modify the treatment effect (so that the transitivity assumption is plausible). Indirect comparisons, in contrast to direct comparisons, are not protected by randomisation and may be confounded by differences between the trials. In our analysis we deemed the following parameters as possible confounders: placebo response, publication year, mean age, baseline severity, percentage of male participants and number randomized based on results from a prior analysis. The plausibility of the transitivity assumption was evaluated by comparing the distribution of these potential effect modifiers across studies grouped by comparison. The impact of the individual parameters was assessed by several metaregressions (appendix 9) and sensitivity analyses (appendix 10).
## Placebo response
To examine the distribution of placebo response over the individual antipsychotics, we examined the change in the "Positive and Negative Symptom Scale of Schizophrenia" (PANSS) points from baseline to endpoint of the placebo arms. As expected the range of placebo response was large between -28 points and 17 points on the PANSS scale. In older studies patients responded only a little in the placebo arm, if at all (e.g. chlorpromazine, median five points PANSS worse). In contrast participants in placebo arms of newer compounds responded considerably (e.g. brexpiprazole, median 14 points PANSS better). The median response in placebo arms was six points on the PANSS scale. The impact of the placebo response was therefore examined in several metaregressions (appendix 9.1) and sensitivity analyses (appendix 10). The boxplot presents the change in the "Positive and Negative Symptom Scale of Schizophrenia"
(PANSS) points from baseline to endpoint of the placebo comparators of the individual antipsychotics. If original data was reported on the "Brief Psychiatry Rating Scale", data was transformed using validated formulas. 5
## Publication year
We examined the distribution of publication year over the individual antipsychotics. The overall range was between 1967 and 2018 with a median of 2000. Obviously, the individual compounds differed considerably with regards to publication year, because some substances were only approved recently. We examined publication year as a potential effect modifier, because it can be a proxy parameter for a number of factors that may have changed over the years (e.g. changes in trial design, monitoring, trial populations etc), that could have possibly influenced the effect sizes.
## Sample size
We examined the distribution of sample size over the individual antipsychotics. The overall sample size range was 4 to 1336 with a median of 50. Overall there were smaller sample sizes in older antipsychotics.
Sample size is accounted for by more weight given to larger studies in meta-analysis. In addition, we inspected the effects of sample size as a potential effect modifier.
## Baseline severity
We Baseline severity values are presented using the "Positive and Negative Symptom rating scale of schizophrenia"=PANSS. If original data was reported on the "Brief Psychiatry Rating Scale", data was transformed using validated formulas. 5
## Mean age
We examined the distribution of mean age over the individual antipsychotics. The overall median mean age was 38 with most antipsychotics between the age of 30 and 40.
## Changes in heterogeneity
Below we present the results from the changes in heterogeneity in each meta-regression model.
## Placebo response
The study publication year ranges between 1967 and 2018. At the same time, the response to placebo (measured as change in PANSS in the placebo arm) increases over time as shown in the graph below, as also shown in a previous publication. 4
## Fig. 9.1.1: change of placebo response over time
The mean change in the PANSS score in the placebo arm was -6.62 with median -6.40 and IQR -12.60 to -2.00.
To address whether and how much the degree of placebo-response had and impact on the results, we fit two network meta-regression models where the standardized mean difference (SMD) was a function of the change in PANSS for the placebo arm as described in the methods (appendix 4).
In one analysis we assumed a common coefficient across drugs, because as shown in [fig_ref] 11. 9: Prolactin elevationTreatments are ranked according to their surface under the curve cumulative... [/fig_ref].1.2 below the direction of the effect (except for cariprazine) was always the same. In the second analysis we used specific coefficients for each compound. In both analyses in studies without a placebo-arm, we fit a hierarchical model where a response to placebo value was stochastically imputed as a prediction from the year of the study publication.
In the following figure 9.1.2 we present both the drug-specific and the mean coefficients β to assess the impact of placebo-response: In [fig_ref] 11. 9: Prolactin elevationTreatments are ranked according to their surface under the curve cumulative... [/fig_ref].1.3 we present the result of the first model using drug specific coefficients. The SMDs presented below are the intercepts from the meta-regression model and they represent the SMDs for an imaginary situation where patients in placebo change their PANSS score from baseline by -6 units which was the median placebo response in the studies. We see that overall the hierarchies do not change much. In [fig_ref] 11. 9: Prolactin elevationTreatments are ranked according to their surface under the curve cumulative... [/fig_ref].1.4 we present the results using a common coefficient for all drugs. Here, the hierarchy of the effect sizes is even less changed than in the analysis with drug specific coefficients:
## Publication year
When the model was adjusted for publication in year 2016, loxapine as an older drug had a lower SMD compared to the unadjusted model. The remaining hierarchy was similar to the unadjusted model. When the model was adjusted for an infinetely large study, no significant changes in the hierarchy were observed.
## Age
When the model was adjusted for mean age 40, some drugs had higher SMDs, but the hierachy from the unadjusted model remained.
## Male participants
When the model was adjusted for 50% male participants, no significant changes in the hierarchy were observed. [fig_ref] 11. 9: Prolactin elevationTreatments are ranked according to their surface under the curve cumulative... [/fig_ref].6: Forest plot overall change in symptoms adjusted for 50% male participants Treatments are ranked according to standardised mean difference (SMD) compared to placebo. Treatments crossing the y-axis are not significantly different from placebo.
## Sponsorship
When the model was adjusted for the situation where all drugs in a trial are sponsored or not sponsoredso that the sponsoring interests do not bias the result towards any of the arms-no significant changes in the hierarchy were observed.
## Appendix 10: sensitivity analyses
To check for factors that could bias the effect we removed studies due to these criteria in several sensitivity analyses.
## Changes in heterogeneity
Below we present the results from the changes in heterogeneity in each sensitivity analysis.
# Sensitivity analysis
## Removing placebo controlled trials altogether
As it can be seen from the table above to remove the placebo groups from the analysis had the most important impact on the results (approximately 55% reduction of the variance). We therefore analysed the impact of removing placebo controlled studies in two ways.
In the first approach we excluded placebo-controlled studies altogether.
## Removing only placebo arms from three or more arm trials
In the second approach we excluded only the placebo arms, but left the antipsychotic arms of three (or more) arm trials from the analysis. After excluding placebo controlled trials, the hierarchy did not change considerably.
## Risk of bias
After excluding trials with high overall risk of bias, the hierarchy did not change significantly.
## Trial duration
After excluding trials with less than four or more than eight weeks duration, the hierarchy did not change significantly. [fig_ref] Figure 1: Appendix 14 Summary of network geometry S4 Provide a brief overview... [/fig_ref].4: Exclusion of trials with less than four or more than eight weeks duration Treatments are ranked according to standardised mean difference (SMD) compared to placebo. Treatments crossing the y-axis are not significantly different from placebo.
## Imputed standard deviations
After excluding trials with imputed standard deviations, loxapine had a lower SMD as in the unadjusted model, but the remaining hierarchy did not change significantly.
## Intention to treat analysis
After excluding studies, that did not use ITT, the hierarchy did not change significantly.
## Unfair dose
Unfairs doses were defined as the olanzapineequivalent 50% lower or higher as the comparator. . For this analysis we excluded studies in which the lowest dose was 50% less than the largest dose olanzapinequivalents. In case a study compared haloperidol 2mg (olanzapine equivalence dose =4mg) to olanzapine 10mg we excluded the study. We calculated the mean olanzapine equivalence doses for each drug arm according to the "International Consensus Study on Antipsychotic dose". 6 [fig_ref] Figure 1: Appendix 14 Summary of network geometry S4 Provide a brief overview... [/fig_ref].7: Exclusion of trials with unfair doses Treatments are ranked according to standardised mean difference (SMD) compared to placebo. Treatments crossing the y-axis are not significantly different from placebo.
## Older studies
We assumed that older studies are different concerning patient charateristics and study quality. To check if this factors influence the results considerably we excluded studies published before 1990. After excluding these studies the hierarchy did not change considerably. Treatments are ranked according to standardised mean difference (SMD) compared to placebo. Treatments crossing the y-axis are not significantly different from placebo.
## Failed trials
In recent years some trials examining new compounds were termed "failed". In these trials well established efficacious compounds as olanzapine did not separate significantly from placebo. After excludsion of these trials the hierarchy did not change considerably. Appendix 12: Quality of life: Results from the pairwise metaanalysis [fig_ref] Figure 1: Appendix 14 Summary of network geometry S4 Provide a brief overview... [/fig_ref] : Results for quality of life
As results for quality of life showed a high inconsistency, we present the result from the pairwise comparisons, separated by drugs. SMD: standardized mean difference,. CI: Confidence interval, SD=standard deviation.
## Appendix 13: evaluation of heterogeneity and inconsistency
We inferred the magnitude of heterogeneity by comparing the estimated 2 to empirical distributions of heterogeneity typically found in meta-analyses. The predictive τ² distribution for mental health outcomes according to Rhodes et al has a median of 0.049 and an IQR of 0.01 to 0.242. Low heterogeneity could be considered when the estimated tau2 is less than the 25% quantile of the empirical distribution, moderate heterogeneity for τ² between 25% and 50% quantile and high heterogeneity for τ² larger than the 50% quantile. We evaluated global consistency under the assumption of a full design-by-treatment interaction model, using the decompose.design function in R package netmeta.
## Outcome
## Negative symptoms
## Depressive symptoms
## Figure 14.3: network plot depressive symptoms
The size of the nodes corresponds to the number of participants randomized to each treatment. Treatments with direct comparisons are linked with a line; its thickness corresponds to the number of trials evaluating the comparison. 14.14 Anticholinergic side effects [fig_ref] Figure 1: Appendix 14 Summary of network geometry S4 Provide a brief overview... [/fig_ref].14: Network plot anticholinergic side effects The size of the nodes corresponds to the number of participants randomized to each treatment. Treatments with direct comparisons are linked with a line; its thickness corresponds to the number of trials evaluating the comparison.
## Quality of life
## All-cause discontinuation
## Weight gain
## Use of antiparkinson medication
## Qtc prolongation
## Across-study bias (publication bias):
Our search was comprehensive including study registers and pharmaceutical registers. So even if we missed small unpublished trials it seems not very likely that they would influence the results strongly. Nevertheless we assessed publication bias with the following procedure. We conducted funnel-plots for antipsychotics vs. placebo and for haloperidol versus all other antipsychotics. Haloperidol was chosen as the reference because it had served as a comparator antipsychotic for many older and newer antipsychotics.
1) Antipsychotics vs. placebo:
We first produced funnel plots for every outcome comparing all antipsychotics against placebo. If an overall publication bias was detected we assumed publication bias for all comparisons except for the following situation: In case more than 10 studies were available for the respective comparison (e.g. haloperidol vs. placebo), and there was no evidence for publication bias by visual inspection, we did not downgrade for publication bias. Funnel plots with less than ten studies are not meaningful ). All other comparisons were downgraded one level for publication bias. This procedure is illustrated below.
## Figure 1: evaluation of publication bis for placebo comparisons
2) Antipsychotics vs. haloperidol:
We produced funnel plots for every outcome comparing haloperidol with all other antipsychotics pooled.
If an overall publication bias was detected we assumed publication bias for all comparisons except for the following situation: In case more than 10 studies were available for the respective comparison (e.g. haloperidol vs. risperidone), and there was no evidence for publication bias by visual inspection, we did not downgrade for publication bias. Funnel plots with less than ten studies are not meaningful . All other comparisons were downgraded one level for publication bias. This procedure is illustrated below.
## Figure2: evaluation of publication bis for placebo comparisons
## Indirectness:
To control for transitivity studies with treatment-resistant patients, predominant negative symptoms, first episode, children or elderly were excluded a priori, as they are known to be different from the adult acute episode patient. Therefore no indirectness was assumed and no comparison was downgraded for this reason.
## Imprecision:
For placebo comparisons the clinically meaningful threshold was set at a standardized mean difference of higher or lower than 0. For dichotomous outcomes an odds ratio lower or higher than 1 was considered clinically meaningful. If the confidence interval crossed the threshold the comparison was downgraded one level. For comparisons of two antipsychotics the clinically meaningful threshold was set at standardized mean differences of -0.1 and 0.1 for continuous outcomes and at odds ratio of 0.8 and 1.25 for dichotomous outcomes. If the confidence interval crossed one threshold the comparison was downgraded one level. Crossing both thresholds resulted in a downgrading of two levels.
## Heterogeneity:
We made use of prediction intervals to prepopulate judgments on heterogeneity and its implications on the quality of the network treatment effects. In particular, we judged the agreement of conclusions based on confidence and prediction intervals in relation to the clinically important effect size.
The same thresholds for clinically meaningful thresholds as above were used and the recommendations automatically provided by CINeMA were followed (http://cinema.ispm.ch/#doc).
## Incoherence:
We assessed incoherence both globally using the design-by-treatment interaction model and locally using the Separate Indirect from Direct Evidence (SIDE, or node-splitting approach). We did not downgrade comparisons that had only direct evidence with respect to incoherence. The same judgement was made for comparisons that had both direct and indirect evidence with the contribution of direct evidence being more than 90%. For comparisons that had only indirect evidence, we downgraded due to incoherence one or two levels depending on whether the p-value of the design by treatment interaction model was between 0.01 and 0.10 or less than 0.01 respectively. We use the rules described in the table below to infer about our confidence regarding incoherence in network treatment effects informed by less than 90% from direct evidence.
## Confidence in evidence for all antipsychotics compared to placebo
## Figure 15.2: confidence in evidence for all antipsychotics compared to placebo
In this figure we present a summary of the confidence in the evidence for the individual drugs compared to placebo according to CINeMA (Confidence in Network Meta-analysis). The 13 outcomes presented in [fig_ref] Figure 2: 3 and 4 [/fig_ref] were considered. The bars present the percentage of outcomes with each evidence level (e.g. for olanzapine 23% of the reported outcomes had a high evidence level, 54% a moderate, 8% a low and 15% a very low one). The antipsychotics with the largest proportion of outcomes ranked with high certainty of evidence are presented on top (e.g. for risperidone 31% of the outcomes had a high level of evidence). As CINeMA does not consider comparisons for which no data are available, we added this information in the white bars (e.g. for thiothixene for 46% of the outcomes no data were available at all). Colour code: green=high, blue=moderate, orange=low, red=very low, white=percentage of outcomes with no data available.
## Cinema for the primary outcome "overall change in symptoms"
## Studies with imputed standard deviation
In case standard deviations were not given by the study authors, we re-calculated them using appropriate formulas 2 . We excluded these studies with imputed standard deviations for this subgroup analysis.
## Placebo
To examine the possible influence of placebo controls we performed the following two sensitivity analyses ("head-to-head studies" and "only active arms"):
## A) head to head studies
We excluded placebo-controlled trials from the data set. Studies with placebo arms were excluded altogether even if they had two or more active arms.
## B) only active arms
We excluded the placebo arms, but kept the remaining active arms of the studies. Two arm placebo controlled studies were excluded.
# Studies using intent-to-treat-analysis
We examined only studies that used intent-to-treat analysis and excluded studies that used completer data. In case of doubt or if the type of data analysis was not clearly stated, we excluded the study in this subgroup analysis.
## Only fair doses
We calculated the mean olanzapine equivalence doses for each drug arm according to the "International Consensus Study on Antipsychotic dose" . For this analysis we excluded studies in which the mean dose of one drug was 50% less than the mean dose of the other in olanzapine equivalents. For example, if a study compared haloperidol 2mg/day (olanzapine equivalence dose =4mg) with olanzapine 10mg/day we excluded the study.
## Older studies
We assumed that older studies are different concerning patient charateristics and study quality. To check if this influenced the results considerably we excluded studies published before 1990 folowwing a reviewer request.
## Failed trials
Three (or more) arm trials (new antipsychotic versus versus placebo and versus an already licensed antipsychotic as an active comparator) in which neither the test drug nor the active comparator was significantly superior to placebo are called "failed trials". Such trials were excluded in a sensitivity analysis following a reviewer request.
Risk of bias item Low risk Unclear risk High risk "identical capsules" would be sufficient)
- No blinding of outcome assessment, but the review authors judge that the outcome measurement is not likely to be influenced by lack of blinding (as we focus on efficacy, subjectively measured by rating scales, this is usually not the case, but if our primary outcome were death, such an objective, undisputable outcome would be less affected by blinding);
- Blinding of outcome assessment ensured, and unlikely that the blinding could have been broken (in our reviews we do not automatically assume that side-effects could break the blind, unless the study authors present evidence for this).
blinding was made (e.g. not even the term "identical capsules" is used)
- Insufficient information to permit judgement of 'Low risk' or 'High risk';
- The study did not address this outcome.
automatically assume that side-effects could break the blind, unless the study authors present evidence for this).
## Incomplete outcome data
- No missing outcome data;
- dropouts with reasons reported
- "reasonable" intention to treat analysis (e.g. at least one dose and one post baseline assessment)
- abstract or poster only
- Insufficient reporting of attrition/exclusions to permit judgement of 'Low risk' or 'High risk' (e.g.
number randomized not stated, dropouts or reasons for dropouts not indicated);
- only completer analysis (unless there were no dropouts) [bib_ref] Efficacy and safety of asenapine in asian patients with an acute exacerbation..., Kinoshita [/bib_ref]
## Risk of bias item
Low risk Unclear risk High risk Selective reporting - The study protocol is available and all of the study's pre-specified (primary and secondary) outcomes that are of interest in the review have been reported in the prespecified way; as we usually do not have the protocol, we will compare what is described in the method section with what is reported in the results.
- The study protocol is not available but it is clear that the published reports include all expected outcomes, including those that were pre-specified - only abstract or poster
- Insufficient information to permit judgement of 'Low risk' or 'High risk'. It is likely that the majority of studies will fall into this category.
- Not all of the study's pre-specified primary outcomes have been reported;
- One or more primary outcomes is reported using measurements, analysis methods or subsets of the data (e.g. subscales) that were not pre-specified;
- One or more reported primary outcomes were not pre-specified (unless clear justification for their reporting is provided, such as an unexpected adverse effect);
- One or more outcomes of interest in the review are reported incompletely so that they cannot be entered in a meta-analysis;
- The study report fails to include results for a key outcome that would be expected to have been reported for such a study (IN OUR CASE USUALLY overall efficacy (continous or dichotomous responder) AND dropouts + dropout reasons inefficacy and adverse events)
## Other bias
The study appears to be free of other sources of bias.
- Insufficient information to assess whether an important risk of bias exists (one reason could be that the study is only available as an abstract or poster); or
- Insufficient rationale or evidence that an identified problem will introduce bias.
- Statistically significant baseline imbalance in an important outcome (in our case primarily overall efficacy)
- Premature termination of the study - Study has been claimed to be fraudulent
[fig] Figure 1: Appendix 14 Summary of network geometry S4 Provide a brief overview of characteristics of the treatment network. This may include commentary on the abundance of trials and randomized patients for the different interventions and pairwise comparisons in the network, gaps of evidence in the treatment network, and potential biases reflected by the network structure. [/fig]
[fig] Figure 2: 3 and 4 [/fig]
[fig] 3: Anticipated or actual start date Give the date when the systematic review commenced, or is expected to commence. date by which the review is expected to be completed. review at time of this submission Indicate the stage of progress of the review by ticking the relevant boxes. Reviews that have progressed beyond the point of completing data extraction at the time of initial registration are not eligible for inclusion in PROSPERO. This field should be updated when any amendments are made to a published record. other relevant information about the stage of the review here. [/fig]
[fig] 1: Response to treatment (study defined) (s) 2. Change in positive symptoms of schizophrenia (s) 3. Change in negative symptoms of schizophrenia (s) 4. Change in depressive symptoms (s) 5. Dropout due to any reason (all-cause discontinuation) (s) 6. Dropout due to inefficacy of treatment (s) 7. Adverse events a) Use of antiparkinson medication (s) b) Akathisia (s) c) Weight gain d) Prolactin levels e) At least one sexual side-effect (s) f) Sedation/somnolence (s) g) Cardiac side-effects, in particular QTc prolongation h) At least one anticholinergic side-effect (s). i) 8. Patient subjective well-being, quality of life (s) +9. Overall functioning (s) + Subjective outcomes are marked with an (s). [/fig]
[fig] Figure 5 2: Boxplot for distribution of publication year across comparisons [/fig]
[fig] Figure 5: 4: Boxplot for distribution of baseline severity across comparisons [/fig]
[fig] Figure 9 2: Forest plot overall change in symptoms adjusted for publication year 2016 Treatments are ranked according to standardised mean difference (SMD) compared to placebo. Treatments crossing the y-axis are not significantly different from placebo. [/fig]
[fig] Figure 9 4: Forest plot overall change in symptoms adjusted for mean age 40 Treatments are ranked according to standardised mean difference (SMD) compared to placebo. Treatments crossing the y-axis are not significantly different from placebo.When the model was adjusted for to a baseline severit of 94 on the Positive and Negative Syndrome Scale, no significant changes in the hierarchy were observed. [/fig]
[fig] Figure 9 5: Forest plot overall change in symptoms adjusted for baseline severity Treatments are ranked according to standardised mean difference (SMD) compared to placebo. Treatments crossing the y-axis are not significantly different from placebo. [/fig]
[fig] Figure 9 7: Forest plot overall change in symptoms adjusted for sponsorship Treatments are ranked according to standardised mean difference (SMD) compared to placebo. Treatments crossing the y-axis are not significantly different from placebo. [/fig]
|
Spontaneous Gastric Perforation in a Case of Collagenous Gastritis
Collagenous gastritis is an extremely rare disease, both in children and adults. Symptoms vary depending on the extent of collagenous changes in the bowel. In most of the children, iron deficiency anemia and abdominal pain are the presenting symptoms. We present a 15-year-old boy with acute abdomen due to gastric perforation the cause of which was collagenous gastritis.
A 15-year-old, previously healthy boy without a history of smoking or drug abuse, admitted to the hospital with acute epigastric pain without dyspnea or heartburn. On physical examination the boy was pale with tachycardia (110/min), but normal blood pressure (110/70 mmHg) and normal capillary oxygen saturation in room air (SPO2 99%). Tenderness in the upper abdomen was present. Laboratory investigations showed mild iron deficiency anemia (haemoglobin concentration 7.1 mmol/l, mean cellular volume 65 fl, serum iron 4 mmol/l, serum ferritin 3 g/l) with normal infection parameters. X-ray revealed free air under the right diaphragm. Abdominal ultrasound and CT scan were performed in search of underlying abnormalities, but none could be detected. At laparotomy, gastric perforation was found, which was repaired after refreshing the margins and tissue was sent for histopathology. Histopathology of the fragments showed ulceration without demonstrable Helicobacter pylori; recognizable mucosa was lacking. Subsequent gastro-duodenoscopy showed macroscopically a nodular aspect of the corpus with remarkably coarse gastric folds and pseudo-polyps [fig_ref] Figure 1: Coarse folds and pseudopolyps in the gastric lumen [/fig_ref].
Histological examination of gastric biopsies showed active chronic inflammation with lympho-plasmacellular infiltration and a thickened sub-epithelial collagenous band, up to 23 μm, consistent with the di-agnosis collagenous gastritis; the duodenal mucosa was normal. The patient was treated with iron supplementation because of persistent anemia. Repeated gastroscopies over the years yielded similar results with collagen bands of up to 22-29 m [fig_ref] Figure 2: Hypertrophic mucosa with irregularly thickened sub-epithelial membrane [/fig_ref] ; colonoscopy revealed normal colonic mucosa, both macroscopically and on histological examination. At the age of 18 year, the patient is asymptomatic, but needed frequent courses of iron replacement therapy.
# Discussion
The first report on collagenous gastritis was published in 1989. [bib_ref] Collagenous gastritis: reports and systematic review, Brain [/bib_ref] [bib_ref] Collagenous gastritis: histopathologic features and association with other gastrointestinal diseases, Leung [/bib_ref] Over 300 patients with collagenous gastritis (40 pediatric) have been reported, the youngest being 9 months old. [bib_ref] Proximal collagenous gastroenteritides: clinical management. A systematic review, Nielsen [/bib_ref] It is twice as common in females. [bib_ref] Collagenous gastritis, sprue and colitis in a 9-month-old infant, Billiémaz [/bib_ref] [bib_ref] Collagenous gastritis, a new spectrum of disease in pediatric patients: two case..., Suskind [/bib_ref] [bib_ref] Collagenous gastritis: a case report, morphologic evaluation, and review, Vesoulis [/bib_ref]. Collagenous gastritis in adults is often accompanied by collagenous sprue and (or) collagenous colitis; in children and adolescents it is usually an isolated phenomenon. [bib_ref] Collagenous gastritis, a new spectrum of disease in pediatric patients: two case..., Suskind [/bib_ref] The etiology has not been completely known yet; an autoimmune background is suggested. [bib_ref] Proximal collagenous gastroenteritides: clinical management. A systematic review, Nielsen [/bib_ref] The presentation is variable in children including upper abdominal pain, anemia, vomiting, upper gastrointestinal bleeding. [bib_ref] Proximal collagenous gastroenteritides: clinical management. A systematic review, Nielsen [/bib_ref] [bib_ref] Collagenous gastritis, a new spectrum of disease in pediatric patients: two case..., Suskind [/bib_ref] [bib_ref] Collagenous gastritis: a case report, morphologic evaluation, and review, Vesoulis [/bib_ref]. Perforation of the stomach as the presenting symptom of collagenous gastritis is unusual. Leung et al presented a 20-year-old man with perforated gastric ulcers preceding the occurrence of collagenous gastritis by four years.Collagenous gastritis was also reported in literature, in a 15-year-old boy with a bleeding gastric ulcer, but without perforation.The diagnosis of collagenous gastritis is based on characteristic endoscopic and histological findings. At endoscopy, especially in children and adolescents, the gastric mucosa usually looks macroscopically inflamed, with thickened folds and a nodular or pseudopolypoid as was also seen in our patient. [bib_ref] Collagenous gastritis, a new spectrum of disease in pediatric patients: two case..., Suskind [/bib_ref] Histologically the sub-epithelial layer is irregularly thickened, not extending around the foveolae and glandular tubes. It consists of collagen; typically, entrapped capillaries are present within this layer and the lamina propria contains an increase in intraepithelial inflammatory cells. [bib_ref] Proximal collagenous gastroenteritides: clinical management. A systematic review, Nielsen [/bib_ref] The collagen layer is normally is 0.30-2.81 μm (mean, 1.33 μm) and does not seem to change significantly in other forms of chronic gastritis but it is grossly increased to between 15 and 94 μm with a diffuse or patchy distribution in collagenous gastritis. [bib_ref] Collagenous gastritis: histopathologic features and association with other gastrointestinal diseases, Leung [/bib_ref] [bib_ref] Collagenous gastritis, sprue and colitis in a 9-month-old infant, Billiémaz [/bib_ref] [bib_ref] Collagenous colitis in children: clinicopathologic, microbiologic, and immunologic features, Camarero [/bib_ref] This phenomenon is pathognomonic; it is found in no other gastrointestinal condition.
There is very limited data on the treatment of collagenous gastritis in children and adults [bib_ref] Collagenous gastritis: reports and systematic review, Brain [/bib_ref] [bib_ref] Proximal collagenous gastroenteritides: clinical management. A systematic review, Nielsen [/bib_ref] [bib_ref] Collagenous gastritis, a new spectrum of disease in pediatric patients: two case..., Suskind [/bib_ref]. Both acid reduction and corticosteroid treatment seem to be mostly ineffective, although many children become symptom-free independent of the type of treatment. Obviously, iron deficiency anemia can successfully be treated with iron supplementation, a treatment that has to be continued for prolonged period. Collagenous gastritis is easily overlooked on account of non-specific features and can lead to gastric perforation. Histological specimens should always be taken during surgical repair, since the diagnosis of collagenous gastritis can only be established on histological findings.
[fig] Figure 1: Coarse folds and pseudopolyps in the gastric lumen. [/fig]
[fig] Figure 2: Hypertrophic mucosa with irregularly thickened sub-epithelial membrane. (200x) [/fig]
[fig] Figure 3: Azan staining of gastric mucosa. Collagen is presented in blue. (400x) [/fig]
|
Near-Infrared Plasmonic Assemblies of Gold Nanoparticles with Multimodal Function for Targeted Cancer Theragnosis
Here we report a novel assembly structure of near-infrared plasmonic gold nanoparticles (AuNPs), possessing both photoacoustic (PA) and photothermal (PT) properties. The template for the plasmonic AuNP assembly is a bioconjugate between short double-strand DNA (sh-dsDNA) and human methyl binding domain protein 1 (MBD1). MBD1 binds to methylated cytosine-guanine dinucleotides (mCGs) within the sequence of sh-dsDNA. Hexahistidine peptides on the engineered MBD1 function as a nucleation site for AuNP synthesis, allowing the construction of hybrid conjugates, sh-dsDNA-MBD1-AuNPs (named DMAs). By varying the length of sh-dsDNA backbone and the spacer between two adjacent mCGs, we synthesized three different DMAs (DMA_5mCG, DMA_9mCG, and DMA_21mCG), among which DMA_21mCG exhibited a comparable photothermal and surprisingly a higher photoacoustic signals, compared to a plasmonic gold nanorod. Further, epidermal growth factor receptor I (EGFR)-binding peptides are genetically attached to the MBD1 of DMA_21mCG, enabling its efficient endocytosis into EGFR-overexpressing cancer cells. Notably, the denaturation of MBD1 disassembled the DMA and accordingly released the individual small AuNPs (<5 nm) that can be easily cleared from the body through renal excretion without causing accumulation/toxicity problems. This DMA-based novel approach offers a promising platform for targeted cancer theragnosis based on simultaneous PA imaging and PT therapy.AuNPs have attracted much attention because synthetic methods are well developed, they exhibit unique and tunable optical (localized surface plasmon resonance (LSPR)) properties, and they are generally considered biocompatible, making them an attractive material for biomedical applications such as cancer therapy and imaging 1-4 . In particular, AuNPs can absorb light and transform it into heat, resulting in both a localized temperature increase and an acoustic wave generation. Thus AuNPs are considered as a promising dual functional agent for photothermal (PT) therapy and photoacoustic (PA) imaging of cancer 5-7 and thus provide highly useful function for cancer theragnosis that simultaneously enables cancer diagnosis and therapy and monitoring of the responses to therapy. The cancer theragnosis can significantly reduce risks and cost in cancer treatment and improve cancer management, and is therefore highly attractive to both clinicians and patients 8-10 .One problem to hamper the clinical application of AuNPs, however, is the accumulation of AuNPs inside the body, which causes nanotoxicity 11,12 . For example, the AuNPs accumulated in liver cause severe damages such as acute inflammation, tissue apoptosis, and abnormal increase of kupffer cells[13][14][15]. To avoid this problem, AuNPs must be sufficiently small, because small-sized AuNPs (<8 nm in diameter) that can pass through glomerular filtration are effectively cleared from the body through renal excretion, and thus result in less accumulation in body tissues[16][17][18]. Additionally, smaller AuNPs (<10 nm) typically show higher photothermal efficiencies, i.e. more efficiently transduce light to heat than larger AuNPs, because smaller AuNPs produce less surface light scattering. Published: xx xx xxxx OPEN www.nature.com/scientificreports/ 2 Scientific REPORTS | 7: 17327 |
According to Mie theory that simulates the photothermal properties of spherical nanoparticles, absorption and scattering properties depend on size of the particles. As particle size increases, spherical particles gradually start to produce more surface light scattering [bib_ref] Size-Dependent Photothermal Conversion Efficiencies of Plasmonically Heated Gold Nanoparticles, Jiang [/bib_ref]. Another important requirement for in vivo theragnostic AuNPs is that they should be able to absorb near-infrared (NIR) light in the wavelength range of 600 to 1,000 nm, which has sufficiently long tissue penetration depth appropriate for in vivo deep tissue imaging and therapy [bib_ref] Near-infrared spectroscopic measurement of myoglobin oxygen saturation in the presence of hemoglobin..., Schenkman [/bib_ref] [bib_ref] Non-invasive in vivo characterization of breast tumors using photon migration spectroscopy, Tromberg [/bib_ref]. Although non-spherical gold nanoparticles such as gold nanorods (AuNRs) or gold nanoshells (AuNSs) are more capable of absorbing the NIR light, it is difficult to synthesize AuNRs or AuNSs that are sufficiently small to be cleared from the body through renal excretion [bib_ref] Small Gold Nanorods with Tunable Absorption for Photothermal Microscopy in Cells, Shibu [/bib_ref] [bib_ref] Scalable routes to gold nanoshells with tunable sizes and response to near-infrared..., Prevo [/bib_ref]. Small spherical AuNPs with a diameter of about 5 nm have a maximum absorption peak at around 520 nm [bib_ref] A New Hydrosol of Gold Clusters .1. Formation and Particle-Size Variation, Duff [/bib_ref] and are therefore not suitable for in vivo deep tissue application. A possible solution is to develop a rod-shaped assembly of small spherical AuNPs that can exhibit efficient plasmon absorption at NIR wavelengths due to plasmon coupling between individual AuNPs [bib_ref] Biosensing with plasmonic nanosensors, Anker [/bib_ref] [bib_ref] Threading plasmonic nanoparticle strings with light, Herrmann [/bib_ref].
In addition, targeted delivery of AuNPs to cancer cells has been another major issue for PT therapy. Higher targeting efficiency reduces the amount of therapeutically ineffective AuNPs that are delivered to the distant region from cancer and hence enhances the performance of PT cancer therapy [bib_ref] Enhanced In Vivo Tumor Detection by Active Tumor Cell Targeting Using Multiple..., Kwon [/bib_ref]. To make AuNPs have a cancer-targeting capability, the surface of AuNPs has been chemically conjugated with cancer cell-binding biomolecules (bio-ligands such as peptides or antibodies having affinity for cancer cells) [bib_ref] Cancer cell imaging and photothermal therapy in the near-infrared region by using..., Huang [/bib_ref] [bib_ref] Surface plasmon resonance scattering and absorption of anti-EGFR antibody conjugated gold nanoparticles..., El-Sayed [/bib_ref]. However, the chemical conjugation is generally performed in a random fashion, and therefore the critical factors affecting cancer-targeting function of AuNPs (i.e. surface density, orientation, and activity of bio-ligands attached on AuNPs) are not well-controllable, which still remains problematic. In this study, these problems was resolved by genetic conjugation that enables highly uniform conjugation of correctly oriented bio-ligands without conformational alteration to protein template used for ordered assembly of AuNPs [bib_ref] Genetic Assembly of Double-Layered Fluorescent Protein Nanoparticles for Cancer Targeting and Imaging, Kim [/bib_ref].
Reportedly, self-assemblies of biomacromolecules such as DNA and proteins have provided structural guides for producing ordered metal nanocrystal/nanowire arrays [bib_ref] Lattice engineering through nanoparticle-DNA frameworks, Tian [/bib_ref] [bib_ref] DNA-based self-assembly of chiral plasmonic nanostructures with tailored optical response, Kuzyk [/bib_ref] : complementary DNA hybridization or avidin-biotin binding mechanism has mediated the construction of a structural template that AuNP superstructures are formed on. This method is typically based on thiol-linked conjugation between the template and small AuNPs (1 to 5 nm) that are formerly synthesized through the reduction of gold salts in the presence of thiol capping ligands, thus leaving the capping ligands on the AuNP surface, which may interfere with further surface modification and/or functionalization of the AuNPs [bib_ref] Size-controlled synthesis of monodispersed gold nanoparticles stabilized by polyelectrolyte-functionalized ionic liquid, Shan [/bib_ref] [bib_ref] Seeding growth for size control of 5-40 nm diameter gold nanoparticles, Jana [/bib_ref]. In this study, we demonstrate a new method of AuNP assembly using a DNA-protein conjugate as a template for on-site synthesis of AuNPs. The DNA-protein conjugate template is prepared by the specific binding between human MBD1 and mCGs on sh-dsDNA, followed by AuNP synthesis on MBD1. The polyhistidine peptide on MBD1 serves as a nucleation site for AuNP synthesis. This procedure does not require the use of thiol capping ligands, and the size of AuNPs and the aspect ratio of their assembled structure are easily controllable. Notably, the sh-dsDNA-MBD1-AuNP assemblies show light absorption at NIR wavelengths, efficient photothermal and photoacoustic properties, and excellent cancer cell targeting capability. Furthermore, they are reversibly disassembled to small AuNPs that can be effectively cleared through renal excretion. The results of this study demonstrate that DMAs hold a promising potential as a platform agent for biocompatible and targeted cancer theragnosis.
# Result
Synthesis and assembly of small AuNPs using DNA-protein conjugate as a template. As shown in [fig_ref] Figure 1: Schematic illustration of [/fig_ref] , an MBD1 variant that is genetically engineered to have N-and C-terminal polyhistidine tags (H 6 ) and C-terminal affibodies binds to the mCG sites of a sh-dsDNA backbone. Three different sh-dsDNAs having different sequence and length (named 5 mCG, 9 mCG, and 21 mCG) were used to examine the effects of mCG location and distance between mCG sites on the AuNP formation [fig_ref] Figure 1: Schematic illustration of [/fig_ref]. The 5-and 9 mCG represent a 60-bp dsDNA backbone containing 5 mCGs with 10-bp spacer and 9 mCGs with 5-bp spacer, respectively. The 21 mCG represent a 139-bp dsDNA backbone containing 21 mCGs with 5-bp spacer. The polyhistidine tags on MBD1 were used as a site for the chemisorption of gold ions (Au + ) and for the growth of AuNP [bib_ref] Au nanowire fabrication from sequenced histidine-rich peptide, Djalali [/bib_ref]. That is, Au + ions chemisorbed on H 6 were metalized by addition of a reducing agent (NaBH 4 ), leading to the formation of sh-dsDNA-MBD1-AuNPs (DMAs), as confirmed by energy dispersive X-ray (EDX) spectroscopy [fig_ref] Figure 1: Schematic illustration of [/fig_ref].
From the analysis of transmission electron microscopy (TEM) [fig_ref] Figure 2: Morphological and optical characteristics of DMAs [/fig_ref] , DMA synthesized using 5 mCG as a template (DMA_5mCG) consists of AuNPs with a diameter of ~4 nm, showing an aspect ratio of ~6.5, whereas 9-and 21 mCG-derived DMAs (DMA_9mCG and DMA_21mCG, respectively) show different aspect ratios , ~28.0 and ~33.4, respectively but consist of the AuNPs with a similar size (~2 nm). It is speculated that the spacer length between mCGs on the DNA template significantly influences the size of AuNPs although the same amount of gold salt was used, which is because an AuNP growth is subject to be confined by adjacent AuNPs. [fig_ref] Figure 2: Morphological and optical characteristics of DMAs [/fig_ref] shows that compared to standard AuNPs (5 nm), light absorption spectra of DMAs was significantly red-shifted, due probably to the confinement of multiple AuNPs within small DMA dimensions and plasmon coupling between adjacent AuNPs. It is notable that DMA_21mCG exhibited broader overall absorption spectra and a stronger NIR absorption at around 800 nm than the other two DMAs, which is because the magnitude of plasmon coupling increases as AuNP number per assembly increases.
When one of DMAs (i.e. DMA_9mCG) was treated with a protein denaturant (6 M guanidine hydrochloride, Gdn-HCl), it was completely disassembled to individual small AuNPs as shown in TEM images of [fig_ref] Figure 2: Morphological and optical characteristics of DMAs [/fig_ref]. This was also confirmed by the disappearance of the broad absorbance at NIR wavelengths [fig_ref] Figure 2: Morphological and optical characteristics of DMAs [/fig_ref] , while the same Gdn-HCl treatment did not cause any change in the absorbance spectra of standard AuNPs (5 nm) (data not shown). The denatured MBD1 is subject to losing specific binding affinity for mCGs on sh-dsDNA backbone, leading to the disassembly of the DNA-MBD1 conjugate, and in turn releasing the individual AuNPs. Additionally, we confirmed that DMA_21mCG was also well disassembled when placed in a solution at 37 °C (i.e. in vivo physiological temperature) for a long period of time (6 h and overnight) [fig_ref] Figure 2: Morphological and optical characteristics of DMAs [/fig_ref]. The effect of temperature on the disassembly of DMA_21mCG was further estimated by increasing the solution temperature Photothermal and photoacoustic properties of DMAs. The potential utility of DMAs as a photothermal agent was evaluated by measuring the temperature elevation of aqueous DMA suspensions (0.02 mg Au per ml solution) under NIR laser irradiation (808 nm, 2 W/cm 2 ), as presented in [fig_ref] Figure 3: Photothermal activities of DMAs [/fig_ref]. After 5 min of NIR irradiation, the temperature of the DMA_5mCG, DMA_9mCG and DMA_21mCG solutions increased to 39.0, 44.2 and 46.5 °C, respectively. When tested under the same condition, the temperature of solutions containing commercial AuNPs (5 nm) and AuNRs (10 nm × 41 nm) increased up to 31.0 and 48.2 °C, respectively, and the temperature increase of pure water was negligible. In photothermal therapy of cancer, the tumor tissue is typically heated to 41 to 47 °C 5 , and therefore DMA_9mCG and DMA_21mCG show a potential as a PT agent at the low concentration used (0.02 mg Au/ml). Also as shown in [fig_ref] Figure 3: Photothermal activities of DMAs [/fig_ref] , upon cessation of NIR irradiation the DMA solutions were cooled down to room temperature within 5 min. By fitting these time-course temperature profiles to a theoretical model (Eqs 1 to 4 in Experimental Methods) 36,37 , we estimated the PT conversion efficiencies (η pc ) of the DMAs. As summarized in [fig_ref] Figure 3: Photothermal activities of DMAs [/fig_ref] , the η pc values of DMA_21mCG, DMA_9mCG, and DMA_5mCG were 14.8%, 12.3%, and 10.4%, respectively, while η pc of the commercial AuNR (10 nm × 41 nm) was 19.0%. This result is in agreement with the previous report that smaller AuNPs have higher photothermal conversion efficiencies than larger AuNPs [bib_ref] Clearance properties of nano-sized particles and molecules as imaging agents: considerations and..., Longmire [/bib_ref]. Among DMAs, DMA_21mCG shows the highest heat generation efficiency due probably to its highest aspect ratio and correspondingly strongest plasmon coupling of AuNPs.
The PA properties of DMAs at various concentrations were estimated under NIR irradiation (808 nm wavelength). As presented in [fig_ref] Figure 4: Photoacoustic activities of DMAs [/fig_ref] , the PA signal intensity monotonically increased with Au concentration, and DMA_21mCG showed the strongest PA signal, which is well explainable by PT efficiencies of DMAs because the PA signal is generated due to thermoelastic expansion (vibration) of AuNPs caused by locally heated media. It is worth noting that DMA_21mCG and DMA_9mCG generated stronger PA signal than AuNR (10 nm × 41 nm), even though the AuNR showed a higher PT efficiency than DMAs [fig_ref] Figure 3: Photothermal activities of DMAs [/fig_ref]. It is speculated that the close proximity between AuNPs on DMAs produced overlapping thermal fields and thus increased the thermal flux, resulting in an amplification of the PA signal [bib_ref] pH-Induced aggregated melanin nanoparticles for photoacoustic signal amplification, Ju [/bib_ref]. [fig_ref] Figure 3: Photothermal activities of DMAs [/fig_ref] indicate that DMA_21mCG has a strong potential as a cancer theragnostic agent based on PA imaging and PT therapy. [fig_ref] Figure 5: Cellular uptake of Cy5 [/fig_ref] ,b, the DMAs containing affibody (peptide ligand with high affinity for EGFR) (aff+) were effectively internalized by EGFR-expressing cancer (A431) cells, while the endocytic uptake of affibody-free DMAs by A431 cells was negligible. Notably, DMA_21mCG (aff+) that contains the highest number of mCG and MBD1 (aff+) exhibited the highest-level endocytic uptake by A431 cells among the three DMAs, and its uptake was almost 8-times higher than that of MBD1 (aff+) only [fig_ref] Figure 5: Cellular uptake of Cy5 [/fig_ref] finding that the multi-valent binding ligands significantly enhance the efficiency of binding to target [bib_ref] Green fluorescent protein nanopolygons as monodisperse supramolecular assemblies of functional proteins with..., Kim [/bib_ref]. Therefore, DMA-21mCG appears to be the best candidate for cancer theragnosis. Based on the previous literature 27, [bib_ref] Genetic Assembly of Double-Layered Fluorescent Protein Nanoparticles for Cancer Targeting and Imaging, Kim [/bib_ref] , the intracellular localization of DMA_21mCG can be determined from fluorescent cell images: punctate fluorescence patterns typically imply the confinement of fluorescent molecules within endosomal compartments such as lysosome, whereas smeared/diffuse fluorescent signals indicate the signaling molecules in cytosol. [fig_ref] Figure 5: Cellular uptake of Cy5 [/fig_ref] clearly shows the punctuate signals in tumor cells, suggesting that the intracellular DMAs are located in endosomal compartments.
## Endocytic uptake of dmas by in vitro cancer cells. as shown in
Cancer cell ablation using photothermal activity of DMAs. The cytotoxicity of DMA_21mCG in A431 cells was evaluated by the Cell Counting Kit-8 (CCK-8) method. As shown in [fig_ref] Figure 6: Cytotoxicity of and photothermal ablation of EGFR-overexpressing A431 cells by DMA_21mCG [/fig_ref] , when treated with DMA_21mCG (aff+) without NIR irradiation for 24 h, more than 95% of cells were viable at the entire range of concentration (up to 125 nM), indicating no cytotoxicity. [fig_ref] Figure 6: Cytotoxicity of and photothermal ablation of EGFR-overexpressing A431 cells by DMA_21mCG [/fig_ref] shows that nearly 75% of the A431 cells was ablated at 2 h after the 20-min irradiation of NIR laser (808 nm, 2 W/cm 2 ) to the cell cultures treated with DMA_21mCG (aff+) (125 nM), and the cancer cell death increased as the duration of NIR irradiation increased, indicating the excellent performance of PT-based cancer cell treatment by DMA_21mCG (aff+). As shown in [fig_ref] Figure 6: Cytotoxicity of and photothermal ablation of EGFR-overexpressing A431 cells by DMA_21mCG [/fig_ref] to e, the fluorescence images taken after staining with Calcein AM and propidium iodide (PI) confirmed that a significant level of cancer cell death occurred after the NIR treatment in the presence of intracellular DMA-21mCG (aff+). These results suggest that DMA_21mCG (aff+) is a promising functional material for targeted cancer therapy.
# Discussion
Cancer theragnosis that simultaneously enables cancer diagnosis and therapy is a next mainstay of clinical cancer treatment. It has crucial advantages over traditional cancer diagnoses and therapies and is highly attractive to both clinicians and patients because it can significantly reduce risks and cost in cancer treatment and improve cancer management. The multi-functional agents with cancer targeting, imaging, and therapeutic capabilities are essential to the cancer theragnosis, and here we report a new type of cancer theragnostic agent based on photothermal and photoacoustic activies of AuNPs. The rod-shaped assemblies of small AuNPs (<5 nm) that exhibit the light absorption at a broad range of NIR wavelengths were developed using MBD1-bound sh-dsDNA as a structural template, where MBD1 protein is biologically conjugated with sh-dsDNA via specific binding between MBD1 and mCG within the DNA sequence. The MBD1 protein was genetically modified by inserting two different foreign peptides to its N-and C-terminus: 1) histidine-rich peptides (H 6 ) to be used as a gold ion (Au 3+ ) chemisorption and gold nucleation site for on-site synthesis of the small AuNPs and 2) cancer cell receptor-binding peptides (EGFR-specific affibodies) to enhance cancer cell-targeting efficiency. Among the resulting DNA-MBD1-AuNP assemblies (DMAs), DMA_21mCG with the highest aspect ratio of rod-shaped assembly of AuNPs showed the highest-level PT and PA activities and the best performance in targeting EGFR-expressing cancer cells, suggesting that DMA_21mCG holds a promising potential as a cancer theragnostic agent. Further, it is worth noting that under the particular conditions for denaturing MBD1, the DMAs are disassembled and release the small individual AuNPs that can easily pass through glomerular filtration in kidney. This suggests that inside the body where proteins are spontaneously denatured, the DMA structure would be eventually atomized into small individual AuNPs that can be cleared from the body through renal excretion, which hence would resolve in vivo accumulation-associated toxicity problems of AuNPs that have been hampering in vivo clinical application of AuNPs.
In summary, we first report the synthesis of rod-shaped assemblies of small AuNPs with multimodality function (PT and PA activities at a broad range of NIR wavelengths), which was performed using genetically modified MBD1-bound sh-dsDNA as a synthetic scaffold. The three different sequences of sh-dsDNA with different length (60, 139 bp), different number of mCG [bib_ref] Nanoparticle-based theranostic agents, Xie [/bib_ref] [bib_ref] Non-invasive in vivo characterization of breast tumors using photon migration spectroscopy, Tromberg [/bib_ref] , and different spacer length between adjacent mCGs (5, 10 bp) were used as a DNA template, where each mCG specifically binds to a single MBD1 protein. Gold ions (Au 3+ ) are chemisorbed to histidine-rich peptide (H 6 ) of engineered MBD1, immediately followed by reduction and nucleation to small AuNPs. The size of individual AuNPs and the aspect ratio of overall rod shape of assemblies were controlled through varying the length of both the spacer between adjacent mCGs and the entire sh-dsDNA backbone, resulting in three different assembly structures of AuNPs, i.e. DMA_5mCG, DMA_9mCG, and DMA_21mCG. Among the three DMAs, DMA_21mCG showed the highest PT and PA activities that are comparable to or much better than a commercial AuNR. Furthermore, DMA_21mCG comprising cancer cell receptor (EGFR)-binding affibody peptides (DMA_21mCG (aff+)) was successfully endocytosed by EGFR-expressing cancer cells (A431) without causing any cytotoxicity. The 20-min irradiation of NIR laser to A431 cells treated with DMA_21mCG (aff+) ablated about 75% of the initial cancer cells within 4 h. Notably, DMA_21mCG (aff+) was disassembled by denaturing MBD1, resulting in the release of the small individual AuNPs that can easily pass through glomerular filtration in kidney. Consequently, this novel approach to synthesize rod-shaped DMA assemblies enables the successful preparation of multimodality agent with cancer targeting, PA-based cancer imaging, and PT-based cancer therapeutic functions. DMAs could be cleared from the body through renal excretion without causing in vivo accumulation-associated toxicity problems. All these features make this novel DMA assemblies an attractive candidate as a cancer theragnostic agent.
# Methods
Preparation of genetically engineered MBD1 and sh-dsDNA backbones. Through polymerase chain reactions (PCRs) using appropriate primers, two gene clones were prepared from previously cloned expression vectors for the synthesis of two MBD1 protein derivatives containing two hexahistidine tags: NH 2 -NdeI-H 6 -MBD1 (MAEDWLDSPALGPGWKRREVFRKSGATAGRSDTYYQSPTGDRIRSKVELTRYLGPAGDLTLFDFK QGIL)-H 6 -XhoI-COOH and NH 2 -NdeI-H 6 -MBD1-XhoI-H 6 -L(G 3 SG 3 TG 3 S G 3 )-(affibody) 2 -ClaI-COOH. These genes were sequentially ligated into the pT7-7 expression vector to construct the following vectors encoding the two engineered MBD1 proteins above: pT7-H6-MBD1-H6 and pT7-H6-MBD1-H6-Affi [fig_ref] Figure 4: Photoacoustic activities of DMAs [/fig_ref]. E. coli Rosetta TM (DE3) (Novagen, Darmstadt, Germany) was transformed with each expression vector above, and ampicillin-resistant transformants were selected. The recombinant E. coli cells were grown in Luria-Bertani media at 37 °C, and the recombinant gene expression was induced by adding Isopropyl-β-D-thiogalactoside (IPTG, 1 mM) to the growing culture when the culture turbidity (optical density at 600 nm, OD 600 ) reached 0.5 to 0.7, followed by further cultivation for 6 h at 37 °C. The synthesized recombinant MBD1 proteins were purified using a Ni-NTA purification system (Qiagen, Hilden, Germany), as described in detail in our previous publication [bib_ref] One-pot approach for examining the DNA methylation patterns using an engineered methylprobe, Kim [/bib_ref]. The purified MBD1 derivatives were analyzed by sodium dodecyle sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) (18%) [fig_ref] Figure 4: Photoacoustic activities of DMAs [/fig_ref].
Three different sh-dsDNA backbones were prepared by a commercial supplier (Xenotech, Daejeon, South Korea): two DNA backbones containing five CG and 9 CG dinucleotides, respectively, on 60-bp DNA fragment and a DNA backbone containing twenty one CG dinucleotides on 139-bp DNA fragment. The full sequences of the three sh-dsDNA backbones are presented in . The cytosine methylation of CG dinucleotides was performed using DNA methyltransferase (New England Biolabs, Massachusetts, USA), resulting in the synthesis of 5 mCG, 9 mCG, and 21 mCG. The degree of methylation was evaluated through analyzing the digestion pattern by HpaII restriction enzyme as described in previous publication [bib_ref] One-pot approach for examining the DNA methylation patterns using an engineered methylprobe, Kim [/bib_ref] (also see [fig_ref] Figure 5: Cellular uptake of Cy5 [/fig_ref]. In all methylated DNA products the degree of methylation was more than 99%.
## Synthesis and characterizations of dmas.
The sh-dsDNA-MBD1 conjugates were prepared as follows: a solution containing purified MBD1 protein (10 µM) with or without affibodies was directly mixed with a solution containing one of methylated DNA backbones (5 mCG, 9 mCG, or 21 mCG) at a molar ratio of sh-dsDNA to MBD1 of 1:200. The mixture was then incubated for 1 h at room temperature, followed by adding 1 mg of AuClP(CH 3 ) 3 (Sigma-Aldrich, St. Louis, MO, USA) to 200 µl of the mixture above (10 µM sh-dsDNA-MBD1 conjugate, 10 mM Tris-HCl, 1 mM ethylenediaminetetraacetic acid (EDTA), 1 mM dithiothreitol (DTT), pH 7.9). After stirring overnight at 20 °C, the mixture was centrifuged at 10,000 rpm for 15 min. The supernatant was mixed with 5 μl of 1 M NaBH 4 (reducing agent), and the mixture was kept under stirring at 20 °C for 10 min to synthesize DMA. Among various types of reducing agents, NaBH 4 was finally selected through preliminary studies to optimize the conditions for Au 3+ reduction and DMA synthesis including the type and concentration of reducing agent, mixing and reduction time, etc. [bib_ref] Superparamagnetic Gold Nanoparticles Synthesized on Protein Particle Scaffolds for Cancer Theragnosis, Kwon [/bib_ref] [bib_ref] Proteinticle/Gold Core/Shell Nanoparticles for Targeted Cancer Therapy without Nanotoxicity, Kwon [/bib_ref].
The surface charge properties of DMAs (DMA_5mCG, DMA_9mCG, and DMA_21mCG) (10 nM DMA, 10 mM Tris-HCl, 1 mM EDTA, 1 mM DTT, pH 7.9) were characterized using a Nano ZS Zetasizer (Malvern Instruments, Worcestershire, UK) equipped with a 633 nm laser. The synthesized DMAs were examined by Transmission Electron Microscopy (TEM) and Energy Dispersive X-ray (EDX) spectroscopy using a CM-200 electron microscope (Philips, CA, USA) operating at an accelerating voltage of 200 kV. TEM and EDX spectroscopy specimens were prepared by dropping a 10 µL of 1 µM DMA suspension onto a carbon-coated copper grid. After waiting for 1 h (to allow the DMA to deposit on the grid surface), the TEM and EDX grid was washed twice with distilled water. For EDX spectroscopy analysis, the count number was plotted as a function of characteristic energy. UV/Vis absorption spectra for DMA samples were measured using a TECAN microplate reader (Infinite M200 Pro, TECAN, Männedorf, Switzerland). [fig_ref] Figure 6: Cytotoxicity of and photothermal ablation of EGFR-overexpressing A431 cells by DMA_21mCG [/fig_ref].
## Photothermal conversion efficiencies (η
The phothothermal conversion efficiencies (η pc ) of the DMA samples were estimated by comparing the measured cooling temperature profiles with the energy balance for the system [bib_ref] Genetic Assembly of Double-Layered Fluorescent Protein Nanoparticles for Cancer Targeting and Imaging, Kim [/bib_ref] [bib_ref] Lattice engineering through nanoparticle-DNA frameworks, Tian [/bib_ref]. where τ s , m w , and C w is system time constant, mass of the medium (water) (0.3 g), and heat capacity of water (4.2 J/g), respectively. As presented in , the value of τ s can be estimated by fitting measured cooling temperature profiles to the equations
[formula] τ θ = − t l n( ) (3) s θ = − − T T T T (4) surr surr max [/formula]
where t, θ, and T is cooling time, dimentionless driving force temperature, and temperature of the system at time t, respectively.
Photoacoustic signal measurement. The PA properties of DMAs (DMA_21mCG, DMA_9mCG, DMA_5mCG), AuNP and AuNR were characterized as follows: Tygon tubes loaded with all samples to be tested were placed across at the center of a water bath and imaged using the procedure described in a previous publication [bib_ref] Amplified Photoacoustic Performance and Enhanced Photothermal Stability of Reduced Graphene Oxide Coated..., Moon [/bib_ref]. An Nd:YAG laser system (Surelite III-10/Surelite OPO Plus, Continuum, Santa Clara, CA, USA) with an energy density of 6 mJ/cm 2 was used as the excitation source. A linear array ultrasound (US) transducer with a frequency range of 5 to 14 MHz was used as a PA signal detector. To examine the effect of Au concentration on the PA signal intensity, 808 nm wavelength laser pulses were delivered under a constant irradiation power (6 mJ/ cm 2 ). The resulting PA signals were recorded using an US transducer. Photoacoustic signals were reconstructed into PA images by averaging 128 sequential frames, which gave a sufficient signal-to-noise ratio. After selecting a region of interest (ROI) in the reconstructed image, the total PA signal intensity from the ROI was quantified using MATLAB.
In vitro cancer cell culture and fluorescence imaging. The cellular uptake efficiencies of MBD1 protein and DMAs were estimated in the culture of EGFR-overexpressing cancer (A431) cells. The A431 cells were cultivated using RPMI Medium 1640 (Gibco, Life technologies, Massachusetts, USA) containing fetal bovine serum (10%) and penicillin (100 U/ml), seeded in 35-mm coverslip bottom dishes at a density of 1 × 10 4 cells/ dish, and further cultivated for 24 h. After washing with phosphate buffer saline (PBS, pH 7.4), A431 cells were incubated in a serum-free medium containing Cy5.5-DMA (aff-) (50 nM), Cy5.5-DMA (aff+) (50 nM), or Cy5.5-MBD1 only (50 nM) for 2 h. After washed with PBS (pH 7.4), the Cy5.5-DMA-or Cy5.5-MBD1-treated cells were fixed in 4% paraformaldehyde and 0.1% glutaraldehyde for 10 min and then counterstained with 4′,6′-diamidino-2-phenylindole hydrochloride (DAPI). After washed once again with PBS, the cells in PBS were imaged using a Confocal Laser Scanning Microscope (LSM 700, Carl-Zeiss, Jena, Germany). Fluorescence measurement was performed using 630-680 nm excitation and 690-740 nm emission long pass filters. To quantify the fluorescence intesntity per cell, fluorescence images were analyzed using the ImageJ software as described in our previous publication 24 .
## Cytotoxicity of dmas and photothermal ablation of cancer cells by dmas. the cytotoxicity of
DMA_21mCG in A431 cells was evaluated by Cell Counting Kit-8 (CCK-8). A431 cells were seeded in a 96-well microplate at a density of 1 × 10 4 cells per well and cultivated for 24 h at 37°C in Roswell Park Memorial Institute (RPMI) Medium 1640 (Gibco, Life technologies, Massachusetts, USA) containing fetal bovine serum (10%) and penicillin (100 U/ml). Cells in different wells were treated with DMA_21mCG at different DMA concentrations (0, 5, 25, 50, 100 and 125 nM) and further incubated for 24 h. Finally, the DMA_21mCG-treated cells were stained using the CCK-8 kit for 2 h, and the absorbance of each well was measured at 450 nm in a microplate reader. The photothermal effect of DMA_21mCG on the ablation of A431 cells was evaluated using the same CCK-8 method. A431 cells were seeded in a 96-well microplate (1 × 10 4 cells/well) and cultivated for 24 h. Cells in different wells were treated with different concentrations of DMA_21mCG (aff+) (0 and 125 nM) and further incubated for 4 h. Then, the cells were irradiated by an NIR laser beam (808 nm, 2 W/cm 2 ) for 0, 5, or 20 min. The CCK-8 reagent was added in each well, and the cells were incubated for 2 h, immediately followed by measuring the absorbance at 450 nm. The cell viability was determined as the ratio of the number of live cells in the DMA-treated cell culture to the number of live cells in non-DMA-treated control. Data were presented as means ± standard deviations for triplicate experiments. The cell viability was also estimated using Live/Dead Double Staining Kit (Sigma-Aldrich, St. Louis, MO, USA). Briefly, A431 cells were seeded in 35-mm coverslip bottom dishes (1 × 10 5 cells per dish) and cultivated for 24 h. DMA_21mCG (aff+) (125 nM) was added to each coverslip, and the cells were incubated for 4 h. Afterwards, the cells were irradiated by an NIR laser (808 nm, 2 W/cm 2 ) for 0, 5, or 20 min. The staining solution was prepared by adding 1 μl of Calcein acetoxylmethyl (Calcein AM) solution (1.5 mg/ml) and 2 μl of propidium iodide (PI) solution (1 mg/ml) to 2 ml PBS. This freshly prepared staining solution was added to each coverslip, and the cells were incubated for 15 min at room temperature. After the staining, the cells were imaged using a fluorescence microscope (Olympus IX81, Tokyo, Japan). Calcein AM (495 nm excitation, 515 nm emission) stains live cells (green), and PI (535 nm excitation, 617 nm emission) stains dead cells (red).
[fig] Figure 1: Schematic illustration of (a) DMA consisting of sh-dsDNA, MBD1, and AuNPs and (b) three sh-dsDNA backbones containing 5, 9, and 21 methylated cytosine-guanine (mCG) dinucleotides: 5 mCG, 9 mCG, and 21 mCG. [/fig]
[fig] Figure 2: Morphological and optical characteristics of DMAs (DMA_5mCG, DMA_9mCG, and DMA_21mCG). (a) TEM images. (b) Light (UV/VIS/NIR) absorption spectra of three DMAs and AuNP (5 nm). (c) Light absorption spectra and TEM images of DMA_9mCG before and after MBD1 denaturation by Gdn-HCl (6 M for 10 min at room temperature). Scientific REPORTS | 7: 17327 | DOI:10.1038/s41598-017-17714-2 [/fig]
[fig] Figure 3: Photothermal activities of DMAs. (a) Time-course change of temperature of aqueous DMA/AuNP/ AuNR solutions under continuous NIR laser irradiation (808 nm, 2 W/cm 2 , 5 min), followed by no irradiation (5 min). Blank water was used as a control. (b) Photothermal (PT) conversion efficiencies of DMAs and AuNR. [/fig]
[fig] Figure 4: Photoacoustic activities of DMAs. (a) Photoacoustic (PA) images. (b) PA signal intensities of aqueous DMA/AuNP/AuNR solutions at various Au concentrations. [/fig]
[fig] Figure 5: Cellular uptake of Cy5.5-labeled DMAs (aff+ and aff−) and MBD1 (aff+ and aff−) in EGFRoverexpressing A431 cancer cell cultures at 37 °C. (a) Confocal fluorescence microscopy images of A431 cells treated with Cy5.5-labeled MBD1 or DMAs with and without affibodies for 1 h. Nuclei were counterstained with DAPI (blue). (b) Fluorescence intensities from the images of (a). Error bars represent standard deviations (n ≥ 7). *p < 0.05. [/fig]
[fig] Figure 6: Cytotoxicity of and photothermal ablation of EGFR-overexpressing A431 cells by DMA_21mCG (aff+). (a) Cytotoxicity of DMA-21mCG in A431 cells treated at different concentrations for 24 h. (b) CCK-8 viability of A431 cells exposed to NIR irradiation (5 and 20 min) after treated by DMA_21mCG (aff+) (125 nM). Measurements were performed in triplicates. Error bars represent standard deviations. Fluorescence images of DMA_21mCG (125 nM)-treated A431 cells, double stained with Calcein AM and PI after an NIR laser irradiation (c) for 0-, (d) 5-, and (e) 20 min. [/fig]
|
Co-Selection of Resistance to Antibiotics, Biocides and Heavy Metals, and Its Relevance to Foodborne Pathogens
Concerns have been raised
# Introduction
Biocides, in the broadest sense, are substances formulated to be harmful to (or to otherwise control) living organisms [bib_ref] Assessment of the Antibiotic Resistance Effects of Biocides; Scientific Committee on Emerging..., Scenihr [/bib_ref]. Many authorities adopt more restrictive definitions, similar to those for a microbicide, for the purposes of considering non-antibiotic antimicrobial agents, albeit still with diverse chemical characteristics and applications. Useful definitions in this mould have been provided by Tumah [bib_ref] Bacterial biocide resistance, Tumah [/bib_ref] and Sheldon [bib_ref] Antiseptic "resistance": Real or perceived threat?, Sheldon [/bib_ref] , which the authors of the present review regard as reasonable in scope and practicality. With few exceptions (such as iodide salts that show some efficacy in the treatment of certain veterinary fungal and bacterial infectionsbiocides are not sufficiently selective to be used within body tissues, but some may be used as antiseptics on body surfaces. Other common applications include use as disinfectants on equipment and surfaces in many environments including farms and hospitals, as decontaminants on carcass surfaces following slaughter, as sporicidal sterilants for medical equipment and as preservatives in pharmaceuticals, cosmetics and food [bib_ref] Assessment of the Antibiotic Resistance Effects of Biocides; Scientific Committee on Emerging..., Scenihr [/bib_ref]. Heavy metals are arbitrarily defined, but commonly-included elements show varying degrees of toxicity for organisms [bib_ref] Heavy metals" a meaningless term? (IUPAC Technical Report), Duffus [/bib_ref] [bib_ref] Heavy metal driven co-selection of antibiotic resistance in soil and water bodies..., Seiler [/bib_ref] with some, such as silver, copper and zinc, being relatively non-toxic to mammalian tissues and systems. These are used as antimicrobial coatings and wound dressings [bib_ref] Silver in medicine: A brief history BC 335 to present, Barillo [/bib_ref] [bib_ref] Bacterial silver resistance: Molecular biology and uses and misuses of silver compounds, Silver [/bib_ref] [bib_ref] Alternative antimicrobial approach: Nano-antimicrobial materials, Beyth [/bib_ref] or (in the case of copper and zinc) in agriculture as in-feed growth promoters and for enteric disease control, particularly in the pig and poultry sectors.
Biocidal substances have been used, in one form or another, for centuries [bib_ref] Silver in medicine: A brief history BC 335 to present, Barillo [/bib_ref] [bib_ref] A brief history of heat, chemical and radiation preservation and disinfection, Hugo [/bib_ref] and the more recent development of versatile biocides with limited toxicity for animal tissues, such as quaternary ammonium compounds (QAC), biguanides and bisphenols, has led to increased use of such compounds to assist improved hygiene [bib_ref] Biocide tolerance in bacteria, Morente [/bib_ref]. Hygienic practices have been a cornerstone of efforts to control antibiotic use (and associated resistance) in clinical and farm environments. In addition, general hygiene and the in-feed use of copper and zinc have been increased in pig and poultry production to improve disease prevention in the wake of the phasing out of antibiotic growth promoters in some parts of the World, notably Europe. Organic acids and phytochemicals such as essential oils are antimicrobial agents that also have been trialled as in-feed alternatives to antibiotic growth promoters and for pathogen control [bib_ref] Butyric acid-based feed additives help protect broiler chickens from Salmonella Enteritidis infection, Fernandez-Rubio [/bib_ref] [bib_ref] Essential oils and aromatic plants in animal feeding-A European perspective. A review...., Franz [/bib_ref] [bib_ref] Effect of butyric acid on the performance and carcass yield of broiler..., Leeson [/bib_ref] [bib_ref] Essential oil and aromatic plants as feed additives in nonruminant nutrition: A..., Zeng [/bib_ref].
In recent decades observations were made among important pathogenic bacteria genera of reduced susceptibility of isolates to commonly-used biocides, particularly in healthcare settings [bib_ref] A hospital outbreak caused by a chlorhexidine and antibiotic-resistant Proteus mirabilis, Dance [/bib_ref] [bib_ref] Glutaraldehyde-resistant Mycobacterium chelonae from endoscope washer disinfectors, Griffiths [/bib_ref] [bib_ref] Triclosan-resistant Staphylococcus aureus, Sasatsu [/bib_ref]. This was often seen in association with an elevated frequency of antimicrobial resistances [bib_ref] Disinfectant resistance in antibiotic-resistant organisms, Alexander [/bib_ref] [bib_ref] Antiseptic and antibiotic resistance in Gram-negative bacteria causing urinary tract infection, Stickler [/bib_ref]. Concerns have consequently been raised that in certain circumstances bacteria may become unsusceptible to some biocide regimes, or that a reduced susceptibility to biocides may foster an elevated degree or frequency of antimicrobial resistance [bib_ref] Assessment of the Antibiotic Resistance Effects of Biocides; Scientific Committee on Emerging..., Scenihr [/bib_ref] [bib_ref] Biocide abuse and antimicrobial resistance-A cause for concern?, Fraise [/bib_ref] [bib_ref] Do biocides select for antibiotic resistance?, Russell [/bib_ref]. Whilst such concerns were initially focussed upon healthcare environments, where biocides and antibiotics were frequently used and where monitoring was commonly practised, the scope of investigations has broadened to include other areas, such as domestic cleaning and hygiene products where the use of biocides has increased [bib_ref] Significance of biocide usage and antimicrobial resistance in domiciliary environments, Bloomfield [/bib_ref] [bib_ref] Potential impact of increased use of biocides in consumer products on prevalence..., Gilbert [/bib_ref].
A major area under particular scrutiny is the livestock food chain, from farms and hatcheries through slaughter facilities and dairy processing plants to food processing, packaging and retail facilities [bib_ref] Assessment of the Antibiotic Resistance Effects of Biocides; Scientific Committee on Emerging..., Scenihr [/bib_ref]. This has partly been driven by alarm over evidence of multi-resistant pathogens, including Salmonella (pathovar Typhimurium DT104 and latterly monophasic Typhimurium variants) [bib_ref] Leakage of emerging clinically relevant multidrug-resistant Salmonella clones from pig farms, Antunes [/bib_ref] [bib_ref] National survey for Salmonella in pigs, cattle and sheep at slaughter in..., Davies [/bib_ref] , livestock-associated meticillin-resistant Staphylococcus aureus (LA-MRSA) [bib_ref] Emergence of methicillin-resistant Staphylococcus aureus (MRSA) in different animal species, Cuny [/bib_ref] and extended-spectrum β-lactamase E. coli [bib_ref] Distribution, numbers, and diversity of ESBL-producing E. coli in the poultry farm..., Blaak [/bib_ref] , on farms and in the food chain. In this context the possible effects of heavy metals have attracted particular attention, owing to the frequent heavy supplementation of copper and zinc in feeds for some animal species.
The present review examines the evidence for reduced biocide and heavy metal susceptibility among pathogens and bacterial communities, for co-selection with antimicrobials and for mechanisms underlying the observed patterns. This is a broad remit, and coverage is necessarily selective, with an ultimate focus on issues in animal production, zoonotic bacteria and the food chain, plus discussion of the notable differences seen between field and laboratory investigations.
## Agents with the potential for co-selection of antibiotic resistance
## Antibiotics
Antibiotics, and related drugs such as sulphonamides, exhibit high antimicrobial potency and selective toxicity that is sufficient in many cases to allow their use as anti-infective agents within body tissues. This is usually because of a highly specific action on a microbial target [bib_ref] Exploiting current understanding of antibiotic action for discovery of new drugs, Chopra [/bib_ref] , which is commonly inhibitory to growth rather than lethal at normal therapeutic concentrations, and which acts in concert with the host's immune system to resolve microbial colonisation or infection over an extended period [bib_ref] Potential impact of increased use of biocides in consumer products on prevalence..., Gilbert [/bib_ref] [bib_ref] Tests for determining in-use concentrations of antibiotics and disinfectants are based on..., Cerf [/bib_ref].
Reduced susceptibility of an organism to an antibiotic may be innate, due for example to features of the microbe's cell envelope, energy metabolism or the presence of an alternative metabolic pathway. Alternatively, reduced susceptibility may be acquired via single-or multi-step mutation that affects the target site and/or the effective concentration of the antibiotic within the cell, or by the acquisition of genetic material encoding a feature such as an inactivating enzyme or an alternative to the target molecule [bib_ref] Mechanisms of bacterial biocide and antibiotic resistance, Poole [/bib_ref].
Bacterial resistance may be promoted by, among other things, sub-therapeutic antibiotic concentrations in certain tissues or organs, such as the gut, and the tendency for antibiotics to induce non-specific mutagenesis in bacteria [bib_ref] Sublethal antibiotic treatment leads to multidrug resistance via radical-induced mutagenesis, Kohanski [/bib_ref]. A key parameter of an antibiotic-microbe combination is the minimum inhibitory concentration (MIC), indicating a concentration of antibiotic preventing growth of the microbe and likely to prove therapeutically effective if achieved in the target tissue(s).
## Biocides
By contrast with antibiotics, biocides have more varied applications, do not operate with the benefit of a concurrent immune response, and commonly have to contend with microbes in protected or resistant states, such as in biofilms or organic matter, in nutrient-or moisture-limited environments, or following sporulation. Consequently they typically are intended to be lethal, usually after a single application, and generally have multiple structural and biochemical targets achieving a gross disruptive effect on micro-organisms [bib_ref] Biocide tolerance in bacteria, Morente [/bib_ref] [bib_ref] Bacterial target sites for biocide action, Maillard [/bib_ref] [bib_ref] Antiseptics and disinfectants: Activity, action, and resistance, Mcdonnell [/bib_ref]. In consequence of the foregoing considerations, "in use" concentrations of biocides typically are multiples of the laboratory-determined minimum lethal concentration and are intended to be rapidly lethal to the target organisms [bib_ref] Potential impact of increased use of biocides in consumer products on prevalence..., Gilbert [/bib_ref] [bib_ref] Biocide use and antibiotic resistance: The relevance of laboratory findings to clinical..., Russell [/bib_ref].
Several recent reviews provide comprehensive comparisons of biocide vs. antibiotic actions and susceptibilities [bib_ref] Bacterial biocide resistance, Tumah [/bib_ref] [bib_ref] Biocide tolerance in bacteria, Morente [/bib_ref] [bib_ref] Significance of biocide usage and antimicrobial resistance in domiciliary environments, Bloomfield [/bib_ref] [bib_ref] Potential impact of increased use of biocides in consumer products on prevalence..., Gilbert [/bib_ref] [bib_ref] Mechanisms of bacterial biocide and antibiotic resistance, Poole [/bib_ref] [bib_ref] Emergence of resistance to antibacterial agents: The role of quaternary ammonium compounds-a..., Buffet-Bataillon [/bib_ref] [bib_ref] Cross-resistance between biocides and antimicrobials: An emerging question, Davin-Regli [/bib_ref]. Authors vary, sometimes markedly, in their assessment of the potential significance of altered biocide susceptibility among bacteria. Gilbert and McBain [bib_ref] Potential impact of increased use of biocides in consumer products on prevalence..., Gilbert [/bib_ref] provide a useful glossary of the often confusing, and sometimes misused, terminology in this field. Classification and activity of biocides are well reviewed by McDonnell and Russell [bib_ref] Antiseptics and disinfectants: Activity, action, and resistance, Mcdonnell [/bib_ref] , whilst Maillard [bib_ref] Bacterial target sites for biocide action, Maillard [/bib_ref] and Ortega Morente et al. [bib_ref] Biocide tolerance in bacteria, Morente [/bib_ref] provide further discussion of known and postulated targets and modes of action for commonly-used biocides.
Mechanisms of biocide action are incompletely understood and appear to be varied, reflecting the diversity of biocides, applications and associated degrees of toxicity for higher organisms. Only one commonly-used agent, the bisphenol triclosan, has a well-documented specific enzyme target: the enoyl acyl carrier protein reductase (FabI and homologues), involved in fatty acid synthesis [bib_ref] The impact of triclosan on the spread of antibiotic resistance in the..., Carey [/bib_ref]. Exposure to triclosan and the development of reduced susceptibility is associated with mutation plus upregulation of fabI expression [bib_ref] Analysis of triclosan-selected Salmonella enterica mutants of eight serovars revealed increased aminoglycoside..., Rensch [/bib_ref]. Some workers believe that all observed effects of triclosan at in-use concentrations are mediated through this specific pathway [bib_ref] Biocide tolerance in bacteria, Morente [/bib_ref] , although in S. aureus a lack of correlation between growth phase and kill rate of triclosan [bib_ref] Triclosan and antibiotic resistance in Staphylococcus aureus, Suller [/bib_ref] suggests that less-specific cell damage may also occur, at least in some Gram-positive organisms.
Most biocides when applied at in-use concentrations appear to affect multiple targets, with membrane effects being a common theme. These may manifest as physical disruption and leakage of cell contents, disruption of membrane ion gradients, and/or permeabilising effects resulting in enhanced uptake of the biocide itself into the cell interior [bib_ref] Biocide tolerance in bacteria, Morente [/bib_ref] [bib_ref] Potential impact of increased use of biocides in consumer products on prevalence..., Gilbert [/bib_ref] [bib_ref] Antiseptics and disinfectants: Activity, action, and resistance, Mcdonnell [/bib_ref] [bib_ref] Development of antimicrobial resistance in Campylobacter jejuni and Campylobacter coli adapted to..., Mavri [/bib_ref]. Effects on cytoplasmic components appear to be similarly diverse and to affect multiple targets. Given the centrality of cell membrane effects and the need for biocides to traverse the cell envelope and cytoplasmic membrane, it is perhaps unsurprising that Gram-negative bacteria are generally less susceptible to many biocides than are Gram-positive organisms [bib_ref] Bacterial biocide resistance, Tumah [/bib_ref] [bib_ref] Biocide tolerance in bacteria, Morente [/bib_ref] , and that some common biocides have lipophilic domains in their molecules, allowing close association with cell membrane phospholipids. These include the biguanide chlorhexidine and cationic surfactants, exemplified by the versatile family of quaternary ammonium compounds (QAC) [bib_ref] Antiseptics and disinfectants: Activity, action, and resistance, Mcdonnell [/bib_ref] [bib_ref] Emergence of resistance to antibacterial agents: The role of quaternary ammonium compounds-a..., Buffet-Bataillon [/bib_ref].
## Heavy metals
The use, particularly in pig and poultry production, of heavy metal-containing compounds for growth promotion and therapy of intestinal disease potentially selects for micro-organisms with reduced susceptibilities to heavy metals and co-selects for reduced susceptibility to other antimicrobials. Unlike biocides (but in common with antibiotic growth promoters) these compounds are used at inhibitory rather than lethal concentrations, thus potentially allowing resistance to emerge within the alimentary tract and the immediate farm environment at a higher frequency than following biocide application.
However, unlike antibiotic growth promoters, such metals are very persistent in the environment. Heavy metals may accumulate in soil, water and sediments from agricultural practices as well as other sources such as aquacultural and marine antifouling treatments or industrial effluent. In one study in the USA around 90% of in-feed copper and zinc was found to be shed in faeces [bib_ref] In-feed use of heavy metal micronutrients in US swine production systems and..., Medardus [/bib_ref] , and it has been estimated in the UK that livestock may in some localities be the major source of environmental contamination by zinc and copper [bib_ref] An inventory of heavy metals inputs to agricultural soils in England and..., Nicholson [/bib_ref]. Thus, such destinations provide further environments (in addition to intestinal contents and faeces) in which co-selection for reduced susceptibilities may occur [bib_ref] Heavy metal driven co-selection of antibiotic resistance in soil and water bodies..., Seiler [/bib_ref] [bib_ref] Cu exposure under field conditions coselects for antibiotic resistance as determined by..., Berg [/bib_ref]. Furthermore, it has been hypothesized that genes coding for reduced susceptibilities to copper and zinc among diverse bacteria have long been selected for in the environment by eukaryote phagocytes such as amoebae, which may use copper-and zinc-mediated bacterial killing mechanisms [bib_ref] Survival in amoeba-a major selection pressure on the presence of bacterial copper..., Hao [/bib_ref].
The modes of action of copper and zinc when used in pig and poultry feed as growth promoters are not known with certainty and probably involve suppression of bacterial fermentation, but possibly also modulation of host responses to bacterial lipopolysaccharide [bib_ref] Zinc and copper in animal feed-Development of resistance and co-resistance to antimicrobial..., Yazdankhah [/bib_ref]. Effects on host growth are not consistent and may be dependent on animal age [bib_ref] Copper resistance in Enterococcus faecium, mediated by the tcrB gene, is selected..., Hasman [/bib_ref] [bib_ref] Effects of copper sulfate, zinc oxide, and neoterramycin on weanling pig growth..., Shelton [/bib_ref] , whilst permitted inclusion rates vary between regulatory regimes and according to age. In the EU, higher inclusion rates are permitted for young piglets and in short-term "medicinal remedies", which for inorganic zinc compounds may result in inclusion rates in excess of thirty times basal requirements [bib_ref] Co-selection of antibiotic and metal resistance, Baker-Austin [/bib_ref]. Generally-permitted levels in the EU are around 3 to 4 times the basal requirements, and at these levels no therapeutic effects on enteric disease would be expected. Supplementation of ruminant feed with metals is much less intensive than for pigs and poultry, and in particular sheep are highly susceptible to copper toxicity [bib_ref] Susceptibility of different bacterial species isolated from food animals to copper sulphate,..., Aarestrup [/bib_ref]. Other heavy metals added to animal feed that may result in elevated environmental deposits include iron, cobalt and manganese [bib_ref] Heavy metal driven co-selection of antibiotic resistance in soil and water bodies..., Seiler [/bib_ref]. Arsenic is a metalloid element that is still permitted as a therapeutic and growth promoting agent in some territories, although its relationship to altered bacterial susceptibilities to other antimicrobials has been little studied [bib_ref] The environmental and public health risks associated with arsenical use in animal..., Silbergeld [/bib_ref].
## Mechanisms linking altered susceptibilities of bacteria to antibiotics, biocides and heavy metals
## Terminology and testing
Terms such as "resistance" and "tolerance" have acquired specific technical meanings in the field of antimicrobials, and looser usage is generally discouraged. Current terminology distinguishes between clinical and microbiological antibiotic resistance. Clinical resistance is present when phenotypic testing of a microbe/antibiotic combination against a clinical breakpoint indicates that therapeutic failure is likely, even with maximal dosing. Microbiological resistance is defined by the presence of an acquired or mutational resistance mechanism to the drug in question, in comparison with a fully-susceptible "wild-type", and may be assessed by genetic analysis or phenotypic testing against a wild-type cut-off value. The clinical resistance scenario is clearly not applicable in the case of biocides, so to avoid ambiguity it is desirable to avoid using "resistance" in relation to the activity of biocides [bib_ref] Significance of biocide usage and antimicrobial resistance in domiciliary environments, Bloomfield [/bib_ref]. Similarly, the non-specific use of the term "tolerance" is discouraged, in deference to its technical meaning of an increase in the minimum bactericidal concentration of a substance without alteration in the MIC. The preferred terminology of many authors concerning variation in the effects of biocides upon bacteria is "reduced/increased susceptibility", or variants thereof. This has been adopted in the present review.
The testing of antibiotic susceptibility by determination of MIC has a sound basis in clinical application, as lethality and speed of killing is less significant than suppression of growth, in the face of immune system activity. However, for biocides the time required for complete killing of the target organisms, commonly measured as a ≥5 log reduction in viable count [bib_ref] Control of Salmonella in food related environments by chemical disinfection, Møretrø [/bib_ref] , is in most cases a more appropriate measure of effectiveness, and a prolongation of this kill time may be viewed as good evidence for reduced susceptibility. For some biocides, such as QAC and chlorhexidine, there appears to be a reasonably consistent relationship between changes in MIC and in kill time against planktonic suspensions of certain bacteria [bib_ref] Antibiotic and biocide resistance in methicillin-resistant Staphylococcus aureus and vancomycin-resistant enterococcus, Suller [/bib_ref] , but for many others correlations between MIC and kill times are poor or not seen [bib_ref] Significance of biocide usage and antimicrobial resistance in domiciliary environments, Bloomfield [/bib_ref] [bib_ref] Triclosan and antibiotic resistance in Staphylococcus aureus, Suller [/bib_ref] [bib_ref] Antibiotic and biocide resistance in methicillin-resistant Staphylococcus aureus and vancomycin-resistant enterococcus, Suller [/bib_ref] [bib_ref] Exposure of Escherichia coli ATCC 12806 to sublethal concentrations of food-grade biocides..., Capita [/bib_ref] [bib_ref] Chloroxylenol-and triclosan-tolerant bacteria from industrial sources-Susceptibility to antibiotics and other biocides, Lear [/bib_ref].
Additional major complications are the substantial differences between kill tests conducted in suspension and on surfaces, and with organisms in differing states of growth, attachment and water stress [bib_ref] Significance of biocide usage and antimicrobial resistance in domiciliary environments, Bloomfield [/bib_ref] [bib_ref] Potential impact of increased use of biocides in consumer products on prevalence..., Gilbert [/bib_ref] [bib_ref] Emergence of resistance to antibacterial agents: The role of quaternary ammonium compounds-a..., Buffet-Bataillon [/bib_ref] [bib_ref] Control of Salmonella in food related environments by chemical disinfection, Møretrø [/bib_ref] [bib_ref] Effect of triclosan on Salmonella typhimurium at different growth stages and in..., Tabak [/bib_ref]. Indeed, there are few "standard" biocide tests that attempt to mimic accurately and repeatably the likely in-use conditions, and much biocide research is conducted using MIC or minimum bactericidal concentration (MBC) tests in suspension or on solid microbiological media. These have the merit of being inexpensive and repeatable, and they lend themselves to mass screening of many compounds, formulations, bacterial strains and conditions. However, putative reductions in susceptibility identified through such tests may prove not to have significant effects on biocide kill times in more realistic tests of activity [bib_ref] Significance of biocide usage and antimicrobial resistance in domiciliary environments, Bloomfield [/bib_ref] [bib_ref] Emergence of resistance to antibacterial agents: The role of quaternary ammonium compounds-a..., Buffet-Bataillon [/bib_ref].
Nonetheless, where evidence of co-resistance is being sought, MIC testing can reveal significant alterations in antibiotic susceptibilities even though MIC values for biocides need to be interpreted with caution.
In terminology to accurately describe linked changes in susceptibility among antimicrobial compounds (co-selection) there is an important distinction between cross-resistance and co-resistance [bib_ref] Emergence of resistance to antibacterial agents: The role of quaternary ammonium compounds-a..., Buffet-Bataillon [/bib_ref] [bib_ref] The broader context of antibiotic resistance: Zinc feed supplementation of piglets increases..., Bednorz [/bib_ref] [bib_ref] Efficacy of biocides used in the modern food industry to control Salmonella..., Condell [/bib_ref]. Links arising as a consequence of physiological adaptations that have effects on the action of a number of agents, are termed cross-resistance. Examples include efflux pump upregulation or over-expression, reduced cell envelope permeability, or (usually in relation to antibiotics) alteration in a target site or acquisition of a neutralising enzyme (such as an extended-spectrum β-lactamase) that prevents the action of agents from more than one class. By contrast, where the mechanisms of reduced susceptibility are dissimilar but are selected together owing to genetic linkage, the phenomenon is termed co-resistance. There is another common definition of cross-resistance, i.e., resistance to related compounds such as members of the same antibiotic family, but this is too narrow to be of much use in the present discussion.
## Identified mechanisms of co-selection
On initial consideration, differences between antibiotics and biocides in respect of targets and of modes and intensity of action, would suggest that there is not likely to be much common ground between the two classes of agents in respect of reduced susceptibilities of micro-organisms. Indeed, antibiotic effects on bacteria can be markedly reduced by single-step mutations in target enzymes, or by the acquisition of neutralising enzymes such as β-lactamases [bib_ref] Mechanisms of bacterial biocide and antibiotic resistance, Poole [/bib_ref]. Clinical doses are carefully controlled to avoid toxicity or other detrimental effects and distribution of antibiotics into certain tissues may be poor. Therefore such specific, single-step changes can result in clinical resistance to antibiotics, or to further selection for higher resistance. Similar mechanisms are seen only rarely against biocides (mutation and upregulation of fabI in relation to triclosan [bib_ref] Analysis of triclosan-selected Salmonella enterica mutants of eight serovars revealed increased aminoglycoside..., Rensch [/bib_ref] being a case), and are considered likely to have little effect in "real world" applications where application concentrations are typically multiples of the lethal dose and several targets are affected simultaneously [bib_ref] Significance of biocide usage and antimicrobial resistance in domiciliary environments, Bloomfield [/bib_ref].
However, there are some phenomena that confer reduced susceptibility both to biocides and to antibiotics and which are either part of the normal adaptive repertoire of micro-organisms (intrinsic) or are readily acquired by mutation or genetic transfer under appropriate conditions. Intrinsic resistance may be specific to a bacterial grouping (such as mycobacteria or spore-formers), may be exhibited following phenotypic adaptations to the environment (such as biofilm formation or nutrient stress responses), or be associated with survival in relatively protected environments, such as phagocytes in the body or protozoa in the soil [bib_ref] Significance of biocide usage and antimicrobial resistance in domiciliary environments, Bloomfield [/bib_ref] [bib_ref] Potential impact of increased use of biocides in consumer products on prevalence..., Gilbert [/bib_ref]. The phenomena of low cell wall permeability and active efflux of toxic molecules by multi-drug transporters, discussed below under adaptive resistance, can be considered part of intrinsic resistance mechanisms in some instances, notably in the case of Pseudomonas aeruginosa, which exhibits resistance to many antibiotics by virtue of small outer membrane porin channels and vigorous efflux activity [bib_ref] Potential impact of increased use of biocides in consumer products on prevalence..., Gilbert [/bib_ref]. Whilst some of these examples may manifest as apparent reduced susceptibility to antibiotic or biocide without there being any long-term change to the organism, they do potentially allow some proportion of the bacterial population to survive an otherwise terminal challenge, increasing the risk of selection of organisms permanently adapted to the antimicrobial agent.
By contrast to intrinsic resistance, acquired or adaptive resistance is associated with changes in the organism caused by mutation and/or altered expression of endogenous genes, or by transfer of mobile exogenous genetic material [bib_ref] Bacterial biocide resistance, Tumah [/bib_ref] [bib_ref] Mechanisms of bacterial resistance to biocides, Russell [/bib_ref]. Adaptation may be a constitutional and essentially non-reversible effect, typically following mutation. Alternatively, it may be a transient and regulated response to the conditions. Triclosan appears to select for mutational adaptation in Stenotrophomonas maltophilia more intensely than does the QAC benzalkonium chloride [bib_ref] Recent advances in the potential interconnection between antimicrobial resistance to biocides and..., Oggioni [/bib_ref] , indicating that biocides differ in their propensity to induce mutational adaptation in any given organism. Some mechanisms can manifest as either stable and constitutional or transient and regulated types of adaptation, with the Multiple Antibiotic Resistance (MAR) phenomenon of Enterobacteriaceae (discussed later in this section) being a case in point. The acquisition of resistance mechanisms by genetic transfer may, depending on the case, present a more or less regulated adaptation. Acquired and adaptive mechanisms associated variously with co-and/or cross-resistance include: biofilm capability, multi drug efflux, altered cell envelope permeability, and target site mutation and over-expression.
Biofilms have an extracellular matrix that provides a diffusion barrier and an enhanced medium for bacterial signalling and genetic exchange, plus a potential site for neutralisation or binding of chemical agents: enzymes degrading triclosan, chlorhexidine and some oxidising agents, among others, are known [bib_ref] Bacterial biocide resistance, Tumah [/bib_ref] [bib_ref] Biocide tolerance in bacteria, Morente [/bib_ref] [bib_ref] Biofilms in vitro and in vivo: Do singular mechanisms imply cross-resistance?, Gilbert [/bib_ref]. Biofilms similarly provide an extracellular site for sequestration of metal ions [bib_ref] Heavy metal driven co-selection of antibiotic resistance in soil and water bodies..., Seiler [/bib_ref] [bib_ref] Co-selection of antibiotic and metal resistance, Baker-Austin [/bib_ref]. Minor elevations in MIC may have increased significance for survival of a bacterium in the context of a biofilm diffusion gradient [bib_ref] Biocide tolerance and the harbingers of doom, Mcbain [/bib_ref] ; furthermore, bacteria in the centre of biofilms are relatively nutrient-deprived and therefore may exhibit stress responses that additionally diminish their susceptibility to chemical attack [bib_ref] Biocide use and antibiotic resistance: The relevance of laboratory findings to clinical..., Russell [/bib_ref] [bib_ref] Biofilms in vitro and in vivo: Do singular mechanisms imply cross-resistance?, Gilbert [/bib_ref]. There is evidence for protection of both Gram-positive and Gram-negative bacteria by biofilm, and elevated MIC values in the order of 10 to 1000-fold have been observed for benzalkonium chloride and triclosan, among other biocides, in respect of activity against biofilmed versus planktonic bacteria including Listeria spp. [bib_ref] Bacterial biocide resistance, Tumah [/bib_ref] [bib_ref] Biocide tolerance in bacteria, Morente [/bib_ref]. Biofilm formation is likely to be part of the normal environment-responsive repertoire for capable organisms, although in some cases there may be selection of mutants with hyper-expressed biofilm capacity [bib_ref] Adaptation to benzalkonium chloride and ciprofloxacin affects biofilm formation potential, efflux pump..., Pagedar [/bib_ref] [bib_ref] Proteomic and phenotypic analysis of triclosan tolerant verocytotoxigenic Escherichia coli O157:H19, Sheridan [/bib_ref].
Active transport of molecules to the periplasmic or extracellular environment via multi-drug efflux pumps is a well-established feature of biocide and antibiotic resistance [bib_ref] Efflux pumps as antimicrobial resistance mechanisms, Poole [/bib_ref] , either as intrinsic or acquired mechanisms. Many Gram-negative organisms encode chromosomal broad-substrate efflux pumps [bib_ref] Does the wide use of quaternary ammonium compounds enhance the selection and..., Hegstad [/bib_ref] , and a variety of multi-drug pumps are similarly-encoded by some Gram-positive organisms including S. aureus [bib_ref] Efflux pumps as antimicrobial resistance mechanisms, Poole [/bib_ref]. Alternatively, acquired resistance may be associated with efflux pumps introduced on mobile genetic elements among both Gram-positive and Gram-negative bacteria, or with mutations causing constitutional over-expression of Gram-negative efflux activity [bib_ref] Potential impact of increased use of biocides in consumer products on prevalence..., Gilbert [/bib_ref].
Multi-drug efflux pumps actively expel a broad range of unrelated and structurally diverse compounds, which includes major biocide classes, dyes and antibiotics, driven by trans-membrane ion gradients (protons or sodium) or ATP hydrolysis. Poole [bib_ref] Mechanisms of bacterial biocide and antibiotic resistance, Poole [/bib_ref] [bib_ref] Efflux pumps as antimicrobial resistance mechanisms, Poole [/bib_ref] and Putman et al. [bib_ref] Molecular properties of bacterial multidrug transporters. Microbiol, Putman [/bib_ref] have provided excellent reviews. Substrates characteristically have a high degree of hydrophobicity, and there is fair evidence that, in many cases at least, the cytoplasmic membrane, rather than the cytoplasm, is the site of initial binding of compounds [bib_ref] Molecular properties of bacterial multidrug transporters. Microbiol, Putman [/bib_ref]. This may contribute to the effectiveness of efflux pumps against primarily membrane-active biocides. Whilst efflux pumps may provide only modest reductions in susceptibility to antimicrobial compounds, they appear to have an additional role in the development of higher-level resistance, either by additive or synergistic action with other mechanisms [bib_ref] Ineffectiveness of topoisomerase mutations in mediating clinically significant fluoroquinolone resistance in Escherichia..., Oethinger [/bib_ref] , or by increasing the frequency of mutational high-level resistance by allowing a greater proportion of organisms to survive antimicrobial exposure [bib_ref] Efflux pumps as antimicrobial resistance mechanisms, Poole [/bib_ref].
There are also efflux pumps that expel heavy metal ions, such as elements of the czc system encoding a pump for zinc, cobalt and cadmium, and pcoA, part of a copper extrusion system [bib_ref] In-feed use of heavy metal micronutrients in US swine production systems and..., Medardus [/bib_ref]. Whereas some efflux pumps for heavy metals may not have broad substrate affinities, there is evidence of co-transport of non-metals in other cases [bib_ref] Co-selection of antibiotic and metal resistance, Baker-Austin [/bib_ref].
Expression of efflux pumps is typically under transcriptional regulation by DNA-binding products of activator and repressor genes [bib_ref] Molecular properties of bacterial multidrug transporters. Microbiol, Putman [/bib_ref]. These may be products of the particular efflux pump gene cluster or of a global regulator. The AcrR repressor and MarA activator of the AcrAB-TolC efflux pump of E. coli and Salmonella spp. are local and global regulatory products, respectively. Plasmid-borne efflux pumps, such as the QacAB QAC exporter of S. aureus, appear to encode similar regulatory apparatus in addition to structural genes [bib_ref] Genetic basis of resistance to quaternary ammonium compounds-The qac genes and their..., Jaglic [/bib_ref]. Transcriptional regulators may themselves be upregulated by substrates or bind them directly, thereby providing routes by which efflux activity may be modulated by the presence of potentially toxic substances [bib_ref] Molecular properties of bacterial multidrug transporters. Microbiol, Putman [/bib_ref]. This protects the cell from expending energy on indiscriminate extrusion of metabolically useful compounds when toxic substrates are not present.
Efflux pumps may be upregulated by biocides, for example chlorhexidine or benzalkonium chloride acting on the MexCD-OprJ pump of P. aeruginosa [bib_ref] Efflux pumps as antimicrobial resistance mechanisms, Poole [/bib_ref] or QAC and other lipophilic cations acting on the QacR repressor [bib_ref] Genetic basis of resistance to quaternary ammonium compounds-The qac genes and their..., Jaglic [/bib_ref]. Thus, strains possessing such efflux pumps and adapted to survive challenge by biocides that upregulate them may have developed a hyper-expressing phenotype that is nonetheless still repressed in the absence of the biocide. However, many substrates do not stimulate their own efflux and this applies to some biocides, for example triclosan in the case of AcrAB-TolC pump. In this scenario, a bacterial strain with efflux-mediated reduced susceptibility to a non-upregulating biocide will likely be a constitutively-expressing mutant, selected from its un-mutated and therefore fully-susceptible peers, and will not therefore downregulate the efflux once the biocide exposure has waned [bib_ref] Biocide abuse and antimicrobial resistance: Being clear about the issues, Gilbert [/bib_ref]. This potentially makes such strains a more persistent risk for cross-resistance, although there will be a fitness cost in the absence of selective pressure. Sometimes, increased efflux may paradoxically result in increased susceptibility to an antimicrobial substance [bib_ref] Molecular properties of bacterial multidrug transporters. Microbiol, Putman [/bib_ref] , possibly as a result of the extrusion of protective substances.
Alterations in cell envelope permeability have substantial potential for cross-resistance, as either intrinsic or adaptive phenomena. Recognised changes associated with reduced susceptibility to biocides relate to the outer membrane of Gram-negative organisms, namely altered fatty acid and protein profiles, and changed surface hydrophobicity and ultrastructure [bib_ref] Antiseptic "resistance": Real or perceived threat?, Sheldon [/bib_ref] [bib_ref] Biocide tolerance in bacteria, Morente [/bib_ref] [bib_ref] Development of antimicrobial resistance in Campylobacter jejuni and Campylobacter coli adapted to..., Mavri [/bib_ref]. However, biocide-adapted bacteria with altered outer membrane proteins do not necessarily show altered permeability in assays [bib_ref] Cross-resistance to antibiotics of Escherichia coli adapted to benzalkonium chloride or exposed..., Langsrud [/bib_ref]. For primarily membrane-active biocides, such as QAC, changes in the cell envelope might also constitute a target site alteration, in addition to a permeability barrier. However, the cytoplasmic membrane appears to be a primary target for QAC and changes here have been little investigated.
Examinations of the role of the cell wall in respect of biocide resistance have also been limited. The waxy, hydrophobic nature of the mycobacterial wall, coupled with narrow cell envelope porins, is associated with resistance of this genus to many hydrophilic biocides [bib_ref] Bacterial biocide resistance, Tumah [/bib_ref] [bib_ref] Biocide tolerance in bacteria, Morente [/bib_ref] [bib_ref] Permeability of the cell wall of Mycobacterium smegmatis, Trias [/bib_ref]. Using the fractional distribution of cell suspensions in aqueous vs. xylene phases Alexander et al. [bib_ref] Disinfectant resistance in antibiotic-resistant organisms, Alexander [/bib_ref] reported consistent and significant changes in cell surface hydrophobicity between disinfectant-susceptible and non-disinfectant-susceptible strains of both Gram-positive (S. aureus) and Gram-negative bacteria. Similarly consistent changes were also noted following higher versus lower challenge concentrations of disinfectant.
It is perhaps unsurprising that some of the above mechanisms may be found operating in concert. For example, many cases of reduced susceptibility to antibiotics and/or biocides show both increased efflux and cell permeability changes [bib_ref] Emergence of resistance to antibacterial agents: The role of quaternary ammonium compounds-a..., Buffet-Bataillon [/bib_ref] [bib_ref] Cross-resistance between biocides and antimicrobials: An emerging question, Davin-Regli [/bib_ref]. Global regulatory responses may drive cross-resistance, by a combination of the mechanisms already described and probably other changes associated with protection and repair of the biosynthetic and metabolic machinery of the cell [bib_ref] Phenotypic and proteomic characterization of multiply antibiotic-resistant variants of Salmonella enterica serovar..., Karatzas [/bib_ref]. These changes may be reversible stress responses or mutational shifts. A well-studied example of the latter is the MAR phenomenon in E. coli and other members of the Enterobacteriaceae. This is a common phenotype associated with antibiotic resistance or controlled exposure to biocides [bib_ref] Do biocides select for antibiotic resistance?, Russell [/bib_ref]. On its own it generates modest increases in MIC values (around four-to eight-fold) for a range of antibiotics [bib_ref] Biocide tolerance in bacteria, Morente [/bib_ref]. Sensitivity to organic solvents, triclosan and pine oil are characteristically decreased, but effects on other biocide MIC values vary according to the particular biocide or biocide mix, and there may indeed be no significant change [bib_ref] Phenotypic and proteomic characterization of multiply antibiotic-resistant variants of Salmonella enterica serovar..., Karatzas [/bib_ref]. Affected systems are under the influence of the marRAB operon, and the phenomenon arises following mutational or substrate-mediated inactivation of the MarR transcriptional repressor, leading to increased expression of the MarA global regulator and causing, among other things, upregulation of the mar operon itself plus the AcrAB-TolC multi-drug efflux pump, and downregulation of the OmpF porin [bib_ref] Molecular properties of bacterial multidrug transporters. Microbiol, Putman [/bib_ref] [bib_ref] Regulation of bacterial drug export systems, Grkovic [/bib_ref] [bib_ref] Overexpression of the marA or soxS regulatory gene in clinical topoisomerase mutants..., Oethinger [/bib_ref]. Substrates that may reversibly activate the mar regulon include salicylate, bile salts and the antibiotics tetracycline and chloramphenicol. It appears that a principal mechanism mediating reversible MarR inactivation in E. coli, and probably other Enterobacteriaceae at least, is liberation of copper from the cell envelope under the action of stressors, with subsequent interaction between copper (II) ions and MarR [bib_ref] The multiple antibiotic resistance regulator MarR is a copper sensor in Escherichia..., Hao [/bib_ref]. Whether this effect is seen under direct copper stress remains to be investigated.
It is clear that other global regulatory systems may exert similar effects, generating cross-resistance in the face of external stimuli. The SoxS (oxidative stress) and Rob regulators in E. coli are homologues of MarA, activating the mar operon as well as other systems [bib_ref] Do biocides select for antibiotic resistance?, Russell [/bib_ref] [bib_ref] Biocide use and antibiotic resistance: The relevance of laboratory findings to clinical..., Russell [/bib_ref] [bib_ref] Molecular properties of bacterial multidrug transporters. Microbiol, Putman [/bib_ref]. Furthermore, inter-bacterial signalling may promote cross-resistance in whole populations. Secretion of tryptophanase by relatively resistant E. coli subjected in a bioreactor to fluoroquinolone or gentamicin was associated with increased efflux, other stress responses and increased fluoroquinolone MIC values among other strains in the same reactor [bib_ref] Bacterial charity work leads to population-wide resistance, Lee [/bib_ref]. This suggests that the spontaneous or regulated development of resistance to antimicrobial challenge by some members of a bacterial community may, via signalling, increase the likelihood of survival of other members of that community by mechanisms common to other cross-resistance phenomena. Potentially this may allow longer survival or more survivors, in turn increasing the likelihood of the development of higher-level resistance.
Co-resistance, arising from the selection of linked genes encoding otherwise unrelated resistance mechanisms to different antimicrobials, is a well-recognised phenomenon in antibiotic use and resistance, such as in the case of the chromosomally-encoded penta-resistance typical of Salmonella Typhimurium DT104 [bib_ref] Epidemic Salmonella typhimurium DT 104-A truly international multiresistant clone, Threlfall [/bib_ref]. However, co-selection via such genetic linkages can occur more broadly, across genes affecting susceptibility to biocides, heavy metals and antibiotics. Co-location of such genes on mobile elements, such as plasmids and transposons [bib_ref] Significance of biocide usage and antimicrobial resistance in domiciliary environments, Bloomfield [/bib_ref] [bib_ref] Antibiotic and antiseptic resistance genes are linked on a novel mobile genetic..., Ciric [/bib_ref] , also raises the possibility of transfer of co-resistance between bacteria. An obvious example is provided by class 1 integrons, which encode a QAC efflux mechanism (qacEΔ1) plus sulphonamide resistance (sul1) and variable other antibiotic resistance genes [bib_ref] Importance of integrons in the diffusion of resistance, Carattoli [/bib_ref]. However, the true significance of QAC efflux in class 1 integrons is debatable, not least because QacEΔ1 is a functionally attenuated Qac pump variant [bib_ref] The 3ʹ-conserved segment of integrons contains a gene associated with multidrug resistance..., Paulsen [/bib_ref]. Variations between bacterial genera in the existence of resistance-related genes and their linkage in the same organism make co-resistance a variable and somewhat arbitrary phenomenon [bib_ref] Genetic basis of resistance to quaternary ammonium compounds-The qac genes and their..., Jaglic [/bib_ref] [bib_ref] Effects of chlortetracycline and copper supplementation on antimicrobial resistance of fecal Escherichia..., Agga [/bib_ref].
## Effects on bacterial fitness
Adaptation for reduced susceptibility to antimicrobial substances often comes with associated costs to the organism. This is particularly significant when the change is mutational resulting in constitutive expression of the resistance mechanism [bib_ref] Potential impact of increased use of biocides in consumer products on prevalence..., Gilbert [/bib_ref]. In the present context, broad substrate efflux pumps provide a good example as they consume cell energy resources and indiscriminately remove some useful metabolic substances from the cell. When a non-upregulating biocide selects for a strain that over-expresses efflux constitutively (such as mutants in MepR or MarR repressors of S. aureus or E. coli, respectively), this strain will subsequently incur the costs of efflux regardless of the presence of the biocide [bib_ref] Significance of biocide usage and antimicrobial resistance in domiciliary environments, Bloomfield [/bib_ref] [bib_ref] Molecular properties of bacterial multidrug transporters. Microbiol, Putman [/bib_ref]. Plasmids encoding resistance to biocides or heavy metals plus antibiotics provide another example, where instability and the fitness cost require a selective pressure for their maintenance [bib_ref] Selection of a multidrug resistance plasmid by sublethal levels of antibiotics and..., Gullberg [/bib_ref]. However, in this latter example, the minimum selective concentration of the agent may be very low in comparison with its MIC [bib_ref] Selection of a multidrug resistance plasmid by sublethal levels of antibiotics and..., Gullberg [/bib_ref].
There may be immediate evidence of fitness costs in the laboratory, in the form of reduced colony size or other growth characteristics [bib_ref] Analysis of triclosan-selected Salmonella enterica mutants of eight serovars revealed increased aminoglycoside..., Rensch [/bib_ref] [bib_ref] Cross-resistance to antibiotics of Escherichia coli adapted to benzalkonium chloride or exposed..., Langsrud [/bib_ref] [bib_ref] Selection of a multidrug resistance plasmid by sublethal levels of antibiotics and..., Gullberg [/bib_ref] , or the cost may be more subtle and seen only in communities where there is competition within and between microbial species. However, co-selection is not uniformly costly and extended exposure may select for compensatory adaptations to restore fitness [bib_ref] Biocide use and antibiotic resistance: The relevance of laboratory findings to clinical..., Russell [/bib_ref]. Furthermore, some resistance adaptations, such as biofilm enhancement, may enhance survival in other environments [bib_ref] Proteomic and phenotypic analysis of triclosan tolerant verocytotoxigenic Escherichia coli O157:H19, Sheridan [/bib_ref].
# Other related phenomena
It is recognised that antibiotic exposure has a non-specific effect on the frequency of mutation and of horizontal gene transfer among bacteria, which may not be associated with any change in susceptibility to the inciting antibiotic. Similarly, the frequency of mutation to antibiotic resistance among Salmonella spp. is elevated following exposure to sub-inhibitory concentrations of phenolic biocides or triclosan [bib_ref] Effect of triclosan or a phenolic farm disinfectant on the selection of..., Randall [/bib_ref]. Non-specific mutagenesis may be a factor in instances where biocide exposure leads to decreased sensitivity to antibiotics but an unchanged, or even reduced, ability to cope with the biocide itself [bib_ref] Development of antimicrobial resistance in Campylobacter jejuni and Campylobacter coli adapted to..., Mavri [/bib_ref] [bib_ref] Triclosan-induced aminoglycoside-tolerant Listeria monocytogenes isolates can appear as small-colony variants, Kastbjerg [/bib_ref]. Furthermore, effects (positive or negative) have been noted with biocides at sub-lethal concentrations in respect of plasmid and phage transfer frequencies [bib_ref] Significance of biocide usage and antimicrobial resistance in domiciliary environments, Bloomfield [/bib_ref] [bib_ref] Potential impact of increased use of biocides in consumer products on prevalence..., Gilbert [/bib_ref] [bib_ref] Distinguishing effects of ultraviolet exposure and chlorination on the horizontal transfer of..., Guo [/bib_ref].
Susceptibility to metals may be reduced by complexing or sequestering the metal ions or by chemical reduction to less toxic states [bib_ref] Heavy metal driven co-selection of antibiotic resistance in soil and water bodies..., Seiler [/bib_ref] [bib_ref] Co-selection of antibiotic and metal resistance, Baker-Austin [/bib_ref] [bib_ref] Escherichia coli mechanisms of copper homeostasis in a changing environment, Rensing [/bib_ref]. Furthermore, there are several identified interactions between antibiotics (or bacterial resistance to antibiotics) and heavy metals, which may have co-selective or counter-selective consequences. These include: formation of complexes between metal cations and certain antibiotics (for example tetracycline) leading to altered absorption and distribution, degrading of β-lactam antibiotics by copper, and metal dependence of certain resistance mechanisms such as β-lactamases and some tetracycline efflux [bib_ref] Metal ion-catalysed hydrolysis of ampicillin in microbiological growth media, Beard [/bib_ref] [bib_ref] Heavy metals in liquid pig manure in light of bacterial antimicrobial resistance, Hölzel [/bib_ref] [bib_ref] Metal ion interaction with penicillins-Part VII: Mixed-ligand complex formation of cobalt (II),..., Mukherjee [/bib_ref].
In order for selection of biocide-insusceptible strains to occur, some proportion of the bacterial population must survive application of biocides, and this is one scenario where the mode of use of biocides would appear to offer few opportunities for survivor selection, compared with antibiotics. However, microbes in stressed or biofilm communities will often show intrinsic resistance and survive a severe challenge [bib_ref] Potential impact of increased use of biocides in consumer products on prevalence..., Gilbert [/bib_ref] [bib_ref] Efficacy of biocides used in the modern food industry to control Salmonella..., Condell [/bib_ref] [bib_ref] Susceptibility of antibiotic-susceptible and antibiotic-resistant hospital bacteria to disinfectants, Rutala [/bib_ref] , as discussed above. Furthermore, use of biocides in the presence of heavy organic soiling or with diluting water containing interfering organic or mineral substances, such as may occur on farms, can be associated with marked reductions in efficacy even at recommended application concentrations [bib_ref] Biocide tolerance in bacteria, Morente [/bib_ref] [bib_ref] Evaluation of commonly-used farm disinfectants in wet and dry models of Salmonella..., Mclaren [/bib_ref]. An allied phenomenon, again of particular relevance to farms but also to sewage treatment and disposal, is the relative environmental persistence of many biocides such as QAC and triclosan and their tendency to bind to organic matter and soil [bib_ref] Emergence of resistance to antibacterial agents: The role of quaternary ammonium compounds-a..., Buffet-Bataillon [/bib_ref] [bib_ref] Does the wide use of quaternary ammonium compounds enhance the selection and..., Hegstad [/bib_ref]. This provides a potential long-lasting low-level exposure of microbes to selective pressure. Concentrations of triclosan in biosolids and sediments have been reported to lie between the MIC of "fully-susceptible" and "tolerant" bacteria [bib_ref] The impact of triclosan on the spread of antibiotic resistance in the..., Carey [/bib_ref].
When considering the effects of biocides and heavy metals on communities of bacteria, changes in the structure and diversity of the community become important elements, such as a shift to predominantly intrinsically-resistant Gram-negative bacteria in the genital tract of catheterised patients subjected to repeated antiseptic washes [bib_ref] Antiseptic and antibiotic resistance in Gram-negative bacteria causing urinary tract infection, Stickler [/bib_ref]. Work with sub-therapeutic antibiotic pressures on pig gut communities in vivo have shown substantial variability in observed effects, with changes in some cases appearing to be minor possibly owing to shifts occurring predominantly at strain or species-level [bib_ref] Temporal changes and the effect of subtherapeutic concentrations of antibiotics in the..., Holman [/bib_ref]. Community-wide changes are often examined in the context of water, soils, sewage and other biosolids, where sub-lethal concentrations of antimicrobials are typical. Phenotypic testing (using cultural and non-cultural methods) and metagenomic approaches allow examination of community-wide antimicrobial sensitivities and the frequency of known antimicrobial resistance genes [bib_ref] Cu exposure under field conditions coselects for antibiotic resistance as determined by..., Berg [/bib_ref] [bib_ref] Coselection for microbial resistance to metals and antibiotics in freshwater microcosms, Stepanauskas [/bib_ref] [bib_ref] Long-term exposure to benzalkonium chloride disinfectants results in change of microbial community..., Tandukar [/bib_ref]. Such studies tend to focus on the effect of a known selection pressure on the frequency of a number of resistance genes, on genetic transfer rates and on sensitivity of the bacterial community to other antimicrobials. The risks of co-resistance arising in specific pathogenic bacteria are considered in relation to shifts in this background.
## Co-resistance in practice: observational and experimental evidence
## Biocides
Many surveys and investigations have involved hospitals or other healthcare environments and S. aureus, particularly the β-lactam-insusceptible meticillin-resistant S. aureus (MRSA) has commonly been targeted. S. aureus strains encode a variety of chromosomal and plasmid-borne efflux pumps, and laboratory chromosomal efflux mutants to single-agent biocides show co-selection for reduced fluoroquinolone sensitivity. However, findings in respect of S. aureus biocide-antibiotic co-resistance have in general been highly variable between studies, between strains and between biocides [bib_ref] Antiseptic "resistance": Real or perceived threat?, Sheldon [/bib_ref] [bib_ref] Potential impact of increased use of biocides in consumer products on prevalence..., Gilbert [/bib_ref] [bib_ref] Triclosan and antibiotic resistance in Staphylococcus aureus, Suller [/bib_ref] [bib_ref] Antibiotic and biocide resistance in methicillin-resistant Staphylococcus aureus and vancomycin-resistant enterococcus, Suller [/bib_ref] [bib_ref] Efflux pumps as antimicrobial resistance mechanisms, Poole [/bib_ref] [bib_ref] Triclosan and antimicrobial resistance in bacteria: An overview, Yazdankhah [/bib_ref] , no doubt affected to some extent by differing methodologies. Apparent biocide tolerance may not be stable and associations seen when comparing MIC values may not be evident when considering kill rates [bib_ref] Antibiotic and biocide resistance in methicillin-resistant Staphylococcus aureus and vancomycin-resistant enterococcus, Suller [/bib_ref] , whilst interpretation can be further hampered by a lack of control isolates to validate an observed specific association.
Other investigations of field strains from healthcare sources have yielded similarly variable findings. An early study examined Gram-negative isolates from urine infections submitted by hospitals or community practitioners and reported MIC values to be above in-use concentrations of some cationic biocides, including chlorhexidine and cetrimide (QAC), among certain genera (Pseudomonas, Proteus, Providencia) associated with multiple antibiotic resistance [bib_ref] Antiseptic and antibiotic resistance in Gram-negative bacteria causing urinary tract infection, Stickler [/bib_ref]. Hospital strains of S. aureus (considered cloxacillin-resistant) and Gram-negative bacteria (gentamicin-resistant) commonly exhibited MIC values for chlorhexidine, cetrimide and chloroxylenol (a phenolic disinfectant) that were above their respective in-use concentrations, whereas corresponding antibiotic-sensitive strains were generally more susceptible [bib_ref] Disinfectant resistance in antibiotic-resistant organisms, Alexander [/bib_ref].
By contrast, and using surface disinfection tests, hospital-derived antibiotic-resistant strains of staphylococci and Gram-negative pathogens did not show significant or consistent differences in kill rates using QAC or phenolic disinfectants, when compared with antibiotic-sensitive laboratory strains [bib_ref] Susceptibility of antibiotic-susceptible and antibiotic-resistant hospital bacteria to disinfectants, Rutala [/bib_ref]. Vancomycin-resistant and -sensitive enterococci showed no differences in MIC values for QAC and chlorhexidine antiseptics [bib_ref] Antibiotic and biocide resistance in methicillin-resistant Staphylococcus aureus and vancomycin-resistant enterococcus, Suller [/bib_ref].
A prospective investigation examined E. coli isolates from bacteraemias in one hospital and found no correlation between source or eventual mortality and the MIC for two QAC compounds [bib_ref] Effect of higher minimum inhibitory concentrations of quaternary ammonium compounds in clinical..., Buffet-Bataillon [/bib_ref]. An association was observed in this study between MIC of the QACs and a trimethoprim-sulphonamide systemic antimicrobial (co-trimoxazole) but QAC MIC values remained below 1% of in-use concentrations. A follow-up integron analysis confirmed the presence of class I integrons, characteristically encoding a QAC efflux pump (qacEΔ1) and sulphamethoxazole resistance (sul1), plus in this case trimethoprim resistance [bib_ref] Importance of integrons in the diffusion of resistance, Carattoli [/bib_ref] [bib_ref] Molecular mechanisms of higher MICs of antibiotics and quaternary ammonium compounds for..., Buffet-Bataillon [/bib_ref].
Several surveys have been performed on biocide (plus or minus antibiotic) susceptibility patterns in bacterial isolates from livestock units. In a study of environmental bacteria recovered from hatcheries in the USA, non-susceptibility to in-use concentrations of QAC, phenolic and glutaraldehyde biocide products was observed among 19 of 350 isolates, these being low-pathogenic Gram-positive and Gram-negative species [bib_ref] Investigation of bacterial resistance to hatchery disinfectants, Willinghan [/bib_ref]. Glutaraldehyde was the most commonly tolerated biocide, which correlated with the frequent and long-term use of this agent in hatcheries. Among 569 Danish livestock isolates examined for MIC in respect of a panel of biocides, no distinct subsets of "low-susceptibility" bacteria were seen, although the MIC ranges were rather wide in some cases: 128-fold and 512-fold ranges for chlorhexidine and benzalkonium chloride, respectively [bib_ref] Susceptibility of different bacterial species isolated from food animals to copper sulphate,..., Aarestrup [/bib_ref]. Using 310 Gram-positive isolates from milking cow teats that had been regularly subjected to iodine or chlorhexidine antisepsis, Martin and Maris [bib_ref] Resistance of 310 Gram-positive strains isolated from milking cow udders, Martin [/bib_ref] demonstrated a significant association among streptococci between reduced susceptibility (elevated MIC) to chlorhexidine and to ampicillin, tetracycline and three aminoglycoside antibiotics.
However, no association was seen between bacteria showing elevated MIC values to chlorhexidine and its use on source animals.
Biocide-manufacturing plants are other environments with bacteria that potentially are regularly exposed to biocides. Isolates from plants manufacturing the phenolic agents triclosan or chloroxylenol were typically of low susceptibility to these agents, or from genera with intrinsically low susceptibility such as Pseudomonas [bib_ref] Chloroxylenol-and triclosan-tolerant bacteria from industrial sources, Lear [/bib_ref]. For those isolates with elevated MIC values to triclosan or chloroxylenol there was no evidence of associated reductions in susceptibilities to other biocides (chlorhexidine, benzalkonium chloride or phenol), nor to any of a panel of antibiotics [bib_ref] Chloroxylenol-and triclosan-tolerant bacteria from industrial sources-Susceptibility to antibiotics and other biocides, Lear [/bib_ref].
Laboratory investigations into co-selection often include "training" of field or laboratory strains by repeated subculture with stepwise increases in the concentration of biocides in liquid or solid media, by the use of biocide concentration gradients in solid media, or by extended culture in biocides at near-inhibitory concentrations. This may yield strains with decreased susceptibility to the "training" biocide and other antimicrobials (usually in terms of increased MIC), which may or may not be stable upon subculture in the absence of the biocide. Alternatively, there may be little or no change in MIC value for the training biocide, but more marked changes in those for certain antibiotics. Findings tend to be highly individual to the particular combination of biocide and bacterial strain being examined [bib_ref] Significance of biocide usage and antimicrobial resistance in domiciliary environments, Bloomfield [/bib_ref]. Whilst being a useful tool for investigating possible mechanisms of co-selection, the relevance to "real world" interactions of such carefully-controlled and prolonged exposure to biocides has often been questioned.
Stepwise training procedures for triclosan may produce dramatic increases in MIC (three or four orders of magnitude) for E. coli and some Klebsiella isolates, whereas changes reported for other species including staphylococcal human skin isolates are far smaller, in the order of one-to ten-fold [bib_ref] Proteomic and phenotypic analysis of triclosan tolerant verocytotoxigenic Escherichia coli O157:H19, Sheridan [/bib_ref] [bib_ref] Adaptive resistance to biocides in Salmonella enterica and Escherichia coli O157 and..., Braoudaki [/bib_ref] [bib_ref] Effects of chronic triclosan exposure upon the antimicrobial susceptibility of 40 ex-situ..., Ledder [/bib_ref] [bib_ref] Selection for high-level resistance by chronic triclosan exposure is not universal, Mcbain [/bib_ref] [bib_ref] The potential for biocide tolerance in Escherichia coli and its impact on..., Sheridan [/bib_ref]. Data for training Salmonella enterica serovars with triclosan is inconsistent, with minor effects in one report [bib_ref] Effects of chronic triclosan exposure upon the antimicrobial susceptibility of 40 ex-situ..., Ledder [/bib_ref] , and MIC values elevated 64-to >512-fold in others [bib_ref] Analysis of triclosan-selected Salmonella enterica mutants of eight serovars revealed increased aminoglycoside..., Rensch [/bib_ref] [bib_ref] Prolonged treatment of Salmonella enterica serovar Typhimurium with commercial disinfectants selects for..., Karatzas [/bib_ref]. For other commonly-tested single-agent biocides such as chlorhexidine, chlorine, peroxyacetic acid and QAC including benzalkonium chloride, training procedures tend to yield up to eight-fold elevations in MIC, with occasionally a decrease in MIC being observed for certain combinations of bacterial strain and biocide [bib_ref] Adaptation to benzalkonium chloride and ciprofloxacin affects biofilm formation potential, efflux pump..., Pagedar [/bib_ref] [bib_ref] Cross-resistance to antibiotics of Escherichia coli adapted to benzalkonium chloride or exposed..., Langsrud [/bib_ref] [bib_ref] Selection for high-level resistance by chronic triclosan exposure is not universal, Mcbain [/bib_ref] [bib_ref] Comparison of antibiotic resistance patterns in Listeria monocytogenes and Salmonella enterica strains..., Alonso-Hernando [/bib_ref] [bib_ref] Resistance to phenicol compounds following adaptation to quaternary ammonium compounds in Escherichia..., Soumet [/bib_ref].
Following such biocide training of strains, allied shifts in susceptibility to other biocides and antibiotics are commonly reported, although variable in nature [bib_ref] The impact of triclosan on the spread of antibiotic resistance in the..., Carey [/bib_ref]. Bradoukai and Hilton [bib_ref] Adaptive resistance to biocides in Salmonella enterica and Escherichia coli O157 and..., Braoudaki [/bib_ref] reported reductions in susceptibility to a panel of antibiotics and to the other test biocides, following stepwise training of E. coli O157 and various Salmonella serovars against triclosan, chlorhexidine and benzalkonium chloride. The authors reported shifts occurring in a minority of strain-antimicrobial pairings, with the pattern varying between species and between Salmonella serovars. Increases in antibiotic MIC values in this study varied between two-and 500-fold. Exposure training of veterinary field isolates and a laboratory strain of E. coli to three QAC compounds yielded elevations of MIC that were in some cases above clinical breakpoints for resistance, for a number of antibiotics: phenicol, tetracycline, fluoroquinolone, β-lactam classes and trimethoprim [bib_ref] Resistance to phenicol compounds following adaptation to quaternary ammonium compounds in Escherichia..., Soumet [/bib_ref]. Stepwise training of E. coli strains with benzalkonium chloride or chloramphenicol resulted in mutual increases in MIC: around six-fold for the QAC training agent, two-fold for the QAC after antibiotic adaptation and over twenty-fold for chloramphenicol after the QAC adaptation [bib_ref] Cross-resistance to antibiotics of Escherichia coli adapted to benzalkonium chloride or exposed..., Langsrud [/bib_ref]. Benzalkonium chloride-adapted strains showed higher survival in lethality tests and increased MIC values, of 1.5-to >20-fold, for antibiotics in most cases.
Training Listeria monocytogenes and Salmonella serovars Enteritidis and Typhimurium with five poultry carcase decontaminants (including chlorine dioxide, acidified sodium chlorite and peroxyacetic acid) resulted in reduction in antibiotic susceptibilities to therapeutic compounds in a minority of cases, past clinical resistance breakpoints in 13 of 300 pairings [bib_ref] Comparison of antibiotic resistance patterns in Listeria monocytogenes and Salmonella enterica strains..., Alonso-Hernando [/bib_ref]. There was no strong pattern, although development of resistance to the aminoglycosides streptomycin and neomycin was common. By contrast, triclosan training of a number of Salmonella enterica serovars was associated with neutral effects on sensitivity or, in the case of the two aminoglycosides tested (gentamicin and kanamycin), increases in susceptibility [bib_ref] Analysis of triclosan-selected Salmonella enterica mutants of eight serovars revealed increased aminoglycoside..., Rensch [/bib_ref]. Reviewing the evidence on peroxyacetic acid, chlorine, chlorite and trisodium phosphate as a poultry carcass and meat decontaminant, the European Food Standards Agency concluded that there was no evidence that their use would lead to acquired reduced susceptibility among contaminating bacteria, nor to acquired resistance to antibiotics.
Shorter single exposures to sub-lethal concentrations of biocides appear to have effects upon antibiotic sensitivity that are similar or related to those observed with stepwise training procedures. Salmonella Enteritidis cells surviving a short exposure to in-use concentrations of chlorine exhibited up to eight-fold increases in MIC values for tetracycline, nalidixic acid and chloramphenicol [bib_ref] Exposure of Salmonella Enteritidis to chlorine or food preservatives increases susceptibility to..., Potenski [/bib_ref]. When the low-pathogenic Gram-negative anaerobe Bacteroides fragilis was exposed to triclosan, benzalkonium chloride, oxidising agents and chlorhexidine only the last was associated with elevated MIC values to antibiotics, these being of the β-lactam and tetracycline classes but not chloramphenicol or a fluoroquinolone [bib_ref] Induction of multiple antibiotic resistance in Bacteroides fragilis by benzene and benzene-derived..., Pumbwe [/bib_ref]. Interestingly, non-biocidal drugs known to induce efflux (salicylate, ibuprofen, paracetamol) exerted a similar effect as chlorhexidine. Similarly, one-time exposure of two E. coli strains to benzalkonium chloride, salicylate, a bile salt and the oxidising stressor methyl viologen resulted in modest increases in MIC values for benzalkonium chloride and chloramphenicol in most cases [bib_ref] Cross-resistance to antibiotics of Escherichia coli adapted to benzalkonium chloride or exposed..., Langsrud [/bib_ref].
In some examinations of biocide training and the co-selection of antibiotic resistance, a role for efflux can be demonstrated or reasonably inferred. Inactivation of the AcrAB-TolC efflux pump increased E. coli susceptibility to triclosan ten-fold, whilst overexpression of AcrAB-TolC or induction of the MAR phenotype increased the triclosan MIC by factors of 3 and 20, respectively [bib_ref] Overexpression of marA, soxS, or acrAB produces resistance to triclosan in laboratory..., Mcmurry [/bib_ref]. Among benzalkonium chloride-adapted E. coli, ethidium bromide efflux was increased for one of two strains in the study by Langsrud et al. [bib_ref] Cross-resistance to antibiotics of Escherichia coli adapted to benzalkonium chloride or exposed..., Langsrud [/bib_ref]. Partial or complete reversal of acquired antibiotic resistances upon application of an efflux pump inhibitor have also been reported in E. coli by Soumet et al. [bib_ref] Resistance to phenicol compounds following adaptation to quaternary ammonium compounds in Escherichia..., Soumet [/bib_ref] and in B. fragilis by Pumbwe et al. [bib_ref] Induction of multiple antibiotic resistance in Bacteroides fragilis by benzene and benzene-derived..., Pumbwe [/bib_ref].
Reduced susceptibility of Salmonella enterica serovars to triclosan was partially reversed by exposure to an efflux pump inhibitor [bib_ref] Analysis of triclosan-selected Salmonella enterica mutants of eight serovars revealed increased aminoglycoside..., Rensch [/bib_ref] , whilst three Salmonella serovars with reduced susceptibilities to benzalkonium chloride, triclosan or erythromycin following training showed a return to susceptibility in the presence of an efflux inhibitor [bib_ref] Mechanisms of resistance in Salmonella enterica adapted to erythromycin, benzalkonium chloride and..., Braoudaki [/bib_ref]. Over-expression of AcrAB was mildly protective of Salmonella against some biocides, but not an oxidising mix [bib_ref] Exposure of Salmonella enterica serovar Typhimurium to high level biocide challenge can..., Whitehead [/bib_ref] , whilst introducing the MAR repressor (MarR) on a plasmid abolished increases in antibiotic MIC associated with chlorine exposure in Salmonella Enteritidis [bib_ref] Exposure of Salmonella Enteritidis to chlorine or food preservatives increases susceptibility to..., Potenski [/bib_ref].
Evidence for efflux in co-selection comes from other sources also. Multidrug efflux pump genes were far more common than antibiotic-specific resistance genes among Gram-negative food isolates with reduced susceptibility to biocides plus antibiotics [bib_ref] Antimicrobial resistance determinants in antibiotic and biocide-resistant gram-negative bacteria from organic foods, Fernández-Fuentes [/bib_ref]. If co-selection had occurred in these bacteria, then the dominant mechanism appeared to be cross-resistance, probably involving efflux.
Where cross-resistance mechanisms are involved, reduced susceptibility to biocides may follow from the development of antibiotic resistance, as well as vice versa. Reciprocal, albeit asymmetric, adaptation to benzalkonium chloride and chloramphenicol by E. coli reported by Langsrud et al. [bib_ref] Cross-resistance to antibiotics of Escherichia coli adapted to benzalkonium chloride or exposed..., Langsrud [/bib_ref] provides a good example, as does an investigation whereby fluoroquinolone-or tetracycline-resistant E. coli or Salmonella strains inoculated onto cloths were killed less efficiently by phenolic biocides than were antibiotic-susceptible strains [bib_ref] Efflux pump activity in fluoroquinolone and tetracycline resistant Salmonella and E. coli..., Thorrold [/bib_ref]. Evidence for efflux activity was presented.
Multi-drug efflux mechanisms often are plasmid-borne among Gram-positive bacteria [bib_ref] Potential impact of increased use of biocides in consumer products on prevalence..., Gilbert [/bib_ref] , and mobile efflux pumps are seen among Gram-negative organisms too. Where efflux genes on mobile elements are accompanied by specific antibiotic resistance genes [bib_ref] Antiseptic and antibiotic resistance plasmid in Staphylococcus aureus that possesses ability to..., Yamamoto [/bib_ref] , cross-resistance and co-resistance may be acquired together. Clinical isolates of E. coli with reduced QAC and antibiotic susceptibilities had integrons encoding both the QacEΔ1 efflux pump and trimethoprim/sulphonamide resistance, and in addition over-expressed the AcrAB-TolC efflux pump [bib_ref] Molecular mechanisms of higher MICs of antibiotics and quaternary ammonium compounds for..., Buffet-Bataillon [/bib_ref]. The Qac efflux pumps are widespread among staphylococci and a range of Gram-negative bacteria, being typically plasmid-borne and often occurring in conjunction with specific antibiotic resistance genes [bib_ref] Biocide tolerance in bacteria, Morente [/bib_ref] [bib_ref] Emergence of resistance to antibacterial agents: The role of quaternary ammonium compounds-a..., Buffet-Bataillon [/bib_ref] [bib_ref] Genetic basis of resistance to quaternary ammonium compounds-The qac genes and their..., Jaglic [/bib_ref]. Although Qac-type pumps will expel a number of lipophilic cationic compounds including QAC and other biocides, evidence for their significance in biocide insusceptibility is conflicting [bib_ref] Genetic basis of resistance to quaternary ammonium compounds-The qac genes and their..., Jaglic [/bib_ref] [bib_ref] The role of the qacA gene in mediating resistance to quaternary ammonium..., Cervinkova [/bib_ref] , and their association with antibiotic resistance may owe more to genetic linkage with resistance genes than to efflux-mediated cross-resistance.
In a study specifically examining biofilm-forming capability, E. coli isolates from dairy equipment that had reduced susceptibility to benzalkonium chloride and/ or ciprofloxacin proved to have superior biofilm capacity, whilst training more susceptible isolates with the antimicrobial agents led to decreased susceptibility, of varying stability, in parallel with improved biofilm formation and evidence of increased efflux activity [bib_ref] Adaptation to benzalkonium chloride and ciprofloxacin affects biofilm formation potential, efflux pump..., Pagedar [/bib_ref]. Improved biofilm capability plus efflux has also been seen with triclosan-adapted E. coli [bib_ref] Proteomic and phenotypic analysis of triclosan tolerant verocytotoxigenic Escherichia coli O157:H19, Sheridan [/bib_ref] , and Salmonella Typhimurium in biofilm were substantially more resistant to killing than were planktonic stationary-phase cells [bib_ref] Effect of triclosan on Salmonella typhimurium at different growth stages and in..., Tabak [/bib_ref].
Using a bacterial community approach to examining co-selection, bioreactors were seeded with river sediment communities and subjected over four years to varying nutrient and benzalkonium chloride concentrations. When exposed to the QAC, community diversity reduced and community-wide MIC values for benzalkonium chloride, ciprofloxacin, tetracycline and penicillin G all increased [bib_ref] Long-term exposure to benzalkonium chloride disinfectants results in change of microbial community..., Tandukar [/bib_ref]. The frequency of efflux genes increased in biocide-exposed populations and an efflux inhibitor reduced the antibiotic MIC values, but interestingly not that of the QAC. Thus, population restructuring, genetic transfer, efflux and other resistance mechanisms were all potentially involved in this model microcosm of long-term sediment contamination by QAC.
## Heavy metals and other agents added to animal feed
Co-selection mechanisms impacting antibiotic susceptibility have been identified for heavy metals that are similar to those for biocides, namely cross-resistance via extracellular adaptations (biofilms especially), cell envelope changes, efflux and global regulatory responses, plus co-resistance via gene linkage [bib_ref] Heavy metal driven co-selection of antibiotic resistance in soil and water bodies..., Seiler [/bib_ref]. Important respects in which heavy metals differ from biocides include their mode of use in agriculture and aquaculture, where they are used either continuously or for an extended period of time and usually at sub-lethal concentrations, and their marked tendency to persist in manure and the environment [bib_ref] Heavy metal driven co-selection of antibiotic resistance in soil and water bodies..., Seiler [/bib_ref]. Indeed, the extended relationship between pathogens and copper in agricultural waste, which may then be mixed with the soil microbiota, has prompted suggestions that this may be an especially advantageous scenario for the development and transfer of multiple resistance determinants in pathogenic bacteria [bib_ref] Heavy metal driven co-selection of antibiotic resistance in soil and water bodies..., Seiler [/bib_ref] [bib_ref] Cu exposure under field conditions coselects for antibiotic resistance as determined by..., Berg [/bib_ref]. By contrast, some heavy metals (particularly mercury and lead) have been associated with counter-selection leading to reduced antibiotic susceptibility, probably via a variety of mechanisms [bib_ref] Heavy metal driven co-selection of antibiotic resistance in soil and water bodies..., Seiler [/bib_ref] [bib_ref] Heavy metals in liquid pig manure in light of bacterial antimicrobial resistance, Hölzel [/bib_ref].
Many metal-antibiotic insusceptibility correlations have been observed in samples retrieved from heavy metal-contaminated environments, and there are identified cross-resistance mechanisms that can effect such changes including efflux (Listeria monocytogenes, Pseudomonas aeruginosa and E. coli) and co-regulation, via robA in E. coli and czcS in Pseudomonas aeruginosa [bib_ref] Co-selection of antibiotic and metal resistance, Baker-Austin [/bib_ref]. In soil long-contaminated by historical industrial use of copper and without recent agricultural activity, both culture and non-cultural viability assays of soil bacteria showed decreased copper and antibiotic susceptibilities, compared with local matched control samples [bib_ref] Cu exposure under field conditions coselects for antibiotic resistance as determined by..., Berg [/bib_ref].
Despite much data on correlations, evidence for causal links remains scant [bib_ref] Heavy metal driven co-selection of antibiotic resistance in soil and water bodies..., Seiler [/bib_ref]. An exception is provided by Stepanauskas et al. [bib_ref] Coselection for microbial resistance to metals and antibiotics in freshwater microcosms, Stepanauskas [/bib_ref] who reported that river water microcosms subjected to heavy metal (nickel or cadmium) stresses yielded significantly increased frequencies of bacteria resistant to antibiotics, and vice-versa. Similarly, copper amendment of soil to concentrations near the European Union limit for sewage-treated soils was followed by a shift towards reduced copper susceptibility in cultured organisms from the soil microbiota and an allied decrease in multiple antibiotic susceptibilities, close to clinical breakpoints in many instances [bib_ref] Copper amendment of agricultural soil selects for bacterial antibiotic resistance in the..., Berg [/bib_ref]. In the latter study, there was an associated change towards Gram-negative isolates, and many such metal-antibiotic susceptibility correlations may reflect a shift in bacterial community structure and diversity towards naturally-resistant organisms, with limited implications for animal and human pathogens. However, there is evidence of co-resistance in other cases [bib_ref] Heavy metal driven co-selection of antibiotic resistance in soil and water bodies..., Seiler [/bib_ref]. A relevant example is the recently-reported linkage between czrC, associated with reduced zinc and cadmium susceptibility, and the SCCmec genomic island characterising human and livestock-associated MRSA strains [bib_ref] Cloning and occurrence of czrC, a gene conferring cadmium and zinc resistance..., Cavaco [/bib_ref]. The copy number of tetracycline-and sulphonamide-resistance genes (tetA, sul1) in DNA extracts from weaner pigs was significantly increased in animals heavily supplemented with zinc [bib_ref] High dietary zinc supplementation increases the occurrence of tetracycline and sulfonamide resistance..., Vahjen [/bib_ref] , although to what extent that reflected selection of insusceptible organisms vs. a community shift is unclear.
Homeostasis of metabolically-important metals may rely on chromosomally-encoded uptake, efflux, chaperoning and detoxification mechanisms, such as is the case with copper [bib_ref] Escherichia coli mechanisms of copper homeostasis in a changing environment, Rensing [/bib_ref] [bib_ref] Copper homeostasis in Enterococcus hirae, Solioz [/bib_ref]. However, there is fair evidence that heavy metals in contaminated sediments and water may commonly exceed selective concentrations [bib_ref] Heavy metal driven co-selection of antibiotic resistance in soil and water bodies..., Seiler [/bib_ref] and resistance to toxic external concentrations of heavy metals appears commonly to require additional plasmid-borne genes [bib_ref] Potential impact of increased use of biocides in consumer products on prevalence..., Gilbert [/bib_ref] [bib_ref] Co-selection of antibiotic and metal resistance, Baker-Austin [/bib_ref] [bib_ref] Escherichia coli mechanisms of copper homeostasis in a changing environment, Rensing [/bib_ref] [bib_ref] Species distribution and heavy metal resistance of Enterobacteriaceae members isolated from Istanbul..., Çardak [/bib_ref] [bib_ref] tcrB a gene conferring transferable copper resistance in Enterococcus faecium: Occurrence, transferability,..., Hasman [/bib_ref]. Under these circumstances co-resistance may be accompanied by enhanced genetic mobility. An example of plasmid-associated co-resistance with contemporary relevance is the tcrB gene associated with reduced copper susceptibility [bib_ref] Copper resistance in Enterococcus faecium, mediated by the tcrB gene, is selected..., Hasman [/bib_ref] , which shares homology with the homeostatic efflux-associated copB of Enterococcus hirae [bib_ref] tcrB a gene conferring transferable copper resistance in Enterococcus faecium: Occurrence, transferability,..., Hasman [/bib_ref]. In Enterococcus faecium, tcrB is plasmid-borne and shows linkage with erm(B) and vanA, encoding resistance to erythromycin and vancomycin, respectively [bib_ref] Copper resistance in Enterococcus faecium, mediated by the tcrB gene, is selected..., Hasman [/bib_ref] [bib_ref] tcrB a gene conferring transferable copper resistance in Enterococcus faecium: Occurrence, transferability,..., Hasman [/bib_ref]. The carriage of vanA is believed to have been promoted by the glycopeptide growth promoter avoparcin, which was banned in the 1990s amid concerns about cross-resistance with human vancomycin-resistant enterococci [bib_ref] vanA-mediated high-level glycopeptide resistance in Enterococcus faecium from animal husbandry, Klare [/bib_ref].
There is some evidence that controlled exposure of bacterial pathogens to essential oils can alter antibiotic susceptibilities, but field studies are lacking [bib_ref] Evaluation of bacterial resistance to essential oils and antibiotics after exposure to..., Becerril [/bib_ref] [bib_ref] Natural extracts stimulate membrane-associated mechanisms of resistance in Gram-negative bacteria, Fadli [/bib_ref] [bib_ref] Habituation to sub-lethal concentrations of tea tree oil (Melaleuca alternifolia) is associated..., Mcmahon [/bib_ref]. Compared with heavy metals, phytochemical and organic acid feed additives are short-lived following ingestion and, given their limited application currently, their significance for microbial resistance on farms is likely to be minimal.
## Observations on specific livestock-associated bacterial groups
## Escherichia coli
Field surveys, although limited in scope, have not revealed strong evidence for links between biocide and antibiotic susceptibilities among E. coli. Avian-pathogenic (APEC) strains from egg laying flocks were examined for MIC and MBC of five biocide components: two aldehydes, a QAC and hydrogen peroxide [bib_ref] Susceptibility of Avian Pathogenic Escherichia coli from laying hens in Belgium to..., Oosterik [/bib_ref]. Despite prevalent resistance to several antibiotics, no subset of strains with reduced biocide susceptibility was evident in this study. McMurray et al. [bib_ref] Overexpression of marA, soxS, or acrAB produces resistance to triclosan in laboratory..., Mcmurry [/bib_ref] reported that examination of 31 antibiotic multi-resistant clinical strains did not reveal common overexpression of AcrAB-TolC. By contrast, 570 isolates from meat in the USA where QAC are commonly used as sanitisers in the food chain, frequently demonstrated elevated MIC values for the four tested QAC [bib_ref] Presence of disinfectant resistance genes in Escherichia coli isolated from retail meats..., Zou [/bib_ref]. There were also significant associations reported in this study between the presence of some plasmid-borne QAC resistance genes [sugE(p), qacEΔ1] and phenotypic antibiotic resistance, consistent with co-resistance. However, the specificity of these associations for QAC-exposed E. coli was not tested against control strains.
Field strains resistant to ciprofloxacin, including those with upregulated efflux (acrAB) or global responses (marAB, soxRS) showed no difference compared with ciprofloxacin-sensitive strains in susceptibility to three commercial compound disinfectants in MIC and competition assays, nor conversely did disinfectant-passaged strains show a significantly increased frequency of mutation to reduced sensitivity on 4x MIC ciprofloxacin media [bib_ref] Farm disinfectants select for cyclohexane resistance, a marker of multiple antibiotic resistance,..., Randall [/bib_ref]. However, the authors reported a decrease in cyclohexane sensitivity, consistent with the MAR phenomenon, among a minority of disinfectant-passaged strains (particularly a phenolic disinfectant) and among these, ciprofloxacin, ampicillin and tetracycline sensitivity typically was reduced. Regardless of changes to cyclohexane or antibiotic sensitivity, disinfectant exposure did not significantly alter the studied strains' MIC values in respect of the disinfectants themselves.
Other laboratory investigations have also emphasised the variable nature of biocide adaptation and of allied effects in respect of antibiotics. Field strains of verocytotoxigenic E. coli could be trained to increase MIC values for single-agent biocides (triclosan and benzalkonium chloride) but no stable MIC changes could be achieved for chlorhexidine or commercial biocide mixes, and no increase in antibiotic sensitivities were seen [bib_ref] The potential for biocide tolerance in Escherichia coli and its impact on..., Sheridan [/bib_ref]. Adaptation of a laboratory strain to sub-lethal concentrations of food-grade biocides was associated with marked variability, according to biocide, in the degree and stability of reduced biocide susceptibility, changes to biofilm formation, cell membrane hydrophobicity and antibiotic sensitivities [bib_ref] Exposure of Escherichia coli ATCC 12806 to sublethal concentrations of food-grade biocides..., Capita [/bib_ref]. The induction of reduced triclosan susceptibility, and associated increases in antibiotic, chlorhexidine and benzalkonium chloride MIC, occurred to differing degrees and at different rates depending on serogroup [bib_ref] Low level of cross-resistance between triclosan and antibiotics in Escherichia coli K-12..., Braoudaki [/bib_ref]. A marked reduction in triclosan susceptibility induced in an O157:H19 strain was associated with mixed effects, including growth inhibition, increased biofilm formation and a non-constitutive triclosan-inducible change in outer membrane proteins [bib_ref] Proteomic and phenotypic analysis of triclosan tolerant verocytotoxigenic Escherichia coli O157:H19, Sheridan [/bib_ref]. This last study also found that efflux inhibition profoundly increased the sensitivity to triclosan, but only at a high dose of the inhibitor (PAβN), when other toxic effects may have influenced the outcome.
Against a background of prevalent multi-drug antibiotic resistance and moderately low copper sensitivity, administration of copper to weaner pigs in feed at a growth-promotion concentration was associated with significantly increasing susceptibility to most tested antibiotics among faecal E. coli [bib_ref] Effects of chlortetracycline and copper supplementation on antimicrobial resistance of fecal Escherichia..., Agga [/bib_ref]. The study also found that copper susceptibility was not decreased, and neither was the frequency of the plasmid-borne copper resistance gene pcoD increased. It is possible that baseline levels of antibiotic and copper resistance were sufficiently high in this study population that the copper supplementation was insufficient to select for reduced copper susceptibility or associated antibiotic resistance genes. In a prospective study of faecal E. coli among 180 weaners, in-feed copper supplementation at a growth-promoting concentration was associated with variable effects on MIC values of antibiotics over time, with the only significant change compared with control being a reduction of MIC for neomycin and chlortetracycline after six weeks [bib_ref] Effects of copper sulfate, zinc oxide, and neoterramycin on weanling pig growth..., Shelton [/bib_ref] ; No significant effects were observed for high concentrations (2000-3000 ppm) of added zinc.
Multivariable modelling of observed copper and zinc concentrations in pig slurry and antibiotic MIC values suggested positive associations between MIC of a minority of antibiotics and one or other of the metals [bib_ref] Heavy metals in liquid pig manure in light of bacterial antimicrobial resistance, Hölzel [/bib_ref]. Negative associations with antibiotic MIC were seen in respect of lead and mercury. Linkage between pcoD and the tetracycline resistance determinant tetB, but not tetA, was observed by Agga et al [bib_ref] Effects of chlortetracycline and copper supplementation on antimicrobial resistance of fecal Escherichia..., Agga [/bib_ref] , emphasising the often arbitrary nature of co-resistance links. A possible effect on the mobility of antibiotic resistance genes in E. coli was observed by Bednorz et al. [bib_ref] The broader context of antibiotic resistance: Zinc feed supplementation of piglets increases..., Bednorz [/bib_ref] , who reported an increased diversity of genotypes and plasmid profiles, and increased multi-drug antibiotic resistance, among pigs supplemented with high levels (2500 mg/kg feed) of zinc.
## Salmonella enterica
Among Danish broiler house isolates from the decade up to 2001, MIC values were elevated for certain serovar-biocide combinations but there was no relationship to likely previous field exposure nor was there any temporal change in MIC among strains persistent in an individual broiler house [bib_ref] Possible associations between Salmonella persistence in poultry houses and resistance to commonly..., Gradel [/bib_ref]. In this study the MAR phenotype was seen very rarely among those strains with elevated MIC, and triclosan-selected variants showing MAR-type reduced sensitivity to cyclohexane did not show elevated MIC to the original tested biocide products. Over 500 isolates from Danish pig slaughterhouses showed MIC values well below in-use concentrations of a QAC, an acid/hydrogen peroxide mix and triclosan, and post-cleaning and disinfection samples recovered Salmonella only from drains [bib_ref] Biocide and antibiotic susceptibility of Salmonella isolates obtained before and after cleaning..., Gantzhorn [/bib_ref]. There was little evidence of associations between increased biocide MIC and antibiotic resistance. Field strains from pigs and poultry in Thailand commonly exhibited the MAR phenotype, but this was not associated with reduced susceptibility to benzalkonium chloride or chlorhexidine among 257 strains [bib_ref] Susceptibilities to antimicrobials and disinfectants in Salmonella isolates obtained from poultry and..., Chuanchuen [/bib_ref]. Møretrø et al. [bib_ref] Control of Salmonella in food related environments by chemical disinfection, Møretrø [/bib_ref] have noted that there is little evidence of resistance by field salmonellae to working concentrations of disinfectants.
Multiple antibiotic resistance was significantly more common among a subset (fewer than 4%) of 428 livestock and human isolates showing modest reductions in triclosan susceptibility [bib_ref] Prevalence of decreased susceptibility to triclosan in Salmonella enterica isolates from animals..., Copitch [/bib_ref]. All of the subset in this report showed elevated expression of acrB and most also had increased efflux of a fluorescent dye, but mutations in fabI were not seen. Similarly, Randall et al. [bib_ref] Association between cyclohexane resistance in Salmonella of different serovars and increased resistance..., Randall [/bib_ref] reported associated lower susceptibilities to AcrAB-TolC efflux substrates (cyclohexane, ampicillin, tetracycline, ciprofloxacin, a QAC and triclosan), but not to non-transported kanamycin, among isolates from poultry, human and environmental sources.
Several laboratory studies have also supported a substantial role for efflux-associated reduced biocide susceptibility and cross-resistance. Serovars Enteritidis, Typhimurium and Virchow with stable laboratory-induced adaptation (2× to 64× the original MIC) to triclosan or benzalkonium chloride reverted to susceptibility when treated with an efflux pump inhibitor [bib_ref] Mechanisms of resistance in Salmonella enterica adapted to erythromycin, benzalkonium chloride and..., Braoudaki [/bib_ref]. However, concurrent changes in cell surface charge and hydrophobicity among the mutants suggest other enabling mechanisms were also present.
For Salmonella Typhimurium, adaptation to benzalkonium chloride was associated with constitutive overexpression of AcrAB-TolC and, when that was experimentally inactivated by mutation and biocide training repeated, with similar expression of AcrEF-TolC [bib_ref] Resistant mechanism study of benzalkonium chloride selected Salmonella Typhimurium mutants, Guo [/bib_ref]. When the same serovar was subjected to severe biocide challenges (five hours at in-use concentrations) survivors showed consistent reduced growth rates and low-level decreased sensitivity to ciprofloxacin, tetracycline and chloramphenicol, although not to the challenge biocide [bib_ref] Exposure of Salmonella enterica serovar Typhimurium to high level biocide challenge can..., Whitehead [/bib_ref]. The authors reported a MAR phenotype with decreased OmpF and increased AcrEF-TolC-mediated efflux. By contrast, survivors of moderate (five hours in broth at 2× to 4× MIC) challenges showed the same or reduced growth rates as the initial inoculum, and the same survival percentage when re-challenged, suggesting that adaptation under these conditions was largely a reversible physiological process.
Other investigations have emphasised the variability in findings, between strains, biocides and experimental methods. Karatzas et al. [bib_ref] Prolonged treatment of Salmonella enterica serovar Typhimurium with commercial disinfectants selects for..., Karatzas [/bib_ref] reported that sequential daily exposure of S. Typhimurium to sub-inhibitory concentrations of commercial compound biocides did not yield elevations in MIC to the biocides but, depending on the biocide, stable elevations in some antibiotic resistances were seen, accompanied in some cases by reductions in MIC to the other disinfectants, in growth parameters, and in invasiveness of a cell monolayer. By contrast, Randall et al. [bib_ref] Commonly used farm disinfectants can select for mutant Salmonella enterica serovar Typhimurium..., Randall [/bib_ref] described single, extended (seven-day) exposure of S. Typhimurium to aldehyde, oxidising, phenolic and QAC-based commercial compound biocides at 2× MIC concentrations, yielding strains with reduced susceptibility at frequencies suggesting single-step mutation events. In this study, loss of susceptibility to the biocides was not associated with decreased growth, or with reduced colonisation and shedding from chicks. Additionally, changes in MIC values, proteome expression and biocide kill time varied by strain and by biocide, with kill time for the phenolic biocide actually being shortened among mutants [bib_ref] Commonly used farm disinfectants can select for mutant Salmonella enterica serovar Typhimurium..., Randall [/bib_ref].
Some of the mutants in the above report by Karatzas et al. [bib_ref] Prolonged treatment of Salmonella enterica serovar Typhimurium with commercial disinfectants selects for..., Karatzas [/bib_ref] showed elevated acrB expression, consistent with the MAR phenotype. Of the tested biocides, oxidising and phenolic/tar acid mixes were associated with less reduction of antibiotic sensitivities than QAC and QAC/aldehyde-based products. Follow-up testing of one representative strain for each of three biocide products (QAC/aldehyde, oxidising, phenolic) revealed outer membrane protein changes, decreased motility and invasiveness for all, whereas expression of AcrAB-TolC and changes in lipopolysaccharides varied between biocides [bib_ref] Phenotypic and proteomic characterization of multiply antibiotic-resistant variants of Salmonella enterica serovar..., Karatzas [/bib_ref]. In the contrasting extended single exposure study [bib_ref] Commonly used farm disinfectants can select for mutant Salmonella enterica serovar Typhimurium..., Randall [/bib_ref] , the MAR phenotype was not observed, but native efflux capability in the form of a functional AcrAB-TolC pump was required for strains with reduced susceptibility to any tested biocide except for the oxidising one. As a final comparison, small-scale in vivo competition and transmission trials were performed using oral dosing of chicks with wild-type S. Typhimurium and derivatives with the MAR phenotype induced after exposure to oxidising or aldehyde biocides [bib_ref] Fitness and dissemination of disinfectant-selected multiple-antibiotic-resistant (MAR) strains of Salmonella enterica serovar..., Randall [/bib_ref]. These showed the MAR mutants to have a fitness disadvantage compared with the parent strain, and to not be preferentially selected by fluoroquinolone treatment.
The variable and sometimes counter-selective effects of biocide exposure were also illustrated by Birošová and Mikulášová [bib_ref] Development of triclosan and antibiotic resistance in Salmonella enterica serovar Typhimurium, Birošová [/bib_ref]. They reported that S. Typhimurium, when selected on solid media at between 0.125× and 2× MIC concentration for reduced triclosan susceptibility, showed variable phenotypes in respect of ciprofloxacin, chloramphenicol and tetracycline resistance, depending to some extent on the intensity and frequency of the selection steps. Furthermore, the same study also reported that ciprofloxacin-resistant strains with the MAR phenotype showed no decrease in triclosan susceptibility (and sometimes a decrease in ciprofloxacin resistance) when passaged on 0.5× MIC triclosan media, by contrast to non-MAR equivalent strains where ciprofloxacin resistance was maintained and triclosan susceptibility decreased.
In tests comparing Salmonella cells in differing states, six of seven commercial compound biocides at half their in-use concentrations and against planktonic cells were fully lethal for members of a collection of 189 Salmonella strains, whilst the remaining biocide was inconsistently lethal for one strain in the collection [bib_ref] Efficacy of biocides used in the modern food industry to control Salmonella..., Condell [/bib_ref]. Surface-dried cells were less susceptible but generally were still effectively killed at in-use concentrations, whereas cells in biofilm were resistant to killing by several of the products even at full concentration and with extended contact times. Møretrø et al. [bib_ref] Control of Salmonella in food related environments by chemical disinfection, Møretrø [/bib_ref] has noted that Salmonella in biofilms are unpredictably insusceptible to biocides. Attempts by Condell et al. [bib_ref] Efficacy of biocides used in the modern food industry to control Salmonella..., Condell [/bib_ref] to adapt strains to the compound products were not successful, but stable increases of up to 4× MIC were seen when using individual components (triclosan, benzalkonium chloride, chlorhexidine); these strains did not show reduced susceptibility to the other agents, but did in some cases show decreased sensitivity to some antibiotics.
Carcass and food decontaminant biocides have been examined specifically in relation to Salmonella. Exposure of pan-susceptible and multi-drug resistant serovars Typhimurium and Newport (plus E. coli O157:H7) to biocides under recommended conditions after being inoculated onto meat showed variations in kill rate that correlated strongly with the particular products and weakly with bacterial species and serovar, but not with the antibiotic resistance phenotype of the strains [bib_ref] Efficacy of chemical interventions against Escherichia coli O157:H7 and multidrug-resistant and antibiotic-susceptible..., Geornaras [/bib_ref]. Examining the lethality (decimal reduction times) for various serovars of three biocides, Capita [bib_ref] Variation in Salmonella resistance to poultry chemical decontaminants, based on serotype, phage..., Capita [/bib_ref] reported significant positive associations between antibiotic resistance and resistance to killing by acidified sodium chlorite or trisodium phosphate. Such associations were present in only some of the serovars examined, and no association could be discerned in the case of citric acid. In another study, training multi-drug resistant strains of diverse serovars with trisodium phosphate, sodium nitrate and sodium hypochlorite resulted in modest elevations in MIC values to the training agent [bib_ref] Effect of sub-lethal concentrations of biocides on the susceptibility to antibiotics of..., Molina-González [/bib_ref]. There were also variable effects on antibiotic sensitivity, exceeding clinical resistance breakpoints for a minority (around 10%) of strain-antibiotic combinations (including aminoglycosides, by contrast to efflux mutants discussed previously) and in variable patterns, suggesting that mechanisms were non-specific in respect of antibiotic structures or targets.
Examining heavy metal susceptibility, Medardus et al. [bib_ref] In-feed use of heavy metal micronutrients in US swine production systems and..., Medardus [/bib_ref] reported that among 349 Salmonella isolates from nine pig units in the USA, an elevated zinc MIC was associated with the occurrence of the czcD-encoded zinc efflux pump but not with the concentration of faecal zinc. By contrast faecal copper concentration was associated with an elevated copper MIC, but not with the occurrence of a copper efflux gene pcoA. The same study reported that serovars were associated both with copper and zinc susceptibility and with patterns of antibiotic resistance; such resistances were not independently associated with copper or zinc susceptibility once serovar was accounted for.
A Portuguese study of monophasic S. Typhimurium variants, mostly of human and pig origin, found that the sil operon, encoding an efflux mechanism for copper and silver, was common among the principal clonal groups, and it exhibited an associated phenotypic resistance to the metals, which was restricted to anaerobic conditions in the case of copper [bib_ref] Metal tolerance in emerging clinically relevant multidrug-resistant Salmonella enterica serotype 4, Mourão [/bib_ref]. Antibiotic resistance genes in this commonly multi-drug-resistant pathovar were co-located with sil, on the chromosome (European clonal group) or a non-transferable plasmid (Spanish and Southern European clonal groups). The authors found that in plasmid-borne cases, many antibiotic resistance genes were located within class I integrons.
## Livestock-associated meticillin-resistant staphylococcus aureus (la-mrsa)
S. aureus is an upper respiratory tract coloniser that is distributed very widely among mammals but which generally shows host-species-specific subtypes [bib_ref] Emergence of methicillin-resistant Staphylococcus aureus (MRSA) in different animal species, Cuny [/bib_ref]. Isolates of clonal complex 398 (CC398) are closely associated with pigs and some other livestock in a number of countries [bib_ref] Spa type distribution in Staphylococcus aureus originating from pigs, cattle and poultry, Hasman [/bib_ref] , and within the last decade CC398 strains carrying the SCCmec genomic island characteristic of MRSA have come to prominence as persistent colonisers of pig herds [bib_ref] Emergence of methicillin-resistant Staphylococcus aureus (MRSA) in different animal species, Cuny [/bib_ref]. These LA-MRSA strains can colonise humans in contact with animals but, unlike community-and healthcare-acquired MRSA, they do not readily spread between people.
A specific association between reduced zinc sensitivity and CC398 MRSA [bib_ref] Decreased susceptibility to zinc chloride is associated with methicillin resistant Staphylococcus aureus..., Aarestrup [/bib_ref] was shown to be a consequence of the frequent presence of czrC, encoding a putative metal transporter, in SCCmec (type V) of both pig and human isolates [bib_ref] Cloning and occurrence of czrC, a gene conferring cadmium and zinc resistance..., Cavaco [/bib_ref]. Among weaner pigs, administration of pharmacological in-feed doses of either zinc or tetracycline increased counts on nasal swabs of naturally-occurring czrC-positive, tetracycline-resistant MRSA [bib_ref] Author's response: Critique of paper on "Effects of tetracycline and zinc on..., Moodley [/bib_ref] [bib_ref] Effects of tetracycline and zinc on selection of methicillin-resistant Staphylococcus aureus (MRSA)..., Moodley [/bib_ref]. A larger and slightly longer-term study following weaners with pharmacological vs. basal in-feed zinc concentrations on a unit with endemic czrC-positive MRSA, found a significant association between the prevalence of MRSA-positive animals and zinc level at four and five weeks of age, but in both groups MRSA-positive animals were similarly infrequent by seven weeks of age [bib_ref] Zinc oxide therapy increases prevalence and persistence of methicillin-resistant Staphylococcus aureus in..., Slifierz [/bib_ref]. A further report by the same group [bib_ref] Methicillin-resistant Staphylococcus aureus in commercial swine herds is associated with disinfectant and..., Slifierz [/bib_ref] described significant and independent associations between zinc oxide supplementation or frequent disinfection of pens and nasal MRSA in nursery age pigs, on 26 units. Carriage of qac genes was universal among MRSA, and reduced susceptibility to benzalkonium chloride was observed amongst a minority of strains. In the same study, czrC was detected in about two-thirds of tested isolates, in association with a lower susceptibility to zinc compared with czrC-negative isolates.
Other chromosomal resistance determinants for heavy metals (mercury and cadmium) have been associated with meticillin resistance in human MRSA strains [bib_ref] Pattern of antibiotic and heavy-metal ion resistance in recent hospital isolates of..., Millar [/bib_ref] [bib_ref] Genetic characterisation of resistance to metal ions in methicillin-resistant Staphylococcus aureus: Elimination..., Poston [/bib_ref] [bib_ref] Genetics of drug resistance in methicillin-resistant Staphylococcus aureus from Australian hospitals, Townsend [/bib_ref]. Moreover, clusters of heavy metal and antibiotic resistance genes have also been found to be plasmid-borne among MRSA strains from humans (ST398, type V SCCmec with czrC) and pigs (ST398, type IV SCCmec without czrC) [bib_ref] Novel erm(T)-carrying multiresistance plasmids from porcine and human isolates of methicillin-resistant Staphylococcus..., Gómez-Sanz [/bib_ref]. Genes associated with resistance to macrolides, lincosamides, streptogramin B and tetracyclines [erm(T), tet(L)] plus, variably, aminoglycosides and trimethoprim (aadD and dfrK) were co-located with copA, cadDX and mco, encoding copper and cadmium transporters and a copper oxidase. The authors of this study also reported transmissible phenotypic co-resistance to the relevant agents.
## Listeria spp.
Resistance to antibiotics, other than intrinsic resistance to many cephalosporins, is uncommon in Listeria monocytogenes, the principal species of public health concern [bib_ref] The sub-species characterization and antimicrobial resistance of Listeria monocytogenes isolated from domestic..., Bae [/bib_ref]. There were no unusual antibiotic resistances identified among 29 strains from a meat processing plant, three of which showed long-term persistence and reduced susceptibility to benzalkonium chloride [bib_ref] Antibiotic susceptibility in benzalkonium chloride-resistant and -susceptible Listeria monocytogenes strains, Ortiz [/bib_ref]. Similarly, antibiotic susceptibility was high among 114 food isolates, 8% of which showed modestly elevated MIC values for benzalkonium chloride [bib_ref] Antimicrobial susceptibility of Listeria monocytogenes from food products, Aarestrup [/bib_ref]. Reduced susceptibility to benzalkonium chloride in field strains has been shown to be plasmid-borne and not linked to antibiotic resistance in one study [bib_ref] Listeria monocytogenes strains selected on ciprofloxacin or the disinfectant benzalkonium chloride exhibit..., Rakic-Martinez [/bib_ref] but associated with efflux and small changes in susceptibility to kanamycin in another [bib_ref] Role of efflux pumps in adaptation and resistance of Listeria monocytogenes to..., Romanova [/bib_ref]. In both these studies, efflux-related reduced susceptibility to this QAC plus, variously, ciprofloxacin or kanamycin has been observed in strains subjected to selection with one or other of these agents. Analysis of gene expression in these studies suggested that the efflux system(s) induced by an antimicrobial agent may differ between strains, thus producing differing patterns of co-selection.
Laboratory exposure of L. monocytogenes to sub-lethal concentrations of triclosan produced aminoglycoside-resistant mutants that showed no change in triclosan MIC but which exhibited, variously, other phenotypic changes including small colony morphology and increased oxygen sensitivity [bib_ref] Triclosan-induced aminoglycoside-tolerant Listeria monocytogenes isolates can appear as small-colony variants, Kastbjerg [/bib_ref] [bib_ref] Sublethal triclosan exposure decreases susceptibility to gentamicin and other aminoglycosides in Listeria..., Christensen [/bib_ref]. The same effect was not seen in these studies when QAC or oxidising compounds were substituted for triclosan, but decreased sensitivity to aminoglycosides has been reported elsewhere, alongside reduced susceptibility to benzalkonium chloride following exposure to this agent [bib_ref] Role of efflux pumps in adaptation and resistance of Listeria monocytogenes to..., Romanova [/bib_ref]. The recent publication of the genomes of five L. monocytogenes strains with reduced susceptibility to benzalkonium chloride [bib_ref] Genome sequences of five disinfectant-resistant Listeria monocytogenes strains from two Iberian pork-processing..., López-Alonso [/bib_ref] may assist elucidation of the mechanisms involved in co-selection.
## Campylobacter spp.
A survey of 42 isolates from human, animal and environmental sources did not show any associations between the (varied) MIC values to single biocides and sensitivities to erythromycin and ciprofloxacin [bib_ref] The prevalence of antibiotic and biocide resistance among Campylobacter coli and Campylobacter..., Mavri [/bib_ref]. In another survey, of 443 field strains, modest increases in multiple antibiotic resistances correlated with lowered susceptibilities to efflux substrates (ethidium bromide, acridine orange, triclosan) in a small minority of isolates [bib_ref] Prevalence of multiple antibiotic resistance in 443 Campylobacter spp. isolated from humans..., Randall [/bib_ref].
A laboratory study training five isolates with five biocides (QACs, chlorhexidine, triclosan and trisodium phosphate) yielded modest, reversible increases in MIC values in 20% of cases, plus reduced MIC in others [bib_ref] Development of antimicrobial resistance in Campylobacter jejuni and Campylobacter coli adapted to..., Mavri [/bib_ref]. The study also reported variable allied decreases in susceptibility to the other tested biocides and antibiotics; investigations suggested that alterations in cell envelope plus efflux mechanisms were involved.
## Enterococcus spp.
Enterococcus faecium and Enterococcus faecalis are leading causes of hospital-acquired infections, commonly with multiple drug resistances [bib_ref] Enterococcus research: Recent developments and clinical challenges, Sundsfjord [/bib_ref]. Schwaiger et al. [bib_ref] Insusceptibility to disinfectants in bacteria from animals, food and humans-Is there a..., Schwaiger [/bib_ref] examined both these species from human bacteraemias and faeces plus the latter species additionally from a variety of livestock, meat and dairy sources, in respect of susceptibility to a QAC (DDAC) and to 22 antibiotics. Most isolates with reduced susceptibility to DDAC (compared with wild-type) were from dairy sources, but resistance to "in use" concentrations were not observed in field isolates or laboratory-trained strains. Associations between reduced DDAC susceptibility and resistance to a few antibiotics were observed among human faecal E. faecium strains only.
Co-transmission of the copper efflux-associated tcrB and erythromycin resistance erm(B) genes has been reported between a sediment-derived livestock species (Enterococcus hirae) and E. faecalis in conjugation experiments [bib_ref] Erythromycin-and copper-resistant Enterococcus hirae from marine sediment and co-transfer of erm(B) and..., Pasquaroli [/bib_ref]. From 569 Danish livestock-derived bacterial isolates tested for susceptibilities to antibiotics, biocides and heavy metals, the only distinct "less susceptible" group of organisms seen were Enterococcus spp. in respect of copper sulphate [bib_ref] Susceptibility of different bacterial species isolated from food animals to copper sulphate,..., Aarestrup [/bib_ref]. Furthermore, among pig-derived E. faecium and E. faecalis, reduced susceptibility to copper and a number of antibiotics was correlated with the intensity of copper supplementation and of antibiotic use in the source countries, and tcrB was specifically present in the copper-insusceptible isolates [bib_ref] Antimicrobial resistance among enterococci from pigs in three European countries, Aarestrup [/bib_ref]. However, as antibiotic and copper use were themselves closely correlated, the extent of actual co-selection could not be determined. Genomic analysis of E. faecalis from copper-supplemented Danish pigs revealed the presence of a chromosomal cluster of copper-insusceptibility genes, including the tcrYAZB operon, in three of six isolates [bib_ref] Genome sequences of copper resistant and sensitive Enterococcus faecalis strains isolated from..., Zhang [/bib_ref]. Tetracycline-and vancomycin-resistance genes (tetM, vanA) were present in all and one of these three "copper-insusceptible" isolates, respectively, whilst only tetM was encoded among the "copper-susceptible" strains, in one of the three isolates. The frequency of individual antibiotic resistances among Enterococcus spp. from Australian pigs reflected prevalent antibiotic use within the industry, whilst reduced susceptibility to zinc (but not to copper) was common among E. faecalis and resistance to vancomycin was present in commensal non-pathogenic species [bib_ref] Antimicrobial and heavy metal resistance in commensal enterococci isolated from pigs, Fard [/bib_ref].
Among E. faecium from Danish farms, tcrB was more frequently detected from the more intensively copper-supplemented livestock species, whilst linkage with erythromycin and vancomycin resistance was seen only among (heavily copper-exposed) pig isolates [bib_ref] tcrB a gene conferring transferable copper resistance in Enterococcus faecium: Occurrence, transferability,..., Hasman [/bib_ref]. Copper fed to feedlot cattle at a growth promotion concentration (10× basal requirement) was associated with modestly but significantly increased frequencies of detection of tcrB-positive and macrolide-resistant erm(B) enterococci with lowered copper susceptibilities, whilst resistances to other screened antibiotics, including vancomycin, were unaffected [bib_ref] Effects of in-feed copper and tylosin supplementations on copper and antimicrobial resistance..., Amachawadi [/bib_ref]. A prospective study of weaner and grower pigs given a heterogeneous inoculum of tcrB-positive and -negative E. faecium reported that exposure at a commercial (versus basal) in-feed concentration of copper was associated with a higher detection frequency of tcrB and of the linked erm(B) and vanA genes [bib_ref] Copper resistance in Enterococcus faecium, mediated by the tcrB gene, is selected..., Hasman [/bib_ref].
## Summary of data on specific bacterial groups
In general, field surveys have provided weak evidence at best of causal associations between reduced susceptibilities to biocides, heavy metals and antibiotics in these bacterial groups, although data in this area is far from comprehensive. There is experimental evidence that multi-drug efflux is necessary, if not sufficient, in many cases of reduced antibiotic susceptibility associated with exposure to biocides. Furthermore, such efflux-dependence of modestly elevated antibiotic MIC values has also been seen in some field survey isolates although it is not possible to establish clear links with biocide use in these cases. Selection for cross-resistant strains, with perhaps efflux and envelope adaptations, appears plausible for Enterobacteriaceae, at least, under field conditions given the survival of cells even under severe challenge in the laboratory. Biofilms have the potential to enhance this substantially.
One overarching theme is the marked variability in results seen between experimental studies, even examining highly similar organisms and biocidal agents. This suggests that several mechanisms may be involved, to a greater or lesser extent, when co-selection occurs. Furthermore, what mix of susceptibilities ultimately emerges is probably rather sensitive to initial conditions of antimicrobial agent, species, strain and exposure.
For heavy metals, associations have been identified between reduced zinc or copper susceptibility, serovar and antibiotic resistance among pig Salmonella isolates, including monophasic Salmonella Typhimurium. These are established foodborne pathogens. Furthermore, there is reasonable evidence of a co-resistance phenomenon involving copper, macrolides and perhaps vancomycin among enterococci of pigs. However, the relevance of this to disease-causing strains in humans remains undetermined.
LA-MRSA shows some evidence of co-resistance consequent on selection by zinc, but this selective pressure appears to be modest in the face of age-related reductions in prevalence of the organism among pigs. How the effect of zinc compares with selection by antibiotics of this multi-resistant organism is unclear, given that LA-MRSA ST398 isolates are frequently resistant to the commonly-used tetracyclines. There is some evidence that QAC biocides might also select for MRSA carriage in nursery-age pigs. Whether selection, by zinc or QAC, at this age may affect the risk of LA-MRSA contamination of foodstuffs is currently unclear. By contrast in L. monocytogenes, a species notable for little antibiotic resistance, reduced biocide susceptibility in environmental field strains doesn't appear typically to be associated with changes in antibiotic sensitivity.
## Field observations vs. laboratory findings
There is a considerable amount of experimental evidence, as discussed in Sections 4 and 5, for co-selection of antibiotic resistance by exposure of common primary and opportunistic bacterial pathogens, to biocides principally, and also to heavy metals. However, this is not matched by compelling evidence of similar effects in farm and other environments regularly exposed to such agents. Nonetheless, biocide susceptibility is commonly measured in terms of MIC and several experimental studies have shown that exposure to biocides, particularly compound products, may result in elevated antibiotic MIC (including the MAR phenotype among Salmonella) without significant change in MIC of the biocide. Thus, there may be some effects of biocides on antibiotic sensitivity among field strains where the association is not evident from testing of MIC values. Further testing, of kill times and of effects of biocides against organisms in stressed states, on surfaces, in biofilms, etc., might reveal more associations than are currently evident. Evidence [bib_ref] Metal tolerance in emerging clinically relevant multidrug-resistant Salmonella enterica serotype 4, Mourão [/bib_ref] that the sil operon improves copper handling by monophasic S. Typhimurium only under anaerobic conditions, such as are encountered in the gut, also indicates that test conditions may substantially affect which associations are seen between heavy metals and antibiotics.
Nonetheless, there are some reasons to be sceptical about the occurrence in real-world environments of phenomena described in laboratory studies. Controlled, extended laboratory exposure of bacterial strains to sub-lethal concentrations of antimicrobial substances is not likely to be a good model for interactions between biocides and organisms in the environment. In-use conditions usually involve transient high-level application, often of a synergistic compound preparation, followed by variable exposure of survivors to residual concentrations. Some biocides are persistent, whilst others biodegrade rapidly, and the effectiveness of others is highly-dependent on dilution, pH or interfering organic substances [bib_ref] Biocide use and antibiotic resistance: The relevance of laboratory findings to clinical..., Russell [/bib_ref].
This variation in effective exposures, given the differences seen even between similar laboratory studies of controlled exposure, is likely to yield unpredictable patterns of altered susceptibility that may also include increased sensitivity to some agents. Certainly, some biocides and heavy metals can have a counter-selective effect on antibiotic resistance genes and/or phenotypes [bib_ref] Heavy metal driven co-selection of antibiotic resistance in soil and water bodies..., Seiler [/bib_ref] [bib_ref] Heavy metals in liquid pig manure in light of bacterial antimicrobial resistance, Hölzel [/bib_ref] [bib_ref] Impact of UV and peracetic acid disinfection on the prevalence of virulence..., Biswal [/bib_ref] [bib_ref] Spatial analysis of antibiotic resistance along metal contaminated streams, Tuckfield [/bib_ref]. Furthermore, certain effects observed in the laboratory may not be as adaptive as they initially appear. For example, the induction of efflux by triclosan binding to the repressor of the SmeDEF multi-drug transporter of S. maltophilia only occurs once triclosan is present at toxic levels [bib_ref] Recent advances in the potential interconnection between antimicrobial resistance to biocides and..., Oggioni [/bib_ref] [bib_ref] The binding of triclosan to SmeT, the repressor of the multidrug efflux..., Hernández [/bib_ref].
Even once significantly reduced susceptibilities to biocide, metal or antibiotic have been generated, the fitness costs discussed previously will militate against survival of many adapted strains in the context of shifting and re-establishing bacterial communities such as is likely in farm environments. Nonetheless some adaptations, such as biofilming capacity, may prove immediately advantageous for survival and there are agents (particularly heavy metals) and circumstances, such as prevalent antibiotic use, where a selective pressure may be maintained for a strain following adaptation for reduced susceptibility. However, the particular circumstances where an appropriate selective pressure is applied at the right time to maintain or intensify a particular reduction in antibiotic sensitivity that has been generated by biocide use may not be especially common. Indeed, instead of co-selection, biocides and antibiotics can have additive or synergistic antimicrobial effects when applied together [bib_ref] Effect of surfactants on antibiotic resistance, Suling [/bib_ref] [bib_ref] Influence of the treatment of Listeria monocytogenes and Salmonella enterica serovar Typhimurium..., Zanini [/bib_ref].
In cases of heavy metal contamination, the antimicrobial selective pressures are likely to be sub-lethal and exerted for a long time, extending to many years in the case of contaminated soil or sediment. It is of interest that investigations into heavily copper-supplemented pigs and heavily contaminated water or soil have provided some of the most convincing data on associations between reduced susceptibility to antibiotic and non-antibiotic (i.e., metal) antimicrobials, in a wide range of bacteria.
# Conclusions
The long history of biocide use in healthcare, agriculture, industry and elsewhere has not yielded any consistent evidence of a generalised problem of reduced susceptibility to such agents. Furthermore, the comparatively brief history of antibiotic use and the rapid emergence of single and multiple antibiotic resistance among exposed bacteria has provided compelling evidence of a causal association between antibiotic consumption and the occurrence of antibiotic resistance. This drives efforts to control medical and veterinary use of antibiotics. Increasing hygiene measures and, in stock rearing, substitution of prophylactic and growth-promoting antibiotics with heavy metal salts has been part of the response to greater constraints on antibiotic use.
Concerns have been raised in recent years that known and suspected links between biocide or heavy metal use and antibiotic resistances may prove significant in perpetuating those resistances. The legitimacy of these concerns has been bolstered by evidence from antibiotic and biocide research that co-selection may yield multiple resistances and that even low-level reduced antibiotic sensitivity can foster the development of high-level resistance given a suitable selective pressure. The heavy metals are one group of agents that, by deliberate application or by contamination, are typically present at non-lethal concentrations and for prolonged periods compared with biocides. In this respect their interactions with bacteria more closely resemble those of antibiotics than of disinfectants or antiseptics.
There is much evidence from laboratory studies that exposure of bacteria to members of one class of antimicrobial agent can affect susceptibilities to other classes. However, data from hospital and other healthcare environments has provided conflicting evidence as to the likely clinical significance of co-selection of antibiotic resistant pathogens by biocides. Data for similar effects on the diverse microbiota on farms and in the food chain is yet more limited than for healthcare-associated pathogens. Again, laboratory studies indicate that such links exist for the important zoonotic pathogens but survey data suggests that, for now, counter-selective processes largely limit the effects of co-selective pressures. For heavy metals, in addition to identified genetic linkages with antibiotic resistance determinants, evidence exists from animal and environmental studies for links between the use of such agents in the pig and poultry industries and increased genetic mobility and co-selection. However, the immediate significance of susceptibility associations in respect of foodborne pathogens is uncertain.
AcknowledgmentsFunding for the review was provided by the UK Department of Environment, Food and Rural Affairs (Defra).Conflicts of InterestThe authors declare no conflict of interest. |
Lost but not forgotten: A population‐based study of mortality and care trajectories among people living with HIV who are lost to follow‐up in Ontario, Canada
ObjectivesSelection as a consequence of volunteer participation in, and loss to follow-up from, cohort studies may bias estimates of mortality and other health outcomes. To quantify this potential, we estimated mortality and health service use among people living with HIV (PLWH) who were lost to cohort follow-up (LTCFU) from a volunteer clinical HIV-infected cohort, and compared these to mortality and health service use in active cohort participants and non-cohort-participants living with HIV in Ontario, Canada.MethodsWe analysed population-based provincial health databases from 1995 to 2014, identifying PLWH ≥ 18 years old; these included data from participants in the Ontario HIV Treatment Network Cohort Study (OCS), a volunteer, multi-site clinical HIV-infected cohort. We calculated all-cause mortality, hospitalization and emergency department (ED) visit rates per 100 person-years (PY) and estimated hazard ratios (HRs) of mortality, adjusting for age, sex, income, rurality, and immigration status.ResultsAmong 23 043 PLWH, 5568 were OCS participants. Compared with nonparticipants, participants were younger and less likely to be female, to be an immigrant and to reside in a major urban centre, and had lower comorbidity. Mortality among active participants, participants LTCFU and nonparticipants was 2.52, 3.30 and 2.20 per 100 PY, respectively. After adjustment for covariates, mortality risk was elevated among participants LTCFU compared with active participants (HR 2.26; 95% confidence interval 1.91, 2.68). Age-adjusted hospitalization rates and ED visit rates were highest among participants LTCFU.ConclusionsMortality risk and use of health care resources were lower among active cohort participants. Our findings may inform health outcome estimates based on volunteer cohorts, as well as quantitative bias adjustment to correct for such biases. People living with HIV who are lost to follow-up 91 0.57 (0.48, 0.69) Canadian born/long-term resident Ref CI, confidence interval; Ref, reference group to which the remaining groups are compared.
# Introduction
The generalizability of findings from cohort studies may be limited if the cohort population differs from the wider population of interest. Selection bias, a form of participation bias, exists when active members of the cohort have significantly different characteristics than those unable or unwilling to participate, and there is reason to believe that this variation could potentially impact outcome risk. Biases can also occur as a result of participant loss to cohort follow-up (LTCFU). In HIV-infected cohort studies, LTCFU is typically defined as failure to attend HIV care for a prespecified time period. In HIV studies conducted in developed countries, LTCFU occurs at rates from 3.5 to 9.8 per 100 patient-years (PY) [bib_ref] Characteristics of and outcomes in HIV-infected patients who return to care after..., Ndiaye [/bib_ref] [bib_ref] Loss to follow-up in an international, multicentre observational study, Mocroft [/bib_ref] [bib_ref] Loss to follow-up in the Australian HIV Observational Database, Mcmanus [/bib_ref] with cumulative LTCFU rates of 20-40%, defined by studies as a lack of engagement in primary care and out-patient visits for 6 or 12 months without returning, after an initial linkage to HIV care [bib_ref] do patients who initiate care stay in care?, Torian [/bib_ref] [bib_ref] Establishment, retention, and loss to follow-up in outpatient HIV care, Fleishman [/bib_ref] [bib_ref] Discontinuation from HIV medical care: squandering treatment opportunities, Samet [/bib_ref].
From a methodological perspective, knowledge of health outcomes of cohort study participants after LTCFU is necessary for the interpretation of research findings from those studies. Time to event analyses of clinical outcomes or death in cohort studies often assume that censoring as a result of LTCFU is independent of the outcome under study. However, such assumptions would bias life expectancy calculations for people living with HIV (PLWH) if mortality rates among individuals who are lost to cohort follow-up (LTCFU) differ substantially from those of individuals still under follow-up [bib_ref] Correcting mortality for loss to follow-up: a nomogram applied to antiretroviral treatment..., Egger [/bib_ref] [bib_ref] How does loss to follow-up influence cohort findings on HIV infection? A..., Lanoy [/bib_ref]. When the mortality rates after LTCFU are not available, studies have estimated life expectancy using assumptions about mortality after LTCFU derived from cohort studies from other countries, from local cohorts, or from sensitivity analyses with a range of possible outcome rates [bib_ref] Life expectancy of HIV-positive individuals on combination antiretroviral therapy in Canada, Patterson [/bib_ref] [bib_ref] Impact of definitions of loss to follow-up on estimates of retention, disease..., Shepherd [/bib_ref] [bib_ref] Heterogeneity in outcomes of treated HIV-positive patients in Europe and North America:..., May [/bib_ref] [bib_ref] Incorporating loss to follow-up in estimates of survival among HIV-infected individuals in..., Verguet [/bib_ref] ; however, determining the extent to which the results are truly reflective of their local setting can be challenging.
Our objective was to estimate mortality and health service use trajectories among volunteer HIV-infected cohort participants, comparing those actively participating to those LTCFU, as well as to nonparticipants. We addressed this objective using data from Ontario, Canada, where population-based data are available for all PLWH in HIV care, and the setting of the long-standing Ontario HIV Treatment Network (OHTN) Cohort Study (OCS).
# Methods
We carried out a population-based retrospective cohort study using data from the OCS and administrative data housed at ICES.
## Ohtn cohort study
The OCS is a voluntary open cohort study of PLWH who have received care at one of 11 clinical sites since 1995 [bib_ref] Cohort profile: the Ontario HIV treatment network cohort study (OCS), Rourke [/bib_ref]. Participants have completed annual questionnaires since 2007, for which they received a modest honorarium. Clinical and demographic data are abstracted from medical charts. We linked OCS data to administrative databases using unique identifiers. The only OCS data beyond health administrative data that were used for this analysis were dates of death and loss to follow-up.
## Administrative data sources
We conducted analyses using health administrative data from the province of Ontario, Canada. We used the following data sets for this study: the Registered Persons Database (RPDB), an electronic registry of all Ontarians eligible for health coverage, to ascertain patient demographic information and date of deaths; 2006 Statistics Canada census data; the Ontario Health Insurance Plan (OHIP) database, which contains physician billing claims for approximately 95% of physician-based visits in this province; the Discharge Abstract Database, which captures all provincial hospitalization discharge data; the National Ambulatory Care Reporting System and Emergency Claims Database (ERCLAIMS) for information on emergency department and selected out-patient visits; the Immigration, Refugee and Citizenship Canada (IRCC) Permanent Resident Database; and the Office of the Register General -Deaths (ORGD) Database for detailed death records. These data sets were linked using unique encoded identifiers and analysed at ICES.
## Study population
To obtain an administrative cohort of PLWH, we applied a validated algorithm to identify all individuals ≥ 18 years receiving HIV care in Ontario between 1 April 1992 and 31 December 2014 [bib_ref] Validation of case-finding algorithms derived from administrative data for identifying adults living..., Antoniou [/bib_ref]. OCS participants were linked to this administrative cohort. OCS participants who were not linked to ICES were excluded from analyses .
For those identified by the above algorithm as PLWH who were also OCS participants, the index date was defined as the earliest of (1) the date of first HIV billing code that triggered the HIV algorithm or (2) the date of entry into the OCS cohort. For non-OCS participants, the index date was the date of the first HIV billing code between 1 April 1992 and 31 December 2014. Patients were followed until death or 31 December 2014, whichever occurred first. We excluded those who presented to HIV care in 1992 or earlier, as their exact date of presentation was unknown. We applied rules to censor patients with long gaps in care (i.e. a period of ≥ 7 years with no health care encounters but with no record of death). Censoring began at 7 years after the last encounter preceding the gap until the earlier of (1) the end of the study period or (2) re-engagement in care. For example, a patient who had regular health care encounters for a period of 5 years after presentation to HIV care but no encounters thereafter would be censored at 12 years of follow-up. If a similar patient re-engaged with the health care system at 14 years of follow-up, he/she would be censored from years 12 to 14 and then re-included in the analysis from year 14 to the end of the study period.
## Classification of cohort participation status
Our primary exposures of interest were OCS participation and LTCFU among OCS participants. OCS participants were classified as LTCFU if there were no available clinicbased data (i.e. no data from their medical chart or interview data) over an 18-month period. Time after LTCFU from the OCS was measured until the earliest of (1) the end of ICES follow-up as a consequence of moving out of province, (2) death or (3) 31 December 2014.
## Outcomes
Our primary outcome was mortality. We used the RPDB to determine the date of death. For OCS participants without evidence of death in the RPDB, we used date of death reported in OCS data.
To understand the care engagement of active OCS participants compared to those who were LTCFU, our secondary outcome was health services use [emergency department (ED) visits, hospital admissions and engagement in care]. We classified out-patient visits as family physician visits, visits to HIV specialists (defined as any visit to an internal medicine or infectious disease specialist), and visits specifically for HIV care (ICD-9, codes 042, 043 and 044). We defined engagement in HIV care after LTCFU from OCS, defined as at least two out-patient visits for HIV care at least 3 months apart. We reported engagement in HIV care in the first year after follow-up and in the most recent calendar year, 2014.
## Covariates
We adjusted for the following demographic information and covariates at the time of care presentation: age, sex, income, rurality (as determined from neighbourhood-level postal codes linked to 2006 Statistics Canada census data) and immigration status [immigrant from country with generalized HIV epidemic and immigrant from country without generalized HIV epidemic]. We used the John Hopkins ACG â System Version 10 (Sun Microsystems Inc., Santa Clara, CA)to categorize medical comorbidity by assigning individuals to one of 32 distinct aggregated diagnosis groups (ADGs) based on condition duration, severity, diagnostic certainty, aetiology, and specialty care involvement. Comorbidity burden was classified as low (≤ 5 ADGs), medium (6-9 ADGs) and high (≥ 10 ADGs).
## Analyses
We compared characteristics between OCS participants and nonparticipants using means and standard deviations and t-tests for continuous variables, and frequency and percentages and v 2 tests for categorical variables. We calculated all-cause mortality rates, and hospitalization and ED rates (crude and age-adjusted to the 2011 Ontario population) per 100 PY, stratified by non-OCS participants, OCS active participants and OCS participants after LTCFU. For all outcomes, observation time was divided into periods to detect temporal trends, and by sex to detect differences between men and women.
We used proportional hazards models with time to death from date of presentation for HIV care. We left truncated time between care presentation and OCS enrolment for OCS participants and time between care presentation and 1995 for non-OCS participants who presented to care prior to 1995. Participants for whom death was not documented during the study period were censored at the later of the last follow-up date in ICES databases or 31 December 2014. LTCFU for OCS participants was modelled with a binary time-dependent covariate, which was set to zero at OCS enrolment and to 1 after a period of 18 months of having no available clinic-based data (i.e. no data from their medical chart no interview data). We estimated hazard ratios (HRs) of mortality for the following covariates: OCS participation, LTCFU, and calendar time period, as well as the interaction of LTCFU and time period and OCS participation and time period. We adjusted for participant demographics at the time of care presentation: age (per 10 years), sex, immigration status (immigrant from a country with generalized HIV epidemic or without generalized HIV epidemic versus long-term resident), rurality (rural versus urban/suburban location) and neighbourhood income quintile (other quintiles versus highest quintile)). SAS version 9.4 (SAS Institute, Cary, NC) was used for all analyses.
Ethics approval for this project was granted by the Ottawa Hospital Research Ethics Board and the Sunnybrook Health Sciences Centre Research Ethics Board. Ethics approval for the OCS was obtained from each study site. Participants signed informed consent at the time of enrolment into the cohort.
# Results
Between 1 January 1995 and 31 December 2014, 23 043 individuals received HIV care in Ontario. During this time period, 6129 PLWH were enrolled into the OCS; of these, 5619 (92%) were linked to the administrative data holdings at ICES, and 5568 presented to care after 1992, for a total of 37 718.68 PY of followup. Rates of linkage to ICES were 94% for active participants, 96.5% for patients who had died, 98% for participants from the site that had closed and 82% for LTCFU participants. As of 31 December 2014, of the 5568 OCS participants linked to ICES, 3219 (57.8%) were active, 951 (17.0%) had died, 219 (3.9%) were enrolled at a site that had closed, and 1179 (21.2%) were LTCFU. Compared with people who remained active in the cohort, those LTCFU were more likely to have presented to care in earlier years, to have a slightly younger age at care presentation, to be a longterm resident of Canada, and to have lower comorbidity (Appendix 2, [fig_ref] Table 1: Characteristics of Ontario HIV Treatment Network Cohort Study [/fig_ref]. After LTCFU from the OCS, 267 participants were recorded as having died (3.3 deaths per 100 PY), for a total of 1218 deaths.
The remaining 17 475 PLWH not linked to the OCS were classified as nonparticipants. At presentation to care, compared with nonparticipants, participants were significantly younger, were less likely to be female, to be an immigrant and to come from a major urban centre, and had lower comorbidity [fig_ref] Table 1: Characteristics of Ontario HIV Treatment Network Cohort Study [/fig_ref]. As of 31 December 2014, 3797 (21.7%) of the 17 475 nonparticipants had died. Crude rates of death among OCS participants and OCS nonparticipants were 2.52 and 2.20 per 100 PY of follow-up, respectively. [fig_ref] Table 2: Age-adjusted mortality rates per 100 patient-years [/fig_ref] shows the age-adjusted mortality rates by calendar year for nonparticipants, active participants and participants after LTCFU. For all three groups, mortality rates declined by > 70% between 1995 and 2014. In all calendar periods, age-adjusted mortality rates were higher for participants LTCFU than for those who remained [fig_ref] Table 3: Proportional hazards model of mortality among people living with HIV in Ontario [/fig_ref]. Older age at entry into care was associated with a greater hazard of mortality, as was living in lower income neighbourhoods. Female sex, residing in an urban location and having immigrated after 1985 were factors associated with a lower hazard of mortality.
Age-adjusted hospitalization rates decreased over calendar time for all groups. Hospitalization rates were highest among OCS participants who were LTCFU. Compared with men, women had higher ageadjusted hospitalization rates among non-OCS participants (P < 0.0001) but similar rates among OCS active participants and OCS participants LTCFU. Similarly, ageadjusted ED visit rates were highest in OCS participants who were LTCFU across all time periods. Rates of ED visits increased over time for all three cohorts, except for OCS active participants, for whom ED visits declined in the 2010-2014 period. Compared with men, women had more ED visits among all groups.
Most of the 1179 individuals who were LTCFU from the OCS but had no record of death continued to access care through out-patient visits at a median rate of 9.12 visits per year [interquartile range (IQR) 4.2-17.2 visits/year] [fig_ref] Table 5: Out-patient visits among Ontario HIV Treatment Network Cohort Study [/fig_ref]. Half of these visits were to family physicians (median 4.5 visits/year; IQR 1.6-9.2 visits/year) and about a third were for HIV-specific care (median 3.1 visits/year; IQR 1.2-6.2 visits/year). Among participants lost to followup across all time periods, 60% remained engaged in HIV care (at least two out-patient visits for HIV care at least
# Discussion
In our setting in Ontario, Canada, we documented higher rates of mortality and increased use of acute health care resources, including hospitalizations and ED visits, among PLWH who become lost to follow-up from a clinical HIVinfected cohort study. OCS participants were LTCFU at a rate of 3.13 per 100 PY; this group had more than twice the rate of mortality and higher rates of acute care use (hospitalizations and ED visits) than participants who remained active in the cohort or individuals who had never joined the cohort. As the OCS is a volunteer cohort, not a population-based study, LTCFU from the cohort was not synonymous with the lack of engagement in HIV care. In fact, approximately half of LTCFU participants remained engaged in HIV care after LTCFU from the cohort. Our findings provide a robust estimate of mortality among those LTCFU that can enhance assumptions made when estimating life expectancy from HIV-infected cohorts.
Other HIV-infected cohort studies have reported mixed findings on the association between LTCFU and mortality, in part as a consequence of different definitions of LTCFU, the calendar year period under study and the degree to which a cohort study is population-based, such that LTCFU implies a lack of engagement in HIV care. A French HIV-infected cohort study reported a mortality rate of 29.8% within a 1-year period among the 5.4% of participants who were LTCFU for ≥ 12 months and who never returned [bib_ref] How does loss to follow-up influence cohort findings on HIV infection? A..., Lanoy [/bib_ref]. This French cohort study described an analysis which incorporated assumptions for mortality among those LTCFU that resulted in a significantly increased 1-year mortality rate among patients without delayed access to care compared with an analysis performed without correction for LTCFU (1.9% versus 0.6%, respectively; P = 0.0012) [bib_ref] How does loss to follow-up influence cohort findings on HIV infection? A..., Lanoy [/bib_ref]. However, for patients with delayed access to care, the 1-year mortality rates were not significantly different between analyses that did and did not correct for LTCFU (7.8% versus 6.5%, respectively; P = 0.29). A US cohort study reported that participants who were ever LTCFU for 12 months had only modestly elevated mortality (1.2 times higher after adjustment for confounders) compared with participants who remained in care after 5 years, perhaps as a consequence of re-engagement in HIV care outside the cohort [bib_ref] Loss to clinic and five-year mortality among HIV-infected antiretroviral therapy initiators, Edwards [/bib_ref]. Finally, an Australian HIV-infected cohort study reported no effect of a gap in care of 12 months or more on mortality rates during the period 1999-2013, also attributable at least in part to the availability of care outside of the cohort [bib_ref] Loss to follow-up in the Australian HIV Observational Database, Mcmanus [/bib_ref].
Results from our study are in contrast with these previously published findings. We observed a doubling of mortality risk among OCS participants who were LTCFU, despite the fact that 60% had HIV care (at least two HIV care visits, 3 months apart) in the first year after LTCFU. The stronger effect of LTCFU on mortality in our study compared to other studies could be attributable to our more stringent definition of LTCFU of no clinic visits for a period of 18 months, rather than the 12-month period used by other studies. In general, higher mortality among LTCFU participants may be attributable to sicker participants leaving the cohort, as evidenced by higher acute care rates after LTCFU, or the loss of the social support and medication adherence support from cohort participation [bib_ref] Healthcare contacts among patients lost to follow-up in HIV care: review of..., Connors [/bib_ref]. The loss of established HIV primary care may also contribute to revolving through ED visits, which has been shown not to constitute high-quality care [bib_ref] Care fragmentation, quality, and costs among chronically ill patients, Frandsen [/bib_ref] [bib_ref] Essential components of effective HIV care: a policy paper of the HIV..., Gallant [/bib_ref]. In a study of patients attending a regional HIV care site in Alberta, Canada [bib_ref] Healthcare contacts among patients lost to follow-up in HIV care: review of..., Connors [/bib_ref] , 9% of 1928 patients seen between 1 January 2010 and 31 August 2014 were LTCFU for ≥ 12 months. Only 20% of the LTCFU Alberta patients received HIV care elsewhere in the province compared to the 60% of OCS patients in our study who received care in the first year after loss to follow-up. LTCFU patients in Alberta had frequent visits to the ED. The LTCFU rate from the Southern Alberta Clinic between 2001 and 2006 was reported as 12% [bib_ref] Unappreciated epidemiology: the churn effect in a regional HIV care programme, Gill [/bib_ref]. The rate of LTCFU from the Canadian Observational Cohort, a cohort collaboration of clinical sites in British Columbia, Ontario (including OCS) and Quebec, between 2000 and 2015 was 11% [bib_ref] Life expectancy of HIV-positive individuals on combination antiretroviral therapy in Canada, Patterson [/bib_ref].
Strengths of our study include the large sample size, the long duration of the study period and the comprehensiveness of the provincial health administrative databases. Nevertheless, our research is subject to limitations. Because we were unable to distinguish between individuals who were not receiving HIV care and those who had moved out of the province, we only classified people as having out-migrated if they had not received care for 7 years. Inclusion of person-time for out-migrants who were not identified by this 7-year rule would result in an underestimate of mortality rates for LTCFU participants, making our results even more striking. The administrative databases did not contain clinical HIV information (e.g. adherence to antiretroviral therapy, CD4 counts or HIV viral loads); thus, we were unable to compare these markers according to participation status. LTCFU participants were less likely than active participants to have been linked to administrative databases, which is likely to have biased our results towards/away from the null depending on whether sicker or healthier patients were more likely to have agreed to have their records linked to administrative data.
This approach and our findings may inform health outcome estimates based on volunteer cohorts and could be used for quantitative bias adjustment to correct for such biases. A consensus statement [bib_ref] The prevention and treatment of missing data in clinical trials, Little [/bib_ref] regarding preferred methods for presentation of longitudinal studies with missing data [bib_ref] Missing data, Ware [/bib_ref] offers guidelines for the conduct of sensitivity analyses when data are missing not at random (MNAR). Statistical methods for handling missing data as a result of LTCFU include inverse probability weighting for the probability of censoring [bib_ref] Nonignorable loss to follow-up: correcting mortality estimates based on additional outcome ascertainment, Schomaker [/bib_ref] , pattern mixture modelling [bib_ref] Pattern mixture models for quantifying missing data uncertainty in longitudinal invariance testing, Sterba [/bib_ref] and selection models [bib_ref] Missing data: our view of the state of the art, Schafer [/bib_ref]. Interpretation of sensitivity analyses requires a sense of the plausibility of different imputation strategies. Our study provides data on mortality and health care utilization after LTCFU from a cohort study, which will be useful for the conduct and interpretation of sensitivity analyses making assumptions about participants who are LTCFU.
Adam McGee. The OHTN Cohort Study is supported by the Ontario Ministry of Health and Long-Term Care. The opinions, results and conclusions are those of the authors only. No endorsement by the Ontario HIV Treatment Network is intended or should be inferred.
Funding: Canadian Institutes of Health Research Grant FRN-TT5-128270.
Data access: The data set from this study is held securely in coded form at ICES. While data sharing agreements prohibit ICES from making the data set publicly available, access may be granted to those who meet prespecified criteria for confidential access, available at www.ices. on.ca/DAS. The full data set creation plan and underlying analytic code are available from the authors upon request, understanding that the programmes may rely upon coding templates or macros that are unique to ICES. *Suppressed because of small cell sizes. SD, standard deviation.
# Appendix 1:
[table] Table 1: Characteristics of Ontario HIV Treatment Network Cohort Study (OCS) participants and nonparticipants at presentation to HIV care [/table]
[table] Table 2: Age-adjusted mortality rates per 100 patient-years (95% confidence intervals) by calendar year period for non-Ontario HIV Treatment Network Cohort Study (OCS) participants, OCS active participants and OCS participants after being lost to cohort follow-up (LTCFU) [/table]
[table] Table 3: Proportional hazards model of mortality among people living with HIV in Ontario [/table]
[table] Table 4: (a) Crude and age-adjusted hospitalization rates per 100 patient-years (95% confidence intervals) by calendar year period and sex for non-Ontario HIV Treatment Network Cohort Study (OCS) participants, OCS active participants and OCS participants after being lost to cohort follow-up (LTCFU). (b) Crude and age-adjusted emergency department rates per 100 patientyears (95% confidence intervals) by calendar year period and sex for non-OCS participants, OCS active participants and OCS participants after being LTCFU [/table]
[table] Table 5: Out-patient visits among Ontario HIV Treatment Network Cohort Study (OCS) participants who were lost to cohort follow-up (LTCFU) (n = 1179) [/table]
[table] Table A1: Characteristics of Ontario HIV Treatment Network Cohort Study (OCS) participants at presentation to HIV care by loss to cohort follow-up status [/table]
|
Live cell visualization of the interactions between HIV-1 Gag and the cellular RNA-binding protein Staufen1
Background: Human immunodeficiency virus type 1 (HIV-1) uses cellular proteins and machinery to ensure transmission to uninfected cells. Although the host proteins involved in the transport of viral components toward the plasma membrane have been investigated, the dynamics of this process remain incompletely described. Previously we showed that the double-stranded (ds)RNA-binding protein, Staufen1 is found in the HIV-1 ribonucleoprotein (RNP) that contains the HIV-1 genomic RNA (vRNA), Gag and other host RNA-binding proteins in HIV-1-producing cells. Staufen1 interacts with the nucleocapsid domain (NC) domain of Gag and regulates Gag multimerization on membranes thereby modulating HIV-1 assembly. The formation of the HIV-1 RNP is dynamic and likely central to the fate of the vRNA during the late phase of the HIV-1 replication cycle.Results: Detailed molecular imaging of both the intracellular trafficking of virus components and of virus-host protein complexes is critical to enhance our understanding of factors that contribute to HIV-1 pathogenesis. In this work, we visualized the interactions between Gag and host proteins using bimolecular and trimolecular fluorescence complementation (BiFC and TriFC) analyses. These methods allow for the direct visualization of the localization of protein-protein and protein-protein-RNA interactions in live cells. We identified where the virus-host interactions between Gag and Staufen1 and Gag and IMP1 (also known as VICKZ1, IGF2BP1 and ZBP1) occur in cells. These virushost interactions were not only detected in the cytoplasm, but were also found at cholesterol-enriched GM1containing lipid raft plasma membrane domains. Importantly, Gag specifically recruited Staufen1 to the detergent insoluble membranes supporting a key function for this host factor during virus assembly. Notably, the TriFC experiments showed that Gag and Staufen1 actively recruited protein partners when tethered to mRNA.Conclusions:The present work characterizes the interaction sites of key components of the HIV-1 RNP (Gag, Staufen1 and IMP1), thereby bringing to light where HIV-1 recruits and co-opts RNA-binding proteins during virus assembly.
# Background
HIV-1 replication is characterized by multiple virus-host interactions that represent fundamental events enabling viral propagation. While Gag is central to assembly, numerous host proteins are also required for the generation of infectious HIV-1 particles [bib_ref] Host factors exploited by retroviruses, Goff [/bib_ref]. The vRNA can both be translated to produce Gag and Gag-Pol or packaged into virions [bib_ref] HIV-1 gag proteins: diverse functions in the virus life cycle, Freed [/bib_ref]. Gag selects the HIV-1 RNA genome (vRNA) for packaging in the cytoplasm. These events involve the regulated assembly of viral ribonucleoprotein (RNP) complexes. This is a prerequisite for successful retroviral vRNA trafficking from the nucleus into the cytoplasm, through the cytoplasm, and then into progeny virions at sites of assembly [bib_ref] Modulating HIV-1 RNA processing and utilization, Mclaren [/bib_ref] [bib_ref] The retrovirus RNA trafficking granule: from birth to maturity, Cochrane [/bib_ref]. Importantly, recent studies show how vRNA transport mechanisms dictate to what extent both the vRNA is translated and to what efficiency Gag is assembled [bib_ref] Retroviral mRNA nuclear export elements regulate protein function and virion assembly, Swanson [/bib_ref] [bib_ref] Distinct intracellular trafficking of equine infectious anemia virus and human immunodeficiency virus..., Jin [/bib_ref]. Studies also suggest that the host factors that interact with viral Gag and RNA might dictate intracellular trafficking events during viral egress (reviewed in [bib_ref] Retrovirus RNA trafficking: from chromatin to invasive genomes, Swanson [/bib_ref].
Initially Gag is synthesized as a precursor molecule, but is then cleaved to give rise to matrix (MA), capsid (CA), nucleocapsid (NC), a late domain (p6) plus two spacer
# Results
## Bimolecular fluorescence complementation (bifc) to visualize gag-staufen1 interactions in live mammalian cells
Recently, the relationship between Staufen1 and precursor Gag molecule (pr55 Gag ) was characterized. While Staufen1 is found predominantly in the cytoplasm at the endoplasmic reticulum [bib_ref] Mammalian staufen is a double-stranded-RNA-and tubulin-binding protein which localizes to the rough..., Wickham [/bib_ref] ; Gag is localized in a punctate, non-uniform pattern throughout the cytoplasm and is enriched at the plasma membrane [bib_ref] Intracellular transport of human immunodeficiency virus type 1 genomic RNA and viral..., Lehmann [/bib_ref]. Here, we used the BiFC assay because it enables live cell visualization of protein-protein interactions. Moreover, it has proven to provide a reliable read-out of protein-protein interaction sites in several cell types and organisms [bib_ref] Distinct intracellular trafficking of equine infectious anemia virus and human immunodeficiency virus..., Jin [/bib_ref] [bib_ref] Detection and characterization of influenza A virus PA-PB2 interaction through a bimolecular..., Hemerka [/bib_ref] [bib_ref] Intracellular interactions between APOBEC3G, RNA, and HIV-1 Gag: APOBEC3G multimerization is dependent..., Friew [/bib_ref] [bib_ref] Anx2 interacts with HIV-1 Gag at phosphatidylinositol (4,5) bisphosphate-containing lipid rafts and..., Harrist [/bib_ref] [bib_ref] Coassembly and complementation of Gag proteins from HIV-1 and HIV-2, two distinct..., Boyko [/bib_ref] [bib_ref] Intermolecular interactions between retroviral Gag proteins in the nucleus, Kenney [/bib_ref]. As a starting point for this part of our research, we studied the interaction between Rev-dependent Gag proteins as depicted in [fig_ref] Figure 1: Gag interactions with host proteins Staufen1 and IMP1 occur in the cytoplasm... [/fig_ref] (top) [bib_ref] Distinct intracellular trafficking of equine infectious anemia virus and human immunodeficiency virus..., Jin [/bib_ref]. Gag multimerized and assembled with high efficiency as shown by strong green fluorescence signals due to Gag-VenusC (VC) and VenusN (VN) BiFC [fig_ref] Figure 1: Gag interactions with host proteins Staufen1 and IMP1 occur in the cytoplasm... [/fig_ref] , bottom panels). Gag-Gag multimerization occurred at the plasma membrane, and numerous Gag-Gag interaction events were also seen within the cytoplasm.
We then characterized the interaction between Gag and Staufen1. These proteins are known to interact in a RNA-independent manner [bib_ref] Identification of Staufen in the human immunodeficiency virus type 1 Gag ribonucleoprotein..., Chatel-Chaix [/bib_ref] and are in close proximity (≈10 nm) as determined by bioluminescence resonance energy transfer experiments [bib_ref] The host protein Staufen1 participates in human immunodeficiency virus type 1 assembly..., Chatel-Chaix [/bib_ref] [bib_ref] Identification of Staufen in the human immunodeficiency virus type 1 Gag ribonucleoprotein..., Chatel-Chaix [/bib_ref] [bib_ref] The host protein Staufen1 interacts with the Pr55Gag zinc fingers and regulates..., Chatel-Chaix [/bib_ref] , thus we expected to observe BiFC; but in addition, we wanted to identify the interaction sites for this virus-host pair. We detected small and large robust BiFC signals in the cytoplasm. Furthermore, a close examination of cells revealed that the Staufen1 and Gag BiFC signals coincided with the plasma membrane periphery [fig_ref] Figure 1: Gag interactions with host proteins Staufen1 and IMP1 occur in the cytoplasm... [/fig_ref] , similar to what was found for Gag. This was observed in over 90% of cells (n > 300) exhibiting BiFC.
We also performed BiFC to identify where Gag and Insulin like growth factor II mRNA binding protein (IMP1) interacted in cells. We chose IMP1 because it is a component of the Staufen1 RNP [bib_ref] Unexpected roles for UPF1 in HIV-1 RNA metabolism and translation, Ajamian [/bib_ref] [bib_ref] Visualization of RNA-protein interactions in living cells: FMRP and IMP1 interact on..., Rackham [/bib_ref] and because IMP1 associates to Gag and is incorporated in HIV-1 [bib_ref] Association of RNA helicase a with human immunodeficiency virus type 1 particles, Roy [/bib_ref] [bib_ref] Insulin-like growth factor II mRNA binding protein 1 associates with Gag protein..., Zhou [/bib_ref]. The co-expression of Gag-VN and IMP1-VC generated intense BiFC signals predominantly in the cytoplasm [fig_ref] Figure 1: Gag interactions with host proteins Staufen1 and IMP1 occur in the cytoplasm... [/fig_ref] , bottom panels) with a detectable amount at the plasma membrane in some cells (not shown; see [fig_ref] Figure 2: Interactions between Gag and cellular proteins Staufen1 or IMP1 at GM1 containing... [/fig_ref]. IMP1-Gag exhibited a very specific and abundant interaction and shared some features with the interaction site that we identified for Staufen1-Gag including well defined cytoplasmic and plasma membrane foci. The Gag-binding domain in IMP1 was mapped to the four KH RNA-binding domains [bib_ref] Insulin-like growth factor II mRNA binding protein 1 associates with Gag protein..., Zhou [/bib_ref]. Therefore we performed BiFC analysis; and as expected, the IMP1-KH(1-4)-Gag interaction was maintained [fig_ref] Figure 1: Gag interactions with host proteins Staufen1 and IMP1 occur in the cytoplasm... [/fig_ref] , top panels) whereas the expression of IMP1-RRM [bib_ref] HIV-1 gag proteins: diverse functions in the virus life cycle, Freed [/bib_ref] , lacking the interaction domain, failed to complement with Gag in this assay [fig_ref] Figure 1: Gag interactions with host proteins Staufen1 and IMP1 occur in the cytoplasm... [/fig_ref] , bottom panels). A variety of other negative controls were performed. For example, the co-expression of bacteriophage coat protein MS2 fused to the VN moiety with either Gag-VC, Staufen1-VC or IMP1-VC did not produce BiFC signals in any cell, demonstrating the specificity of the method (Additional file 1: [fig_ref] Figure 1: Gag interactions with host proteins Staufen1 and IMP1 occur in the cytoplasm... [/fig_ref]. Furthermore, we expressed the BiFC Gag moieties along with pNL4.3 proviral DNA at a 1:5 molar ratio in order to demonstrate that the resulting BiFC signals are specific and not due to artifacts created by Gag-VC/VN overexpression or changes in the kinetics of viral particle assembly, consistent with an earlier report [bib_ref] Imaging the biogenesis of individual HIV-1 virions in live cells, Jouvenet [/bib_ref]. BiFC will only be positive in cells expressing pNL4.3 in the presence of Rev. Importantly, this experimental set up, that also includes the expression of the full complement of viral genes, leads to identical BiFC signals (Additional file 1: [fig_ref] Figure 1: Gag interactions with host proteins Staufen1 and IMP1 occur in the cytoplasm... [/fig_ref]. Finally, BiFC analyses were performed in Jurkat T cells, and again, robust Gag-Gag and Gag-Staufen1 BiFC signals were evident at the periphery of T cells [fig_ref] Figure 1: Gag interactions with host proteins Staufen1 and IMP1 occur in the cytoplasm... [/fig_ref].
## Association of gag and cellular factors at gm1-containing lipid rafts on the plasma membrane
Our earlier reports indicated that Staufen1 associates with vRNA and Gag in both cells and virus [bib_ref] Identification of Staufen in the human immunodeficiency virus type 1 Gag ribonucleoprotein..., Chatel-Chaix [/bib_ref] [bib_ref] The double-stranded RNA-binding protein Staufen is incorporated in human immunodeficiency virus type..., Mouland [/bib_ref]. Our recent data suggest that this host protein modulates Gag multimerization on membranes [bib_ref] The host protein Staufen1 participates in human immunodeficiency virus type 1 assembly..., Chatel-Chaix [/bib_ref]. Gag preferentially mediates viral assembly at specific sites on the plasma Interactions between Gag and cellular proteins Staufen1 or IMP1 at GM1 containing lipid rafts on the plasma membrane as determined by BiFC. (A) HeLa cells were co-transfected with pCMV-Rev and Rev-dependent Gag-VN and Gag-VC plasmids. At 24 hr post-transfection lipid raft staining in live cells was performed. Images were captured using laser scanning confocal microscopy to detect the co-localization patterns of oligomerizing Gag molecules and lipid raft microdomains (indicated by CT-B that binds the pentasaccharide chain of the raft marker protein, GM1). membrane called lipid raft microdomains [bib_ref] Multimerization of human immunodeficiency virus type 1 Gag promotes its localization to..., Lindwasser [/bib_ref] [bib_ref] Depletion of cellular cholesterol inhibits membrane binding and higher-order multimerization of human..., Ono [/bib_ref] that are composed of cholesterol and sphingolipids, and contain several other components such as ganglioside GM1, glycophosphatidylinositol-anchored (GPI-anchored) proteins, tyrosine kinases of the Src family and others. Because the Gag and Staufen1 interaction occurs also on well defined plasma membrane structures [fig_ref] Figure 1: Gag interactions with host proteins Staufen1 and IMP1 occur in the cytoplasm... [/fig_ref] , we next determined the nature of these interaction domains using BiFC concomitant with live cell lipid raft staining. We transfected cells with Gag-VN and Gag-, Staufen1-or IMP1-VC BiFC constructs, and at 24 hr stained lipid rafts using AlexaFluor 594-labeled cholera toxin subunit B (CT-B) as described in Materials and Methods. As a reference condition for the association and multimerization of Gag on the plasma membrane, we again used Gag-VN and Gag-VC in BiFC [fig_ref] Figure 2: Interactions between Gag and cellular proteins Staufen1 or IMP1 at GM1 containing... [/fig_ref]. We observed an almost complete co-localization of CT-B label and oligomerizing Gag, which is in accordance with previously published work [bib_ref] Human immunodeficiency virus type 1 virological synapse formation in T cells requires..., Jolly [/bib_ref] [bib_ref] Requirement for an intact T-cell actin and tubulin cytoskeleton for efficient assembly..., Jolly [/bib_ref] [bib_ref] Role of lipid rafts in virus replication, Ono [/bib_ref]. Furthermore, BiFC between Gag-VN and Staufen1-VC followed by GM1 labeling revealed that a substantial part of these interactions also occurred at lipid rafts [fig_ref] Figure 2: Interactions between Gag and cellular proteins Staufen1 or IMP1 at GM1 containing... [/fig_ref]. Likewise, a proportion of the Gag-IMP1 BiFC signals coincided with lipid raft domains, although as reported above, the plasma membrane localization was not as marked [fig_ref] Figure 2: Interactions between Gag and cellular proteins Staufen1 or IMP1 at GM1 containing... [/fig_ref].
Staufen1's abundance at the plasma membrane was puzzling since it is normally distributed in the cytoplasm co-localizing with the endoplasmic reticulum [bib_ref] Mammalian staufen is a double-stranded-RNA-and tubulin-binding protein which localizes to the rough..., Wickham [/bib_ref] [bib_ref] Staufen recruitment into stress granules does not affect early mRNA transport in..., Thomas [/bib_ref]. Therefore, to determine if Staufen1 is recruited by Gag to lipid raft domains, we co-transfected HeLa cells with the BiFC plasmid pair Gag-VN and Gag-VC, and at 24 hr post-transfection, we fixed and then stained the cells for endogenous Staufen1 [fig_ref] Figure 2: Interactions between Gag and cellular proteins Staufen1 or IMP1 at GM1 containing... [/fig_ref]. The BiFC signals between Gag-VN/Gag-VC were observed at the plasma membrane and were preserved following fixation. Notably, abundant staining for endogenous Staufen1 coincided with the majority of the Gag BiFC signals in whole cells [fig_ref] Figure 2: Interactions between Gag and cellular proteins Staufen1 or IMP1 at GM1 containing... [/fig_ref] , left panel) and in the expanded region on the right [fig_ref] Figure 2: Interactions between Gag and cellular proteins Staufen1 or IMP1 at GM1 containing... [/fig_ref] , Gag-Gag positive cell) whereas in Gag-Gag negative cells, Staufen1 was dispersed in the cytoplasm. Thus, Staufen1 is recruited presumably by Gag to plasma membrane lipid rafts. Finally, we performed a similar analysis for endogenous Staufen1 in Jurkat T cells. Upon expression of Gag-VC and Gag-VN, endogenous Staufen1 coincided with Gag BiFC signals at cell-to-cell contact sites in Gag-expressing Jurkat T cells [fig_ref] Figure 2: Interactions between Gag and cellular proteins Staufen1 or IMP1 at GM1 containing... [/fig_ref].
We also observed Staufen1-Gag BiFC signals at intracellular domains marked by CT-B. These sites appeared to be vesicular in nature and were observed in ~75% of all cells examined (in >200 cells in 6 experiments; [fig_ref] Figure 2: Interactions between Gag and cellular proteins Staufen1 or IMP1 at GM1 containing... [/fig_ref] , white arrow). These sites of interaction with CT-B staining represent either rapidly internalized raft membrane domains or sites of raft biogenesis/synthesis [bib_ref] Structure and origin of ordered lipid domains in biological membranes, Brown [/bib_ref] [bib_ref] Functions of lipid rafts in biological membranes, Brown [/bib_ref]. To characterize them further, we performed time lapse imaging of live cells. The structures were mostly immobile, but several were dynamic showing characteristics of membranes that were capable of fusion, fission, detachment and subsequent trafficking towards the plasma membrane (Additional file 2: [fig_ref] Figure 2: Interactions between Gag and cellular proteins Staufen1 or IMP1 at GM1 containing... [/fig_ref]. This result suggests that Gag passes through intracellular lipid raft membrane domains on its way to the plasma membrane.
## Biochemical fractionation of lipid rafts
We performed a detergent-free membrane flotation assay to purify lipid rafts with the advantages that fewer insoluble aggregates form, and the purifications are met with less contamination from non-raft cellular membranes [bib_ref] A simplified method for the preparation of detergent-free lipid rafts, Macdonald [/bib_ref] ; [fig_ref] Figure 3: Staufen1 co-fractionates with Gag in lipid rafts [/fig_ref]. We either mock transfected HeLa cells or co-transfected them with Gag-VN and Gag-VC plasmids to reproduce our BiFC conditions above. Alternatively, cells were transfected with a Rev-dependent GagΔNC/p6 construct [bib_ref] Basic residues in the nucleocapsid domain of Gag are required for interaction..., Lingappa [/bib_ref] as a negative virus assembly control. At 24 hr post-transfection cells were lysed, washed and processed for fractionation. Eighteen fractions from each gradient were probed for the raft marker Caveolin-1 and for Staufen1, IMP1 and Tuberin (TSC2). Gag was detected with either an anti-GFP (recognizing VC of Gag-VC; [fig_ref] Figure 3: Staufen1 co-fractionates with Gag in lipid rafts [/fig_ref] or with an anti-p24 [fig_ref] Figure 3: Staufen1 co-fractionates with Gag in lipid rafts [/fig_ref] antibody. Lipid rafts and associated proteins such as Caveolin-1 accumulated principally in fractions #2 to #6 in all conditions (mock and in the presence of Gag or truncated Gag proteins). The cytoplasmic protein TSC2 principally sedimented to fractions #12 to #18 [fig_ref] Figure 3: Staufen1 co-fractionates with Gag in lipid rafts [/fig_ref] -D, representing non-membrane fractions), but small amounts were detected in association with rafts in the upper fractions as reported [bib_ref] Tuberin is a component of lipid rafts and mediates caveolin-1 localization: role..., Jones [/bib_ref]. In mock conditions, a fraction of Staufen1 sedimented to the lipid rafts (≈6.5% of total Staufen1; [fig_ref] Figure 3: Staufen1 co-fractionates with Gag in lipid rafts [/fig_ref]. In the presence of Gag however, a notable three-fold increase of Staufen1 (≈19%) fractionated to lipid raft fractions (Figure 3C &3E) with ≈14% of total Gag sedimenting within these fractions. In addition, a shift of IMP1 was observed within the gradient when Gag was expressed. Approximately 10% (vs ≈6% in mock conditions) and ≈31% (vs ≈5% in mock conditions) of IMP1 was found in lipid raft and intermediary fractions (#7-#10), respectively. This was consistent with our imaging data that detected a small proportion of IMP1 in the lipid rafts in Gag-expressing cells [fig_ref] Figure 1: Gag interactions with host proteins Staufen1 and IMP1 occur in the cytoplasm... [/fig_ref]. Of interest is the observation that IMP1 overexpression disrupts the association of mature Gag products on membranes [bib_ref] Insulin-like growth factor II mRNA binding protein 1 associates with Gag protein..., Zhou [/bib_ref] ; therefore, the abundance of IMP1 on lipid raft domains might be underrepresented. LAMP-3/ CD63 reactivity co-sedimented to these intermediary fractions (M.P.M. & A.J.M., data not shown) and further study of these membrane domains will be necessary. As a control we expressed the assembly defective GagΔNC/p6 which can not bind to several host proteins like Staufen1, IMP1 and HP68 (ABCE1), for example [bib_ref] Identification of Staufen in the human immunodeficiency virus type 1 Gag ribonucleoprotein..., Chatel-Chaix [/bib_ref] [bib_ref] Insulin-like growth factor II mRNA binding protein 1 associates with Gag protein..., Zhou [/bib_ref] [bib_ref] Basic residues in the nucleocapsid domain of Gag are required for interaction..., Lingappa [/bib_ref]. Neither GagΔNC/p6 nor Staufen1 sedimented to any great extent to the lipid raft fractions [fig_ref] Figure 3: Staufen1 co-fractionates with Gag in lipid rafts [/fig_ref] indicating a dependence on Gag for the enhanced recruitment of Staufen1 to lipid rafts. These biochemical data reflect our imaging data that showed a recruitment of Staufen1 to lipid raft microdomains in the majority of Gag-transfected cells [fig_ref] Figure 2: Interactions between Gag and cellular proteins Staufen1 or IMP1 at GM1 containing... [/fig_ref].
## Depletion of membrane cholesterol by hydroxy-propyl-βcyclodextrin (hβcd) reduces gag-gag and gag-staufen1 membrane bifc
Since cholesterol is essential for lipid raft structure and function, its depletion should disrupt BiFC signals occurring at these sites. Indeed, cholesterol depletion from the rafts reduces the total amount of associated Gag and more specifically the presence of higher-order Gag multimers on the plasma membrane [bib_ref] Depletion of cellular cholesterol inhibits membrane binding and higher-order multimerization of human..., Ono [/bib_ref]. Thus, to confirm that the plasma membrane domains where we find Gag-Staufen1 BiFC are lipid rafts, we depleted cholesterol from membranes using HβCD. At various time points between 0 and 25 minutes, lipid rafts were detected by CT-B staining; and the BiFC signals were imaged by laser scanning confocal microscopy. Multimerized Gag demonstrated strong association with GM1-containing lipid microdomains before addition of HβCD [fig_ref] Figure 4: Time-dependent depletion of cholesterol from lipid rafts leads to the disruption of... [/fig_ref] , t = 0 min). We then perfused cells with HβCD and collected images from live cells at time points thereafter. Timelapse imaging revealed a significant decrease in CT-B staining at all time points after 15 minutes indicating that HβCD was effective at disrupting lipid rafts. The BiFC signals for Gag-Gag also dramatically decreased over time [fig_ref] Figure 4: Time-dependent depletion of cholesterol from lipid rafts leads to the disruption of... [/fig_ref]. The decreases in Gag-Gag and Gag-Staufen1 BiFC signals were likely due to the disruption of plasma membrane assembly domains thereby preventing both the accumulation of Gag and the bimolecular interactions. Because we took multiple laser scans of the same cells, we attempted to rule out any effect photobleaching might have in the BiFC and CT-B signals captured at the later time points. We therefore transfected cells with a Rev-dependent Gag construct and at 24 hr treated them with HβCD over an extended period of time (0-45 min). Cells were subsequently fixed, and immunofluorescence was performed to obtain snapshots of the distributions of Gag and of the raft marker protein, Caveolin-1 [fig_ref] Figure 4: Time-dependent depletion of cholesterol from lipid rafts leads to the disruption of... [/fig_ref]. Indeed, both Caveolin-1 and Gag staining were diminished over the course of this experiment. Therefore, we can conclude that photobleaching is not significant in these experiments, and also that the scaffold for interactions between Gag and host proteins is disrupted by cholesterol depletion. Finally, whereas the BiFC signals for both the Staufen1-Gag and IMP1-Gag interactions were observed at time 0 (data not shown but refer to [fig_ref] Figure 2: Interactions between Gag and cellular proteins Staufen1 or IMP1 at GM1 containing... [/fig_ref] , these substantially decreased over time following HβCD treatment indicating that intact lipid rafts contribute to these bimolecular interactions [fig_ref] Figure 4: Time-dependent depletion of cholesterol from lipid rafts leads to the disruption of... [/fig_ref].
## Effects of modulating staufen1 levels on the distribution of gag-gag bifc complexes
To characterize the role of Staufen1 in the trafficking and formation of Gag-Gag assembly complexes in live cells, we depleted or overexpressed Staufen1 using siStaufen1 or a Staufen1-HA cDNA, respectively [bib_ref] Identification of Staufen in the human immunodeficiency virus type 1 Gag ribonucleoprotein..., Chatel-Chaix [/bib_ref] [bib_ref] The double-stranded RNA-binding protein Staufen is incorporated in human immunodeficiency virus type..., Mouland [/bib_ref]. The depletion of Staufen1 was efficient and reduced expression levels to less than 1/6 while the over-expression increased cellular Staufen1 levels approximately 6-fold [fig_ref] Figure 5: Staufen1 depletion decreases plasma membrane-associated Gag and results in intracellular clustering of... [/fig_ref]. In cells transfected with a control siRNA (siNS) coexpressing Gag-VN and Gag-VC, Gag BiFC was found at the plasma membrane and in discrete cytoplasmic domains as shown earlier [fig_ref] Figure 5: Staufen1 depletion decreases plasma membrane-associated Gag and results in intracellular clustering of... [/fig_ref]. However, in Staufen1-depleted cells, in addition to the Gag-Gag BiFC signals observed at the plasma membrane, strong signals were observed at cytoplasmic juxtanuclear regions [fig_ref] Figure 5: Staufen1 depletion decreases plasma membrane-associated Gag and results in intracellular clustering of... [/fig_ref]. When Staufen1-HA was over-expressed, we observed that the Gag-Gag BiFC punctae were abundant and well defined, and we could not detect any marked changes in plasma membrane association of Gag-Gag BiFC compared to the control siNS condition [fig_ref] Figure 5: Staufen1 depletion decreases plasma membrane-associated Gag and results in intracellular clustering of... [/fig_ref]. Neither the depletion nor the over-expression of Staufen1 caused any significant redistribution of ABCE1, a host factor that interacts with NC domain of Gag and is involved in assembly [bib_ref] Basic residues in the nucleocapsid domain of Gag are required for interaction..., Lingappa [/bib_ref] [bib_ref] Host ABCE1 is at plasma membrane HIV assembly sites and its dissociation..., Dooher [/bib_ref] , in relationship to the localization of Gag-Gag BiFC or gag RNA signals (Additional file 3: [fig_ref] Figure 3: Staufen1 co-fractionates with Gag in lipid rafts [/fig_ref]. Likewise, the depletion of the Staufen1 63 kDa isoform alone [bib_ref] Novel Staufen1 ribonucleoproteins prevent formation of stress granules but favour encapsidation of..., Abrahamyan [/bib_ref] or the depletion of UPF1 [bib_ref] Unexpected roles for UPF1 in HIV-1 RNA metabolism and translation, Ajamian [/bib_ref] did not result in intracellular Gag BiFC signals (M.P.M., Lara Ajamian and A.J.M., data not shown).
We noticed earlier that Gag-Gag BiFC occurred on sometimes circular, membrane-like structures. Moreover, Gag, Staufen1 and vRNA traffic on endosomal membranes and in a manner that is dependent on endosomal vesicle positioning [bib_ref] Intracellular transport of human immunodeficiency virus type 1 genomic RNA and viral..., Lehmann [/bib_ref] [bib_ref] Endosomal trafficking of HIV-1 gag and genomic RNAs regulates viral egress, Molle [/bib_ref]. Therefore to determine the nature of the Gag structures, we co-expressed RFP fusion marker proteins Rab5 (early endosomes), Rab7 (late endosomes), Rab9 (Golgi/endoplasmic reticulum), LAMP1 (late endosomes) and Caveolin-1 (lipid rafts/ caveosomes) in Staufen1-depleted cells. This analysis revealed that the Gag-Gag BiFC signals were on membranes that bore characteristics of endosomal membranes/vesicles, consistent with recent work [bib_ref] Intracellular transport of human immunodeficiency virus type 1 genomic RNA and viral..., Lehmann [/bib_ref] [bib_ref] Endosomal trafficking of HIV-1 gag and genomic RNAs regulates viral egress, Molle [/bib_ref] ; [fig_ref] Figure 6: Gag localizes near Rab7-, Rab9-and LAMP1-containing membranes at cytoplasmic and juxtanuclear sites... [/fig_ref]. Specifically, the Gag-Gag BiFC multimers that coalesced intracellularly upon Staufen1 depletion coincided to varying extents with late endosomal membrane components LAMP-1-, Rab7-and on Rab9-containing membranes. While Staufen1 depletion did not influence the patterns of Rab7, Rab5, endoplasmic reticulum or Golgi staining (M.P.M. and A.J.M., data not shown), we nevertheless attempted to identify the origin of the Gag BiFC signals in Staufen1-depleted cells. To do this, we overexpressed two trans-dominant negative (TDN) proteins that efficiently block endocytosis at various steps as described [bib_ref] Intracellular transport of human immunodeficiency virus type 1 genomic RNA and viral..., Lehmann [/bib_ref]. These include Eps15-TDN-GFP and Rab7-TDN-GFP. These have been shown to block early endocytic events as shown by a block to Transferrin receptor recycling. Identical distributions of the Gag (expressed here as Rev-dependent Gag only [bib_ref] Basic residues in the nucleocapsid domain of Gag are required for interaction..., Lingappa [/bib_ref] signals were obtained when clathrin-dependent and clathrinindependent (M.P.M. and A.J.M., data not shown) endocytosis was blocked in Staufen1-depleted cells (Additional file 4: [fig_ref] Figure 4: Time-dependent depletion of cholesterol from lipid rafts leads to the disruption of... [/fig_ref] and data not shown). Therefore, we propose that the localization of the BiFC signals obtained in Staufen1-depleted cells was unlikely to be due to endocytosed Gag-Gag multimers (see Discussion).
## Trimolecular fluorescence complementation (trifc) to visualize specific gag-rna interactions in living mammalian cells
Gag-Gag and Gag-Staufen1 interactions as visualized by BiFC were found consistently in the cytoplasm and at plasma membrane lipid raft domains. Because both Gag and Staufen1 bind mRNAs, we therefore explored how the association to mRNA would influence the Gag-Gag and Gag-Staufen1 associations. To do this, we employed TriFC, a method that allows the detection of protein-RNA or protein-protein-RNA interactions in live cells [bib_ref] Visualization of RNA-protein interactions in living cells: FMRP and IMP1 interact on..., Rackham [/bib_ref]. To validate this assay we took advantage of the well characterized Gag-vRNA association mediated by the psi RNA packaging signal sequence. Thus, we generated constructs that expressed mRNAs containing either the complete packaging signal psi (pGL3MS2site-psi; including SL2 and SL3 sequences), or delta psi domain in which 19 nucleotides are deleted from SL3 (pGL3MS2site-Deltapsi; [fig_ref] Table 1: List of plasmids used in this study [/fig_ref] ; [bib_ref] Identification of a sequence required for efficient packaging of human immunodeficiency virus..., Lever [/bib_ref]. The mRNA reporters, based on pGL3MS2 site/basic [bib_ref] Visualization of RNA-protein interactions in living cells: FMRP and IMP1 interact on..., Rackham [/bib_ref] , were also designed to contain an MS2 binding site that provides an efficient tethering of MS2-VN fusion protein. The K D of this interaction falls in the nanomolar range [bib_ref] Mutants of the bacteriophage MS2 coat protein that alter its cooperative binding..., Lecuyer [/bib_ref] , and it likely occurs rapidly and is stable. As shown in [fig_ref] Figure 7: TriFC analysis for studying RNA-protein interactions in live cells [/fig_ref] in order for the successful complementation of both VN and VC fragments the simultaneous expression of three plasmids is required. In the first case the VN is expressed as a fusion MS2-VN protein and can tether to the RNA reporter in proximity to the psi sequence [fig_ref] Figure 7: TriFC analysis for studying RNA-protein interactions in live cells [/fig_ref]. hGag-VC will interact with the psi domain of the reporter to allow for fluorescence complementation [fig_ref] Figure 7: TriFC analysis for studying RNA-protein interactions in live cells [/fig_ref] ; or when the psi is mutated, fluorescence complementation will not occur [fig_ref] Figure 7: TriFC analysis for studying RNA-protein interactions in live cells [/fig_ref]. Using this assay, we found intense TriFC signals in the juxtanuclear region and cytoplasm [fig_ref] Figure 7: TriFC analysis for studying RNA-protein interactions in live cells [/fig_ref]. In contrast, we did not detect TriFC when the mRNA reporter harboring the mutated psi (delta psi) was expressed [fig_ref] Figure 7: TriFC analysis for studying RNA-protein interactions in live cells [/fig_ref] , and this was found in at least 5 experiments and having examined >400 viable cells. This pGag-VC [bib_ref] Distinct intracellular trafficking of equine infectious anemia virus and human immunodeficiency virus..., Jin [/bib_ref] pCMV-Rev NIH pSVGagRRE-R [bib_ref] Requirements for incorporation of Pr160gag-pol from human immunodeficiency virus type 1 into..., Smith [/bib_ref] pGag-NC/p6(Tr361) [bib_ref] Basic residues in the nucleocapsid domain of Gag are required for interaction..., Lingappa [/bib_ref] pGL3MS2 site/basic [bib_ref] Visualization of RNA-protein interactions in living cells: FMRP and IMP1 interact on..., Rackham [/bib_ref] pGL3-βActin zipcode [bib_ref] Visualization of RNA-protein interactions in living cells: FMRP and IMP1 interact on..., Rackham [/bib_ref] pGL3MS2site-psi This study pGL3MS2site-Delta-psi This study pEps15-TDN-GFP [bib_ref] Intracellular transport of human immunodeficiency virus type 1 genomic RNA and viral..., Lehmann [/bib_ref] pRab5-TDN-GFP [bib_ref] Intracellular transport of human immunodeficiency virus type 1 genomic RNA and viral..., Lehmann [/bib_ref] mRFP-Rab5 Addgene plasmid 14437, [bib_ref] Rab7 associates with early endosomes to mediate sorting and transport of Semliki..., Vonderheit [/bib_ref] dsRed-rab7 Addgene plasmid 12661, [bib_ref] Rab proteins mediate Golgi transport of caveolainternalized glycosphingolipids and correct lipid trafficking..., Choudhury [/bib_ref] dsRed-rab9 WT Addgene plasmid 12677, [bib_ref] Rab proteins mediate Golgi transport of caveolainternalized glycosphingolipids and correct lipid trafficking..., Choudhury [/bib_ref] Cav1-mRFP Addgene plasmid 14434, [bib_ref] Assembly and trafficking of caveolar domains in the cell: caveolae as stable,..., Tagawa [/bib_ref] LAMP1-RFP Addgene plasmid 1817,result was in accordance with our expectation because packaging into virions of the virus-specific RNA harboring this deletion was less than 2% of that of the wild-type virus [bib_ref] Identification of a sequence required for efficient packaging of human immunodeficiency virus..., Lever [/bib_ref]. These findings indicate that the association of Gag with mRNA reporter molecules is sufficient to sequester them into granules that bear hallmark appearance of RNP containing proteins and mRNA [bib_ref] Visualization of RNA-protein interactions in living cells: FMRP and IMP1 interact on..., Rackham [/bib_ref]. The experiments demonstrated the biological relevance and specificity of the interactions between Gag and the HIV-1 vRNA psi packaging domain detected by TriFC. When full-length Venus was expressed in the context of the fusion MS2-Venus and co-expressed with the reporter mRNA bearing MS2 binding domains, a uniformly distributed fluorescence signal was obtained throughout the cell, in contrast to what we have obtained above (Additional file 5: [fig_ref] Figure 5: Staufen1 depletion decreases plasma membrane-associated Gag and results in intracellular clustering of... [/fig_ref]. Additional negative controls included the expression of a psi mRNA reporter (pGL3MS2site-Delta-psi), the expression of Staufen1-or IMP1-VC fusion proteins [fig_ref] Figure 7: TriFC analysis for studying RNA-protein interactions in live cells [/fig_ref] &7F) or the exclusion of the VN or VC expression vectors in the transfections (M.P.M. and A.J.M., data not shown). We also expressed an mRNA reporter that contains the β-Actin zipcode RNA sequence, and we did not detect fluorescence complementation [fig_ref] Figure 7: TriFC analysis for studying RNA-protein interactions in live cells [/fig_ref] , whereas a strong fluorescence signal was obtained using the cognate binding protein IMP1-VC [fig_ref] Figure 7: TriFC analysis for studying RNA-protein interactions in live cells [/fig_ref] ; [bib_ref] Visualization of RNA-protein interactions in living cells: FMRP and IMP1 interact on..., Rackham [/bib_ref]. Because Gag expression is Rev-independent and may traffic aberrantly in this set of experiments, we modified the TriFC assay and employed a Rev-dependent Gag expressor. We coexpressed the luciferase mRNA reporter harboring the psi domain (or a mutated form as described above), MS2-VN to tether to the MS2 RNA loops, Staufen1-VC, a Revdependent Gag expressor [bib_ref] Basic residues in the nucleocapsid domain of Gag are required for interaction..., Lingappa [/bib_ref] and pCMV-Rev as depicted (Additional file 6: [fig_ref] Figure 6: Gag localizes near Rab7-, Rab9-and LAMP1-containing membranes at cytoplasmic and juxtanuclear sites... [/fig_ref]
## Trifc visualization of protein-protein recruitment and interactions on mrna template
Using TriFC, we next evaluated whether Gag, while tethered to the mRNA, could recruit Staufen1 and other host proteins. MS2-Staufen1, MS2-hGag and MS2-Gag(C36S) expression constructs were created for this purpose. When mRNA reporters are co-expressed with MS2-VN and MS2-fusion proteins, these molecules will tether to the same MS2 RNA-binding site (MS2BS) on the mRNA. This is a native property of the bacteriophage coat protein that binds the MS2BS hairpin as a dimer [bib_ref] Crystal structure of the MS2 coat protein dimer: implications for RNA binding..., Ni [/bib_ref] [bib_ref] The three-dimensional structures of two complexes between recombinant MS2 capsids and RNA..., Valegard [/bib_ref]. In addition, in this system we expressed a second protein of interest as a fusion with VC. When MS2-hGag was cotethered with MS2-VN to the mRNA reporter without any additional protein binding domains [bib_ref] Visualization of RNA-protein interactions in living cells: FMRP and IMP1 interact on..., Rackham [/bib_ref] , hGag-VC was recruited to the mRNA as evidenced by bright TriFC signals [fig_ref] Figure 8: Gag recruits host proteins Staufen1 and IMP1 while tethered to mRNA [/fig_ref]. At the later time point (40 hr), the punctae were larger, more abundant and were better defined. Nevertheless, the signal intensity and abundance of TriFC signals generally correlated with protein expression levels [fig_ref] Figure 8: Gag recruits host proteins Staufen1 and IMP1 while tethered to mRNA [/fig_ref]. mRNA-tethered MS2-hGag recruited both Staufen1-VC and IMP1-VC [fig_ref] Figure 8: Gag recruits host proteins Staufen1 and IMP1 while tethered to mRNA [/fig_ref] &8C). TriFC signals had apparent sizes ranging from 0.2 to 1.25 μm in diameter and were distributed in the cytoplasm of host cells, similar to what we found for hGag proteins (see Discussion). Because NC is the interacting domain for APOBEC3G [bib_ref] APOBEC3G is incorporated into virus-like particles by a direct interaction with HIV-1..., Alce [/bib_ref] , Staufen1 [bib_ref] Identification of Staufen in the human immunodeficiency virus type 1 Gag ribonucleoprotein..., Chatel-Chaix [/bib_ref] , and IMP1 When we tethered MS2-VN and MS2-Staufen1 to the mRNA via MS2BS and expressed Gag-VC, TriFC was readily detected, indicating that Staufen1, when bound to mRNA, recruits Gag [fig_ref] Figure 8: Gag recruits host proteins Staufen1 and IMP1 while tethered to mRNA [/fig_ref]. These results demonstrate that Gag potentially recruits cellular factors while bound to an mRNA; and likewise, RNA-binding host factors can recruit Gag to mRNA. The results also support the notion that the HIV-1 RNP, containing at its core the precursor Gag and the genomic RNA, will selectively engage cellular factors such as Staufen1 and IMP1 during assembly.
# Discussion
## Gag-staufen1 interactions in living cells
Staufen1 was previously found as a component of HIV-1 particles [bib_ref] The double-stranded RNA-binding protein Staufen is incorporated in human immunodeficiency virus type..., Mouland [/bib_ref] [bib_ref] The anti-HIV-1 editing enzyme APOBEC3G binds HIV-1 RNA and messenger RNAs that..., Kozak [/bib_ref]. We have uncovered additional roles for this host factor including one in promoting Gag multimerization and assembly and another in the selection of vRNA for encapsidation [bib_ref] The double-stranded RNA-binding protein Staufen is incorporated in human immunodeficiency virus type..., Mouland [/bib_ref] [bib_ref] Novel Staufen1 ribonucleoproteins prevent formation of stress granules but favour encapsidation of..., Abrahamyan [/bib_ref]. In the present study, we extend our understanding of Staufen1's role in HIV-1 replication by characterizing its interactions with Gag using two powerful, live cell fluorescence complementation techniques. We demonstrate that Staufen1 interacts with Gag, that Staufen1 recruits Gag when it is bound to mRNA and likewise, Gag recruits Staufen1 when bound to the cognate psi packaging signal. The results highlight Staufen1's involvement in the HIV-1 RNP in the assembly of HIV-1.
## Recruitment of staufen1 by gag to lipid rafts and to mrna
Our earlier work demonstrated that when we modulated Staufen1 levels in cells, Gag-Gag interactions increased, and these accumulated in detergent-resistant complexes [bib_ref] The host protein Staufen1 participates in human immunodeficiency virus type 1 assembly..., Chatel-Chaix [/bib_ref]. Here, we used BiFC to evaluate where Staufen1 and Gag interact. This approach, which employs native Gag [bib_ref] The double-stranded RNA-binding protein Staufen is incorporated in human immunodeficiency virus type..., Mouland [/bib_ref] hr later laser scanning confocal microscopy was used to assess TriFC. Bright fluorescence signals in cytoplasmic punctae were detected indicating that the interaction between hGag-VC and the psi RNA domain occurred. This condition represented a positive control for the TriFC system employed here. TriFC signals were not detected when HeLa cells were transfected with the mutated psi reporter (pGL3MS2site-Delta-psi) (D). In order to determine the specificity of the TriFC assay, HeLa cells were co-transfected with plasmids expressing either Staufen1-VC (E) or IMP1-VC (F) along with MS2-VN and pGL3MS2site-psi. TriFC was absent in all cells examined indicating that Staufen1 and IMP1 do not bind the psi RNA. In parallel experiments, Gag did not associate with mRNA reporter bearing β-Actin zipcode sequence (pGL3-βActin zipcode) (G) whereas IMP1 did (H), as expected [bib_ref] Visualization of RNA-protein interactions in living cells: FMRP and IMP1 interact on..., Rackham [/bib_ref]. Phase contrast images are shown on the left of each panel in (C)-(H). The size bars are equal to 10 μm. sequences, has become increasingly popular to visualize Gag in cells [bib_ref] HIV-1 matrix dependent membrane targeting is regulated by Gag mRNA trafficking, Jin [/bib_ref]. While Staufen1 and Gag are shown to associate in cytoplasmic compartments, our results also reveal that these interactions occur on membranes and at the plasma membrane where Gag multimerizes during assembly [bib_ref] Distinct intracellular trafficking of equine infectious anemia virus and human immunodeficiency virus..., Jin [/bib_ref] [bib_ref] The host protein Staufen1 participates in human immunodeficiency virus type 1 assembly..., Chatel-Chaix [/bib_ref]. Whereas Staufen1 is usually found to localize on reticulotubular structures and the endoplasmic reticulum [bib_ref] Mammalian staufen is a double-stranded-RNA-and tubulin-binding protein which localizes to the rough..., Wickham [/bib_ref] , we show that Staufen1 is recruited from the cytoplasm to lipid rafts at the plasma membrane where it interacts with Gag. Consistently, Staufen1 colocalizes with Gag and vRNA at this location [bib_ref] Novel Staufen1 ribonucleoproteins prevent formation of stress granules but favour encapsidation of..., Abrahamyan [/bib_ref] , and the distribution of the Staufen1-interacting partner, UPF1, also shuttles to the plasma membrane domain in HIV-1 expressing cells [bib_ref] Unexpected roles for UPF1 in HIV-1 RNA metabolism and translation, Ajamian [/bib_ref]. Here, Gag expression alone appears to be the driving force behind Staufen1 recruitment to GM1-positive plasma membrane domains that serve as the main platforms for viral assembly. Gag's ability to recruit host factors and bind psi RNA concomitantly, via the same protein domain (NC), reveals a rather multifaceted property of Gag and further strengthens the implication of Staufen1 in HIV-1 assembly [bib_ref] The host protein Staufen1 participates in human immunodeficiency virus type 1 assembly..., Chatel-Chaix [/bib_ref]. Furthermore, time-lapse confocal imaging showed that Gag-Staufen1 foci are mobile and dynamic and are able to merge with similar structures and separate over time into smaller particles that traffic to and anchor at the plasma mem-brane (Additional file 2: [fig_ref] Figure 2: Interactions between Gag and cellular proteins Staufen1 or IMP1 at GM1 containing... [/fig_ref]. Consistently, distinct populations of Gag and Staufen1 (and the vRNA) traffic on endosomal membranes in cells [bib_ref] Intracellular transport of human immunodeficiency virus type 1 genomic RNA and viral..., Lehmann [/bib_ref] [bib_ref] Endosomal trafficking of HIV-1 gag and genomic RNAs regulates viral egress, Molle [/bib_ref]. Moreover, the Gag-vRNA interactions as well intermediary Gag assembly domains have been found in juxtanuclear domains [bib_ref] Kinesin KIF4 regulates intracellular trafficking and stability of the human immunodeficiency virus..., Martinez [/bib_ref] [bib_ref] HIV-1 Gag-RNA interaction occurs at a perinuclear/centrosomal site; analysis by confocal microscopy..., Poole [/bib_ref] [bib_ref] Trafficking of HIV-1 RNA is mediated by heterogeneous nuclear ribonucleoprotein A2 expression..., Levesque [/bib_ref] [bib_ref] Identification of an intracellular trafficking and assembly pathway for HIV-1 gag, Perlman [/bib_ref]. Thus, we propose that Staufen1 is hijacked by Gag shortly after its synthesis in order to assist in trafficking and assembly.
There are potentially two caveats with respect to the TriFc method used here. First is the use of codon-optimized hGag, that when expressed, could result in deviations in transport and assembly. Nevertheless, several recent studies have utilized Rev-independent hGag expressors to uncover new information on intracellular Gag trafficking and Gag interacting partners [bib_ref] Distinct intracellular trafficking of equine infectious anemia virus and human immunodeficiency virus..., Jin [/bib_ref] [bib_ref] Anx2 interacts with HIV-1 Gag at phosphatidylinositol (4,5) bisphosphate-containing lipid rafts and..., Harrist [/bib_ref] [bib_ref] HIV-1 matrix dependent membrane targeting is regulated by Gag mRNA trafficking, Jin [/bib_ref]. Furthermore, we show that by employing a Rev-dependent Gag expressor in a modified TriFC analyses nearly identical TriFC signals were obtained, revealing that codon-optimization of hGag did not have a significant bearing on the results. The second caveat is the possibility for an alternative interpretation that includes the order in which the protein-protein and protein-RNA interactions occur. We are claiming that either Gag or Staufen1 recruits the other partner while bound to mRNA, and this is supported by BiFC and endogenous Staufen1 staining [fig_ref] Figure 2: Interactions between Gag and cellular proteins Staufen1 or IMP1 at GM1 containing... [/fig_ref]. However, the bimolecular interaction between Gag and Gag and that found between Staufen1 and Gag could represent the initial event after which the bimolecular complex co-traffics to the mRNA substrate. Even though the K D for MS2-MS2BS interaction is in the low nanomolar range, further work will be necessary in order to confidently differentiate between some of these possibilities.
## Involvement of staufen1 in the anterograde trafficking of gag
We recently showed that a population of Staufen1 associates along with Gag and vRNA on endosomal membranes [bib_ref] Intracellular transport of human immunodeficiency virus type 1 genomic RNA and viral..., Lehmann [/bib_ref]. Likewise, RNPs translocate within cells by making use of machineries that direct traffic of cellular membranes and vesicles (reviewed in [bib_ref] The role of membranes and membrane trafficking in RNA localization, Cohen [/bib_ref]. In this study we wanted to characterize the possible roles of Staufen1 in the transport and distribution of multimerizing Gag-Gag that we can readily visualize in live cells. To resolve this, we modulated Staufen1 expression levels in cells that expressed Gag-VN/Gag-VC. When we depleted Staufen1, Gag-Gag BiFC signals were found at the plasma membrane but also in the cytoplasm at juxtanuclear regions on what appeared to be endosomal membrane vesicles. We suspect that the Gag-Gag BiFC signals are due to coalescing HIV-1 RNPs that tether to endosomal membrane populations that may not traffic properly. These were identified in earlier work [bib_ref] Intracellular transport of human immunodeficiency virus type 1 genomic RNA and viral..., Lehmann [/bib_ref] and were found to form in an HIV-1-dependent manner containing Staufen1, Gag (and Gag multimers) and vRNA reactivity later named Staufen1 HIV-1-dependent ribonucleoproteins (SHRNPs; [bib_ref] Novel Staufen1 ribonucleoproteins prevent formation of stress granules but favour encapsidation of..., Abrahamyan [/bib_ref]. The data shown here are consistent with the enhanced Gag-Gag multimerization in Staufen1depleted cells that we found using another biophysical technique (bioluminescence resonance energy transfer) [bib_ref] The host protein Staufen1 participates in human immunodeficiency virus type 1 assembly..., Chatel-Chaix [/bib_ref] , but also reveal some of the principal locations of these events (at juxtanuclear and raft domains) [bib_ref] Identification of Staufen in the human immunodeficiency virus type 1 Gag ribonucleoprotein..., Chatel-Chaix [/bib_ref] [bib_ref] Novel Staufen1 ribonucleoproteins prevent formation of stress granules but favour encapsidation of..., Abrahamyan [/bib_ref]. It is possible that, during viral egress, Gag mediates the formation of its own mixed type of membranes bearing several endosomal marker proteins similar to those that other viruses engineer [bib_ref] Virus factories: associations of cell organelles for viral replication and morphogenesis, Novoa [/bib_ref]. Of interest are data from recently published work in which the accumulation of HIV-1 Gag and cholesterol-enriched membranes was observed in juxtanuclear late endosomal compartments in Niemann-Pick type C1-deficient cells [bib_ref] Deficiency of niemann-pick type C-1 protein impairs release of human immunodeficiency virus..., Tang [/bib_ref]. Furthermore, the targeted depletion of the suppressor of cytokine signalling 1 reduced the transport of Gag to the plasma membrane, leading to accumulations of it near the nucleus. In addition, strong co-localization between another late endosomal marker, CD63, and Gag was observed as a result of Rab9 depletion [bib_ref] Rab9 GTPase is required for replication of human immunodeficiency virus type 1,..., Murray [/bib_ref]. These domains appear to be important for steps in assembly including Gag-Gag and Gag-vRNA associations [bib_ref] Intracellular transport of human immunodeficiency virus type 1 genomic RNA and viral..., Lehmann [/bib_ref] [bib_ref] HIV-1 Gag-RNA interaction occurs at a perinuclear/centrosomal site; analysis by confocal microscopy..., Poole [/bib_ref] [bib_ref] Trafficking of HIV-1 RNA is mediated by heterogeneous nuclear ribonucleoprotein A2 expression..., Levesque [/bib_ref] , vRNA encapsidation [bib_ref] Novel Staufen1 ribonucleoproteins prevent formation of stress granules but favour encapsidation of..., Abrahamyan [/bib_ref] and Gag trafficking [bib_ref] Identification of an intracellular trafficking and assembly pathway for HIV-1 gag, Perlman [/bib_ref] [bib_ref] Human Ubc9 contributes to production of fully infectious human immunodeficiency virus type..., Jaber [/bib_ref] and in one case, for Gag degradation. Importantly, we showed that the cytoplasmic accumulations of Gag did not form because of endocytosed Gag or viral particles derived from the plasma membrane (Additional file 4: [fig_ref] Figure 4: Time-dependent depletion of cholesterol from lipid rafts leads to the disruption of... [/fig_ref]. These results suggest that the Staufen1 plays roles in the anterograde transport of Gag, but is also a host factor involved in the formation of the HIV-1 RNP during viral egress and assembly.
# Conclusions
In the present study, we demonstrate that the intermolecular associations between the host mRNA-binding protein Staufen1 and HIV-1 Gag occur in the cytoplasm and at the plasma membrane in HeLa cells and T lymphocytes. Furthermore, we demonstrate that Staufen1 is recruited by Gag to lipid raft microdomains and present results that indicate Staufen1 has potentially novel functions in the intracellular trafficking of HIV-1 Gag during viral egress.
# Methods
## Cell culture and transfections
HeLa cells and Jurkat T cells were maintained under standard conditions at 37°C with in Dulbecco's modified Eagle's medium or RPMI, respectively, supplemented with 10% fetal bovine serum (FBS), 100 units/ml penicillin and 100 mg/ml streptomycin (Invitrogen). Lipofectamine 2000 (Invitrogen) was used for the DNA transfections of Jurkat T cells or HeLa cells according to the manufacturer's protocol. 1 day before transfection 1.5-1.8 × 10 5 or 4-5 × 10 5 HeLa cells were seeded in 6-well plates or 60-mm dishes, respectively to have ~60% confluency when transfected. 10 6 Jurkat T cells were transfected in 25 ml tissue culture flasks.
## Bifc and trifc analysis
For the BiFC and TriFC experiments, HeLa cells were plated on poly-D-lysine-coated 18 mm micro-glass coverslips (VWR) or 40 mm Bioptechs cover glasses 20-24 hr prior to transfection. Cells were transfected with 1.5 μg to 3 μg plasmid DNA per 18 mm and 3 μg to 5 μg per 40 mm depending on the design of the experiment. Jurkat T cells were transfected in suspension. At 24 hr post-transfection coverslips (HeLa cells) were mounted on a Bioptechs FCS3 imaging perfusion chamber (Bioptechs, Inc.), continuously perfused with fresh medium and warmed using a heating element (Bioptechs, Inc.). For Jurkat T cells, cell adherence to coverslips was promoted using poly-Dlysine (Sigma-Aldrich, Inc). Live cell imaging was performed at 37°C using either Zeiss Pascal LSM5 confocal microscope (Carl-Zeiss, Inc.) with 63×1.3 oil immersion objective or inverted Leica fluorescence microscope. Time-lapse images were captured using 63×1.3 oil immersion objective mounted on a motorized Leica microscope equipped with a PerkinElmer ERS spinning disk confocal system with heated stage and chamber to maintain the cells at 37°C and CO 2 . Images were collected at the times shown or as stated otherwise, for the indicated period. Two-dimensional data sets were deconvoluted using AutoDeblur (MediaCybernetics, Inc.) and compiled using Imaris 6.3.1 software (Bitplane, Inc.). In some experiments, BiFC and TriFC signals were measured in two-dimensional single confocal planes using Imaris software.
## Immunofluorescence and fluorescence in situ hybridization (if/fish) analyses
Laser scanning confocal microscopy was performed using a Leica microscope equipped with a PerkinElmer ERS spinning disk or a Carl-Zeiss LSM5 Pascal confocal microscope . Combined 2-and 3-colour IF/FISH co-analyses were performed exactly as described [bib_ref] Intracellular transport of human immunodeficiency virus type 1 genomic RNA and viral..., Lehmann [/bib_ref]. Images were captured at 512×512 or at 1024×1024 pixels resolution.
## Sds-page, western blot analysis and antibodies
SDS-PAGE and western blotting were performed as described earlier [bib_ref] Novel Staufen1 ribonucleoproteins prevent formation of stress granules but favour encapsidation of..., Abrahamyan [/bib_ref]. Antibodies used included: rabbit anti-Caveolin-1 (Santa Cruz); rabbit anti-Tuberin (TSC1, Abcam); rabbit anti-p24 (Intracell); mouse anti-p24 (NIH); rabbit anti-IMP1 (gift from Finn Nielsen, Rigshospitalet, Copenhagen, Denmark); mouse or rabbit anti-Staufen1 (gifts from Luc DesGroseillers, Université de Montréal, Canada and Graciela Boccaccio, University LeLoir, Argentina); rabbit anti-ABCE1 (a gift from Jais Lingappa, University of Washington); mouse anti-GFP (Roche) and mouse anti-Luciferase (Sigma-Aldrich).
## Lipid rafts staining
To visualize the distribution of insoluble membrane microdomains we used Vybrant Lipid Raft Labeling Kit (Invitrogen) according to the manufacturer's protocol. The method relies on the binding of a red-fluorescent AlexaFluor 594 conjugate of cholera toxin subunit B (CT-B) to the pentasaccharide chain of plasma membrane ganglioside GM1 (a lipid raft marker). An antibody that recognizes CT-B is then used to cross-link the CT-Blabeled lipid rafts into distinct patches on the plasma membrane, which we visualized by microscopy. The coverslips with stained live cells were mounted on Bioptechs FCS3 live cell perfusion chamber and were visualized by laser scanning confocal microscopy.
## Isolation and analysis of detergent-free lipid rafts
We used a simplified method for the fractionation of lipid rafts that does not require the use of detergent [bib_ref] A simplified method for the preparation of detergent-free lipid rafts, Macdonald [/bib_ref]. All procedures were carried out with RNAse-free equipment and materials and on ice. Briefly, for each of the cases, two 175 cm 2 of HeLa cells were transfected with corre-sponding plasmids. 24 hr later they were washed 3 times and scraped into buffer B1 (20 mM Tris-HCl pH7.8, 250 mM sucrose, 1 mM CaCl 2 , 1 mM MgCl 2 and RNAse out (Invitrogen) 1 μl/5 ml), centrifuged for 5 min at 1500 rpm and resuspended again in 1 ml of buffer B1 containing complete protease inhibitor cocktail (Roche). The cells were homogenized in RNAse free 1.5 ml Eppendorf tubes with plastic pestles, centrifuged at 1000 rpm for 10 min and the supernatant (S1) was collected. The pellet was lysed in 1 ml of B1 and homogenized again. Following centrifugation at 500 × g for 10 min, the second supernatant (S2) was collected and mixed with S1 (total ~2 ml). 2 ml of 50% OptiPrep (diluted in B1 without Ca 2+ and Mg 2+ to reach 50%) was added to S1 and S2. The resulting 4 ml of 25% OptiPrep mixture was first poured in 12 ml centrifugation tubes (Beckman Coulter). A step gradient 0-20% was then created using 1.6 ml of 20%, 15%, 10%, 5% and 0% OptiPrep mixtures. The samples were centrifuged in a Beckman ultracentrifuge for 90 min at 52 000 × g with rotor SW41. After centrifugation 0.66 ml fractions (18 in total) starting from the top of the tube were collected and 150 μl were used for western blot analysis. Optical densities of resultant bands were quantified using ImageJ software (freeware from the NIH) as described [bib_ref] Trafficking of HIV-1 RNA is mediated by heterogeneous nuclear ribonucleoprotein A2 expression..., Levesque [/bib_ref].
## Cholesterol depletion
For the complete depletion of cellular cholesterol in live cells, cells were first stained with Vybrant Lipid Raft Labeling Kit (Invitrogen). Coverslips were then mounted in Bioptechs environmental chamber and by perfusion DMEM medium was exchanged with DMEM containing 30 mM 2-hydroxypropyl-β-cyclodextrin (HβCD) [bib_ref] Lipid rafts and HIV pathogenesis: virionassociated cholesterol is required for fusion and..., Liao [/bib_ref]. Pictures were taken at t = 0 min before addition HβCD and at different time points after addition as indicated in the figures.
## Plasmid expression constructs and sirnas
The plasmids pMS2-Venus, pMS2-VN, pStaufen1-VC, pIMP1-VC, pFMRP-VC, pGL3-basic/site and pGL3-βActin zipcode have been described previously [bib_ref] Visualization of RNA-protein interactions in living cells: FMRP and IMP1 interact on..., Rackham [/bib_ref]. To generate pMS2-Staufen1, pMS2-humanized(h)Gag and pMS2-Gag(C36S) vectors Staufen1, hGag and Gag(C36S) sequences were amplified by PCR from pcDNA-Staufen1-TAP [bib_ref] The composition of Staufen-containing RNA granules from human cells indicates their role..., Villace [/bib_ref] , pCMV55M1-10 [24] and pNL4.3(C36S) [bib_ref] Noninfectious human immunodeficiency virus type 1 mutants deficient in genomic RNA, Gorelick [/bib_ref] and replaced the VN sequence from pMS2-VN between XhoI and NotI. To generate phGag-VenusC and pGag(C36S)-VenusC, hGag and Gag(C36S) (amplified from pNL4.3(C36S) [bib_ref] Noninfectious human immunodeficiency virus type 1 mutants deficient in genomic RNA, Gorelick [/bib_ref] PCR fragments replaced IMP1 sequence from pIMP1-VenusC between NheI and XhoI. The Rev-responsive Gag expressor, pSV-GagRRE-R was generously supplied by David Rekosh (University of Virginia, U.S.A.; [bib_ref] Requirements for incorporation of Pr160gag-pol from human immunodeficiency virus type 1 into..., Smith [/bib_ref].
For the generation of mRNA reporter expressing plasmids pGL3MS2site-psi or pGL3MS2site-Delta-psi the fragments containing vRNA packaging signal psi (52 bp) and delta psi (33 bp) were amplified from plasmids HxBRU and plasmid pHXBAP1 [bib_ref] The double-stranded RNA-binding protein Staufen is incorporated in human immunodeficiency virus type..., Mouland [/bib_ref] and cloned in pGL3MS2site/basic vector between NheI and XhoI. All plasmids were purified using the Sigma GeneElute Maxiprep kit. pCMV-Rev was obtained from the National Institutes of Health AIDS Research and Reference Reagent Program. The complete list of plasmids used in this work is given in [fig_ref] Table 1: List of plasmids used in this study [/fig_ref].
siRNAs used to deplete Staufen1 (targeting both isoforms -55 and 63 kDa, siStaufen) or the higher molecular weight isoform (Staufen1-63 kDa) were described earlier [bib_ref] Identification of Staufen in the human immunodeficiency virus type 1 Gag ribonucleoprotein..., Chatel-Chaix [/bib_ref] [bib_ref] Novel Staufen1 ribonucleoproteins prevent formation of stress granules but favour encapsidation of..., Abrahamyan [/bib_ref].
# Additional material
[fig] Figure 1: Gag interactions with host proteins Staufen1 and IMP1 occur in the cytoplasm and at the plasma membrane of transfected HeLa and Jurkat T cells as determined by BiFC. (A) Top -schematic representation of BiFC method. Bottom -Rev-dependent Gag-VN and Gag-VC were co-transfected with pCMV-Rev in HeLa cells. At 24 hr post-transfection, cells were imaged by laser scanning confocal microscopy to detect BiFC. The white arrows indicate plasma membrane concentrated accumulations of Gag-Gag BiFC signals. (B) Gag-VN and Staufen1-VC (top panels) or Gag-VN and IMP1-VC (bottom panels) interactions identified by BiFC. BiFC signals for these interacting pairs were mainly detected in the cytoplasm (indicated by white arrows) and at or near the plasma membrane. (C) Interactions between Gag-VN with IMP1-KH(1-4)-VC (top) and with IMP1-RRM(1-2)-VC (bottom) as determined by BiFC analysis. Evidence for interaction is demonstrated by a green fluorescence signal. (D) The interaction between Gag-VN and Gag-VC (top) or Gag-VN and Staufen1-VC (bottom) was determined by BiFC in Jurkat T cells. Magnified sections demonstrate details on the shapes of BiFC signals/complexes. The size bars are equal to 10 μm. [/fig]
[fig] Figure 2: Interactions between Gag and cellular proteins Staufen1 or IMP1 at GM1 containing lipid rafts on the plasma membrane as determined by BiFC. (A) HeLa cells were co-transfected with pCMV-Rev and Rev-dependent Gag-VN and Gag-VC plasmids. At 24 hr post-transfection lipid raft staining in live cells was performed. Images were captured using laser scanning confocal microscopy to detect the co-localization patterns of oligomerizing Gag molecules and lipid raft microdomains (indicated by CT-B that binds the pentasaccharide chain of the raft marker protein, GM1). (B) Gag-VN/Staufen1-VC BiFC signals and CT-B staining in live cells. (C) Gag-VN/IMP1-VC BiFC signals and CT-B staining in live cells. (D) HeLa cells or (E) Jurkat T cells were co-transfected with pCMV-Rev and Rev-dependent Gag-VN and Gag-VC plasmids. At 24 hr post-transfection the cells were fixed in 4% paraformaldehyde (T cells were attached to poly-D-lysine coated coverslips before fixation), permeabilized in 0.2% Triton and stained for endogenous Staufen1 and p17 to detect Gag (in Jurkat T cells only; (E), Gag is presented in blue). BiFC signals also identify Gag-Gag oligomers in fixed cells. Magnifications of cells on right show endogenous Staufen1 in Gag-Gag BiFC-negative (box 1) and positive (box 2) HeLa (D) or Jurkat T (E) cells. The size bars are equal to 10 μm. [/fig]
[fig] Figure 3: Staufen1 co-fractionates with Gag in lipid rafts. (A) A detergent-free method for the isolation and fractionation of lipid rafts. HeLa cells were mock transfected (B) or co-transfected with pCMV-Rev and Rev-dependent Gag-VN and Gag-VC (C) or Rev-dependent GagΔNC/p6 (D). At 24 hr post-transfection, cells were harvested and fractionated on Optiprep gradients: lipid rafts fractionated in fractions #2-6 and non-membrane associated proteins in the bottom fractions. Western blotting identified Caveolin-1 (Cav-1), Staufen1 (two isoforms: St-55, St-63), IMP1, TSC2, Gag-VC (in C) and p24 (to detect GagΔNC/p6 in (D)) in each fraction. (E) The relative quantities of Staufen1 in each fraction were measured using ImageJ software (NIH) (the sum of all fractions per condition = 1.0). [/fig]
[fig] Figure 4: Time-dependent depletion of cholesterol from lipid rafts leads to the disruption of Gag-Gag, Gag-Staufen1 or Gag-IMP1 BiFC. (A) HeLa cells were co-transfected with pCMV-Rev, Gag-VN and Gag-VC. At 24 hr post-transfection the cells were stained with the Vybrant Lipid Raft Labeling Kit and treated with cholesterol disrupting drug HβCD (final concentration 30 mM). Pictures were taken at the indicated times post-HβCD treatment. (B) HeLa cells were co-transfected with pCMV-Rev and the Rev-dependent Gag [46] and at 24 hr post-transfection were treated with HβCD for different periods of time (as indicated, for up to 45 min). The cells were then fixed, stained for Caveolin-1 and Gag and visualized by laser scanning confocal microscopy. (C) Hela cells were co-transfected with pCMV-Rev and either Gag-VN and Staufen1-VC or Gag-VN (top panels) and IMP1-VC (lower panels). At 24 hr post-transfection lipid rafts were identified in live cells using the Vybrant Lipid Raft Labeling and treated with HβCD for up to 25 min. The BiFC signals were determined at t = 0 (not shown) and at the latest time point of t = 25 min. The size bars are equal to 10 μm. [/fig]
[fig] Figure 5: Staufen1 depletion decreases plasma membrane-associated Gag and results in intracellular clustering of Gag-Gag BiFC signals. HeLa cells were co-transfected with pCMV-Rev, Gag-VN and Gag-VC plasmids with control non-silencing siRNA (siNS), Staufen1 siRNA (siStaufen1) or Staufen1-HA. At 24 hr post-transfection cells were harvested for western blotting for Staufen1, Gag, Caveolin-1 and TSC2 (as loading controls) (A) or stained for lipid rafts in live cells. BiFC signals and lipid raft (CT-B) staining were captured by laser scanning confocal microscopy in cells transfected with siNS (B), siStaufen1 (C) or Staufen1-HA (D). Black and white images of lipid rafts (CT-B) and Gag-Gag BiFC signals and merged colour representations are shown. The insets are magnifications of boxed areas. The size bars are equal to 10 μm. [/fig]
[fig] Figure 6: Gag localizes near Rab7-, Rab9-and LAMP1-containing membranes at cytoplasmic and juxtanuclear sites in Staufen1-depleted cells. HeLa cells were transfected with pCMV-Rev, Gag-VN and Gag-VC with either siNS or siStaufen1 siRNAs and one of the following plasmids: (A) Rab5-RFP, (B) Rab7-RFP, (C) Rab9-RFP, (D) LAMP1-RFP or (E) Caveolin-1-RFP. At 24 hr post-transfection, the distributions of Gag-Gag BiFC and RFP fusion proteins were visualized in live cells by laser scanning confocal microscopy. The insets show zoomed boxed regions of cells to demonstrate the levels of co-localization of Gag-Gag BiFC signals with either of membrane marker proteins. White arrows identify Gag-Gag BiFC aggregates. The size bars are equal to 10 μm. [/fig]
[fig] Figure 7: TriFC analysis for studying RNA-protein interactions in live cells. (A) Depiction of the TriFC analysis employed in the study. The mRNAreporter molecule contains HIV-1 vRNA packaging signal psi in close proximity to the integrated MS2 RNA-binding site (MS2BS). MS2-VN binds MS2BS to tether to the mRNA molecule. The C-terminal moiety of Venus (VC) is expressed as hGag-VC fusion and binding of hGag-VC to the RNA packaging sequence (psi domain) will bring both VN and VC Venus moieties in close enough proximity to generate a fluorescent signal by VN-VC complementation. (B) A 19-basepair deletion from SL3 of psi prevents the binding between hGag-VC and psi RNA and does not yield fluorescence complementation. (C) HeLa cells were grown on coverslips and transfected with pGL3MS2site-psi, MS2-VN and Gag-VN. [/fig]
[fig] Figure 8: Gag recruits host proteins Staufen1 and IMP1 while tethered to mRNA. HeLa cells were co-transfected with mRNA-reporter pGL3-basic/site, MS2-VN, MS2-Gag or MS2-Staufen1 and hGag-VC, Staufen1-VC or IMP-VC as indicated. TriFC analysis was performed at 24 and 40 hr post-transfection. MS2 RNA-tethered hGag recruits Gag (A), Staufen1 (B) and IMP1 (C) to generate TriFC signals in cells. MS2 RNA-tethered Staufen1 recruits hGag (D). Western blotting analysis for Staufen1, p24 (to identify Gag), and GFP (that recognizes C terminal part of Venus) and Luciferase (reporter mRNA, pGL3-basic/site expressed Luciferase) confirmed expression of DNA constructs. Negative controls included transfections that omitted either the VC fusion proteins (bottom left in panels (A)-(D)) or the bridging MS2 molecule (MS2-hGag, MS2-Gag(C36S) or MS2-Staufen1; data not shown). The size bars are equal to 10 μm. [/fig]
[table] Table 1: List of plasmids used in this study. [/table]
|
Subsets and Splits