text
stringlengths 66
377k
|
|---|
Current standards and ethical landscape of engineered tissues—3D bioprinting perspective
Tissue engineering is an evolving multi-disciplinary field with cutting-edge technologies and innovative scientific perceptions that promise functional regeneration of damaged tissues/organs. Tissue engineered medical products (TEMPs) are biomaterial-cell products or a cell-drug combination which is injected, implanted or topically applied in the course of a therapeutic or diagnostic procedure. Current tissue engineering strategies aim at 3D printing/bioprinting that uses cells and polymers to construct living tissues/organs in a layer-by-layer fashion with high 3D precision. However, unlike conventional drugs or therapeutics, TEMPs and 3D bioprinted tissues are novel therapeutics and need different regulatory protocols for clinical trials and commercialization processes. Therefore, it is essential to understand the complexity of raw materials, cellular components, and manufacturing procedures to establish standards that can help to translate these products from bench to bedside. These complexities are reflected in the regulations and standards that are globally in practice to prevent any compromise or undue risks to patients. This review comprehensively describes the current legislations, standards for TEMPs with a special emphasis on 3D bioprinted tissues. Based on these overviews, challenges in the clinical translation of TEMPs & 3D bioprinted tissues/organs along with their ethical concerns and future perspectives are discussed.
# Introduction
Tissue engineered products including naturally derived or synthetic biomaterial-based scaffolds with/without autologous or allogeneic cells are utilized to replace or restore the functions of damaged tissues and organs.Advanced therapy medicinal products (ATMPs) are considered as medicines for the treatment of disease or injuries in humans using genes, tissues or cells. ATMPs are classified into categories such as gene therapy medicinal products (GTMP), cell therapy medicinal products (CTMP), tissue-engineered medicinal products (TEMPs) and sometimes a combination of these categories.In some countries like the US, a group of biological medicinal products are categorized as Human Cells, Tissues, and Cellular and Tissue-based Product (HCT/P).ATMPs have several applications in the clinical arena to improve the quality of patients life. Organ transplantation is the terminal-stage Current standards and ethical landscape of engineered tissues-3D bioprinting perspective treatment strategy for conditions like myocardial infarction, acute liver failure, chronic kidney disease, and Type 1 diabetes, etc. and may replace the need for TEMPsHowever, shortage of organ donors, high cost and immunological complications limit the affordability and clinical success of organ transplantation around the globe.Thus, there is a need for artificial organ fabrication through tissue engineering approaches to overcome the limitations of organ transplantation.Although various approaches have been used to develop engineered tissues, only a limited number of TEMPs are approved and commercialized for clinical uses. This may be attributed to the fact that engineered tissues have high risks, more uncertainties in safety, efficacy and very expensive for commercialization.TEMPs are the latest therapeutics that involve one or more complex manufacturing processes, varied constituents and different characteristic features, which demands unique standards and regulations for approval processes.Global regulatory agencies evaluate the quality, safety, efficacy and cost-effectiveness of TEMPs for healthcare applications and regulate through various steps involved from commercialization to clinical practice.Tissue engineered medical products are regulated based on exclusively framed legislations in different countries. These legislations have the same set of objectives and rules but follow different regulatory terms/approval processes for commercialization.TEMPs submitted under clinical transformation approval need to be manufactured in Good Manufacturing Facilities (GMP) setup using clinical-grade raw materials with defined Quality Attributes (QAs) and also need to be tested clinically with Good Clinical Practice (GCP) guidance.Regulatory authorities of medicinal product approval generally focus on the criteria mentioned above and more specifically, on its benefits and risk involvement.In the current decade, TEMPs fabricated through conventional tissue engineering approaches have a wide range of potential applications, such as congenital heart disease (CHD), heart valve replacement, bone fracture healing, severe burn injuries, cartilage defect, acute/chronic liver problems, etc., are now considered as a safer and efficient clinical solution.shows the classification of ATMPs with emphasis on TEMPs, which falls under the main scope of this review. Further, bottom-up and reverse engineering approaches for fabricating patient-specific three-dimensional cellular/biomaterial scaffold with autologous cells (stem cells/other cells) using 3D bioprinting technique has become a better strategy in fabricating organs with biomimetic geometries and other physiological features of native tissues without any immune complications.3D printing has greatly evolved as an efficient technology especially for various applications in regenerative medicine. The progress and applications of 3D printing and bioprinting in regenerative medicine were broadly represented in. Hence, computerized fabrication of multi-functional human 3D organs could facilitate the ease of organ transplantation and 3D tissue/organ model fabrication for pre-clinical drug/biological testing in the future.
From the above perspectives, this review mainly focuses on all the legislations, regulations, guidelines and standards followed worldwide for clinical trial approval and commercialization of the tissue engineered products. Further, the current tissue fabrication process with the advancement of additive manufacturing approaches for engineering personalized 3D printed tissues, components & types of 3D bioprinting, and the potential of bioprinting technology to fabricate tissues/organs for transplantation applications are also discussed. Additionally, the characteristics of bioprinted constructs, their applications and how they stand in comparison with tissue engineered medical products from an ethical and regulatory standpoint are elaborated. Further insights into the regulatory aspects of tissue engineered and bioprinted organs/tissues that greatly hamper the widespread clinical translations are discussed. Additionally, the ethical issues revolving around tissue engineered and bioprinted products and challenges in commercial success are also discussed. Finally, future perspectives of tissue engineered constructs are discussed and compared with the projected future of 3D bioprinting in regenerative medicine applications.
## Tissue engineered medical products (temps)
Tissue engineering (TE) is an interdisciplinary field that employs the principles of life sciences and engineering to restore and regenerate the physiological activity of damaged tissues or organs. This has led to the rapid development of tissue engineered medical products (TEMPs) using conventional tissue engineering strategies for diagnostic and therapeutic applications in various tissues such as skin, cartilage, bone, blood vessels, heart valves, etc., using components such as cells, scaffolds, biomolecules, processed tissues and their derivatives.Notably, there are several TEMPs for cartilage defects and skin substituents which are currently in phase II and III clinical trial stages and so far provided the confirmations on therapeutic value, regenerative activity, safety and longterm biological effects.Moreover, the increase in clinical demands for organ transplantation, specifically for aging populations has propelled the exponential need for TEMPs on a larger scale and at affordable prices. However, limitations in conventional fabrication techniques such as lack of three-dimensional architecture, cellular positioning at the desired locations, variable cellular density, template requirement, difficulty in complex shape fabrication suggest the usage of advanced fabrication techniques to engineer tissues or organs for transplantation applications.An additive manufacturing approach: 3D printed TEMPS 3D printing has now gained huge attention from many users of different fields due to the layer-by-layer deposition, feasibility in fabricating complex models, ease of operation, low cost and multi-material deposition.Various 3D printing techniques have been employed to develop patient-specific scaffolds and tissues using polymers such as polylactic acid (PLA), titanium, ceramics, polycaprolactone (PCL), polyurethane, etc. and most of them are evaluated for durability and functionality in both in vitro and in vivo studies for tissues such as bone, heart valves, skin, intestines, etc.Currently in clinical surgery, 3D printing technology plays a major role by providing different technical and visual or physical support in the form of pre-operative surgical guides (cutting/drilling/ planning), surgical tools, custom specific implants and prostheses. 28 3D printing offers a wide range of materials selection approach to create 3D structures as support or major implant for spinal surgery, maxillofacial, cranial, dental, orthopedic surgery, etc. Undoubtedly, for all these surgical approaches, medical surgeons collaborate with additive manufacturing based medical device companies or bioengineers to assist in the design and 3D printing for specific requirements. Additionally, 3D planning software makes patient-specific virtual surgery model to provide the clear-cutting plane and drilling trajectories to limit damages for nerves and blood vessels and improves positioning accuracy to place the implant.Subsequently, the printing of different sized models can help to overcome the limitation of using cadaver-based surgical planning where specificity and availability are a major issue.In a recent study of tissue regeneration to mimic native tissue structure, 3D printed bone graft made from PCL impregnated chitosan loaded with rabbit bone marrow stem cells showed improved differentiation activity and increased expression of osteogenic specific genes such as alkaline phosphatase (ALP), collagen type I (COL1), osteocalcin (OCN), and Runt-related transcription factor (RUNX2) after 14 days in vitro. This graft was also subcutaneously implanted in nude mice and observed that the 3D printed scaffold exhibited stronger osteogenesis and bone-matrix formation after 3 weeks of surgery.
## 3d bioprinting of tissues and organs
3D bioprinting technique is a 3D tissue fabrication technique where the cells are integrated into a cross-linkable hydrogel matrix called bioink to create 3D tissue equivalent constructs in the desired pattern.3D bioprinting requires essential components such as 3D imaging, CAD/ CAM software, bioink, and bioprinter to carry out the fabrication process. Bioink is a combination of cells, polymers (biomaterials) and signaling molecules like growth factors with adequate viscoelastic and cell supportive functionalities that are suitable for the printing of tissues and organs in a bioprinter platform.Further, the development of novel biomaterials or chemical modification of existing materials helps to customize bioinks for tissuespecific applications. Moreover, bioinks are deposited in a layer-by-layer manner at desired locations to enable the fabrication of vascularized 3D structures for tissue or organ regeneration applications. Bioink made of fibrous proteins such as collagen and fibrin contains more than 90% water, allowing them to create cell supportive and biocompatible tissue constructs.Natural polymers such as collagen, Matrigel, gelatin, alginate, agarose, methylcellulose, fibrinogen and synthetic polymers such as pluronics, carbopol, nanoclay, hydroxyapatite and poly(Nisopropylacrylamide) have been used as bioinks for the fabrication of various tissues like cardiac patches, bone, cartilage, and cornea.Bioprinters capable of achieving ideal print speed, human-scale resolution, and also equipped to operate multi-material print heads could be beneficial in the fabrication of patient-specific tissues or organs. Bioprinting is divided into three types based on the bioink dispensing mechanism such as inkjet bioprinting, laser-assisted bioprinting and extrusion-based bioprinting. Inkjet bioprinting requires less viscous bioinks to dispense through micron-sized nozzles either by thermal or piezoelectric stimulus.Laser-assisted bioprinting methods employ focusing the laser pulse on the metalcoated plate with low-medium viscosity bioinks to create 3D tissues.These two printing systems have good printing resolution (20-40 µm) and cell viability (>90%). However, these methods are not widely preferred for fabricating human scale 3D tissues due to the difficulties in layer-by-layer stacking ability beyond a particular build volume. Extrusion bioprinters can dispense a variety of bioinks with a wide range of viscosities either through pneumatic or mechanical screw-based mechanisms. In recent years researchers have successfully fabricated various 3D tissues and organ models as a proof-of-concept using extrusion bioprinting.Further, the features and limitations of these various bioprinting techniques such as inkjet, extrusion and laser-assisted bioprinting methods are briefly tabulated in. Printing of multiple cell types with variable celland Ding et al.Koch et al.and Guillotin et al.Lee et al.and Gaetani et al.densities (1 × 10 6 cells/mL to 1 × 10 8 cells/mL) at desired locations made 3D bioprinting one of the most advanced techniques for tissue or organ fabrication.A comprehensive literature review of 3D tissues/organs fabricated using different bioprinting strategies for tissue regeneration application has been tabulated in.
## Regulatory roles and responsibilities of approving authorities
The main goal of regulatory agencies is to produce laws and regulations by following different regulatory frameworks to allow safer medicinal products for clinical trials and marketing purposes. Globally tissue-based products (TEMPs) are identified differently based on the categorizing condition for the regulatory approval. Likewise, TEMPs have been considered as drugs in the EU, a medical device in Japan and biologic/combined products in the U.S through their regulations/directives and further their approval pathways are decided by assessing the quality, safety and efficacy through nonclinical and clinical studies.The regulatory agencies of different countries that deal with clinical trial approval and permission for commercialization are listed in . Identically, Gene and Cell-based Therapies (GCTs) regulatory frameworks are outlined to assess product characteristics, evidentiary and non-evidentiary reasons for approval and post-marketing risk management studies.It is identified that clinical translation of products is approved based on confirmed evidential benefit results for US and EU applications while regulations of Japan follow nonconfirmed beneficial results. US, EU and Japan regulatory agencies are now approving Gene and Cell-based Therapies (GCTs) with scientific uncertainties and safety risks to allow the clinical transformation based on medical needs.Oftentimes, approval processes for new products are based on the standards available on the technical details of preexisting relevant products.
PROVENGE was the first commercialized cell therapy treatment for prostate cancer treatment in the United States and Transcyte for third degree skin replacement graft was the first FDA approved tissue graft in 1997.Activskin, a skin graft was the first approved TEMP in China by 2013.In 2001, the first cell therapy and tissue engineered product (chondrocyte-based product) was successfully commercialized in South Korea. Relatively, South Korea has more number of cell therapies in clinical practice as on 2018 status. 1 Totally 10 ATMPs were approved by the European Medicines Agency (EMA) in the Europe Union and likewise, the first commercialized tissue engineered product in Europe is ChondroCelect (for treating cartilage defects) in 2009, but this product was discontinued in 2016 due to commercial reason from its marketing authorization TiGenix NV.As discussed earlier, in comparison with the last decade, more tissue-based therapies are approved by the authorities in consideration to produce these high therapeutically efficient TEMPs for the unmet medical needs.As a result, several products for skin, cartilage, bone, heart, neural and several other complex organs have been commercialized in recent times.
## Current regulations for temps
Recently updated tissue engineered product regulations, approval requirements and rules enforced so far in U.S, India and European Union for the currently available TEMPs are shown in .
US based regulations. Unites States (US) authorized Food and Drug Administration (FDA) is a regulatory agency, which enforces Food, Drug and Cosmetic Act (FDCA), Kefauver-Harris Amendments, Medical Device Amendments and Public Health Service Act (PHS) for rules and conditions to evaluate product safety and efficacy in the United States of America (USA) medical care.In this current decade, to utilize the safer health care innovations with an involvement to technically advance innovative medicinal therapy, several expedited approval pathways, alternative pathways with exemptions and special situation based decisions have been structured and enacted into legislations for further approval at a short period of time.Recently FDA guidance expedited program includes Fast Track designation and Breakthrough Therapy designation were created through regenerative medicine advanced therapy (RMAT) designation program by 21st Century Cures Act for the sponsors of regenerative medicine products.The regulatory criteria for human cells, tissues, and cellular & tissue-based products (HCT/P) issued by the FDA of the US are followed by the Center for Biologics Evaluation and Research (CBER) under section 361 of the Public Health Service Act (PHS) provided in 21 CFR (Code of federal regulations) Parts 1270 and 1271 regulations.Additionally, for some HCT/Ps not found in the mentioned criteria of 1271.10(a), it could be regulated as drugs, devices or biologics under section 351 of the Public Health Service Act (PHS) and Federal Food, Drug, and Cosmetic Act (FD&C 23Act).Due to the interdisciplinary landscape, HCT/Ps products with some exemptions such as cryopreserved femoral vein for AV shunt, cultured cells on biomaterial/decellularized scaffolds are regulated as a medical device with tissues or combination products through Center for Devices and Radiological Health (CDRH).These regulations include HCT/P criteria for minimal manipulation (possibilities of processing method to induce alteration characteristics of the cells or tissue to replace/restore the function), homologous usage (without the combination of other articles which could give rise to new clinical application), its ability of with or without systemic effect having a dependency on cells metabolic activity.Moreover, this can also regulate the usage of only FDA-registered resource materials for medical application. Further, it also states that manufacturers of HCT/P should register and list the product for approval from the authorities for the license, permission of preclinical, clinical data approval evaluation. To facilitate updating, electronic submission is also possible using electronic HCT/P establishment registration (eHCTERS) system with all the submission of requirements. FDA is also providing special protocol assessment (SPA) guidance by the Center for Drug Evaluation and Research (CDER) and the Center for Biologics Evaluation and Research (CBER) to help in study designs animal studies and clinical studies and trials. Under section 505(b) (5)(B) of the FD&C Act, under SPA, drug stability or animal efficacy protocols should apply for approval with study designs and statistical analysis. Submission of SPA request is a process that involves (i) Informing FDA of an upcoming request; (ii) Timing of a request for review process up to 45 days which includes documents submission and resubmission of additional required documents; (iii) The request format should contain the cover letter with bold block letters mentioning "REQUEST FOR SPECIAL PROTOCOL ASSESSMENT"; (iv) The SPA should submit to the appropriate CDER or CBER division, using standard or electronic submission. FDA will start analyzing the submissions to see whether it is suitable under SPA and communicate decision by mail within 45 days of the review timeline.EU based regulations. In the EU (European Union), classification of tissue engineered products are clearly explained and regulated well and Regulation (EC)No 1394/2007 was designed and amended on December 30, 2008 for Advanced therapy and medicinal products (ATMP including cell therapy, gene therapy and tissue engineered medicinal products) for evaluation of product quality, efficacy and safety.These directives state that the applicants producing product specifications should provide the product details clearly including, name, composition, quantitative & qualitative details, clinical particulars, its interaction with other substance & molecule, precautions steps if any, usage during pregnancy & lactation details, pharmacological, pharmacodynamics & pharmacokinetic properties, pre-clinical safety data, shelf life, storage condition and its procedure, marketing details, etc. In addition, packaging should mention the marketing authorization, manufacture batch number 2004/23/EC, expiry date, method of use, special warnings, etc.Further, all the clinical data analysis, proper GMP and GLP activities are expected. Particularly, conditional approval has been described by the regulation (EC)N 507/2006 with some difference in the regulatory framework of EU. 17 EU regulatory implements adaptive approach regulatory pathways mostly for the cellular and tissue-based products. Besides, other pathways are also there such as standard pathway, conditional pathway, adaptive licensing and PRIority MEdicines (PRIME) are framed to regulate products of different categories.India based regulations. In India, Drug and Cosmetic Act 1940 and Rules 1945 were enacted to regulate drug approval by Central Drugs Standard Control Organization (CDSCO). 66 Especially Cell Biology Based Therapeutic Drug Evaluation Committee (CBBTDEC) by CDSCO was formed in 2010 for cell therapy related clinical approval.Additionally in India, the Indian Council of Medical Research (ICMR) involve in providing guidelines for conducting biomedical and clinical research activities for clinical regulation. Briefly, the Drugs & Cosmetics Act 1940 was legislated by Indian law in concern with the importing, manufacturing and marketing drugs and cosmeticsIn China, the regulatory approach is mainly focused on the final product utility in treating the medical need and it does not mainly depend on the raw material and processing method. A cohort study conducted by Coppens on the approval of GCTs between 2008 and 2017 by US, EU and Japan regulatory concluded that Japan showed higher acceptance of GCT in consideration with uncertainties and safety risks followed by EU and US. In the majority of cases, unmet medical needs are considered for its regulatory approval by the EU and Japan clinical transformation regulatory system. While considering post-marketing characteristics, product safety, quality and risks are analyzed in Japan and EU, but in US safety is the main focus.However, US and EU have many alternative pathways and regulations to facilitate the approval process relatively faster. Japan follows time limited approval pathways to overcome some of the already existing limitations.Regulations for 3D printed/bioprinted TEMPs 3D printed (additive manufactured) products in medical applications are fabricated as whole device/part which could be utilized as prosthesis or an implant or other medical assistance devices.In recent times, 3D printed prostheses for hip, knee, skull, jaw bone or joint implant, limb prostheses, orthopedic implants, heart valves, etc., are developed and currently under extensive clinical research. These 3D printed prostheses have the potential to create personalized implants of complex structures with precisely controlled material and structural properties within short time when compared to other implant fabrication methods.Further, additive manufactured metal implants have added advantage in controlling internal porous structures to promote biological fixation with long-term stability. Although 3D-printed implants (including metal implants) are devoid of cellular materials, they are also regulated as medical devices.However, bioprinted TEMPs with complex cell source and bioink composition for regenerative medicinal application could be regulated either as biologics or drugs/medical devices based on the regulatory authorities. 82 3D bioprinted tissues fall between the categories of living materials & technology and hence do not fall directly into the existing categories of regulations. In order to bring this to light, experts have developed a concept called "bio-objects," which means that all kinds of biotechnologies fall "in-between" the existing categories of living and non-living matter.These "bio-objects" do not fall into the standard regulatory frameworks which are based on the clinically wellestablished medicinal product categories. Consequently, these products require new regulations/laws for clinical trials and commercialization.
In the US, personalized 3D printed medical devices are regulated under the medical device category or with some custom device exemptions to evaluate their safety and efficacy through pre-market/post-market requirements.However, generic medical devices are categorized into 16 special domains (based on their usage and risk) and each device is considered among one of the three regulatory classes based on its safety level and efficacy. These medical devices are classified through de novo classificationa risk-based classification for the regulations as mentioned in. Based on these categorical classifications, 510 k regulation is required for class I and II devices while for class III, premarket approval application (PMA) is required for the FDA approval process. All these classification systems are subject to common control requirements of Food, Drug and Cosmetic Act (FDCA) and additional market approval as a design control model from regulatory agencies.The FDA has provided guidelines for the 3D printed materials under two different considerations such as Design and Manufacturing Considerations (device design/ patient match design, software process, material, printing process & control, post-process and validation) and Device Testing Considerations (device description and measurements, material characterization and mechanical testing) to ensure product quality, efficacy and determination of classification for regulations.Currently, 3D printed medical devices are regulated through premarket notification, New Drug Application (considering as new drug) and Biologics License Application (considering as biologics).The FDA has so far approved the AXIOM 20 3D printer for manufacturing medical devices and one 3D printed drug named Spritam ® tablets for epilepsy treatment.On the other hand, EU follows legislation such as AIMDD 90/385/EE (Active Implantable Medical Device Directive), MDD 93/42/EEC (Medical Device Directive) and IVDMDD 98/79/EC (In Vitro Diagnostic Medical Device Directive) for 3D printed medical device products. Further, as per the standards mentioned in directive 93/42/ EEC, medical devices are classified into four categories such as Class I (Noninvasive devices), Class IIa, Class IIb and Class III ranked from lowest to higher risk. Higher rank classes require higher safety assessment levels and such classification levels are decided based on the consideration of patient contact duration, degree of invasiveness and place of the implantation/contact in human body part. These rules have been followed by effectively implementing as a set of 18 rules presented in. A titanium based (Ti-6Al-4V) 3D printed implant called iFuse-3D is clinically used as structural support for sacroiliac joint to promote bone growth. This implant has received market clearance under conventional medical devices regulations in both US and EU. Post-market evaluation of this implant has assessed patient safety and efficacy of the product similarly to a machined implant. The clinical outcomes had revealed that iFuse-3D reduced the pain related complaints, thereby avoided secondary corrective surgeries. In addition, this implant did not show any unanticipated clinical complaints after treatment and hence is efficient in creating patient-specific implants using 3D printing technology.Likewise, Ackland et al. fabricated personalized 3D printed prosthetic temporomandibular joint (TMJ) for a 58-year female patient. This prosthetic was modeled using patient musculoskeletal modeling for assessing implant stress and strain, applied load, screw stress and other physiological loadings. Further, the modeled prosthetic was analyzed, fabricated and finally implanted into the patient to study its efficacy. It was observed that a normal jaw opening distance of 40.0 mm was achieved with less pain after 6 months of postoperative surgery. These results demonstrate the effectiveness of personalized 3D printed complex joint replacement prosthetics developed through several modeling and computational analysis.
## Universal standards for tissue engineered products
Standards for TEMPs are established to ensure product characteristics and all associated analytical procedures to ensure common quality results worldwide with improved repeatability and reliability of the data. The International Organization of Standards (ISO) and the American Society for Testing and Materials (ASTM) International are the recognized organizations for presenting standardization of biomaterials and medical devices.ASTM International is an organization that provides resources, technical expertise to develop universal standards and guidelines for a wide range of developing fields. It comprises several committees and subcommittees with respect to specific technical areas to frame the standards and guidelines including material characteristics, test methods, system, etc. Specifically for tissue engineered medical products, the F04 committee on "Medical and Surgical Materials and Devices" with other subcommittees named Division IV "Tissue-Engineered Medical Products (TEMPs)" was established in 1997 and the up to date standards published by these committees.Further, ASTM has established the F42 committee on "Additive Manufacturing Technologies" in 2009 with eight technical committees to formulate standards for 3D printed medical products.Similarly, ISO comprises a Technical Committee (TC) and Sub Committee (SC) to provide standards and guidelines for different fields. In particular, for fields of tissue engineering and regenerative medicine, technical committee 150 "Implants for surgery" with subcommittee SC7 "Tissue-engineered medical products," which was established in 2001 covers TEMPs based on published standards (International organization of standards/Technical committee/Subcommittee ISO/TC 150/SC7). Other technical committees such as TC194 (Biological and clinical evaluation of medical devices), TC210 (Quality management and corresponding general aspects for medical devices), TC212 (Clinical laboratory testing and in vitro diagnostic test systems), TC261 (Additive manufacturing) and TC266 (Biomimetics) are established with several standards which are widely utilized for health care applications.Periodically, each committee could establish the new standards or revised version of published standards for evaluation. For example, TC150 has published 165 standards and TC 194 has published 32 ISO standards in 2020. An informative list of ASTM and ISO standards followed by world regulatory authorities for commercializing medical products are tabulated. These standards and updated protocols could help to establish a uniform approach and assessment procedure for validating medical products including tissue engineering products while ensuring high quality and more safety to humans.This is imperative because cellular and acellular synthetic/natural constructs need specific standards for characterization and in vitro, in vivo and clinical trial analyses with respect to the raw material used.TC 261 published ISO/ASTM 52910:2018 standard "Additive manufacturing-Design-Requirements, guidelines and recommendations," which describes the design consideration that needs to be taken while manufacturing different types of products or components through the additive manufacturing approach. This standard will provide necessary guidance to the engineers and students to carry out the basic design of additive manufactured products for clinical translation. These design guidelines facilitate the manufacturers to decide on optimal cost, quality and delivery time. Further, this could also be applied to identify the potential of additive manufacturing techniques by assessing the choice of material, its availability, build volume and print volume comparison.ISO in liaison with the governmental, non-governmental and international organizations have also been facilitated to further raise the standards with more expertise in the field. For instance, ISO 261 and ASTM F47 have jointly framed general standards (framing additive manufacturing requirements, guidelines, safety, definition and concepts) and specific/broad category of material based standards (framing material, process or application details) to continuously improve worldwide standards and quality of additive manufactured products without any confusionsThis also enables the development of Prosthetics and orthotics-terms relating to the treatment and rehabilitation of persons having a lower limb amputation ISO 8549-1:2020
Prosthetics and orthotics-vocabulary-part 1: general terms for external limb prostheses and external orthoses ISO 8549-2:2020
Prosthetics and orthotics-vocabulary-part 2: terms relating to external limb prostheses and wearers of these prostheses ISO 8548- Prosthetics and orthotics-limb deficiencies-part 1: method of describing limb deficiencies present at birth ISO 8548-2:2020
Prosthetics and orthotics-limb deficiencies-part 2: Method of describing lower limb amputation stumps ISO 8548- Prosthetics and orthotics-limb deficiencies-part 3: Method of describing upper limb amputation stumps ISO 8548- Prosthetics and orthotics-limb deficiencies-part 4: Description of causal conditions leading to amputation ISO 8548- Prosthetics and orthotics-limb deficiencies-part 5: Description of the clinical condition of the person who has had an amputation ISO 29783- Prosthetics and orthotics-vocabulary-part 1: normal gait ISO 29783-2:2015
Prosthetics and orthotics-vocabulary-part 2: Prosthetic gait ISO 8549-3:2020
Prosthetics and orthotics-vocabulary-part 3: terms relating to orthoses ISO 8549-4:2020
Prosthetics and orthotics-vocabulary-part 4: terms relating to limb amputation classification of standards) with specific description level (field, group, subgroup) with codes for better understanding and followed as required.
## Clinical research guidelines
International Organizations such as the World Health Organization, Council for International Organizations of Medical Sciences (CIOMS), US Department of Health and Human Services, National Institute of Health, Council of Europe and Nuffield Council on Ethics are working to provide guidance for the research on the health care field on ethics, safety and quality medical therapy/products development.The main objective is to ensure that guidelines should be clear, complete and should take into consideration of all regulatory aspects of the country. To ensure human health with safer medicinal products, several guidelines are being published in collaboration with a wide variety of technical experts of the field to provide all the detailed procedures and requirements with ethical concerns relevant to the particular topic.Additionally, an International standard for conducting clinical trials named "Good Clinical Practice" was issued by International Conference on Harmonization (ICH).Furthermore, clinical trials should be conducted under the ethical principles in the Declaration of Helsinki, earlier risk and beneficial assessment, rights, safety and well-being of participants analysis, planned data storage and with approved protocols. All the published, new and revised versions of the guidelines for tissue engineered product regulation of the US, India and European Union are highlighted inwith a brief description of the guidelines and their scope. The objectives of these regulations are based on providing maximum benefits with minimum risk and hence EC (Ethical Committee), researchers and other stakeholders are always dependent on benefitrisk assessments.These guidelines helps researchers, industrial manufacturers, marketing authorization members and other involved members about the regulatory need, approach and analysis. Importantly, these guidelines have to be regularly revised, modified and improved envisaging tissue engineered product commercialization in the future. On a similar note, ICMR in India formulated the first guidelines, "Ethical guidelines for biomedical research on human participants" in 2006 to conduct human trial research with proper safety measures and effective utilization of study for further applications. It states to conduct human research with four basic principles that include person's beneficence, non-maleficence, respect and justice to protect participant dignity, safety, rights and wellbeing. Despite various ambiguity and lack of clarity in 2006 guidelines, the following are some technical factors to be considered while revising in future:i. Possibility of allowing clinical trial for drugs, which are under approval stage by Drug controlled general of India (DCGI).
## Current status toward clinical translation of engineered tissue constructs
In a recent report, it was observed that upon the interest of Tissue Engineered (TE) product research & development, there have been 66 on-going or approved TE clinical trials (biomaterial/stem cell) between 2011 and 2018 in the United States, indicating a trajectory of TE product growth toward clinical testing.The global tissue engineering market report 2019 had concluded that North America (43%) and Europe (35%) holds the largest marketing shares of TE products and further reported that the worldwide tissue engineering market reached 14,000 million USD in 2018 and it is expected to reach 55,200 Million USD by 2025.On the other hand, bioprinted tissues have major obstacles toward clinical translation due to the degradation profile of the construct, immunogenicity of the bioinks, shape fidelity, durability and difficulty in the incorporation of vascularity and less affordability for the targeted applications.However, 3D bioprinted tissue constructs have several advantages and are gaining momentum toward effective ways to overcome these existing challenges. As an example, the average treatment cost of a single burn patient would be approximately around $88,000, whereas large-scale production of 3D bioprinted skin might cost less than this threshold.Another issue with the clinical practice of 3D bioprinted engineered tissues would be the timeline for developing matured tissues or organs with respect to patients need. Specifically, developing simple tissues such as tubular constructs (blood vessels, heart valves) and skin requires a shorter time, whereas complex tissues or organs such as bladder, liver, kidneys, etc., may require longer time owing to high cell density, granularity, vascularization and additional time for tissue maturation in vitro.Yet several manufacturers are utilizing this technique for other applications such as cosmetics, tumor models, drug screening and discovery.
To date, there is no single bioprinted FDA-approved product for clinical use that is commercially available in the market. Organovo, a 3D bioprinting company, has successfully bioprinted a liver tissue patch and implanted it in a mice model. This liver patch showed better engraftment, fluid retention and other liver-specific functionality for only up to 35 days. As a result, the manufacturer has committed to continue testing the patch further on large animals before testing on humans.Patient-specific tooth constructs (8 mm × 8 mm × 20 mm) with human dental pulp stem cell (hDPSCs) were printed by Jonghyeuk Han et al. using PCL, gelatin, hyaluronic acid, glycerol and fibrinogen. 5 mg/mL fibrinogen was used to mimic the pulp portion of teeth and 20 mg/mL of fibrinogen was used to print dentin in the PCL support. hDPSCs cultured in high concentrated fibrinogen exhibited odontogenic differentiation by expressing DMP1 and DSPP markers with higher mineralization. In contrast, low concentrated fibrinogen located in the pulp region maintained cells in the undifferentiated state.These studies showed the efficacy of 3D printing and cell therapies in fabricating patient-specific models to treat patients with cartilage defects, pancreatic, heart problems and other major therapies in the future. However, more clinical trials need to do to improve the patient safety and efficiency of the product. Nonetheless, the engineered or bioprinted tissues have various ethical scrutiny while translating from bench/pre-clinical side to bedside, which needs to be addressed in the early stages of research thereby thwarting lavish expense of time and money.
## Major problems in clinical transformation of engineered and bioprinted tissues
Worldwide regulatory considerations are not standardized with a common approach for the TEMPs category and hence there is a need for amendments in current standards and regulations. The technical enforcement of regulatory unions on framing the guidelines, reviewing and decision making may differ across the world based on the regulatory framework considered for its assessment.However, revisions in legislations and adaption of standards to the current technical developments are still lagging in many cases, thus making it a challenge for clinical translation of TEMPs.This is because the regulatory bodies take a huge time to comprehend the evolving transformations of new technologies and develop appropriate regulations/laws. Further, the experts of these bodies may not be able to make regulations by anticipating future products and technologies. However, these regulatory bodies may be advocated to form a multi-disciplinary research team that includes scientists, legal experts and social scientists. This team could be a part of the regulatory from the start of the development of a new technology and to form suitable regulations. This idea is developed as RRI: Responsible Research and Innovation which advocates this type of multi-disciplinary integration from the inception of new products development.As an alternative to overcome these challenges, most regulatory bodies are involved in an adaptive approach for its approval of cellular therapy products, gene therapy products, tissue engineered products and additive manufactured medical devices. TEMPs are complex and nonreliable products due to the incorporation of natural biomaterials and cells, which cause different host reactions to the product in a population. Thus, from a product point of view, scientific uncertainties of TEMPs cause a massive challenge and executional difficulties for regulatory authorities to approve medicinal therapies and products. In addition, most of the clinical trials have ended up with minimal efficiency and/or inducing high risk in patients' health, thus it gets automatically rejected during the approval process. Nevertheless, these limitations make it difficult to formulate common guidelines for characterizing the material/product and it heavily affects the commercialization of high-risk category products to be legally approved for clinical practice all over the world.Though suitable TEMPs are designed and studied well, delay in their market approval is due to the many different procedures and licensing steps involved in the regulations. It includes pre-clinical assessment studies (animal model, efficacy study, toxicology study), clinical trial design (experimental group and endpoint design, GCP and GMP maintenance), clinical outcome and approval decision after risk-based quality and efficacy assessments. However, less availability of preclinical results adds to the challenges on approval of engineered tissue under the high-risk categories. In these perspectives, the major challenges faced by researchers, sponsors and other non-medical regime partners are mainly due to stringent regulations, poor knowledge in the requirements to conduct clinical trials, inadequate laboratory practice guidelines and lack of funds that eventually make the clinical translation of tissue-based products more difficult. Furthermore, the unavailability of prior relevant standards also limits most of the clinical trial approval of TEMPs.
A pilot study conducted by Davies et al. in the United Kingdom along with cardiology, neurology, ophthalmology, orthopedic surgery, plastic and reconstructive surgery clinical specialists, have identified the major barrier for clinical transformation as efficacy and cost-effectiveness in addition to other observations including safety, regulation, methodology, visibility and patient characteristics for cell-based therapy clinical treatment. Moreover, it was also noted that inconsiderate processing techniques involved in manufacturing reduce the possibilities of approval through regulatory pathways.In another study, Hanna et al. stated that the number of ATMPs approved clinical studies in the EU has increased every year from 12 to 150 numbers between 2004 and 2014. Among the collected data with a total of 939 clinical trials, only 6.9% were in end-stage study (in phase III), which can take another 5 years for commercialization and the remaining 92.2% were in the early stages (phase I, combined phase I, II, and III stages). This may be due to the availability of sponsors (commercial/non-commercial) for the study, lack of market access and launching strategies and financial budget.Culme-Seymour et al. studied the cost-effective way to deliver engineered trachea for improving the health of three patients (two children and one adult). A patient could undergo the treatment procedure with multiple charges ranging from US$500,000 to US$600,000 based on treatment, availability and follow-up period. It also includes clinical charges (related to the patient's care at the hospital), clinical testing charges (charges due to the undergone procedures and analysis prior to and after treatment) and charges for engineered air transplant tissue and its maintenance. An estimated total manufacture cost for stem cell engineered trachea replacement is approximately $27,490, which could be considered as a cost-effective treatment in comparison with the quality of life regained after the treatment under severe conditions with available surgical procedures. 147
## Ethical concerns on temps, 3d printed and bioprinted tissues
## Cell source, processing procedures and their cost
Complex aspects of tissue engineering and bioprinting make its therapeutic potential ethically complicated compared to biologics and drugs.One of the major ethical concerns in TEMPs and bioprinted constructs is the cell source and its processing procedures.Cells and engineered tissue-based therapies are considered as drugs or biologics in few countries like Japan and the EU union. In these countries, cells and engineered tissue-based therapies are regulated under common legislations. In contrast, cell only therapy (autologous or stem cell) has different regulations when compared to 3D-printed/bioprinted/engineered tissues in US and other countries.The human body has different types of cells such as stem cells, cardiac cells, bone cells, nerve cells, fat cells, blood cells, muscle cells, etc. Among these stem cells, mesenchymal and induced pluripotent stem cells isolated from healthy adult donors are widely used as cell sources in TEMPs and bioprinted constructs due to their high differentiation potential toward any particular lineage of interest and negligible immunogenicity. Further, the use of embryonic stem cells (ESCs) isolated from blastocysts was considered unethical in some countries due to its collection procedure from living or aborted fetuses (either single cell or cluster of cells). These cells should not be intended to use for cloning humans, germ cells, embryos and human-animal chimeras.However, there may be limitations such as the large risk of patient's compliance, less therapeutic efficiency after implantation, high cost and compromised safety.There are several successful short-term in vivo studies, but long-term studies have indicated that these cells may lead to the risk of teratoma and cancer related issues.Numerous studies suggested that MSCs isolated from tissues such as the umbilical cord, bone marrow, adipose tissue, etc., have the potential to differentiate irrepressibly, suppress the immune system and produce new blood vessels, thereby promoting tumor growth and metastasis.In a study, hMSCs were intravenously injected into female BALB/c 4T1 tumor induced mice model and showed an increase in the tumor growth and metastasis, TGF-beta, interleukins (IL-4, IL-10) and decrease in interferon-gamma production, thereby enhancing the immunosuppressive environment for 35 days.In the literature, few works reported the use of MSCs, ESCs and iPSCs isolated from mouse, murine, human and chicks as bioink components for the printing of various tissues such as skin, cornea, bone, heart, etc. In many reports, primary cells and other cell types were used for bioprinting applications. For example, Michael et al. have used laser-assisted bioprinter and developed a cellularized skin substitute containing fibroblasts and keratinocytes on Matriderm for full-thickness wounds. The skin substitute was implanted in vivo and showed better integration with the host skin and neovascularization in the affected area within 11 days post-surgery.In another study, mesenchymal stromal cell lines (D1 cells) along with collagen and nano-hydroxyapatite were bioprinted in situ on mice calvaria defect model and showed enhanced bone volume to total volume using micro-CT experiments after 2 months of treatment.In conclusion, most of these bioprinted constructs were tested in vivo for its better engraftment with host tissues, vascularization and regeneration efficiency, but these are only initial studies and would have to go a long way through pre-clinical and clinical evaluations.
Patient-specific cells may be another choice to be used as a cell source for both seeding on TEMPs and bioprinting organs to avoid the ethical issues on cell sources. It is possible to isolate stem cells from patients with informed consent, thereby providing an immediate solution for the shortage of organ donors and limiting the use of immunosuppressants.However, the cost of stem cell collection procedure, in vitro culturing and fabrication procedures would require GMP sophisticated laboratories, expensive consumables, equipment, quality assurance testing, researchers, surgeons and biomedical engineers would be highly expensive and hence only benefits high income communities.Yet, customized and personalized products developed using the bioprinting technique may reduce the cost as they are 3D bioprinted on a very small scale.On the whole, 3D bioprinted organs would be more beneficial for the recipients waiting for donors. We hope that it may reduce the black market for organs or human transplantable materials if the cost gets reduced.However, clinical translation of 3D bioprinting may be a huge challenge while considering the social stratification, global inequalities and overall cost.Further, implementation of this newer technology into the existing healthcare and insurance systems requires careful considerations.
## Variability after implantation
An important facet to be considered after implantation of TEMPs in patients' body is that it exhibits an inherent variability due to the dynamic response of metabolically active cells in the host ECM environment, which may develop immune rejections or mismatch within the patient's body.In some cases, migration of the impregnated cells from TEMPs to other parts of the host body through the bloodstream may occur and this may cause unnecessary adverse events such as cancer or any unknown diseases. Some common ethical issues and risks associated with TEMPs are microbial contamination(re-infection), insufficient sterility and toxicity due to the presence of allogenic sources, dysfunction of bioactive motifs, over reactions of growth factors and toxicity of cryo-preservatives.Further, clinical trials on 3D bioprinted tissues require a more individual centric approach (transplantation medicine) than conventional normal trials. For example, these products have to be tested in an individual subject and should be continued on other subjects only after evaluating the clinical response in the first subject.Furthermore, 3D bioprinted tissue/organs could cause a high risk to the subject and hence such trials should be done when there is no alternative option for the chosen subject.
## Ownership of products
Finally, the ownership of products or bioprinted organs would be an issue for the donors, recipients, researchers, doctors and product manufacturing companies. A few known issues are (i) the donors must be informed about the future and further applications of their donated stem cells; (ii) safety issues related to the blueprint of patient's organ since advancement in genetic engineering may develop a cloned organs or simple biological building blocks (though it may take several decades to commence) and (iii) the owner of developed product for a patent will be an issue either for researcher, patient or company. 165
# Conclusion
Fabrication of personalized living tissues/organs is now more possible and realistic with the approaches in 3D-bioprinting. In spite of the availability of adaptive regulations, only a very few TEMPs have been licensed. Even though scientific requirements and pre-clinical standards are constantly revised, several tissue engineered products and 3D bioprinted tissues are yet to enter clinical trials. This could be attributed to the fact that the current legislation and standards are more suited for conventional drugs/ therapeutics and therefore fails to fully absorb the complexities involved in the latest medicinal products. Hence, it is imperative for the regulatory authorities and other parties to learn the background of these advanced products and adapt the complexities to facilitate the approval processes for TEMPs and 3D bioprinted tissues. Further, the clinicians and researchers should learn the scientific requirements, background of standards, necessary training on clinical trials and GMPs to fulfill the needs of regulatory bodies and global standards. In conclusion, considering the nature of TEMPs and 3D bioprinted tissues, it is essential to make sure that the products are within shelflife time at the time of clinical procedure and tested with well-planned protocols to assess the safety and long-term efficacy.
## Future perspectives
In order to anticipate the future of 3D bioprinting, it is imperative to understand the current advancements achieved in the successful development of 3D bioprinted tissues such as skin, cartilage and bone. This achievement is due to the clinically most relevant features attained in printing these constructs such as suitable cell types, desired functional aspects, biomimetic resolution and other required associated cues, including vascular network. Therefore, successfully 3D bioprinted skin, cartilage and bone could serve as a yardstick to envision the future prospects and commercialization scope of fabricated constructs. shows the 3D bioprinted tissues/organs, and bioprinted organ-on-chip developments along with anticipated years for regulatory approval and commercialization. On considering the medical device, ISO and other regulations in force, we anticipate that bioprinted 3D tissue models and organ-on-chip may not undergo stringent regulations and hence could be envisaged for commercialization possibilities in the next 5-8 years. In contrary, commercialization chances for 3D bioprinted tissues and organs are anticipated only in the forthcoming decades due to complex nature and multiple biological compositions. Additionally, a collaborative approach between academia, industry and hospitals is needed to draw regulatory amendments, which could further improve the clinical translation of bioprinted organs/tissue.
Despite the benefits of creating fully functional soft and hard 3D tissues with neovascularization ability, additional challenges include culturing the printed heterogeneous tissues/organs with a different supporting medium in a bioreactor or with any special set up for maturation. Further, the highest possible printing resolution with a laser bioprinter is 20 µm while small capillaries of native tissues are 3 μm in diameter. In addition to the small diameter, the complexity of vascular networks with the multi-level organization is yet to be achieved in 3D bioprinting. Other challenges are the requirement of a huge number of cells for printing, the need for extensive research on optimal bioink composition with good printing capabilities, shape fidelity and the demand for sophisticated advanced complementary bioprinting technologies with ensured environment and sterility maintenance. Additionally, bioprinting has several other difficulties such as ethical issues, safety, affordability, and large-scale production for successful translation into clinical practice. Marketing approval for the bioprinted tissue equivalents is yet to be achieved with the current/revised legislation and hence focus is needed toward these directions to utilize 3D bioprinting for the fabrication of organs for transplantation needs. As a nextgeneration fabrication approach, the 4D bioprinting concept with an additional ability of printed constructs to undergo complete maturation in response to dynamic external stimuli may be followed. In the future, several extensive researches on smart bioink composition with excellent shape fidelity, feasibility for large scale production, potential bioprinters and faster clinical trials for the developed bioprinted constructs will definitely pave the way for clinical success and commercialization scope as compared to conventional TEMPs.
## Declaration of conflicting interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article. |
Consequences of partially recessive deleterious genetic variation for the evolution of inversions suppressing recombination between sex chromosomes
The evolution of suppressed recombination between sex chromosomes is widely hypothesized to be driven by sexually antagonistic selection (SA), where tighter linkage between the sex-determining gene(s) and nearby SA loci is favored when it couples male-beneficial alleles to the proto-Y chromosome, and female-beneficial alleles to the proto-X. Despite limited empirical evidence, the SA selection hypothesis overshadows several alternatives, including an incomplete but often-repeated "sheltering hypothesis" that suggests that expansion of the sex-linked region (SLR) reduces homozygous expression of partially recessive deleterious mutations at selected loci. Here, we use population genetic models to evaluate the consequences of deleterious mutational variation for the evolution of neutral chromosomal inversions expanding the SLR on proto-Y chromosomes. We find that SLR-expanding inversions face a race against time: lightly loaded inversions are initially beneficial, but eventually become deleterious as they accumulate new mutations, and must fix before this window of opportunity closes. The outcome of this race is strongly influenced by inversion size, the mutation rate, and the dominance coefficient of deleterious mutations. Yet, small inversions have elevated fixation probabilities relative to neutral expectations for biologically plausible parameter values. Our results demonstrate that deleterious genetic variation can plausibly drive recombination suppression in small steps and would be most consistent with empirical patterns of small evolutionary strata or gradual recombination arrest.K E Y W O R D S : Chromosomal inversion, indirect selection, mutation, recombination suppression, sex chromosomes.
Sex chromosomes have evolved from homologous pairs of autosomes repeatedly within many eukaryotic lineages across the tree of life (reviewed by [bib_ref] Sex determination: why so many ways of doing it?, Bachtrog [/bib_ref]. A striking feature of many sex chromosome systems is the evolution of recombination suppression, which profoundly influences the long-term fate of the chromosomes. Once recombination stops, the subsequent evolution of sequence divergence and functional degeneration of the non-recombining region of the sex-limited chromosome, and possibly dosage compensation to retain adequate gene expression levels in both sexes, can all contribute to the gradual evolution of sex chromosome heteromorphy [bib_ref] A dynamic view of sex chromosome evolution, Bachtrog [/bib_ref] [bib_ref] Steps in the evolution of heteromorphic sex chromosomes, Charlesworth [/bib_ref] [bib_ref] The accumulation of sexually antagonistic genes as a selective agent promoting the..., Rice [/bib_ref] [bib_ref] Evolution of the y sex chromosome in animals, Rice [/bib_ref].
The initial loss of recombination between sex chromosomes is widely hypothesized to be caused by selection favoring linkage disequilibrium between the sex-determining gene(s) and nearby loci experiencing sex-differences in selection (e.g., sexually antagonistic loci with alleles that have opposite fitness effect between sexes) [bib_ref] Sex differences in fitness and selection for centric fusions between sex-chromosomes and..., Charlesworth [/bib_ref] [bib_ref] The evolution of sex dimorphism in recombination, Lenormand [/bib_ref] [bib_ref] Evolutionary potential for genomic islands of sexual divergence on recombining sex chromosomes, Otto [/bib_ref] [bib_ref] The accumulation of sexually antagonistic genes as a selective agent promoting the..., Rice [/bib_ref] [bib_ref] Evolution of the y sex chromosome in animals, Rice [/bib_ref]. Indeed, even though strong empirical support for the sexual antagonism hypothesis remains elusive, it continues to overshadow a variety of alternatives that have received less theoretical or empirical attention [bib_ref] A neutral model for the loss of recombination on sex chromosomes, Jefferies [/bib_ref] [bib_ref] Why do sex chromosomes stop recombining?, Ponnikas [/bib_ref].
Several of these alternative hypotheses revolve around the idea that a chromosomal rearrangement -typically an inversion -expanding the male-limited region of a Y chromosome (or female-limited region of a W) may be selectively favored due to a form of heterozygote advantage, or "sheltering," arising from the combination of wild-type and partially recessive deleterious alleles that it captures [bib_ref] No amicable divorce? challenging the notion that sexual antagonism drives sex chromosome..., Ironside [/bib_ref] [bib_ref] Why do sex chromosomes stop recombining?, Ponnikas [/bib_ref]. In fact, a tangle of at least three distinct sheltering hypotheses have been described in varying detail, ranging from loose verbal models to mathematical and simulation models. First, [bib_ref] Inbreeding, heterozygote advantage and the evolution of neo-x and neo-y sex chromosomes, Charlesworth [/bib_ref] cautiously suggested that selection might favor linkage between the sex-determining locus and multiple selected loci exhibiting associative overdominance due to partially recessive deleterious variation, an idea that was echoed in subsequent reviews [bib_ref] No amicable divorce? challenging the notion that sexual antagonism drives sex chromosome..., Ironside [/bib_ref] [bib_ref] Why do sex chromosomes stop recombining?, Ponnikas [/bib_ref]. A later verbal model byproposed that partially recessive deleterious mutations in partial linkage with the sex-determining region could cause selection for recombination arrest to avoid homozygous expression in recombinant genotypes. Most recently,modeled the evolution of inversions expanding the sex-linked region on Y chromosomes under deleterious mutation pressure using constant fitness effects for inversion genotypes depending on the alleles they initially capture, with accompanying individual-based simulations (we review the various hypotheses in detail in Appendix A of the Supporting Information). Each of these models proposes, in some way, that an inversion linking alleles at selected loci to the heterozygous male-determining allele on a Y chromosome will reduce the homozygous expression of partially recessive deleterious alleles at those loci. This sheltering effect is hypothesized to cause higher fitness for inverted relative to non-inverted Y chromosomes.
The red thread running through each of these sheltering hypotheses is that deleterious mutational variation is pervasive throughout the genome [bib_ref] The effect of deleterious mutations on neutral molecular variation, Charlesworth [/bib_ref] [bib_ref] Our load of mutations, Muller [/bib_ref] and the fate of new inversions expanding the sexlinked region may therefore be strongly influenced by the random sample of that variation that they happen to capture. Moreover, deleterious genetic variation is known to have important implications for the evolution of inversions on autosomes: autosomal inversions undergo a complex time-dependent selection process that can result in diverse evolutionary outcomes, including fixation, extinction, and balanced polymorphism, depending on the set of alleles they initially capture [bib_ref] Selection of new inversions in multi-locus genetic systems, Charlesworth [/bib_ref] [bib_ref] Impact of chromosomal inversion length on fixation probability, Connallon [/bib_ref] [bib_ref] Frequency changes of new inversions in populations under mutation-selection equilibria, Nei [/bib_ref].
However, there is a crucial difference between autosomal inversions and those expanding the sex-linked region on a Y chromosome. By capturing the dominant sex-determining factor (or expanding the chromosomal region already linked to it), the latter are prevented from occurring in both X and Y chromosomes. At first glance, this might appear to enforce heterozygosity of partially recessive deleterious alleles captured by the inversion, thereby paving the way for inversion fixation as suggested by earlier studies [bib_ref] No amicable divorce? challenging the notion that sexual antagonism drives sex chromosome..., Ironside [/bib_ref]. As we demonstrate below, this is an oversimplification. In fact, when partially recessive deleterious genetic variation is present, inversions expanding the sex-linked region (SLR hereafter) on Y chromosomes experience fundamentally different time-dependent selection processes compared to autosomal ones. Careful consideration of these time-inhomogeneous selection processes is necessary to fully understand the evolutionary dynamics of inversions contributing to recombination suppression between sex chromosomes.
In this paper, we develop a population genetic model to describe the evolutionary dynamics of new inversion mutations that capture the dominant sex-determining factor in randomly mating populations, while explicitly considering the consequences of standing deleterious mutational variation. We briefly describe the deterministic frequency dynamics predicted by the model before turning our attention to the fixation probabilities for inversions of different lengths, which we calculate using stochastic Wright-Fisher simulations. Our results illuminate two important features of inversions on Y chromosomes expanding the SLR: (i) because partially recessive deleterious mutations segregate outside the SLR on both X and Y chromosomes, mutations initially captured by an inversion are expressed as homozygotes at a rate equal to their frequency in X chromosomes, so that inversions capturing even a single deleterious allele will carry a permanent, but temporally dynamic, deleterious mutation load; (ii) inversions initially capturing fewer than the average number of deleterious mutations over the chromosomal segment they span will initially be beneficial, but this selective advantage erodes over time as new mutations accumulate on descendent copies of the inversion until the benefit becomes smaller than the permanent load carried by the inversion; at this point the overall fitness effect of the inversion becomes irreversibly deleterious and its chances of fixation negligible. Hence, the evolutionary fate of SLR-expanding inversions is a race against time; initially beneficial inversions must fix before their selective advantage decays and their window of opportunity closes permanently. The outcome of this race can be non-intuitive and is determined jointly by the size of new inversions as well as several key population genetic parameters, including the deleterious mutation rate, dominance and selection parameters, and population size. We close by discussing the implications of our findings for existing theories of recombination arrest between sex-chromosomes.
# Methods and results
## Overview of the model
Consider a population of diploid, randomly mating individuals with discrete generations, in which sex is determined genetically by a dominant male-determining factor (as in a maleheterogametic X-Y system). The model is equally applicable to female heterogametic Z-W systems if male/female labels are reversed. The order of life history events proceeds as follows: fertilization, mutation, selection, then meiosis. We model idealized sex chromosomes that can be divided into two regions: (i) a non-recombining SLR that could be limited to just the sexdetermining (SD) gene(s), or a more extensive region harboring the SD gene(s); and (ii) a pseudoautosomal region (PAR) in which recombination can still occur, and functional homologs of any genes are still present on both X and Y chromosomes. Hence, our models are most applicable to genetic systems in which the evolution of recombination suppression between sex chromosomes is incomplete, and the PAR accounts for a sizeable fraction of the proto sex chromosomes (e.g., [bib_ref] Steps in the evolution of heteromorphic sex chromosomes, Charlesworth [/bib_ref].
We model the evolution of new chromosomal inversion mutations arising on a Y chromosome that would expand the SLR were they to fix among the Y chromosomes in the population. Our goal is to predict how new inversions will respond to indirect selection against deleterious mutations segregating within the population at the loci they span. To isolate these indirect selection effects, we assume that the inversion itself is neutral (i.e., inversions cause no breakpoint effects or meiotic dysfunction; [bib_ref] Selection on inversion breakpoints favors proximity to pairing sensitive sites in drosophila..., Corbett-Detig [/bib_ref] [bib_ref] Inversion breakpoints and the evolution of supergenes, Villoutreix [/bib_ref]. For simplicity, we also assume that loci within the SLR do not contribute to indirect selection on inversions. This second assumption can be justified in different ways: (i) there has been sufficient differentiation and functional degeneration within the SLR on Y chromosomes that few functional genes remain in this region; (ii) the SLR is small relative to the length of inversions, such that there are few loci other than the sex-determining genes within the SLR; and (iii) any loci within the SLR that are captured by an inversion will contribute minimally to indirect selection favoring suppressed recombination because they are already fully sex-linked.
Our model relies on several other important simplifying assumptions. First, we assume that inversions completely suppress recombination between inverted-Y and X chromosomes over the chromosomal region they span (in fact, genetic exchange could occur via double crossovers or gene conversion; Krimbas & Pow-ell 1992, [bib_ref] Pervasive gene conversion in chromosomal inversion heterozygotes, Korunes [/bib_ref]. Second, we assume that new inversions occur rarely enough that the evolutionary fate of a given inversion is independent of others (i.e., we assume "strong selection, weak mutation" with respect to inversions; Gillespie 1991). Our results therefore preclude the possibility that multiple inversions segregate simultaneously within the population. Third, we assume that deleterious alleles segregate at mutation-selection balance outside of the SLR, with no epistasis, and no linkage disequilibrium among loci or with the SLR prior to a new inversion mutation. This requires strong purifying selection against deleterious variants relative to mutation or genetic drift. Finally, we assume that fitness is multiplicative over the loci spanned by the inversion.
Below, we develop a deterministic model describing the frequency dynamics of new inversions expanding the SLR on Y chromosomes in the presence of partially recessive deleterious mutational variation and illustrate important features of the model predictions. We emphasize the role of joint changes in the inversion frequency as well as deleterious allele frequencies on X chromosomes. For simplicity, we present deterministic results for the idealized case where mutation and selection coefficients at all loci spanned by the inversion are equal. We then present Wright-Fisher (W-F) simulations that incorporate stochastic fluctuations in inversion frequencies due to random gamete sampling in a finite population. We present most of the mathematical details in Appendix B. Computer code needed to reproduce the simulations and main figures is available on GitHub (https://github. com/colin-olito/shelteringOnSexChrom) and a version of record is archived on Zenodo [bib_ref] Consequences of partially recessive deleterious genetic variation for the evolution of inversions..., Olito [/bib_ref].
## Deterministic model
We first define two useful terms: the total number of selected loci located outside of the SLR on the chromosome arm containing it (i.e., within the PAR), n tot , and the length of a new inversion, x, expressed as the proportion of the PAR that the new inversion links to the ancestral SLR. Assuming that the selected loci are distributed uniformly along the chromosome arm, the number of loci spanned by a new inversion of length x will be n = n tot x. Each of the n loci are assumed to be diallelic, with a wild-type allele (A) at the i th locus that mutates to a deleterious variant (a) at a rate µ i per meiosis (we ignore backmutation from a → A), with locus-specific genotypic relative fitnesses of w i,AA = 1, w i,Aa = 1 − h i s i , w i,aa = 1 − s i . Deleterious alleles segregate at each locus at their mutation-selection balance equilibrium frequency ofq i = µ i /(h i s i ). A new inversion mutation will capture a random sample of the standing deleterious variation at these n loci, which can be conveniently divided into two classes: loci where the inversion initially captures a deleterious allele (denoted by a superscript D) and those where it captures a wild-type allele (denoted by a superscript W ).
## B r i e f c o m m u n i c at i o n
To describe the deterministic frequency dynamics of an SLR-expanding inversion, it is necessary to track allele frequency changes for these two classes of loci, D and W , within four different chromosome classes: X's in ovules/eggs, X's in pollen/sperm, non-inverted Y's, and inverted Y's, which we denote X f , X m , Y , and Y I respectively (see [bib_ref] Selective maintenance of recombination between the sex chromosomes, Otto [/bib_ref]. The frequencies at time t for each of the n loci can be described using the following notation:
[formula] q D,i Xf ,t , q D,i Xm,t , q D,i Y,t , q D,i Y I ,t , and q W,i Xf ,t , q W,i Xm,t , q W,i Y,t , q W,i Y I ,t [/formula]
where q refers to the deleterious allele frequency at the i th locus. Note that, by assumption, q D,i Y I = 1 for all t because all descendent copies of the inversion will carry deleterious alleles at D loci. Likewise, q W,i Y I = 0 for t = 1 but increases due to new mutations on descendent copies of the inversion. We present the full development of the recursions in Appendix B.
Under the simplifying assumption that mutation and selection parameters are constant across all loci captured by the inversion (µ i = µ, s i = s, h i = h), the number of deleterious alleles initially captured by the inversion, r, will be Poisson distributed: r ∼ Poisson(U x/hs). In this idealized case, the deleterious allele frequencies at all loci within a given chromosomal class will follow the same trajectory (e.g., each inversion-captured locus that is initially free of a deleterious allele will evolve the same as other loci in the inversion with the same, mutation-free initial state). That is, q W,i class,t = q W class,t , and q D,i
[formula] class,t = q D class,t , where class ∈ {X f , X m , Y, Y I }. [/formula]
This allows us to define the following simplified recursion for the frequency of an inversion that initially captures r deleterious alleles in terms of the allele frequencies:
[formula] Y I t+1 = Y I t 1 − s hp D Xf ,t + q D Xf ,t r 1 − s h p W Y I t q W Xf ,t + q W Y I t p W Xf ,t + q W Y I t q W Xf ,t n−r /w Y (1) wherew Y = Y I t 1 − s hp D Xf ,t + q D Xf ,t r 1 − s h p W Y I t q W Xf ,t + q W Y I t p W Xf ,t + q W Y I t q W Xf ,t n−r + (1 − Y I t ) 1 − s h p D Xf ,t q D Y,t + q D Xf ,t p D Y,t + q D Xf ,t q D Y,t r 1 − s h p W Xf ,t q W Y,t + q W Xf ,t p W Y,t + q W Xf ,t q W Y,t n−r ,(2) [/formula]
and we have used the convention p · ·,t = 1 − q · ·,t to simplify notation. The deterministic frequency dynamics of the inversion can now be fully described by a system of eight recursions corresponding to the deleterious allele frequencies in each of the seven
[formula] relevant loci × chromosome classes (q D Xf , q D Xm , q D Y , and q W Xf , q W Xm , q W Y , q W Y I ; see Appendix B [/formula]
) and the frequency of the inversion, Equations (1) and (2).
Equation (1) offers immediate insight into the different contributions of D vs. W loci to relative fitnesses of inversion genotypes. The fitness effects of D loci depend solely on the frequency of deleterious alleles in X chromosomes in ovules/eggs (selection terms in bracketed expressions with an exponent of r involve only p D Xf ,t and q D Xf ,t ) because all descendent copies of the inversion already carry a deleterious allele at these loci. Under recurrent mutation, q D Xf ,t will always be nonzero, and so it is immediately clear that D loci will impart a permanent fitness cost to the inversion. Meanwhile, W loci depend jointly on the frequency of deleterious alleles in ovule/egg-derived X chromosomes and the accumulation of new deleterious mutations on descendent copies of the inversion at these loci (selection terms involve p W Xf ,t , q W Xf ,t , and q W Y I ,t ). Compared to the fitness of standard Y chromosomes (the second expression in square brackets in Equation 2) it is clear that any temporary fitness advantage of new inversions must come from the initially low frequency of deleterious alleles at W loci among inverted Y chromosomes (q W Y I ,t ). The key questions become: when, and for how long, does the temporary fitness benefit from W loci outweigh the permanent load associated with D loci? and does it result in elevated fixation probabilities for new SLR-expanding inversions?
## Deterministic frequency dynamics
When deleterious alleles are approximately codominant (i.e., h i ≈ 1/2), most purifying selection occurs in heterozygotes and an inversion initially loaded with even a single deleterious allele (i.e., when r > 0) will not invade (see [bib_ref] Impact of chromosomal inversion length on fixation probability, Connallon [/bib_ref]. However, when deleterious mutations are partially recessive (0 < h i < 1/2), as expected by theory and supported by empirical data [bib_ref] Inferences about the distribution of dominance drawn from yeast gene knockout data, Agrawal [/bib_ref] [bib_ref] Gene expression drives the evolution of dominance, Huber [/bib_ref] [bib_ref] Fitness landscapes: an alternative theory for the dominance of mutation, Manna [/bib_ref] , lightly loaded inversions expanding the SLR on Y chromosomes can deterministically rise to high frequency under restrictive conditions. that are initially beneficial because they capture fewer deleterious mutations than the population average over the chromosomal segment they span. The time course of inversion fitness relative to non-inverted Y chromosomes illustrates both the decay of the initial fitness benefit as new mutations accumulate on descendent copies of the inversion-bearing chromosome at W loci (inversions that initially have relative fitness greater than one), and the permanent deleterious load due to D loci (all inversions with r > 0 eventually become deleterious). The tipping-point where the relative fitness of loaded inversions drops below one occurs when the transient benefit to the inversion of initially capturing fewer-than-average deleterious mutations no longer compensates for the cost of being fixed for those few mutations. The deterministic inversion frequency dynamics reflect these changes in relative fitness and highlight that the deterministic outcome for initially unloaded inversions is to fix among Y chromosomes in the population. When deleterious mutations are strongly recessive (h ≈ 0.01), inversions capturing several deleterious mutations can deterministically rise to intermediate to high frequencies before the decline in their initial fitness advantage results in them becoming deleterious and crashing to extinction . The restrictive conditions under which an initially loaded inversion is expected to deterministically 'fix' occur when all deleterious mutations are strongly recessive and a large inversion captures very few of them (see Figs. S1-S3 in Appendix C).
Importantly, the load carried by inversions due to D loci is temporally dynamic because the frequencies of partially recessive deleterious alleles on X (and non-inverted Y) chromosomes change over time in response to the inversion frequency . We illustrate this for representative cases of inversions initially loaded with relatively few deleterious mutations (r = 1 in panels G,H; r = 10 in panel I) by showing the deterministic frequency dynamics for three important loci × chromosome classes: W loci on the inversion (q W YI ), and both W and D loci on X chromosomes in ovules/eggs (q W Xf and q D Xf ). The red line shows the accumulation of deleterious mutations on the inversion at W loci, which causes the decline in initial fitness benefit due to these loci. The black solid and dotted lines show the corresponding changes in deleterious allele frequency among ovule/egg-derived X chromosomes at W loci (q W Xf ) and D loci (q D Xf ), respectively. Selection against deleterious alleles at D loci on ovule/egg-derived X chromosomes intensifies as the initially beneficial inversion increases in frequency because all inverted-Y-bearing sons who inherit a deleterious allele from their mother's X chromosome will be homozygous for the deleterious allele at D loci. This intensified selection drives the frequency of deleterious alleles at D loci on X chromosomes (q D Xf ) down to lower levels relative to the equilibrium prior to the inversion. Nevertheless, the deleterious load due to D loci persists because new mutations continually arise on X and non-inverted Y chromosomes. When the transient benefit of the inversion due to W loci can no longer compensate for this load, the inversion becomes deleterious and declines in frequency, at which point the deleterious allele frequencies return to the pre-inversion equilibrium. These dynamics are especially evident when deleterious mutations are strongly recessive .
An overview of the deterministic dynamics for inversions of different lengths initially loaded with different numbers of deleterious alleles is presented in Appendix C, Figures S1-S3. Of particular note in the Supporting Information figures is that lightly loaded large inversions have higher initial fitness than smaller ones, but are still not expected to increase to high frequencies unless deleterious mutations are strongly recessive. The time dependent dynamics can also result in faster accumulation of strongly recessive deleterious mutations on SLR-expanding inversions .
## Wright-fisher simulations
While the deterministic frequency dynamics presented above clearly illustrate the shifting balance between the time-dependent selection processes on inversions due to W and D loci, they do not consider either the stochastic process of gamete sampling present in finite populations that can result in loss of rare inversions, nor the likelihood that a new inversion captures a given number of mutations. In reality, larger inversions are more likely to capture a greater number of deleterious mutations (see Figs. S1-S3), and the combined effects of W and D loci on inversion relative fitness will depend on allele frequency dynamics at each of the n loci. To begin tackling the effects of drift, we use Wright-Fisher simulations carried out in R (R Core Team 2020) to estimate the fixation probability for a single Y chromosome bearing an SLR-expanding inversion as a function of inversion length. We make the (strong) simplifying assumption that deleterious allele frequencies at the n loci spanned by the inversion change deterministically, while the inversion itself is subject to genetic drift due to random gamete sampling. This approach qualitatively captures the effects of time-dependent selection on inversion fixation probability in large populations, where ephemeral indirect selection effects are most likely to be important. However, it also makes our simulation results not applicable to smaller populations where deleterious allele frequency dynamics are dominated by drift rather than selection. Unfortunately, relaxing this assumption using an individual based model becomes computationally intractable for relevant values of selection and dominance coefficients due to the high replication necessary to reliably estimate fixation probabilities (see below).
In each generation, haplotype frequencies are censused among gametes prior to fertilization, following mutation and indirect selection due to segregating deleterious alleles at each locus, with a total effective population size of N. We use our exact deterministic recursions (Appendix B) to generate predictions for the gene frequencies at each of the n loci after indirect selection for each locus × chromosome class. The realized frequency of the inversion among Y chromosomes in each generation was then simulated using pseudo-random binomial sampling during the gametic phase, with the number of Y chromosomes representing the number of trials (N/2), and the deterministic frequency predictions representing the probabilities of sampling the inversion among pollen/sperm. For each replicate simulation, the number of loci spanned by each new inversion was n = n tot x and the number of deleterious alleles initially captured by a new inversion of length x was drawn from a Poisson distribution with mean and variance U x/hs. Fixation probabilities were estimated from the outcomes of at least 100 × N/2 replicate simulations (this was increased to 500 × N/2 for larger inversions because of their low fixation probabilities). For comparison, we also estimated the fixation probability of autosomal inversions using a similar procedure but implementing multinomial pseudo-random sampling of adult genotypes (e.g., .
## Inversion fixation probabilities
We focus on the effect of inversion length on fixation probability because the fitness benefits and costs scale differently with inversion length for autosomal and SLR-expanding inversions. In both cases, the initial fitness benefit of capturing wild-type alleles at more loci increases with inversion length, but so does the probability of capturing more deleterious alleles and carrying a larger permanent deleterious load. For autosomal inversions, these countervailing effects of inversion size cancel out when deleterious mutations are not strongly recessive and population sizes are sufficiently large, resulting in expected fixation probabilities that are independent of inversion length and approximately equal to the initial frequency of the inversion [bib_ref] Impact of chromosomal inversion length on fixation probability, Connallon [/bib_ref]. With substantial (albeit partially recessive) fitness effects in heterozygotes and small population sizes, there is a fixation bias toward small autosomal inversions, which have a maximum fixation probability approximately equal to that of a typical neutral allele, 1/2N [fig_ref] Figure 2: Fixation probabilities estimated from Wright-Fisher simulations plotted as a function of inversion... [/fig_ref]. In larger populations, the fixation probability becomes increasingly independent of inversion size, particularly when there are fewer average deleterious mutations per standard chromosome arm [fig_ref] Figure 2: Fixation probabilities estimated from Wright-Fisher simulations plotted as a function of inversion... [/fig_ref]. We focus our analysis on deleterious alleles with equal dominance coefficients of h i = h = 0.25, which corresponds roughly to the average dominance coefficient of deleterious mutations estimated from empirical studies [bib_ref] Fitness landscapes: an alternative theory for the dominance of mutation, Manna [/bib_ref] [bib_ref] Mutation load: The fitness of individuals in populations where deleterious alleles are..., Agrawal [/bib_ref] , but explore the effects of strong recessivity in Appendix D (see Figs. S5 and S6).
For inversions expanding the SLR on Y chromosomes, the fitness costs scale differently with inversion size for reasons that were highlighted by our deterministic model: inversions capturing fewer than the average number of deleterious alleles are initially beneficial and can therefore rise in frequency until they accumulate enough deleterious mutations that they eventually become deleterious. However, because the capture of even a single deleterious allele imparts a permanent deleterious load on a new inversion, initially unloaded inversions benefit most from the time-dependent selection process. Consequently, there is a strong fixation bias toward smaller SLR-expanding inversions [fig_ref] Figure 2: Fixation probabilities estimated from Wright-Fisher simulations plotted as a function of inversion... [/fig_ref] up to two orders of magnitude larger than the neutral expectation of 2/N) because these are most likely to initially capture no deleterious alleles (see Figs. S1-S3). Nevertheless, larger SLR-expanding inversions that are lightly loaded are still more likely to go to fixation than similarly sized autosomal ones.
The strength of mutation relative to selection, which influences the average number of deleterious mutations per standard (non-inverted) arrangement chromosome, has a similar effect on the fixation probability for autosomal versus SLR-expanding inversions. A higher average number of deleterious mutations on non-inverted chromosomes (higher U/hs) decreases the fixation probability of larger inversions [fig_ref] Figure 2: Fixation probabilities estimated from Wright-Fisher simulations plotted as a function of inversion... [/fig_ref] because they are more likely to initially capture deleterious mutations and will accumulate new mutations more rapidly, despite having a greater initial fitness benefit over non-inverted chromosomes. The av-erage deleterious load also influences the threshold inversion size where the fixation probability of SLR-expanding inversions drops below 2/N, with higher average numbers of deleterious mutations per standard-arrangement chromosome (large U/hs) corresponding to smaller threshold sizes [fig_ref] Figure 2: Fixation probabilities estimated from Wright-Fisher simulations plotted as a function of inversion... [/fig_ref].
Despite the involvement of rather complicated timedependent selection processes, the different evolutionary dynamics of autosomal vs. SLR-expanding inversions in our models can be explained rather simply: the homozygous expression of partially recessive deleterious mutations initially captured by autosomal inversions increases with inversion frequency, while for SLR-expanding inversions they are expressed at a rate equal to their frequency on X chromosomes in ovules/eggs, which remains small (≤q i ) and even decreases with inversion frequency. Due to their lower, temporally dynamic, permanent deleterious load, lightly loaded SLR-expanding inversions have a window of time during which they can fix before their fitness benefit decays and they become deleterious. Initially unloaded inversions benefit most from the time-dependent selection process, resulting in elevated fixation probabilities for smaller inversions that can be significantly higher than that of a neutral allele.
# Discussion
Our model predictions have several important implications. First, our deterministic model clarifies that the notion proposed by several earlier sheltering hypotheses-that linkage to the permanently heterozygous male-determining factor will prevent or reduce the homozygous expression of deleterious mutations, thereby favoring recombination suppression-is an oversimplification. Deleterious alleles initially captured by an SLRexpanding inversion are not prevented from being expressed as homozygotes, as suggested by earlier verbal and mathematical arguments [bib_ref] No amicable divorce? challenging the notion that sexual antagonism drives sex chromosome..., Ironside [/bib_ref]. Rather, they are expressed as homozygotes at a rate equal to their frequencies on X chromosomes in ovules/eggs, which change with inversion frequency. The deleterious load carried by these inversions is temporally dynamic but still permanent: initially beneficial inversions capturing even one mutation will eventually become deleterious. This is the critical difference between SLR-expanding and autosomal inversions, for which the homozygous expression of any captured deleterious alleles increases with inversion frequency because more individuals are homozygous for the inverted karyotype, which effectively prevents even lightly loaded inversions from ever fixing [bib_ref] Impact of chromosomal inversion length on fixation probability, Connallon [/bib_ref] [bib_ref] Frequency changes of new inversions in populations under mutation-selection equilibria, Nei [/bib_ref].
Second, our simulation results suggest that small inversions on proto-Y chromosomes are more likely to contribute to recombination suppression when selection is indirect and attributable to associations between inversions and deleterious variation. The elevated fixation probabilities of small inversions predicted by our models suggest that a relatively simple alternative to the sexually antagonistic selection hypothesis-a suitably located neutral inversion combined with segregating deleterious variationcan drive the evolution of recombination suppression between sex chromosomes in small steps. Whether our model predictions are consistent with empirical patterns of evolutionary strata size remains to be seen, but our results suggest that the length of SLR-expanding inversions may offer insight into the selective processes driving their fixation. In particular, large evolutionary strata generally appear inconsistent with the inversion fixation probabilites expected under the sheltering hypothesis [bib_ref] Impact of chromosomal inversion length on fixation probability, Connallon [/bib_ref] [bib_ref] The role of genic selection in the establishment of inversion polymorphism in..., Santos [/bib_ref] [bib_ref] The origins of inversion polymorphisms, Van Valen [/bib_ref].
It remains difficult to determine the relative importance of different hypotheses for suppressed recombination between sex chromosomes for several reasons. First, knowledge about the mutation rate and length distribution of new inversions in natural populations is limited, as is the number of well-described evolutionary strata (although genomic data are starting to shed light on these; [bib_ref] Impact of chromosomal inversion length on fixation probability, Connallon [/bib_ref]. Second, our models make several simplifying assumptions. Generalizing to other cases is beyond the scope of the present paper, but important questions remain for future work. For example, allowing non-uniform distributions of dominance and selection coefficients will likely further reduce the size of SLR-expanding inversions that ultimately become fixed, since this is the case for autosomal inversions [bib_ref] Impact of chromosomal inversion length on fixation probability, Connallon [/bib_ref]. Yet, a few loci segregating for highly recessive deleterious alleles (e.g., h i ≈ 0.01) could have the opposite effect, although these may generally have strongly deleterious effects that would oppose this [bib_ref] Gene expression drives the evolution of dominance, Huber [/bib_ref] [bib_ref] The integrative biology of genetic dominance, Billiard [/bib_ref] ; see . Lastly, our W-F simulations made the strong simplifying assumption that deleterious allele frequencies changed deterministically, while inversions were subject to genetic drift. Although the time-dependent indirect selection effects in our model will be most important in large populations, additional simulation studies are needed to determine at what population sizes they become relevant. Yet, a recent simulation study of Y recombination arrest and sex-specific regulatory evolution sheds some light on the robustness of our model predictions . In the absence of regulatory evolution (with only partially recessive deleterious mutations, similar to the senario we investigate here), found higher fitness variance of large SLR-expanding inversions, but faster degradation of large inversions and a fixation bias towards small inversions when N = 10 4 , results that are consistent with our model predictions.
Most notably, we have limited our attention to neutral inversions capturing deleterious mutations, without considering those with beneficial effects or capturing loci under sexually antagonistic selection that could cause linkage disequilibrium with the SLR. However, the presence of partially recessive deleterious ge-netic variation should influence the fixation probabilities of differently sized inversions under these other selection scenarios (e.g., Olito & Abbott 2020) by making the conditions for fixation of small inversions easier to satisfy. We have also focused our analysis on single randomly mating populations and selected loci initially at linkage equilibrium with the SLR. However, both inbreeding and prior linkage disequilibrium with the SLR feature prominently in previous sheltering hypotheses [bib_ref] Inbreeding, heterozygote advantage and the evolution of neo-x and neo-y sex chromosomes, Charlesworth [/bib_ref] ; see Appendix A for further details). We briefly address these possibilities using mathematical models in Appendices D and E, where we show that they do not result in indirect positive selection for recombination-suppressing inversions. However, a more complete treatment of these problems is probably warranted.
Overall, our results highlight the importance of confronting seemingly intuitive hypotheses and verbal arguments with mathematical models. Despite the intuitive appeal of the sexually antagonistic selection hypothesis for recombination suppression between sex chromosomes, our predictions suggest that a simple alternative-the presence of partially recessive deleterious genetic variation on proto-sex chromosomes and a suitably located inversion-is plausible under some conditions and deserves more careful consideration. Yet, our models also show that apparently simple verbal arguments, like the earlier sheltering hypotheses for recombination suppression, also deserve careful scrutiny because the details of the genetic system and emergent selection processes are anything but intuitive.
## Supporting information
Additional supporting information may be found online in the Supporting Information section at the end of the article. : Overview of deterministic fitness and frequency dynamics for initially beneficial inversions of different sizes initially loaded with different numbers of deleterious alleles [fig_ref] Figure 2: Fixation probabilities estimated from Wright-Fisher simulations plotted as a function of inversion... [/fig_ref] : Overview of deterministic fitness and frequency dynamics for initially beneficial inversions of different sizes initially loaded with different numbers of deleterious alleles. : Overview of deterministic fitness and frequency dynamics for initially beneficial inversions of different sizes initially loaded with different numbers of strongly recessive deleterious alleles (h = 0.01). : Comparison of deterministic trajectories of deleterious allele frequencies at W loci on descendent copies of autosomal (q Y I,Aut o W ; black lines) and SLR-expanding (q W Y I , red lines) inversions, divided by their corresponding equilibrium frequency (q). : Fixation probabilities of new SLR-expanding inversions on Y chromosomes plotted as a function of inversion length when deleterious mutations are more strongly recessive (hi = 0.1) than the average dominance coefficient of deleterious mutations estimated from empirical studies [bib_ref] Fitness landscapes: an alternative theory for the dominance of mutation, Manna [/bib_ref] [bib_ref] Mutation load: The fitness of individuals in populations where deleterious alleles are..., Agrawal [/bib_ref] [bib_ref] Gene expression drives the evolution of dominance, Huber [/bib_ref]. : Fixation probabilities of new SLR-expanding inversions on Y chromosomes plotted as a function of inversion length when deleterious mutations are strongly recessive (hi = 0.01). : Reproduction of from [bib_ref] Inbreeding, heterozygote advantage and the evolution of neo-x and neo-y sex chromosomes, Charlesworth [/bib_ref] using the 3-locus model of an SLR-expanding inversion
[fig] Figure 1, Figure 1: illustrates key features of the deterministic dynamics for inversions initially capturing different numbers of partially recessive deleterious mutations for three dominance scenarios (h i = h = {0.25, 0.1, 0.01}). All examples show inversions Illustration of deterministic fitness and frequency dynamics for inversions initially loaded with different numbers of deleterious alleles. Results are shown for inversions of length x = 0.2 three different dominance scenarios (h = {0.25, 0.1, 0.01} corresponding to each column of panels, left to right). Panels (a)-(c) show the fitness of SLR-expanding inversions on a Y chromosome relative to the average fitness of all Y chromosomes (color coded lines). Points indicate when the corresponding inversions dropped below a frequency of 10 − 5, where they became effectively extinct, while stars indicate when an inversion reached a frequency of (1 − 10 −5 ), at which point they were considered to have fixed. Panels (d)-(f) show the inversion frequency dynamics (color coded lines) and illustrate that despite being initially beneficial, lightly loaded inversions (i) are not expected to deterministically rise to high frequencies unless deleterious mutations are strongly recessive; and (ii) will eventually become deleterious and crash to extinction (see also Figs. S1-S3 in Appendix C). Panels (f)-(i) illustrate the deleterious allele frequency dynamics at W loci on the inversion (q W Y I ; red line), and both W and D loci on X chromosomes in ovules/eggs (q W X f and q D X f ; black solid and dotted lines, respectively) for the representative case of inversions initially loaded with relatively few deleterious alleles (r = 1 for G,H; r = 10 for I). Results were generated using the following parameter values: s = 0.01, U = 0.02, x = 0.2, n tot = 10 4 . [/fig]
[fig] Figure 2: Fixation probabilities estimated from Wright-Fisher simulations plotted as a function of inversion length for Autosomal (panels a and b) and SLR-expanding (panels c and d) inversions on Y chromosomes. Point shapes indicate different chromosome-arm wide mutation rates relative to selection (i.e., different values of U), which influence the average numbers of deleterious mutations carried by a standard-arrangement chromosome (U/hs). Dashed horizontal lines indicate the corresponding expected fixation probability for a neutral allele for the same population size, and hence correspond to values of 1/2N for autosomal inversions, and 2/N for Y-linked inversions. Other parameter values were set to: h = 0.25, s = 0.01, n tot = 10 4 . [/fig]
|
Rectal Foreign Body: A Successful Removal at the Bedside and Detailing of a Stepwise Management
Objective:Management of emergency care Background:A rectal foreign body (RFB) can be stigmatizing for patients and present a dilemma for the treating physician. Removal can be challenging owing to the variety of objects introduced. The goals of therapy are to safely remove the RFB and to minimize injury to the bowel.Case Report:A 22-year-old man was referred from a district hospital to our institution after being unable to remove a selfinflicted RFB after sexual gratification. He was hemodynamically stable with a soft and nontender abdomen. A mass was felt in the suprapubic region. Abdominal radiography revealed a well-defined radiolucent object in the pelvic region, which was consistent with a lubricant bottle. No sign of bowel obstruction or perforation was observed. The RFB was successfully retrieved by a combination of transrectal digital manipulation and directed gentle abdominal pressure, allowing for descent of the RFB and transanal traction at the bedside. Various approaches have been described for removal of a RFB, from simple bedside strategies to open surgery for complicated cases. Endoscopy and minimally invasive techniques have also demonstrated a role in formulating a tailored approach.Conclusions:We describe a successful retrieval of an RFB at the bedside, avoiding unnecessary open surgery.
# Background
Rectal foreign body (RFB) insertion is not uncommon, yet it causes significant morbidity for the people involved. Removal of an RFB presents a challenge to the managing physician owing to the variety of objects introduced and difficulty in retrieval. It is common among middle-aged men, but there are cases among women as well. Many reasons underlie RFB insertion, including sexual gratification, concealment, sexual assault, accidental causes, and anal erotism. Among many possible objects, glass bottles and sex devices are the most common rectally inserted objects. The presentation may be immediate because of an individual's inability to remove the object or delayed because of embarrassment or complications such as obstruction, bleeding from the rectum, or colonic perforation. Several techniques have been described for RFB removal, but the shape, size, and location of the object often determine the method for retrieval. Here, we describe a case of successful bedside retrieval of an RFB, avoiding unnecessary open surgery.
## Case report
A 22-year-old man was referred from a district hospital to our institution after being unable to remove a lubricant bottle self-inserted into his rectum. He reported he had inserted the bottle for sexual gratification together with his partner. His previous medical history was unremarkable. Upon initial assessment, he was hemodynamically stable. An abdominal examination demonstrated no peritonitis. However, a cylindrical mass was felt in the suprapubic region. Rectal examination revealed a hard object that could be felt just at the tip of a finger. Abdominal radiographyrevealed a well-defined radiolucent object in the pelvic region, consistent with a lubricant bottle. There was no proximal colon dilatation or pneumoperitoneum.
Manual retrieval via a transanal approach was attempted at bedside. The patient was given 50 μg fentanyl intravenously and placed in a lateral decubitus position. The anal canal and the examiner's fingers were generously lubricated. The edge of the RFB was palpated at the rectum and disimpacted ventrally from the sacral curve. Directed gentle right iliac fossa pressure was applied to assist the foreign body in negotiating the distal fold of Houston and sliding distally to the examiner's fingers. The end of the object was secured with 2 fingers and gentle traction was applied to successfully retrieve the object. The foreign body was a bottle of a lubricant measuring 20 cm long. After the extraction, a digital rectal examination revealed good sphincter tone and proctoscopy showed no mucosal defect. The patient was observed in the surgical ward overnight and was discharged in good condition the next day.
# Discussion
An RFB may pose a challenge for diagnosis and treatment because it is considered taboo, especially in Asian countries. Patients with an RFB are usually ashamed and may conceal the problem owing to stigma and thus not provide an accurate clinical history. Thorough history taking, physical examination, and radiography are warranted if the index of suspicion is high. Our patient was referred to our institution by a district hospital. Such a presentation definitely could not be managed in a district environment, and it required a tertiary hospital for selective treatment options.
Several techniques have been described for the extraction of an RFB. A retained RFB can be classified as high-or lowlying depending on its location relative to the rectosigmoid junction. Clinically, a low-lying RFB is considered when it is palpable on a digital rectal examination, which could allow its extraction at bedside. In stable patients, less-invasive procedures such as transanal extraction by hand or forceps should be attempted first, and if unsuccessful, removal could require surgery. The patient should be placed in a lithotomy position to facilitate abdominal manipulation to apply the Valsalva maneuver. The success rate depends on the size of the clinician's hand and the adequacy of anal sphincter relaxation. Passing a Foley catheter proximal to the object and inflating the balloon to break any suction effect may allow traction of the foreign body as well. We were fortunate that we managed to retrieve the retained RFB at the bedside. Owing to the fact that it was low-lying in lateral decubitus position, it could be grasped by the examiner's hand with concomitant abdominal manipulation.
High-lying retained rectal foreign body usually necessitates endoscopic or surgical intervention. The endoscopic extraction technique involves the use of a flexible endoscope to extract a foreign body that is situated more proximally. This technique provides visualization of the mucosa, and a polypectomy snare can be used to help extract the foreign body. The endoscope can also be used to evaluate for mucosal injuries after successful extraction. A limitation of endoscopic removal is the size of instruments afforded by the endoscope, making removal of a larger RFB significantly more difficult. This limitation can be overcome by the use of a single-incision laparoscopic surgery (SILS) port transanally, thus permitting visualization, insufflation, and deployment of 2 working forceps simultaneously.
An algorithm to simplify the management of an RFB has been proposed by Amran et al. After the initial assessment and diagnosis, patients with signs of perforation should undergo laparotomy with prior resuscitation and antibiotics. Patients without signs of peritonitis may undergo manual evacuation or endoscopic removal if the RFB is located more proximally. Should these attempts fail, removal under general anesthesia can be attempted. Instruments such as forceps, vacuum, or even an SILS approach may aid removal. If all else fails, a laparotomy is warranted. Another indication for surgical intervention is acute bleeding.
Attempt could be made by "milking" the foreign body distally into the lower rectum and to remove it transanally. This procedure was done successfully by Ng et alin a case in which a foreign body similar to that in our case (lubricant bottle) was inserted rectally in a young woman. Fortunately, no mucosal injury, perforation, or contamination occurred in both cases, and resection or colostomy was unnecessary. If conservative measures are unsuccessful, extraction via a controlled colotomy is indicated, followed by primary repair. In the presence of perforation without gross contamination, primary anastomosis of a short-segment resection could be performed in an otherwise healthy bowel wall. In the presence of gross contamination, a Hartmann procedure is advisable.
# Conclusions
A high index of suspicion is needed for accurate diagnosis of an RFB. Early removal of an RFB should be attempted, employing a stepwise approach in an effort to preserve bowel integrity. Available facilities and the location and dimensions of the RFB should be considered when planning for its removal. |
Fast and strong amplifiers of natural selection
## Evolutionary graph theory
Here we formally define Moran Birth-death process on a population structure, the key quantities of fixation probability and timescale of fixation, and the notions of amplifiers and strong amplifiers.
Moran process on a population structure. The population structure is described by a connected graph (network) G = (V, E), where V is a set of |V | = N nodes and E is a set of edges, possibly including self-loops, where each edge uv is assigned a weight w uv . The nodes of G represent sites, each site hosting one individual. The individuals are of two types: residents with fitness normalized to 1, and mutants with (relative) fitness advantage r > 1. The set of nodes occupied by mutants is called a configuration. Given a configuration S we denote by F(S) the total fitness of the population. The edges of G represent connections between adjacent sites, marking where an individual could place its offspring, and the weights represent the strengths of those interactions. The Moran Birth-death process is a stochastic (random) process that acts in discrete time steps until either all nodes are occupied by mutants (fixation occured) or by residents (extinction occurred): At each time step, first ("Birth") an individual is selected randomly proportionally to its fitness and it produces a copy of itself. Denote the corresponding site by u. Second ("death"), a site v adjacent to u is selected randomly proportionally to the weight w uv of the edge uv, and the individual at site v gets replaced by a copy of the individual at site u.
Fixation probability, Extinction probability, Fixation time. Given a graph G with N nodes, a mutant advantage r > 1 and a single node v, we denote by fp(G, r, v) the fixation probability of the Moran Birth-death process, starting with a configuration containing a single mutant with fitness r occupying node v, against a background population of residents with fitness normalized to 1. Similarly, we denote by ep(G, r, v) = 1−fp(G, r, v) the extinction probability.
Concerning the duration of the process, we denote by FT(G, r, v) the (unconditional) fixation time, that is, the (expected) number of steps the process makes until either the mutants or the residents fixate, divided by the population size N . (This quantity, without rescaling by the factor of N , is sometimes also called the absorption time.) Note that the fixation time counts all the steps of the process, including those where the set of nodes occupied by mutants might not have changed due to an offspring replacing an individual of the same type. The argument for rescaling the fixation time by the population size N is that Moran process can be viewed as a discretization of a continuous-time Markov chain where a node with fitness F reproduces after a time given by an exponentially distributed random variable with mean 1/F (by listing only the timepoints in which a reproduction event has happened). In this continuous viewpoint, a population with double the size undergoes twice as many reproduction events per unit time. Thus we prefer to measure the fixation time in generations rather than in individual steps.
Finally, for completeness we briefly mention another quantity that is sometimes used to measure time. We denote by CFT(G, r, v) the conditional fixation time, that is, the (expected) number of steps the process makes until the mutants fixate, conditioning on the fact they do, divided by the population size N . In other words, when computing the conditional fixation time, we average over only those evolutionary trajectories that eventually reach fixation of the mutant. We note that conditional fixation time is typically longer than the (unconditional) fixation time, since the trajectories leading to fixation are typically longer than those leading to extinction. Similarly, for strong amplifiers (defined below), the fixation time and the conditional fixation time are typically asymptotically equivalent, since the fixation event has probability 1.
Uniform and temperature initialization. Averaging over the |V | = N possible locations of the initial mutant, we denote by
[formula] fp(G, r) = 1 N v∈V fp(G, r, v) [/formula]
the fixation probability under uniform initialization. This is meaningful under the assumption that mutations arise spontaneously, with the same rate at each node. If one assumes that mutations arise during reproduction, one prefers to study the weighted average
[formula] fp T (G, r) = 1 N v∈V (L(v) + T (v)) · fp(G, r, v), where L(v) = wvv v w vv [/formula]
is the probability that when a node v reproduces it self-loops, and T (v) = u∈V \{v} wuv v w uv is the temperature of node v. Intuitively, the temperature of a node v counts how many times per generation is v replaced by its neighbors, on average, when all individuals are residents. Therefore the temperature initialization can be thought of as follows: Consider a uniform population of residents, sample a random reproduction event, and then place a mutant on the node being replaced. The extinction probability ep(G, r), ep T (G, r) and the fixation time FT(G, r), FT T (G, r) under uniform and temperature initialization are defined analogously.
Complete graph K N , amplifiers. The well-mixed ("unstructured") population consisting of N individuals is modeled by a complete graph K N . It is known [bib_ref] Population structure determines the tradeoff between fixation probability and fixation time, Tkadlec [/bib_ref] that for any fixed r > 1 we have
[formula] fp(K N , r) = 1 − 1/r 1 − 1/r N → N →∞ 1 − 1/r and FT(K N , r) = (1 + 1/r) · log N + o(log N ) [/formula]
as N → ∞. This gives a natural baseline to which one can compare the other population structures. In particular, given r > 1, a population structure A N with N nodes is called an r-amplifier if it increases the fixation probability as compared to the complete graph of the same size (i.e. fp(A N , r) > fp(K N , r)). A population structure that is an r-amplifier for all r > 1 is called an amplifier. (We note that other authors sometimes further require that a graph A N satisfies fp(A N , r) < fp(K N , r) for all r < 1 to be called an amplifier.)
## Results for large n
Superamplfication and timescale of fixation concern the behavior of population structures in the limit of a large population size N → ∞. First we recall the standard mathematical notation for asymptotic results.
Notation for asymptotic results. In the limit N → ∞, we consider the standard mathematical notation of o(·), O(·), Ω(·), ω(·), and Θ(·) that denote asymptotically strictly less than, less than or equal to, greater than or qual to, strictly greater than, and asymptotically equal to (up to a constant factor), respectively. Finally, we use the approximation sign x ≈ y to state that y = x + o(x).
As four examples, we will write:
[formula] - 1 N = o(1) since lim N →∞ 1 N /1 = 0; - √ N = ω(log N ) since lim N →∞ ( √ N )/(log N ) = ∞; - 1 2 N (N + 1) = Θ(N 2 ) since lim N →∞ 1 2 N (N + 1)/(N 2 ) = 1 2 is a non-zero constant; and - 1 2 N (N + 1) = 1 2 N 2 + 1 2 N ≈ 1 2 N 2 , since 1 2 N = o( 1 2 N 2 ) [/formula]
. For detailed treatment see .
We say that an event happens with high probability if the probability p N that the event happens tends to 1 as N tends to infinity (equivalently, if 1 − p N = o(1)).
Strong amplifiers and Timescale of fixation. In order to compare the behavior of population structures for large N , we consider sequences {G N } ∞ N =1 of graphs of increasing population size. Regarding probability, such a sequence {G N } ∞ N =1 is called a strong amplifier (also known as a superamplifier ) if, for any fixed r > 1, it satisfies fp(G N , r) → N →∞ 1. Strong amplification is, in a sense, the strongest possible form of amplification -an arbitrarily minute (but fixed) fitness advantage r > 1 guarantees that a single mutant with that advantage will fixate with high probability, as the population size N grows large.
Regarding the duration of the process, we define the timescale of fixation by focusing on how the fixation time dependends on N and ignoring both the lower order terms and the possible constants that depend on r but not on N . For instance, for Complete graphs K N we have FT(G N , r) = (1 + 1/r) · log N + o(log N ), hence for any fixed r > 1 we can write FT(G N , r) = Θ(log N ). Formally, given a sequence {G N } ∞ N =1 and r > 1 the timescale of fixation is a function T(G N , r) such that FT(G N , r) = Θ(T(G N , r)). If the same function T(G N , r) applies for all r > 1, as is the case for instance for the Complete graphs, we omit the parameter r and write T(G N ).
## Selection reactors
In this section we introduce Selection reactors. First we describe their underlying (unweighted) graph structure.
Unweighted selection reactor. Given integers N , n the unweighted selection reactor USR N (n) is an unweighted graph with N nodes such that:
1. There are n hub nodes and N − n leaf nodes.
2. Every two hub nodes are connected, every hub node is connected to every leaf node, and every leaf node has a self-loop.
Firing in and out. Due to the symmetry of selection reactors, any step of the process that changes the mutant configuration is one of the following four types:
1. A hub node replaces a leaf node: We say the hub node fires out.
## 2.
A hub node replaces another hub node: We say the hub node does not fire out (or stays).
## 3.
A leaf node replaces a hub node: We say the leaf fires in.
## 4.
A leaf node replaces itself due to a self loop: We say the leaf does not fire in (or loops).
Moreover, we say that a fire-in event is a resident fire-in when the individual reproducing is a resident (and similarly for mutant and/or fire-out events).
Selection reactors. Now we can formally define (weighted) Selection reactors. Given integers N , n and two real numbers p in , p out ∈ (0, 1), the selection reactor SR N (n, p in , p out ) is a weighted graph with N nodes such that:
1. There are n hub nodes and N − n leaf nodes.
2. Every two hub nodes are connected by an edge with weight w h , every leaf node has a self-loop with weight w l , and every hub node is connected to every leaf node by an edge with weight w a = 1. The numbers w h and w l are chosen such that:
(a) For any leaf node, once it is selected for reproduction, it fires in the hub (as opposed to self-looping) with probability p in . In other words, we expect to see (N − n) · p in fire-in events per generation.
(b) For any hub node, once it is selected for reproduction, it fires out to some leaf (as opposed to replacing another hub node) with probability p out . In other words, we expect to see n · p out fire-out events per generation.
It is readily checked that the properties 2a and 2b are satisfied by assigning weight w h = (N −n)(1−p out )/((n−1)p out ) to each edge within the hub, and by assigning weight w l = n · (1 − p in )/p in to each self-loop, respectively.
## Mathematical tools
Here we introduce the mathematical notions and tools required for our proofs.
Martingales. Note that Moran process on a population structure G = (V, E) can be represented as an absorbing Markov chain, whose states correspond to all the possible configurations S ⊆ 2 V (i.e. subsets of nodes occupied by mutants) and whose transition probabilities can be expressed in terms of the weights of the edges. We analyze such Markov chains using the machinery of martingales: Intuitively, a martingale is a "potential function" φ : 2 V → R that assigns a real number to every configuration S ∈ 2 V such that, in expectation, the value of the potential does not change in one step of the Moran process. Formally, we consider the sequence {S t } ∞ t=0 of random variables where S 0 is the (possibly random) initial configuration of the process and S t is the configuration after t steps. We say that a function φ is a martingale if
[formula] E[φ(S t+1 ) | S t ] = φ(S t ) [/formula]
Similarly, we say that a function ψ is a sub-martingale if, in expectation, it does not decrease, that is if
[formula] E[ψ(S t+1 ) | S t ] ≥ ψ(S t ) [/formula]
Example martingale. As an example, consider Moran process running on a complete graph G = K N and any r ≥ 1. Given any configuration S, let p + S (resp. p − S ) be the probability that in the next step we gain (resp. lose) one mutant. Let F(S) = N + (r − 1) · |S| be the total fitness of the population. Then we have
[formula] p + S = r · |S| F(S) · N − |S| N − 1 and p − S = N − |S| F(S) · |S| N − 1 , [/formula]
hence p + S = r · p − S . Now consider a function φ c that counts the number of mutants in the current configuration. Formally φ c is defined as
[formula] φ c (S t ) = |S t |. [/formula]
We claim that φ c is a martingale when r = 1 and a sub-martingale when r > 1: Indeed,
[formula] E[φ c (S t+1 ) | S t ] = φ c (S t ) + p + St · (+1) + p − St · (−1) = φ c (S t ) + (r − 1) · p − St , [/formula]
where the last term is 0 when r = 1 and non-negative when r > 1.
Specific mathematical tools. We make use of the following standard results.
[formula] Proposition 1 (Azuma-Hoeffding inequality). Suppose {S t } ∞ t=0 [/formula]
is a sub-martingale and there exists d ≥ 0 such that |S k+1 − S k | ≤ d for any k ≥ 0. Then for any t ≥ 0 and any real ε > 0 we have
[formula] P[S t − S 0 ≤ −ε] ≤ exp −ε 2 2t · d 2 . [/formula]
Proposition 2 (Markov's inequality). Let X be nonnegative random variable with expectation µ = E[X] and let λ > 0 be a real number. Then
[formula] P[X > λ · µ] ≤ 1 λ . Proposition 3 (Limit for 1/e). Suppose f (N ) = o(N ). Then lim N →∞ (1 − 1/N ) f (N ) = 1. [/formula]
Proposition 4 (Asymptotics of the harmonic series). For any N ≥ 1 we have log N ≤ N i=1 1/i ≤ log N + 1.
## Supplementary note 2: results
Here we restate our two main results from the main text and present the high level ideas behind their proofs. Conceptually, our first main result shows that in order to achieve strong amplification, some asymptotic slowdown is necessary. In other words, there are no strong amplifiers that would be only constant factor slower than complete graphs.
## Theorem 1. for any strong amplifier {g
[formula] N } ∞ N =1 there exists an increasing unbounded function β(N ) such that T(G N ) ≥ log N · β(N ). [/formula]
To complement this, our second main result shows that strong amplification can be achieved with an arbitrarily small slowdown as compared to the complete graph. The slowdown is given by a function β(N ) which can increase arbitrarily slowly as long as it is unbounded.
[formula] Theorem 2. For any increasing unbounded function β(N ) there exists a strong amplifier {G N } ∞ N =1 such that T(G N ) ≤ log N · β(N ). [/formula]
In fact, our two main results (Theorem 1 and Theorem 2) can be viewed as immediate corollaries of two slightly more general theorems (Theorem A and Theorem B, respectively). Next we state those two more general theorems and we provide the high level ideas behind their proofs. The full formal proofs appear in Section 3.
## Proof idea for theorem 1
Theorem 1 follows directly from Theorem A which, for any fixed r > 1 and any amplifier (not necessarily a strong amplifier), gives a lower bound on fixation time that is inversely proportional to the extinction probability.
Theorem A. Fix r > 1.
## If g is an r-amplifier under uniform initialization then
[formula] FT(G, r) ≥ (r − 1) 2 r 4 (r + 1) · 1 ep(G, r) · log N · r − 1 r + 1 − 1 = Ω 1 ep(G, r) · log N . [/formula]
2. If G is an r-amplifier under temperature initialization then
[formula] FT T (G, r) ≥ (r − 1) 2 r 4 (r + 1) · 1 ep T (G, r) · log N · (r − 1) 2 r(r + 1) 2 − 1 = Ω 1 ep T (G, r) · log N . [/formula]
Note that this indeed immediately implies Theorem 1:
[formula] When {G N } ∞ N =1 [/formula]
is a strong amplifier, then for any fixed r > 1 the extinction probability tends to zero in the limit N → ∞, thus FT(G N , r) = Ω 1 ep(G,r) · N log N = ω(N log N ) and the slowdown as compared to the complete graph is asymptotically non-negligible.
Proof Idea. The proof of Theorem A rests on two ideas. The first idea is that the fixation time on any graph G can be bounded from below in terms of temperatures of its nodes (see Lemma 1) by noting that, on any trajectory to fixation, each node (except for the node where the initial mutant appeared) has to be replaced at least once by one of its neighbors: In particular, if many nodes have a relatively low temperature ("cold nodes") then the fixation time is relatively long since we must wait until each of the cold nodes is replaced at least once (this corresponds to a coupon-collector-like process with many rare coupons).
The second idea is that when fp(G, r) is relatively high then there exist many initial nodes u for which ep(G, r, u) is relatively low. However, when ep(G, r, u) is low then T (u) has to be low too -otherwise u would be likely to be replaced by one of its neighbors before reproducing even once. Thus any graph G with high fixation probability inevitably contains many cold nodes and the general bound given by Lemma 1 becomes particularly strong for strong amplifiers.
## Proof idea for theorem 2
Theorem 2 immediately follows from Theorem B which states that selection reactors with suitably tuned parameters are strong amplifiers whose fixation time is asymptotically arbitrarily close to the fixation time of the complete graphs. We believe that the proof of Theorem B can be modified to apply to sparse incubators introduced in [5] but for convenience in what follows we focus solely on Selection Reactors. Specifically, for any function α : N → N that is both o(N 1/2 ) and ω(1), and for any integer N , let
[formula] SR α N = SR N (n = N/α(N ) , p in = 1/α 2 (N ), p out = 1/α 2 (N )) [/formula]
be the selection reactor with n = N/α N hub nodes, = N − n ≈ N leaf nodes in the periphery, and with fire-in and fire-out rates both equal to p in = p out = 1/α 2 (N ). With this notation we have the following theorem.
Theorem B (Reactors are fast strong amplifiers). For any function α : N → R that is both o(N 1/2 ) and ω(1), and for any fixed r > 1, the selection reactors {SR α } ∞ N =1 satisfy the following, under both uniform and temperature initialization:
[formula] 1. (strong amplifiers) fp(SR α N , r) → N →∞ 1. 2. (fast) FT(SR α N , r) = O(log N · α 5 (N )). [/formula]
Note that this indeed immediately implies Theorem 2 as one can take α(N ) = 5 β(N ). Before we present the high level idea behind the proof of Theorem B, we need to introduce more notions. Recall that once a leaf node is selected for reproduction, it fires in the hub with probability p in . Similarly, a hub node, once selected, fires out with probability p out . We will make extensive use of a key parameter h = p in /n p out / = · p in n · p out that characterizes the bias towards the hub along any fixed edge that connects a leaf node with a hub node. Our proofs rely on the fact that h = ω(1) and on several other properties, namely:
[formula] n = ω(1), n = o(N ), p out = o(1), N p in = o(n), h = o(n). [/formula]
For the parameters stated in Theorem B, we have = Θ(N ) and p in = p out = 1/α 2 (N ) and thus h = Θ(α(N )) and all the properties are satisfied.
Proof idea for probability. Here we sketch the intuition behind our proof that Selection reactors from Theorem B are strong amplifiers. The argument proceeds in five steps. We show that:
1. Whenever the hub contains j ∈ [1, n/2] mutants, the size of the mutant population in the hub is more likely to increase than to decrease, see Lemma 4. This is because the hub by itself is biased towards gaining mutants (due to r > 1), and the leaves do not change this as they interact with the hub only rarely.
2. With high probability, the initial mutant appears at a leaf (for both uniform and temperature initialization), see Lemma 5. This is a simple computation.
3. With high probability, that initial mutant repeatedly fires in the hub before being eliminated and, by the result of step 1, this establishes a mutant subpopulation in the hub that has a superconstant size, see Lemma 6.
4. By Item 1 that subpopulation of mutants in the hub grows to size n/2, with high probability, see Lemma 7. This is because the hub is biased towards gaining mutants.
## 5.
Once there are j ≥ n/2 mutants in the hub, fixation on the whole graph occurs with high probability, see Lemma 8. This is because, at this point, the mutants have established such a large subpopulation that an extinction is extremely unlikely.
Proof idea for time. Here we sketch the intuition behind our proof that Selection reactors from Theorem B are fast strong amplifiers. Recall that a configuration S is the subset of nodes that are occupied by mutants. We say that a step of the process is active if it changes the configuration (due to a mutant replacing a resident or vice versa).
The idea is to decompose the process into the active process that tracks the active steps and the waiting process that tracks the steps in which the configuration S stays the same (either due to a node self-looping or due to a node replacing a neighbor of the same type). The argument then proceeds in four steps:
1. We define a certain potential function ψ that assigns a non-negative integer to each configuration and satisfies ψ(S) = O( · h) for any configuration S.
## 2.
For the active process, we argue that, in expectation, the value of the potential function increases by at least a constant in each active step (and always by at most O(h) in one step). This then implies that, in expectation, the active process attains each potential value k only a few times (namely R k = O(h 2 ) times), see Lemma 9 and Lemma 10.
3. For the waiting process, we establish an upper bound on the number W k of (waiting) steps that the process makes until an active step occurs, when the current configuration S is assigned potential value k = ψ(S), see Lemma 11 and Lemma 12.
4. By linearity of expectation, this implies that, in expectation, the process makes in total at most k R k · W k = O(h 2 ) · k W k steps. Summing up our upper bounds for W k from previous item gives the result.
## Supplementary note 3: proofs
In the next two subsections we present the formal proofs of our two main theorems.
## No superamplifiers are as fast as the complete graph
The goal of this section is to prove Theorem A. First, we present a lower bound on the fixation time for an arbitrary graph G in terms of the temperatures of its nodes. The bound is based on the fact that, on any trajectory to fixation, each node has to be replaced by one of its neighbors at least once. This reduces to a coupon collector process where nodes with lower temperature correspond to coupons that are more rare. Flajolet et al [bib_ref] Birthday paradox, coupon collectors, caching algorithms and selforganizing search, Flajolet [/bib_ref] found a closed form expression for the expected number of steps of such a coupon collector process, for any given list of temperatures. Here we present a simplified (weaker) bound that takes into account only those nodes whose temperature is lower than a given threshold ("cold" nodes). The proof idea is similar to the proof of Theorem 1 in [bib_ref] Population structure determines the tradeoff between fixation probability and fixation time, Tkadlec [/bib_ref]. Lemma 1. Let G be a graph, u a node of G and r > 1. Suppose that at least k nodes of G different from u have temperature at most t each. Then FT(G, r, u) ≥ fp(G, r, u) · log k r · t .
Proof. Consider a modified Moran process M that is identical with the standard Moran process with a mutant initialized at u, except that whenever the mutants go extinct we instead again initialize a single mutant at u and continue the process. Clearly, the modified process M always terminates with the mutants fixating. We claim that the expected fixation time is given by FT (G, r, u) = 1 fp(G,r,u) · FT(G, r, u): Indeed, for M denote by p = fp(G, r, u) the fixation probability and by E = ET(G, r, u), C = CFT(G, r, u), T = FT(G, r, u), the (expected) extinction time, conditional fixation time, and the fixation time, respectively. Then we have T = p · C + (1 − p) · E. Denoting by T = FT (G, r, u) the fixation time of M , by linearity of expectation we have
[formula] T = p · C + (1 − p) · (E + T ), [/formula]
since the first mutant either fixates, in which case M does C steps on average and then terminates (probability p), or goes extinct, in which case M does E steps on average and then restarts (probability 1 − p). This now rewrites as the desired
[formula] T = p · C + (1 − p) · E 1 − (1 − p) = T p . [/formula]
Let C ⊆ V be a set of k nodes with temperature at most t each (the "cold" nodes). Since the process M eventually reaches fixation, for each i = 0, . . . , k − 1 it transitions from some configuration with i mutants in C to some configuration with i + 1 mutants in C at least once. Any time the process is in a configuration with i mutants in C, the probability p + i that it successfully gains another mutant in C satisfies p + i ≤ r N · (k − i) · t , because each of the remaining k − i resident nodes in C has temperature at most T (v) ≤ t and thus it is replaced with probability at most replace(v) ≤ r N t . Therefore the expected number of generations spent in such configurations is
[formula] t i ≥ 1 N · 1/p + i ≥ 1/(r(k − i) · t ) [/formula]
. By linearity of expectation, summing over i = 0, . . . , k − 1 we get
[formula] FT(G, r, u) = fp(G, r, u) · FT (G, r, u) ≥ fp(G, r, u) · k−1 i=0 t i ≥ fp(G, r, u) · k−1 i=0 1 r(k − i) · t ≥ fp(G, r, u) · log k r · t . [/formula]
Remark. We remark that Lemma 1 yields lower bounds on fixation time for various specific graph families and initialization schemes. For instance, star graphs S N contain N −1 = Θ(N ) nodes of temperature 1/(N −1) = Θ(1/N ) each, and therefore for each node u we have
[formula] FT(S N , r, u) ≥ fp(S N , r, u) · log(N − 2) r · 1 N −1 = fp(S N , r, u) · Ω(N log N ). [/formula]
For uniform initialization we have fp(S N , r) → 1 − 1/r 2 = Θ(1), thus averaging over the initial nodes we obtain FT(S N , r) = Ω(N log N ) which is asymptotically tight [bib_ref] Population structure determines the tradeoff between fixation probability and fixation time, Tkadlec [/bib_ref]. Similarly, dense incubator D N (see [bib_ref] Asymptotically optimal amplifiers for the moran process, Goldberg [/bib_ref] for their definition) are superamplifiers that contain Θ(N ) nodes of temperature Θ(N −2/3 ) each, thus FT(D N , r) = Ω(N 2/3 log N ).
In order to prove Theorem A using Lemma 1 we aim to show that any superamplifier has many nodes with low temperature. To that end, we use two lemmas. The first one (Lemma 2) allows us to look for nodes u with small extinction probability ep(G, r, u) (instead of nodes with low temperature T (u)). The second one (Lemma 3) states that for any graph G that amplifies (under either uniform or temperature initialization), a constant portion of its nodes have a small extinction probability ep(G, r, u). Lemma 2. Fix r > 1. Let G be a graph and u its node. Denote by ep u = ep(G, r, u) the extinction probability of a single mutant starting at node u. Then
[formula] T (u) ≤ r · ep u ·(1 − L(u)) 1 − ep u ≤ r · ep u 1 − ep u . [/formula]
Proof. Assume that the configuration of the Moran process at some step t is S t = {u}. Let p (resp., q) be the probability that |S t+1 | = 2 (resp., |S t+1 | = 0). The event |S t+1 | = 2 occurs when u is selected for reproduction and it does not self-loop. The event |S t+1 | = 0 occurs when u is replaced by one of its neighbors. Hence
[formula] p = r F(S t ) · (1 − L(u)) and q = v =u 1 F(S t ) · w vu v w vv = 1 F(S t ) · T (u). [/formula]
Note that if |S t+1 | = 2 and |S t+1 | = 0, then we have S t+1 = S t = {u}, therefore the extinction probability from u is lower-bounded by
[formula] ep u ≥ q q + p ≥ T (u) T (u) + r · (1 − L(u)) [/formula]
.
Clearing the denominators and using 1 − L(u) ≤ 1 we get the desired bounds.
Lemma 3. Fix r > 1 and set c r = 1 2 (r + 1) > 1. Consider a graph G.
## (uniform initialization) let
[formula] S = {u ∈ V : ep(G, r, u) ≤ c r · ep(G, r)}. Then |S| ≥ N · r−1 r+1 . 2. (Temperature initialization) Similarly, let S T = {u ∈ V : ep(G, r, u) ≤ c r · ep T (G, r)}. If G is an r-amplifier under temperature initialization then |S T | ≥ N · (r−1) 2 r(r+1) 2 . [/formula]
Proof. We prove the two claims independently (but along the same lines).
1. Let X be a random variable that denotes the extinction probability of a node of V chosen uniformly at random. Note that X is non-negative, and
[formula] E[X] = u∈V 1 N · ep(G, r, u) = ep(G, r). [/formula]
By Markov's inequality, we have
[formula] P[X > c r · ep(G, r)] ≤ E[X] c r · ep(G, r) = 1 c r . [/formula]
Finally, note that
[formula] 1 − 1 c r ≤ P[X ≤ c r · ep(G, r)] = |S| |V | = |S| N [/formula]
and thus |S| ≥ N · (1 − 1/c r ) = N · r−1 r+1 as claimed. 2. Likewise, let Y be a random variable that denotes the extinction probability of a node of V chosen according to the temperature initialization. Proceeding as before,
[formula] E[Y ] = 1 N u∈V (T (u) + L(u)) · ep(G, r, u) = ep T (G, r) [/formula]
and since Y is non-negative, by Markov's inequality, we have
[formula] P[Y > c r · ep T (G, r)] ≤ 1 c r . hence 1 − 1 c r ≤ P[Y ≤ c r · ep T (G, r)] = 1 N u∈S T T (u) + L(u) . [/formula]
Fix u ∈ S T and denote by ep u = ep(G, r, u) its extinction probability. By definition of S T we have ep u ≤ c r · ep T (G, r), and since G is an amplifier under temperature initialization, we have ep T (G, r) ≤ 1/r, therefore ep u ≤ r+1 2r . Plugging this into the bound given by Lemma 2 we obtain
[formula] T (u) + L(u) ≤ r · ep u ·(1 − L(u)) 1 − ep u + L(u) (by Lemma 2) ≤ r · r+1 2r (1 − L(u)) 1 − r+1 2r + L(u) (since it is increasing in ep u ) = r(r + 1) r − 1 − L(u) · r(r + 1) r − 1 − 1 ≥ 0 for any r > 1 ≤ r(r + 1) r − 1 . [/formula]
This yields
[formula] r − 1 r + 1 = 1 − 1 c r ≤ 1 N u∈S T T (u) + L(u) ≤ 1 N · |S T | · r(r + 1) r − 1 , [/formula]
hence |S T | ≥ N · (r−1) 2 r(r+1) 2 as claimed.
We are now ready to prove Theorem A.
Theorem A. Fix r > 1.
## If g is an r-amplifier under uniform initialization then
[formula] FT(G, r) ≥ (r − 1) 2 r 4 (r + 1) · 1 ep(G, r) · log N · r − 1 r + 1 − 1 = Ω 1 ep(G, r) · log N . [/formula]
2. If G is an r-amplifier under temperature initialization then
[formula] FT T (G, r) ≥ (r − 1) 2 r 4 (r + 1) · 1 ep T (G, r) · log N · (r − 1) 2 r(r + 1) 2 − 1 = Ω 1 ep T (G, r) · log N . [/formula]
Proof. Both claims are proved in essentially the same way.
1. Let c r = r+1 2 > 1 and consider the set
[formula] S = {u ∈ V : ep(G, r, u) ≤ c r · ep(G, r)} [/formula]
of nodes with not too high an extinction probability ("cold nodes"). By Lemma 3 we have |S| ≥ N · r−1 r+1 . Moreover, for any node u ∈ S we have
[formula] T (u) ≤ r · ep(G, r, u) 1 − ep(G, r, u) (by Lemma 2) ≤ r · c r · ep(G, r) 1 − c r · ep(G, r) (since u ∈ S) ≤ r · r+1 2 · ep(G, r) 1 − r+1 2 · 1 r (since ep(G, r) ≤ 1/r) = r 2 (r + 1) r − 1 · ep(G, r), [/formula]
where throughout the computation all denominators are positive. Therefore we can apply Lemma 1 to all nodes
[formula] u ∈ V with k = |S| − 1 = N · r−1 r+1 − 1 and t = r 2 (r+1) r−1 · ep(G, r) to get FT(G, r) = 1 N u∈V FT(G, r, u) ≥ 1 N u∈V fp(G, r, u) · log k r · t (by Lemma 1) = fp(G, r) · log k r · t ≥ 1 − 1 r · log(|S| − 1) r · r − 1 r 2 (r + 1) · ep(G, r) (since fp(G, r) ≥ 1 − 1/r) = (r − 1) 2 r 4 (r + 1) · 1 ep(G, r) · log N · r − 1 r + 1 − 1 = Ω 1 ep(G, r) · log N . [/formula]
2. Likewise, let c r = r+1 2 > 1 and consider the set
[formula] S T = {u ∈ V : ep(G, r, u) ≤ c r · ep T (G, r)} [/formula]
of nodes with not too high an extinction probability ("cold nodes"). By Lemma 3 we have |S T | ≥ N · (r−1) 2 r(r+1) 2 . By exactly the same computation as before, for any node u ∈ S we have
[formula] T (u) ≤ r · ep(G, r, u) 1 − ep(G, r, u) (by Lemma 2) ≤ r · c r · ep T (G, r) 1 − c r · ep T (G, r) (since u ∈ S T ) ≤ r · r+1 2 · ep T (G, r) 1 − r+1 2 · 1 r (since ep T (G, r) ≤ 1/r) = r 2 (r + 1) r − 1 · ep T (G, r), [/formula]
where again all denominators are positive.
Therefore we can apply Lemma 1 to all nodes u
[formula] ∈ V with k = |S T |−1 = N · (r−1) 2 r(r+1) 2 −1 and t = r 2 (r+1) r−1 ·ep T (G, r) to get FT T (G, r) = 1 N u∈V T (u) + L(u) · FT(G, r, u) ≥ 1 N u∈V T (u) + L(u) · fp(G, r, u) · log k r · t (by Lemma 1) = fp T (G, r) · log k r · t ≥ (r − 1) 2 r 4 (r + 1) · 1 ep T (G, r) · log N · (r − 1) 2 r(r + 1) 2 − 1 = Ω 1 ep T (G, r) · log N . [/formula]
## Selection reactors are superamplifiers
The goal of this section is to prove Item 1 of Theorem B. Let S i,j be a configuration consisting of i ∈ [0, ] mutants among the leaves and j ∈ [0, n] mutants in the hub.
The following lemma states that whenever the hub consists mostly of residents, we are at least a constant r = 1 2 (r + 1) > 1 more likely to gain a mutant there rather than to lose one. It holds since p out = o(1) and · p in = o(n). Lemma 4. Consider the Moran process starting from a configuration S i,j with 1 ≤ j ≤ n/2. Let S i ,j be the first subsequent configuration with j = j. Then, for all large enough , we have
[formula] P[j = j + 1] P[j = j − 1] ≥ 1 2 (r + 1) > 1. [/formula]
Proof. Let p + (i, j) (resp. p − (i, j)) be the probability that when in configuration S i,j , the number j of mutants in the hub increases (resp. decreases) in a single step. Let F = F(S i,j ) = N + (r − 1)(i + j) be the total fitness. Straightforward computation gives
[formula] p + (i, j) ≥ rj F · (1 − p out ) · n − j n and p − (i, j) ≤ n − j F · (1 − p out ) · j n A + F · p in · j n B [/formula]
where in p + (i, j) we counted only the "within-hub" interactions and in p − (i, j) we bounded by the worst case when all leaf nodes are residents. We claim that B = o(A): Indeed, we have n − j ≥ 1 2 n and 1 − p out = 1 − 1/α(N ) 2 = Θ(1), hence the claim is equivalent with p in = o(n) which is true in our case as p in = /α(N ) 2 and n = /α(N ). Therefore,
[formula] lim N →∞ p + (i, j) p − (i, j) ≥ lim N →∞ rj F · (1 − p out ) · n−j n n−j F · (1 − p out ) · j n = r. [/formula]
In particular, for all large enough N the ratio is higher than a smaller positive constant 1 2 (r + 1) and thus also
[formula] P[j = j + 1] P[j = j − 1] ≥ min i p + (i, j) p − (i, j) ≥ 1 2 (r + 1). [/formula]
The next lemma states that, with high probability, the first mutant appears in a leaf node. It holds since n = o( ) and p in = o(1).
## Lemma 5.
Under both uniform and temperature initialization, the first mutant is initialized at a leaf node with high probability.
Proof. This is straightforward computation. For uniform initialization, the first mutant appears at a leaf with probability /N → N →∞ 1.
Under temperature initialization, any leaf node u satisfies L(u) + T (u) ≥ L(u) = 1 − p in → N →∞ 1, hence the overall fixation probability is at least /N → N →∞ 1 too.
The next lemma states that, with high probability, the initial mutant produces a mutant subpopulation in the hub of a superconstant size. It holds since h = ω(1) and h = o(n 3 ). Lemma 6. Consider the Moran process starting from a configuration S i,j with i ≥ 1. Then, with high probability, the process reaches a configuration S i ,j with j ≥ h 1/3 .
## Proof. fix a leaf u that hosts a mutant in the initial configuration
[formula] S i,j . [/formula]
First, we claim that, with high probability, u fires in the hub √ h times before being replaced by a resident: The probability p + that in a single step u is selected for reproduction and fires in the hub satisfies p + ≥ 1 N · p in . The probability p − that in a single step u is replaced by a resident firing out from the hub satisfies p − ≤ n N · p out · 1 . Hence as long as u contains a mutant, the probability that u fires in the hub before being replaced by a resident satisfies
[formula] p = p + p + + p − ≥ pin npout 1 + pin npout = h 1 + h > 1 − 1 h . [/formula]
The probability q that u fires in the hub at least √ h times before being replaced satisfies q ≥ p
[formula] √ h = (1 − 1/h) √ [/formula]
h . Since h = ω(1), by Proposition 3 this event indeed happens with high probability as claimed.
Next, we claim that after √ h mutant fire-ins, the hub contains at least h 1/3 mutants, with high probability. Note that this claim is not obvious, for at least two reasons: First, if the hub was biased towards losing mutants (which it is not), it could happen that each invading mutant gets swiftly eliminated. Second, if the hub was very small (e.g. a constant size), it could happen that the fire-ins from u repeatedly replace the same node.
To prove the second claim, run the process until either the hub contains at least h 1/3 mutants, or u has fired in the hub √ h times. Let t be the (random, stopping) time when the earlier of those two conditions occurs and let p t be the probability that the configuration S i ,j at time t still satisfies j < h 1/3 . Consider the potential φ(S i,j ) = j that counts the number of mutants in the hub. As in the proof of Lemma 4, when j ≤ h 1/3 ≤ n/2, the hub together with all the resident fire-ins is still biased towards gaining mutants (or not biased either way when j = 0). Hence in steps when u does not fire in, the potential φ does not decrease in expectation. Moreover, in steps when u does fire in, the potential increases in expectation by at least (n − j)/n ≥ 1 − h 1/3 /n = 1 − o(1), where the last equality holds due to h = o(n 3 ). Hence
[formula] h 1/3 ≥ E[φ(S i ,j )] = (1 − p t ) · h 1/3 + p t · φ(Si,j) ≥0 + √ h · (1 − o(1)) =ω(h 1/3 ) ≥ (1 − p t ) · h 1/3 + p t · ω(h 1/3 ), thus p t = o(1) as desired. [/formula]
The next lemma states that once the hub contains a superconstant number of mutants, mutants eventually gain majority in the hub, with high probability. It holds since h = ω(1) and n = ω(1). Lemma 7. Let S i,j be a configuration with j ≥ h 1/3 . Then, with high probability, the process reaches a configuration S i ,j with j ≥ n/2.
Proof. By Lemma 4, for any j ∈ [1, n/2] the mutant subpopulation in the hub is a constant r = 1 2 (r + 1) more likely to increase rather than decrease. Hence the probability p that, starting with j mutants, the mutant subpopulation in the hub increases to size n/2 before being eliminated, can be lower-bounded by the absorption probability of a 1-dimensional random walk with a constant forward bias r as
[formula] p ≥ 1 − 1/(r ) j 1 − 1/(r ) n/2 . [/formula]
Since both j and n/2 are ω(1), we have
[formula] lim N →∞ 1 − 1/(r ) j 1 − 1/(r ) n/2 = 1 1 = 1. [/formula]
The next lemma states that once the mutants have majority in the hub, fixation on the whole graph occurs with high probability. It holds since h = ω(1) and h = o(n). Lemma 8. Let S i,j be a configuration with j ≥ n/2. Then, with high probability, the process reaches the configuration S ,n .
Proof. We will show that, in fact, once we have j = ω(h) mutants in the hub, fixation on the whole graph occurs with high probability (this suffices since h = o(n)). For any configuration S i,j , define its "star-like" potential by
[formula] φ(S i,j ) = β i · γ j , [/formula]
where β = h+r hr 2 +r → 1 r 2 and γ = hr+1 hr+r 2 → 1 are positive real numbers less than one. In particular, we have φ(S 0,0 ) = 1 and φ(S ,n ) → N →∞ 0. Also, since j = ω(h) we have φ(S i,j ) ≤ γ j = (1 − Θ(1/h)) ω(h) → N →∞ 0. As we verify below, the potential is chosen such that it does not change in expectation when the reproduction event happens along an edge that connects a hub node with a leaf node. Moreover, for reproduction events that happen along edges within the hub, the potential decreases in expectation. Thus the function φ is a sub-martingale and by Doob's optional stopping theorem, the expectation φ ∞ = E[φ(S T )] of the potential at the (random, stopping) time T when the process terminates satisfies
[formula] φ(S i,j ) ≥ φ ∞ = fp i,j ·φ(S ,n ) + (1 − fp i,j ) · φ(S 0,0 ), [/formula]
where fp i,j is the fixation probability starting from the configuration S i,j . Rearranging the terms and dividing by the (positive) expression φ(S 0,0 ) − φ(S ,n ), this gives
[formula] fp i,j ≥ φ(S 0,0 ) − φ(S i,j ) φ(S 0,0 ) − φ(S ,n ) → N →∞ 1 − 0 1 − 0 = 1. [/formula]
To verify our claim, it suffices to check contributions of three different types of edges:
(a) "within-hub" edges connecting a mutant and a resident: When this edge is used for reproduction, it is r times more likely that we gain a mutant rather than lose it. Denoting the current potential by x, we need to show that r r + 1
x · γ + 1 r + 1
x
[formula] · 1 γ ≤ x. [/formula]
This reduces to (1 − γ)(rγ − 1) ≥ 0 which is true because the first term is positive and the second one is positive for any large enough population size (recall that h = ω(1)).
(b) "resident leaf -mutant hub" edge: The edge is used in the direction towards the hub with probability p 1 = 1 F · p in · 1 n and in the direction towards the leaf with probability p 2 = r F · p out · 1 . Hence we need to verify that
[formula] p 1 p 1 + p 2 · x · 1 γ + p 2 p 1 + p 2 · x · β = x. [/formula]
## This reduces to showing
[formula] h h + r · 1 γ + r h + r · β = 1 [/formula]
which is true since the left-hand side rewrites as h h + r · hr + r 2 hr + 1 + r h + r · h + r hr 2 + r = hr(h + r) + (h + r) (h + r)(hr + 1) = 1.
(c) "mutant leaf -resident hub" edge: Similarly as in the previous case, we get p 1 = r F · p in · 1 n and p 2 = 1 F · p out · 1 . Then we need to verify hr hr + 1 · γ + 1 hr + 1
[formula] · 1 β = 1 [/formula]
which is done by rewriting the left-hand side as
[formula] hr hr + 1 · hr + 1 hr + r 2 + 1 hr + 1 · hr 2 + r h + r = h h + r + r h + r = 1. [/formula]
We are now ready to prove Item 1 of Theorem B.
Theorem B (Reactors are fast strong amplifiers). For any function α : N → R that is both o(N 1/2 ) and ω(1), and for any fixed r > 1, the selection reactors {SR α } ∞ N =1 satisfy the following, under both uniform and temperature initialization:
[formula] 1. (strong amplifiers) fp(SR α N , r) → N →∞ 1. 2. (fast) FT(SR α N , r) = O(log N · α 5 (N )). [/formula]
Proof of Theorem B, Item 1. Each of the following four steps happens with high probability: By Lemma 5, the first mutant appears at a leaf node. From there, by Lemma 6, this mutant establishes a mutant subpopulation in the hub with superconstant size j ≥ h 1/3 . From there, by Lemma 7, the mutant subpopulation in the hub grows until the mutants have majority in the hub. From there, by Lemma 8, the mutants fixate.
## Selection reactors are fast superamplifiers
The goal of this section is to prove Item 2 of Theorem B. Denote the configuration with i mutant leaf-nodes and j mutant hub-nodes by S i,j . Let h 0 = h be the parameter h = ( · p in )/(n · p out ) rounded down to the nearest integer. We define ψ(S i,j ) = h 0 · i + j. Intuitively, each mutant leaf-node is worth h 0 , each mutant hub-node is worth 1, and the potential of a configuration is the total worth of its mutant nodes. The potential function ψ is a rescaled and rounded version of the function ψ (S) = v∈S 1/ deg(v) which was used in several earlier works such as [bib_ref] Approximating fixation probabilities in the generalized moran process, Díaz [/bib_ref] [bib_ref] Phase transitions of the moran process and algorithmic consequences, Goldberg [/bib_ref].
First we show that, in expectation, each active step increases the potential by at least a constant c (that depends on r > 1). The claim is essentially the same as Lemma 2 inand Lemma 67 in [bib_ref] Phase transitions of the moran process and algorithmic consequences, Goldberg [/bib_ref]. Lemma 9. Fix a selection reactor SR( , n, p in , p out ) and r > 1. Fix a configuration S different from S 0,0 and S ,n . Sample a random active step of the process and denote the new configuration by S = S. Let
[formula] c 1 = r − 1 r + 1 , c 2 = r − 1 − r 2 h + r = r − 1 − O(1/h), c 2 = r − 1 r − r − 1 rh + 1 = r − 1 r − O(1/h), [/formula]
be positive constants and let c = min{c 1 , c 2 , c 3 }. Then
[formula] E[ψ(S ) | S] ≥ ψ(S) + c. [/formula]
Proof. We say that an edge is active if its two endpoints are occupied by one mutant and one resident. There are three types of active edges:
1. Type 1 edges ("hub-hub") that connect a mutant in the hub to a resident also in the hub;
2. Type 2 edges ("hub-leaf") that connect a mutant in the hub to a resident in a leaf; and 3. Type 3 edges ("leaf-hub") that connect a mutant in a leaf to a resident in the hub.
Each active edge can be used during a replacement event in either of its two directions. Fix an active edge uv with u a mutant and v a resident and suppose it has type i (where 1 ≤ i ≤ 3). Below we show that, assuming the edge uv is selected for replacement (in one of its two directions), the potential increases at least by c i in expectation. Since in any one step of the active process, some active edge is used for replacement, we conclude that each active step increases the potential by at least min i {c i } = c, in expectation. We treat the three types of edges separately. For two nodes x, y let p x→y be the probability that in a single step of the Moran process, node x replaces node y. Let F = + n + (i + j)(r − 1) be the total fitness of the population at a configuration S i,j .
1. Type 1 edges ("hub-hub"): If the mutant at a hub-node u replaces a resident at hub-node v, the potential increases by 1, otherwise it decreases by 1. We have
[formula] p u→v = rj F · (1 − p out ) · n − j n and p v→u = n − j F · (1 − p out ) · j n , [/formula]
thus, in expectation, the potential increases by
[formula] p u→v · (+1) + p v→u · (−1) p u→v + p v→u = r − 1 r + 1 = c 1 . [/formula]
2. Type 2 edges ("hub-leaf"): If the mutant at a hub-node u replaces a resident at leaf-node v, the potential increases by h 0 , otherwise it decreases by 1. We have
[formula] p u→v = rj F · p out · − i and p v→u = − i F · p in · j n , [/formula]
thus, in expectation, the potential increases by
[formula] p u→v · (+h 0 ) + p v→u · (−1) p u→v + p v→u = h 0 · r − h h + r ≥ (h − 1)r − h h + r = r − 1 − r 2 h + r = c 2 [/formula]
3. Type 3 edges ("leaf-hub"): If the mutant at a leaf-node u replaces a resident at hub-node v, the potential increases by 1, otherwise it decreases by h 0 . We have
[formula] p u→v = ri F · p in · n − j n and p v→u = n − j F · p out · i , [/formula]
thus, in expectation, the potential increases by
[formula] p u→v · (+1) + p v→u · (−h 0 ) p u→v + p v→u = hr − h 0 hr + 1 ≥ hr − h hr + 1 = r − 1 r − r − 1 hr + 1 = c 3 . [/formula]
In particular, for any fixed r > 1 and as N → ∞ we have h → ∞ and thus c = c 1 = (r − 1)/(r + 1) is a constant that only depends on r.
Let R k be the expected number of times the active process attains a potential value k. The next lemma bounds R k from above. Lemma 10. Fix an integer k. Run the active process starting from any configuration. Then the expected total number of visits to configurations T that satisfy ψ(T ) = k is O(h 2 ).
Proof. Run the active process until the potential value k is attained for the first time (if it is never attained, the claim is trivial). Denote the corresponding configuration by S 0 and for t ≥ 0 let S t be a random variable denoting the configuration after t active steps, starting from S 0 (S 0 = S 0 ). Consider a function ξ(S t ) = ψ(S t ) − c · t, where c is the positive constant from Lemma 9. We claim that the function ξ is a sub-martingale with differences bounded by h 0 + c: Indeed, since
[formula] E[ξ(S t+1 ) | S t ] = E[ψ(S t+1 ) − c · (t + 1) | S t ] = E[ψ(S t+1 ) | S t ] − c · (t + 1) ≥ ψ(S t ) + c − c · (t + 1) (by Lemma 9), = ψ(S t ) − c · t = ξ(S t ), [/formula]
the sequence is a sub-martingale and since
[formula] |ξ(S t+1 ) − ξ(S t )| = |ψ(S t+1 ) − ψ(S t ) − c| ≤ |ψ(S t+1 ) − ψ(S t )| + c ≤ h 0 + c, [/formula]
its differences are upper-bounded by h 0 + c. By Azuma's inequality [bib_ref] Weighted sums of certain dependent random variables, Azuma [/bib_ref] we can bound the probability that after precisely t steps the potential is at most its initial value as
[formula] P[ψ(S t ) ≤ ψ(S 0 )] = P[ξ(S t ) − ξ(S 0 ) ≤ −ct] ≤ exp −c 2 t 2 2t · (c + h 0 ) 2 ≤ e −c 2 2(c+h 0 ) 2 t . [/formula]
Summing the infinite geometric series over t = 0, . . . , ∞ we obtain an upper bound on the expected number of visits to states T with φ(T ) ≤ φ(S 0 ), and thereby also on the number of visits to states with φ(T ) = φ(S 0 ). Specifically, we denote the exponent by x = −c 2 2(c+h0) 2 and note that since x ∈ (−1/2, 0), we have e x ≤ 1 + x/2. Upon summing the series we obtain the desired
[formula] ∞ t=0 (e x ) t = 1 1 − e x ≤ 1 1 − (1 + x/2) = − 2 x = 4 (c + h 0 ) 2 c 2 = O(h 2 ). [/formula]
Next we move our attention to the waiting process. Given a configuration S i,j , let p i,j be the probability that the next step of the process is active. The number W i,j of waiting steps the process makes when at S i,j before we observe an active steps then satisfies W i,j = 1/p i,j . The next lemma establishes a lower bound on p i,j .
Lemma 11. Consider a selection reactor with n ≤ and any configuration S = S i,j different from S 0,0 and S ,n . Then the probability p i,j that the next step of the process is active satisfies
[formula] p i,j ≥ p in 2r · n · (i(n − j) + ( − i)j) . [/formula]
Proof. We account for the fire-in events only. Let F = + n + (r − 1)(i + j) be the total fitness. Since n ≤ we have F ≤ 2r . The probability p r that a resident fires in the hub and replaces a mutant equals p r = − i F · p in · j n = p in F · n · ( − i)j and the probability p m that a mutant fires in the hub and replaces a resident equals
[formula] p m = ri F · p in · n − j n ≥ p in F · n · i(n − j) [/formula]
Hence p i,j ≥ p r + p m ≥ pin 2r ·n · (i(n − j) + ( − i)j) as desired. Next, for each k ∈ [1, · h 0 + n − 1] let W k = max{W i,j | ψ(S i,j ) = k} be the largest possible number of waiting steps we can expect to encounter, when at a configuration with potential value equal to ψ(S i,j ) = k. The following lemma bounds the sum h0+n−1 k=1 W k from above.
Lemma 12. Suppose n ≤ /4. Then
[formula] h0+n−1 k=1 W k = O N log N · h p in . [/formula]
Proof. Split the configuration space into subsets where each subset X k contains configurations S i,j with k = ψ(S i,j ) = h 0 · i + j. First, within each slice, we identify the configuration where the bound on p i,j from Lemma 11 is weakest. To that end, it suffices to analyze the expression t(i, j) = i(n − j) + ( − i)j for a fixed k = h 0 · i + j. Note that t(i, j) = t( − i, n − j). Plugging in j = k − h 0 · i we get t(i, j) = i(n − (k − h 0 i)) + ( − i)(k − h 0 i) = 2h 0 · i 2 − i · (2k + h 0 − n) + k which is a quadratic function of i with a positive coefficient 2h 0 by the leading term. Its minimum over real numbers is attained for i 0 = (2k + h 0 − n)/4h 0 . Note that when k ≤ 1 2 ( h 0 + n) − n then we have i 0 ≥ k/h 0 which implies that the corresponding j 0 satisfies j 0 ≤ 0, thus the minimum of t(i, j) over real numbers is attained at a point outside of the allowed configuration space. For such k we thus get t(i, j) ≥ t(k/h 0 , 0) = n h 0 · k.
Similarly, when k ≥ 1 2 ( h 0 + n) + n we have i 0 ≤ (k − n)/h 0 and the corresponding j 0 satisfies j 0 ≥ n. For such k we thus get t(i, j) ≥ t((k − n)/h 0 , n) = n h 0 · ( h 0 + n − k).
Finally, for the 2n + 1 intermediate values k ∈ [ 1 2 ( h 0 + n) − n, 1 2 ( h 0 + n) + n], note that any corresponding configuration S i,j satisfies i ∈ [ 1 2 − 3n 2h0 , 1 2 + 3n 2h0 ]. Since t(i, j) is non-decreasing in j when i ≤ 1 2 and non-increasing in j when i ≥ 1 2 , together with the symmetry t(i, j) = t( − i, n − j) this implies that
[formula] t(i, j) ≥ t 1 2 − 3n 2h 0 , 0 ≥ t 1 8 , 0 = 1 8 n, [/formula]
where for convenience in the second inequality we used a bound n ≤ /4 which holds for all large enough . It remains to compute the sum over all k ∈ [1, h 0 + n − 1]. Using Lemma 11, the above analysis and the fact that h 0 = o( ) we can write
[formula] h0+n−1 k=1 W k ≤ 2r n p in · h 0 n · 1 2 ( h0+n)−n k=1 1 k + h 0 n · h0+n−1 k= 1 2 ( h0+n)+n 1 h 0 + n − k + (2n + 1) · 8 n ≤ 2r p in · (h 0 · log( h 0 ) + h 0 · log( h 0 ) + 5) = O p in · h 0 log = O N log N · h p in . [/formula]
We are now ready to prove Item 2 of Theorem B.
Theorem B (Reactors are fast strong amplifiers). For any function α : N → R that is both o(N 1/2 ) and ω(1), and for any fixed r > 1, the selection reactors {SR α } ∞ N =1 satisfy the following, under both uniform and temperature initialization:
1. (strong amplifiers) fp(SR α N , r) → N →∞ 1.
2. (fast) FT(SR α N , r) = O(log N · α 5 (N )).
Proof of Theorem B, Item 2. For each k ∈ [1, h 0 + n − 1], consider the "slice" X k consisting of configurations S i,j with k = ψ(S i,j ) = h 0 · i + j. Let R k be the expected number of times a slice X k is visited (by an active step) and let W k be an upper bound on the expected number of steps before the slice is left, once it is visited. Rescaling from steps to generations we then have
[formula] FT(SR α N , r) ≤ 1 N h0+n−1 k=1 R k · W k . [/formula]
By Lemma 10, for any k we have R k = O(h 2 ). Hence by Lemma 12 we have
[formula] FT(SR α N , r) = O(h 2 ) · 1 N · h0+n−1 k=1 W k = O(log N · h 3 /p in ). [/formula]
Since we have h = Θ(α(N )) and p in = 1/α 2 (N ), this gives the desired bound O(log N · α 5 (N )). |
Endocrinology in the time of COVID-19
We provide guidance on prevention of adrenal crisis during the global COVID-19 crisis, a time with frequently restricted access to the usual level of healthcare. Patients with adrenal insufficiency are at an increased risk of infection, which may be complicated by developing an adrenal crisis; however, there is currently no evidence that adrenal insufficiency patients are more likely to develop a severe course of disease. We highlight the need for education (sick day rules, stringent social distancing rules), equipment (sufficient glucocorticoid supplies, steroid emergency self-injection kit) and empowerment (steroid emergency card, COVID-19 guidelines) to prevent adrenal crises. In patients with adrenal insufficiency developing an acute COVID-19 infection, which frequently presents with continuous high fever, we suggest oral stress dose cover with 20 mg hydrocortisone every 6 h. We also comment on suggested dosing for patients who usually take modified release hydrocortisone or prednisolone. In patients with adrenal insufficiency showing clinical deterioration during an acute COVID-19 infection, we advise immediate (self-)injection of 100 mg hydrocortisone intramuscularly, followed by continuous i.v. infusion of 200 mg hydrocortisone per 24 h, or until this can be established, and administration of 50 mg hydrocortisone every 6 h. We also advise on doses for infants and children.
## Introductory remarks
This guidance has been drawn up to inform clinicians and healthcare staff in their quest to provide guidance on the optimal management of patients with adrenal insufficiency under the circumstances of an acute global healthcare capacity crisis due to COVID-19, the viral illness caused by the novel corona virus SARS-CoV-2.
For the purposes of this guidance, we define primary adrenal insufficiency (PAI) as all patients with loss of function of the adrenal itself, mostly either due to autoimmune adrenalitis, that is, Addison's disease described by the eponymous Thomas Addison, or other causes including congenital adrenal hyperplasia, bilateral adrenalectomy and adrenoleukodystrophy. The overwhelming majority of PAI patients suffer from both glucocorticoid and mineralocorticoid deficiency. Our guidance similarly applies to patients with secondary adrenal insufficiency (SAI) due to hypothalamic or pituitary disease; these patients typically suffer from glucocorticoid deficiency, in the majority in combination with deficiency of other hypothalamic-pituitary axes. Similarly, the same precautionary rules apply to patients with tertiary adrenal insufficiency due to chronic exogenous glucocorticoid therapy for treatment of other conditions. Patients at risk of tertiary adrenal insufficiency are those treated with prednisoloneequivalent doses of greater than 5 mg daily for longer than 4 weeks.
## Are patients with adrenal insufficiency at an increased risk of covid-19?
Yes, patients with adrenal insufficiency are at increased risk of COVID-19; they are at an increased risk of catching this infection and they have a higher risk of complications due to the potential for an adrenal crisis to be triggered by the infection. There is currently no evidence, however, suggestive of a higher likelihood of a severe course of disease in patients with AI falling ill with COVID-19.
- Risk of adrenal crisis during acute illness: Patients with adrenal insufficiency are at risk to develop a potentially life-threatening adrenal crisis if experiencing major stress, such as an acute illness. This requires administration of increased doses of glucocorticoid replacement to prevent and, if already in progress, treat the adrenal crisis [bib_ref] Society for Endocrinology Endocrine Emergency Guidance: emergency management of acute adrenal insufficiency..., Arlt [/bib_ref] [bib_ref] How to avoid precipitating an acute adrenal crisis, Wass [/bib_ref]. Adrenal crises are regularly observed in patients with PAI and SAI [bib_ref] Adrenal crisis in treated Addison's disease: a predictable but under-managed event, White K & Arlt [/bib_ref] [bib_ref] Frequency and causes of adrenal crises over lifetime in patients with 21-hydroxylase..., Reisch [/bib_ref] [bib_ref] High incidence of adrenal crisis in educated patients with chronic adrenal insufficiency:..., Hahner [/bib_ref] and contribute to the observed increased mortality in these patients.
## - increased risk of infections in adrenal insufficiency:
Patients with PAI including Addison's disease and congenital adrenal hyperplasia have been shown to be at an increased risk of infections [bib_ref] Drug prescription patterns in patients with Addison's disease: a Swedish population-based cohort..., Björnsdottir [/bib_ref] [bib_ref] Increased infection risk in Addison's disease and congenital adrenal hyperplasia, Tresoldi [/bib_ref] ; this has also been shown for patients with SAI due to hypothalamic-pituitary disease [bib_ref] Exploring inpatient hospitalizations and morbidity in patients with adrenal insufficiency, Stewart [/bib_ref]. Furthermore, respiratory infections have been shown to contribute to the increased mortality observed in patients with PAI [bib_ref] Premature mortality in patients with Addison's disease: a populationbased study, Bergthorsdottir [/bib_ref] [bib_ref] Mortality in patients with diabetes mellitus and Addison's disease: a nationwide, matched,..., Chantzichristos [/bib_ref]. In addition, patients with PAI have been shown to have significantly decreased natural killer cell cytotoxicity [bib_ref] Primary adrenal insufficiency is associated with impaired natural killer cell function: a..., Bancos [/bib_ref] , an important function of the innate immune system in fighting viral infections. Therefore, patients with PAI can be assumed to be at an increased risk of infection with COVID-19. Patients receiving supraphysiologic, immunosuppressive doses of exogenous glucocorticoids for the treatment of another condition are at even higher risk of infection.
## How should we manage patients with an established diagnosis of adrenal insufficiency?
## A. prevention mode
All patients with established adrenal insufficiency should be provided with adequate self-management support to enable them to manage their conditions adequately and safely. Self-management support can be facilitated and communicated by mailshot, video, text, email phone call or videoconferencing, as appropriate. This should follow the 3E framework for self-management support (Educate, Equip and Empower).
## Educate
- Ensure that all patients (and their families/partners/ carers) are educated in the use of 'the sick day rules', that is, the need to increase their usual glucocorticoid replacement dose during intercurrent illness and the need to self-inject hydrocortisone and call for emergency medical assistance when the oral medication cannot be reliably absorbed due to vomiting or diarrhoea and/or the presence of severe and major illness or trauma. General sick day rules for patients with adrenal insufficiency are described in detail in recently published clinical guidelines [bib_ref] Consensus statement on the diagnosis, treatment and follow-up of patients with primary..., Husebye [/bib_ref] [bib_ref] Diagnosis and treatment of primary adrenal insufficiency: an Endocrine Society Clinical Practice..., Bornstein [/bib_ref] [bib_ref] Guidelines for the management of glucocorticoids during the peri-operative period for patients..., Woodcock [/bib_ref] ; see also https:// endo-ern.eu/wp-content/uploads/2019/03/20190312-Stressinstructie-addisoncrisis-hydrocortison-ENG-Endo-ERN-approved.pdf. However, for the purposes of this guidance, we have revised the generic sick day rules, having in mind patients with an acute COVID-19 infection, which frequently presents with high fever over sustained periods of time (see [fig_ref] Table 1: Suggested management and hydrocortisone stress dose cover in patients with adrenal insufficiency... [/fig_ref] and section B).
- Patients with adrenal insufficiency are at increased risk of COVID-19, albeit not as high as in patients undergoing cancer treatment or taking high doses of potent immunosuppressive drugs. All patients with adrenal insufficiency should 'observe stringent social distancing'. If they are working, they should either work from home or work under conditions that allow very stringent social distancing at all times. This means that adrenal insufficiency patients should not work in situations that do now allow them to keep their safe distance, as is the case, for example, for healthcare workers, carers and supermarket cashier staff. It will be important to provide patients with letters stating this fact to ensure their employers are informed and can adjust working conditions as appropriate.
## Equip
- Ensure that the patient has 'sufficient supplies of oral glucocorticoid preparations' (usually hydrocortisone, but also cortisone acetate, prednisolone or prednisone). Ensure that patients who usually take modified release hydrocortisone preparations have a sufficient supply of immediate release, regular oral hydrocortisone for emergency use, for example, by prescribing an extra 4-week supply of hydrocortisone 10 mg three times daily. In patients with PAI including congenital adrenal hyperplasia also ensure sufficient mineralocorticoid supplies (fludrocortisone). Consider issuing prescriptions of 3-month hydrocortisone supplies every 2 months and arrange for them to be dispensed by mail; this will ensure that the patient has access to sufficient extra glucocorticoid doses in case of intercurrent illness. Patients should understand that they must continue taking their glucocorticoid replacement under all - Ensure that the patient is in possession of an 'up-todate hydrocortisone emergency self-injection kit' and that the patient and a relative/partner/friend is confident in self-administration of the injection. Consider refreshing knowledge by talking through the procedure over the phone and providing links to training videos (https://www.addisonsdisease.org.uk/ the-emergency-injection-for-the-treatment-of-adrenalcrisis and https://www.adrenals.eu/animations/how).
## Empower
- Ensure that all patients are in possession of a 'steroid emergency card' or equivalent written instructions for healthcare staff on how to treat the patient in a major stress situation that prevents self-management.shows the recently issued UK version of the steroid card, developed further from a version originally proposed by a Swedish group (18) They should monitor how much urine they pass; the excretion of only little amounts of dark, concentrated urine indicates insufficient hydration, which should prompt further increased oral fluid intake. Importantly, patients with adrenal insufficiency and an acute suspected or confirmed COVID-19 infection should also 'immediately take a double hydrocortisone morning dose and then increase their hydrocortisone replacement to 20 mg four times daily', that is, 20 mg hydrocortisone every 6 h, for example, at 0600 h, 1200 h, 1800 h and 2400 h [fig_ref] Table 1: Suggested management and hydrocortisone stress dose cover in patients with adrenal insufficiency... [/fig_ref]. In children, their usual daily dose should be trebled and administered orally [bib_ref] Consensus statement on the diagnosis, treatment and follow-up of patients with primary..., Husebye [/bib_ref] [bib_ref] Diagnosis and treatment of primary adrenal insufficiency: an Endocrine Society Clinical Practice..., Bornstein [/bib_ref] [bib_ref] Guidelines for the management of glucocorticoids during the peri-operative period for patients..., Woodcock [/bib_ref] , the personal experience of the authors is that an acute COVID-19 infection is associated with significant and persistent acute inflammation and often continuous high fever, which in our view requires a more evenly spaced glucocorticoid cover throughout day and night. We have based our suggested doses on a three-compartment model of oral hydrocortisone delivery [bib_ref] Modelling oral adrenal support, Smith [/bib_ref] , drawing from experimental data from the Prevention of Adrenal Crisis in Stress (PACS) study [bib_ref] Prevention of adrenal crisis: cortisol responses to major stress compared to stress..., Prete [/bib_ref] and a previous study on oral hydrocortisone pharmacokinetics [bib_ref] Modified-release hydrocortisone to provide circadian cortisol profiles, Debono [/bib_ref]. This modelling indicates that the mere doubling of the regular glucocorticoid dose could leave patients with prolonged periods of glucocorticoid deficiency during an acute and highly inflammatory infection such as COVID-19.
- Under no circumstances should patients hesitate to contact medical emergency services, 'if the clinical signs and symptoms of COVID-19 significantly worsen'. Patients (or their carers) should 'contact medical emergency services without delay and immediately administer their hydrocortisone emergency injection (100 mg i.m.)'. If for any reason they cannot administer the injection, they should immediately take 50-100 mg hydrocortisone orally, if possible, while waiting for medical emergency services to arrive. If need be, patients and their carers should consider making their own way to hospital and continue to take 50 mg hydrocortisone every 6 h. 'Signs and symptoms indicating clinical deterioration in patients affected by COVID-19', which typically occur 7-10 days after onset of the first COVID-associated symptoms, include: ⚬ feeling very dizzy on sitting or standing ⚬ feeling very thirsty despite drinking regularly ⚬ feeling very cold ⚬ shaking uncontrollably ⚬ becoming drowsy, confused or difficult to wake up - Following emergency injection of 100 mg hydrocortisone by self-injection or medical emergency personnel, the patients should be maintained on 'major stress dose hydrocortisone, that is, 200 mg over 24 h, preferably in the hospital setting administered via continuous i.v. infusion', which ensures that intermittent troughs in cortisol levels are avoided- Co-incident type 1 diabetes is found in around 10% of patients with PAI [bib_ref] Addison's disease: a survey on 633 patients in Padova, Betterle [/bib_ref] [bib_ref] Clinical and immunological chracteristics of autoimmune Addison disease: a nationwide Swedish multicenter..., Dalin [/bib_ref] [bib_ref] Management of endocrine disease: disease burden and treatment challenges in patients with..., Chantzichristos [/bib_ref]. The clinical experience is that diabetic patients affected by COVID-19 quickly struggle to maintain glycaemic control, with significantly increased insulin requirements, and are more prone to diabetic ketoacidosis. A recent guidance on managing 'type 1 and type 2 diabetes - Patients on stable replacement usually undergo annual checks of electrolytes and plasma renin to ensure adequacy of mineralocorticoid replacement, but during the COVID-19 crisis blood checks should be reserved for patients with clinical signs of hypotension, such as dizziness when standing up. Blood pressure self-measurement, for example, after sitting for at least 5 min and then again after standing up for a minute, should be encouraged; patients should also be taught how to take their heart rate and be advised which readings should prompt to contact their specialist care team for further advice (such as resting heart rate >100/min and systolic blood pressure <100 mmHg; otherwise healthy and well patients to remeasure after 1 h before contacting medical staff). Many healthcare centres have established bloodletting centres in convenient locations away from hospitals that can be used if a blood test is considered urgent.
- Routine glucocorticoid replacement therapy is monitored based on the patient's clinical performance and ability to cope with daily stress and does not require laboratory evaluation [bib_ref] Diagnosis and management of adrenal insufficiency, Bancos [/bib_ref] , thus history taking and discussions can easily take place via teleconferencing.
## Disclaimer
Due to the emerging nature of the COVID-19 crisis, this document is not based on extensive systematic review or meta-analysis, but on rapid expert consensus. The document should be considered as guidance only; it is not intended to determine an absolute standard of medical care. Healthcare staff need to consider individual circumstances when devising the management plan for a specific patient.
# Declaration of interest
Wiebke Arlt is the Editor-in-Chief of the European Journal of Endocrinology. W A was not involved in the review or editorial process for this paper, on which she is listed as an author. The other authors have nothing disclose.
# Funding
This guidance did not receive any specific grant from any funding agency in the public, commercial or not-for-profit sector.
[table] Table 1: Suggested management and hydrocortisone stress dose cover in patients with adrenal insufficiency and suspected or confirmed COVID-19 infection. Request medical advice on the suspected COVID-19 infection Onset of signs and symptoms of 'clinical deterioration' (dizziness; intense thirst; shaking uncontrollably; drowsiness, confusion, lethargy; vomiting; severe diarrhoea; increasing shortness of breath, respiratory rate >24/min, difficulty speaking • Immediately inject (patient or carer) 100 mg hydrocortisone per i.m. injection in adults and adolescents (25 mg in infants, 50 mg in school children) • Call for emergency medical attention for treatment and transfer to hospital, consider making their own way to hospital • If patients cannot be taken or kept in hospital, then they should take 50 mg hydrocortisone every 6 h orally at home; if possible, they should receive i.v. hydrocortisone and an isotonic saline infusion in the admissions unit At hospital On regular ward or intensive care ward, irrespective of whether breathing unaided or supported by continuous positive airway pressure (CPAP) respiration or mechanically ventilated • Hydrocortisone 100 mg per iv injection in adults and adolescents, followed by continuous iv infusion of 200 mg hydrocortisone/24 h (alternatively 50 mg every 6 h per i.v. or i.m. bolus injection) • Infants and children should receive an initial parenteral injection of 50 mg hydrocortisone/m 2 (usually 25 mg in infants and 50 mg in children) followed by 50 mg/24 h in infants and 100 mg/24 h in children • Pause fludrocortisone in adults • Continuous i.v. fluid resuscitation with isotonic saline; regularly check urea and electrolytes Recovery; improving respiratory function, reducing or normal temperature • Gradual tapering of stress dose hydrocortisone down to double regular replacement dose at time of discharge (endocrinologist to advise) • Re-start usual fludrocortisone dose in adults when total daily hydrocortisone dose <50 mg https://eje.bioscientifica.comcircumstances and that there is no need to increase the dose unless they fall ill. [/table]
|
Levels of selected oxidative stress markers in the vitreous and serum of diabetic retinopathy patients
[bib_ref] Antioxidant therapy for retinal disease, Kiang [/bib_ref] [bib_ref] Role of advanced glycation end products (AGEs) and oxidative stress in diabetic..., Yamagishi [/bib_ref] [bib_ref] Sepici-Dinçel A. Antioxidant enzymes and diabetic retinopathy, Yildirim [/bib_ref] [bib_ref] Antioxidant status lipid peroxidation and nitric oxide end products in patients of..., Bhatia [/bib_ref] [bib_ref] The change of oxidative stress products in diabetes mellitus and diabetic retinopathy, Pan [/bib_ref] [bib_ref] Lipid peroxidation in proliferative vitreoretinopathies, Verdejo [/bib_ref]
# Methods
## Participants in the study:
The study included two groups of patients who underwent vitrectomy during the study period. The study group consisted of patients with type 2 diabetes with proliferative diabetic retinopathy-the PDR group (n=37, average±SD age=68.90±11.65)-and it was compared to the second group, which consisted of patients with nondiabetic eye disorders-the NDED group (n=50, average±SD age=61.24±11.94)-having anatomic vitreoretinal disorders (macular hole, retinal detachment, epiretinal membrane). There were no statistically significant differences between mean demographic parameters, age range, or sex of the two groups. The diagnosis of type 2 diabetes was based on World Health Organization criteria, and proliferative diabetic retinopathy was classified according to the modified Airlie House classification of diabetic retinopathy. All study participants were recruited and examined at the Department of Ophthalmology, Sveti Duh Clinical Hospital in Zagreb. Non-inclusion criteria encompassed subjects previously treated with intravitreal steroids or anti-VEGF therapy, subjects who had previously undergone vitreoretinal surgery, and subjects with other retinal diseases (senile macular degeneration, central retinal artery occlusion and/or veins and branches), subjects on systemic corticosteroid therapy or cytostatics, all subjects with poorly controlled cardiovascular status, pregnant women, and women unable to exclude the possibility of pregnancy with certainty.
The study was approved by the Ethics Committee of the Sveti Duh Clinical Hospital, and was conducted at the Department of Ophthalmology of the hospital from June 2012 to February 2013. All applicable guidelines were followed: Basics of Good Clinical Practice, Helsinki Declaration, Croatian Healthcare Act, and Patient Rights Act. All patients gave their informed consent to participate in the study, after receiving detailed information from their ophthalmologist.
Surgical procedure: Vitreous samples (1.5-2.0 ml) were obtained by the standard vitreoretinal aspiration procedure, pars plana vitrectomy, and all procedures were conducted by the same vitreoretinal surgeon. Vitreal samples were taken immediately after setting the trocars (before any other surgical manipulation) and before turning on the infusion system to avoid dilution (filtered air was insufflated to retain volume). Samples were centrifuged within one hour after surgery for 10 min at 15000 ×g at 4 °C. After centrifugation, the liquid portion (vitreous) was separated and stored at −80 °C until analysis. Blood samples for serum analysis (5 ml) were simultaneously collected from the cubital vein, in vials containing EDTA, and were centrifuged at 1200 ×g at 4 °C. Serum samples were stored at −80 °C. At the time of analysis, all samples were dissolved at room temperature and, if necessary, centrifuged again. All analyses, sample loading, and reagent mixture additions were performed in an array manner using equal sample loading by multichannel pipettes.
Chemicals: Chemicals used in the biochemical analysis were purchased from Sigma Chemical Company. Bovine heart cytochrome C (Type VI) and human blood SOD (Type I, lyophilized powder, 2400 U/mg protein), BSA, xantine, xantine oxidase, 2-thiobarbituric acid, dodecyl sulfate sodium salt, and 1,1,3,3-tetrametoxypropane were purchased from Sigma (St. Louis, MO). All other chemicals were of analytical grade: OxiSelect™ AOPP Assay Kit, Lipid Hydroperoxide Assay kit (Cayman Chemical Company, Ann Arbor, MI), VEGF kit (Enzo Life Sciences, Farmingdale, NY), and glucose assay kit (GOD-PAP method, HUMAN™).
Glucose assay: Glucose levels in serum were measured by a glucose liquicolor (GOD-PAP method) HUMAN™ enzymatic colorimetric test for glucose detection following the manufacturer's instructions. Briefly, 1000 µl of reagent mixture (glucose oxidase, 4-aminoantipyrinephenol, peroxidase, and mutarotase) was added to 10 µl of the sample. Absorbance was measured at 500 nm with a Libro S22 spectrophotometer (Biochrom, UK). Calculation was made by dividing absorbance of the sample by absorbance of a supplied glucose standard solution and multiplying by factors of 100 and 5.55. The glucose concentration was expressed as mmol/l. VEGF assay: Vascular endothelial growth factor (VEGF) was measured using a VEGF ELISA kit (human; Enzo Life Sciences, USA). All kit reagents were prepared according to the manufacturer's instructions. A polyclonal antibody against human VEGF-16VEGF labeled with the enzyme horseradish peroxidase was added to samples. The measured optical density was read at 450 nm and was directly proportional to the concentration of human VEGF in the standards or samples. The concentration of VEGF was calculated from a calibration curve and expressed as pg/l.
Advanced oxidation protein products (AOPP) assay: AOPP was assayed with the OxiSelect™ AOPP Assay Kit according to the manufacturer's instructions. AOPP content was determined by comparing the test sample with the chloramine standard curve. Briefly, 200 μl samples or standards were added to separate wells of the microtiter plate (Biorad, Hercules, CA). A total of 10 μl chloramine reaction initiator was added. The absorbance of each well was recorded immediately on a spectrophotometric plate reader using a wavelength of 340 nm. Results were calculated according to a standard curve and expressed as µM.
Lipid hydroperoxide (LPO) assay: Lipid hydroperoxide was assayed with the Lipid Hydroperoxide Assay kit (Cayman Chemical Company) by direct measurement of redox reaction with iron ions according to the kit manual. Briefly, the solution LPO Assay Extract R was added to test samples. To this mixture, 1 ml cold chloroform solution was added and blended. A total of 450 µl of the chloroform-methanol mixture was added to 500 µl of extract chloroform sample, followed by a 50 µl mixture of chromogen, which turned purple. Absorbance was measured at 500 nm with a Spectro UVD-3500 spectrophotometer (Labomed Inc., Los Angeles, CA). From the calibration curve of LPO, we calculated the concentration of LPO in each sample according to the formula specified by the manufacturer and expressed as µM.
Malondialdehyde (MDA) assay: The presence of lipid peroxidation was determined by measuring the concentration of malondialdehyde (MDA). A total of 200 μl supernatant was mixed with 200 μl 8.1% aqueous sodium dodecyl sulfate, 1.5 ml 20% aqueous acetic acid (pH 3.5), and 1.5 ml 0.81% aqueous thiobarbituric acid and heated for 60 min at 95 °C. After cooling samples on ice, absorbance was measured at 532 nm and 600 nm with a Libro S22 spectrophotometer (Biochrom, Cambridge, UK). The total absorbance was determined using the formula A total =A 532 -A 600 . An array of known concentrations of tetramethoxypropane was used for creating the calibration curve using the same protocol as for the homogenized samples. MDA values are presented as nmol/mL.
Glutathione assay (GSH) assay: The glutathione assay is a modification of the method first described by Tietze [bib_ref] Enzymic method for quantitative determination of nanogram amounts of total and oxidizedglutathione..., Tietze [/bib_ref]. Briefly, in a 96-well plate, 40 μl 10 mM 5-5′-dithiobis [2-nitrobenzoic acid] (DTNB, Ellman's Reagent) was added to 20 μl sample supernatant pre-treated with 40 µl 0.035M HCL. This mixture was incubated for 10 min. DTNB reacts with GSH to form chromospheres. The absorbance of these chromogens was measured at 412 nm in an ELISA plate reader (BIORAD). Then, 100 µl reaction mixture (9980 µl 0.8 mM NADPH and 20 µl gluthatione reductase, 0.2 U/ ml) was added and the absorbance was read at 412 nm every minute for 5 min. The results were calculated from the standard curve of array of dilutions of glutathione (GSH). Concentrations are presented as µmol/ml.
Total superoxide dismutase (SOD) assay: The measure of SOD activity is calculated from the percentage of inhibition of the reaction of xantine oxidation by xantine oxidase (optimized reaction ratio ΔA/ min≈0.025), which creates a superoxide anion as a substrate for SOD. The superoxide anion not used by the enzyme SOD oxidizes the cytochrome. For determination of SOD activity, 25 µl of undiluted sample were mixed with 1.45 ml of the reaction mix (cytochrome C, 0.05 mM; xantine, 1 mM mixed to a 10:1 ratio with addition of DTNB). To this mixture, 20 µl xantine oxidase 0.4 Uml −1 was added to start a reaction. The reaction was measured over 3 min at 550 nm. The absorbance and percentage of inhibition were compared to the calibration curve created with different dilutions of SOD. Enzyme values are presented as U/ml.
# Statistical analysis:
Statistics were based on non-parametric methods, due to the small sample size and non-normal data distribution (verified by the Kolmogorov-Smirnov test). The Mann-Whitney U test was used to analyze numerical variables between groups. Spearman's rank correlation was used for correlation analysis within and between measured parameters in the vitreous and serum of both groups of patients (NDED and PDR). Analyses were performed using SPSS version 17 (SPSS Inc., Chicago, IL). The level of statistical significance was set at p≤0.05.
# Results
A comparison of gender-dependent differences [fig_ref] Table 1: Measured paraMeTers and sTaTisTical differences wiThin and beTween genders in non-diabeTic paTienTs... [/fig_ref] revealed that there were no statistically significant gender differences in the measured parameters in the vitreous or serum, except a single detected difference in the NDED group, where serum MDA was significantly lower (p≤0.05) in male than in female patients. However, within the same gender groups, there were significant differences (p≤0.05) between the NDED and PDR groups in the majority of measured parameters [fig_ref] Table 1: Measured paraMeTers and sTaTisTical differences wiThin and beTween genders in non-diabeTic paTienTs... [/fig_ref].
Trends of statistical differences in the measured parameters between the NDED and PDR groups by age distribution [fig_ref] Table 2: Mean values of Measured paraMeTers by age groups and The correlaTion analysis... [/fig_ref] were similar to trends within the sexes [fig_ref] Table 1: Measured paraMeTers and sTaTisTical differences wiThin and beTween genders in non-diabeTic paTienTs... [/fig_ref]. The correlation of the measured parameters and age of patients included in the study [fig_ref] Table 2: Mean values of Measured paraMeTers by age groups and The correlaTion analysis... [/fig_ref] revealed a significant negative correlation (p≤0.05) of serum MDA and SOD activity with increased age in PDR patients, and serum VEGF was positively correlated with increasing age.
The analysis of serum glucose of NDED patients and PDR patients at the time of the procedure [fig_ref] Figure 1: Average serum glucose concentrations in non-diabetic patients with eye disorders [/fig_ref] revealed that glucose levels were as expected, i.e., were significantly higher (p≤0.05) in PDR patients than in NDED patients. However, serum glucose values did not correlate significantly with changes of the measured parameters in either group [fig_ref] Table 3: correlaTion of Measured paraMeTers wiTh seruM glucose in non-diabeTic paTienTs wiTh eye... [/fig_ref] , with the exception of a slight significant positive correlation with MDA in the NDED group.
Vitreous VEGF levels were significantly increased (p≤0.05), almost 10-fold, in the PDR group compared to NDED patients [fig_ref] Figure 2: Relationships of vitreous and serum VEGF values in nondiabetic patients with eye... [/fig_ref]. Serum VEGF of the PDR group was also increased significantly (p≤0.05; [fig_ref] Figure 2: Relationships of vitreous and serum VEGF values in nondiabetic patients with eye... [/fig_ref]. However, since the scale on [fig_ref] Figure 2: Relationships of vitreous and serum VEGF values in nondiabetic patients with eye... [/fig_ref] is set for vitreal VEGF, the serum VEGF range is provided here (serum VEGF: NDED range 2.0-104 pg/ml; PDR range 5.0-760 pg/ml).
The analysis of advanced oxidized protein product (AOPP) levels [fig_ref] Figure 3: Relationships of vitreous and serum AOPP values in nondiabetic patients with eye... [/fig_ref] revealed no statistical differences between NDED and PDR patients. Serum AOPP levels in NDED patients were only slightly significantly (p≤0.05) higher than in the vitreous. However, serum AOPP levels were significantly higher (p≤0.05) in PDR than in NDED patients.
Lipid peroxidation (LPO; [fig_ref] Figure 4: Relationships of vitreous and serum LPO values in nondiabetic patients with eye... [/fig_ref] was almost equal in the vitreous and serum in the NDED group. In the PDR group, LPO values were slightly but significantly higher (p≤0.05) in serum than in the vitreous. In a comparison between groups, LPO values were significantly (p≤0.05) higher in the PDR group, with an almost five-fold increase of the peroxydized lipid concentration. The LPO analysis marked the most prominent change of all the analyzed oxidative stress markers.
Similarly, MDA [fig_ref] Figure 5: Relationships of vitreous and serum MDA values in nondiabetic patients with eye... [/fig_ref] is also a lipid peroxidation marker. It was established that serum levels were higher (p≤0.05) than vitreous values in the PDR group. In the NDED group, serum and vitreous MDA values showed no significant difference. However, vitreous and serum MDA values were significantly higher (p≤0.05) in the PDR than in the NDED group.
The activity of total superoxide dismutase (SOD; [fig_ref] Figure 6: Relationships of vitreous and serum SOD activity levels in non-diabetic patients with... [/fig_ref] was slightly but significantly (p≤0.05) lower in serum than in the vitreous of the NDED subjects. In PDR patients, the vitreal SOD activity was slightly but significantly (p≤0.05) lower than in NDED patients, while serum SOD activity was significantly higher (p≤0.05) compared to NDED patients.
Glutathione levels (GSH) were significantly (p≤0.05) higher in the serum than in the vitreous of both NDED and PDR groups; however, no significant differences were detected in the serum or vitreous GSH levels between the groups [fig_ref] Figure 7: Relationships of vitreous and serum GSH values in nondiabetic patients with eye... [/fig_ref]. The Spearman correlation coefficients of the measured parameters in the NDED subjects are presented in and for PDR patients in [fig_ref] Table 5: correlaTion analysis of Measured paraMeTers in viTreous and seruM of proliferaTive diabeTic... [/fig_ref]. In the NDED group, the only significant correlation between the measured parameters was the negative correlation between GSH and SOD in serum , ρ=-0.627; p<0.001). In PDR, however, several correlations occurred as a consequence of the pathological condition and new allostasis [fig_ref] Table 5: correlaTion analysis of Measured paraMeTers in viTreous and seruM of proliferaTive diabeTic... [/fig_ref]. The significantly (p≤0.05) strongest ones (ρ ≤ ±0.600) were between vitreous LPO and serum SOD (Table 5, [fig_ref] Figure 8: The strongest correlations between the measured markers in non-diabetic patients with eye... [/fig_ref] , serum SOD and LPO (Table 5, [fig_ref] Figure 8: The strongest correlations between the measured markers in non-diabetic patients with eye... [/fig_ref] , vitreous LPO and serum LPO (Table 5, [fig_ref] Figure 8: The strongest correlations between the measured markers in non-diabetic patients with eye... [/fig_ref] , and vitreous VEGF and serum SOD (Table 5, [fig_ref] Figure 8: The strongest correlations between the measured markers in non-diabetic patients with eye... [/fig_ref].
# Discussion
Investigating the dynamics and correlative ties of oxidative stress markers in homeostatically linked physiologic fluids, i.e., the vitreous and serum of patients having proliferative diabetic retinopathy (PDR), might reveal possible markers and their ratios for disease control and monitoring progression.
This study of the links and correlations between the measured parameters indicated that all parameters assayed in the serum or eye of NDED subjects had ρ (Spearman correlation coefficient) values near zero. Conclusively, markers in the vitreous or serum were physiologically independent in the NDED group . Conversely, in PDR patients, diabetic retinopathy and increased oxidative stress resulted in several statistically significant correlations between the measured parameters in the vitreous and serum [fig_ref] Table 5: correlaTion analysis of Measured paraMeTers in viTreous and seruM of proliferaTive diabeTic... [/fig_ref]. This is likely due to the fact that the blood-retinal barrier becomes more permeable in diabetes, as has been shown experimentally [bib_ref] Diabetes-induced superoxide anion and breakdown of the blood-retinal barrier: role of the..., El-Remessy [/bib_ref].
Age [fig_ref] Table 2: Mean values of Measured paraMeTers by age groups and The correlaTion analysis... [/fig_ref] was found to be a partially contributing factor, i.e., with increasing age of diabetic patients, there was a significant increase of serum VEGF and a reduction of serum SOD and MDA, but age did not influence the vitreal changes. One must remember that the overall serum SOD activity and MDA levels were 3-4 times higher in PDR than in NDED patients, and the correlation with age occurs in such a higher range of pathophysiological allostasis [fig_ref] Figure 5: Relationships of vitreous and serum MDA values in nondiabetic patients with eye... [/fig_ref] and [fig_ref] Figure 6: Relationships of vitreous and serum SOD activity levels in non-diabetic patients with... [/fig_ref].
The analysis of glucose levels [fig_ref] Figure 1: Average serum glucose concentrations in non-diabetic patients with eye disorders [/fig_ref] shows that, at the time of analysis in diabetic patients (fasting patients on the day of operation), it may be statistically but is not prominently higher. Parameters were not significantly correlated to glucose levels at the time of measurement [fig_ref] Table 3: correlaTion of Measured paraMeTers wiTh seruM glucose in non-diabeTic paTienTs wiTh eye... [/fig_ref]. Rather, it is the long period of fluctuations of all biochemical parameters over time that results in the imbalance of the antioxidant defense system.
VEGF was found to be increased in the vitreous of PDR patients [fig_ref] Figure 2: Relationships of vitreous and serum VEGF values in nondiabetic patients with eye... [/fig_ref] , similar to [bib_ref] Vitreous and serum levels of vascular endothelial growth factor and platelet-derived growth..., Praidou [/bib_ref] [bib_ref] Increased levels of vascular endothelial growth factor and advanced glycation end products..., Endo [/bib_ref] [bib_ref] Proliferative diabetic retinopathy and relations among antioxidant activity oxidative stress and VEGF..., Izuta [/bib_ref]. However, in this study, a moderate yet significant positive correlation (ρ=0.357 and p=0.001; [fig_ref] Table 5: correlaTion analysis of Measured paraMeTers in viTreous and seruM of proliferaTive diabeTic... [/fig_ref] was found between the vitreous and serum VEGF in the PDR group. We note that, in diabetes, microvascularization in other tissues also contribute to an increase in serum VEGF, though this significant correlation between serum and vitreous levels may suggest an important physiologic link between these humoral compartments [bib_ref] Diabetes-induced superoxide anion and breakdown of the blood-retinal barrier: role of the..., El-Remessy [/bib_ref]. As such, serum levels could perhaps reflect intraocular changes.
AOPPs are structurally similar to advanced glycation end-product (AGE) and exert similar biologic activity. Serum AOPP levels are elevated in patients with renal complications, atherosclerosis, and diabetes [bib_ref] The role of advanced oxidation protein products and total thiols in diabetic..., Baskol [/bib_ref]. The present study revealed no statistically significant differences in vitreous AOPP between the NDED and PDR patients [fig_ref] Figure 1: Average serum glucose concentrations in non-diabetic patients with eye disorders [/fig_ref]. The established difference between higher serum than vitreal values of AOPP can be explained by the higher protein content in serum than in the vitreous, and higher values of serum oxidized proteins in PDR than in NDED patients. Data on AOPP levels in the vitreous of diabetic patients with PDR disease is scarce, though serum AOPP increases in diabetes have been reported, together with an increase in oxidized albumin and protein carbonyls [bib_ref] The role of advanced oxidation protein products and total thiols in diabetic..., Baskol [/bib_ref] [bib_ref] The redox state of human serum albumin in eye diseases with and..., Oettl [/bib_ref] [bib_ref] Elevated protein carbonyl and HIF-1 alpha levels in eyes with proliferative diabetic..., Loukovaara [/bib_ref] [bib_ref] Role of hyperglycemia-mediated erythrocyte redox state alteration in the development of diabetic..., Choudhuri [/bib_ref]. The correlation analysis showed that AOPP increased independently from other parameters, since it was not significantly correlated to changes in other parameters in either the vitreous or serum. Accordingly, the AOPP method used here does not reflect vitreous oxidative stress and is more specific for plasmatic oxidative stress products. Glutathione (GSH), a small peptide, showed no prominent changes. There are reports of unchanged thiol levels in PDR patients [bib_ref] The role of advanced oxidation protein products and total thiols in diabetic..., Baskol [/bib_ref]. Others, however, have reported total thiol levels and GSH depletion in serum and vitreous in human subjects with PDR [bib_ref] Interleukin-8, nitric oxide and glutathione status in proliferative vitreoretinopathy and proliferative diabetic..., Cicik [/bib_ref] , though most of the reports are on animal models [bib_ref] Influence of glutathione on the electroretinogram in diabetic and non-diabetic rats, Wright [/bib_ref] [bib_ref] Telmisartan ameliorates neurotrophic support and oxidative stress in the retina of streptozotocin-induced..., Ola [/bib_ref] [bib_ref] Effect of lipid-hydroperoxide-induced oxidative stress on vitamin E, ascorbate and glutathione in..., Kamegawa [/bib_ref].
The results and comparison of lipid peroxidation markers using direct measures of lipid peroxidation (LPO) and indirect measures of the lipid peroxidation byproduct malondialdehide (MDA) indicated that the LPO measurement method [fig_ref] Figure 4: Relationships of vitreous and serum LPO values in nondiabetic patients with eye... [/fig_ref] showed a more prominent difference (5-7 times higher increase in serum and vitreous) than the MDA method (2-3 times increase in serum and vitreous) in the PDR group [fig_ref] Figure 5: Relationships of vitreous and serum MDA values in nondiabetic patients with eye... [/fig_ref]. There are no literature reports assessing lipid peroxidation (LPO) using the method described here, though earlier studies measuring lipid peroxidation using other markers yielded significantly increased results [bib_ref] Effect of lipid-hydroperoxide-induced oxidative stress on vitamin E, ascorbate and glutathione in..., Kamegawa [/bib_ref]. The present results on MDA changes in the vitreous of diabetic patients with PDR are in accordance with the studies by Mandal et al. [bib_ref] Oxidative stress-associated neuroretinal dysfunction and nitrosative stress in diabetic retinopathy, Mandal [/bib_ref] and Manciano et al. [bib_ref] Lipid peroxidation and total antioxidant capacity in vitreous, aqueous humor, and blood..., Mancino [/bib_ref] , which showed an increase in vitreal MDA in PDR patients. They also indicated that lipid peroxidation in diabetic patients with PDR is a highly pronounced process in the humoral parts of the organism, and lipid peroxidation appears to be highly responsible for induced oxidative stress in diabetic patients [bib_ref] Lipid peroxidation and antioxidant status in patients with diabetic retinopathy, Gupta [/bib_ref] [bib_ref] Elevated lipid peroxides induced angiogenesis in proliferative diabetic retinopathy, Saxena [/bib_ref]. The LPO method is more sensitive compared to methods that measure lipid peroxidation byproducts, such as MDA. LPO offers a better picture of the extent of lipid peroxidation than MDA. This was confirmed by the significant correlations between LPO in the vitreous and serum. MDA was the only serum variable in the study group to remain independent from the other parameters. Its ρ value remained similar as in the NDED group. This indicates that the MDA change in serum does not reflect changes in the vitreous but rather depends on systemic sources from other tissues. In addition, the determined ρ value for LPO showed a significant positive association with increased LPO concentrations in the eye and in serum, but also with vitreous MDA and SOD and serum VEGF, AOPP, SOD, and especially with serum LPO. These results indicate that the increase in overall serum LPO levels, compared to normal serum levels, is approximately twice the increase in lipid peroxidation in the eye. Given that previous studies [bib_ref] Effect of lipid-hydroperoxide-induced oxidative stress on vitamin E, ascorbate and glutathione in..., Kamegawa [/bib_ref] [bib_ref] Elevated lipid peroxides induced angiogenesis in proliferative diabetic retinopathy, Saxena [/bib_ref] have shown a significant correlation between the increase of LPO in the vitreous and an increase in the expression of vascular endothelial growth factors (VEGF), it appears that LPO determination in serum might be a good predictor for the onset of oxidative stress within the vitreous. Furthermore, [fig_ref] Table 5: correlaTion analysis of Measured paraMeTers in viTreous and seruM of proliferaTive diabeTic... [/fig_ref] and [fig_ref] Figure 8: The strongest correlations between the measured markers in non-diabetic patients with eye... [/fig_ref] show that a pronounced correlation change was found for serum LPO and serum SOD activity. A relatively strong and physiologically significant correlation was recorded between serum SOD activity and vitreous LPO levels. Unlike the high serum SOD activity, the ocular activity of SOD shown in [fig_ref] Figure 6: Relationships of vitreous and serum SOD activity levels in non-diabetic patients with... [/fig_ref] is likely to decrease slightly compared to NDED patients. Yildirim et al. [bib_ref] Sepici-Dinçel A. Antioxidant enzymes and diabetic retinopathy, Yildirim [/bib_ref] reported no changes in serum SOD activity in PDR patients. This can be explained by the saturation of the ocular enzyme by its substrate (superoxide radical), which is a free radical whose concentration is elevated during oxidative stress. On the other hand, SOD activity was positively correlated with vitreous VEGF [bib_ref] Extracellular SOD and VEGF are increased in vitreous bodies from proliferative diabetic..., Izuta [/bib_ref].
SOD activity could be important in diabetics with PDR, as indicated elsewhere [bib_ref] Extracellular SOD and VEGF are increased in vitreous bodies from proliferative diabetic..., Izuta [/bib_ref] [bib_ref] Catalase/superoxide dismutase (SOD) and catalase/paraoxonase (PON) ratios may implicate poor glycemic control, Sözmen [/bib_ref] [bib_ref] Vascular protection: superoxide dismutase isoforms in the vessel wall, Faraci [/bib_ref] [bib_ref] Diabetic retinopathy superoxide damage and antioxidants, Santos [/bib_ref] [bib_ref] Hyperglycemia increases mitochondrial superoxide in retina and retinal cells, Du [/bib_ref]. This study shows that it can be monitored with LPO, as they concomitantly change in both the eye and in serum. For the exact dynamics of oxidative markers and serum-vitreous relations, a multi-year study would be required on the same diabetic patients from early diabetes onset to late retinopathy changes. Nevertheless, studies of the final proliferative state offer reliable initial findings of candidate markers [bib_ref] Proliferative diabetic retinopathy and relations among antioxidant activity oxidative stress and VEGF..., Izuta [/bib_ref] [bib_ref] Extracellular SOD and VEGF are increased in vitreous bodies from proliferative diabetic..., Izuta [/bib_ref] [bib_ref] Sivashanmugham M. Homocysteinethiolactone and paraoxonase: novel markers of diabetic retinopathy, Barathi [/bib_ref]. Based on the results of this study, it can be concluded that further research on concomitant changes and ratios of vitreal and serum LPO and SOD activity could be promising as possible indicators of oxidative change in the eye, if performed alongside other linked ratios [bib_ref] Catalase/superoxide dismutase (SOD) and catalase/paraoxonase (PON) ratios may implicate poor glycemic control, Sözmen [/bib_ref] [bib_ref] Sivashanmugham M. Homocysteinethiolactone and paraoxonase: novel markers of diabetic retinopathy, Barathi [/bib_ref]. These conclusions are supported by the significant correlations reported in this study. The correlations suggest that monitoring their mutual alterations might be informative during PDR development and should be taken into consideration in further research, including animal studies and human studies of diabetic retinopathy development in asymptomatic patients with no clinical signs.
# Acknowledgments
This work was conducted without any commercial conflicts of interest and was supported by project no. 119-0000000-1255 of the Ministry of Science and Education of the Republic of Croatia and the Department of Animal Physiology, Faculty of Science, University of Zagreb and the Sveti Duh Clinical Hospital, Zagreb, Department of Ophthalmology and Medical Laboratory of Diagnostics. We are grateful to Pero Hrabac, PhD for assistance with statistics. All of the authors declare they have no conflict of interest. Authors contribution: VB Šarić participated in developing the study design, collecting and analyzing data, writing the manuscript. I. Landeka participated in biochemical analysis and method development. B Šarić participated in collecting data and writing the manuscript. M. Barberić participated in data measurement. L. Andrijašević participated in writing and editing the manuscript, research result evaluation. N Oršolić is the project leader and contributed in finalizing the manuscript. D. Đikić participated in developing the study design, biochemical analysis and analyzing data, critical evaluation of results, and writing the manuscript and, together with B. Cerovski, is supervisor of the PhD thesis of the first author VB Šarić.
[fig] Figure 1: Average serum glucose concentrations in non-diabetic patients with eye disorders (NDED) and proliferative diabetic retinopathy (PDR) patients. NDED group (n=50); PDR group (n=37). a Columns are significantly different (p≤0.05). [/fig]
[fig] Figure 2: Relationships of vitreous and serum VEGF values in nondiabetic patients with eye disorders (NDED) and proliferative diabetic retinopathy (PDR) patients. NDED group (n=50); PDR group (n=37). a Boxes are significantly different (p≤0.05). [/fig]
[fig] Figure 3: Relationships of vitreous and serum AOPP values in nondiabetic patients with eye disorders (NDED) and proliferative diabetic retinopathy (PDR) patients. NDED group (n=50); PDR group (n=37). a,b,c Boxes bearing the same superscript letter are significantly different (p≤0.05). [/fig]
[fig] Figure 4: Relationships of vitreous and serum LPO values in nondiabetic patients with eye disorders (NDED) and proliferative diabetic retinopathy (PDR) patients. NDED group (n=50); PDR group (n=37). a,b,c Boxes bearing the same superscript letter are significantly different (p≤0.05). [/fig]
[fig] Figure 5: Relationships of vitreous and serum MDA values in nondiabetic patients with eye disorders (NDED) and proliferative diabetic retinopathy (PDR) patients. NDED group (n=50); PDR group (n=37). a,b,c Boxes bearing the same superscript letter are significantly different (p≤0.05). [/fig]
[fig] Figure 6: Relationships of vitreous and serum SOD activity levels in non-diabetic patients with eye disorders (NDED) and proliferative diabetic retinopathy (PDR) patients. NDED group (n=50); PDR group (n=37). a,b,c,d Boxes bearing the same superscript letter are significantly different (p≤0.05). [/fig]
[fig] Figure 7: Relationships of vitreous and serum GSH values in nondiabetic patients with eye disorders (NDED) and proliferative diabetic retinopathy (PDR) patients. NDED group (n=50); PDR group (n=37). a,b,c,d Boxes bearing the same superscript letter are significantly different (p≤0.05). [/fig]
[fig] Table 4: correlaTion analysis of Measured paraMeTers in viTreous and seruM of non-diabeTic paTienTs wiTh eye disorders (nded) paTienTs. are significantly correlated (p≤0.05) Values above the diagonal empty cells (/) represent the Spearman correlation coefficient (ρ); values below the diagonal empty cells represent the level of statistical significance (p≤0.05) of the correlation. [/fig]
[fig] Figure 8: The strongest correlations between the measured markers in non-diabetic patients with eye disorders (NDED) and proliferative diabetic retinopathy (PDR) patients; A) LPO vitreous and SOD serum, B) LPO serum and SOD serum, C) LPO vitreous and LPO serum, D) VEGF vitreous and SOD serum. [/fig]
[table] Table 1: Measured paraMeTers and sTaTisTical differences wiThin and beTween genders in non-diabeTic paTienTs wiTh eye disorders (nded) and proliferaTive diabeTic reTinopaThy (pdr) paTienTs.* values are statistically different (p≤0.05) from the same parameter of non-diabetic patients with eye disorders (NDED) within the same gender group (column), i.e., males in NDED versus PDR; females in NDED versus PDR. # values are statistically different between genders within non-diabetic patients with eye disorders (NDED) and with proliferative diabetic retinopathy (PDR) patients (p≤0.05). [/table]
[table] Table 2: Mean values of Measured paraMeTers by age groups and The correlaTion analysis of Measured paraMeTers wiTh [/table]
[table] Table 3: correlaTion of Measured paraMeTers wiTh seruM glucose in non-diabeTic paTienTs wiTh eye disorders (nded) and proliferaTive diabeTic reTinopaThy (pdr) paTienTs. [/table]
[table] Table 5: correlaTion analysis of Measured paraMeTers in viTreous and seruM of proliferaTive diabeTic reTinopaThy # values are significantly correlated (p≤0.05) Values above the diagonal empty cells (/) represent the Spearman correlation coefficient (ρ); values below the diagonal empty cells represent the level of statistical significance (p≤0.05) of the correlation. [/table]
|
‘Practical’ resources to support patient and family engagement in healthcare decisions: a scoping review
Background: Extensive literature exists on public involvement or engagement, but what actual tools or guides exist that are practical, tested and easy to use specifically for initiating and implementing patient and family engagement, is uncertain. No comprehensive review and synthesis of general international published or grey literature on this specific topic was found. A systematic scoping review of published and grey literature is, therefore, appropriate for searching through the vast general engagement literature to identify 'patient/family engagement' tools and guides applicable in health organization decision-making, such as within Alberta Health Services in Alberta, Canada. This latter organization requested this search and review to inform the contents of a patient engagement resource kit for patients, providers and leaders. Methods: Search terms related to 'patient engagement', tools, guides, education and infrastructure or resources, were applied to published literature databases and grey literature search engines. Grey literature also included United States, Australia and Europe where most known public engagement practices exist, and Canada as the location for this study. Inclusion and exclusion criteria were set, and include: English documents referencing 'patient engagement' with specific criteria, and published between 1995 and 2011. For document analysis and synthesis, document analysis worksheets were used by three reviewers for the selected 224 published and 193 grey literature documents. Inter-rater reliability was ensured for the final reviews and syntheses of 76 published and 193 grey documents. Results: Seven key themes emerged from the literature synthesis analysis, and were identified for patient, provider and/or leader groups. Articles/items within each theme were clustered under main topic areas of 'tools', 'education' and 'infrastructure'. The synthesis and findings in the literature include 15 different terms and definitions for 'patient engagement', 17 different engagement models, numerous barriers and benefits, and 34 toolkits for various patient engagement and evaluation initiatives. Conclusions: Patient engagement is very complex. This scoping review for patient/family engagement tools and guides is a good start for a resource inventory and can guide the content development of a patient engagement resource kit to be used by patients/families, healthcare providers and administrators.
# Background
Patient-centred care implies that patients and their families are the focus of the health care system as recipients of its services, programs and delivery approaches [bib_ref] |Promoting patient-centered care: a qualitative study of facilitators and barriers in healthcare..., Luxford [/bib_ref]. The aim is to ensure that service delivery and decisions are made around the principles of patient and familycentred healthcare, and focus on safety compliance, 'best practice' or evidence-based interventions, policies and positive health outcomes [bib_ref] Valuing people as individuals: development of an instrument through a survey of..., Coyle [/bib_ref]. The outcome of this model is that patients and their families are actively and meaningfully engaged in discussions and decisions concerning policies, programs, service delivery and implications of the care provided. The challenge for healthcare institutions is to locate or develop, and implement the mechanisms, tools and resources essential for preparing and supporting patients, their families, healthcare providers and healthcare administrators to effectively and successfully practice patient engagement. This challenge became the 'Call to Action' by the Canadian Health Services Research Foundation to health authorities and institutions between 2009 and 2013.
One institution which took up this challenge was Alberta Health Services (AHS), the provincial health organization in Alberta, Canada. AHS identified patient engagement as one of its core principles aligned with quality and safety of health service delivery, and created a Patient Engagement Framework for the organization. In this Framework, patient engagement was defined as "a broad two way practice guided by a set of principles, processes and activities that provide an opportunity for stakeholders to be involved in meaningful interactions. Engagement considers and incorporates the values and needs of patients, clinicians, and communities into health services decision making to enhance transparency and accountability". In line with this definition, AHS defined 'patient' in the broadest sense, as "all individuals including clients, residents and members of the public who receive or have requested healthcare or services from AHS and its health care providers". In 2009 AHS established the Patient Engagement Department to help advance patient engagement throughout the organization. As part of this work, Patient Engagement staff identified the need for a Resource Kit for patients, providers (e.g. clinicians), staff and leaders (e.g. administrators) to guide them in how to initiate and implement patient engagement within AHS.
Questions were posed regarding the development of a resource kit that would provide the basics as well as some advanced knowledge, skills, tools and resources needed for engaging patients/families. If health care organizations, and more specifically health care providers and administrators, want to involve patients and families in providing advice or input on improving care delivery or access, what are some effective approaches to do this? What preparation is needed to effectively engage patients and families in AHS decisions? Are there existing models, tools and guides which can be used or adapted?
These questions were incentives for conducting a systematic scoping literature review, of both published and grey literature, as described by Arksey and O'Malley [bib_ref] Scoping studies towards a methodological framework, Arksey [/bib_ref]. The intent of this scoping review was to identify what tools and guides, education and infrastructure resources existed to engage patients and families in healthcare delivery and other decision-making processes. The ultimate goal was to have an inventory of existing materials which would inform the contents for a patient engagement resource kit to be used by patients, providers and leaders in their efforts to successfully implement and evaluate patient engagement initiatives across AHS. This paper describes the systematic scoping literature review and synthesis of the information into relevant thematic clusters which can be considered for incorporation into a patient engagement resource kit. The actual resource kit development and its pilot or evaluation to determine the quality of the tools and guides for patient engagement is an ongoing process within AHS, and is, therefore, not part of this review or paper.
# Methods
The scoping review was selected as the most suitable method for this study, because by definition it is used "to map rapidly the key concepts underpinning a research area and the main sources and types of evidence available, and can be undertaken as stand-alone projects in their own right, especially where an area is complex or has not been reviewed comprehensively before". The refined approach of Arksey and O'Malley [bib_ref] Scoping studies towards a methodological framework, Arksey [/bib_ref] was applied, which includes: (1) identifying the research question/s, which is/are generally broad in nature; [bib_ref] Valuing people as individuals: development of an instrument through a survey of..., Coyle [/bib_ref] identifying topic-relevant studies through a search which is as comprehensive as possible; (3) selecting and rejecting studies using a set of inclusion/exclusion criteria, based on familiarity with the literature; (4) reviewing the sorted and sifted data through charting based on key issues and themes; and (5) analyzing the results through thematic analysis, and reporting findings descriptively and numerically. As a final step, a consultation exercise is recommended involving key stakeholders to inform and validate study findings. Based on the study research questions posed, which was the first step in the scoping review, the overall intent of this study was to provide a comprehensive exploration of the diverse published and grey literature to select the patient engagement tools, education and infrastructure resources that would become the content of a resource kit that would be useful to patients/families, providers and leaders within AHS in their efforts to implement and evaluate patient engagement. The actual resource kit development and its pilot or evaluation to determine the quality of the tools and guides for patient engagement is not part of this review or paper, and therefore, the final step of consultation in the scoping review process is part of the ongoing process within AHS, and reported separately. The following outlines the process taken with the various steps for a scoping review:
Searching and selecting the published and unpublished (Grey) literature
Searching the published and grey literature required the expertise of a health research librarian. Before the application of search terms, parameters for the published and unpublished grey literature databases or search engines were established through the specific inclusion criteria which limited the literature search to the dates selected and English language preference.
The published literature was then searched through Embase, Medline, ERIC, Web of Science and ProQuest using the terms 'patient engagement', 'patient involvement', and 'patient participation'. This search was refined through the application of specific search terms such as 'healthcare', 'decision making', 'advisory committees', and other terms pertaining to more specific organizational engagement approaches, education/training, tools, guides and infrastructure supports and evaluation considerations. Some journals known to contain patient engagement studies were hand searched (e.g. Journal of Participatory Medicine and International Journal for Quality in Health Care). [fig_ref] Table 1: Search terms and general inclusion/exclusion criteria for published and unpublished [/fig_ref] contains the search terms applied to the databases. [fig_ref] Figure 1: See legend on next page [/fig_ref] depicts the flowchart for the screening process applied to the published literature.
The grey literature was searched using the same and additional search terms (toolkits, resources, education, and supports) applied to Google and MSN search engines. Specific countries were also added to the list of search terms -United States (U.S.), Australia and the United Kingdom (U.K.) and Europe were selected because these countries are well-known in the literature and at conferences for their more advanced patient engagement practices and resources. Canada was selected because it was the country where this study took place, and also because patient engagement was a relatively new approach in Canadian healthcare systems at the time, but had some examples worth noting. Websites were manually searched in addition to the Google and MSN searches. [fig_ref] Table 1: Search terms and general inclusion/exclusion criteria for published and unpublished [/fig_ref] also contains the search summary for the grey literature and [fig_ref] Figure 1: See legend on next page [/fig_ref] depicts the flowchart for the screening process involved with this type of literature.
## Article review, selection and analysis
All selected published abstracts were read and rated independently by three of the five members of the project literature review working group, thereby ensuring inter-rater reliability through consensus and minimization of bias. Each reviewer rated the abstracts independently using the inclusion/exclusion criteria identified in [fig_ref] Table 1: Search terms and general inclusion/exclusion criteria for published and unpublished [/fig_ref] , by choosing 'yes' include the source, or 'no' do not include it, or 'maybe' include it. Sources that did not have an abstract were automatically included for review. The 'majority rule' was used to decide whether to include or exclude a particular article abstract. All abstracts were discussed by the three reviewers. If there were discrepancies in ratings of an article or item, then the reviewers discussed the source as a group to reach consensus on the rating. To this stage of the screening and selection reduction process, 224 published and 193 grey or unpublished items were selected for more detailed analysis and screening.
In order to chart and begin analysis of the content within the final 224 published articles selected as per the search terms, inclusion and screening criteria, a literature analysis worksheet was designed to gather specific information from each article which other review worksheets would not be able to provide. Rather than examining the methodology, sample sizes or quality of study, as in designed or validated review worksheets or guides, the worksheet designed for this scoping review was intended to gather information on the type of patient engagement (general or specific), type of information (e.g. tools, guides, resources, support, preparation), level within organization (e.g. governance, organization/system, programs), targeted study participants (e.g. patients, staff, leaders), study setting of the article (e.g. region, hospital) and level of engagement (e.g. information sharing, consultation, involvement, active engagement, partnership). Through this process, additional articles were rejected, reaching a final number of 74 to be analysed.
A similar worksheet was used with the 193 grey literature documents selected for review. In addition to topic categories being identified such as patient recruitment, engagement levels, or evaluation, the grey literature analysis worksheet incorporated checklists for identifying specifically whether items reviewed were reports, websites, toolkits, resources, education materials, evaluation and other items for each country selected. All 193 grey literature items were retained for final analysis.
All of the retained articles or items and their review worksheets (74 for published articles and 193 for grey items) were analyzed using content thematic analysis, a well-known qualitative analysis approach. The initial codes from this analysis were derived from the initial patient engagement terms and processes identified for the search and applied to all the articles/items and the worksheets. The themes were identified from the most commonly used codes across all of the articles/items and review worksheets. Specific patient engagement items categorized as tools, education or infrastructure were further identified and classified as inclusive of each theme. Summary tables of these various classified areas were constructed for reporting purposes.
## Limitations or challenges identified
Several limitations or challenges were identified as part of the study process. The first was the language and terminology specific to "patient engagement". The initial search of the published literature using 'patient engagement' resulted in very few hits. More hits were identified with "patient participation" and "patient involvement". Other terms inclusive of patient engagement were identified as the search continued, and included "consumer participation/ involvement/engagement", "citizen participation/involvement/engagement", and "public participation/ involvement/engagement". This meant that the search, if Applying 'patient engagement', 'patient involvement', 'patient participation' to those below:
Applying 'patient engagement', 'patient involvement', 'patient participation' initially followed with 'healthcare', 'decision making', advisory committees', and other terms pertaining to more specific engagement approaches, education/training, tools, resources, and infrastructure supports and evaluation considerations.
- 'knowledge and skills…'
- 'design, delivery and/or evaluation processes.
- '…readiness for meaningful engagement'.
- 'Process and impact of patient engagement'. Also applied terms to specific countries known for patient/public engagement strategies.
- '…resources (e.g. tools), needed by patients, providers, staff and leaders for effective patient engagement/involvement/ participation Australia United Kingdom and rest of Europe United States - '…preparation (e.g. education) needed by all stakeholders….'
- 'support (e.g. infrastructure) needed by everyone for ….'
- '…in patient-centered system redesign; …to build capacity; …to "pilot" a resource kit; …evaluate the impact'
- 'Patient engagement/involvement/ participation resource kit'
- '…best practices…validated tools…proven approaches'
- 'Recruit patients…'
- '…spectrum or level of engagement…ready for successful engagement'
- '…knowledge and skills needed
- '…engagement at different levels within the organization (governance, to program-based planning and evaluation)
- '…collaboration and partnering…'
- '…organizational policies and practices…'
- '…resources…'
- '…capacity…'
Inclusion criteria for articles Inclusion criteria of literature
[formula] - From 1995-2010 - From 1995-2010 [/formula]
- Written in English - Written in English
- Abstracts containing one or more of the key search terms or areas as identified in the research proposal.
- Contains one or more of the key engagement-related and healthcare terms or areas.
- Studies that refer to the involvement of patients at the program or governance levels
- From United States, Canada, Australia, and Europe (with special emphasis on United Kingdom).
## Exclusion criteria of articles exclusion criteria of literature
- Incorporate public or consumer engagement in areas outside of healthcare.
- Incorporates public or citizen engagement or similar areas rather than patient engagement, and in areas outside of health.
- Refer to involvement of the patients outside of the governance or program level.
- Refers to the involvement of the patients outside of the governance/ program level.
- Involve patients in their own treatment and care aspects or in personal healthcare decision making. limited to only "patient engagement" would not be comprehensive for this scoping literature review. Hence, all related or similar terms were searched, yielding many more hits for selection. A manuscript on this topic was published by the study team [bib_ref] The many faces of patient engagement, Gallivan [/bib_ref].
Another limitation was found with the semantics and overlap of different terms for "resources", "resource kit", "tool kit", "guides" and "resource tool kit", all related to education/information sources, training packages, models, strategies, approaches or processes, guides, workbooks and other such sources on patient engagement in the broadest sense. Regardless of overlap, all terms needed to be searched for optimal access to all relevant sources for this review.
Another challenge presented with the integration of the published and grey literature rather than keeping these two sets of literature separate for categorizing and clustering of categories into themes. There is no evidence to suggest there is a difference in which approach is used for the identification of the themes destined for use in aligning the contents of the "resource kit for patient engagement". The themes were selected because they best described the common findings from both published and grey literature.
The larger study focusing on the development and evaluation of a practical patient engagement resource kit for patients/families, providers and leaders received ethics approval through three separate ethics review boards in Alberta, Canada: the Community Research Ethics Board of Alberta (Protocol 3 1015); University of Alberta Health Ethics Review (Pro00018481_CLS6); and Calgary Health Research Ethics Board -Conjoint Health Ethics Board (I.D. # E-23545). Kit. Within each of the seven themes were sub-themes which were clustered into three main groups more in line with the expected content for a patient engagement Resource Kittools (including models, engagement approaches, barriers, benefits, evaluation guides), education (including stakeholder roles and expectations, defining and understanding patient engagement, overcoming barriers or turning them into enablers, using evaluation results to improve process and outcomes), and infrastructure (supports, finances, institutional support, capacity, resources). Through the synthesis of the literature, 15 different terms and definitions for 'patient engagement' were found along with 37 different engagement toolkits and frameworks (See [fig_ref] Table 2: Various toolkits for engagement and evaluation of engagement [/fig_ref] , and 17 models, each unique to target groups, with or without tools or guides, and some with specific process or evaluation frameworks or models (original or adapted).
# Results
## Definition of patient engagement
Upon reviewing the literature, it was evident that there were a number of different terms and definitions for 'patient engagement'. This topic was published as a separate manuscript [bib_ref] The many faces of patient engagement, Gallivan [/bib_ref]. Although the term 'patient engagement' was commonly used in discussions related to patients interacting and being meaningfully involved in health care initiatives, it was rarely used in the literature. Fourteen other terms were identified and defined by 26 different sources on the concepts of patient engagement including Citizen Engagement; Consumer Engagement; Involvement; Meaningful Patient Involvement; Participation; Patient and Public Engagement; Patient and Public Involvement (PPI); Patient-Centred Care; and Patient Involvement. Not only was the terminology for 'patient engagement' confusing, so was trying to define it. The terms 'involvement', 'engagement', and 'participation' were often used interchangeably. Forbat et al. [bib_ref] Patient and public involvement: models and muddles, Forbat [/bib_ref] concluded that "a range of ways of conceptualizing involvement are used interchangeably in policy and practice without due recognition of the very different meanings of public consultation, patient/carer involvement in treatment decision-making, and patient/carer involvement in service design and development" [bib_ref] Patient and public involvement: models and muddles, Forbat [/bib_ref] , p. 2547. As well, the terms 'meaningful' and 'involvement' did not mean the same to all stakeholders. "Meaningful involvement is not considered as an objective in itself in the contexts of projects and therefore have not been carefully planned for, resourced and evaluated" [bib_ref] Assessing the involvement of the patient communicty in European commission co-funded health..., Sanna [/bib_ref] , p. 270. The European Patient's Forum highlighted that "while there is diversity about the manner in which to interpret and implement patient involvement into the healthcare system, there is still a common challenge concerning the concept of meaningful patient involvement", p. 105.
## Stakeholder roles and expectations
A lack of consensus and understanding about terminology, the goals and expectations and roles and responsibilities of stakeholders were perceived as barriers to achieving meaningful and successful patient engagement. Forbat et al. concluded that "one of the greatest barriers to truly integrating patient involvement into health services, policy and research is the conceptual muddle with which involvement is articulated, understood and actioned" [bib_ref] Patient and public involvement: models and muddles, Forbat [/bib_ref] , p. 2547. For the purposes of informing a resource kit for 'patient engagement' , consistent terminology was viewed as an important consideration in setting up patient engagement initiatives. Choosing 'patient engagement' , for example, included all forms of involvement, participation, collaboration and engagement, and 'patient' in this context included families or family caregivers as well as others in the public domain. Different role descriptions and terms of reference for specific patient engagement initiatives in the literature articulated how patients and their families would be engaged or were expected to engage, and what the objectives, expectations or outcomes of the initiative were anticipated to be regarding patient engagement [bib_ref] Patient and public involvement: models and muddles, Forbat [/bib_ref] [bib_ref] Assessing the involvement of the patient communicty in European commission co-funded health..., Sanna [/bib_ref]. Expectations between patients and their families and organizations must be coordinated, hence, the differences in toolkits, frameworks and approaches, as identified in [fig_ref] Table 2: Various toolkits for engagement and evaluation of engagement [/fig_ref].
## Meaningful and appropriate engagement
There was a desire to ensure patients and their families provided their perspectives to help design and improve health services; however this was not easy. Broad representation of individuals with a variety of health related experiences would ensure a responsive approach to the needs of service users. Finding the right patient or consumer without an 'axe to grind' and who could represent the ordinary patient was the goal [bib_ref] Facilitating consumer participation: an approach to finding the 'right' consumer, Happell [/bib_ref]. Including one or more patients who had contextual knowledge and experience related to the engagement activity would be of most benefit, for example, "a consumer who has undergone transplant surgery, would be far better able to advise on the needs of consumers in these circumstances, than they would to provide advice on proposed changes to mental health legislation" [bib_ref] Facilitating consumer participation: an approach to finding the 'right' consumer, Happell [/bib_ref] , p. 128. Asking the right questions, such as, who wanted to be involved, and who should be involved, were integral to helping the organization meet its goals for public involvement and accountability [bib_ref] Bringing 'the public' into health technology assessment and coverage policy decisions: from..., Abelson [/bib_ref]. How to Develop a Community Based Patient Advisory CouncilThe overview provided by Chafe et al. [bib_ref] Deciding whether to engage the public on health care issues, Chafe [/bib_ref] indicated that the public wanted to be involved in varying degrees, with 25 percent of the public wanting to be involved in healthcare decision making but less than 10 percent wanting to be involved in difficult funding decisions, as they feel ill-equipped to contribute.
Attitudes toward patient engagement were not universal. Patient and provider perspectives of health, treatment, role, and organizational attributes differed and guided personal attitudes towards involvement. Farrellsuggested it was the attitudes of those in the care relationship, providers and patients, who played a central role in engagement; while a patient centered approach was one that "makes patients feel that they matter, that professionals are being honest with them and that meaningful discussion is possible". Staff that reacted to patients and their families in an "impatient, patronizing or disrespectful manner" inhibited future engagement opportunities (p. 23).
Based on feedback from participants of engagement opportunities, their involvement experience had been positive but they "become increasingly impatient when they perceive themselves to be a rubber stamp for decisions that are already taken", p. 14. Abelson and Eyles, believed that exposure to the health system through engagement opportunities enabled participants to have a greater understanding of the complexity of health sector decisionmaking and increased respect for decision makers.
Descriptions of the approach to involving stakeholders in decision making at the National Institute for Health and Clinical Excellence (NICE) stressed that "involving people is a serious business" [bib_ref] Involving stakeholders in healthcare decisions -the experience of the national institute for..., Culyer [/bib_ref] , p. [bib_ref] Effectiveness of strategies for informing, educating and involving patients, Coulter [/bib_ref] , and required open-mindedness as well as willingness to change and accommodate. The Institute for Patient and Family Centred Caresuggested that incorporating patients as 'champions' for engagement within the health system helped to avoid tokenism; for example, the patient advisor was involved in all stages of a project or initiative (from planning to evaluation).
## Models of engagement
A variety of patient engagement models were used by organizations, health systems and governments in the countries selected for this review, as shown in [fig_ref] Table 2: Various toolkits for engagement and evaluation of engagement [/fig_ref]. Since Arnstein's Eight Ladders of participationwas first developed to understand and explain citizen involvement, many adaptations have been created to clarify the meaning of involvement. The literature recognized 19 different models of engagement, some models being adaptations of original models. Aside from Arnstein's Ladder, one of the more popular and adapted models was from the International Association for Public Participationor IAP2 which used 'inform', 'consult', 'involve', 'collaborate' and 'empower' as levels of involvement. One adaptation of the IAP2 model was found in Health Canada's Policy Toolkit for Public Involvement in Decision Making using communication (inform or educate), listening (gather information), consulting (discuss), engaging and partnering. Using the levels of engagement for IAP2, there were some tested and documented methods of engagement under each level which can inform the Patient Engagement Resource Kit (depicted in [fig_ref] Table 3: Methods of engagement for each engagement level across the international association for... [/fig_ref].
## Benefits and barriers to patient engagement
Patient engagement was generally considered beneficial to the health care system in its policy and planning activities, but barriers were also identified. It was because of the known benefits and the management or resolution of barriers that specific enablers of patient engagement could also be identified. A 2010 Cochrane Review of patient involvement found that there were potential benefits and barriers or enablers to patient engagement across all levels of involvement, but there was also a "lack of research that reliably investigates whether consumer involvement achieves these intentions and, if so, which methods of consumer involvement are most effective", p. 4. Other studies, however, demonstrated the benefits for patients and decision makers at various levels to have patients engaged in face-to-face discussions and decisions concerning healthcare and health product decisions or issues [bib_ref] Service User Involvement in the Irish Health Service: A Review of the..., Mcevoy [/bib_ref]. The sharing of information, experiences and concerns between patients and decision makers was more than educational; it was also informative for healthcare recommendations.
One of the overarching benefits of patient engagement was that it enabled the health system to address the right issues in an appropriate way, design programs, policy and planning activities closely tailored to the needs of both individuals and special populations; achieve better results; and validate outcomes [bib_ref] Assessing the involvement of the patient communicty in European commission co-funded health..., Sanna [/bib_ref] [bib_ref] Service User Involvement in the Irish Health Service: A Review of the..., Mcevoy [/bib_ref]. General benefits found in the literature at both an individual and organizational level included better health and treatment outcomes, more appropriate and relevant services, increased legitimacy and credibility of decision making, increased sense of dignity and selfworth, and improved service user satisfaction. It was "assumed that input from consumers in the planning of health care can lead to more accessible and acceptable health services", p. 4.
Many of the benefits and barriers noted in the literature were case or project specific, or derived from stakeholder feedback. For instance, Howe et al. [bib_ref] Can the patient be on our team? an operational approach to patient..., Howe [/bib_ref] categorized barriers to patient engagement in patient safety initiatives as interpersonal, intrapersonal and cultural. Others categorized barriers as resources (e.g. time and cost), service user or patient issues, organizational issues, or systemwide barriers. Both Kovacs Burnsand Frankish et al. [bib_ref] Challenges of citizen particpation in regional health authorities, James [/bib_ref] identified broad barriers to participation in health care decision-making. Many of these challenges or barriers were related to values, assumptions and expectations underlying patient engagement in health authorities, structures and processes associated with decision-making. Kovacs Burns concluded that "Each of these challenges must be anticipated and managed in accordance with the partnership and process aspects that are part of the engagement framework".
More specific barriers/challenges identified by patients or families, care providers and leaders or administrators, are summarized in .
These were categorized generally into the following areas:
Legal -In general, although a high level of individual patient rights existed regarding healthcare, there was a gap when it came to 'patient involvement' as a right. Political -The lack of or poor political commitment to patient involvement at all levels in the healthcare system and especially at the policy decision level was one of the strongest barriers. Bureaucracy, including administrative procedures, reporting and technical skills required for some engagement activities [bib_ref] Service User Involvement in the Irish Health Service: A Review of the..., Mcevoy [/bib_ref] as well as the "lack of political will and government commitment to ensuring stakeholder engagement in decisions that concern policies or programs", p. 11, were difficult barriers to overcome. Administrative -Patient involvement was seen as inconvenient and time-consuming interrupting the smooth operation of a hierarchical, bureaucratic organization, especially if there was little or no knowledge about practices of involvement. Professional -Despite progress towards acceptance of a more important role for patients, attitudes of health professionals remained a strong barrier [bib_ref] Patient and public involvement: models and muddles, Forbat [/bib_ref]. Negative attitudes might manifest through professionals disengaging, not sharing information or resources, or exerting their power [bib_ref] Challenges of citizen particpation in regional health authorities, James [/bib_ref]. Much of this negativity could stem from professionals feeling threatened if they had to seek advice from expert patients; that it was a significant change from the medical model they were used to; or that it might question the role of health professionals [bib_ref] Service User Involvement in the Irish Health Service: A Review of the..., Mcevoy [/bib_ref] [bib_ref] Can the patient be on our team? an operational approach to patient..., Howe [/bib_ref]. Communication -Language, in terms of health literacy and especially with the use of technical terms, was a barrier to patient involvement. Personal -Characteristics of patients like ethnicity, age, disease and other relevant aspects might lead to discrimination, and therefore lower opportunities for involvement. Other considerations for patient and family involvement included their willingness to partcipate, commitments and time, transportation, wellness and health, language and communication, and fear of health care being jeopardized. Resources -There were two key aspects: a) the added value of patient involvement had not been quantified in economic terms and, thus had not been adequately compensated and b) meaningful patient involvement required resources.Focus groupForums for debatePatient advisory councils/ committees [bib_ref] Involving stakeholders in healthcare decisions -the experience of the national institute for..., Culyer [/bib_ref] [bib_ref] Service User Involvement in the Irish Health Service: A Review of the..., Mcevoy [/bib_ref] Citizen juryWebsitePatient surveys Health panelsExpert patientsConsumer managed project/service [bib_ref] Facilitating consumer participation: an approach to finding the 'right' consumer, Happell [/bib_ref] Press releasesFeedback and complaints (i.e. interviews, comment cards etc.) [bib_ref] Patient and public involvement: models and muddles, Forbat [/bib_ref] Shadowing patientsCharretteCitizen's panelsMail outsStory-tellingWorkshops [bib_ref] Service User Involvement in the Irish Health Service: A Review of the..., Mcevoy [/bib_ref] Constituent assemblyConsensus conferenceFact sheets [bib_ref] |Promoting patient-centered care: a qualitative study of facilitators and barriers in healthcare..., Luxford [/bib_ref] Social media (Facebook, Twitter, etc.)Public meetings Delphi ProcessDeliberative pollingHotline [bib_ref] Facilitating consumer participation: an approach to finding the 'right' consumer, Happell [/bib_ref] Planning meetings/ Forums RetreatsSearch conferenceDisplays and exhibitionsSuggestion boxes Round tablesStudy circlesPresentationsPatient diaries [26] Impact assessmentsStudy groups [bib_ref] Service User Involvement in the Irish Health Service: A Review of the..., Mcevoy [/bib_ref] Mystery shopping Ethics committeesSustainable community developmentWorld café Think tanksTown hall meetingsRevolving conversationEvaluation of patient engagement
Although the literature identified a variety of patient engagement toolkits and guides (Table 2 -A. General Toolkits), the evaluation of them and/or perspectives on those that have been evaluated regarding their use and effectiveness requires further study. An overview of patient engagement evaluation frameworks or algorithms used, evaluation of specific methods, and pilot testing templates for use in community health partnerships were not part of this scoping literature review. [fig_ref] Table 2: Various toolkits for engagement and evaluation of engagement [/fig_ref] also contains a list of evaluation toolkits used in various patient or citizen engagement initiatives, most applicable in evaluating the process and outcomes of engagement to make improvements. Anton et al. [bib_ref] Involving the public in NHS service planning, Anton [/bib_ref] examined the development of an assessment framework for public involvement and found that their multi-method study identified a lack of consensus for how public involvement should be evaluated.
They believed that evaluation in this sense was contextual and should be tailored to meet the needs of the engagement opportunity and its intended purpose [bib_ref] Involving citizens and patients in health research, Venuta [/bib_ref]. When evaluating the impact of public involvement policies, Wait and Nolte [bib_ref] Public involvement policies in health: exploring their conceptual basis, Wait [/bib_ref] , suggested that it "remains difficult to evaluate, partly due to many policies [being] short lived or very recent. Usually no timeframe or evaluative framework is specified for their assessment" (p. 157). The authors suggested further work was needed to define patient engagement policy objectives and to understand the dynamics of the various stakeholders within the health system in an effort to move healthcare systems closer to those which were responsive to the needs of patients, their families and the public. Caytonsuggested that there was "limited evidence to support the argument that patient involvement improves outcomes" and references Coulter and others who have argued that "patient experience, that is, what happened Benefits and barriers to patient engagement for patients, providers, leaders and institutions Barriers Benefits
## Patient barriers patient benefits
Personal and professional commitmentsHelps improve communicationsPatients seen as having the time and resources to participatenot always the case [bib_ref] Challenges of citizen particpation in regional health authorities, James [/bib_ref] Better understanding of health servicesHealth status and self-confidence [bib_ref] Patient and public involvement: models and muddles, Forbat [/bib_ref] Commitment to contributeTime to deal with diagnosisPatients meet other patientsFinancial considerationsneed expenses paidBecome empowered and valued for expertise and skillsTime availability & time for projectNot seeing direct personal benefit'Involvement fatigue'Meeting times (daytime meetings and work)
## Provider barriers provider benefits
Negative attitudes toward patient involvement [bib_ref] Challenges of citizen particpation in regional health authorities, James [/bib_ref] Builds trust and better communication between patients and staffLack knowledge of patient involvementProvides information about patient experience to inform planning and service improvementDismissive of how patients can contribute and not forthcoming with information/resources [bib_ref] Challenges of citizen particpation in regional health authorities, James [/bib_ref] Helps to provide accessible and responsive services based on local experience and needDifficulties/unwillingness to explain complex terminology [bib_ref] Challenges of citizen particpation in regional health authorities, James [/bib_ref] Enhances patient confidence in health systemFeel threatened by possible reduction of influence, and significant change from medical-model [bib_ref] Deliberations about deliberative methods: issues in the design and evaluation of public..., Abelson [/bib_ref] Difficulties in relinquishing power [bib_ref] Facilitating consumer participation: an approach to finding the 'right' consumer, Happell [/bib_ref] Affect on clinician/patient relationshipLeader/Instituion barriers Leader/Institution benefits Negative attitudes toward patient involvement [bib_ref] Deliberations about deliberative methods: issues in the design and evaluation of public..., Abelson [/bib_ref] More appropriate, better quality and relevant services [bib_ref] Service User Involvement in the Irish Health Service: A Review of the..., Mcevoy [/bib_ref] Lack of knowledge of how patients may be involved -little training or guidance for professionals in partnership working or joint decision-makingService responsive to patients' needs [bib_ref] Facilitating consumer participation: an approach to finding the 'right' consumer, Happell [/bib_ref] Tokenism [bib_ref] |Promoting patient-centered care: a qualitative study of facilitators and barriers in healthcare..., Luxford [/bib_ref] [bib_ref] Facilitating consumer participation: an approach to finding the 'right' consumer, Happell [/bib_ref] [bib_ref] Challenges of citizen particpation in regional health authorities, James [/bib_ref] Policy, research, practice and patient information that includes consumers' ideas or addresses their concernsLeadership may be questioned either wayOrganization is participative, accountable and transparentto them, rather than how satisfied they say they are, is a better measure of success" (p. 2). Parsons et al.echoed this sentiment as shown by the challenges they identified in measuring patient engagement in primary care settings. From the literature, key evaluation components were identified: participation or response rates of consumers, consumer influence on decisions, health care outcomes or resource utilization, consumers' or professionals' satisfaction with the involvement process or resulting products, cost, critical factors for success, and limitations of methods or processes. Part of a multistep process, these evaluation components were used to determine whether the engagement opportunity process was effective, the intended goal was achieved, and the engagement outcome had any contextual idiosyncrasies [bib_ref] Consumer Carer and Community Engagement Framework: for Health Services, Hosptials and WA..., Health [/bib_ref]. Rather than evaluation being a step that happens at the end of the engagement opportunity, Sheedysuggested that "Integrating these considerations into the planning process at the outset will save time and frustration at the end, and enable better learning from the process as it is taking place" (p. 33).
## Engagement resourcing
When costing a public dialogue opportunity, the Center for Public Dialogue suggested it was "impossible to provide a 'standard' cost estimate" as each engagement opportunity is unique. Considered part of a four-step process, costing for public dialogue opportunities included: consultation planning, testing of materials, implementation, and analysis and evaluation, p. 25. Engagement costs might include travel, accommodation, rental space, printing, translation, courier, long distance calls, time of department staff and remuneration of participants.
Considered a challenge to effective service user engagement, McEvoy [bib_ref] Service User Involvement in the Irish Health Service: A Review of the..., Mcevoy [/bib_ref] suggested that adequately resourcing patient engagement enables patients and families (service users) the opportunity to contribute through engagement programs, which if not resourced properly might become tokenistic exercises. The Improvement Leaders' Guide for Involving Patients and Carers suggested three important considerations: time, financing, and training and support.
Sheedy recommended that prior to engaging, all of the components for a successful engagement opportunity should be in place, including time, resources, and capacity. When planning a public or citizen engagement, Sheedy suggested that timing was everything, "while not all citizen engagement projects are time intensive, working with citizens will usually take longer than consulting experts", p. 22. As current budgets did not routinely include funds for engagement, Sheedy also suggested ensuring budget components like transportation, compensating for lost work time, and building internal staff capacity are factored in. All of these aspects were viewed as the necessary 'infrastructure' for decreasing barriers for patients to be able to participate and for enhancing their engagement experience.
# Discussion
This systematic scoping review of the published and grey literature provided a glimpse of the complexity of patient/ family engagement in health care decision making processes. The literature depicted a large diversity of terms and definitions [bib_ref] The many faces of patient engagement, Gallivan [/bib_ref] , tools and approaches to patient engagement in different settings, contexts, and purposes [fig_ref] Table 2: Various toolkits for engagement and evaluation of engagement [/fig_ref] , and barriers and benefits to consider . These were captured under seven content themes based on the analysis of selected published and grey literature -'definition of patient engagement','stakeholder roles and expectations', meaningful and appropriate engagement', 'models of engagement', 'benefits and barriers to patient engagement', 'evaluation of patient engagement', and 'engagement resourcing'. These themes could be useful for identifying or naming the content sections of the patient engagement resource kit. In addition, as outcomes of the scoping review the themes identified what knowledge, skills, tools, guides, models and resources existed and could be used or adapted by patients/families, providers or leaders for initiating and developing patient and family engagement. With synthesis of the literature, the sub-themes and information within each of the seven themes were clustered into three main groups more in line with the expected content for a patient engagement resource kittools, education and infrastructure.
This scoping literature review provided the opportunity to consider the most appropriate term and definition of patient engagement out of 15 choices, to meet the needs of stakeholders in their roles and expectations, and to identify different models of engagement with tools and guides to adopt or adapt to meet the needs of patient engagement within healthcare organizations such as Alberta Health Services. Over 37 toolkits were identified, of which 14 were general all-purpose patient-related engagement toolkits, and six were patient engagement evaluation toolkits or guides [fig_ref] Table 2: Various toolkits for engagement and evaluation of engagement [/fig_ref]. There were also 19 different engagement models, one or more of which could be adapted for use within AHS.Each of the thematic areas within the resource kit also included educational components. For example, key lessons should be heeded from the benefits and barriers to patient engagement . Creating awareness, clarifying terminology, providing training and tools and guides to understand and become knowledgeable about patient engagement while implementing it, were considered keys to its success. Addressing these lessons with the development of education and awareness tools would help to ensure the success of patient engagement initiatives and strategies. Training and education for all stakeholders, including patients, would help to achieve common language and common grounds for understanding the benefits of and the 'how-to' of patient engagement. Providers who have received education would be less threatened and more knowledgeable about how to involve patients and what resources it would take to do so.
Some sources suggested that one of the biggest cultural obstacles to patient engagement was a result of staff and decision-makers making assumptions that patients do not have the knowledge about healthcare operations to be involved in its decision-making processes [bib_ref] Can the patient be on our team? an operational approach to patient..., Howe [/bib_ref]. "This is a cultural issue that has largely arisen from professionalization and specialization that leads experts to believe that non-experts have nothing or little to contribute" [bib_ref] Patient and public involvement: models and muddles, Forbat [/bib_ref] , p. 25. Farrell said it best:
Staff education and training will always be integral to any program of change within the [health system]. Initiatives to improve patient and public involvement must address the knowledge, attitudes and skills of professionals and staff at all levels of the service. Hands-on experience could be a powerful way to change attitudes as this opens people's eyes to the real potential involvement but formal training and education is also crucial, p. 39.
If education could increase awareness of the benefits and help to eliminate the barriers to patient engagement, then meaningful and effective patient engagement was more likely to occur. Without evaluation of patient engagement initiatives including tools, guides, approaches or process used, and resources to support the process and outcomes, it would be impossible to know what aspects were working well, or which ones needed improvement or changes. Obtaining direct patient feedback on the engagement experience was practical for understanding and improving the process and outcomes for patients and others.
Engagement resourcing of tools and educational or other activities was crucial to engagement success and should be considered when planning any stakeholder engagement. As the engagement of patients and their families was a new area of study, more will continue to be discovered on how best to resource departments invested in patient engagement, the activity itself and the mobilization of the lessons into practice. This was seen as too important to be happening off the side of someone's desk. Dedicated resources should be applied to ensure the health system had the capacity to be responsive to the needs of patients and their families.
The term infrastructure was chosen here for the financial and human resourcing and related supports needed for patient engagement. This term was chosen for several reasons, although the literature does use alternative terms Lansky described the current state of patient engagement 'infrastructure' within the United States, and believed that a lack of national health information and an absence of a centralized management and finance system limited the ability for patients to play a more active role. In efforts to build infrastructures which helped to cultivate engagement opportunities, Abelson & Eyles suggested, "building social capital and civic infrastructure is largely a matter of removing the constraints that often truncate that self-organizing process and of improving the space it needs to flourish", p. 19. Like Nova Scotia's Community Health Boards and Saskatchewan's Citizen's Advisory Councils, Alberta legislated the creation of 12 Community Health Councils to provide feedback directly from local Albertans, including patients and their families, about what was working well in the health care system and areas in need of improvement.
Other practical examples of 'infrastructure' supporting patient engagement included the emergence of organizational policies, legislation and national health plans, as demonstrated by the Queensland Government endorsement of a Community Engagement Improvement Strategy, England's Health and Social Care Act 2001 which "places a legal duty on health care organizations to make arrangements to involve and consult patients and the public and to develop an ongoing relationship rather than a consultation being a one off"
# Conclusions
This scoping review on patient engagement, including the identification of key themes and resources relevant to making patient engagement successful, offered a wealth of information that would not only assist in the development of a resource kit for patient engagement but also provide other significant suggestions and recommendations to be considered in the patient engagement process. As follow up to this scoping review and the identification of themes and content (tools, education and infrastructure) items, is the actual development and pilot of the Resource Kit including evaluating the strengths/effectiveness as well as weaknesses of each theme and the content items within each theme. This latter will be an ongoing process within AHS, and the Resource Kit will be remain a dynamic initiative to keep items current and of practical use.
The following were some of the key highlights from this scoping literature review:
This scoping literature review was comprehensive and unique compared with other literature reviews found on patient engagement. This review included both published and grey literature, and analyzed both in the context of their contributions towards patient engagement in the broadest sense, considering tools, guides, barriers or benefits, and other attributes. Therefore, this scoping review will be filling a gap in the literature concerning patient engagement. One of the key findings with the search process was that the term 'patient engagement' by itself was inadequate for searching the published and grey literature. Other more commonly used and related terms had to also be part of the search criteria; otherwise, the search would not be adequate, and might not provide access to key sources of patient engagement tools and guides for patients, staff and leaders. This could be interpreted to mean that caution must be taken to not narrow the search terms too quickly but rather be more inclusive as different groups, organizations and researchers have slightly different preferences for 'patient engagement' terminology and definitions. The selected relevant literature within this scoping review contributed to one or more clustered areas as tools/resources/approaches, education, and support/infrastructure for each of the identified user groupspatients, providers, leaders and AHS Patient Engagement Department. There was a mix of literature relevant across the user groups as well as some specifically targeted literature for specific user groups. This literature will inform the Resource Kit proposed to provide patients, providers and leaders with the information and tools to make patient engagement meaningful and successful. Patient engagement is not easy and was in fact quite complex-the literature identified it as being very challenging as it presented with many potential barriers to anticipate or consider in addition to the benefits. From the literature, it was clear what the barriers and benefits were with only some having been evaluated as to their impact. Keeping track of the barriers (legal, political, administrative, professional, communication, personal or resources) as a frame of reference, would help identify tools, guides and strategies to avoid or deal with them before they became a fatal flaw in the patient engagement initiative or process. The goal would be to deal with the barriers and forge a clear path for enablers of patient engagement. More research was needed to evaluate the benefits of patient engagement in different settings and contexts. Having an evaluation plan for each patient engagement initiative at the outset would help establish what and how measures were taken and made for both patient engagement processes and outcomes. Evaluation guides and tools were part of this literature review. From an evaluation point of view, Coulter and Ellins conducted a review of the literature for evidence on patient focused quality interventions identifying that while the results of most reviews were positive (beneficial effect), "choosing appropriate criteria to evaluate patient focused interventions is difficult…and the lack of standardized outcome measures hampers comparison of results" [bib_ref] Effectiveness of strategies for informing, educating and involving patients, Coulter [/bib_ref] , p. 24. It was recommended that this scoping review and the Resource Kit be dynamic and updated accordingly, especially if it was on the AHS Patient Engagement Department website for patients, providers and leaders to access and implement.
[fig] Figure 1: See legend on next page.) [/fig]
[table] Table 1: Search terms and general inclusion/exclusion criteria for published and unpublished (Grey) literature search [/table]
[table] Table 2: Various toolkits for engagement and evaluation of engagement [/table]
[table] Table 3: Methods of engagement for each engagement level across the international association for public participation (IAP2) spectrum [/table]
|
Brunner's Gland Adenoma–An Uncommon Cause for Intussusception and Gastric Outlet Obstruction
Brunner's gland adenoma is extremely uncommon small bowel tumors with an incidence of ,0.01% and account for less than 1% of all gastrointestinal tumors. They are branched acinotubular glands found within the submucosal layer and located between the pyloric ring and the major duodenal papilla. Brunner's glands produce an alkaline secretion containing viscous mucin to protect the duodenum from acidic gastric chyme. Although these lesions are usually asymptomatic and are incidentally discovered on upper gastrointestinal endoscopy, they may occasionally present with symptoms of upper gastrointestinal hemorrhage, duodenal obstruction, and more rarely with biliary fistulation or intussusception. We present an atypical case of a large 9-cm Brunner's gland adenoma causing duodenojejunal intussusception in a 44-year-old Chinese man, who presented with long-standing epigastric pain, nausea, and vomiting.
# Introduction
A 44-year-old Chinese man with no previous comorbidities, no chronic medications, and no surgical history presented to the emergency department with an episode of self-limiting acute epigastric pain, nausea, and vomiting for 2 days. [bib_ref] Surgical management of giant Brunner's gland hamartoma: Case report and literature review, Stewart [/bib_ref] [bib_ref] Brunner's gland adenoma-a rare cause of gastrointestinal bleeding: Case report and systematic..., Sorleto [/bib_ref] [bib_ref] Brunner's gland hamartoma presenting as gastric outlet obstruction: Unusual presentation and review..., Gupta [/bib_ref] His bloating was worse after meals and increased in relation to the portion of food consumed. He reported long-standing intermittent epigastric discomfort with bloating for 3 months. His vital signs were within normal limits. The physical examination revealed no remarkable findings. His abdomen was soft without any palpable abdominal mass, guarding, or rebound tenderness. Laboratory investigations did not reveal any significant abnormality.
## Case report
Upper gastrointestinal endoscopic examination performed showed large amount of food residue in the stomach, although the patient was adequately fasted. His stomach was capacious, with a distorted antrum. The first part of the duodenum was distorted and dilated with
# Discussion
Brunner's glands are named after the Swiss anatomist Johann Conrad Brunner, who first described them in 1,688.Hyperplasia of Brunner's glands to form a lesion greater than 1 cm is described as a Brunner's gland adenoma (BGA) or Brunner's gland hamartoma. [bib_ref] Large Brunner's gland adenoma: Case report and literature review, Rocco [/bib_ref] Brunner's glands are located predominantly in the first part of the duodenum and produce alkaline secretions containing mucus, pepsinogen, and urogastrone in response to acid stimuli. Their size and number decrease progressively in the distal parts of the duodenum. Most BGAs are in the duodenal bulb (57%), followed by second (27%) and three times a day (7%) parts of the duodenum. They are seldom found in the pyloric canal (5%), jejunum (2%), or proximal ileum (2%). [bib_ref] Large Brunner's gland adenoma: Case report and literature review, Rocco [/bib_ref] The etiology and pathogenesis of BGAs are not fully understood. Suggested causes include factors causing chronic local irritation, such as Helicobacter pylori infection, Billroth II reconstruction, or chronic pancreatitis. [bib_ref] Brunner's gland hyperplasias and hamartomas in association with Helicobacter pylori, Destek [/bib_ref] [bib_ref] Brunner[apos ]s gland hamartoma simulating a pancreatic mass with duodenal obstruction, Mumtaz [/bib_ref] BGA is a benign lesion, but there have been case reports describing neoplastic lesions arising from the Brunner's gland. [bib_ref] Carcinoma of duodenum arising from Brunner's gland, Akino [/bib_ref] Most patients with BGA are asymptomatic and are only incidental during endoscopy or imaging studies. Symptomatic patients with BGA tend to have adenomas larger than 2 cm in size and commonly present with gastrointestinal bleeding or obstructive symptoms. [bib_ref] Brunner's gland hamartomas: Clinical presentation and pathological features of 27 cases, Levine [/bib_ref] Less common presenting symptoms include intussusception, biliary fistula, recurrent pancreatitis, or obstructive jaundice. To date, there have been few published case reports of BGA associated with acute or chronic gastric outlet obstruction and intussusception.
The diagnosis of BGA can be challenging. Imaging findings are usually nonspecific and may suggest adenoma, adenocarcinomas, lymphoma, carcinoid tumors, gastrointestinal stromal tumors, leiomyomas, leiomyosarcomas, and metastatic disease. [bib_ref] Gastrointestinal bleeding from a Brunner's gland hamartoma: Characterization by endoscopy, computed tomography,..., Block [/bib_ref] Being submucosal in origin, traditional endoscopic punch biopsies are usually negative. Because a Brunner's gland may be covered by normal mucosa, confirmatory diagnosis is only established after a deep endoscopic or a surgical biopsy, with consequent histological examination of the excised mass. [bib_ref] Gastric outlet obstruction caused by Brunner's gland hyperplasia: Case report and review..., Krishnamurthy [/bib_ref] Treatment options include endoscopic removal of small and/or pedunculated stalked lesion, endoscopic mucosal resection with piecemeal technique for larger adenomas, and surgical resection where endoscopic therapy is not feasible. [bib_ref] Successful endoscopic resection of large pedunculated Brunner's gland hamartoma causing gastrointestinal bleeding..., Jung [/bib_ref] Endoscopic management is the preferred initial mode of therapy. Where feasible, endoscopic ultrasound can provide further information about the depth of involvement and any submucosal vasculature that may be associated with the BGA. Cases of BGA with an intact muscularis propria layer can be managed endoscopically with endoscopic submucosal dissection. 14 In our case, surgery was preferred because of a large lesion complicated with intussusception.
[fig] Figure 1: EGD findings: Left: Antrum. Right: Duodenum as seen on entry of esophagogastroduodenoscopy. EGD, esophagogastroduodenoscopy. [/fig]
[fig] Figure 2: CTAP: Distended gastric cavity with the presence of intussusception, demonstrating pathognomonic findings of a targetlike mass. CTAP, computed tomography of the abdomen and pelvis. prominent valvulae conniventes (Figure 1), which are normally only seen starting in the distal length of the duodenum closer to the duodenojejunal flexure. 4 Computed tomography showed a long segment duodenojejunal intussusception, seen as a target-like mass, with proximal dilatation of the stomach (Figure 2). Subsequent surgical laparotomy revealed a large polyp measuring 9.0 3 5.0 3 4.0 cm, arising from the first part of the duodenum (Figure 3). Histology showed features of a benign Brunner's gland adenoma (BGA) consisting of multiple nodules of closely packed Brunner's glands, separated by fibromuscular septa. No mitotic figures were observed (Figure 4). Our patient underwent an uneventful postoperative recovery with resolution of symptoms. [/fig]
[fig] Figure 4: Histological findings. Histopathological appearance of a giant Brunner's gland adenoma stained with Hematoxylin and Eosin. Area of magnification showing lobules of hyperplastic Brunner's glands separated by fibromuscular septa. [/fig]
[fig] Figure 3: Postoperative specimen showing a large polyp measuring 9.0 3 5.0 3 4.0 cm arising from the first part of the duodenum. [/fig]
|
Peripheral perfusion index predicting prolonged ICU stay earlier and better than lactate in surgical patients: an observational study
Background: Peripheral perfusion index (PPI) is an indicator reflecting perfusion. Patients undergoing long time surgeries are more prone to hypoperfusion and increased lactate. Few studies focusing on investigating the association between PPI and surgical patients' prognoses. We performed this study to find it out. , we retrospected all surgical patients who were transferred to ICU, Xinyang Central hospital, Henan province, China. Inclusive criteria: age ≥ 18 years old; surgical length ≥ 120 min. Exclusive criteria: died in ICU; discharging against medical advice; existing diseases affecting blood flow of upper limbs, for example, vascular thrombus in arms; severe liver dysfunction. We defined "prolonged ICU stay" as patients with their length of ICU stay longer than 48 h. According to the definition, patients were divided into two groups: "prolonged group" (PG) and "non-prolong group" (nPG). Baseline characteristics, surgical and therapeutic information, ICU LOS, SOFA and APACHE II were collected. Besides we gathered data of following parameters at 3 time points (T0: ICU admission; T1: 6 h after admission; T2: 12 h after admission): mean artery pressure (MAP), lactate, heart rate (HR), PPI and body temperature. Data were compared between the 2 groups. Multivariable binary logistic regression and ROC (receiver operating characteristic) curves were performed to find the association between perfusion indictors and ICU LOS.Results: Eventually, 168 patients were included, 65 in PG and 103 in nPG. Compared to nPG, patients in PG had higher blood lactate and lower PPI. PPI showed significant difference between two groups earlier than lactate (T 0 vs T 1 ). The value of PPI at two time points was lower in PG than nPG(T0: 1.09 ± 0.33 vs 1.41 ± 0.45, p = 0.001; T1: 1.08 ± 0.37 vs 1.49 ± 0.41, p < 0.001). Increased lactate T1 (OR 3.216; 95% CI 1.253-8.254, P = 0.015) and decreased PPI T1 (OR 0.070; 95% CI 0.016-0.307, P < 0.001) were independently associated with prolonged ICU stay. The area under ROC of the PPI T1 for predicting ICU stay> 48 h was 0.772, and the cutoff value for PPI T1 was 1.35, with 83.3% sensitivity and 73.8% specificity.Conclusions: PPI and blood lactate at T 1 (6 h after ICU admission) are associated with ICU LOS in surgical patient. Compared to lactate, PPI indicates hypoperfusion earlier and more accurate in predicting prolonged ICU stay.
# Introduction
Hemodynamic changes in postoperative patients are complex and various. Many patients would manifest increased level of blood lactate after long-term surgeries. Ensuring adequate perfusion is essential for recovery and better prognosis.
Many indicators have been proved reliable in reflecting perfusion. Among them, blood lactate is one of the widest used indexes guiding fluid management and resuscitation. But it still has some drawbacks. The most evident one is lactate is delayed in reflecting real-time perfusion. As we all know, from the beginning of hypoperfusion to elevated lactate, there is a time interval. If we use lactate as an indicator of hypoperfusion, we will take measures later than the real start of inadequate perfusion, which is why we turn to a more timely parameter: peripheral perfusion index (PPI).
Some researchers have proved PPI is a reflector of hypoperfusionand is associated with patients' mortality. But most of these studies are about septic shock. Apart from that, no study has been about comparing lactate with PPI, trying to figure out which one is superior in predicting patients' outcomes.
If we can prove PPI is comparable or superior to lactate in reflecting postoperative patients' prognoses, we will take beneficial interventions earlier because PPI is more instant than lactate in mirroring perfusion state. That is why we performed this study to investigate the role of PPI in predicting postoperative patients' prognoses and to compare PPI with lactate, finding out their strengths and limitations.
# Patients and methods
## Patients
The Institutional Research and Ethics Committee of the Xinyang Central Hospital approved this study. Written informed consent was waived since it was a retrospective study.
This was a retrospective study of all surgical patients who were admitted to ICU in Xinyang Central Hospital (Henan province, China) from January 1st, 2019 to September 30th, 2019. The inclusive criteria are following:1) age ≥ 18 years old; 2) surgical length ≥ 120 min. Exclusive criteria: 1) died in ICU; 2) discharging against medical advice; 3) existing diseases affecting peripheral blood flow, for example, vascular thrombus in arms; 4) severe liver dysfunction.
Prolonged ICU stay was defined as ICU LOS longer than 48 h (from the time of arrival in ICU). Patients were divided into two groups according to the definition: "prolong group" (PG) and "non-prolong group" (nPG).
## Data collection
Before we commenced data collection, we had estimated the minimum sample size we needed to obtain (supplement.1). The electronic medical records and anaesthesia charts of all recruited patients were reviewed to collect the following information: gender, age, comorbidities(CAD: coronary artery disease; COPD: chronic obstructive pulmonary disease; CKD: chronic kidney disease; HTN: hypertension; PMI: pacemaker implantation), surgical types and length, volume of intraoperative fluid infusion, fluid balance of 0-6 h and 0-12 h in ICU, the application of vasoactive agent in the first 12 h in ICU, SOFA(Sequential Organ Failure Assessment), APACHE II(Acute Physiology and Chronic Health Evaluation) on 1st day in ICU and the length of ICU stay.
We also collected each patient's perfusion associated parameters at three different time points: ICU admission(T0), 6 h after ICU admission(T1), 12 h after ICU admission(T2). The perfusion related variables include mean artery pressure (MAP), heart rate (HR), peripheral perfusion index (PPI) and body temperature (axillary). MAP, HR and PPI would be obtained by monitors continuously (Philips IntelliVue MP50; Koninklijke Philips, The Netherlands) and exported to our medical records database automatically. PPI was captured by the pulse oximeter and blood pressure was measured using the oscillometric non-invasive technique or via radial arterial cannulation using pulse contour analysis.
As for PPI, we used its absolute value in every 15 min to calculate the averaged value for each hour. All PPIs were averaged values (in the first, sixth and twelfth hour in ICU).
# Statistical analysis
This is a retrospective study only involving in-hospital information. We assumed missing data were limited to a small number of observations and adopted likewise deletion to handle missing data.
Descriptive analyses were performed for the nPG and PG groups. Continuous variables are expressed as means ± standard deviations, and categorical variables are expressed as absolute values and percentages. For the continuous variables, the data were analyzed using Student's t-test, the Mann-Whitney U test or the Kruskal-Wallis test depending on the data distribution and the number of variables. The categorical variables were analyzed using chi-square or Fisher's exact tests.
Before we performed the multivariate logistic regression analysis, univariate analysis was performed for each potential risk factor of prolonged ICU stay. We defined P < 0.05 in univariate analyzation as the threshold of a variable qualified to be included in multivariate logistic regression model. As for indicators associated with prolonged ICU stay, the receiver operating characteristic (ROC) analyses were performed to test each indicator's efficacy in predicting prolong ICU stay. All comparisons were two-tailed, and P < 0.05 was required to exclude the null hypothesis. The statistical analysis was performed using IBM SPSS Statistics, Version 20.0 (Armonk, NY: IBM Corp).
# Results
From January to September 2019, a total of 252 surgical patients were transferred to ICU, Xinyang Central Hospital, Henan Province, China. 84 patients were excluded for several reasons and 169 patients were recruited finally. They were divided into 2 groups according to their ICU LOS: 65 patients in "prolonged group", 103 patients in "non-prolonged group". Main reasons for each patient in PG were showed in supplement 2. The ICU LOS was 66.63 ± 13.48 h in PG and 35.21 ± 8.19 h in nPG, P < 0.001. The flow chart of patients' enrollment was showed in.
## Baseline characteristics
We analyzed the basic characteristics (age, gender, APA-CHEII and SOFA on 1st day) between the two group, and no significant differences were observed. We calculated the number of patients who have different comorbidities (CAD: coronary artery disease; COPD: chronic obstructive pulmonary disease; CKD: chronic kidney disease; HTN: hypertension; PMI: pacemaker implantation) and no differences were found between the 2 groups. More details were displayed in the.
## Surgical and therapeutic information
The surgical length of PG and nPG was 208.78 ± 63.58 min and 197.72 ± 54.33 min, respectively, with P value being 0.100. For prolonged and non-prolonged patients, amount of intraoperative fluid infusion was 1687.86 ± 465.75 ml and 1575.00 ± 419.97 ml(P = 0.295) while amount of fluid input in the first 24 h in ICU was 2357.90 ± 409.88 ml and 2115.01 ± 238.77 ml(P = 0.006), respectively. Considering types of surgery could influence postoperative perfusion, we collected data of surgical types (abdomen, chest, lung, esophagus, gynecology, great artery, others). None of them were significantly different between two groups. Vasoactive agent is another factor influencing peripheral prefusion. The number of patients with application of vasoactive drugs in the 1st 12 h in ICU were 7 out of 65 in PG and 15 out of 103 in nPG (P = 0.478). More details were displayed in the.
## Hemodynamic and perfusion indicators
At three time points (T 0 , T 1 , T 2 ), we did not see any differences on HR, MAP and body temperate between the 2 groups. When patients arrived in ICU, there were no differences on the immediate (T 0 ) lactate between PG(3.08 ± 0.72 mmol/L) and nPG(3.09 ± 0.73 mmol/L). But after six(T 1 ) and twelve hours(T 1 ), lactate in prolong group was higher than no-prolong group(T1: PG/nPG, 3.13 ± 0.67/2.62 ± 0.59 mmol/L, P = 0.009; T2: PG/nPG, 3.25 ± 0.44/2.02 ± 0.42 mmol/L, P < 0.001).
Compared to lactate, PPI showed earlier indication: the value was lower in PG than nPG at T 0 (1.09 ± 0.33 vs 1.41 ± 0.45, P = 0.001) as well as T 1 (1.08 ± 0.37 vs 1.49 ± 0.41, P < 0.001). But the difference disappeared after 12 h (T 2 , PG/nPG: 1.17 ± 0.27/1.23 ± 0.33, P = 0.392). More information was presented in.
## Univariate and multivariable analysis
Univariate regression analysis was performed for all potential risk factors of prolonged ICU stay. Only two variables were found significantly related: PPI (T 0 , T 1 ), lactate (T 1 , T 2 ). More details showed in. Because T 1 was the only time point at which both PPI and lactate were significantly associated with prolonged ICU stay, we performed the multivariable logistic regression using PPI T1 and lactate T1 .
The multivariable logistic regression analysis demonstrated that lactate T1 (OR 3.216; 95% CI 1.253-8.254, P = 0.015) and PPI T1 (OR 0.070; 95% CI 0.016-0.307, P < 0.001) were independent predictors of prolonged ICU stay in surgical patients. In, more details of the logistic regression model were showed.
# Discussion
The most important finding of this research is association between PPI and length of ICU stay. If PPI can reach 1.35 within the first 6 h in ICU, it gives physicians a hint that this patient could discharge from ICU within 48 h, with 83.3% sensitivity and 73.8% specificity. Besides, we also proved that PPI indicated prolonged ICU stay more superior and earlier than lactate.
Why PPI can predict prolonged ICU stay and is superior to lactate
Peripheral perfusion index (PPI) is an indicator mirrorring inadequate perfusion in critical patients. It is measured using pulse co-oximetry technology which is characterized by being real-time and noninvasive. PPI is defined as "the ratio of pulsatile blood flow to the non-pulsatile blood flow", mirroring the strength of blood flow and quality of perfusion at sensor site, reflecting perfusion state of the body part. Compared to blood lactate testing, PPI is real-time and non-invasive.
Many researches have proved quality of perfusion is essential for critical patients in the process of recovery. Poor perfusion prolongs patients' inhospital stay and increases mortality and morbidity. That is why PPI, a regional perfusion indicator, is related to length of ICU stay. Lower PPI means poorer perfusion and longer ICU stay. But there is another question: why PPI is superior to lactate, an index that has been widely validated and used. Before we figure it out, we need to know alterations of blood flow during hypoperfusion.
At the initial stage of hypoperfusion, peripheral vessels contract in order to return enough blood to heart. In this phase, macro vital signs, like HR and BP, are normal. Because macro circulation is stable, blood lactate does not increase. But PPI, an indicator of regional perfusion, will decrease because vascular contraction reduces reginal blood flow. This makes PPI superior to other macro parameters, for example lactate, in alerting physicians to hypoperfusion.
We can see this superiority of PPI from our results. At T 0 , PPI showed significant differences between PG and nPG, but lactate was not statistically different. After 6 h, lactate became significantly different between PG and nPG at T 1 . There was a delay in lactate comparing to PPI. In order to find out more details about tendency of lactate in two groups, we performed Wilcoxon signedrank test (abnormal distribution) and found out that lactate showed no obvious decreasing trend in prolonged group while dramatically reduced in non-prolonged group (see supplement3). PPI was higher in nPG than PG both at T 0 and T 1 . Considering that PPI has been proved effective in reflecting perfusion state, the different tendencies in two groups probably result from PPI. From the. In order to confirm lactate is not real-time enough to predict poor prognosis, we performed a similar multivariable logistic regression (see supplement 4). In that model, lactateT 0 is not associated with prolonged ICU stay while PPI T0 is relevant with it. We also plotted ROC curves of PPI T1 and lactate T1 for predicting prolonged ICU stay. The AUC of PPI T1 and lacatate T1 are 0.772 and 0.677 respectively. These results showed us that PPI predict prolonged ICU stay earlier and better.
## The role of ppi in managing postoperative patients' treatment
Patients undergoing long time surgeries tend to manifest high level of blood lactate. It is known to us all that hyperlactacidemia is a strong indicator for poor prognosis. If we want to reduce blood lactate, we need to ensure adequate perfusion. According to our conclusions, intensivists should pay more attention to PPI, an invasive and real-time indicator, which has been proved superior to lactate in predicting patients' prognosis. In a previous study, PPI < 1.4 is a sign of hypoperfusion in critical patients. Our conclusion is that PPI < 1.35 after 6 h in ICU predicts LOS ICU longer than 48 h with 83.3% sensitivity and 73.8% specificity, which is similar to Lima AP's study.
In clinic practice, intensivists should be aware of postsurgical patients' alterations of PPI. If PPI shows obvious trend of declination, doctors should be cautious of hypoperfusion and poor prognosis. They need to find out the reason of its decrease and rule out non-perfusion relevant factors, trying to restore adequate perfusion.
## Factors that influence ppi
One of the reasons limiting the use of PPI is that it is easily influenced by other factors. Low temperature and vascular diseases are the two main confounding factors which could affect value of PPI. Besides, the position can also affect PPI. If we place patients' arms on unsuitable positions, for example, placing arms under a weight, it will have negative effects on regional perfusion. Before we apply PPI into clinic practice, we need to keep in mind that lower value does not mean poor perfusion unless we rule out other confounding factors. In this study, we excluded all patients with diseases affecting regional blood flow. In our routine nursing, we give warm-keeping service to all the patients, for example covering blankets. If necessary, we provide warm blower to maintain patients' body temperature. Usually, we do not place any weights on patients' arms and all patients are in supine positions unless they require special treatments, for example prone position for acute respiratory distress syndrome. All patients enrolled in our study were supine.
What is new about PPI in our study compared to previous ones Many studies proved PPI's predictive role in hypoperfusionand its association with mechanically ventilated patients' mortality. What is novel in our study is that we compared PPI with lactate, which has not been showed in previous study and defined another important aspect of PPI: association with length of ICU stay.
Before our study, there is an important research carried out by Van Genderen MEabout exploring the superiority of different peripheral perfusion indicators. In that study, the siginificance of parameters relecting the state of peripheral perfusion in predicting the occurrence of post-operative complications had been proved. But it did not compare these indicators with the most widely used parameters relefcting perfusion: blood lactate. If we want to popularize a new parameter, we must compare its property with the extensively used one. That is why we perform our study, and luckily, we proved PPI's strenghth over lactate. In addition to that study, there is another research, written by Su L etc., about the role of PPI in predicting mortality of patients with mechanical ventilation. Unlike Su's research, mainly focusing on patients with respiratory failure requiring mechanical ventilation, our study is about applying PPI in postoperative management.
# Limitations
This study has several limitations. First, its sample size is small. More studies on PPI with bigger sample size should be carried out in the future. Second, patients who died in ICU or discharged against medical advice were excluded. This would cause the patients included in our study less severe. As we can see from the results, most patients' PPI could reach to around 1.2 within 12 h. Besides, the highest average blood lactate in both groups was less than 4.0 mmol/L. These phenomena indicate patients enrolled in the study were not so serious. It is better to confine our conclusions to surgical patients who are not in life-threatening conditions. As for patients on the edge of death, predicting their LOS ICU is much more complicated. Third, our studying period is only 12 h. The reason why we focus on a short time is that we want to foresee patients' outcomes in early clinic phase. It is worth to note that many complications could appear in the late stage, these complications may postpone the discharge. A previous study had confirmed PPI alteration was associated with development of severe complications. So, we have confidence that decreased PPI is associated with prolonged ICU stay even in the long term.
# Conclusions
For surgical patients transferred to ICU, PPI and lactate at T1(6 h after ICU admission) are associated with length of ICU stay. Compared to lactate, PPI reflects prolonged ICU stay earlier and better than lactate. The value of PPI at T1 lower than 1.35 has 83.3% sensitivity and 73.8% specificity in predicting LOS ICU longer than 48 h.
## Supplementary information
Supplementary information accompanies this paper at https://doi.org/10. 1186/s12871-020-01072-0.
Additional file 1.
## Additional file 2.
Additional file 3. |
DNA-SIP based genome-centric metagenomics identifies key long-chain fatty acid-degrading populations in anaerobic digesters with different feeding frequencies
Fats, oils and greases (FOG) are energy-dense wastes that can be added to anaerobic digesters to substantially increase biomethane recovery via their conversion through long-chain fatty acids (LCFAs). However, a better understanding of the ecophysiology of syntrophic LCFA-degrading microbial communities in anaerobic digesters is needed to develop operating strategies that mitigate inhibitory LCFA accumulation from FOG. In this research, DNA stable isotope probing (SIP) was coupled with metagenomic sequencing for a genome-centric comparison of oleate (C 18:1 )-degrading populations in two anaerobic codigesters operated with either a pulse feeding or continuous-feeding strategy. The pulse-fed codigester microcosms converted oleate into methane at over 20% higher rates than the continuous-fed codigester microcosms. Differential coverage binning was demonstrated for the first time to recover population genome bins (GBs) from DNA-SIP metagenomes. About 70% of the 13 C-enriched GBs were taxonomically assigned to the Syntrophomonas genus, thus substantiating the importance of Syntrophomonas species to LCFA degradation in anaerobic digesters. Phylogenetic comparisons of 13 C-enriched GBs showed that phylogenetically distinct Syntrophomonas GBs were unique to each codigester. Overall, these results suggest that syntrophic populations in anaerobic digesters can have different adaptive capacities, and that selection for divergent populations may be achieved by adjusting reactor operating conditions to maximize biomethane recovery.
# Introduction
Anaerobic digestion is an attractive biotechnology for recovering renewable energy as biomethane (CH 4 ) during organic waste treatment. Worldwide interest in anaerobic digestion has led to the implementation of~9000 full-scale facilities in Europe and over 3000 facilities in the US by 2012 [bib_ref] Bioenergy and biofuels: History, status, and perspective, Guo [/bib_ref]. The conversion of organic matter into biomethane relies upon an intricate balance of multiple microbial trophic groups, including hydrolyzing and fermenting bacteria, syntrophic acetogenic bacteria and methanogenic archaea [bib_ref] Diversity and dynamics of microbial communities in engineered environments and their implications..., Briones [/bib_ref]. Despite the relatively widespread adoption of anaerobic digestion, the variety of its contributing genetic and metabolic systems still remain relatively underexplored [bib_ref] Genome-centric resolution of microbial diversity, metabolism and interactions in anaerobic digestion, Vanwonterghem [/bib_ref] , owing to the vast phylogenetic and metabolic diversity [bib_ref] 454 pyrosequencing analyses of bacterial and archaeal richness in 21 full-scale biogas..., Sundberg [/bib_ref] and general non-culturability of a majority of the microorganisms present in digesters [bib_ref] Linking microbial community structure, interactions and function in anaerobic digesters using new..., Vanwonterghem [/bib_ref]. An improved understanding of the key ecological niches within anaerobic digesters and the metabolic and physiological attributes of microbial populations occupying those niches could help to facilitate new process designs and operational strategies that maximize biomethane recovery from organic waste.
Adding fats, oils and greases (FOG) as a cosubstrate during the anaerobic digestion of livestock manure or wastewater treatment solids has gained popularity to increase biomethane recovery due to the high biological methane potential of FOG [bib_ref] Increased biogas production at wastewater treatment plants through co-digestion of sewage sludge..., Luostarinen [/bib_ref] [bib_ref] Determining the limits of anaerobic co-digestion of thickened waste activated sludge with..., Wang [/bib_ref]. During anaerobic degradation, FOG is rapidly hydrolyzed into glycerol and long-chain fatty acids (LCFA) [bib_ref] Mechanism of inhibition caused by long-chain fatty acids in anaerobic digestion process, Hanaki [/bib_ref]. Because the degradation of LCFA in methanogenic environments relies on low hydrogen, formate and acetate levels for favorable thermodynamics, LCFA are converted into methane through an obligatory syntrophic partnership between acetogenic β-oxidizing bacteria and methanogenic archaea [bib_ref] Energetics of syntrophic cooperation in methanogenic degradation, Schink [/bib_ref] [bib_ref] Ecophysiology of syntrophic communities that degrade saturated and unsaturated long-chain fatty acids, Sousa [/bib_ref]. So far, only seven isolated bacterial species are capable of β-oxidizing LCFA (412 carbons in length) in syntrophy with methanogens, and all belong to the families of Syntrophomonadaceae and Syntrophaceae [bib_ref] Thermosyntropha tengcongensis sp. nov., a thermophilic bacterium that degrades long-chain fatty acids..., Zhang [/bib_ref]. However, the generally slow growth rates of syntrophic bacteria coupled with the toxicity of LCFA has limited the number of cultured representatives obtained [bib_ref] Identification and cultivation of anaerobic, syntrophic long-chain fatty acid-degrading microbes from mesophilic..., Hatamoto [/bib_ref].
Biomethane recovery from FOG can be limited by the reduction of microbial activity attributed to the toxicity of LCFAs at high concentrations [bib_ref] Mechanism of inhibition caused by long-chain fatty acids in anaerobic digestion process, Hanaki [/bib_ref] [bib_ref] Effects of free long-chain fatty acids on thermophilic anaerobic digestion, Angelidaki [/bib_ref] [bib_ref] Bactericidal effect of long chain fatty acids in anaerobic digestion, Rinzema [/bib_ref] [bib_ref] Co-digestion of grease trap sludge and sewage sludge, Davidsson [/bib_ref] [bib_ref] Increased biogas production at wastewater treatment plants through co-digestion of sewage sludge..., Luostarinen [/bib_ref] [bib_ref] Semi-continuous anaerobic co-digestion of thickened waste activated sludge and fat, oil and..., Wan [/bib_ref] [bib_ref] Determining the limits of anaerobic co-digestion of thickened waste activated sludge with..., Wang [/bib_ref]. Our recent work indicated that the maximum FOG loading rate to an anaerobic codigester was dependent on the LCFAdegrading population abundance and specific LCFA bioconversion kinetics [bib_ref] Microbial community adaptation influences long-chain fatty acid conversion during anaerobic codigestion of..., Ziels [/bib_ref] [bib_ref] Long-chain fatty acid feeding frequency in anaerobic codigestion impacts syntrophic community structure..., Ziels [/bib_ref]. The codigester LCFA feeding frequency was found to impact the composition and diversity of β-oxidizing Syntrophomonas species, with LCFA pulse feeding leading to higher bioconversion kinetics in comparison to semi-continuous LCFA feeding [bib_ref] Long-chain fatty acid feeding frequency in anaerobic codigestion impacts syntrophic community structure..., Ziels [/bib_ref]. However, those detected differences in the syntrophic community structure between codigesters were based on time-series 16S rRNA gene amplicon sequencing, and thus the impact of feeding frequency on the active fraction of the community remained unknown. A comparison of the active LCFA-degrading populations selected with different feeding frequencies would improve our understanding of relationships between operating parameters and physiological diversity within digester syntrophic populations.
Stable isotope probing (SIP) is a cultivationindependent method that can be used to elucidate links between microbial activity (function) and identity within environmental samples [bib_ref] Stable-isotope probing as a tool in microbial ecology, Radajewski [/bib_ref] [bib_ref] Stable isotope probinglinking microbial identity to function, Dumont [/bib_ref]. DNA-SIP relies on the incorporation of heavy isotopes (for example, 13 C) into microbial DNA during growth on labeled substrates, and thus acts as a 'filter' to enrich the DNA of active populations [bib_ref] When metagenomics meets stable-isotope probing: progress and perspectives, Chen [/bib_ref]. DNA-SIP has been combined with shotgun metagenomic sequencing approaches to discover novel enzymes and operons within microbial communities [bib_ref] Methanotrophic bacteria in oilsands tailings ponds of northern Alberta, Saidi-Mehrabad [/bib_ref] , and to identify new functional traits of microbial taxa [bib_ref] SIP metagenomics identifies uncultivated Methylophilaceae as dimethylsulphide degrading bacteria in soil and..., Eyice [/bib_ref]. Recently-developed differential coverage binning techniques for recovering population genomes from shotgun metagenomes [bib_ref] Genome sequences of rare, uncultured bacteria obtained by differential coverage binning of..., Albertsen [/bib_ref] [bib_ref] MetaBAT, an efficient tool for accurately reconstructing single genomes from complex microbial..., Kang [/bib_ref] [bib_ref] mmgenome: a toolbox for reproducible genome extraction from metagenomes, Karst [/bib_ref] could be amenable to DNA-SIP due to the expected DNA sequence composition divergence between heavy density fractions of 12 C-and 13 C-incubated samples [bib_ref] Targeted metagenomics of active microbial populations with stable-isotope probing, Coyotzi [/bib_ref]. However, there have been no reported studies coupling differential coverage binning with DNA-SIP to recover population genome bins from metagenome sequence data thus far.
The objective of this study was to apply DNA-SIP combined with shotgun metagenomic sequencing for a genome-centric comparison of the active oleate (C 18:1 LCFA) degrading populations in two parallel anaerobic codigesters operated with different oleate feeding frequencies. Differential coverage binning was employed to retrieve population genome bins from the DNA-SIP 13 C and 12 C-metagenomes. 16S rRNA gene amplicon sequencing was also conducted on 13 C and 12 C heavy DNA fractions to verify key populations involved in converting oleate into biomethane within the anaerobic codigesters.
# Materials and methods
## Anaerobic bioreactor operation
To examine the selective impacts of feeding frequency on the microbial community, two parallel anaerobic codigesters treating manure and oleate were operated at 35°C for over 200 days, as described previously [bib_ref] Long-chain fatty acid feeding frequency in anaerobic codigestion impacts syntrophic community structure..., Ziels [/bib_ref]. The digesters had a 4-liter working volume and a 20-day hydraulic and solids retention time (SRT), and were fed with blended manure feedstock collected from a nearby dairy farm. Manure was fed daily at a VS loading rate of 1.3 ± 0.2 g VS/l-d, and an aqueous solution of sodium oleate (497%, Tokyo Chemical Industry Co.) was added to the influent feed at 10% of the daily feed volume. The daily average oleate organic loading rate (OLR) to both codigesters was identical, but their oleate feeding frequency differed. One codigester received the oleate feed semi-continuously every 6-h, and is termed the continuous-fed (CF) codigester. The other codigester received the oleate feed in a pulse every 2 days, and is termed the pulse-fed (PF) codigester. At day 205 of operation, the oleate OLR to both codigesters was reduced from 3.2 to 1.8 g COD/l-d to promote stable digestion, as acetate was accumulating in the CF codigester effluent by day 200 [bib_ref] Long-chain fatty acid feeding frequency in anaerobic codigestion impacts syntrophic community structure..., Ziels [/bib_ref]. The resulting oleate concentration in the feed was 44 mM after day 205. From days 205-230, the pH of both codigesters averaged 7.5, and effluent VFA (acetate) was below 400 mg l − 1 . The codigesters were not fed between days 228-230 to reduce the level of any background substrates before commencing SIP incubations on day 230. At that time, total effluent VFA and LCFA (liquid+sorbed) levels were below 70 mg l − 1 .
Stable isotope probing incubations with anaerobic digester sludge Sludge was sampled directly from the PF and CF codigesters on day 230. Microcosms were established in 35 ml glass serum bottles pre-purged with N 2 :CO 2 (80:20), amended with 10 ml of codigester sludge (0.17 g VS total), sealed and capped with butyl rubber septa, and purged again with N 2 :CO 2 (80:20). All microcosms were maintained at 35°C, and stirred at 150 rpm on an orbital shaker. Microcosms were grouped into duplicate vials that were supplemented with either 12 C-sodium oleate or 13 Clabeled sodium oleate at an initial concentration of 8 mM. The 13 C-labeled sodium oleate was universally labeled at all 18 carbons (498% atom purity, Cambridge Isotope Laboratories, Tewksbury, MA, USA). Triplicate un-supplemented controls were incubated in parallel to measure background methane production. Gas production was measured based on the headspace pressure increase using a digital manometer (Series 490 A, Dwyer Instruments), and the headspace methane content was measured with a GC-FID (SRI 8610C) equipped with a Supelco Alumina Sulfate Plot column (50 m, 0.53 mm i.d.). Headspace pressure and composition were sampled about every 5 h using sterile syringes. Background methane production was subtracted from oleate-amended assays to measure the overall rate of substrate conversion.
After the first addition of oleate was converted at 495% efficiency into methane over 48 h , 12 C-and 13 C-labeled sodium oleate were resupplemented at 8 mM after purging the headspace with N 2 :CO 2 (80:20). Following another 48-h for oleate conversion into methane, a third supplement of 8 mM 12 C-and 13 C-labeled sodium oleate was added to the microcosms after purging the headspace with N 2 :CO 2 (80:20). The third incubation lasted 48 h, after which biomass from all bottles was sacrificed for DNA extraction. In total, 240 μmol of either 12 C-or 13 C-labeled oleate was added to the microcosms over the course of 6 days (144 h).
DNA extraction and density-gradient centrifugation Anaerobic codigester sludge samples were collected for DNA extraction at the start of the SIP incubation (time zero), as well as from each 12 C-oleate and 13 Coleate fed microcosm after 6 days of incubation. DNA was extracted and subjected to density-gradient centrifugation and fractionation, as described in detail in the Supplementary Information.
## Quantitative pcr
To further detect differences in DNA buoyant density (BD) values between the 12 C-and 13 C-incubated microcosms, quantitative PCR (qPCR) was conducted on DNA from gradient fractions with BDs ranging from 1.68 to 1.75 g ml − 1 . The qPCR targeted 16S rRNA gene fragments of the syntrophic LCFA β-oxidizing bacterial genus of Syntrophomonas using primers and probes developed by [bib_ref] Monitoring the dynamics of syntrophic β-oxidizing bacteria during anaerobic degradation of oleic..., Ziels [/bib_ref]. More detailed descriptions of the reaction Cumulative methane production (minus blank controls) for the microcosms fed with 12 C-and 13 C-labeled oleate over three repeated batch feeding periods. The number above each plot indicates the batch oleate feed round. The black dashed line shows the theoretical methane potential of the added oleate (25.5 ml CH 4 ; based on 2.9 g COD/g oleate, 8 mM concentration and 35°C temperature). The black solid line represents the predicted methane production based on non-linear model fitting with a modified Gompertz equation . Error bars represent the standard deviation of the biological replicates.
conditions, primers and probes, and calibration standards are given in the Supplementary Information.
16S rRNA gene amplicon sequencing DNA from 'heavy' gradient fractions with BDs ranging from 1.70 to 1.73 (4 fractions) were pooled at equal volumes and subjected to 16S rRNA gene amplicon sequencing for each microcosm . Briefly, 16S rRNA gene fragments were amplified with barcoded and indexed universal prokaryotic V4-V5 primers 515 F-Y (5′-GTGYCAGCMGCCGC GGTAA) and 926 R (5′-CCGYCAATTYMTTTRAGT TT) [bib_ref] Every base matters: assessing small subunit rRNA primers for marine microbiomes with..., Parada [/bib_ref] , and the products were pooled and sequenced on a 2x300 bp Illumina MiSeq run by the United States Department of Energy (DOE) Joint Genome Institute (JGI). Raw sequence data is available through the JGI Portal (http://jgi.doe.gov) under project ID number 1105527.
Amplicon reads were denoised into their exact sequences using the DADA2 pipeline v.1.2.1 [bib_ref] DADA2: High-resolution sample inference from Illumina amplicon data, Callahan [/bib_ref] before clustering into 99.5% OTUs with UPARSE v.8.1 [bib_ref] UPARSE: highly accurate OTU sequences from microbial amplicon reads, Edgar [/bib_ref] , as described in detail in the Supporting Information.
## Metagenomic sequencing, assembly and annotation
Pooled volumetric fractions of heavy DNA from density-gradient centrifugation , as well as codigester DNA samples from time zero, were used to generate an Illumina sequence library with an average insert size of 250 bp that was sequenced on an Illumina HiSeq-2500 with paired-end 150 bp reads at the DOE JGI. The metagenomic sequencing produced an average of 88 ± 11 M reads for SIP DNA samples and 140 ± 13 M reads for time zero DNA samples. Reads were trimmed and screened to remove adapters as well as filter sequences with a minimum Phred quality of 12, minimum length of 51 and no ambiguous bases using BBDuk v.36.02 (https://sourceforge.net/projects/bbmap/). Trimmed and screened paired-end reads from the codigester time zero metagenomes were individually assembled into contigs using MEGAHIT v.1.0.3 [bib_ref] MEGAHIT: an ultra-fast single-node solution for large and complex metagenomics assembly via..., Li [/bib_ref] with default parameters and the kmer list: 23, 43, 63, 83, 103, 123. The assembled contigs were annotated with the Integrated Microbial Genomes Metagenome (IMG/M) platform [bib_ref] IMG/M: integrated genome and metagenome comparative data analysis system, Chen [/bib_ref] by the DOE JGI, and KEGG annotation information (KO and EC numbers) were extracted. Raw sequence reads and assemblies are available through the .
## Metagenomic binning
Quality-filtered reads from all metagenomes, including the heavy gradient fractions of the 12 C-and 13 Coleate fed microcosms and from codigester time zero samples, were individually mapped to contigs from the time zero assemblies using BBMap v.36.02 (https://sourceforge.net/projects/bbmap/) in pairedend mode with default parameters. SAMtools v.1.4 [bib_ref] The Sequence Alignment/Map format and SAMtools, Li [/bib_ref] was used to sort and index the mapping files and extract contig coverage information across samples. The mapping data and coverage information were used to bin contigs into population genome bins (GBs) separately for each codigester sample set using MetaBat v.0.26.3 [bib_ref] MetaBAT, an efficient tool for accurately reconstructing single genomes from complex microbial..., Kang [/bib_ref] in 'superspecific' mode. CheckM v.1.0.6 was used to evaluate the level of bin completeness and contamination based on domain-level single-copy marker genes [bib_ref] CheckM: assessing the quality of microbial genomes recovered from isolates, single cells,..., Parks [/bib_ref]. Taxonomic classification of GBs was conducted by extracting a set of 107 conserved marker genes within open reading frames using mmgenome [bib_ref] mmgenome: a toolbox for reproducible genome extraction from metagenomes, Karst [/bib_ref] , running BLASTP on the extracted marker genes against the RefSeq v.52 protein database with a maximum e-value cutoff of 1e-5, and finding the least common ancestor (LCA) of the top 5 blast hits for each marker gene with MEGAN v.6.4 [bib_ref] MEGAN analysis of metagenomic data, Huson [/bib_ref]. The finest taxonomic ranking was assigned for which at least 70% of the LCA results agreed. Phylogenetic trees were generated for GBs that were over 40% complete using PhyloPhlAn v.1.3 based on 400 conserved marker genes [bib_ref] PhyloPhlAn is a new method for improved phylogenetic and taxonomic placement of..., Segata [/bib_ref] , as well as with UPGMA clustering of distance estimates from in silico DNA-DNA hybridization with the Total copies of Syntrophomonas 16S rRNA genes measured by qPCR for each density-gradient fraction recovered from isopycnic separation of DNA from 13 C-incubated microcosms and 12 C-controls for both anaerobic codigesters. The filled circles indicate gradient fractions that were pooled for subsequent 16S rRNA gene amplicon sequencing and metagenomic sequencing. Both biological replicate microcosms are shown, and each point represents an average of duplicate technical replicates.
Genome-to-Genome Distance Calculator v.2.1 [bib_ref] Digital DNA-DNA hybridization for microbial species delineation by means of genome-to-genome sequence..., Auch [/bib_ref].
Statistical analysis Statistically significant differences in raw read counts of OTUs (from 16S rRNA gene amplicon libraries) and GBs (from shotgun metagenome libraries) within the 12 C and 13 C DNA-SIP samples were detected with the negative binomial Wald test in DESeq2 v.1.14.1 separately for each codigester data set [bib_ref] Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2, Love [/bib_ref]. Raw sequence counts of GBs were obtained from parsing the mapping files. An adjusted P-value of 0.01 and a log 2 abundance increase greater than 2.0 (or 44-fold increase) were used to identify features (for example, OTUs or GBs) that significantly increased between the 12 C and 13 C DNA-SIP samples.
Regularized log-transformed sequence counts of 16S rRNA gene amplicon OTUs were determined with DESeq2 for use in principal component analysis. Permutational multivariate ANOVA (ADONIS) was conducted on Euclidian distance values from transformed sequence counts using the vegan package v.2.4.2 in R [bib_ref] The vegan package, Oksanen [/bib_ref].
# Results
Conversion of oleate into methane and enrichment of 13 C DNA Metabolism of 12 C-and 13 C-labeled oleate was confirmed with 97% and 95% COD recoveries as methane in the headspace for the PF codigester and CF codigester microcosms, respectively . Biokinetic modeling of methane production (described in showed that the maximum methane production rate increased with each subsequent batch feeding of 12 C-and 13 C-labeled oleate (Supplementary . The maximum methane production rate in the PF codigester microcosms was 26%, 21% and 17% higher than the CF codigester microcosms for the first, second and third batch incubation rounds, respectively (Supplementary .
Total DNA concentrations were measured in 24 density-gradient fractions to detect BD shifts after the consumption of 240 μmol 12 C-or 13 C-labeled oleate after 6 days. The heavy density fractions with BDs from 1.70 to 1.725 g ml − 1 contained 2.2-times and 1.6-times more DNA in the 13 C-incubated samples than in the 12 C-controls for the PF and CF codigester microcosms, respectively (Supplementary . To further confirm the enrichment of 13 C-labeled DNA, 16S rRNA genes of the LCFA-degrading bacterial genus Syntrophomonas were quantified in density fractions between 1.66 and 1.76 g ml − 1 . The four heavy density-gradient fractions with BDs ranging from 1.70 to 1.725 g ml − 1 contained over 12-times more Syntrophomonas 16S rRNA genes in the 13 C-incubated samples than that in the 12 C-controls for both codigesters .
The four heavy gradient fractions showing high levels of 13 C incorporation were therefore pooled for each microcosm for subsequent 16S rRNA gene amplicon sequencing and shotgun metagenomic sequencing .
16S rRNA gene amplicon sequencing of DNA-SIP samples The microbial community profiles of pooled heavy density-gradient fractions were assessed through paired-end 16S rRNA gene amplicon sequencing. After qualify-filtering, denoising, and merging paired-end amplicon reads, the number of sequences retained per sample averaged 164 000 ± 39 000. Principal component analysis (PCA) showed that community profiles of heavy density fractions from the 12 C-and 13 C-incubated microcosms were distinctly clustered for both codigesters (Supplementary [fig_ref] Figure 3: Relative fractions of the 11 most abundant genera in the heavy gradient... [/fig_ref]. The grouping of community profiles was significant within both codigester sample sets according to the oleate isotope (ADONIS, Po0.05). Additionally, samples within each isotope group were significantly grouped according to the codigester sludge source (ADONIS, Po0.05), indicating that the 13 C-enriched communities from each codigester biomass were distinct.
Differential abundance analysis was conducted to detect 16S rRNA gene OTUs that significantly increased in 13 C-incubated heavy gradient fractions relative to 12 C-controls. A total of 59 differentially abundant OTUs were identified in the PF codigester samples, 42 of which were classified to the Syntrophomonas genus . Within the CF codigester samples, a total of 40 differentially abundant OTUs were identified, with 36 classified as Syntrophomonas (Supplementary [fig_ref] Figure 5: Cumulative coverage values of KEGG ECs potentially involved in LCFA degradation [/fig_ref]. Of the total number of differentially abundant Syntrophomonas OTUs identified, 19 were shared between the codigesters while the remaining OTUs were unique to either codigester (Supplementary . About 48 and 35% of the total 16S rRNA gene counts of differentially abundant Syntrophomonas OTUs were attributed to unique OTUs within the 13 C-samples for the CF and PF codigesters, respectively . The relative fraction of Syntrophomonas 16S rRNA gene counts increased on average 2.6-and 2.9-times in the 13 C-incubated libraries relative to the 12 C-controls for the PF codigester and CF codigester, respectively [fig_ref] Figure 3: Relative fractions of the 11 most abundant genera in the heavy gradient... [/fig_ref]. Cumulative Syntrophomonas genus 16S rRNA gene counts increased on average 5.0-and 5.5-times in the 13 C-incubated samples relative to the 12 C-controls for the PF codigester and CF codigester, respectively (Supplementary . In contrast, the relative abundance of all other bacterial genera decreased in the 13 C 16S rRNA gene libraries relative to 12 C-controls for both codigesters, with the exception of the Ruminococcaceae NK4A214 group in the PF codigester [fig_ref] Figure 3: Relative fractions of the 11 most abundant genera in the heavy gradient... [/fig_ref]. Other differentially abundant OTUs were detected belonging to the genera: Thermovirga, Aminivibrio, Candidatus Cloacamonas, Anaerofustis, Syntrophothermus and Ruminococcaceae NK4A214 group, along with unclassified members of Planctomycetes, Spirochaetae, Synergistes, Actinobacteria and Bacteriodetes [fig_ref] Figure 5: Cumulative coverage values of KEGG ECs potentially involved in LCFA degradation [/fig_ref]. No methanogenic archaeal OTUs were identified as differentially abundant in either codigester sample set. Nonetheless, the relative abundance of Methanosaeta 16S rRNA genes increased from 2.2% in the 12 C-controls to 3.2% in the 13 C-samples for the CF codigester libraries [fig_ref] Figure 3: Relative fractions of the 11 most abundant genera in the heavy gradient... [/fig_ref]. Cumulative 16S rRNA gene counts of Methanosaeta increased by 1.7-and 2.7times in the 13 C-incubated samples relative to 12 Ccontrols for the PF codigester and CF codigester libraries, respectively, and Methanobacterium read counts increased by 1.7-and 1.9-times, respectively (Supplementary .
## Dna-sip metagenomic sequencing
Metagenomes from the PF and CF codigester sludges before SIP incubation (that is, time zero) were assembled to produce contigs for differential coverage binning. The PF and CF codigester time zero metagenome assemblies amounted to 536 947 079 bp and 680 484 169 bp within contigs longer than 1 kb, and had maximum contig lengths of 692 259 bp and 485 835 bp, respectively. Over 200 GBs were recovered through differential coverage binning for each codigester DNA-SIP sample set. Differential abundance analysis of GB read counts in the 13 Cincubated metagenomes relative to 12 C-controls using DESeq2 showed that 16 and 18 GBs were significantly enriched in the 13 C-metagenomes of the PF and CF codigesters, respectively (Supplementary . Due to the low completeness of some enriched GBs , a threshold of 40% genome completeness was established for further phylogenetic analysis. Assigned taxonomies and genomic characteristics of differentially abundant GBs with over 40% completeness are summarized in .
Approximately 75% of the differentially abundant GBs (6 of 8) in the PF codigester samples were taxonomically assigned to Syntrophomonas, along with 71% (5 of 7) of the differentially abundant GBs in the CF codigester . Both codigesters contained one differentially abundant GB that could only be assigned at the phylum-level to Firmicutes. In addition, one differentially abundant GB in the PF codigester was classified to the family-level of Syntrophomonadaceae, and one GB in the CF codigester was assigned to the phylum Candidatus Parcubacteria . Multi-gene alignments based on a set of 400 conserved marker genes showed that some differentially abundant GBs from both codigesters were closely related, while each codigester also contained a set of unique 13 C-enriched GBs . GBs that formed sister groups based on both the conserved marker gene alignments and UPGMA clustering of in silico DNA-DNA hybridization distance values included Syntrophomonas sp. PF07 and Syntrophomonas sp. CF14, as well as Syntrophomonas sp. PF15 and Syntrophomonas sp. CF22 . Syntrophomonas sp. PF103 was consistently clustered with the reference genomes of Syntrophomonas wolfei subsp. wofei and Syntrophomonas wolfei subsp. methylbutyrica . Firmicutes sp. CF263 was placed on a branch outside of Syntrophomonadaceae-affiliated genomes based on marker gene alignments , yet it clustered with the reference genome of Syntrophomonas zehnderi OL-4 based on in silico DNA-DNA hybridization (Supplementary . Both GBs of Firmicutes sp. PF36 and Candidatus Parcubacteria sp. CF250 were placed as outgroups in both phylogenetic clustering approaches , Supplementary .
## Continuous−fed codigester
Pulse−fed Codigester . In comparison to the CF codigester, the PF codigester had a greater number of GBs that were enriched above a log 2 change of 3.0 in the 13 C-incubated samples .
To assess the functional potential of differentially abundant GBs in 13 C-incubated samples, coverage values were extracted and summarized for KEGG ECs within the LCFA degradation pathway (KEGG Pathway Map 00071). Long-chain fatty acid CoA ligase (EC:6.2.1.3), which is responsible for activating free-fatty acids to acyl-CoA thioesters , had the highest EC coverage in differentially abundant GBs within 13 C-samples from both codigesters [fig_ref] Figure 5: Cumulative coverage values of KEGG ECs potentially involved in LCFA degradation [/fig_ref]. The EC categories with the next highest coverage were enoyl-CoA hydratase (EC:4.2.1.17) and acetyl-CoA acetyltransferase (EC:2.3.1.9). Genes encoding for the entire fatty acid β-oxidation cycle were identified within the differentially abundant GBs for both codigesters [fig_ref] Figure 5: Cumulative coverage values of KEGG ECs potentially involved in LCFA degradation [/fig_ref]. However, genes for acyl-CoA dehydrogenase (EC: 1.3.99.-) were observed in the PF codigester GBs but not in the CF codigester GBs [fig_ref] Figure 5: Cumulative coverage values of KEGG ECs potentially involved in LCFA degradation [/fig_ref]. The GBs in the PF codigester also had substantially higher gene coverage for 3-hydroxyacyl-CoA dehydrogenases (EC:1.1.1.35 and EC:1.1.1.36) in 13 C-incubated samples than the CF codigester GBs [fig_ref] Figure 5: Cumulative coverage values of KEGG ECs potentially involved in LCFA degradation [/fig_ref]. Additionally, genes encoding for Δ 2 ,Δ 4 -dienoyl-CoA reductase (EC:1.3.1.34) involved in the metabolism of unsaturated fatty acids [bib_ref] An alternative pathway of oleate β-oxidation in Escherichia coli involving the hydrolysis..., Ren [/bib_ref] were detected in the differentially abundant GBs for both codigesters, and had a coverage that was an order of magnitude higher in the PF codigester GBs (Supplementary .
Syntrophic conversion of LCFA under methanogenic conditions involves interspecies electron transfer between Bacteria and Archaea, which commonly involves hydrogen and/or formate as electron shuttles [bib_ref] Electron transfer in syntrophic communities of anaerobic bacteria and archaea, Stams [/bib_ref]. Several formate dehydrogenases and hydrogenases were detected in the enriched GBs of both codigesters (Supplementary . Formate dehydrogenases (EC:1.2.1.2, EC:1.2.1.43) were more abundant than hydrogenases (EC:1.12.1.2, EC:1.12.7.2, EC:1.12.1.3), and had similar coverage in the differentially abundant GBs from both codigesters (Supplementary . In contrast, hydrogenases (EC:1.12.1.2, EC:1.12.1.3) had lower coverage in the CF codigester GBs relative to the PF codigester (Supplementary .
# Discussion
This research conducted a genome-centric comparison of the active populations between two anaerobic codigesters that were either pulse-fed or continuously-fed with oleate. DNA-SIP was combined with shotgun metagenomic sequencing and differential coverage binning to provide novel insights into active LCFA-degrading populations, and highlighted the potential for physiological Taxonomic classification and characteristics for genome bins that were significantly enriched (Po0.01, log 2 increase42.0) in 13 Cmetagenomes relative to 12 C-controls for both codigester sample sets Based on CheckM output.
DNA-SIP metagenomics of syntrophic populations RM Ziels et al diversity at the trophic level of syntrophy within anaerobic digesters. As the average specific growth rate (μ) in the codigesters was 0.05 d − 1 based on the 20-d SRT, higher growth rates would have been needed for cell duplication and sufficient DNA labeling during the 6-d incubation time, which could have been promoted by batch feeding oleate at initial concentrations of 8 mM in the SIP incubations. A growth rate of 0.11 d − 1 would be required for cell doubling within a Phylogenetic overview of differentially abundant genome bins identified in both anaerobic codigester DNA-SIP metagenomes. The tree was constructed based on a concatenated alignment of conserved marker genes within the metagenomic contigs of each bin using PhyloPhlAn v.1.3 [bib_ref] PhyloPhlAn is a new method for improved phylogenetic and taxonomic placement of..., Segata [/bib_ref]. The color of each genome bin node represents the codigester biomass source, and the height of the outer bars represent the genome bin coverage in the heavy gradient fractions of the 13 C-incubated samples and 12 C-controls. The tree was illustrated using GraPhLan v.0.9.7 [bib_ref] Compact graphical representation of phylogenetic data and metadata with GraPhlAn, Asnicar [/bib_ref]. 6-d period, which is within the range of maximum specific growth rates (μ max ) for co-cultures of syntrophic fatty acid oxidizing bacteria and methanogenic archaea of 0.10 to 0.19 d − 1 [bib_ref] Role of syntrophic microbial communities in high-rate methanogenic bioreactors, Stams [/bib_ref]. While a longer incubation time would have contributed to further DNA labeling, longer incubation times in DNA-SIP can be problematic because they can lead to cross-labeling of peripheral populations due to endogenous cell decay [bib_ref] Methodological considerations for the use of stable isotope probing in, Neufeld [/bib_ref]. In contrast to our study, previous studies employing DNA-SIP for syntrophic propionate-and butyratedegrading communities in soil utilized incubation times ranging from 17 to 69 days [bib_ref] Stableisotope probing of microorganisms thriving at thermodynamic limits: syntrophic propionate oxidation in..., Lueders [/bib_ref] [bib_ref] Fatty acid-oxidizing consortia along a nutrient gradient in the Florida Everglades, Chauhan [/bib_ref] [bib_ref] Syntrophomonadaceaeaffiliated species as active butyrate-utilizing syntrophs in paddy field soil, Liu [/bib_ref] [bib_ref] Syntrophic oxidation of propionate in rice field soil at 15 and 30°C..., Gan [/bib_ref] [bib_ref] Direct interspecies electron transfer accelerates syntrophic oxidation of butyrate in paddy soil..., Li [/bib_ref]. These longer incubation times may have been necessitated due to the low active biomass fractions in soil, which would result in low substrate conversion rates. The higher active syntrophic biomass fraction in the anaerobic codigester sludges in this study promoted more rapid oleate bioconversion in the microcosms, which permitted efficient DNA-SIP labeling within a 6-day incubation period. DNA-SIP based metagenomics could thus be potentially applied with relatively short SIP incubation times to study other biological wastewater treatment process that have relatively high active biomass fractions.
Syntrophic fatty acid conversion requires that free energy from catabolism is shared between syntrophic bacteria and methanogenic archaeal partners [bib_ref] Energetics of syntrophic cooperation in methanogenic degradation, Schink [/bib_ref] [bib_ref] Electron transfer in syntrophic communities of anaerobic bacteria and archaea, Stams [/bib_ref]. In this study, all of the 16S rRNA gene OTUs that were differentially abundant in 13 C-incubated DNA-SIP samples for both codigesters belonged to the domain Bacteria, while no Archaea OTUs were differentially abundant in 13 C-incubated samples. Likewise, all of the differentially abundant GBs recovered from the DNA-SIP shotgun metagenomes were assigned to bacterial taxa. However, the higher cumulative 16S rRNA gene read counts of Methanosaeta and Methanobacterium in 13 C-incubated samples relative to 12 C-controls indicates that they were likely methanogenic partners involved in syntrophic oleate degradation in both codigesters, but the increases in their OTU abundances were not statistically significant. The relatively high level of 12 C-background substrates for methanogenesis (from the digester inoculum) likely diluted the 13 C-labeled DNA of methanogenic archaea, as was also observed with protein-SIP in anaerobic digester communities [bib_ref] Identification of syntrophic acetate-oxidizing bacteria in anaerobic digesters by combined protein-based stable..., Mosbaek [/bib_ref].
To the best of our knowledge, this is the first study to couple DNA-SIP with differential coverage binning to recover population GBs from environmental metagenomes. Differential abundance analysis of GB read counts was an effective technique coupled to DNA-SIP metagenomic binning to identify GBs that were significantly enriched within the heavy gradient fractions of 13 C-incubated samples versus 12 Ccontrols. The taxonomic assignment of differentially abundant GBs agreed with the results of 16S rRNA gene amplicon sequencing, as over 70% of the GBs belonged to the genus Syntrophomonas. This finding corroborates previous indications that Syntrophomonas are key syntrophic LCFA-degrading bacteria in mesophilic anaerobic digesters [bib_ref] Microbial diversity of mesophilic methanogenic consortium that can degrade longchain fatty acids..., Shigematsu [/bib_ref] [bib_ref] Identification and cultivation of anaerobic, syntrophic long-chain fatty acid-degrading microbes from mesophilic..., Hatamoto [/bib_ref] [bib_ref] Molecular assessment of complex microbial communities degrading long chain fatty acids in..., Sousa [/bib_ref] [bib_ref] Ecophysiology of syntrophic communities that degrade saturated and unsaturated long-chain fatty acids, Sousa [/bib_ref] [bib_ref] Monitoring the dynamics of syntrophic β-oxidizing bacteria during anaerobic degradation of oleic..., Ziels [/bib_ref]. However, this is the first study to determine the relative contribution of Syntrophomonas toward LCFA degradation using DNA-SIP. The finding that some phylogenetically distinct Syntrophomonas GBs and 16S rRNA gene OTUs were unique to each codigester further suggests that microbial communities involved in LCFA conversion can be physiologically diverse, and that the LCFA feeding frequency selected for unique active LCFA-degrading syntrophic taxa. Syntrophic bacteria have traditionally been considered metabolicspecialists [bib_ref] Syntrophy in anaerobic global carbon cycles, Mcinerney [/bib_ref] [bib_ref] Genomic insights into syntrophy: the paradigm for anaerobic metabolic cooperation, Sieber [/bib_ref] , which has led to the general conclusion that they have low functional redundancy in anaerobic digesters [bib_ref] Bacterial community structures are unique and resilient in full-scale bioenergy systems, Werner [/bib_ref] [bib_ref] Microbial management of anaerobic digestion: exploiting the microbiome-functionality nexus, Carballa [/bib_ref]. A previous genome-centric analysis of anaerobic digester metagenomes found low diversity within the trophic level of syntrophic acetogenesis [bib_ref] Genome-centric resolution of microbial diversity, metabolism and interactions in anaerobic digestion, Vanwonterghem [/bib_ref]. Using transcriptomic read mapping, [bib_ref] Untangling the effect of fatty acid addition at species level revealed different..., Treu [/bib_ref] could only identify two active Syntrophomonas GBs in an anaerobic digester after pulsing oleate. In this study, 12 active Syntrophomonas GBs were identified using targeted DNA-SIP metagenomic analysis. DNA-SIP based metagenomic binning may thus improve our resolution of anaerobic digester trophic groups by enriching metagenome libraries with actively growing population genomes [bib_ref] Targeted metagenomics of active microbial populations with stable-isotope probing, Coyotzi [/bib_ref] , which may be especially useful to study low-abundance groups like syntrophic bacteria [bib_ref] Role of syntrophic microbial communities in high-rate methanogenic bioreactors, Stams [/bib_ref]. The recent recovery of 236 microbial GBs from anaerobic digester metagenomes led to conclusion that Syntrophomonas were among a core essential microbial group that is present independent of digester operational conditions [bib_ref] Deeper insight into the structure of the anaerobic digestion microbial community; the..., Treu [/bib_ref]. Our finding that the anaerobic digester feeding frequency can impact the genomic composition of Syntrophomonas thus has implications for developing operating strategies that steer the core digester community to obtain better biokinetics and process stability.
In prior research, it was observed that the frequency of LCFA feeding led to different biokinetics between the PF and CF codigesters, which were correlated with individual Syntrophomonas taxa abundances between the two systems [bib_ref] Long-chain fatty acid feeding frequency in anaerobic codigestion impacts syntrophic community structure..., Ziels [/bib_ref]. In this study, biokinetic modeling of methane production in the DNA-SIP microcosms fed with either 13 C-or 12 C-oleate showed that the PF codigester sludge had higher maximum oleate conversion rates than CF codigester sludge during all three DNA-SIP batch degradation periods. This result further corroborates that biological selection from the LCFA feeding frequency altered LCFA bioconversion kinetics in the anaerobic codigesters [bib_ref] Long-chain fatty acid feeding frequency in anaerobic codigestion impacts syntrophic community structure..., Ziels [/bib_ref]. The different oleate bioconversion kinetics observed between the two codigesters could have been attributed to differences in the functional gene repertoires of their active LCFAdegrading populations. The differentially abundant GBs in the PF codigester had more redundant KEGG ECs (for example, multiple ECs for a single reaction step) in the LCFA degradation pathway compared to the CF codigester GBs, which could indicate that pulse feeding led to a higher level of metabolic flexibility. The PF codigester GBs also had a greater distribution of formate dehydrogenase and hydrogenase encoding genes, which have been suggested to afford syntrophs with metabolic flexibility in fluctuating environmental conditions [bib_ref] Flexibility of syntrophic enzyme systems in desulfovibrio species ensures their adaptation capability..., Meyer [/bib_ref] [bib_ref] The importance of hydrogen and formate transfer for syntrophic fatty, aromatic and..., Sieber [/bib_ref]. The greater metabolic flexibility identified in the 13 C-enriched GBs of the PF codigester in this study suggests that pulse feeding could promote a greater physiological adaptive capacity by supporting a higher level of functional redundancy within the ecological niches of anaerobic digesters, which in turn may have supported its higher oleate loading tolerance [bib_ref] Long-chain fatty acid feeding frequency in anaerobic codigestion impacts syntrophic community structure..., Ziels [/bib_ref]. A greater tolerance to overloading and toxicity has also been observed in other pulse-fed anaerobic digesters compared to continuous-fed systems [bib_ref] Growth kinetics and competition between methanosarcina and methanosaeta in mesophilic anaerobic digestion, Conklin [/bib_ref] [bib_ref] Repeated pulse feeding induces functional stability in anaerobic digestion, De Vrieze [/bib_ref] [bib_ref] Changing feeding regimes to demonstrate flexible biogas production: effects on process performance,..., Mulat [/bib_ref]. Although stochastic processes cannot be ruled out because the bioreactors were not replicated, the fact that these bioreactors were inoculated with the same seed source and Syntrophomonas was the predominant group impacted by oleate addition for over 230 days [bib_ref] Long-chain fatty acid feeding frequency in anaerobic codigestion impacts syntrophic community structure..., Ziels [/bib_ref] indicates that the feeding frequency likely imparted a deterministic selective pressure on the syntrophic community structure.
In summary, this study demonstrated the application of DNA-SIP combined with metagenomics and differential coverage binning to identify active populations in an engineered biological treatment process. The results indicated that the operating conditions, specifically the feeding frequency, of the methanogenic bioreactors impacted the genomic composition of active syntrophic populations. The feeding frequency also impacted the level of functional gene redundancy within the active populations, which had implications for increased physiological adaptation using a pulse feeding strategy. Thus, DNA-SIP metagenomics is a powerful tool to help understand how operating strategies may improve the performance of biological wastewater treatment processes, as was shown here for increased methane recovery from high-strength organic wastes.
[fig] Figure 3: Relative fractions of the 11 most abundant genera in the heavy gradient fractions of 13 C-incubated samples and 12 C-controls for sample sets from both codigesters, based on 16S rRNA gene amplicon sequencing. Relative fractions were determined using DESeq2 normalized read counts for all OTUs(Love et al., 2014), and were then aggregated at the genus level. Both biological replicate microcosms are shown.The largest GB coverage fold-change in the 13 Cincubated samples was observed for the sister group of Syntrophomonas sp. PF15 & Syntrophomonas sp. CF22, which were not closely related to previouslysequenced Syntrophomonas isolates(Figure 4; Supplementary Table 2). Other GBs that had high fold-changes in the 13 C-incubated samples (log 2 change 43.0) included Syntrophomonas sp. PF75, Syntrophomonas sp. PF62, Syntrophomonas sp. PF103, Firmicutes sp. PF36, Syntrophomonas sp. PF111, and Syntrophomonas sp. CF196 (Figure 4; Supplementary [/fig]
[fig] Figure 5: Cumulative coverage values of KEGG ECs potentially involved in LCFA degradation (KEGG map 00071), based on the coverage of all differentially abundant genome bins in the 13 C-incubated DNA-SIP metagenomes. Values from duplicate biological replicates are shown for both codigesters, and the size of each marker is proportional to the log 10 of the EC read coverage. [/fig]
|
Tuning multiple imputation by predictive mean matching and local residual draws
Background: Multiple imputation is a commonly used method for handling incomplete covariates as it can provide valid inference when data are missing at random. This depends on being able to correctly specify the parametric model used to impute missing values, which may be difficult in many realistic settings. Imputation by predictive mean matching (PMM) borrows an observed value from a donor with a similar predictive mean; imputation by local residual draws (LRD) instead borrows the donor's residual. Both methods relax some assumptions of parametric imputation, promising greater robustness when the imputation model is misspecified.Methods:We review development of PMM and LRD and outline the various forms available, and aim to clarify some choices about how and when they should be used. We compare performance to fully parametric imputation in simulation studies, first when the imputation model is correctly specified and then when it is misspecified.Results:In using PMM or LRD we strongly caution against using a single donor, the default value in some implementations, and instead advocate sampling from a pool of around 10 donors. We also clarify which matching metric is best. Among the current MI software there are several poor implementations.Conclusions: PMM and LRD may have a role for imputing covariates (i) which are not strongly associated with outcome, and (ii) when the imputation model is thought to be slightly but not grossly misspecified. Researchers should spend efforts on specifying the imputation model correctly, rather than expecting predictive mean matching or local residual draws to do the work.
# Background
The presence of missing data is a common issue in medical research, leading to reduced precision and sometimes bias in parameter estimates. Multiple imputation (MI) can alleviate these issues and is popular approach to dealing with missing data [bib_ref] Multiple imputation: review of theory, implementation and software, Harel [/bib_ref] [bib_ref] Much ado about nothing: A comparison of missing data methods and software..., Horton [/bib_ref] [bib_ref] Multiple imputation using chained equations: Issues and guidance for practice, White [/bib_ref].
It is impossible to know for certain how data went missing. In thinking about the process there are three important scenarios [bib_ref] Inference and missing data, Rubin [/bib_ref] :
1. Missing completely at random (MCAR). The probability of data being missing does not depend on observed or unobserved data. 2. Missing at random (MAR). Conditional on observed data, the probability of data being missing does not depend on unobserved data. MCAR is a special case of MAR. 3. Missing not at random (MNAR). Conditional on observed data, the probability of data being missing still depends on unobserved data.
Researchers analysing incomplete datasets should consider the process by which data may have gone missing, and perform analyses that are valid given this assumption.
MI involves specifying a parametric model for the missing data given the observed data and drawing missing values from the posterior predictive distribution M > 1 times. This model is henceforth referred to as the imputation model. The M filled-in datasets are analysed identically according to the model that would have been used in the absence of missing data. We term this model http://www.biomedcentral.com/1471-2288/14/75 the analysis model. The M parameter estimates are then combined using 'Rubin's rules'.
Multiple imputation can provide valid inference given any of the above mechanisms, although standard software implementations impute assuming MAR (MCAR) by default.
If the imputation model is specified correctly, Rubin's rules lead to consistent parameter estimation and confidence intervals that fully incorporate uncertainty due to missing data [bib_ref] Multiple imputation: a primer, Schafer [/bib_ref]. For imputing a covariate it is advisable to include in the imputation model (i) variables thought to predict missingness, (ii) variables associated with the variable being imputed, and (iii) the outcome variable of the analysis model [bib_ref] Multiple imputation using chained equations: Issues and guidance for practice, White [/bib_ref] [bib_ref] Using the outcome for imputation of missing predictor values was preferred, Moons [/bib_ref].
One of the biggest challenges for users of MI is specifying the imputation model correctly. This is not always easy to do, even for seemingly simple analyses: for instance when the analysis model contains nonlinear functions of incomplete covariates [bib_ref] Multiple imputation of missing covariates with non-linear effects and interactions: an evaluation..., Seaman [/bib_ref].
Predictive mean matching (PMM) [bib_ref] Missing-data adjustments in large surveys, Little [/bib_ref] and local residual draws (LRD) [bib_ref] Alternative methods for CPS income imputation, David [/bib_ref] are methods for drawing imputations that relax some of the assumptions of parametric imputation. In doing so they may improve robustness of inference with missing data to misspecification of the imputation model. These methods are outlined briefly below and described further in the Methods section.
For an incomplete variable x, an imputation model is fitted with parameters α and covariates z. Parametric imputation proceeds by drawing α from its posterior distribution, before drawing missing values of x from the posterior predictive distribution conditional on the draw α * . The draws of the imputation model parameters make parametric imputation 'proper' [bib_ref] Multiple imputation: a primer, Schafer [/bib_ref] and may be taken parametrically or by the approximate Bayesian bootstrap [bib_ref] Multiple imputation for interval estimation from simple random samples with ignorable nonresponse, Rubin [/bib_ref].
PMM and LRD differ from parametric imputation as follows. Let h index observations with x observed and j index observations with x missing. For all h, the linear predictor α obs z h is calculated, and for all j, the linear predictor α mis z j is calculated (α obs and α mis will be defined in the Methods section). Observed values close to the linearpredicted value are selected as the donor pool. Often, but not always, the donor pool is fixed as containing k candidate donors. One of these is selected at random to 'donate'. PMM imputes the donor's x h . LRD adds the donor's residual to the recipient's linear predictor.
In the remainder of this article, we give technical details of these methods reviewing their development and the various forms available, along with the rationale for their use. Two simulation studies on PMM and LRD are then described and reported: in the first, the imputation model is correct; in the second, the imputation model is misspecified. We illustrate various approaches to imputing a missing covariate for a cohort study in ovarian cancer. We finish with a discussion and some conclusions. This article describes the rationale for PMM and LRD, and their development and evaluation in previous work. They are evaluated further in some simple and then more challenging settings. Our focus is on incomplete continuous covariates, though in principle both methods may be used to impute ordinal or categorical covariates. We aim to clarify some choices about how PMM and LRD should be implemented and when they should be used.
# Methods
## The development of predictive mean matching and local residual draws
In this section, we provide a technical description of PMM and LRD, review the development of the various flavours available -of which there are several -and clarify some details. summarises software implementations of PMM and LRD, as of February 2014, and provides some details on options for changing the default values, if available.
Both PMM and LRD begin by calculating a predictive distance δ hj , which can be thought of as a measure of match quality. For all j the k observations minimising |δ hj | are identified where
[formula] δ hj = α mis z j − α obs z h ,( 1 ) [/formula]
and one of these is selected at random. For PMM [bib_ref] Missing-data adjustments in large surveys, Little [/bib_ref] the imputed value x * j is taken as x h . For LRD [bib_ref] Partially parametric techniques for multiple imputation, Schenker [/bib_ref] the imputed value x * j is
[formula] x * j = α mis z j + x h − α obs z h .( 2 ) [/formula]
## Defining the matching distance
Little initially introduced PMM, suggesting the calculation of δ hj such that α mis = α obs =α [bib_ref] Missing-data adjustments in large surveys, Little [/bib_ref]. In the same article, it was noted that this did not allow for uncertainty about α: in parametric imputation a draw α * is taken before imputing x * j conditional on α * . The use of α mis = α * was noted as a remedy. A third metric was introduced by Heitjan and Little where α mis = α obs = α * [bib_ref] Multiple imputation for the fatal accident reporting system, Heitjan [/bib_ref].
We refer to these distance measures as follows:
[formula] Type 0 matching δ hj =αz j −αz h(3) [/formula]
Type 1 matching
[formula] δ hj = α * z j −αz h(4) [/formula]
Type 2 matching
[formula] δ hj = α * z j − α * z h(5) [/formula]
The designation is mnemonic according to the number of * symbols appearing on the right hand side, and types 1 and 2 correspond to the designation used by the ice command in Stata [bib_ref] Multiple imputation of missing values: update, Royston [/bib_ref] and the aregimpute function of the R package Hmisc. Note that with a single incomplete variable δ hj type 0 and type 2 are the same. It is often difficult to determine the type of matching being used in previous work. Type 0 matching was used by David et al. [bib_ref] Alternative methods for CPS income imputation, David [/bib_ref] and Little [bib_ref] Missing-data adjustments in large surveys, Little [/bib_ref] , and was compared to type 2 by Schenker and Taylor [bib_ref] Partially parametric techniques for multiple imputation, Schenker [/bib_ref]. Type 1 matching was described by Little [bib_ref] Missing-data adjustments in large surveys, Little [/bib_ref] , and White, Royston and Wood [bib_ref] Multiple imputation using chained equations: Issues and guidance for practice, White [/bib_ref]. Type 2 matching has been used comparatively more (see for example [bib_ref] Partially parametric techniques for multiple imputation, Schenker [/bib_ref] [bib_ref] Multiple imputation for the fatal accident reporting system, Heitjan [/bib_ref] [bib_ref] Assessing secular trends in blood pressure: a multiple-imputation approach, Heitjan [/bib_ref] [bib_ref] Multiple imputation in public health research, Zhou [/bib_ref] [bib_ref] Multiple imputation in practice: comparison of software packages for regression models with..., Horton [/bib_ref] [bib_ref] A comparison of imputation methods in a longitudinal randomized clinical trial, Tang [/bib_ref] [bib_ref] Survival analysis using auxiliary variables via non-parametric multiple imputation, Hsu [/bib_ref] [bib_ref] Multiple imputation techniques in small sample clinical trials, Barnes [/bib_ref] [bib_ref] A comparison of multiple imputation and fully augmented weighted estimators for Cox..., Qi [/bib_ref].
## Defining the donor pool
There are three broad approaches to defining the donor pool. The first is to use a fixed number of donors k; the second is to define some δ max so that any h for whom |δ hj | < δ max are in the donor pool for j. This is sometimes termed 'caliper matching'. A third approach uses k = n h , the number of observations for which x is observed, but is more likely to select those with small d hj [bib_ref] Multiple imputation using an iterative hot-deck with distance-based donor selection, Siddique [/bib_ref] [bib_ref] MIDAS: a SAS macro for multiple imputation using distance-aided selection of donors, Siddique [/bib_ref] ; see the next section.
David et al. imputed income, initially using global residual draws [bib_ref] Alternative methods for CPS income imputation, David [/bib_ref] , setting k to the number of observations with x observed. However, the results were unsatisfactory to the authors and so δ max = $2, 000 was instead used.
The notion of selecting from a pool of potential donors was apparently not present in the work of Little [bib_ref] Missing-data adjustments in large surveys, Little [/bib_ref] , who matched to the nearest donor only. Heitjan and Little introduced a pool of k = 5 potential donors [bib_ref] Multiple imputation for the fatal accident reporting system, Heitjan [/bib_ref] ; subsequent to that article authors have largely used fixed k > 1.
Schenker and Taylor noted the problem with defining δ max , that it is possible for a recipient to have no donors with α obs z h lying within α mis z j ± δ max . They suggested an adaptive method for choosing k, which involved defining δ max , but if k = 0 or 1 to set k = 2.
## Sampling from the donor pool
The most common method is to randomly sample an observation from the donor pool, for example [bib_ref] Much ado about nothing: A comparison of missing data methods and software..., Horton [/bib_ref] [bib_ref] Partially parametric techniques for multiple imputation, Schenker [/bib_ref] [bib_ref] Multiple imputation for the fatal accident reporting system, Heitjan [/bib_ref] [bib_ref] Multiple imputation in public health research, Zhou [/bib_ref] , however some more sophisticated methods have also been proposed.
Moriarity and Scheuren suggested the use of 'constrained' matching [bib_ref] A note on rubin's statistical matching using file concatenation with adjusted weights..., Moriarity [/bib_ref] , where each h can only donate x h once. Note that this is only feasible with less that half of values missing. An alternative, 'slightly constrained' matching, penalises any h that has already donated by reducing the probability of subsequent donation. Durrant and Skinner used a slightly constrained matching in a simulation study, and found it to be less biased than using a fixed value of k [bib_ref] Using missing data methods to correct for measurement error in a distribution..., Durrant [/bib_ref].
Siddique and Belin proposed a version of PMM that allows any h to donate [bib_ref] Multiple imputation using an iterative hot-deck with distance-based donor selection, Siddique [/bib_ref] , but with the probability of imputing x h for individual j proportional to a function of |δ hj |. A 'closeness' parameter was introduced which could be altered to augment the probability of selecting the closest donors. This was later published as a SAS macro [bib_ref] MIDAS: a SAS macro for multiple imputation using distance-aided selection of donors, Siddique [/bib_ref].
## Notes on lrd
LRD has received far less attention than PMM. This is possibly because of the attraction that, by always borrowing observed values, PMM always imputes observable values, while LRD may not. Conversely, LRD does have the ability to impute values outside the range of observed data, and so may deal better with values that are missing in tails of a distribution.
For LRD there is a second metric to consider, unnoticed in the literature. We note the following imputation types, named correspondingly to match types:
[formula] Type 0 imputation x * j =αz j + (x h −αz h ) Type 1 imputation x * j = α * z j + (x h −αz h ) Type 2 imputation x * j = α * z j + (x h − α * z h ). [/formula]
With parametric imputation, x * j are drawn from a distribution centred at α * z j . Of the above imputation metrics, only type 1 achieves this, while types 0 and 2 draw from a distribution centred atαz j . Schenker and Taylor [bib_ref] Partially parametric techniques for multiple imputation, Schenker [/bib_ref] , and Barnes et al. [bib_ref] Multiple imputation techniques in small sample clinical trials, Barnes [/bib_ref] are unclear as to the imputation type used in their work.
## Rationale for pmm and lrd
Use of PMM and LRD is typically motivated by the notion that they provide a degree of robustness when the imputation model is misspecified, for example if the normality assumption is in question, residuals are heteroscedastic, or associations are non linear. [fig_ref] Figure 1: Bivariate plots are of x vs [/fig_ref] demonstrates how PMM and LRD may guard against these problems in 150 simulated observations, of which 50 are missing x, which is imputed once. The top panels show a dataset with skewed residuals, the middle panels show a dataset exhibiting heteroscedasticity, and the bottom panels show a quadratic relationship. Missing values are MCAR and imputed once by parametric draws (left panels), PMM (centre panels, type 1 matching with k = 3) and LRD (right panels, type 1 matching with k = 3).
Because the data are MCAR, the missing values are a random sample of the observed values; imputed values should thus bear a close resemblance to the observed. With non-normal residuals, parametric imputation does a poor job of preserving the bivariate distribution of y and x, while PMM and LRD do a better job. In the middle panels, parametric imputation again imputes one or two values that do not match the distribution of the observed data well, while PMM borrows from the individual with the lowest observed value of x five times. The most stark illustration of the difference between methods is given in the lower panels, where parametric imputation seems to do a very poor job of preserving the association in observed data but PMM and LRD do well by contrast. http://www.biomedcentral.com/1471-2288/14/75
## Some settings where pmm and lrd may fail
While PMM and LRD are generally advocated as methods to improve the imputation model, there are also potential weaknesses.
The price to pay for the additional flexibility supplied by PMM and LRD is that x * j are not formally draws from the posterior predictive distribution of the imputation model; there is thus no guarantee that Rubin's rules will be appropriate for inference.
The main specific concerns about PMM are around donor sparseness: when there are few donors with a predictive mean close to the predictive mean of a missing observation. It is clear that when |δ hj | is large, matches are of poor quality and so imputed values may be inappropriate. This may occur are when there are few observations with x observed, and under departures from MCAR.
A second pitfall for PMM arises when δ hj has the same sign for all h in the donor pool for j, which will introduce a bias in the imputed values, with consequences for estimation. Again, LRD does not necessarily suffer this bias provided the direction and magnitude of residuals are appropriate.
## Simulation studies
Two simulation studies are reported below. The first compares various forms of PMM and LRD in a setting ideally suited to parametric imputation. The second compares them in a setting where parametric imputation is likely to fail. Both studies aim to evaluate type 1 versus type 2 matching, and to comment on appropriate choices of k.
## Simulation design: correctly specified imputation model
In the first study, we simulate 500 observations on two variables y and x where x ∼ N(0, 1) and y|x is normal in the complete data. The analysis model of interest is a linear regression y i ∼ N(β 0 + βx i , 100). http://www.biomedcentral.com/1471-2288/14/75
Three different strengths of y-x association are simulated: β = 0, β = 3.33 and β = 10, corresponding to R 2 values of 0, 0.99 and 0.5 respectively.
Throughout, y is complete and x is incomplete. Three missingness mechanisms are invoked: MCAR, and two different MAR mechanisms. Let π denote the probability that x is missing. Under MCAR, π = 0.25. The MAR mechanisms are simulated via the linear logistic model logit(π ) = γ 0 +γ 1 y i , such that observations with large values of y are more likely to have values of x missing. Let R be a binary variable indicating whether x is not missing or missing. Values of γ 0 and γ 1 were chosen such that 25% of observations are missing and comparison of R with y returns an area under the ROC curve of 0.65 ('weak' MAR) and 0.75 ('strong' MAR).
The imputation model is
[formula] x h ∼ N(α 0 + α 1 y h , σ 2 ),( 6 ) [/formula]
which is correctly specified. M = 10 imputations [bib_ref] Multiple imputation: a primer, Schafer [/bib_ref] are used for each of the following methods:
- Parametric imputation using posterior draws.
- PMM with type 1 and type 2 matching and, for each match type, k = 1, 3, 5 and 10. - LRD with type 1 and type 2 matching (type 1 imputation throughout), for each match type k = 1, 3, 5, 10 and 20 (20 comes from the expectation that LRD will suffer less than PMM with larger donor pools).
The imputed datasets are analysed and estimates combined using Rubin's rules. All imputations were produced using the ice command in Stata [bib_ref] Multiple imputation of missing values: update, Royston [/bib_ref]. The various MI methods are compared to analysis of the complete data, a gold standard, and analysis of the complete cases, which any imputation method must improve upon to be worthwhile.
The whole simulation process is repeated 1,000 times. Bias, coverage of confidence intervals, and a measure of (in-)efficiency, the standard deviation of β over 1,000 replications (henceforth the 'empirical standard error'), are summarised. Stata version 13 was used for all simulations.
## Simulation design: misspecified imputation model
The simulation results described above evaluate PMM and LRD in a setting where we have a gold-standard imputation method. The simulation design described in this section relates to a setting where the ideal imputation method is unclear: the presence of x and x 2 in the analysis model means it is difficult to find a compatible model for imputing x|y [bib_ref] for the Alzheimer'sDiseaseNeuroimagingInitiative*: Multiple imputation of covariates by fully conditional specification: accommodating..., Bartlett [/bib_ref]. Here, PMM and LRD are expected to perform better than parametric imputation.
A very similar setup to the previous section is used. The key difference is that true model for the data is x ∼ N(1, 1) and y ∼ N(βx 2 , 10 2 ). Three values of R 2 used are again 0, 0.1 and 0.5. This gives a j-shaped relationship between y and x.
The analysis model is a normal errors linear regression,
[formula] y i ∼ N(β 0 + β 1 x i + βx 2 i , σ 2 ). [/formula]
The intercept and linear term are estimated even though their true values are zero. The imputation model is (6), as in the previous section. Note that no full probability model exists that accommodates both the imputation model and the analysis model [bib_ref] Multiple imputation for an incomplete covariate that is a ratio, Morris [/bib_ref] ; this is the definition of an incompatible imputation model. Missing data are induced in the way described above. [fig_ref] Figure 2 y: vs [/fig_ref] shows y and x in six typical simulated datasets representing the two non-zero strengths of association and three missingness mechanisms.
## Ovarian cancer example
To demonstrate PMM and LRD in practice, we provide a simple analysis of a real partially observed dataset. Clark and Altman developed a prognostic model for time to death in 1,189 individuals with epithelial ovarian cancer [bib_ref] Developing a prognostic model in the presence of missing data, Clark [/bib_ref] , of whom 842 died. Ten of the covariates considered for this model were incomplete, and complete cases analysis included just 518 patients. Using this dataset, we compare some of the approaches of our simulations. One of the covariates considered by Clarke and Altman was albumin in g/dL, and was missing in 392 patients. In this dataset albumin has mean 38, standard deviation 5.3, and moderate skew of -0.52. Our analysis model is a Cox model with age in years (which is complete), albumin and albumin-squared as covariates [bib_ref] Regression models and life tables, Cox [/bib_ref].
The approaches compared are as follows:
1. Complete cases. Analyse the subset of 797 patients with observed albumin. 2. Parametric imputation where albumin is imputed from a normal errors linear model. 3. PMM with type 2 matching and k = 1. 4. PMM with type 1 matching and k = 10. 5. LRD with type 2 matching and k = 1. 6. LRD with type 1 matching and k = 20.
The choice of settings for PMM and LRD is to reflect some of the extremes explored in our simulations. All imputation models include as covariates age, death (yes/no) and the Nelson-Aalen estimate of the cumulative hazard function [bib_ref] Imputing missing covariate values for the cox model, White [/bib_ref]. For each imputation method M = 100 imputations were used to keep the impact of Monte Carlo error small. After imputation, albumin 2 was passively imputed by squaring the imputed value of http://www.biomedcentral.com/1471-2288/14/75 albumin [bib_ref] Multiple imputation using chained equations: Issues and guidance for practice, White [/bib_ref]. The Cox model was fitted in each imputed dataset and estimates combined according to Rubin's rules. Results for bias are given in [fig_ref] Figure 3: Bias under a correctly specified imputation model, according to method [/fig_ref]. Complete cases is unbiased under MCAR and with β = 0, but becomes increasingly biased under the MAR mechanisms. Parametric imputation is unbiased in all scenarios as would be expected, because the imputation model is correctly specified. LRD appears to be unbiased throughout. PMM suffers a small downwards bias for k = 10 under strong MAR. However, the magnitude of this bias is miniscule, and it is still a vast improvement on complete cases analysis. The type of matching does not appear to have any influence on bias.
# Results
## Simulation results: correctly specified imputation model
Coverage results are given in . Again, parametric imputation performs well. PMM and LRD both tend towards under-coverage. This is worse with type 2 matching than type 1, though increasing k alleviates problems for both types. For type 2 matching, coverage is worse with smaller β.
The empirical standard errors of methods are given in . Complete data analysis has the lowest standard errors, while complete cases and parametric imputation also tend to be low. PMM and LRD have the largest standard errors with β = 0 and MAR. There is a strong effect of k on empirical SE, with larger values of k never inferior to smaller values.
Taking these results together, it appears that the largest values of k used are optimal. There is no implication for bias with LRD, and for PMM the bias is miniscule. Coverage is always improved through larger values of k, as is efficiency. Type 1 matching provides better coverage than type 2 for both PMM and LRD. In scenarios where type 1 and 2 matching have comparable coverage, efficiency is also similar, although slightly lower for type 1 matching. The results for comparable forms of PMM and LRD are indistinguishable. These results can be interpreted in terms of the probability of repeated donation: if a donor is selected for many individuals within an imputation, this will lead to inefficiency; if a donor is repeatedly used by the same individuals across imputations this will lead to inefficiency and underestimation of the between-imputation variance. http://www.biomedcentral.com/1471-2288/14/75
## Results: misspecified imputation model
Results are presented in [fig_ref] Figure 6: Bias under a misspecified imputation model, according to method [/fig_ref] , with the design of plots following those presented in the previous section.
Parametric imputation now suffers a large bias for nonnull associations, in the worst scenarios being more than half of the true value for β. With β = 0 and MAR, PMM and LRD have a very slight downwards bias for small k with type 1 matching. This is not present with type 2 matching. With β > 0 PMM and LRD always alleviate the bias seen with parametric imputation. With the 'modest' strength of association, β = 3.3, both methods have least bias with k = 1; as k increases there is a modest downwards bias under strong MAR only. In the extreme case of β = 10 PMM and LRD introduce a very serious degree of bias, particularly under MAR: PMM is biased away from zero and LRD towards it. To understand this bias, consider the imputed values for [fig_ref] Figure 2 y: vs [/fig_ref]. For PMM there will be a vertical spike of imputed values at the tails of the x distribution, while for LRD the imputed value in both tails will lie parallel to the slope of the (linear) imputation model, attenuating the degree of curvature in imputed values.
For many of the settings considered, the bias of complete cases analysis is smaller than for any of the imputation methods. For β = 10 this initially appears surprising, but occurs because the strong association between y and x comes close to the assumption required for complete cases analysis to be valid, that the probability of x i being missing is conditionally independent of y i given x i [bib_ref] Regression with imputed covariates: A generalized missing-indicator approach, Dardanoni [/bib_ref].
The coverage of imputation methods is also often poor . Parametric imputation gives coverage greater than 95% when β = 0 and much lower -close to 0% in one scenario -with β > 0. With β = 0, PMM and LRD give slight over-coverage with type 1 matching, while type 2 matching gives under-coverage. For both types of matching, coverage rates increase slightly as k increases, as seen previously with a correctly specified imputation model. With a non-zero association between y and x and MAR, coverage can become extremely poor for all forms of PMM and LRD. For strong MAR, increasing k appears to slightly alleviate problems, while for weak MAR it adds to them. With β = 3.3 coverage for PMM and LRD are very similar, but with β = 10 PMM tends to give better coverage. Again, although PMM and LRD can improve upon parametric imputation the majority of the time, problems are not 'solved', and in the majority of settings considered complete cases analysis has better coverage.
Comparison of empirical standard errors is largely unhelpful in this context because some methods have large degrees of bias. However, it is worth noting from that PMM and LRD are less efficient than complete cases for all settings considered here. [fig_ref] Table 2: Comparison of coefficients for albumin and albumin-squared in the ovarian cancer data [/fig_ref] displays the log hazard ratio (HR) and 95% confidence intervals for albumin and albumin 2 , according to method. Albumin is coded in units of 100 g/dL and http://www.biomedcentral.com/1471-2288/14/75 centred at its mean. The log hazard ratios and confidence intervals for albumin are very similar for all methods. For albumin 2 , the log HR is smallest for complete cases and parametric imputation, and largest for type 1 matching with large k (for both PMM and LRD). Note that if the inclusion of the squared term depended on its significance at the 5% level, analysis using complete cases or after parametric MI would lead to its exclusion, which is not the case for PMM and LRD.
## Ovarian cancer example: results
Despite confidence intervals being of similar length for larger and smaller values of k, the simulation results in tell us that the coverage properties are rather different, and we should favour those using the larger values of k. Albumin is coded in units of 100 g/dL and mean-centred.
# Discussion
We have aimed to assess the performance of imputation by PMM and LRD in settings where they should perform well, and where they may perform badly. The simulation studies presented have shown that these methods can be adequate when the imputation model is correctly specified, and are an improvement over parametric imputation when the imputation model is misspecified. Nonetheless, with a misspecified imputation model, a strong association between the incomplete covariate and outcome, and data missing at random, performance can become extremely poor. The simulation studies described and reported above involved a single incomplete covariate and a single continuous outcome. In this setting, type 2 matching is equivalent to type 0, failing to acknowledge uncertainty about the parameter of the imputation model. They demonstrate that the performance of PMM and LRD can be acceptable when the imputation model is specified correctly. When the imputation model is misspecified, they are usually an improvement over parametric imputation but can be poor nonetheless.
The design of the second simulation study was intended to provide a tough test for both methods, particularly the specific MAR mechanism used. If the mechanism had worked in the opposite direction and the sign γ 1 had been negative, missing values would have occurred at lower http://www.biomedcentral.com/1471-2288/14/75 values of y, which is one standard deviation from the mean of x.
In using PMM or LRD it is generally preferable to use type 1 matching rather than type 2 (or 0). Larger values of k also tend to be better in terms of coverage and efficiency. For the scenarios investigated, the largest values of k investigated were 10 (PMM) and 20 (LRD). However in much larger datasets with tens of thousands of observed data points, much larger values of k might be considered.
PMM has a cosmetic advantage over LRD that it always imputes observable values meaning it is attractive for imputing non-continuous variables. shows that at the time of writing, this is impossible in the majority of software implementations. Only aregimpute in R and ice in Stata have type 1 matching and allow the user to specify k. Further, ice is the only existing software implementation of LRD.
The main problems with PMM are related to donor sparsity -with few donors in the vicinity of an incomplete case, the imputed values may lead to bias. This also applies to LRD when the imputation model is misspecified. Donor sparsity is expected when there is a large proportion of missing data, under MAR, and in the tails of distributions. PMM also suffers from bias when δ hj has the same sign for all donors in the pool.
In general, the recent work by Bartlett et al. [bib_ref] for the Alzheimer'sDiseaseNeuroimagingInitiative*: Multiple imputation of covariates by fully conditional specification: accommodating..., Bartlett [/bib_ref] may be more fruitful for multiple imputation of incomplete covariates where the analysis model contains nonlinear functions of these. We also note the recent method of Vink and van Buuren as an alternative approach to imputing squares [bib_ref] Multiple imputation of squared terms, Vink [/bib_ref].
# Conclusions
We conclude that PMM and LRD may have a role for imputing covariates when the imputation model is thought to be slightly misspecified, but researchers should focus attention on specifying the imputation model correctly, for example using the recent method described in [bib_ref] for the Alzheimer'sDiseaseNeuroimagingInitiative*: Multiple imputation of covariates by fully conditional specification: accommodating..., Bartlett [/bib_ref] , rather than expecting PMM or LRD to do the hard work.
[fig] Figure 1: Bivariate plots are of x vs. z values in a single imputed dataset. Observed x in purple circles; imputed in blue crosses. Left to right: normal errors parametric imputation, PMM and LRD (type 1 matching with k = 3). Top to bottom: Non-normal residuals, heteroscedasticity and non-linearity. These scenarios represent problems for a linear normal errors imputation model. [/fig]
[fig] Figure 2 y: vs. x in typical simulated datasets with a misspecified imputation model, across various simulation settings. [/fig]
[fig] Figure 3: Bias under a correctly specified imputation model, according to method. [/fig]
[fig] Figure 4, Figure 5: Coverage of 95% confidence intervals under a correctly specified imputation model, according to method. http:Empirical standard error of methods under a correctly specified imputation model, according to method. [/fig]
[fig] Figure 6: Bias under a misspecified imputation model, according to method. [/fig]
[fig] Figure 7, Figure 8: Coverage of 95% confidence intervals under a misspecified imputation model, according to method. http:Empirical standard error of methods under a misspecified imputation model, according to method. [/fig]
[table] Table 2: Comparison of coefficients for albumin and albumin-squared in the ovarian cancer data [/table]
|
Comparison of the Fractional Exhaled Nitric Oxide Levels in Adolescents at Three Schools Located Three Different Distances from a Large Steel Mill
Objectives. Exposure to ambient metals and air pollutants in urban environments has been associated with impaired lung health and inflammation in the lungs. Fractional exhaled nitric oxide (FeNO) is a reliable marker of airway inflammation. In this study, we aimed to compare the FeNO levels of three schools that have different distances from iron and steel industry zone for assessing the effects of heavy metals and air pollution on their respiratory health. Methods. Pulmonary function test and FeNO measurements were evaluated in 387 adolescents in three schools which have different distance from plant. Results. FeNO levels were significantly higher in School I ( = 142; 18.89 ± 12.3 ppb) and School II ( = 131; 17.68 ± 7.7 ppb) than School III ( = 114; 4.28 ± 3.9 ppb). Increased FeNO concentration was related to the distance of iron and steel industry zone in young adults. Conclusion. The FeNO concentrations in school children were inversely proportional to the distance from the steel mill. There are needed some studies that can evaluate the safe distance and legislation must consider these findings.
# Introduction
Fractional exhaled nitric oxide (FeNO) measurements are a simple, highly reproducible, noninvasive method to indicate airway inflammation and have been investigated extensively in asthma [bib_ref] Reproducibility of exhaled nitric oxide measurements in healthy and asthmatic adults and..., Kharitonov [/bib_ref]. Currently, FeNO is viewed as a marker of pulmonary inflammation in asthma [bib_ref] Exhaled biomarkers, Kharitonov [/bib_ref].
Various cells in the lower airways synthesise NO via the oxidation of l-arginine by NO synthase. NO modulates vasomotor and bronchomotor tone and acts as a proinflammatory mediator; thus, it may cause oxidative tissue damage. As such, exhaled NO directly reflects the proinflammatory cytokine mechanisms of central importance in the pathophysiology of asthma and allergy [bib_ref] Nitric oxide in health and disease of the respiratory system, Ricciardolo [/bib_ref].
There is growing evidence of significant associations between FeNO levels and air pollution (i.e., particulate matter, ambient NO, and carbon monoxide) or occupational exposure (i.e., organic solvents and heavy metals) [bib_ref] Association of recent exposure to ambient metals on fractional exhaled nitric oxide..., Rosa [/bib_ref] [bib_ref] Air pollution, airway inflammation, and lung function in a cohort Study of..., Barraza-Villarreal [/bib_ref] [bib_ref] Association between exhaled nitric oxide, ambient air pollution and respiratory health in..., Fischer [/bib_ref] [bib_ref] Association between total blood mercury and exhaled nitric oxide in US adults, Min [/bib_ref].
Considering the simplicity of the measurements, especially with portable NO analysers, FeNO can be used to screen large populations cost effectively. Hence, FeNO may serve as a subclinical inflammation marker for monitoring the environmental health effects of these contaminants [bib_ref] Air pollution and increased levels of fractional exhaled nitric oxide in children..., Flamant-Hulin [/bib_ref]. The use of this biomarker in population-based epidemiological research has great potential for assessing the impact of changing real-world mixtures of ambient air pollutants on the respiratory health of children [bib_ref] Longitudinal effects of air pollution on exhaled nitric oxide: the Children's Health..., Berhane [/bib_ref]. Many studies have confirmed that the dust emitted in the production of iron and steel is an important source of ambient air particulate matter in some areas [bib_ref] Characterization and source identification of heavy metals in ambient PM 10 and..., Dai [/bib_ref]. The smelting process in the iron and steel industry is known to be a major source of ambient nickel, iron, vanadium, lead, copper, and zinc, and an association between the airborne concentrations of nickel, vanadium, zinc, iron, and FeNO has been reported [bib_ref] Association of recent exposure to ambient metals on fractional exhaled nitric oxide..., Rosa [/bib_ref]. Another study showed that the concentrations of all measured heavy metals in an iron and steel industry zone were 1-3.5 times higher than those measured in other areas [bib_ref] Characterization and source identification of heavy metals in ambient PM 10 and..., Dai [/bib_ref].
Although the effects of pollutants from industrial plants are well known, less is known about how much they affect people living near the industrial plants and what safe distances are. Therefore, this study compared the FeNO levels of children at three schools located at different distances from an iron and steel industry zone to assess the effects of heavy metals and air pollution on their respiratory health.
# Materials and methods
The study was carried out in three schools located at three different distances from the Karabük Heavy Steel and Iron Plant. This factory established in 1937 is one of the country's largest industrial plants.
The annual steel production capacity is about 1.5 million tons. Schools were determined randomly among all schools in the province and named as School I, the nearest, 3.3 km, from the plant, School II, 8.8 km from the plant, and School III, the furthest, 27.7 km, from the plant [fig_ref] Figure 1: The location of the schools, according to the heavy industry area [/fig_ref]. Address based school registration system is used in our country. Therefore, children go to the nearest school from their homes.
The study was carried out simultaneously by three different teams, using the same device. There was no highway with heavy traffic close to the schools. There was no facility or factory causing air pollution near 5 km from the schools except for School I. All students were evaluated by chest diseases specialist before participating the study. The participants underwent pulmonary function test (PFT), in accordance with the directives of the American Thoracic Society. Informed consent forms were obtained from the parents of these children. Infection, allergy symptoms, and smoking at home were questioned and those who were identified within last two weeks were excluded. If there was no allergic and atopic history, the participant was included in the study without being subjected to the allergy test. The exclusion criteria are as follows:
(1) asthma, allergic rhinitis, and active pulmonary tuberculosis,
(2) upper and/or lower respiratory tract infection,
[formula] (3) immunologic-rheumatic disorders,(4) [/formula]
those that have menstrual cycles of female students, (5) smoking (active or second hand),
those who do not want to participate.
All processes were performed before lunch and at the end of the course for avoiding the effects of foods containing nitrates, like salad greens, caffeine, consumption of alcoholic beverages, and water. All measurements were made between 12:00 and 14:00 and the wind direction was 3 km/h from the west. The amounts of airborne particulate matter (PM 10 ) and SO 2 are given in [fig_ref] Figure 2: The amount of airborne PM 10 and SO 2 [/fig_ref].
FeNO was measured according to the American Thoracic Society/European Respiratory Society (ATS/ERS) guidelinesusing a handheld device NObreath (NObreath5 FeNO Monitor from Bedfont Scientific Ltd., England). An important characteristic of device is its high sensitivity down to the level of a few parts per billion (ppb) (<5 ppb). Before measuring, each child was asked to rinse their mouth with 10% NaHCO3. Three measurements were performed which were among 30 seconds relaxed tidal breathing for each child.
During the measurements students were in sitting position. The measurements of FeNO plateau of at least 3 sec and 6-15 sec expiration duration were taken into consideration. Nose latch was not used during measurements. Participants who were unable to perform the practice or actual blows after a few attempts were excluded from the study. And also ethical approval and local permissions were obtained for this study.
Statistical Analysis. The SPSS version 21 (IBM Corporation, Armonk, NY, USA) program was used for statistical analysis. We used the independent-samples -test and ANOVA for comparison of parametric data among groups. For the corrections of age, sex, and BMI we used general linear model. The limit for statistical significance was accepted as < 0.05.
# Results
501 students were enrolled in this study. Students who were missing the criteria and did not want to participate in the study were ruled out. FeNO measurements of the remaining 387 students were performed. The study population comprised 182 males and 205 females. The mean age of the total study population was 15.5 + 1.42 (13 to 18 years). There were significant difference in terms of age ( < 0, 001) but there were no significant differences in body mass index (BMI) and gender between groups (resp., = 0.294, = 0.533).
Characteristics of study population are seen in [fig_ref] Table 1: Characteristics of patients included in the study [/fig_ref].
Between the groups of children, FeNO values of students at Schools I and II were significantly higher than School III ( < 0.001 and < 0.001, resp.). There were no significant differences between School I and School II ( = 0.335). Also, there were statistically significant decreases in FeNO value when groups moved away from plant ( < 0.001). FeNO value of School III was lower fourfold than others. Results were shown in [fig_ref] Table 2: The comparison of FeNO value among Schools I, II, III [/fig_ref] and [fig_ref] Figure 3: FeNO levels of three schools [/fig_ref].
There were no statistically significant differences in pulmonary function tests between groups. Nonetheless, predicted peak expiratory flow (PEF) values were significantly lower in the closest school to the factory (School I) as compared with Schools II and III ( = 0.002), even though
# Discussion
Heavy metals inhaled with atmospheric particles may adversely affect human health by accumulating in the lungs [bib_ref] Inhaled particles and lung cancer. Part A: mechanisms, Knaapen [/bib_ref] [bib_ref] Panel studies of air pollution on children's lung function and respiratory symptoms:..., Li [/bib_ref]. There is strong evidence that children living in areas with high air pollution levels or high traffic densities have chronically increased levels of exhaled NO. One of the most important sources of heavy metals is steel plants. To our knowledge, no studies have examined the association between FeNO and the air pollution produced by these plants.
The current study showed that the FeNO concentrations in school children were inversely proportional to the distance from the steel mill. The highest FeNO levels were in children attending the school nearest to the plant and lowest levels were in the school farthest from the steel plant.
Most studies of the relationship between air pollution and high FeNO levels have examined asthmatic children [bib_ref] Air pollution and exhaled nitric oxide in Dutch schoolchildren, Graveland [/bib_ref]. Although the relationship in asthmatic children is well known, fewer studies have examined healthy children. A study conducted in the Netherlands that measured FeNO and ambient air pollution levels for 7 weeks in healthy children found that the increases in the levels of various air pollutants were positively associated with the levels of FeNO [bib_ref] Association between exhaled nitric oxide, ambient air pollution and respiratory health in..., Fischer [/bib_ref]. In a study conducted in Seattle, ambient particulate matter less than 2.5 m in diameter (PM 2.5 ) was found to be associated with exhaled NO in children with and without asthma [bib_ref] Air pollution and exhaled nitric oxide in Dutch schoolchildren, Graveland [/bib_ref]. A study conducted in healthy children aged 9-11 years in New York found a relationship between exposure to ambient metals and increased FeNO [bib_ref] Association of recent exposure to ambient metals on fractional exhaled nitric oxide..., Rosa [/bib_ref]. Another study performed at 25 schools in the Netherlands found a positive association between ambient PM 10 concentrations on the day exhaled NO was measured [bib_ref] Measurement of offline exhaled nitric oxide in a study of community exposure..., Koenig [/bib_ref]. A French study found significant positive associations between schoolyard and classroom PM 2.5 levels and exhaled NO in children with and without asthma [bib_ref] Air pollution and increased levels of fractional exhaled nitric oxide in children..., Flamant-Hulin [/bib_ref]. Similar results have been obtained in studies performed in adults [bib_ref] Association between total blood mercury and exhaled nitric oxide in US adults, Min [/bib_ref] [bib_ref] Air pollution is associated with increased level of exhaled nitric oxide in..., Van Amsterdam [/bib_ref] [bib_ref] Increased exhaled nitric oxide on days with high outdoor air pollution is..., Steerenberg [/bib_ref].
Several studies have shown that the harmful effects of airborne pollutants are caused by oxidative stress and inflammation [bib_ref] Air pollution and health: bridging the gap from sources to health outcomes:..., Solomon [/bib_ref] [bib_ref] Biological effects of diesel exhaust particles (DEP). III. Pathogenesis of asthma like..., Sagai [/bib_ref]. The exposure of human lung epithelial cells to ambient fine particles in vitro showed that increased reactive oxygen species (ROS) production and inflammatory processes were initiated by the release of mediators such as IL-6, IL-8, and TNF-alpha [bib_ref] Diesel exhaust particles and their effect on induced cytokine expression in human..., Takizawa [/bib_ref]. Despite the wide variation in particle composition, evidence suggests that the secretion of cytokines by lung epithelial cells exposed in vitro to particles from different sources is regulated by common cell-signalling pathways. These findings imply that particlemediated injury follows common pathogenic mechanisms. Apart from oxidative stress and inflammation, particlebound materials deposited in the lungs interact directly with pulmonary cells and damage them. Soluble oxygentransferring materials and electrophiles (such as transition metals) can cause local damage and cross the cell membranes of pulmonary epithelial cells [bib_ref] Air pollution and health: bridging the gap from sources to health outcomes:..., Solomon [/bib_ref].
Pulmonary function test (PFT) measures were not associated with the levels of FeNO in the current study. In a study of healthy children, Fischer et al. reported that although there was a significant relationship between ambient pollution and FeNO, there was no relationship with the PFT results [bib_ref] Association between exhaled nitric oxide, ambient air pollution and respiratory health in..., Fischer [/bib_ref]. Therefore, researchers have suggested that the subclinical and asymptomatic respiratory system effects of FeNO are not reflected in PFT results.
All of the abovementioned studies showed that the exhaled NO level was a useful biomonitor of individual exposure to air pollutants and to airborne particles.
Air pollution from iron-and steel-making operations has always been an environmental concern. The characterisation of heavy metals in ambient air particulate matter emissions from steel plants has been reported for many countries worldwide, including the United Kingdom, Spain, Poland, Italy, Turkey, Australia, and Korea [bib_ref] Characterization and source identification of heavy metals in ambient PM 10 and..., Dai [/bib_ref]. In a study of the spatial distribution of air pollution, Kara et al. determined that an integrated pollution index was high even as far as 10 km from a heavily industrialised region [bib_ref] Spatial distribution and source identification of trace elements in topsoil from heavily..., Kara [/bib_ref]. In our study, the FeNO levels were significantly elevated in the students at two schools: one was 3.3 km from the plant and the other was 8.8 km.
Despite national and international laws and restrictions on the waste from industrial plants, little is known about how the spatial distribution distance of air pollutants, which is determined by geographical factors such as temperature, humidity, and wind, affects the health of children. Therefore, there are no regulations or protective mechanisms for people living close to industrial plants. Our study showed that the health of students in school almost 9 km from the plant was affected negatively. While industrial plants are often originally built outside residential areas, they become surrounded by urban areas as cities expand as a result of increasing population and urban sprawl. Few studies have examined the safe distances from industrial plants. Ours is the first study to do so for FeNO.
One limitation of our study was that we did not measure the ambient particulate matter in the schools. Because all of the measurements were performed on the same day, this did not affect our findings regarding the distance to the plant. A second limitation is that the heavy metal levels in the particles were not determined.
In summary, it is necessary to draw attention to the effects of atmospheric heavy metal pollution on people living near industrial plants. Studies need to evaluate safe distances and legislation must consider these findings.
# Ethical approval
This study was approved by the Abantİzzet Baysal University Clinical Researches Ethics Committee, 2014/129.
[fig] Figure 1: The location of the schools, according to the heavy industry area. [/fig]
[fig] Figure 2: The amount of airborne PM 10 and SO 2 . [/fig]
[fig] Figure 3: FeNO levels of three schools. [/fig]
[table] Table 1: Characteristics of patients included in the study. [/table]
[table] Table 2: The comparison of FeNO value among Schools I, II, III. <0.001 § * Unadjusted values, adjusted by age, and § adjusted by age, sex, and BMI. Data were showed as mean ± sd. [/table]
|
Pseudo and resistant hypertension: A chaotic perspective
Systemic blood pressure (BP) may oscillate for homeostatic needs (equilibrium by constancy) or just as shifts in other intrinsic and extrinsic variables known as allostatic changes. This transitory pressure often rises alerts physicians to out-of-control hypertension or even hypertensive crisis. There is a very complex theory underlying these stochastic phenomena, which physicists and mathematicians translate into a single word: chaos. These changes happen according to a stochastic probabilistic pattern that presumes chaotic but somewhat predictable and nonlinear modeling of BP-related dynamics as a mathematical approach. Based on the chaos theory, small changes at the initial BP (baseline overtime) values could disturb the homeostasis leading to extreme BP chaotic shifts. These almost insignificant oscillations may also affect other variables and systems, leading to the misdiagnosis of hypertension, "out-of-control" BP levels, and resistant hypertension (RHT). Thus, these unpredictable and transient increases in BP values may be improperly diagnosed as the white coat and masked or resistant hypertension. Indeed, the interference of the chaos in any phenotype of (true or false) hard to control BP is not considered in clinical settings. This review provides some basic concepts on chaos theory and BP regulation. Besides pseudoresistant hypertension (lack of adherence, circadian variations, and others (white-coat, masked, early morning effects or hypertension), chaotic changes can be responsible for out-of-control hypertension.K E Y W O R D Schaos, Lorenz's attractor, refractory hypertension, resistant hypertension, stochastic system
## Ta b l e 1 premises that characterize chaotic behavior (chaos theory)
## Periodicity final determined by onset predictability (polynomial) tendency to go back to the beginning cyclicity
No (Aperiodic) Yes Partially (Imprecise) Yes (Close to, but not exactly) Yes self-restoring systems. [bib_ref] Dynamics and stability of a new class of nonlinear integrated models with..., He [/bib_ref] [bib_ref] Integer-dimensional fractals of nonlinear dynamics, control mechanisms, and physical implications, He [/bib_ref] A plan is stochastic self-restoring if it sustains (nonlinear autoregressive integrated conditions): A random force or an unpredictable disturbance that may cause a deviation from equilibrium; A restoring force that reduces the negative (or positive) deviation from equilibrium via its upward (downward) component;
A resistance force that prevents rapid change in response to the perturbations.
Systemic BP may oscillate to maintain homeostatic needs and the body constancy, or just as shifts in other intrinsic and extrinsic (allostatic equilibrium) variables and systems. These latter changes happen according to a stochastic probabilistic pattern, which means "randomly determined" 3 that may be statistically analyzed but may not be predicted precisely. This approach requires a mathematic nonlinear dynamics regressive analysis based on the chaos theory. [bib_ref] The adaptive value and clinical significance of allostatic blood pressure variation, James [/bib_ref] This critical review reproaches some crucial topics on resistant hypertension and chaotic or complex BP system.
## Pseudo-resistant hypertension
Despite advances in diagnosis and management strategies, uncontrolled hypertension remains a challenging problem and a primary cause of death for 7.5 million people each year globally. [bib_ref] Resistant hypertension: detection, evaluation, and management: a scientific statement from the, Carey [/bib_ref] Out-of-office BP is a more significant predictor of renal and cardiac morbidity and mortality compared to in-office readings. [bib_ref] Effects of intensive versus standard office-based hypertension treatment strategy on white-coat effect..., Ghazi [/bib_ref] [bib_ref] Adverse prognostic value of persistent office blood pressure elevation in white coat..., Mancia [/bib_ref] BP oscillation should lead to (false) diagnoses such as pseudo resistance, including white-coat and masked hypertension. [bib_ref] The white-coat effect is an independent predictor of myocardial ischemia in resistant..., Modolo [/bib_ref] [bib_ref] Effects of intensive versus standard office-based hypertension treatment strategy on white-coat effect..., Ghazi [/bib_ref] [bib_ref] Adverse prognostic value of persistent office blood pressure elevation in white coat..., Mancia [/bib_ref] [bib_ref] Cardiovascular outcome in treated hypertensive patients with responder, masked, false resistant, and..., Pierdomenico [/bib_ref] Thus, understanding some major concepts on "general systems," BP regulation, and "chaos theory" can help physicians treat these clinical conditions.
To understand the occasional increases in BP, some definitions of "general systems" and "chaos theory" are below 3,6,13-18 :
Homeostasis: self-regulation processes to maintain stability while adjusting to a dynamic equilibrium by continuous changes;
Allostatic: state of internal and physiological equilibrium maintained by an organism in response to actual or perceived environmental stressors;
Stochastic: property of a random probability distribution;
Chaos: random or unpredictable behaviors in complex systems governed by deterministic laws. Deterministic chaos suggests a paradox connecting randomness/unpredictability and deterministic processes.
## The general system theory and chaos
In 1925, Ludwig von Bertalanffy,not satisfied with the physical and deterministic approaches to Biology, proposed an organismic conception (Organismic Biology) emphasizing the consideration of the organism as a group or system. The biological systems may be the cells, organisms, or populations presenting the common characteristic of being composed of many other systems in interaction; these mechanisms were nominated cum plicate (Greek: complicated) systems. [bib_ref] The wisdom of the body, Cannon [/bib_ref] Fundamentally, these hard-to-understand subsystems work jointly to produce coherent behaviors (constancy or equilibrium). This initial concept led to many articles, books, and conferences on "general system theory" in many areas of knowledge. Thus, the human organism should be a system of much smaller subsystems with common characteristics.Actually, this most profound intuition concerning real-life "cum plicate" systems historically dates back to Heraclitus (about 540 BC) and Claude Bernard (1813-1878) with the concept of Homeostasis. This term was perfected and coined later by Cannon 20 :
Homeostasis results from the response to a system perturbation and occurs as a retro alimentation looping called feedback mechanisms, well known nowadays as positive or negative stimuli. The concepts above gained space in many other areas of knowledge as a new paradigm called "general systemic thought".A nonlinear or chaotic system behavior of almost the totality of the existing systems, including BP control, has grown since the 1960s. The complex nonlinear systems obey the chaos theory, which studies the foresight and order of the complex (chaotic) systems, although random. [bib_ref] The adaptive value and clinical significance of allostatic blood pressure variation, James [/bib_ref] The antique determinism and complete predictability do not have to space in the chaotic theory because of its nonlinear expression. [bib_ref] The adaptive value and clinical significance of allostatic blood pressure variation, James [/bib_ref] [bib_ref] From crude cardiovascular signals to chaos, Persson [/bib_ref] Later on, chaotic systems and outcomes were included in the chaos theory. [bib_ref] From crude cardiovascular signals to chaos, Persson [/bib_ref] [bib_ref] Is the normal heart rate "chaotic" due to respiration?, Wessel [/bib_ref] For five decades, theoretical arguments were presented seeking the consideration of the human body as a nonlinear dynamic deterministic system and, therefore, dependent on the laws of chaos. [bib_ref] Is the normal heart rate "chaotic" due to respiration?, Wessel [/bib_ref] [bib_ref] Chaos as an intermittently forced linear system, Brunton [/bib_ref] [bib_ref] A history of chaos theory, Oestreicher [/bib_ref] [bib_ref] Nonequilibrium chaos of disordered nonlinear waves, Skokos [/bib_ref] Accepting such ideas without the restrictions of the traditional, linear, perfect, and immutable determinism in all sciences seemed closer to human thought and the universe . Thus, a partial fusion of classical determinism and entropic chaos has occurred, but homeostasis, general systems, allostasis, milieu internet, and equilibrium still have space in human physiology and medicine. [bib_ref] A history of chaos theory, Oestreicher [/bib_ref]
## Blood pressure as a nonlinear variable
Nonlinear behavior is present in almost the totality of the existing systems, including biological ones. [bib_ref] The adaptive value and clinical significance of allostatic blood pressure variation, James [/bib_ref] [bib_ref] Chaos as an intermittently forced linear system, Brunton [/bib_ref] [bib_ref] Nonequilibrium chaos of disordered nonlinear waves, Skokos [/bib_ref] [bib_ref] Using synchronization of chaos to identify the dynamics of unknown systems, Sorrentino [/bib_ref] [bib_ref] Exploiting the ambulatory blood pressure monitoring via chronobiometric and chaosbiometric methods for..., Cugini [/bib_ref] [bib_ref] Chaos in the cardiovascular system: an update, Wagner [/bib_ref] In this scenario, BP is a major complex variable, ranging between randomness linearity, and health-disease, using the heart-rate variability, using techniques of the chaotic domain. [bib_ref] The adaptive value and clinical significance of allostatic blood pressure variation, James [/bib_ref] [bib_ref] Monitoring and surveillance of chronic non-communicable diseases: progress and capacity in high-burden..., Alwan [/bib_ref] [bib_ref] The control mechanisms of heart rate dynamics in a new heart rate..., He [/bib_ref] [bib_ref] Blood pressure variability: a new therapeutic target on the horizon, Cuspidi [/bib_ref] Some authors use it to calculate a deterministic critical value to the concept of risk, superior to the habitually limited time and frequency domains. [bib_ref] Chaos in the cardiovascular system: an update, Wagner [/bib_ref] [bib_ref] Blood pressure variability: a new therapeutic target on the horizon, Cuspidi [/bib_ref] Finally, the chaotic, discontinuous, and uncertainty of nature, always an enigma to the researchers, has been integrated into biological and health sciences.
Hypervolemia and autonomic nervous system imbalance are the most relevant factors for RHT and refractory hypertension. Obesity, endothelial dysfunction, hyperaldosteronism, sleep apnea, arterial stiffness, and inflammation are also involved in this complex syndrome.
The nonlinearity of BP and its chaotic nature may not be considered during the efforts to reduce the BP in truly RHT subjects or "false out-of-/control." These later patients are usually normotensives by 24 h-ambulatory blood pressure monitoring or home blood pressure monitoring, but the BP increase (office or night) may be concomitant with chaotic variations. [bib_ref] Is the normal heart rate "chaotic" due to respiration?, Wessel [/bib_ref] [bib_ref] The control mechanisms of heart rate dynamics in a new heart rate..., He [/bib_ref] The main premises of chaos 6 are shown in shifts may result in an unpredictable effect known as the "butterfly effect". [bib_ref] Chaos as an intermittently forced linear system, Brunton [/bib_ref] [bib_ref] Factors associated with therapeutic inertia in hypertension: validation of a predictive model, Redón [/bib_ref] In addition, intrinsic variables in the BP regulation system depend on other extrinsic conditions; (iii) confirming and treating true RHT require surveys to point out an RHT prevalence between 6 and 12% of the general population of the hypertensives in comparison to later rates reached 21%. 1,27,28 Often, we undervalue the effects of F I G U R E 1 Orchestra and storm: variations of blood pressure according to "normal" levels (normotension, controlled resistant hypertension (RHT), white-coat effect) and hypertension (masked RHT, true RHT) in individuals taking four or more classes of antihypertensive drugs. Comparison between blood pressure values obtained by the office and ambulatory blood pressure monitoring methods. All the interchangeable possibilities may occur in normotensive, pseudoresistant, and true hypertensive patients. These transient modalities of blood pressure changes illustrate the RHT as an unstable, stochastic, and probably chaotic behavior of the variable in this syndrome. Emergency crises can happen anytime and from any of these modalities randomized processes. In addition, intrinsic variables in the BP regulation influence other extrinsic conditions. Confirming and treating true RHT requires surveys to point out an RHT prevalence between 6 and 12% of the general population of the hypertensives or later rates observed around 21%. [bib_ref] Resistant hypertension: detection, evaluation, and management: a scientific statement from the, Carey [/bib_ref] Finally, the BP technique overestimated the prevalence of uncontrolled RHT in approximately 33% of the patients emphasizing the importance of obtaining accurate BP measurements. [bib_ref] Resistant hypertension: detection, evaluation, and management: a scientific statement from the, Carey [/bib_ref] Due to inadequate BP technique and adherence, clinical inertia over-or underestimates RHT diagnosis. [bib_ref] Factors associated with therapeutic inertia in hypertension: validation of a predictive model, Redón [/bib_ref] Thus, falsely high or "normal" BP readings can be responsible for misdiagnosing RHT. Specialists must pay attention to BP oscillations, always aiming for "controlled levels" and cardiovascular equilibrium by stochastic self-restoring disease mechanisms (allostasis). These harmful and spurious BP oscillations may be due to some failures in the stochastic chaos process of BP regulation and not a part of an allostatic self-restoring condition in RHT individuals. In this sense, the crucial issue is to avoid maintaining a patient's lifetime with a mistaken diagnosis and prognosis based only on the initial, punctual, or historical BP measurements. It is a medical obligation to question the BP values periodically, caring out correctly at the medical office, at home, and even by ambulatory blood pressure monitoring. Finally, reviewing some pivotal contents on general systems and deterministic nonlinear processes is critical to the better comprehension of outlier and unstable BP values in these hard-to-control patients.
## Both allostatic and stochastic processes in homeostasis
## Blood pressure and chaos
BP is a nonlinear dependent variable (y) related to many other influences and factors that aim at cardiovascular homeostasis in a chaotic (complex) general system. Office BP is the gold standard for the screening, diagnosis, and management of hypertension. However, optimal diagnosis and successful management of hypertension cannot be exclusively obtained by a handful of conventionally acquired BP readings.
BP and blood flow patterns in humans are variable, allowing energyefficient responses to diverse stimuli from outside (environmental) and
inside (individual's daily, postural, metabolic, emotional). [bib_ref] Integration and control of circulatory function, Guyton [/bib_ref] Pressureflow regulation is a significant component of virtually all integrated physiologic responses and can be systemic or organ-selective. [bib_ref] Integration and control of circulatory function, Guyton [/bib_ref] Usually, the most crucial factor in BP regulation is the level of outflow of the sympathetic nervous system, which affects immediate (seconds, minutes) and long-term (weeks to months) cardiovascular and BP responses. [bib_ref] Integration and control of circulatory function, Guyton [/bib_ref] BP variation is the result of normal and abnormal discharges from the central nervous system (e.g., posterior hypothalamus); however, abnormalities of feedback mechanisms (parasympathetic reflexes) lead to clinical abnormalities.Besides all these participants in BP control, many other components in the blood/plasma/serum, cellular and subcellular levels, and other extrinsic interferences integrate the fine-tuned adjustments to get stable and optimal pressor values. [bib_ref] Adding home and/or ambulatory blood pressure to office blood pressure for cardiovascular..., Mancia [/bib_ref] However, some "small shifts and mistakes" may probably happen in this well-tuned equilibrium and, analogously, turn a calm, silent and blue sky into an unstable, dark, and noisy tempest. [bib_ref] Relationship of autonomic imbalance and circadian disruption with obesity and type 2..., Boer-Martins [/bib_ref] As stated in the chaos theory, small changes at the initial condition (BP) are decisive to determine the duration, strength, disarranges, and damages (hypertensive disease) to the general system. Indeed, this "storm" in the BP system is not predictable by usual mathematical modeling, probabilistic calculus, or well-established statistical methods. [bib_ref] Dynamics and stability of a new class of nonlinear integrated models with..., He [/bib_ref] [bib_ref] Integer-dimensional fractals of nonlinear dynamics, control mechanisms, and physical implications, He [/bib_ref] The key to previewing BP values over time is a nonlinear autoregressive integrated process that applies Newton's second law to stochastic self-restoring systems.
## Consequences of blood pressure dynamics in a chaotic system
Poincare was the first scientist to glimpse the possibility of chaos. A deterministic system exhibits aperiodic behavior that depends sensitively on the initial conditions, rendering long-term prediction. [bib_ref] Chaos and physiology: deterministic chaos in excitable cell assemblies, Elbert [/bib_ref] Human organisms work as complex (meaning, chaotic) nonlinear systems and almost all the totality of systems known in the Universe. The Analytical techniques derived from chaos theory can help characterize the stability and complexity of blood pressure control, which may provide essential measures for predicting cardiovascular risk. Chaos is located in electroencephalogram data, R-R intervals from electrocardiograms, and cellular level, but only a few studies deal with chaos in sustained hypertension. [bib_ref] Complexity and "chaos" in blood pressure after baroreceptor denervation of conscious dogs, Wagner [/bib_ref] According to these premises, as with almost all biological systems in humans, stochastic BP is a form of maintaining the body's homeostasis. For example, heart-rate amplitude depends on parasympathetic and sympathetic nervous activities; when the autonomic activity remains unchanged, heart-rate amplitude during resting reflects basal metabolism. Thus, heart rate parameter alterations suggest that agerelated decreased heart-rate variability, ultra-reduced heart-rate variability in heart failure, and ultra-elevated heart-rate variability in STsegment alterations refer to age-related decreased basal metabolism, impaired myocardial metabolism, and sympathetic nervous system hyperactivity triggered by myocardial ischemia, respectively. [bib_ref] How to cite this article: Moreno H. Pseudo and resistant hypertension: A..., Lee [/bib_ref] Finally, occasional and circumstantial BP measurements taken and rated in pseudo and true RHT can avoid under-and overdiagnoses in out-of-control subjects.
## Concluding remarks
When failing the best pharmacological treatment in patients with true refractory hypertension and even in the hard to control hypertension, clinicians need to understand the possibility of chaos as taking part in BP arises.
# Acknowledgments
The author thanks Dr. Juan Carlos Yugar-Toledo, Dr. Silvia Elaine Ferreira-Melo, and Dr. Bruno Rodrigues for the extensive review of the manuscript.
[fig] Figure 1: Figure 1. [/fig]
|
Duffing oscillation-induced reversal of magnetic vortex core by a resonant perpendicular magnetic field
Nonlinear dynamics of the magnetic vortex state in a circular nanodisk was studied under a perpendicular alternating magnetic field that excites the radial modes of the magnetic resonance. Here, we show that as the oscillating frequency is swept down from a frequency higher than the eigenfrequency, the amplitude of the radial mode is almost doubled to the amplitude at the fixed resonance frequency. This amplitude has a hysteresis vs. frequency sweeping direction. Our result showed that this phenomenon was due to a Duffing-type nonlinear resonance. Consequently, the amplitude enhancement reduced the vortex core-switching magnetic field to well below 10 mT. A theoretical model corresponding to the Duffing oscillator was developed from the Landau-Lifshitz-Gilbert equation to explore the physical origin of the simulation result. This work provides a new pathway for the switching of the magnetic vortex core polarity in future magnetic storage devices.A magnetic vortex state 1,2 is a ground state of a magnetic nanostructure that consists of a perpendicularly magnetized core and in-plane curling magnetizations around the core(Fig. 1a). Because of its importance in fundamental physics, research on the vortex state is an important emerging topic in magnetism studies 1-7 and it has a high potential for application in high-density data storage devices 6-13 . A magnetic vortex state is energetically fourfold degenerated, which is determined by its polarity and chirality, where the polarity refers to the perpendicular direction of the core magnetization, p core (with a value of 11 for the upward direction, or -1 for the downward direction) and the chirality, c, refers to the curling direction of the in-plane magnetization (a value of 11 or -1 for the counterclockwise or clockwise directions, respectively). Obviously, the success of a magnetic vortex device will critically depend on the question of how to control the vortex polarity and chirality effectively. Much effort has been invested recently in developing various methods for reversing the vortex polarity and chirality with a low magnetic field. While the chirality can be reversed easily with a weak field of ,50 mT (ref. 13), the magnetic field required to reverse the vortex core is on the order of 500 mT, which is too large for practical use in device applications[14][15][16][17].To reduce the vortex core-reversal field, an alternative approach used a dynamic field 6-10,13,16-23 . The chirality reversal field can be almost halved when a short-pulsed field is applied 13 . A promising result is also reported for an AC oscillating magnetic field set at the vortex resonance frequency, so that the vortex excitation could assist its polarity reversal. A representative example of such an approach is the vortex gyration excitation, in which the vortex core exhibits a spiral motion as an AC magnetic field is tuned on at the gyration eigenfrequency 6-10,19,21 . Core switching occurs subsequently through vortex-antivortex creation and annihilation 6 as the core's moving speed exceeds a critical value 19 . The core reversal field can be reduced in such a manner to values far below 10 mT (refs 6,7,19).However, this method contains a fundamental problem for applications. After the core reversal and turning off the field, the core gyration exponentially decays to its initial position. The decay radius is comparable to the lateral size of the sample and the relaxation takes a few hundred nanoseconds 21 . This is a severe obstacle to reading the polarity 16 .Recently, Wang and Dong 16 and Yoo et al. 17 found a new method of vortex core flip from numerical simulation. They demonstrated that the vortex core polarity could be switched in a radial excitation mode 16,17,24,25 by a perpendicular AC magnetic field. In contrast to the gyration mode-assisted switching, which involves the vortex core motion, the radial mode-assisted core switching involves only axial symmetric oscillations, thus preserving the vortex core position. Obviously, the radial mode-assisted core switching has a completely different mech-OPEN
anism from the gyration mode-assisted core switching. The underlying mechanism of the radial mode-assisted core switching was not clearly shown by the simulation. The critical field obtained by the radial mode in these studies is of the order of 20 mT (ref [bib_ref] Radial-spin-wave-mode-assisted vortex-core magnetization reversals, Yoo [/bib_ref] , larger than the gyration mode-assisted core reversal.
In this work, we studied the underlying mechanism of the radial mode oscillation and outlined a new pathway to reduce the coreswitching field further down to the mT range, which was more comparable to the critical field of the gyration-assisted core switching. In addition to micromagnetic simulations [bib_ref] User's Guide, Version 1.2a4, Donahue [/bib_ref] , we also established a dynamical equation for the radial mode oscillation from the Landau-Lifshitz-Gilbert (LLG) equation [bib_ref] Dynamics of domain walls in magnetic nanostrips, Tretiakov [/bib_ref]. This equation clearly explores the nonlinear behavior of the radial mode and the critical field reduction.
For direct comparison of the critical field reduction, the simulation structure was set as described by Yoo et al. [bib_ref] Radial-spin-wave-mode-assisted vortex-core magnetization reversals, Yoo [/bib_ref] [fig_ref] Figure 1 |: Magnetic vortex and core reversal induced by sweeping frequency [/fig_ref]. According to previous studies, the radial modes are classified by the node number n . The first mode has one node, the vortex core, which means that the magnetization does not oscillate temporally at the vortex core, but the other parts almost uniformly oscillate. The second mode has two nodes; one is the vortex core and the other a concentric circle. Yoo et al. [bib_ref] Radial-spin-wave-mode-assisted vortex-core magnetization reversals, Yoo [/bib_ref] studied the resonance frequency of the individual radial mode and obtained the eigenfrequencies with the same sample structure as in this study: GHz for the first mode (n 5 1), 15.2 GHz for the second mode (n 5 2), and 20.7 GHz for the third mode (n 53). They also showed the vortex core polarity reversal using the first mode with an oscillating external field of 20 mT.
To reduce the radial mode-induced critical field below 10 mT, we stimulated the first mode of the radial oscillation with a different method; that is, sweeping of the external field frequency. The field was sinusoidal with amplitude of 9 mT and the field frequency f was slowly varied from 14.0 to 6.0 GHz over 40 ns. [fig_ref] Figure 1 |: Magnetic vortex and core reversal induced by sweeping frequency [/fig_ref] shows the magnetization oscillation during frequency sweeping with time. The normalized magnetization along the thickness direction m z and the external magnetic field, H z , were plotted together. The term ,m z . means the spatial average over the entire disk. The magnetization oscillation has the same frequency despite the phase difference. From this oscillation, we can get the oscillation amplitude of magnetization, I z , in the thickness direction, which is half the difference between the nearest maximum and minimum values of the ,m z . oscillation. After reaching an external field frequency of 6.0 GHz, the frequency sweeping direction was reversed and f returned to 14.0 GHz.
In [fig_ref] Figure 1 |: Magnetic vortex and core reversal induced by sweeping frequency [/fig_ref] , I z is shown as a function of f. It is interesting to note that an external field of 9 mT can reverse the vortex core polarity. In downward sweeping of the frequency, the almost uniform magnetization oscillation was observed on the disk except for the core conserving its width (inset of [fig_ref] Figure 1 |: Magnetic vortex and core reversal induced by sweeping frequency [/fig_ref]. This uniform oscillation was maintained before I z reached the maximum amplitude of 0.28 when f was 8.7 GHz. After reaching this critical amplitude, the uniform oscillation collapsed and converged into the disk center that generated a breathing motion of the core. Such a breathing generated a strong exchange field when the core was compressed, and then core polarization switching occurred [bib_ref] Sub-nanosecond switching of vortex cores using a resonant perpendicular magnetic field, Wang [/bib_ref] [bib_ref] Radial-spin-wave-mode-assisted vortex-core magnetization reversals, Yoo [/bib_ref]. Amplitude fluctuations near 8.5 GHz and 10.5 GHz are transition effects discussed below. In contrast to downward sweeping, the upward frequency sweeping did not reach the amplitude of 0.28, so the vortex maintained its polarity. This means that one cycle of frequency sweeping generated one core reversal. It is notable that the amplitude obtained with the fixed-field frequencies was the same as the upward sweeping. The fixed-field frequency amplitudes were determined by amplitude saturation after turning on the external oscillating field. To reverse the core polarity with the upward sweeping oscillation and fixedfrequency oscillation, a larger field was required for achieving the sufficient oscillation amplitude. From this sweeping frequency simulation, it was verified that the critical field was reduced to below 10 mT and this reduction was only observed in downward sweeping because of the hysteresis behavior of the frequency.
To study this hysteresis behavior, we constructed a simplified model. The magnetization oscillation of the first mode can be represented by only two variables, h and Q, because with the exception of the core region, the magnetization oscillates almost uniformly [bib_ref] Radial-spin-wave-mode-assisted vortex-core magnetization reversals, Yoo [/bib_ref]. [fig_ref] Figure 2 |: Simplified model for the first radial mode [/fig_ref] shows the definition of h and Q in the core-free model. The angle h represents the magnetization tilting toward the normal direction and Q denotes tilting in the radial direction. The initial magnetization is then described as h 5 0 and Q 5 0. Note that the initial magnetization of the vortex state could have two values of Q, 0 and p, because of the chirality of the vortex. The two states are energetically equivalent, so in this paper we only considered Q 5 0 as the initial state. Using these two variables, the magnetization state of each component was described as M r 5 M S cosh sinQ, M Q 5 M S cosh cosQ, and M z 5 M S sinh. After inserting these components into the LLG equation [bib_ref] Dynamics of domain walls in magnetic nanostrips, Tretiakov [/bib_ref] and solving, we could derive the following coupled equations:
[formula] _ Q~cM S N z {N r sin 2 Q À Á sin hzcH sin vtz a cos h _ h,ð1Þ_ h~{cM S N r cos h sin Q cos Q{a cos h _ Q:ð2Þ [/formula]
Here, the over dot means the time derivative, c is the gyromagnetic ratio, N z and N r are determined by the demagnetizing energy along thê z andr directions, respectively, H is the external field amplitude, and v (5 2pf) is the angular frequency of the external field. In equations (1) and (2), all the parameters are known except for N z and N r . To obtain these two parameters, we reduced equations (1) and (2) assuming that N z ? N r because of the thin structure, and this condition also results in h = 1, after which simplified equations are derived:
[formula] _ Q~cM S N z hzcH sin vtza _ h,ð3Þ_ h~{ 1 2 cM S N r sin 2Q{a _ Q:ð4Þ [/formula]
We carried out a simulation that solved equations (3) and (4). Gaussian-shaped external field pulses were applied sequentially to the nanodisk to excite the first mode and the interval between pulses was tuned for precise resonant amplification [bib_ref] Resonant amplification of vortex-core oscillations by coherent magnetic-field pulses, Yu [/bib_ref]. We turned off the external field pulse and observed the relaxation process [fig_ref] Figure 2 |: Simplified model for the first radial mode [/fig_ref]. The angles h and Q converged exponentially to zero with a spiral trajectory; h was obtained from sinh 5 ,m z (t).2,m z (0)., where, t is time and Q was determined by the averaged value over the entire disk. From this spiral relaxation, _ Q was plotted as a function of h [fig_ref] Figure 2 |: Simplified model for the first radial mode [/fig_ref] and sin2Q was plotted with _ h [fig_ref] Figure 2 |: Simplified model for the first radial mode [/fig_ref]. The damping effect was negligible because a 5 0.01 (=1) and during relaxation, H 5 0. Thus, two linear relations [fig_ref] Figure 2 |: Simplified model for the first radial mode [/fig_ref] determined the values N z 5 0.85 and N r 5 0.14. These values are reasonable because N z 1 N r < 1 for the first mode was composed with almost uniform magnetization.
Next, we solved equations (3) and (4) to obtain the oscillation amplitudes versus the external field frequencies. The time derivative of the equations and the elimination of the _ h and € h terms yielded the following equation of motion of Q:
[formula] (1za 2 )€ Qz v 2 0 2 sin 2QzC(N z zN r ) _ Q {C(2N r sin 2 Q) _ Q{cHv cos vt~0:ð5Þ [/formula]
Here, v 0 5 cM S ffiffiffiffiffiffiffiffiffiffi ffi N z N r p 5 2p 3 10.7 GHz and C 5 acM S . If we neglect the damping term and the external field term, equation (5) exactly corresponds to the simple plane pendulum, € Wzv 2 0 sin W~0 with W 5 2Q, so it was natural for the main dynamics of the radial mode of the vortex to be the same as that of the plane pendulum. Directly solving equation (5) was not easy, so we simplified equation [bib_ref] Symmetry breaking in the formation of magnetic vortex states in a permalloy..., Im [/bib_ref]. Firstly, we compared the third and the fourth terms of equation (5), both terms have _ Q dependence, and we neglected the fourth term which is much smaller than the third term. Then, we expanded sin 2Q in the second term by using a Taylor expansion such as sin x 5 x 2 x 3 /3!1x 5 /5!… and we adopted up to the third order of this expansion. These simplifications produced the following equation.
[formula] € Qzv 2 0 Q(1zbQ 2 )zd _ Q{Fv cos vt~0:ð6Þ [/formula]
This is the well-known Duffing oscillator equation [bib_ref] Noise-enabled precision measurements of a Duffing nanomechanical resonator, Aldridge [/bib_ref] that describes a nonlinear oscillation induced by the Q variation-dependent spring constant. If b 5 0, equation (6) becomes a harmonic oscillator. In this study, b 5 -2/3 due to the Taylor expansion of sin 2Q. Other parameters were determined similarly: the dissipative constant d 5 acM S (N z 1 N r ) and the external force F 5 cH. It is well known that we can solve the Duffing equation by using the van der Pol transformation [bib_ref] Noise-enabled precision measurements of a Duffing nanomechanical resonator, Aldridge [/bib_ref] [bib_ref] Nonlinear switching dynamics in a nanomechanical resonator, Unterreithmeier [/bib_ref]. A solution of equation (6), the frequency-dependent amplitude Q 0 , exhibits hysteresis behavior [fig_ref] Figure 3 |: Comparison between solutions of the Duffing equation and simulations [/fig_ref]. We set H to be 8 mT. In contrast to a harmonic oscillator showing a symmetric resonant peak, the Duffing oscillator exhibits an asymmetric resonance peak and the curve is divided into stable and unstable solutions that are well known as the foldover effect. If we apply the external field with a starting frequency of 14 GHz and monotonically reduce the frequency, the oscillation amplitude would follow the stable solution line until it meets the maximum amplitude point Q max . After passing the maximum point, the amplitude was drastically reduced to follow the stable line. A similar behavior was also expected for the case of increasing frequency, but the amplitude did not reach Q max because the stable line was connected up to 10 GHz. This hysteresis property was also shown as the phase d variation of Q oscillation with respect to the field oscillation [fig_ref] Figure 3 |: Comparison between solutions of the Duffing equation and simulations [/fig_ref]. The micromagnetic simulation result showed almost the same hysteresis behavior [fig_ref] Figure 3 |: Comparison between solutions of the Duffing equation and simulations [/fig_ref] with the Duffing oscillator, but it showed fluctuation when the amplitude www.nature.com/scientificreports and the phase jumped from one stable solution to the other stable line. Note that in [fig_ref] Figure 3 |: Comparison between solutions of the Duffing equation and simulations [/fig_ref] the amplitude is shown with Q. It differs from [fig_ref] Figure 1 |: Magnetic vortex and core reversal induced by sweeping frequency [/fig_ref] , which was plotted with I z (5 sinh 0 ), but the maximum values of Q 0 and h 0 can be converted through the relation N z sin 2 h max 5 N r sin 2 Q max because the demagnetization energy is alternatively transferred between theẑ andr directions.
We observed the Duffing-type oscillation and critical field reduction in other radial modes. [fig_ref] Figure 4 |: Critical fields for the core reversal H c with respect to the... [/fig_ref] presents the minimum field for core reversal, H c , as a function of frequency f. To obtain the critical field with fixed frequency, we fixed the external field frequency and varied the amplitude of the field from 0 with a 10 mT/ns rate. Frequency dependent critical field had three local minima and each minimum corresponded to the first, second, and third mode. The obtained H c values for modes with fixed frequency were respectively 20 mT for n 5 1, 78 mT for n 5 2 mode, and 127 mT for the n 5 3 mode. The critical fields were also determined by the sweeping frequency method. For the second mode, the frequency started from 17 GHz with 20.2 GHz/ns sweeping rate; for the third mode, it started from 24 GHz with the same rate. Then, the critical field was 8.5 mT for n 5 1, 37 mT for n 5 2, and 76 mT for n 5 3, which were almost half the values obtained with fixed frequency. The sweeping rate can change the critical field but the variation is small. For example, n 5 1 mode, H c 5 8.3 mT with 0.1 GHz/ns rate and 8.6 mT with 0.4 GHz/ns rate. The critical field is obtained with 0.1 mT field interval for n 5 1 and 1 mT interval for n 5 2, 3. We confirmed that these reduced critical fields were not achieved through the upward sweeping of the field frequency because of the Duffing-type nonlinear oscillation. In addition, further reduction of the critical field was achieved by using a square wave type external field. For n 5 1 mode, we obtained 6.3 mT for the vortex core reversal.
We tested the scalability of the radial mode-induced core reversal. When the radius of the disk was 120 nm, the critical field obtained by the frequency sweeping method was 9.3 mT. The core of a disk with radius 250 nm reverses its polarity with 12 mT external field. Increasing the radius, the critical field also increases. This scalability is an important property for developing data storage devices. Contrary to the radial mode-induced polarity switching, the critical field with the gyration-induced polarity switching exhibits inverse radius dependence 19 as well as the chirality reversal [bib_ref] Dynamic switching of the spin circulation in tapered magnetic nanodisks, Uhlíř [/bib_ref].
Finally, we point out the chaotic behavior and the phase commensurability in the radial mode oscillation for further studies. Petit-Watelot et al. observed the chaos and phase-locking phenomenon in the vortex gyration with the core reversal [bib_ref] Commensurability and chaos in magnetic vortex oscillations, Petit-Watelot [/bib_ref]. We observed similar behavior in radial mode oscillation. It is expected that a nonlinear oscillator with a sufficiently large driving force would exhibit chaotic motion. We confirmed this chaotic behavior in the radial mode of the vortex. When the oscillating field strength was smaller than H c , a plot of the variable with respect to its time derivative, for example v _ m z w versus ,m z ., showed a circular trajectory. But when the field was larger than H c , this plot becomes complex in the phase space, which manifests its chaotic behavior. [fig_ref] Figure 5 |: Chaotic behaviors of the radial mode oscillation on the phase space [/fig_ref] shows examples of the chaos in the radial mode. The frequency was fixed at 13.5 GHz. When H 5 60 mT , H c [fig_ref] Figure 5 |: Chaotic behaviors of the radial mode oscillation on the phase space [/fig_ref] , it showed a closed circular trajectory, but when H 5 90 mT . H c the trajectory was not closed [fig_ref] Figure 5 |: Chaotic behaviors of the radial mode oscillation on the phase space [/fig_ref]. Further increases in the field resulted in closed trajectories. However, the trajectories were not a simple circle. To close the trajectory, 14 cycles of field oscillation are needed [fig_ref] Figure 5 |: Chaotic behaviors of the radial mode oscillation on the phase space [/fig_ref] and during these 14 cycles, the core reversed four times. In the case of H 5 120 mT, core reversal occurred twice in five field oscillations [fig_ref] Figure 5 |: Chaotic behaviors of the radial mode oscillation on the phase space [/fig_ref] , implying that the core reversal rate was related to the chaotic behavior. Thus, to describe the radial mode of vortex including its chaotic behavior, the core polarity-related term 32 is needed in the motion equation.
In summary, we studied the nonlinear resonance of the radial mode of the vortex and found that the oscillation mode corresponding to the Duffing-type nonlinear oscillator exhibited a hysteresis behavior with respect to the external field frequency. Through the hysteresis effect, we can achieve hidden amplitude that is almost double that obtained with fixed field frequency and this amplitude multiplication effect reduces the critical field below 10 mT. In addition, we pointed out the chaotic behavior of the radial mode for further studies. We think that to complete the study on vortex dynamics, it is timely to start research on the nonlinear behavior in radial modes, as well as in other oscillations of the magnetic vortex.
# Methods
Micromagnetic simulations. We performed a micromagnetic simulation using the OOMMF code [bib_ref] User's Guide, Version 1.2a4, Donahue [/bib_ref] for numerical calculation of the LLG equation. The simulation structure is a circular disk with a diameter of 160 nm and 7 nm thickness, t d , as shown in [fig_ref] Figure 1 |: Magnetic vortex and core reversal induced by sweeping frequency [/fig_ref]. The structure is divided into 2 3 2 3 7 nm 3 unit cells. Uniform magnetization was assumed along the thickness direction. The material parameters were chosen to resemble a typical permalloy with no crystal anisotropy. The saturation magnetization, M S , was 8.6 3 10 5 A/m, the exchange stiffness, A, was 1.3 3
[fig] Figure 1 |: Magnetic vortex and core reversal induced by sweeping frequency. (a), Disk structure used for simulations and the initial vortex state. The in-plane curling magnetization directions are shown as arrows. At the disk center, a perpendicularly magnetized core exists. (b), The magnetization oscillation induced by the external magnetic field H z applied to the thickness direction (ẑ). The field strength was 9 mT with a sinusoidal form; the frequency was varied with 0.2 GHz/ns rate; ,m z . is the spatially averaged magnetization component in theẑ direction normalized by the saturation magnetization. (c), I z , the amplitude of ,m z . oscillation, as a function of external field frequency f. The amplitude obtained from downward sweeping (red line) can achieve sufficient amplitude for core polarity (p core ) reversal but from upward sweeping (blue line) it cannot. Black circles denote saturated amplitude per each frequency. This fixed frequency result also does not obtain the maximum amplitude for p core switching. Snap shots show the magnetization states. The overlapped images represent one cycle of the oscillation and single images exhibit the reversal process of the core. [/fig]
[fig] Figure 2 |: Simplified model for the first radial mode. (a), The angle h is tilting toward the perpendicular direction and Q represents the radial directional tilt. (b), A spiral trajectory during relaxation; before the relaxation, several Gaussian pulses were applied to excite the magnetization oscillation. (c),(d), Relations between two coupled variables, h and Q, which were obtained from (b). [/fig]
[fig] Figure 3 |: Comparison between solutions of the Duffing equation and simulations. The external field was set to be 8 mT. (a), Solutions of equation (6), the Duffing equation. The plane pendulum asymptotically corresponds to the Duffing oscillator (inset). The solution comprises a stable and an unstable line and where the stable line is disconnected, the amplitude Q 0 and the phase d show drastic jumps. Q 0 has the maximum value Q max at the zero phase (d 5 0). (b), Micromagnetic simulation results with the same external field show similar hysteresis behavior. [/fig]
[fig] Figure 4 |: Critical fields for the core reversal H c with respect to the field frequency f. The fixed f results were obtained by amplitude sweeping simulations and the results show three local minima that correspond to each radial mode (n 5 1, 2, 3); ( . ) denotes the critical fields determined by the sweeping frequency method; dashed lines represent the maximum amplitude-frequency points in the sweeping method. [/fig]
[fig] Figure 5 |: Chaotic behaviors of the radial mode oscillation on the phase space. The external field frequency was 13.5 GHz. Phase space plots of ,m z . are shown in (a-d). We used different colors, red and blue, to distinguish the core polarization. (e-h), Poincaré sections with zero phase (d 5 0). To obtain a steady state trajectory, simulation data from 10 to 20 ns was plotted, except for (b) and (f). We plotted (b) and (f) with the data from 10 to 500 ns. (a), When the applied field strength is smaller than H c , the trajectory resembles a closed circle and the Poincaré section is a single point (e). (b), The trajectory fills some area of the phase space with the increased field. (c),(d), Further increase of the field results in closed trajectories, but they do not show a simple circle. The numbers of points in the Poincaré section represent the number of field cycles needed to close the trajectories. [/fig]
|
Cost-effectiveness of rotavirus vaccination in children under five years of age in 195 countries: A meta-regression analysis
a b s t r a c tBackground: Rotavirus caused an estimated 151,714 deaths from diarrhea among children under 5 in 2019. To reduce mortality, countries are considering adding rotavirus vaccination to their routine immunization program. Cost-effectiveness analyses (CEAs) to inform these decisions are not available in every setting, and where they are, results are sensitive to modeling assumptions, especially about vaccine efficacy. We used advances in meta-regression methods and estimates of vaccine efficacy by location to estimate incremental cost-effectiveness ratios (ICERs) for rotavirus vaccination in 195 countries. Methods: Beginning with Tufts University CEA and Global Health CEA registries we used 515 ICERs from 68 articles published through 2017, extracted 938 additional one-way sensitivity analyses, and excluded 33 ICERs for a sample of 1,418. We used a five-stage, mixed-effects, Bayesian metaregression framework to predict ICERs, and logistic regression model to predict the probability that the vaccine was cost-saving. For both models, covariates were vaccine characteristics including efficacy, study methods, and countryspecific rotavirus disability-adjusted life-years (DALYs) and gross domestic product (GDP) per capita. All results are reported in 2017 United States dollars.
# Introduction
Mortality due to diarrheal diseases has declined by 60% since the year 2000, but diarrheal diseases remain one of the leading causes of death for children under 5 every year, and were responsible for an estimated 497,434 childhood deaths in 2019 [bib_ref] Global burden of 369 diseases and injuries in 204 countries and territories,..., Vos [/bib_ref]. Rotavirus is a leading etiology of diarrheal disease mortality globally, responsible for 30% percent of diarrheal deaths among children under 5 [bib_ref] Rotavirus vaccination and the global burden of rotavirus diarrhea among children younger..., Troeger [/bib_ref].
Some of the recent decline in mortality has been attributed to rotavirus vaccination, which averted an estimated 28,000 deaths in 2016 [bib_ref] Rotavirus vaccination and the global burden of rotavirus diarrhea among children younger..., Troeger [/bib_ref]. By the end of 2018, Gavi, The Vaccine Alliance, supported rotavirus vaccination programs in 45 of 57 eligible countries, covering 39% of children in those countries with a full course of a rotavirus vaccine. By April 2020, a total of 107 countries introduced rotavirus vaccines to their routine immunization schedules.
As more countries consider introducing rotavirus vaccination into their routine immunization programs, cost effectiveness of the vaccine can inform their decisions. However, cost effectiveness analyses (CEAs) are not available in 53 countries. Among the 142 countries with existing CEA, 49 have estimates that are based on a single study, meaning that the results are sensitive to the modeling assumptions of the study authors and do not consider the cumulative evidence. For countries with more than one CEA, results can vary greatly. For example, the 20 published incremental cost-effectiveness ratios (ICERs) for Peru vary from $132 to $2,438 per DALY averted. Similarly, in Bangladesh, the 19 published ICERs vary from $23 to $1,543 per DALY averted. Because of this variability, efforts to synthesize CEA results and transfer them to other settings are growing. Two systematic reviews synthesized published evidence of CEA on rotavirus vaccines [bib_ref] A Systematic Review of Economic Evaluation Methodologies Between Resource-Limited and Resource-Rich Countries:..., Thiboonboon [/bib_ref] [bib_ref] Global economic evaluations of rotavirus vaccines: A systematic review, Kotirum [/bib_ref]. The first compared methodological approaches between high-income (HIC) and lowincome (LIC) countries [bib_ref] A Systematic Review of Economic Evaluation Methodologies Between Resource-Limited and Resource-Rich Countries:..., Thiboonboon [/bib_ref]. The second summarized study characteristics and findings, identifying vaccine efficacy as an important source of variability [bib_ref] Global economic evaluations of rotavirus vaccines: A systematic review, Kotirum [/bib_ref].
More recent cost-effectiveness research relied on meta-analysis to synthesize evidence from studies conducted in multiple countries. Haider et al. used a meta-regression approach to estimate the cost-effectiveness of the rotavirus vaccine across 29 LIC and lower-middle income (LMIC) countries [bib_ref] Systematic Review and Meta-Analysis of Cost-effectiveness of Rotavirus Vaccine in Low-Income and..., Haider [/bib_ref]. They produced pooled estimates for the LICs and LMICs under consideration, rather than providing country-specific estimates. Further, their analysis incorporated only three covariates (Gross Domestic Product [GDP], vaccine coverage, and literacy rate) [bib_ref] Systematic Review and Meta-Analysis of Cost-effectiveness of Rotavirus Vaccine in Low-Income and..., Haider [/bib_ref]. Jit et al. investigated costeffectiveness across five European countries using a single model and incorporating country-specific rotavirus burden. They found that results varied across settings due to differences in rotavirus burden and vaccine price, among other factors [bib_ref] The costeffectiveness of rotavirus vaccination: Comparative analyses for five European countries and..., Jit [/bib_ref]. developed and applied a meta-regression approach to CEA to the human papilloma virus (HPV) vaccine [bib_ref] Costeffectiveness of HPV vaccination in 195 countries: A meta-regression analysis, Rosettie [/bib_ref]. They exploited oneway sensitivity analyses extracted from published studies and used covariates for study methods, vaccine characteristics (coverage, cost, type, booster, and target group), and country-specific variables for GDP per capita and HPV burden to predict incremental cost effectiveness ratios (ICERs) across 195 countries.
In this paper we apply meta-regression methods developed by Rosettie et al. to estimate the cost-effectiveness of rotavirus vaccination in 195 countries. We identify and quantify sources of heterogeneity in published CEAs from study methods, and vaccine characteristics, well as economic and epidemiologic conditions in each country (i.e. GDP per capita, rotavirus burden). We focus specifically on vaccine efficacy, because it differs by location [bib_ref] Global burden of 369 diseases and injuries in 204 countries and territories,..., Vos [/bib_ref] [bib_ref] A Systematic Review of the Effect of Rotavirus Vaccination on Diarrhea Outcomes..., Lamberti [/bib_ref] and has been identified as a driver of heterogeneity [bib_ref] Global economic evaluations of rotavirus vaccines: A systematic review, Kotirum [/bib_ref] , but has not yet been used in any meta-analyses of a vaccinepreventable disease.
# Data and methods
## Cost-effectiveness data
We used cost-effectiveness data through 2017 from two comprehensive registries of all published CEA maintained by Tufts University's Center for Evaluation of Risk and Value in Health: (1) CEA registry with cost per quality-adjusted life year (QALY) results [bib_ref] Skills of the Trade: The Tufts Cost-Effectiveness Analysis Registry, Thorat [/bib_ref] , and (2) Global Health CEA registry with cost per disabilityadjusted life year (DALY) results [bib_ref] A Systematic Review of Cost-Effectiveness Studies Reporting Cost-per-DALY Averted, Neumann [/bib_ref]. As previously reported, they searched PubMed for English-language articles using keywords ''QALYs", ''quality-adjusted" [bib_ref] Skills of the Trade: The Tufts Cost-Effectiveness Analysis Registry, Thorat [/bib_ref] and ''cost-utility analysis" for the CEA registry, and ''disability-adjusted" or ''DALY" for the Global Health CEA registry [bib_ref] A Systematic Review of Cost-Effectiveness Studies Reporting Cost-per-DALY Averted, Neumann [/bib_ref]. Abstracts were screened to identify original estimates, and assigned an International Classification of Dis-eases, 10th Revision (ICD-10) code to each article, [bib_ref] Skills of the Trade: The Tufts Cost-Effectiveness Analysis Registry, Thorat [/bib_ref] [bib_ref] A Systematic Review of Cost-Effectiveness Studies Reporting Cost-per-DALY Averted, Neumann [/bib_ref] which we subsequently mapped to 2017 Global Burden of Disease, Injury and Risk Factor Study [fig_ref] Table 1: Descriptive statistics of cost-effectiveness results and peer-reviewed articles on rotavirus vaccines included... [/fig_ref]. Combined, these two registries include 519 incremental cost-effectiveness ratios (ICERs) across 142 countries.
This study complies with the Guidelines for Accurate and Transparent Health Estimates Reporting (GATHER) (Supplementary [fig_ref] Table 2: Predicted incremental cost-effectiveness ratios aggregated to super-region level compared to range of... [/fig_ref].
## Data standardization, mapping, and extraction
Building on the registry data, we performed several tasks to create covariates and extend the dataset, as described in the Supplementary Sections 3 and 4, and previously reported [bib_ref] Costeffectiveness of HPV vaccination in 195 countries: A meta-regression analysis, Rosettie [/bib_ref]. Briefly, we extracted data on 5 study methods variables with missing values in the registry data whenever possible: cost discount rate, QALY/DALY discount rate, health outcome measure, perspective, and time horizon. We collapsed perspectives into two categories: healthcare payer and health sector versus all other categories. To account for economic and epidemiological differences across countries, each ICER was mapped to GBD categories for at least one age group, sex, location, and cause, which were used to pull GDP per capita and rotavirus burden measured as DALYs from diarrhea attributable to rotavirus per 100,000 population for children under 5 for the study year [bib_ref] Rotavirus vaccination and the global burden of rotavirus diarrhea among children younger..., Troeger [/bib_ref]. DALYs combine the effects of mortality and morbidity where a death from rotavirus in a child aged 2 to 4 years is measured as 81.81 years of life lost, and an episode of rotavirus is measured as roughly 0.002 years lived with disability, i.e. 5/365 years duration times an average disability weight of 0.12. We extracted data on four vaccine characteristics: vaccine type (monovalent, pentavalent, or both), coverage, cost, and efficacy. All comparisons were relative to a null comparator defined as no intervention (n = 1,245, 92.6% of final sample), standard of care (n = 98, 7.3%) or placebo (n = 2, 0.1%). When registry entries did not correspond to these categories, we extracted data to recalculate them. Finally, we extracted ICERs and covariates from oneway sensitivity analyses, identifying the covariate that differed, and the reference analysis. ICER values, vaccine costs, and GDP per capita were adjusted to 2017 US dollars ($US).
## Inclusion and exclusion criteria
Of the 519 initial cost-effectiveness results from the Tufts database, we excluded two articles that compared modeling methods (two ratios), or did not report vaccine type (two ratios) [fig_ref] Figure 1: Study flow diagram of study selection in logistic and meta-regression models [/fig_ref]. We extracted an additional 938 cost-effectiveness results from one-way sensitivity analyses for seven variables. We also excluded 35 ratios due to missing data, reporting errors, or because the ICER was 0 [fig_ref] Figure 1: Study flow diagram of study selection in logistic and meta-regression models [/fig_ref].
We conducted a logistic regression analysis (see below) to estimate the probability that a cost-effectiveness result was costsaving using 1,418 cost-effectiveness results, of which 73 were cost-savings. We conducted the meta-regression analysis with a total of 1,345 ICERs (excluding the cost-savings results) from 142 countries and 68 studies.
# Statistical analysis
To estimate ICERs across countries, we adopted a Bayesian mixed-effects meta-regression framework that employs regularization and outlier detection to select covariates and estimate their associations with the observed ICER. We used those estimated associations to predictions ICERs for 195 countries. This framework consists of five stages described in Supplementary Section 5 and previously reported [bib_ref] Costeffectiveness of HPV vaccination in 195 countries: A meta-regression analysis, Rosettie [/bib_ref] [bib_ref] Trimmed constrained mixed effects models: formulations and algorithms, Zheng [/bib_ref].
The first stage estimated prior distributions for seven preselected covariates in which one-way sensitivity analyses were done in the published literature: cost discount rate, DALY/QALY discount rate, perspective, vaccine type, vaccine coverage, vaccine cost, and vaccine efficacy. The aim of the sensitivity analyses was to reduce omitted variable bias and identify prior distributions for variables that were highly correlated, such as log-vaccine cost and log-GDP per capita. The one-way sensitivity analyses differed by one covariate and no unmeasured covariates from their corresponding reference analysis. We therefore matched each sensitivity analysis with its corresponding reference analysis and fit separate linear models to estimate the effect of the difference between covariate values in reference and sensitivity analyses on the difference in the corresponding log-ICERs. The estimated regression coefficients and standard errors from these models were then used as Gaussian prior distributions for these covariates in all subsequent models.
The second stage estimated the association between log-ICER and log-GDP per capita, using a spline ensemble that can estimate a non-linear response curve. The estimates included log-rotavirus burden as a covariate. This stage had a robust statistical approach to outlier detection, with 10% of observed log-ICERs being classified by the model as outliers and trimmed from the dataset ahead of further modeling steps. The resulting spline-transformed GDP per capita variable was used in subsequent stages.
In the third stage, we used a generalized Lasso approach to select additional covariates to include in the final model, thereby balancing covariate selection from a priori knowledge versus statistical approaches. The seven covariates from the first stage, and the spline-transformed log-GDP per capita variables were preselected. Candidate covariates for the model were: whether or not the study used a lifetime time horizon, whether the outcome mea- sure was QALYs or DALYs, and log-rotavirus burden, all of which were selected for the final meta-regression model. In the fourth stage, we created Gaussian priors for the spline transformed GDP per capita variable, and the two variables selected in stage 3 to safeguard against overfitting in the final estimates. Given that we had coefficient estimates for these variables from Stage 3, we standardized the variables, and used a grid-search and 10-fold cross validation to select the standard deviation for Gaussian priors. The selected prior standard deviation for the standardized covariates minimized the mean squared error for predicting data in the hold-out set.
In the fifth stage, we fit a linear mixed model that accounted for between-study variability and the fact that studies often reported multiple ICERs as part of sensitivity analyses. The model has a random intercept for each article, and covariates selected in the third stage of the analysis using priors estimated from the first and fourth stages, respectively. The parameter estimates for the covariates were used to predict ICERs for 195 countries.
Using this modeling framework, we explore three different parameterizations of efficacy to determine whether or not its inclusion improves model fit. First, we include efficacy as a main effect. Second, we define effective coverage as the product of coverage and efficacy (scaled to be between 0 and 100%), where covered population is the primary beneficiary of the intervention. Finally, we estimate a model without efficacy. All model comparisons are based on R-squared and root mean square error (RMSE) fit statistics, with the best fitting model used for final results. The model with efficacy as the main effect had the best fit (Rsquare = 0.962, RMSE = 0.607), and was used for final results.
We estimated the probability that rotavirus vaccination was cost-saving using a mixed effects logistic regression as described in Supplementary Section 6. The model was fit subsequent to the meta-regression and included the same final set of covariates, but used the full data set with 1,418 results instead of the subset of the data selected in the trimming stage of the analysis. As with the meta-regression, we estimate a random intercept for each article.
We use the estimates from the meta-regression and logistic regression to predict both the ICER and probability the vaccine was cost-savings for 195 countries as a function of country-level covariates. The predicted ICER with adjustment is the product of the predicted ICER and (one minus the cost-saving probability). The predicted cost-saving probabilities were small, as were the adjustments to the predicted ICERs.
Our predictions assume 90% vaccine coverage (the GAVI target), lifetime time horizon, healthcare payer perspective, 3% burden and cost discount rates, monovalent vaccine type, and DALYs as the outcome measure. Our estimate of country-specific vaccine efficacy comes from GBD [bib_ref] Global burden of 369 diseases and injuries in 204 countries and territories,..., Vos [/bib_ref]. Because the rotavirus vaccine does not have a single market price, we used the cost of all required doses based on the 2017 cost per dose reported in the WHO's Market Information for Access (MIA4) and aggregated by Linksbridge(see Supplementary Section 7). We used the listed UNICEF price for the 57 Gavi-eligible countries, and Pan American Health Organization (PAHO) for 22 countries eligible for PAHO Revolving Fund.
We predict both the mean ICER and 95% UI, which are then adjusted by the probability that the intervention is cost-saving. We report two ratios: 1) the adjusted mean predicted ICER to GDP per capita, and the upper bound of the UI of the adjusted predicted ICER to GDP per capita. While GDP per capita has been criticized as a cost-effectiveness threshold as outdated [bib_ref] Cost-effectiveness thresholds: pros and cons, Bertram [/bib_ref] [bib_ref] Thresholds for the costeffectiveness of interventions: alternative approaches, Marseille [/bib_ref] , and not an accurate measure of a population's preference for spending on health improvements [bib_ref] Understanding and improving the one and three times GDP per capita cost-effectiveness..., Robinson [/bib_ref] , the ratios may provide a useful screening tool [bib_ref] Understanding and improving the one and three times GDP per capita cost-effectiveness..., Robinson [/bib_ref] and place the results in the context of the country's economy.
# Results
The sub-Saharan Africa region reported more results than any other region (4 4 0), followed by high-income countries (3 3 9), Southeast Asia, East Asia and Oceana (1 7 7), Latin America and Caribbean (1 3 7), Central Europe, Eastern Europe, and Central Asia (1 0 7), South Asia (83), and North Africa and the Middle East (62). [fig_ref] Table 1: Descriptive statistics of cost-effectiveness results and peer-reviewed articles on rotavirus vaccines included... [/fig_ref]. A total of 521 ICERs were reported from the healthcare payer perspective, followed by 478 from the limited societal, 329 societal, and 17 from the healthcare sector perspectives. Most ICERs (1,029) measured health outcomes with DALYs, and used a 3% discount rate (1,058). Median vaccine efficacy across studies was 82% (Interquartile Range [IQR] 64-87). Median vaccine coverage was 75% (IQR 70-90), and median vaccine cost was $7.41 (IQR $3.61 -$59.30) which included all doses.
In meta-regression analysis, vaccine efficacy and log-DALYs per capita were negatively associated with the log-ICER, and logvaccine cost and log-GDP per capita were positively associated with it (Supplementary . A 10 % increase in burden would reduce the ICER by 4.2%, and a 10% increase in vaccine price would increase the ICER by 6.4%. Including vaccine efficacy in the analysis slightly improved model fit; R 2 increased from 0.960 to 0.961, while Root mean-squared error fell from 0.613 to 0.607. The random effects variance for the model including efficacy was also lower (0.576 including efficacy versus 0.593). We did not observe similar improvements for models of effective coverage (Supplementary . We report estimates for the model including efficacy in the main text, and estimates for the model without efficacy in .
The ranges of observed ICERs within countries were broad due to differences in methods and intervention characteristics. The ranges of observed ICERs within super-regions were broad for the same reasons as well as differences in rotavirus burden and GDP among countries. Despite these differences, the final metaregression model fit the data well, as shown in the comparison of the observed and fitted ICERs globally and by super-region [fig_ref] Figure 2: Observed log-ICER vs fitted log-ICER overall and by GBD super-region [/fig_ref].
Among the seven GBD super-regions [fig_ref] Table 2: Predicted incremental cost-effectiveness ratios aggregated to super-region level compared to range of... [/fig_ref] , Sub-Saharan Africa and South Asia had the lowest population-weighted mean ICERs among GBD super-regions. Among the 46 countries making up the sub-Saharan African region, the mean predicted ICER was $251 per DALY averted (90% UI 38-903), while for the 5 countries making up the South Asia region, the mean predicted ICER was $294 per DALY averted (95% UI 45-1 062). The 34 countries making up the high-income region had the highest ICER ($40 914 per DALY averted; 95% UI 6 382-153 130).
Predicted ICERs varied across countries [fig_ref] Table 3: Predicted incremental cost-effectiveness ratios by country adjusted for cost-saving probabilities [/fig_ref] ; [fig_ref] Figure 3: Predicted ICERs from meta-regression analysis by country [/fig_ref]. Among all 195 countries, the lowest mean adjusted ICERs were observed in Central African Republic (2017 US$ 85 per DALY averted; 95% UI 13-302), and Chad ($120 per DALY averted; 95% UI 18-428), which have the two highest burdens worldwide. The highest mean predicted ICERs were observed in the US ($70,599 per DALY averted; 95% UI 11,030-263,858), and Luxembourg ($46,158 per DALY averted; 95% UI 7,256-169,808), which are in the bottom 8% of rotavirus burden worldwide. Predicted ICERs may be outside the range of observed ICERs, due to differences in parameters in the published and predicted results, such as parameters such as vaccine cost and efficacy.
Among countries eligible for support from Gavi, The Vaccine Alliance, the mean ICER was $255 per DALY averted (95% UI: $39 -$918). Among countries eligible for the PAHO revolving fund, the mean ICER was $2,464 per DALY averted (95% UI: $382 -$3,118).
Considering these results in the context of each country's economy, 21 countries had adjusted ICERs with upper UIs less than onehalf times GDP per capita [fig_ref] Figure 3: Predicted ICERs from meta-regression analysis by country [/fig_ref]. Among them, 15 were in the sub-Saharan Africa region, 4 in Southeast Asia, East Asia, and Oceana, one in South Asia (India), and one in North Africa and the Middle East (Sudan). An additional 43 countries had upper UIs that were between 0.5 and one times the GDP per capita, including 17 in sub-Saharan Africa, 13 in Latin America and the Caribbean, six in Southeast Asia, East Asia, and Oceana, three in North Africa and the Middle East, two in South Asia, and one each in high-income (Norway) and Central Europe, Eastern Europe, and Central Asia (Tajikistan).
# Discussion
To our knowledge, this is the most comprehensive metaregression analysis of CEA of rotavirus vaccination conducted, building on previous work that focused on 29 LICs and LIMCs [bib_ref] Systematic Review and Meta-Analysis of Cost-effectiveness of Rotavirus Vaccine in Low-Income and..., Haider [/bib_ref] , and previous efforts to leverage a meta-regression approach to synthesize all available evidence on HPV and provide costeffectiveness estimates for 195 countries worldwide [bib_ref] Costeffectiveness of HPV vaccination in 195 countries: A meta-regression analysis, Rosettie [/bib_ref]. An asterisk (*) denotes that the total number of articles may exceed 68 since some articles examined multiple regions, vaccine characteristics, and cost-effectiveness analyses characteristics. QALY = quality-adjusted life-year, DALY = disability-adjusted life-year, IQR = interquartile range. Unlike previous approaches, our approach facilitates the transferability of published estimates across settings, while accounting for differences between settings. Currently, a Ministry of Health in a country where there are no cost effectiveness estimates available may look to estimates from neighboring countries. However, these estimates may vary considerably within a country, and there would be no quantitative way to adjust them to account for differences in economic activity, burden of disease, vaccine efficacy or cost. For example, there are no published cost-effectiveness estimates in Equatorial Guinea, while neighboring Cameroon has estimates that range from $10 to $87 per DALY averted [fig_ref] Figure 2: Observed log-ICER vs fitted log-ICER overall and by GBD super-region [/fig_ref]. Relying on estimates from Cameroon would likely underestimate the costeffectiveness of the intervention because of differences in GDP per capita ($15,803 in Equatorial Guinea vs $1,671 in Cameroon) and corresponding vaccine costs ($22.53 vs $4.61, respectively). This is consistent with the findings from Jit et al, who noted that different vaccine prices contributed to variability in CEA results [bib_ref] The costeffectiveness of rotavirus vaccination: Comparative analyses for five European countries and..., Jit [/bib_ref]. Using a meta-regression approach and accounting for the drivers of variability in cost-effectiveness, our estimates are $1,714 per DALY averted (95% UI: $267 -$6,306) in Equatorial Guinea.
Our approach also facilitates decision-making for policy-makers in a country equipped with multiple estimates. For example, Bangladesh's 19 published estimates ranged from $23 to $1,543 per DALY averted owing to differences in modeling assumptions that cannot be easily synthesized in a policy-making setting. Our approach quantifies the effects of important drivers of variability in ICERs, producing an estimated ICER and simple linear equation that allows policy-makers to leverage the estimated effects and Our results provide encouraging evidence in favor of introducing rotavirus vaccination worldwide. All countries had predicted mean ICERs below three times GDP per capita, with the exception of the United States and Somalia. Taking uncertainty into account, the upper bound of the UI of the predicted ICER was below one times GDP per capita in 64 countries, but exceeded three time GDP per capita in five countries (the US, UK, Portugal, Greece, and Somalia).
Nevertheless, much progress needs to be made in order to meet the Global Vaccine Action Plan's target of 90% coverage by 2030. Of the 195 countries considered here, only 24 had reached 90% coverage in 2017 [bib_ref] Measuring routine childhood vaccination coverage in 204 countries and territories, 1980-2019: a..., Galles [/bib_ref]. Twelve countries (Zambia, Eritrea, Senegal, The Gambia, Uzbekistan, Sao Tome and Principe, Ghana, Burkina Faso, Mozambique, Burundi, Rwanda, and Nicaragua) are LICs, two (Palestinian Territories and Morocco) are LMICs, five (Armenia, Fiji, Guyana, Mauritius, Paraguay) are UMICs, and five (Luxembourg, Norway, Saudi Arabia, Bahrain, and Qatar) are HICs. These countries all have different economies but a shared commitment to rotavirus vaccination, suggesting that the large scale-up and delivery of vaccinations across the rest of the world is possible.
Our work here has a number of limitations. First, we are unable to account for all differences between the studies used in our metaanalysis. Doing so would require more detailed descriptions of methodologies used in each article. Second, although we used the full sample of results to estimate the probability that the result was cost-saving, and adjusted all predicted ICERs by the predicted cost-saving probability, we did not propagate the correlation between the logistic regression and the meta-regression models. The probability that the rotavirus vaccine was cost-saving at current vaccine prices was low however, meaning that the correlation would not have substantially changed our results. Third, we applied Lasso in Stage 3 of the analysis framework to identify essential features from a subset instead of all covariates. Future research may improve the estimates by extracting sensitivity analyses for a more covariates, applying selection criteria to the variables in Stage 1, or improving methods for variable selection in the presence of collinearity in Stage 3. Fourth, we did not assess validity of our predictions in terms of model calibration and discrimination. CEA researchers working with decision analytic models have initiated research on predictive validity, and this field will have implications for the quality of the published results that are the data for in our meta-regression analysis. Fifth, uncertainty of reported ICERs is largely based on sensitivity analyses of different assumed values for input parameters.
Despite these limitations, our work is the most comprehensive set of country-specific estimates of rotavirus vaccination. In producing these estimates, our meta-regression approach synthesizes all available evidence, quantifies uncertainty to facilitate decisionmaking, and incorporates country-specific estimates of rotavirus vaccine efficacy. Our results support introducing or expanding rotavirus vaccination, which remains a critical step in reducing rotavirus burden.
## Declaration of competing interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
[fig] Figure 1: Study flow diagram of study selection in logistic and meta-regression models. * For articles with more than one ratio, some ratios may be excluded and others included. CEA = cost-effectiveness analysis, ICER = incremental cost-effectiveness ratio, QALY = quality-adjusted life-year, DALY = disability-adjusted life-year. [/fig]
[fig] Figure 2: Observed log-ICER vs fitted log-ICER overall and by GBD super-region.Fig. 2legend. Observed log-ICER vs fitted log-ICER overall and by GBD super region. The minimum and maximum values for Peru (PER), Bangladesh (BGD), and Cameroon (CMR) are shown to highlight differences in observed log-ICERs across studies within a given country. ICER = incremental cost-effectiveness ratio. [/fig]
[fig] Figure 3: Predicted ICERs from meta-regression analysis by country.Fig. 3 legend. (A) Predicted ICERs from meta-regression analysis by country in 2017 USD per DALY averted; (B) Predicted ICERs relative to seven categories of GDP per capita ranging from <0.5 to >3.0 times GDP; C) GDP category in which the upper bound of the 95% UI falls. ICER = incremental cost-effectiveness ratio, UI = uncertainty interval, DALY = disability-adjusted life-year. [/fig]
[table] Table 1: Descriptive statistics of cost-effectiveness results and peer-reviewed articles on rotavirus vaccines included the analysis.Number of ratios reported in Tufts registries plus sensitivity analyses extracted (%)Number of articles reported in Tufts registries plus sensitivity analyses* (%) [/table]
[table] Table 2: Predicted incremental cost-effectiveness ratios aggregated to super-region level compared to range of input data by super-region from Tufts registries. [/table]
[table] Table 3: Predicted incremental cost-effectiveness ratios by country adjusted for cost-saving probabilities. [/table]
|
Accuracy of toric intraocular lens implantation using automated vs manual marking
Background: Accurate alignment of toric intraocular lens (TIOL) to steep corneal astigmatic axis is important to achieve effective postoperative results. The authors compare the accuracy of astigmatism correction using automated and manual marking in TIOL implantation during cataract surgery. Methods: One hundred thirty-two eyes with nuclear density from Grade 2 to 4 were randomly subdivided into 2 groups (automated and manual marking). All patients underwent manual marking and the steep axis was compared to SensoMotoric Instruments (SMI). After phacoemulsification, 62 patients underwent toric IOL implantation using the SMI and 70 patients underwent toric IOL implantation using manual marking. Intraoperative measurement was the steep axis difference. Clinical measurements included preoperative and postoperative best corrected visual acuity (BCVA), and TIOL axis.Results: The intraoperative steep axis difference between SMI and manual marking was 7.86 ± 6.4 degrees. The difference between the preoperative steep axis and the postoperative TIOL axis using SMI (3.63 ± 1.12 degrees) was significantly lower than that using manual marking (8.29 ± 2.23 degrees) (P < 0.05).Conclusions:The steep axis measurements may be different when using SMI vs. manual marking. The SMI is more accurate than manual marking for TIOL implantation during cataract surgery.
# Background
With developments in cataract surgery, only correction of refractive errors with removal of the opaque lens and implantation of an intraocular lens is not the major goal of the cataract surgery. An estimated 35% of cataract patients have 1.00 diopter (D) or more than 1.00 diopter of preexisting corneal astigmatism, and 15 to 20% of them have 1.5 D or more than 1.5 D of corneal astigmatism.
Limbal corneal relaxing incisions and implantation of a toric intraocular lens (IOL) have been used for correction of astigmatism in cataract patients.
Because 1°of off-axis rotation results in a loss of up to 3.3% of lens cylinder power, marking the accurate axis is most important for successful toric IOL implantation. Several marking methods have been used, for example, marking at the 3-, 6-, and 9-o'clock positions using a toric reference marker and at the 3 and 9 o'clock positions using a horizontal slit beam in the slit lamp. An anterior segment photograph was used to identify several reference vessel points and axis marking points. Digital overlay imaging was also used to evaluate the alignment of toric IOL.
Recently, an eye tracking systems called SensoMotoric Instruments (SMI, Teltow, Germany) has been used to visualize the steep corneal axis through the operating microscope during toric IOL implantation.
To the best of our knowledge, there are no studies evaluating the efficacy of SMI in cataract surgery with toric IOL implantation.
Here in, we compare the outcomes of toric IOL implantation marked with SMI and manual marking 2technique in the cataract patients with corneal astigmatism.
# Methods
This prospective randomized study comprised 132 eyes of 132 patients with cataract and coexisting corneal astigmatism more than 1.5 diopters (D) who were randomly assigned to undergo phacoemulsification and posterior chamber toric intraocular lens implantation at the Seoul St. Mary's Hospital between February 2014 and December 2017. The study protocol followed the guidelines of the Declaration of Helsinki and the institutional review board of Bucheon St. Mary's Hospital. Patients provided written informed consent after receiving an explanation of the surgical systems used in the study.
One hundred thirty-two eyes were randomly divided into two groups (SMI and manual marking). All patients underwent manual marking and the steep axis was compared to SMI intraoperatively. After phacoemulsification, 62 patients underwent toric IOL implantation using SMI (group 1) and 70 patients underwent toric IOL implantation using manual marking (group 2).
Corneal pathology, pseudoexfoliation, history of ocular trauma and intraoperative complications such as posterior lens capsule rupture, lens dislocation, and ocular inflammation were excluded.
Uncorrected visual acuity (UCVA), best corrected visual acuity (BCVA), keratometer, refractometer, and corneal topography by a Scheimpflug imaging system (Pentacam; Oculus, Wetzlar, Germany) were assessed preoperatively.
## Intraoperative & postoperative measurements
Intraoperative measurements included the steep axis difference between SMI and manual marking. The postoperative parameters measured at 1-day postoperatively, 1-month postoperatively, and 2-months postoperatively were uncorrected visual acuity (UCVA), best corrected visual acuity (BCVA), keratometer, refractometer, and corneal topography by a Scheimpflug imaging system (Pentacam). The visual acuity measurements had been recorded by masked personnel.
## Surgical technique
For SMI marking, reference image of the steep axis was taken using a reference unit in the outpatient department. The reference image and data of the steep axis were transferred to the SMI system in the operating room.
For manual marking, the axis-marking procedure was performed with topical anesthesia. The patient was seated at the surgical table and instructed to gaze at a distant target. Using a toric reference marker (AE-2793S; ASICO LLC, Westmont, Illinois), the corneal limbus was marked at the 3-, 6-, and 9-o'clock positions. Next, with the patient lying on the surgical table, the steep axis was marked using a Mendez ring (K3-800; Katena, Denville, New Jersey). The steep axis difference between SMI and manual marking was measured in all 132 eyes.
Phacoemulsification was performed by the same surgeon (C.K.J.) using the Intrepid Infiniti system as described previous publication. After phacoemulsification, sodium hyaluronate 1% (Healon®) was injected into the anterior chamber, and a toric intraocular lens (IOL) was inserted in the capsular bag. A SN6AN (Alcon, Ft. Worth, TX, USA) was implanted in the capsular bag using an injector system. In 62 randomized patients, the axis marks on the toric IOLs were aligned with the steep corneal meridian determined by the SMI, and in another 70 randomized patients, it was determined by manual marking under the protection of an ophthalmic viscosurgical device (OVD), which was subsequently removed through aspiration. The wound was not sutured.
# Statistical analysis
All data are expressed as the mean ± standard deviation. Pairwise comparisons of treatment group categorical variables were performed using the Mann-Whitney U test and continuous variables were analyzed using the unpaired t test. The analyses were performed using SPSS for Windows software (version 16.0, SPSS, Inc.). A P value of less than 0.05 was considered statistically significant.shows the characteristics of the patients in each group. There were no statistically significant differences between the two groups according to age, preoperative mean astigmatism, and uncorrected and best corrected visual acuity (UCVA & BCVA) (P > 0.05).
# Results
## Preoperative parameters
## Intraoperative parameters
The intraoperative steep axis difference between SMI and manual marking was 7.86 ± 6.4 degrees. There was a significant axis difference between the two groups (P < 0.05) .
## Postoperative tiol axis differences
The difference between the preoperative steep axis and the 1-day postoperative TIOL axis using SMI (3.63 ± 1.12 degrees) was significantly lower than that using manual marking (8.29 ± 2.23 degrees) (P < 0.05) . The axis difference of group 1 (4.75 ± 1.37 degrees) was significantly lower than that of group 2 at the postoperative 2 months (8.93 ± 2.17 degrees) (*: P < 0.05). There was no significant difference in TIOL rotation between the two groups (P > 0.05) .
## Uncorrected and best corrected visual acuity (log mar)
Uncorrected visual acuity (logMAR) in the SMI group was significantly higher than that in the manual marking group (P < 0.05). But, there was no significant difference in the BCVA between the two groups (P > 0.05).
# Discussion
Microincision and small incision cataract surgery enables the almost astigmatically neutral phacoemulsification incisions. Newly developed intraocular designs and materials have improved the stability and predictability of Intraocular Lenses (IOLs).
Astigmatic error has become the most important cause of low uncorrected visual acuity, as the phacoemulsification technique has improved. Astigmatism less than 0.5 D did not degrade visual acuity after cataract surgery. Patients with > 0.75 D of corneal astigmatism had better visual outcome with implantation of toric IOLs compared to monofocal IOLs.
Because of the innovations in IOL technology and phacoemulsification technique, implantation of toric IOLs is the best method for correction of large corneal astigmatism more than 1.5 Diopters in patients requiring cataract surgery.
Accurate alignment of toric IOL to steep corneal astigmatic axis is important to achieve effective postoperative results. Inaccurate alignment of the toric IOL occurs due to wrong alignment during the operation or postoperative IOL rotation.
Many methods are used to mark the steep corneal astigmatic axis. When the patient is sitting, the preoperative marking of the horizontal corneal meridian is the most important step. Marking of the horizontal meridian can be done by slit lamp-assisted marking with a horizontal slit beam, a pendular marker, and a nonpendular marker.
Nowadays, digital image guidance such as SMI ensures accurate toric IOL alignment without manual marking during cataract surgery.
Elhofi AH et al. compared the clinical outcome of digital and manual marking for toric intraocular lens (IOL) alignment. The mean postoperative uncorrected distance visual acuity (UCDVA) for the digitalmarking group (0.12 + 0.12 logMAR) was higher than Data represent mean ± standard deviation SMI: SensoMotoric Instruments There was no statistically significant difference in initial characteristics between the two groups by the unpaired t test (P > 0.05, unpaired t-test) Intraoperative astigmatic axis difference between SensoMotoric Instruments and manual marking. a: dot line-SMI axis, thick line-manual axis. b: The difference in the steep astigmatic axis is 7.86 ± 6.4 degrees. There was a significant axis difference between the two groups (P < 0.05 by the Mann-Whitney U test) that for the manual-marking group (0.18 + 0.14 logMAR) (P = 0.104). The mean deviation from targeted induced astigmatism (TIA) for the digital-marking group (0.10 + 0.08 D) was lower than that for the manualmarking group (0.20 + 0.14 D) (P = 0.001). Also, the mean postoperative toric IOL misalignment measured by the slitlamp for the digital-marking group (2.48 + 1.968) was lower than that for the manual-marking group (4.338 + 2.728) (P = 0.003). However, there is no study comparing digital-marking and manual-marking in the same patients.
In our study, all patients underwent SMI and manual marking, and we compared manual marking to SMI intraoperatively. We calculated the mean axis difference between the two methods. There was a significant axis difference (7.86 ± 6.4 degrees) between the two groups (P < 0.05) .
The accuracy of marking steep axis using SMI was significantly better than that using manual marking after 1 day and 2 months (Table 2) (P < 0.05). The visual acuity (logMAR) in the SMI group was also significantly better than that in the manual marking group at the postoperative 2 months (P < 0.05).
Atchison DA et al. described that small levels of crossed cylinder blur produces losses in visual acuity that are dependent on the cylinder axis. The difference of the uncorrected visual acuity (logMAR) between two groups at the postoperative 2 months may occur because of the axis difference between the preoperative steep axis and the postoperative TIOL axis.
To the best of our knowledge, this is the first study comparing manual marking and SMI in the same patients during the operation. Our results showed that SMI causes a large difference in the steep astigmatic axis compared to manual marking. Also, the postoperative effectiveness of the toric IOL axis and uncorrected visual acuity for SMI were better than those for manual marking. However, there was no significant difference in TIOL rotation and best corrected visual acuity between the two groups (P > 0.05).
The only limitation of this study was that we provided the results obtained by a single surgeon. A multivariate evaluation should be performed in the future.
# Conclusions
Accurate marking of the steep astigmatic axis is important to achieve accurate alignment of the toric IOL and good postoperative uncorrected visual acuity. SMI would be a better instrument that provides accurate preoperative marking, intraoperative toric IOL alignment, and better postoperative results. Axis difference in Toric Intraocular Lens (TIOL) between the aimed and postoperative axis 1 day after cataract surgery. The difference between the preoperative steep axis and the 1 day postoperative TIOL axis using SMI was significantly lower than that using manual marking (P < 0.05 by the Mann-Whitney U test) Data represent mean ± standard deviation SMI SensoMotoric Instruments Axis difference †: The difference between the preoperative steep axis and the postoperative TIOL axis The axis difference of group 1 was significantly lower than that of group 2 (*: P < 0.05). And, the uncorrected visual acuity (logMAR) of group 1 was significantly higher than that of group 2 (*: P < 0.05) Axis rotation of the Toric Intraocular Lens (TIOL) between postoperative 1 day and 2 months after cataract surgery. There was no significant difference in TIOL rotation between SensoMotoric Instruments and manual marking (P > 0.05 by the Mann-Whitney U test) |
Automated diagnosis of optical coherence tomography imaging on plaque vulnerability and its relation to clinical outcomes in coronary artery disease
This study sought to develop a deep learning-based diagnostic algorithm for plaque vulnerability by analyzing intravascular optical coherence tomography (OCT) images and to investigate the relation between AI-plaque vulnerability and clinical outcomes in patients with coronary artery disease (CAD).A total of 1791 study patients who underwent OCT examinations were recruited from a multicenter clinical database, and the OCT images were first labeled as either normal, a stable plaque, or a vulnerable plaque by expert cardiologists. A DenseNet-121-based deep learning algorithm for plaque characterization was developed by training with 44,947 prelabeled OCT images, and demonstrated excellent differentiation among normal, stable plaques, and vulnerable plaques. Patients who were diagnosed with vulnerable plaques by the algorithm had a significantly higher rate of both events from the OCT-observed segments and clinical events than the patients with normal and stable plaque (log-rank p < 0.001). On the multivariate logistic regression analyses, the OCT diagnosis of a vulnerable plaque by the algorithm was independently associated with both types of events (p = 0.047 and p < 0.001, respectively). The AI analysis of intracoronary OCT imaging can assist cardiologists in diagnosing plaque vulnerability and identifying CAD patients with a high probability of occurrence of future clinical events.Despite the advances in medical science and healthcare practice, patients presenting with acute coronary syndrome (ACS) or with a history of myocardial infarction (MI) still have a significantly high rate of recurrent cardiovascular events. In particular, more than 20% of ACS patients who were successfully treated by percutaneous coronary intervention (PCI) will have secondary cardiovascular events within 3 years due to the worsening of previously treated culprit lesions or the progression of untreated nonculprit coronary plaques 1 . More recently, the PROSPECT II study showed that 13.2% of recent (within the past 4 weeks) MI patients who were treated by successful PCI had adverse events within 4 years 2 .A vulnerable plaque is defined as a precursor coronary atherosclerotic plaque that may cause future coronary events by plaque rupture and subsequent intraluminal thrombosis 3 . A vulnerable plaque is generally characterized as a plaque with typical pathohistological features such as a large amount of lipid accumulation, inflammatory cell infiltration, intraplaque angiogenesis, and the presence of a thin fibrous cap that covers a lipid core. Coronary angiography is, however, not able to characterize the plaque component and the vulnerability of the plaque. The use of intravascular imaging is one recent approach to this problem. Optical coherence tomography (OCT) is a novel intravascular imaging modality with a high spatial resolution that is comparable with the histopathology, OPEN www.nature.com/scientificreports/ and vulnerable plaques, which are diagnosed by OCT, have been shown to be associated with the subsequent progression of coronary artery stenosis, as well as future major adverse cardiac events 4,5 . For the secondary prevention of coronary artery disease (CAD), patients may have the opportunity to have their coronary lesions evaluated directly by intravascular imaging. However, the detailed analysis of OCT images in daily practice is difficult to perform because of the large numbers of OCT images that need to be evaluated. In addition, OCT image interpretation requires a highly skilled cardiologist due to the complex morphological configurations of the lesions and the coexisting imaging artifacts. To realistically address these problems, a possible solution is the application of artificial intelligence (AI) for the image analysis.The purposes of the study were (1) to develop a deep learning-based diagnostic algorithm for coronary plaque vulnerability by analyzing intravascular OCT images; (2) to test the diagnostic accuracy of the algorithm for plaque vulnerability; and (3) to investigate the relation between AI diagnosis of plaque vulnerability and clinical outcomes by using OCT-observed segments from patients with CAD.
# Methods
Study population. The study patients were recruited from the OCT clinical database that was obtained from three university hospitals in Japan: Kawasaki Medical School, Nara Medical University, and Wakayama Medical University. From these databases, patients who underwent intracoronary OCT imaging during coronary angiography or PCI from 2010 to 2019 were screened (n = 6625) and were enrolled in this study when they fulfilled the following criteria: (1) patients who underwent OCT imaging of nonculprit lesions if they were diagnosed to have clinically overt CAD and (2) patients who had OCT imaging of angiographically normal or minor stenosis (< 25%) segments if they had no significant coronary stenosis. Patients who did not have OCT imaging of nonculprit lesions (n = 3735) were excluded from this study. Patients with poor OCT image quality or with severe calcification (n = 1099) were also excluded from this study.
From the database, 1791 patients whose OCT examinations matched the inclusion criteria were eventually identified, and they were randomly assigned to dataset 1 for the development of the AI algorithm (n = 1689, training: validation = 8:2) and to dataset 2 for the testing of the developed program (n = 102). In dataset 1, 1450 patients who had complete long-term clinical outcome data for at least one month after the index OCT examination were enrolled in the follow-up study (dataset 3). The study flow chart is shown in [fig_ref] Figure 1: Study flow chart [/fig_ref].
This study complied with the Declaration of Helsinki and was approved by the institutional ethical review board of Kawasaki Medical School (IRB number: 3438-1). Written, informed consent was waived by the institutional review board (ethical review board of Kawasaki Medical School) because of the retrospective design of the study; however, informed consent was obtained in the form of an opt-out on the website. Those patients who www.nature.com/scientificreports/ rejected consent to the study were excluded. This study was called the TACUMI study (The Automated diagnosis of Coronary vUlnerable plaque using Medical artificial Intelligence).
## Selection of the coronary segment for analysis and oct image analysis. a nonculprit lesion in a
CAD patient was defined as a plaque with a diameter stenosis of 25-75% on angiographic visual estimation and that was at least 10 mm away from the stented lesions, and lesions from either the culprit vessel or nonculprit vessels were included. Normal segments from non-CAD patients were selected from the proximal coronary segments with angiographically normal or minor stenosis (< 25%). The OCT images of the patient's coronary arteries were recorded using an FD-OCT system (Dragonfly OPTIS and ILUMIEN OPTIS; Abbott Vascular, St. Paul, MN, USA) with a motorized catheter pullback system (36 mm/s). The plaque characterization of each OCT image was classified as normal, a stable plaque, or a vulnerable plaque using the previously established OCT criteria [fig_ref] Figure 2: Clinical OCT images and their corresponding raw images [/fig_ref] [bib_ref] Consensus standards for acquisition, measurement, and reporting of intravascular optical coherence tomography..., Tearney [/bib_ref]. The OCT definition of a normal arterial wall was characterized by a thin (less than 300 μm) intima containing no lipid or calcification. The OCT definition of a stable plaque was characterized as the presence of intimal thickening in fibrous, fibrocalcific plaques (calcification arc ≤ 90°) or in a thick-cap (fibrous cap thickness > 100 μm) fibroatheroma. The OCT definition of a vulnerable plaque was characterized by a plaque with fibrous cap thickness < 100 μm overlying a lipid-rich plaque (lipid arc > 90°) [bib_ref] Consensus standards for acquisition, measurement, and reporting of intravascular optical coherence tomography..., Tearney [/bib_ref]. The OCT images were labeled as being normal, a stable plaque, or a vulnerable plaque by 3 independent OCT expert cardiologists (Kume T, Soeda T, and Kubo T) who were blinded to the patient's information. A consensus reading was obtained when there was concordance among the 3 independent readers. The interobserver reliability of the OCT diagnosis among the 3 OCT expert cardiologists was high (kappa coefficient = 0.81, 0.86, 0.97, respectively). This study allowed one lesion per patient. Therefore, if the patient had not only a normal segment but also had stable or vulnerable plaques, the stable or vulnerable plaque was assigned as the representative plaque of the patient. If the www.nature.com/scientificreports/ patient had more than two independent plaques with both stable and vulnerable characteristics, the vulnerable plaque was assigned as the representative plaque of the patient. If the patient had more than two independent vulnerable plaques, the lesion with the most characteristics of a vulnerable plaque was assigned as the representative plaque of the patient. The mean number of OCT frames per patient was 26 ± 16 (5.3 ± 3.3 mm).
## Development of the deep learning models.
For the deep learning-based classification of the plaque characteristics, three different CNN (convolutional neural network)-based deep learning models, Inception-v3, DenseNet-121, and EfficientNet-B4, were pretrained with the ImageNet dataset 7 , and they were then trained on the prelabeled OCT images by fine-tuning. All of the models had one fully connected (FC) layer that was connected with the global average pooling, and the sizes of the FC layers were 2084, 1024, and 1280 for the Inception-v3, DenseNet-121, and EfficientNet-B4 models, respectively. The input data put into the deep learning model were resized raw OCT image data (299 × 299 × 1), and the output was the probabilities of the three types of plaques: normal, stable, and vulnerable. The details of the development protocols and the external validation are shown in the Supplemental Data.
Testing the diagnostic capability of the AI algorithm. The diagnostic capability of the developed AI algorithm for plaque characterization was compared to the assessments made by general cardiologists in 102 patients in dataset 2. A total of 1,173 OCT frame images (normal: 425, stable plaque: 374, vulnerable plaque: 374) were randomly rearranged. The AI algorithm and four general cardiologists classified the test images as normal, a stable plaque, or a vulnerable plaque. The diagnostic accuracy was calculated as the number of true positives and true negatives using the OCT expert cardiologists' diagnosis as the reference, and the value was divided by the total number of OCT images that were analyzed.
The AI diagnosis and its relation to clinical outcomes. The relation between the deep learning-based plaque characterization and the long-term clinical outcomes was retrospectively evaluated in 1450 patients (dataset 3) and was compared with those based on the diagnoses made by OCT expert cardiologists. As far as the long-term clinical outcomes, two independent parameters were configured. The first one was the presence of a coronary event from the OCT-observed segments, including the need for clinically driven revascularization and the angiographic progression of CAD with a diameter stenosis > 75%. Angiographic progression of CAD was defined as lesions with less than 75% stenosis at baseline that progressed to a stenosis of 75% or more at the follow-up. Another outcome parameter was a composite of the clinical events including cardiac death, noncardiac death, and any clinically driven coronary revascularization, including ACS and a recurrence of the ischemia.
Statistical analysis. The degree of agreement with the OCT labeling as normal, a stable plaque, or a vulnerable plaque among the 3 independent OCT expert cardiologists (interobserver variabilities) was quantified by the kappa test of concordance 8 . All of the variables were entered into a univariate analysis. Chi-squared tests were used for evaluating the categorical variables, and Student's t tests were used to evaluate the continuous variables. If significant differences were recorded among the groups on the univariate analysis, a post hoc analysis using Tukey's honestly significant difference test was used to determine the differences between the groups. The event-free data over the follow-up period were evaluated with a Kaplan-Meier analysis. Variables with a p value < 0.05 on the univariate analysis were included in the multivariate logistic regression analysis to identify the independent factors that were associated with the events from the OCT-observed segment and the composite of the clinical events. The areas under the receiver operating characteristic curves (AUCs) were used to evaluate the diagnostic ability of the model for the plaque characteristics. The AUC was calculated for one of the three classes (normal, stable, vulnerable) and for the other two classes. The AUC was also used to compare the predictive ability of events from the OCT-observed segment and the composite of the clinical events between an AI-diagnosed vulnerable plaque and an expert-diagnosed vulnerable plaque. The mean ± SD is reported for normally distributed data. A p value < 0.05 was considered significant. Statistical analysis was performed using JMP (version 14 for Windows; SAS Institute, Cary, NC, USA).
# Results
AI algorithm for diagnosing the OCT plaque characteristics. The classification accuracies of the three deep learning models (Inception-v3, DenseNet-121, and EfficientNet-B4) for dataset 2 were 91.2%, 92.4%, and 91.9%, respectively. The results for the AUC, Brier score, F1 score, log loss, and calibration plots for each model are shown in Supplemental [fig_ref] Table 1: Clinical characteristics of the patients in dataset 3 according to the classification... [/fig_ref] and Supplemental [fig_ref] Figure 4: Diagnostic accuracy of the developed AI algorithm [/fig_ref]. DenseNet-121, which performed the best on these measures, was used in the following analyses. When the majority decision method (Supplemental [fig_ref] Figure 2: Clinical OCT images and their corresponding raw images [/fig_ref] was applied to the prediction labels of dataset 2, the accuracy increased to 94.0%.
Gradient-weighted class activation mapping (Grad-CAM) was used to visualize the regions of interest for the AI algorithm in predicting a diagnosis [bib_ref] Visual explanations from deep networks via gradient-based localization, Selvaraju [/bib_ref]. [fig_ref] Figure 3: Grad-CAM analysis and t-SNE visualization of the last hidden layer for the... [/fig_ref] -f shows the representative Grad-CAM results for each class of plaque. In the figure, the attention level is indicated with a heatmap, and the level increases from blue to red in rainbow colors. The areas in red roughly correspond to the areas that doctors focus on while making a diagnosis. Furthermore, the image distribution of dataset 2 was visualized using t-distributed stochastic neighbor embedding (t-SNE) [fig_ref] Figure 3: Grad-CAM analysis and t-SNE visualization of the last hidden layer for the... [/fig_ref].
Testing of the diagnostic capability of the AI algorithm. A total of 1173 independent OCT frame images from dataset 2 were used to test the diagnostic capability of the developed AI algorithm, and the OCT images were compared with the diagnoses from the general cardiologists. The clinical profiles of dataset 2 were www.nature.com/scientificreports/ consistent with those of dataset 1 (age: 70.9 ± 10.9 vs. 68.0 ± 11.7 years old, male: 77.5% vs. 73.9%, respectively). Using the diagnosis made by the OCT expert cardiologists as the reference, the diagnostic accuracy of the algorithm was 94.0%, compared to an average of 83.8% among the 4 general cardiologists [fig_ref] Figure 4: Diagnostic accuracy of the developed AI algorithm [/fig_ref] Classification of the patients with the AI algorithm. A total of 1450 patients with complete outcome data (dataset 3) were classified into three groups based on the OCT plaque diagnosis by the AI algorithm: normal (n = 435), stable plaque (n = 465), and vulnerable plaque (n = 550). In this study population, the diagnostic accuracy of the developed algorithm was 90.6% using the diagnosis made by the OCT expert cardiologists as the reference. The baseline clinical characteristics of the study population based on the plaque diagnosis by the developed algorithm are summarized in [fig_ref] Table 1: Clinical characteristics of the patients in dataset 3 according to the classification... [/fig_ref]. Age, sex, hypertension, diabetes mellitus, dyslipidemia, current smoker, prior myocardial infarction, prior PCI, and the index clinical presentation were significantly different among the three groups. The age of patients with vulnerable plaques was significantly higher than that of patients with a normal OCT registered lesion. The estimated glomerular filtration rate (eGFR), LDL-Cho, HDL-Cho, and the medications at baseline were significantly different among the three groups. Patients with vulnerable plaques had higher levels of LDL-Cho and lower levels of HDL-Cho than the patients with a normal OCT registered segment. These clinical characteristics were similar to those diagnosed by OCT expert cardiologists (Supplemental [fig_ref] Table 2: Events from the OCT-observed segments and the composite of clinical events [/fig_ref]. The analyzed lesions were identified in the left anterior descending artery (49.3%), in the left circumflex artery (18.2%), in the right coronary artery (31.0%), and in the left main trunk (1.5%).
The relation between AI diagnosis of plaque vulnerability and clinical outcomes. The median duration between the index OCT examination and the determination of the clinical outcome was 530 days (interquartile range 310-1105 days). The numbers of OCT-observed segment events and the composite of the clinical events were 31 and 187, respectively. The 31 events from the OCT-observed segments consisted of clinically driven coronary revascularization (n = 18: 3 ACS and 15 recurrent angina pectoris) and angiographic progression (diameter stenosis > 75%) of the predetermined segment (n = 13). The Kaplan-Meier analysis showed that the patients with AI-diagnosed vulnerable plaques had significantly higher cumulative rates of events from the OCT-observed segments than It is interesting to note that for stable and vulnerable plaques, the attention is on a part of the fibrous cap overlying the lipid component, whereas for normal plaques, the attention is given to the whole vessel wall in the attention map output by Grad-CAM. The high-dimensional features obtained by DenseNet-121 are dimensionally compressed by t-SNE and are represented as two-dimensional data (g). A total of 1,173 images of the test dataset that was obtained from 102 patients are displayed. Normal (N), stable plaque (S), and vulnerable plaque (V) images are represented as green, yellow, and red dots, respectively. The normal, stable plaque, and vulnerable plaque clusters are clearly observed. Some stable plaque data are included in the normal cluster, which is consistent with the results of the confusion matrix [fig_ref] Figure 4: Diagnostic accuracy of the developed AI algorithm [/fig_ref] www.nature.com/scientificreports/ the patients with AI-diagnosed normal and stable plaques [fig_ref] Figure 5: Clinical outcomes in patients with CAD diagnosed by the AI algorithm [/fig_ref]. Sixteen lesion-related events occurred in the patients with AI-diagnosed vulnerable plaques, whereas two events occurred in patients with AI-diagnosed stable plaques and normal plaques during the follow-up period. On the multivariate logistic regression analyses, only the AI diagnosis of an OCT vulnerable plaque was independently associated with the events from the OCTobserved segments (p < 0.001) [fig_ref] Table 2: Events from the OCT-observed segments and the composite of clinical events [/fig_ref]. Furthermore, the patients with AI-diagnosed vulnerable plaques had significantly higher cumulative rates for the composite clinical outcomes than the patients with AI-diagnosed normal and stable plaques [fig_ref] Figure 5: Clinical outcomes in patients with CAD diagnosed by the AI algorithm [/fig_ref] , and AI-vulnerable plaques and eGFR were independently associated with the composite of the clinical events (p = 0.047 and p < 0.001, respectively) [fig_ref] Table 2: Events from the OCT-observed segments and the composite of clinical events [/fig_ref].
The Kaplan-Meier analyses that were obtained by the OCT diagnosis of the AI algorithm were comparable to those diagnosed by the OCT expert cardiologists . The AUC value of the AI vulnerability for events from the OCT-observed segments was 0.719, indicating a moderate accuracy, and this accuracy (0.742) was comparable to the AUC of the OCT expert-vulnerable plaques. In addition, the AUCs of the AI vulnerability and the expert vulnerability for the composite clinical events were 0.590 and 0.592, respectively.
# Discussion
The salient results of the present study are as follows: (1) the deep learning-based AI algorithm that was developed in this study demonstrated excellent differentiation among normal, stable plaques, and vulnerable plaques, and its diagnostic capability was better than that of general cardiologists; and (2) CAD patients with AI-classified vulnerable plaques had significantly higher rates of coronary lesion-related and clinical events than patients with AI-classified normal coronary arteries and stable plaques.
Patients with CAD have a high risk for recurrent cardiovascular events, not only in lesions that were previously treated with PCI, but also from untreated, nonculprit lesion sites 1,2 . For lesions that were previously treated with PCI, improvements in the drug-eluting stents and antiplatelet therapy have significantly decreased the risk for PCI-related secondary events [bib_ref] Drug-eluting or bare-metal stents for percutaneous coronary intervention: A systematic review and..., Piccolo [/bib_ref]. On the other hand, the rate of cardiovascular events from untreated nonculprit lesions remains very high, despite intensive strategies for risk modification. Accordingly, the effective identification of patients who are at a high risk of developing events from nonculprit lesions is crucial regarding the long-term care of CAD patients. OCT is a high-resolution intravascular imaging technique that uses www.nature.com/scientificreports/ www.nature.com/scientificreports/ near-infrared spectroscopy, and it can differentiate plaque instability in vivo [bib_ref] Clinical significance of lipid-rich plaque detected by optical coherence tomography: A 4-year..., Xing [/bib_ref] [bib_ref] Thin-cap fibroatheroma and microchannel findings in optical coherence tomography correlate with subsequent..., Uemura [/bib_ref]. However, the routine analysis of coronary lesions with OCT has practical limitations because of the large numbers of OCT images that need to be evaluated, and OCT image interpretation requires substantial experience. To address this practical issue, we sought to develop an AI algorithm, as an accurate and automatic diagnostic tool, for the vulnerability of nonculprit coronary plaques and evaluate the relation between AI diagnosis of plaque vulnerability and clinical outcomes in patients with CAD. Various medical image analyses using deep learning have been studied as diagnostic aids for clinicians, and these imaging modalities include CT images, retinal fundic images, pathological images, mammography images, and skin images [bib_ref] International evaluation of an AI system for breast cancer screening, Mckinney [/bib_ref] [bib_ref] Dermatologist-level classification of skin cancer with deep neural networks, Esteva [/bib_ref]. Recently, Liu and colleagues reported the application of an AI program to OCT for coronary plaque characterization [bib_ref] Automated detection of vulnerable plaque for intravascular optical coherence tomography images, Liu [/bib_ref]. Based on their relatively small dataset, they could detect vulnerable plaques with a good diagnostic accuracy (detection quality score 88.46%) using cross-sectional images (generated by a polar coordinate transformation) of the vessels. More recently, Min and colleagues reported the application of an AI program to OCT for identifying thin-cap fibroatheroma using 602 coronary lesions from 602 angina patients and a developed AI program could accurately detect a thin-cap fibroatheroma (AUC = 0.96) [bib_ref] Detection of optical coherence tomography-defined thin-cap fibroatheroma in the coronary artery using..., Min [/bib_ref]. The present study used raw OCT data instead of cross-sectional transformation images. Although this process is not intuitive for humans, it is thought that information loss occurs due to the polar transformation. After a polar coordinate transformation, the information in an image is sparser in the outer part of the image than in the center. Therefore, the information loss is more significant in the outer part of the image. This can be complemented by bilinear or bicubic processing, but these techniques are not sufficient. When DenseNet-121 was trained using the polartransformed dataset, the accuracy for the test data (dataset 2) was 89.5%, which was 2.9 percentage points lower in accuracy than when using the raw data. In addition, by using a graph convolutional neural network (GCN), it is possible to use the raw data of OCT images as the polar coordinate data without any information loss. This is a promising method for the future. However, the OCT polar coordinate image used in the present study had 488,064 nodes (984 × 496) in the raw data and 89,401 nodes (299 × 299) in the resized images for input into the CNN model, which is a very large data size for GCN computation and requires a high computational cost for training. Another notable feature of this study was that OCT images were prelabeled based on the consensus of three OCT experts. According to such a novel approach to training, the AI algorithm demonstrated a very high diagnostic capability (92.4%), and this diagnostic rate was better than that of that from general cardiologists. The time between inputting a single OCT image into DenseNet-121 and obtaining the result was 35 ms; of course, with a more powerful GPU, it could be even faster. Such technology could support clinicians in decision-making and can reduce the burden of patient care. In addition, since this is a very short amount of time, there are potential applications for determining the plaque vulnerability by using a deep learning analysis of angiographic images in real time during coronary angiography and PCI.
In this study, the relation between AI diagnosis of plaque vulnerability and clinical outcomes of CAD patients was also tested. To the best of our knowledge, no previous study has used a well-defined cohort of OCT image samples with a well-established large clinical database to assess the clinical implications of medical AI with www.nature.com/scientificreports/ respect to its relation to future cardiac events. Several studies have attempted to use intravascular imaging for prognostic stratification. The PROSPECT I and II, AtheroRemo-IVUS, and LRP studies have shown the ability of grayscale IVUS, radiofrequency IVUS, and near-infrared spectroscopy IVUS to predict events in patients presenting with ACS and in stable patients undergoing an index PCI [bib_ref] A prospective natural-history study of coronary atherosclerosis, Stone [/bib_ref] [bib_ref] Identification of vulnerable plaques and patients by intracoronary near-infrared spectroscopy and ultrasound..., Erlinge [/bib_ref] [bib_ref] In vivo detection of high-risk coronary plaques by radiofrequency intravascular ultrasound and..., Cheng [/bib_ref] [bib_ref] Identification of patients and plaques vulnerable to future coronary events with near-infrared..., Waksman [/bib_ref]. These IVUS studies have demonstrated the importance of identifying high-risk lesions in patients with ischemic heart disease. We recently showed that vulnerable plaques that were diagnosed with OCT were associated with the subsequent progression of coronary stenosis, as well as future major adverse cardiac events [bib_ref] Clinical significance of lipid-rich plaque detected by optical coherence tomography: A 4-year..., Xing [/bib_ref] [bib_ref] Thin-cap fibroatheroma and microchannel findings in optical coherence tomography correlate with subsequent..., Uemura [/bib_ref]. In the present study, CAD patients with AI-diagnosed OCT vulnerable plaques had significantly worse lesion-specific outcomes. Furthermore, the presence of AI-diagnosed vulnerable plaques was also associated with a higher incidence of composite clinical outcomes, including noncardiac death, in CAD patients. The reason for this finding is unclear, but the comorbidities of CAD patients, such as chronic inflammatory diseases and malignant tumors, are known to correlate with enhanced plaque vulnerabilities in coronary arteries [bib_ref] Ischaemic heart disease and Cancer: Competing malignant conditions, Murphy [/bib_ref]. However, the prognostic value of AI vulnerability for the lesion-specific events was higher (AUC = 0.719) than that for the composite of the clinical events (AUC = 0.590). These data suggested that AI diagnosis of plaque vulnerability might be more useful for evaluating the prognostic value form the OCT-observed segments rather than the composite of the clinical events including cardiac death, noncardiac death and any clinically driven coronary revascularization. In the present study, patients with vulnerable plaques diagnosed by AI algorithm had a large number of coronary risk factors suggesting that this new technology could differentiate high-risk patients who should receive intensive medical management without the need for expertise and experience with OCT. Furthermore, we will be able to adapt this new technology to noninvasive imaging modalities, including coronary computed tomography and magnetic resonance imaging, for predicting cardiovascular events in primary care settings in the future. In addition, the automatic diagnosis of the plaque vulnerability by an AI program will provide useful information not only for the prognostic stratification but also for planning the PCI treatment strategies for patients in the cardiac catheterization laboratory. The plaque characteristics of the stent edge landing zone may affect the stent length selection. Interventional cardiologists prefer to position the edges of the stents into reference segments that are normal or that are at least less diseased segments. A stent landing edge should be avoided in segments with a vulnerable plaque even if they have an angiographically normal appearance because of the possible development of stent edge dissection and stent edge restenosis. Therefore, a quick diagnosis of the plaque vulnerability by an AI program is very useful in the cardiac catheterization laboratory when interventional cardiologists select the stent edge landing zone. From this perspective, integration of an AI program in an OCT imaging apparatus might automatically and efficiently promote the practice of PCI and provide risk stratification of individual CAD patients. Furthermore, the automatic determination and localization of vulnerable plaques with an AI program might help us change the PCI strategy, as well as the medical treatment, in individual CAD patients.
Limitations. Several limitations of the present study should be addressed. First, this was a retrospective cohort analysis, which may have selection bias. In addition, information about the clinical characteristics, including the lipid profiles, at the follow-up was lacking. However, a large number of CAD patients with complete baseline clinical characteristics were recruited. Second, OCT images with severe calcification (calcification arc > 90°) were excluded from this study because the impact of severe calcification detected by OCT in de novo lesions on clinical outcomes remains controversial. Third, lesions with OCT images that showed a fresh thrombus and/or plaque rupture indicating culprit lesions were excluded from this study. However, all of the registered lesions were not functionally tested using a pressure guidewire. Therefore, coronary angiograms have limitations in determining whether a lesion is culprit or nonculprit. Fourth, to test the diagnostic capability of the AI program, four general cardiologists who had different years of experience were recruited. Shibutani et al. recently evaluated the effect of the observers' years of experience on the interpretation of OCT images with reference to the histopathological findings [bib_ref] Interobserver variability in assessments of atherosclerotic lesion type via optical frequency domain..., Shibutani [/bib_ref]. They reported that the interpretation ability of OCT varied significantly among observers, and it is possible that there was a significant selection bias of the observers in the present study. Fifth, the prognosis value of the AI vulnerability for the clinical outcomes in CAD patients was not actually higher than that of the OCT expert vulnerability (AUC of AI vulnerability: 0.590, AUC of expert vulnerability: 0.592). However, the automatic analysis of numerous OCT images by AI-assisted computer systems without the input by expert cardiologists may be able to substantially distinguish patients who are at a high risk without additional workload to the medical staff and hospital resources. In this study, the predictive capability of the algorithm on the clinical outcomes was not validated in an external independent patient cohort, so a future prospective study is needed to confirm the present findings. Last, the clinical event rates, especially the incidence of acute coronary syndrome, were low in the present study.
# Conclusions
An AI program for analyzing intracoronary OCT imaging can assist cardiologists in diagnosing coronary plaque vulnerability and identifying CAD patients with a high probability for the development of future cardiovascular events. The clinical application of an AI system could reduce the medical workload and promote the individualized care of CAD patients based on its prognostic value of clinical outcomes.
## Data availability
All data generated or analyzed during this study are available from the corresponding author on reasonable request.
[fig] Figure 1: Study flow chart. OCT optical coherence tomography, AI artificial intelligence. Scientific Reports | (2022) 12:14067 | https://doi.org/10.1038/s41598-022-18473-5 [/fig]
[fig] Figure 2: Clinical OCT images and their corresponding raw images. OCT of a normal (a), a stable plaque (b), and a vulnerable plaque (c) diagnosed by OCT expert cardiologists. The clinical OCT images in (a-c) were generated from corresponding raw data (d-f) by a polar coordinate transformation. [/fig]
[fig] Figure 3: Grad-CAM analysis and t-SNE visualization of the last hidden layer for the three types of OCT imaging. Normal (a), stable (b), and vulnerable (c) OCT images. (d-f) Images of (a-c) overlaid with the attention map output by Grad-CAM. Expert cardiologists usually differentiate vulnerable plaque from stable plaque based on the thickness of the fibrous cap overlying the lipid component in OCT image. [/fig]
[fig] Figure 4: Diagnostic accuracy of the developed AI algorithm. Diagnostic accuracies of plaque vulnerability between the developed AI algorithm (a) and the general cardiologists (b) compared to the reference of the OCTexpert diagnosis. Receiver operating characteristic (ROC) curves for the AI algorithm for the differentiation of normal vessels (c), stable plaques (d), and vulnerable plaques (e). The dot plots represent the diagnostic accuracy of each general cardiologist. Scientific Reports | (2022) 12:14067 | https://doi.org/10.1038/s41598-022-18473-5 [/fig]
[fig] Figure 5: Clinical outcomes in patients with CAD diagnosed by the AI algorithm. Kaplan-Meier curves of the event-free survival from the OCT-observed segments (a) and the composite of the clinical events (b) according to classification by the AI algorithm. Patients with OCT-diagnosed vulnerable plaques showed higher cumulative rates for both endpoints than the patients with OCT-diagnosed normal and stable plaques.Scientific Reports | (2022) 12:14067 | https://doi.org/10.1038/s41598-022-18473-5 [/fig]
[table] Table 1: Clinical characteristics of the patients in dataset 3 according to the classification by the AI algorithm. The values are presented as the means ± SD. *p < 0.05 vulnerable vs. normal. † p < 0.05 stable vs. normal. ‡ p < 0.05 vulnerable vs. stable. PCI percutaneous coronary intervention, eGFR estimated glomerular filtration rate, HbA1c hemoglobin A1c, LDL-Cho low-density lipoprotein cholesterol, HDL-Cho high-density lipoprotein-cholesterol, BNP brain natriuretic peptide, ACEI angiotensin-converting enzyme inhibitor, ARB angiotensin II receptor blocker. [/table]
[table] Table 2: Events from the OCT-observed segments and the composite of clinical events. CI confidence interval, PCI percutaneous coronary intervention, eGFR estimated glomerular filtration rate, LDL-Cho low-density lipoprotein cholesterol, HDL-Cho high-density lipoprotein-cholesterol, ACEI angiotensin-converting enzyme inhibitor, ARB angiotensin II receptor blocker. [/table]
|
Internal Carotid Artery Dissection in Brazilian Jiu-Jitsu
Carotid artery dissection is a significant cause of stroke in young patients. It may be asymptomatic and go undiagnosed, or minimal transient manifestations may follow, commanding a higher index of suspicion than ordinarily exists to avoid misdiagnosis. Reported herein is a 27-year-old man who suffered extracranial internal carotid artery dissection while practicing a Brazilian Jiu-Jitsu submission maneuver. The patient's condition suddenly deteriorated one week later due to distal embolization and stroke. Despite endovascular treatment, with stenting of the cervical carotid artery, neurologic deficits remained. Of note, the objective in martial arts, which is to kill or incapacitate, has yet to be fully tempered in transitioning to sport. Brazilian Jiu-jitsu, a relatively new and fast-growing form of martial art, places emphasis on submission maneuvers. Related injuries are not common knowledge and are poorly described in the literature. This account is intended to shed light on the risk of this discipline. Through education and improved supervision, vascular injuries of this nature and the potentially lethal or disabling consequences may thus be prevented in young athletes.
# Introduction
Although carotid artery dissection is implicated in only 2.5% of all strokes, [bib_ref] Ischemic stroke in patients under age 45, Bogousslavsky [/bib_ref] it is among the leading causes of stroke in patients < 45 years old. [bib_ref] Ischemic stroke in patients under age 45, Bogousslavsky [/bib_ref] umentation as yet, but the submission maneuvers that are practiced may predispose to certain injuries. [bib_ref] Assessment of injuries during Brazilian Jiu-Jitsu Competition, Scoggin [/bib_ref] This report describes a circumstance in which internal carotid artery dissection was directly attributable to a Brazilian jiu-jitsu maneuver. Emergency computed tomography scan showed hypodensity in the left cerebral hemisphere .
## Case report
# Discussion
In the realm of martial arts, cervical vascular dissection typically involves the vertebral artery, having been reported with mixed martial arts, [bib_ref] Case report on vertebral artery dissection in mixed martial arts, Slowey [/bib_ref] karate, [bib_ref] De-afferented state syndrome (locked-in syndrome) following sudden cervical sprain trauma during a..., Pentore [/bib_ref] wrestling, 4) judo, [bib_ref] Vertebral-artery dissection following a judo session: a case report, Lannuzel [/bib_ref] kickboxing, [bib_ref] Endovascular treatment of a ruptured intracranial dissecting vertebral aneurysm in a kickboxer, Malek [/bib_ref] and kung-fu. [bib_ref] Vertebral artery dissection during Kung-Fu training, Pacei [/bib_ref] Although carotid artery dissection has also been described in taekwondo, [bib_ref] Traumatic internal carotid artery dissection associated with taekwondo, Pary [/bib_ref] karate [bib_ref] Acute aphasia and hemiplegia during karate training, Meairs [/bib_ref] and mixed martial arts, [bib_ref] Injury profile of mixed martial arts competitors, Mcclain [/bib_ref] direct bodily impact (i.e., kicks and punches) may be responsible, which is not characteristic of Brazilian jiu-jitsu.
As such, vulnerability of the extracranial carotid ar- Despite inherent challenges, early diagnosis of cervical arterial dissection improves patient prognosis.
## 2)
This patient scenario underscores the imperative for neurosurgeons and other sports physicians to consider arterial dissection when evaluating symptomatic athletes.
Headache and neck pain are common, and pain may be the only indication. [bib_ref] Cervical artery dissection: pathology, epidemiology and management, Kim [/bib_ref] Other signs and symptoms are cranial nerve palsy, Horner's syndrome, pulsatile tinnitus, ataxia, vertigo, and dizziness. 2)4) However,
given the frequency of asymptomatic carotid stenosis, patients with risk factors who practice contact sports should be screened for this condition as well. [bib_ref] Acute aphasia and hemiplegia during karate training, Meairs [/bib_ref] A thorough history is fundamental, but because the diagnosis is confirmed through imaging, early studies may be prudent in this patient subset. of intracerebral occlusions in an acute phase is also a safe and effective option, even following extracranial carotid artery stenting. [bib_ref] Emergent extracranial internal carotid artery stenting and mechanical thrombectomy in acute ischaemic..., Mishra [/bib_ref] It was contraindicated in this patient since he was beyond the 6-hour time window and had major involvement (over one-third) of the territory supplied by middle cerebral artery, heightening the risk of intracerebral hemorrhage.We acknowledge that no sport is considered completely safe, and the martial arts especially have evolved over millennia expressly as means to kill and disable. [bib_ref] Morbidity and mortality in the martial arts: a warning, Oler [/bib_ref] The transition to sport is a recent phenomenon that appears to be lacking in safety standards and regulations. [bib_ref] Morbidity and mortality in the martial arts: a warning, Oler [/bib_ref] Arterial dissection is a potentially devastating and underrecognized problem in these healthy young enthusiasts. 2)4)7) Awareness of the risks must be increased and better supervision implemented to prohibit prolonged or overly vigorous moves.
## 12)21)
Coaching and training staff must also caution participants against a sense of immunity, urging prompt medical attention for injuries sustained. A grasp of potential consequences is critical in preventing such injuries, helping as well to raise the index of suspicion, prompt earlier and accurate diagnosis, and thus improve outcomes. [bib_ref] Stroke without dissection from a neck holding manoeuvre in martial arts, Mccarron [/bib_ref]
[fig] 5: 10)23) In terms of the various cervical vascular injuries chronicled in the literature that are due to blunt trauma, sports-related events are almost entirely confined to case reports or limited case series. Brazilian Jiu-Jitsu is a specific style of martial art that has gained popularity in recent years. 20)21) There is little actual doc- [/fig]
[fig] Figure 2, Figure 3: Due to the nature of trauma sustained and a clinical suspicion of dissection, conventional cerebral angiography was performed on an emergency basis, using a Berenstein 5F diagnostic catheter (Merit Medical Systems Inc., South Jordan, UT, USA) over a Radiofocus Guide wire 0.035 inch × 260 cm (Terumo Corp., Tokyo, Japan). Subsequently, decreased filling of the left middle cere-bral artery, with several thrombotic occlusions of M2 and M3 segments (Fig. 2A), and a dissection narrowing the origin of left internal carotid artery were observed (Fig. 2B). The hydrophilic guidewire was positioned in the left external carotid artery, enabling replacement of the diagnostic catheter by an Epsylar 6F introducer sheath (OptiMed, Ettlingen, Germany), which was introduced into distal common carotid artery as a guiding catheter. A 5 mm Spider embolic protection device (ev3 [Covidien], Plymouth, MN, USA) was opened within left internal carotid artery, and angioplasty was done, deploying a Protégé RX self-expanding stent 6-8-40 mm (ev3 [Covidien]) to cover the entire length of the dissection(Fig. 2C, D). The procedure was uneventful, using heparin for the duration,followed by acetylsalicylic acid (200 mg/day) and clopidogrel (75 mg/day) in the intensive care unit. Intracranial thrombolysis was not attempted. During the first day of hospitalization, level of consciousness declined (Glasgow Coma Scale score, 8), Cerebral angiogram: (A) anteroposterior view of occluded M2 middle cerebral artery segments (arrows); (B) left common carotid artery in lateral view, showing narrowing at internal carotid artery origin (arrows); (C) stent placement; and (D) postoperative control images confirming adequate arterial patency, with mural compression of thrombus.calling for tracheal intubation. Another emergency computed tomography scan showed left hemispheric infarction, with hemorrhagic transformation and midline shift. An intraparenchymal catheter was im-Artistic rendering of Brazilian Jiu-Jitsu maneuver known as Rear Naked Choke or Lion Killer: combined neck extension and head rotation (as a defense measure) stretches the compressed internal carotid artery at its origin, causing dissection. [/fig]
|
Hypercoagulability in Patients with Non-Alcoholic Fatty Liver Disease (NAFLD): Causes and Consequences
Citation: Tripodi, A.; Lombardi, R.; Primignani, M.; La Mura, V.; Peyvandi, F.; Fracanzani, A.L. Hypercoagulability in Patients with Non-Alcoholic Fatty Liver Disease (NAFLD): Causes and Consequences.
# Introduction
Coagulation is a tightly regulated and complex humoral/cellular mechanism, which in normal conditions allows blood fluidity within the vascular system but helps to stop bleeding at the site of a blood vessel wall injury. The main components of coagulation are the procoagulant and the anticoagulant factors. The procoagulants start their action soon after the formation of a complex between tissue factor and plasma factor VIIa at the site of a vessel wall injury, when tissue factor that is normally hidden into the cell membranes meets circulating blood and starts a series of iterative activations leading to thrombin generation and, ultimately, to the conversion of fibrinogen to fibrin . Procoagulants are physiologically contrasted by the naturally occurring anticoagulants (i.e., antithrombin, protein C, protein S, and the tissue factor pathway inhibitor) . In normal conditions the pro-and anticoagulants are perfectly balanced, and therefore, unwanted thrombin generation and fibrin deposition do not occur. However, there may be clinical conditions in which the balance between the two drivers is perturbed to such . Schematic representation of the iterative coagulation activation starting from the complex TF-FVIIa formation and leading to thrombin (IIa) generation and fibrin formation. Among the othe functions (not shown in the scheme), thrombin activates FXI, FXIII, FV, FVIII, and protein C (see also [fig_ref] Figure 3: Schematic representation of the activation of protein C on the membrane of... [/fig_ref]. Arrows refer to factors inhibited by the antithrombin-heparin complex or by the activated protein C-protein S complex or TFPI. The suffix "a" denotes activated coagulation factors HMWK, high molecular weight kininogen; PK, prekallicrein; Roman numbers refer to coagulation factors; TFPI, tissue factor pathway inhibitor; AT, antithrombin; TF, tissue factor.
## Hypercoagulability
Hypercoagulability is defined as a procoagulant imbalance due to increased levels o procoagulants, decreased levels of anticoagulants, or both. Other circumstances where hypercoagulability may occur in the absence of apparent coagulation derangement is the presence of one of the following: (i) Elevated circulating microvesicles that stem from monocytes, platelets, or endothelial cells, which upon activation release, into the circulation, massive amounts of small cytoplasmic vesicles possessing, on their membranes, the procoagulant asset of the parent cells (e.g., tissue factor from monocytes negatively charged phosphatidylserine from platelets, etc.), which is, therefore disseminated into the systemic circulation [bib_ref] Tissue factor-bearing microparticles and thrombus formation, Zwicker [/bib_ref]. This may dramatically increase the state o plasma hypercoagulability, as shown by the increased risk of cardiovascular events [2 and venous thromboembolism associated with elevated levels of microvesicles [bib_ref] Circulating microparticles and risk of venous thromboembolism, Bucciarelli [/bib_ref]. (ii Ongoing inflammation, which through the massive release of proinflammatory cytokines perturbs the balance between pro-and anticoagulants leading to hypercoagulability [bib_ref] Two-way interactions between inflammation and coagulation, Levi [/bib_ref] (iii) Elevated circulating chromatin substances, collectively known as neutrophi extracellular traps (NETs), which are the hallmark of heightened neutrophil activationNETs support, on their surface, the activation of coagulation factors and are considered to be risk factors for venous thromboembolism [bib_ref] Circulating nucleosomes and neutrophil activation as risk factors for deep vein thrombosis, Van Montfoort [/bib_ref] [bib_ref] Extracellular histones increase plasma thrombin generation by impairing thrombomodulin-dependent protein C activation, Ammollo [/bib_ref].
## How to assess hypercoagulability
In principle, hypercoagulability can be assessed by performing basic coagulation tests such as: (i) the prothrombin and activated partial thromboplastin time (PT and APTT); (ii) the measurement of individual procoagulants, or (iii) the measurement o individual naturally occurring anticoagulants (i.e., antithrombin, protein C, protein S, or . Schematic representation of the iterative coagulation activation starting from the complex TF-FVIIa formation and leading to thrombin (IIa) generation and fibrin formation. Among the other functions (not shown in the scheme), thrombin activates FXI, FXIII, FV, FVIII, and protein C (see also [fig_ref] Figure 3: Schematic representation of the activation of protein C on the membrane of... [/fig_ref]. Arrows refer to factors inhibited by the antithrombin-heparin complex or by the activated protein C-protein S complex or TFPI. The suffix "a" denotes activated coagulation factors. HMWK, high molecular weight kininogen; PK, prekallicrein; Roman numbers refer to coagulation factors; TFPI, tissue factor pathway inhibitor; AT, antithrombin; TF, tissue factor.
## Hypercoagulability
Hypercoagulability is defined as a procoagulant imbalance due to increased levels of procoagulants, decreased levels of anticoagulants, or both. Other circumstances where hypercoagulability may occur in the absence of apparent coagulation derangement is the presence of one of the following: (i) Elevated circulating microvesicles that stem from monocytes, platelets, or endothelial cells, which upon activation release, into the circulation, massive amounts of small cytoplasmic vesicles possessing, on their membranes, the procoagulant asset of the parent cells (e.g., tissue factor from monocytes, negatively charged phosphatidylserine from platelets, etc.), which is, therefore, disseminated into the systemic circulation [bib_ref] Tissue factor-bearing microparticles and thrombus formation, Zwicker [/bib_ref]. This may dramatically increase the state of plasma hypercoagulability, as shown by the increased risk of cardiovascular events [bib_ref] Cellular origins and thrombogenic activity of microparticles isolated from human atherosclerotic plaques, Leroyer [/bib_ref] and venous thromboembolism associated with elevated levels of microvesicles [bib_ref] Circulating microparticles and risk of venous thromboembolism, Bucciarelli [/bib_ref]. (ii) Ongoing inflammation, which through the massive release of proinflammatory cytokines perturbs the balance between pro-and anticoagulants leading to hypercoagulability [bib_ref] Two-way interactions between inflammation and coagulation, Levi [/bib_ref]. (iii) Elevated circulating chromatin substances, collectively known as neutrophil extracellular traps (NETs), which are the hallmark of heightened neutrophil activation. NETs support, on their surface, the activation of coagulation factors and are considered to be risk factors for venous thromboembolism [bib_ref] Circulating nucleosomes and neutrophil activation as risk factors for deep vein thrombosis, Van Montfoort [/bib_ref] [bib_ref] Extracellular histones increase plasma thrombin generation by impairing thrombomodulin-dependent protein C activation, Ammollo [/bib_ref].
## How to assess hypercoagulability
In principle, hypercoagulability can be assessed by performing basic coagulation tests such as: (i) the prothrombin and activated partial thromboplastin time (PT and APTT); (ii) the measurement of individual procoagulants, or (iii) the measurement of individual naturally occurring anticoagulants (i.e., antithrombin, protein C, protein S, or tissue factor pathway inhibitor). However, none of these approaches, if used as standalone laboratory tools, are suitable to account for the complex interaction between the pro-and anticoagulant Biomedicines 2022, 10, 249 3 of 12 drivers operating in vivo. PT and APTT are global coagulation tests, responsive to most of the procoagulants (they are, in fact, abnormally prolonged in congenital hemorrhagic deficiencies such as hemophilia and other congenital hemorrhagic disorders). PT and APTT are, however, not equally responsive to the action of naturally occurring anticoagulants [bib_ref] Acquired coagulation disorders: Revisited using global coagulation/anticoagulation testing, Tripodi [/bib_ref]. They are, in fact, normal in patients with congenital deficiency of antithrombin, protein C, or protein S conditions where they should be shortened, since carrier patients generate much more thrombin than healthy subjects. Hence, on the one hand, PT and APTT are unfit to represent the process of coagulation occurring in vivo. On the other hand, naturally occurring anticoagulants, if measured as standalone parameters, although useful for identifying hypercoagulability in patients with congenital deficiency, would not account for the balance represented by the presence of their counterpart (i.e., the procoagulants) that could be variably increased or decreased depending on the circumstances, especially in acquired coagulopathies. Finally, none of the above approaches account for the contribution to hypercoagulability provided by the presence of inflammation, high procoagulant microvesicles, or NETs. In contrast, hypercoagulability can be conveniently investigated by the last generation of global coagulation procedures, called thrombin generation, which are much more representative of the coagulation process that occurs in vivo than the basic tests, i.e., PT, APTT, and individual pro-or anticoagulants, when taken as standalone measurements.
## Thrombin generation procedures
Thrombin generation defines a global coagulation procedure that, upon in vitro activation of coagulation in platelet-poor plasma by small amounts of tissue factor and negatively charged synthetic phospholipids (triggers), has been designed to continuously monitor the generation of thrombin [bib_ref] Calibrated automated thrombin generation measurement in clotting plasma, Hemker [/bib_ref]. The quality and quantity of triggers have been chosen to mimic as much as possible the conditions operating in vivo [bib_ref] Thrombin generation assessed as endogenous thrombin potential in patients with hyper-or hypo-coagulability...., Chantarangkul [/bib_ref]. The procedure can be run in a completely automated fashion and the system constructs a thrombin generation curve (thrombogram), characterized by an ascending arm, which records the thrombin concentration generated as a function of the action of the procoagulants and a descending arm, which records the thrombin neutralized by the action of the naturally occurring anticoagulants [fig_ref] Figure 2: Schematic representation of the thrombin generation curve [/fig_ref].
Biomedicines 2022, 10, x FOR PEER REVIEW 3 of 12 tissue factor pathway inhibitor). However, none of these approaches, if used as standalone laboratory tools, are suitable to account for the complex interaction between the pro-and anticoagulant drivers operating in vivo. PT and APTT are global coagulation tests, responsive to most of the procoagulants (they are, in fact, abnormally prolonged in congenital hemorrhagic deficiencies such as hemophilia and other congenital hemorrhagic disorders). PT and APTT are, however, not equally responsive to the action of naturally occurring anticoagulants [bib_ref] Acquired coagulation disorders: Revisited using global coagulation/anticoagulation testing, Tripodi [/bib_ref]. They are, in fact, normal in patients with congenital deficiency of antithrombin, protein C, or protein S conditions where they should be shortened, since carrier patients generate much more thrombin than healthy subjects. Hence, on the one hand, PT and APTT are unfit to represent the process of coagulation occurring in vivo. On the other hand, naturally occurring anticoagulants, if measured as standalone parameters, although useful for identifying hypercoagulability in patients with congenital deficiency, would not account for the balance represented by the presence of their counterpart (i.e., the procoagulants) that could be variably increased or decreased depending on the circumstances, especially in acquired coagulopathies. Finally, none of the above approaches account for the contribution to hypercoagulability provided by the presence of inflammation, high procoagulant microvesicles, or NETs. In contrast, hypercoagulability can be conveniently investigated by the last generation of global coagulation procedures, called thrombin generation, which are much more representative of the coagulation process that occurs in vivo than the basic tests, i.e., PT, APTT, and individual pro-or anticoagulants, when taken as standalone measurements.
## Thrombin generation procedures
Thrombin generation defines a global coagulation procedure that, upon in vitro activation of coagulation in platelet-poor plasma by small amounts of tissue factor and negatively charged synthetic phospholipids (triggers), has been designed to continuously monitor the generation of thrombin [bib_ref] Calibrated automated thrombin generation measurement in clotting plasma, Hemker [/bib_ref]. The quality and quantity of triggers have been chosen to mimic as much as possible the conditions operating in vivo [bib_ref] Thrombin generation assessed as endogenous thrombin potential in patients with hyper-or hypo-coagulability...., Chantarangkul [/bib_ref]. The procedure can be run in a completely automated fashion and the system constructs a thrombin generation curve (thrombogram), characterized by an ascending arm, which records the thrombin concentration generated as a function of the action of the procoagulants and a descending arm, which records the thrombin neutralized by the action of the naturally occurring anticoagulants [fig_ref] Figure 2: Schematic representation of the thrombin generation curve [/fig_ref]. The main parameters of the thrombogram are (i) the lag time, defined as the time needed for the first amounts of thrombin to appear; (ii) the time-to-peak, defined as the time needed to reach the peak thrombin; (iii) the peak thrombin, defined as the greatest The main parameters of the thrombogram are (i) the lag time, defined as the time needed for the first amounts of thrombin to appear; (ii) the time-to-peak, defined as the time needed to reach the peak thrombin; (iii) the peak thrombin, defined as the greatest amount of generated thrombin; (iv) the area under the curve, called endogenous thrombin potential (ETP), which accounts for the net amount of thrombin that can be generated under the driving forces of the procoagulants opposed by the anticoagulants operating in plasma. Prolonged or short lag time and time-to-peak can be considered to be indexes of hypo-or hypercoagulability, respectively. Conversely, low or high peak thrombin and ETP can be considered to be indexes of hypo-or hypercoagulability, respectively. The above parameters can be recorded in the presence or absence of exogenously added thrombomodulin (TM) [bib_ref] Modified endogenous thrombin potential with results expressed as ratio of values with-to-without..., Tripodi [/bib_ref]. TM is the physiological activator of protein C, which is, in turn, the physiological inhibitor to factor VIIIa and factor Va [fig_ref] Figure 3: Schematic representation of the activation of protein C on the membrane of... [/fig_ref].
Biomedicines 2022, 10, x FOR PEER REVIEW 4 of 12 amount of generated thrombin; (iv) the area under the curve, called endogenous thrombin potential (ETP), which accounts for the net amount of thrombin that can be generated under the driving forces of the procoagulants opposed by the anticoagulants operating in plasma. Prolonged or short lag time and time-to-peak can be considered to be indexes of hypo-or hypercoagulability, respectively. Conversely, low or high peak thrombin and ETP can be considered to be indexes of hypo-or hypercoagulability, respectively. The above parameters can be recorded in the presence or absence of exogenously added thrombomodulin (TM) [bib_ref] Modified endogenous thrombin potential with results expressed as ratio of values with-to-without..., Tripodi [/bib_ref]. TM is the physiological activator of protein C, which is, in turn, the physiological inhibitor to factor VIIIa and factor Va [fig_ref] Figure 3: Schematic representation of the activation of protein C on the membrane of... [/fig_ref]. TM is, however, located on endothelial cells and much less in plasma, and therefore, any assay meant to assess coagulation in plasma (namely PT, APTT, or protein C/protein S) cannot account for the optimal activation of protein C and, hence, for the inhibition of factor VIIIa, both representing potent regulators of thrombin generation. Hence, the evaluation of ETP measured in the presence or absence of TM or as the ratio between the two values (called the ETP-TM ratio) can be considered to be the best in vitro representation of the coagulation process operating in vivo [bib_ref] Modified endogenous thrombin potential with results expressed as ratio of values with-to-without..., Tripodi [/bib_ref]. In principle, the ETP-TM ratio represents an index of the resistance of plasma to the anticoagulant action of TM, which, in turn, depends on the imbalance between factor VIII and protein C, the two major regulators of thrombin generation. Consequently, the ETP-TM ratio is a reliable index of plasma hypercoagulability (the higher the ratio the greater the hypercoagulability) [bib_ref] Modified endogenous thrombin potential with results expressed as ratio of values with-to-without..., Tripodi [/bib_ref].
## General features of nafld
Hypercoagulability is one of the three components identified by Virchow (i.e, hypercoagulability, endothelial dysfunction, and reduced/turbulent blood flow) that, alone or in combination, can trigger venous thromboembolism. Interestingly, hypercoagulability is very often associated with clinical conditions characterized by an increased risk of cardiovascular diseases, either arterial or venous, or both. The above clinical conditions include type II diabetes [bib_ref] Hypercoagulability in patients with type 2 diabetes mellitus detected by a thrombin..., Tripodi [/bib_ref] , obesity [bib_ref] Body mass index reduction improves the baseline procoagulant imbalance of obese subjects, Tripodi [/bib_ref] , liver cirrhosis [bib_ref] An imbalance of pro-vs. anti-coagulation factors in plasma from patients with cirrhosis, Tripodi [/bib_ref] , Cushing disease, inflammatory bowel diseases [bib_ref] Increased thrombin generation in inflammatory bowel diseases, Saibeni [/bib_ref] [bib_ref] Anti-TNF-α Treatment Reduces the Baseline Procoagulant Imbalance of Patients With Inflammatory Bowel..., Tripodi [/bib_ref] , and others. Owing to its clinical characteristics (see below), non-alcoholic fatty liver disease (NAFLD) could be the prototype of the composite diseases that could be investigated for hypercoagulability.
NAFLD is a common clinical condition with an estimated prevalence of 23% in Europe, 24% in the USA, 32% in the Middle East, and 27% in Asia [bib_ref] Global burden of NAFLD and NASH: Trends, predictions, risk factors and prevention, Younossi [/bib_ref]. Recent trend analyses have shown that, in the period 1991-2019, NAFLD increased from 21.9 to 37.3% (yearly increase of 0.7% (p < 0.0001)) [bib_ref] global NAFLD prevalence: A systematic review and meta-analysis, Le [/bib_ref]. NAFLD is relatively more common in men with advancing age but is also observed in obese children and adolescents with sequels similar to those observed in adults [bib_ref] Genderspecific differences in adipose distribution and adipocytokines influence adolescent nonalcoholic fatty liver..., Ayonrinde [/bib_ref] [bib_ref] Body mass index in childhood and adult risk of primary liver cancer, Berentzen [/bib_ref] ; in women, the prevalence of NAFLD becomes similar . Schematic representation of the activation of protein C on the membrane of endothelial cells and the mechanism of action of the complex of activated protein C/protein S in inhibiting factor Va and factor VIIIa. Instrumental to the mechanism are two endothelial receptors, thrombomodulin (TM) that binds thrombin (Th) and endothelial protein C receptor (EPCR) that binds plasma protein C (PC). The proximity of the above receptors localizes the conversion of the substrate (PC) by the enzyme (Th) into activated PC (APC). APC in complex with its plasmatic cofactor protein S (PS) eventually inhibits factors Va and VIIIa, thus, downregulating thrombin production.
TM is, however, located on endothelial cells and much less in plasma, and therefore, any assay meant to assess coagulation in plasma (namely PT, APTT, or protein C/protein S) cannot account for the optimal activation of protein C and, hence, for the inhibition of factor VIIIa, both representing potent regulators of thrombin generation. Hence, the evaluation of ETP measured in the presence or absence of TM or as the ratio between the two values (called the ETP-TM ratio) can be considered to be the best in vitro representation of the coagulation process operating in vivo [bib_ref] Modified endogenous thrombin potential with results expressed as ratio of values with-to-without..., Tripodi [/bib_ref]. In principle, the ETP-TM ratio represents an index of the resistance of plasma to the anticoagulant action of TM, which, in turn, depends on the imbalance between factor VIII and protein C, the two major regulators of thrombin generation. Consequently, the ETP-TM ratio is a reliable index of plasma hypercoagulability (the higher the ratio the greater the hypercoagulability) [bib_ref] Modified endogenous thrombin potential with results expressed as ratio of values with-to-without..., Tripodi [/bib_ref].
## General features of nafld
Hypercoagulability is one of the three components identified by Virchow (i.e, hypercoagulability, endothelial dysfunction, and reduced/turbulent blood flow) that, alone or in combination, can trigger venous thromboembolism. Interestingly, hypercoagulability is very often associated with clinical conditions characterized by an increased risk of cardiovascular diseases, either arterial or venous, or both. The above clinical conditions include type II diabetes [bib_ref] Hypercoagulability in patients with type 2 diabetes mellitus detected by a thrombin..., Tripodi [/bib_ref] , obesity [bib_ref] Body mass index reduction improves the baseline procoagulant imbalance of obese subjects, Tripodi [/bib_ref] , liver cirrhosis [bib_ref] An imbalance of pro-vs. anti-coagulation factors in plasma from patients with cirrhosis, Tripodi [/bib_ref] , Cushing disease, inflammatory bowel diseases [bib_ref] Increased thrombin generation in inflammatory bowel diseases, Saibeni [/bib_ref] [bib_ref] Anti-TNF-α Treatment Reduces the Baseline Procoagulant Imbalance of Patients With Inflammatory Bowel..., Tripodi [/bib_ref] , and others. Owing to its clinical characteristics (see below), nonalcoholic fatty liver disease (NAFLD) could be the prototype of the composite diseases that could be investigated for hypercoagulability.
NAFLD is a common clinical condition with an estimated prevalence of 23% in Europe, 24% in the USA, 32% in the Middle East, and 27% in Asia [bib_ref] Global burden of NAFLD and NASH: Trends, predictions, risk factors and prevention, Younossi [/bib_ref]. Recent trend analyses have shown that, in the period 1991-2019, NAFLD increased from 21.9 to 37.3% (yearly increase of 0.7% (p < 0.0001)) [bib_ref] global NAFLD prevalence: A systematic review and meta-analysis, Le [/bib_ref]. NAFLD is relatively more common in men with advancing age but is also observed in obese children and adolescents with sequels similar to those observed in adults [bib_ref] Genderspecific differences in adipose distribution and adipocytokines influence adolescent nonalcoholic fatty liver..., Ayonrinde [/bib_ref] [bib_ref] Body mass index in childhood and adult risk of primary liver cancer, Berentzen [/bib_ref] ; in women, the prevalence of NAFLD becomes similar to that of men after menopause [bib_ref] Sex Differences in Nonalcoholic Fatty Liver Disease: State of the Art and..., Lonardo [/bib_ref]. NAFLD is a significant global public health burden to health care systems across the world and is expected to increase along with the incidence of such metabolic derangements typically observed in patients with type II diabetes, obesity, or dyslipidemia. Furthermore, being the most common liver disease, NAFLD is expected to be the leading cause of liver transplantation in the near future [bib_ref] Nonalcoholic steatohepatitis is the second leading etiology of liver disease among adults..., Wong [/bib_ref].
NAFLD ranges from simple steatosis, defined by an accumulation of fat greater than 5% of liver weight, to the progressive form of non-alcoholic steatohepatitis (NASH), whose hallmark is inflammation and hepatocellular death that can progress to fibrosis, up to the most severe form of metabolic cirrhosis. Interestingly, this life-threatening progression of liver disease severity has been associated with an increased risk of atherosclerosis (i.e., coronary artery and abdominal aortic calcification and the presence and progression of carotid intima-media thickness [bib_ref] Associations between nonalcoholic fatty liver disease and subclinical atherosclerosis in middle-aged adults:..., Van Wagner [/bib_ref] [bib_ref] Fatty liver is an independent predictor of early carotid atherosclerosis, Pais [/bib_ref] , as well as high risk of cardiovascular diseases [bib_ref] Non-alcoholic fatty liver disease and risk of incident cardiovascular disease: A meta-analysis, Targher [/bib_ref] , the latter accounting for the higher rate of total mortality observed in NAFLD as compared with the general population. NAFLD has also been associated with the occurrence of venous thromboembolism (including deep vein thrombosis of the lower limbs, pulmonary embolism, and splanchnic vein thrombosis) [bib_ref] Increased risk of venous thromboembolism in hospitalized patients with cirrhosis due to..., Stine [/bib_ref] [bib_ref] High prevalence of nonalcoholic fatty liver in patients with idiopathic venous thromboembolism, Di Minno [/bib_ref]. Hypercoagulability is considered to be a risk factor that alone or (more likely) in combination with other genetic (e.g., prothrombotic polymorphisms) and/or circumstantial risk factors (e.g., old age, cancer, recent surgery, metabolic syndrome, oral contraceptive intake, and/or hormonal replacement therapy, etc.) may increase the risk of arterial and venous thrombotic complications. In this respect, the general features and clinical characteristics of NAFLD make it a prototype to look at regarding an association with hypercoagulability as a cause for the cardiovascular risk in metabolic diseases.
Environmental and genetic factors contribute to the pathogenesis of NAFLD that, in addition to being a very prevalent disease in the general population, is burdened by a series of complications, not only hepatic, which determine morbidity and mortality through multiple mechanisms not yet completely understood. For example, in addition to liver diseases, type II diabetes, which is per se a risk factor of cardiovascular diseases, is frequently associated with NAFLD [bib_ref] Nonalcoholic fatty liver disease is associated with an almost twofold increased risk..., Ballestri [/bib_ref] and may increase, by two times, the risk of developing NASH and severe liver-related complications (cirrhosis, liver failure, and hepatocellular carcinoma). In addition, patients with established type II diabetes also show a high prevalence of fatty liver. All in all, type II diabetes and NAFLD are closely associated, as they share several cardiometabolic risk factors and their clinical evolution is strongly influenced by each other. The prevalence of imaging-defined NAFLD in diabetes is about 60-75%, and that of liver fibrosis, which in NAFLD has been defined as the strongest predictor of long-term adverse clinical outcomes, is about 5-20% [bib_ref] The global epidemiology of NAFLD and NASH in patients with type 2..., Younossi [/bib_ref]. Interestingly, hypercoagulability detected by thrombin generation has been shown in patients with type II diabetes and seems to be mediated by increased factor VIII and the presence of procoagulant microvesicles [bib_ref] Hypercoagulability in patients with type 2 diabetes mellitus detected by a thrombin..., Tripodi [/bib_ref]. Other relevant extrahepatic manifestations in NAFLD are neoplasms of the gastrointestinal tract [bib_ref] The risk of incident extrahepatic cancers is higher in non-alcoholic fatty liver..., Allen [/bib_ref] [bib_ref] The association between nonalcoholic fatty liver disease and esophageal, stomach, or colorectal..., Lee [/bib_ref] and chronic kidney disease [bib_ref] NAFLD as a driver of chronic kidney disease, Byrne [/bib_ref] , which share the same risk factors as type II diabetes, as well as metabolic syndrome, obesity, hypertension, and obstructive sleep apnea [bib_ref] Association between non-alcoholic fatty liver disease and obstructive sleep apnea, Umbro [/bib_ref] , even in the absence of obesity or metabolic syndrome and hypothyroidism [bib_ref] Low Thyroid Function in Nonalcoholic Fatty Liver Disease Is an Independent Predictor..., Kim [/bib_ref] , which are associated with a higher risk for overall and cardiovascular mortality, polycystic ovary syndrome [bib_ref] Polycystic ovary syndrome with hyperandrogenism as a risk factor for non-obese non-alcoholic..., Kim [/bib_ref] , and psoriasis [bib_ref] Prevalence, characteristics and severity of non-alcoholic fatty liver disease in patients with..., Miele [/bib_ref]. Many risk factors influence survival in NAFLD/NASH, whose leading causes of death are cardiovascular diseases. A similar association has been observed for liver transplant outcome. Indeed, liver transplant recipients with NASH have significantly more fatal and nonfatal post-transplant cardiovascular events than recipients without NASH [bib_ref] Outcomes of liver transplantation for nonalcoholic steatohepatitis: A systematic review and meta-analysis, Wang [/bib_ref] [bib_ref] Patients transplanted for nonalcoholic steatohepatitis are at increased risk for postoperative cardiovascular..., Vanwagner [/bib_ref].
Among the above, the relatively high prevalence of cardiovascular diseases is the most relevant for this review and will be taken into consideration in this article. The proposed mechanisms for the development of cardiovascular diseases in patients with NAFLD include genetic predisposition, chronic inflammation, endothelial dysfunction, oxidative stress, and hemostatic alterations in the balance between pro-and anticoagulant factors [bib_ref] Non-alcoholic fatty liver disease and cardiovascular risk: Pathophysiological mechanisms and implications, Francque [/bib_ref]. In recent decades, single nucleotide polymorphisms in the genes that regulate lipid handling and secretion have been associated with fatty liver and its progressive forms.
## Hypercoagulability in patients with nafld
The presence of hemostatic alterations has been hypothesized to underlie both arterial cardiovascular injury (presence of arterial thrombosis (e.g., myocardial infarction, cerebrovascular stroke) and venous thromboembolism (e.g., increased rate of deep vein thrombosis, pulmonary embolism, and portal vein thrombosis)) [bib_ref] Increased risk of venous thromboembolism in hospitalized patients with cirrhosis due to..., Stine [/bib_ref]. However, the association of NAFLD with hypercoagulability, although surmised based on epidemiological observations, has not been widely evaluated.
The present review article aims to overview the causes and possible consequences of hypercoagulability in NAFLD to foster research work that may help to manage this important aspect in this group of patients.
## Basic tests of coagulation in nafld
Typically, in patients with NAFLD, such basic tests of coagulation as the PT and APTT are within normal range or slightly prolonged, yet of negligible clinical significance [bib_ref] Procoagulant imbalance in patients with non-alcoholic fatty liver disease, Tripodi [/bib_ref]. This is not unexpected if one considers that these tests are not truly representative of the process that occurs in vivo (see above). In contrast, individual pro-and anticoagulants show variable degrees of abnormalities. Previous studies have shown that some procoagulant factors (namely factors VIII, IX, XI, and XII) are increased [bib_ref] Increased coagulation factor VIII, IX, XI and XII activities in non-alcoholic fatty..., Kotronen [/bib_ref] , whereas little information has been reported on the levels of naturally occurring anticoagulants.
In a relatively large study [bib_ref] Procoagulant imbalance in patients with non-alcoholic fatty liver disease, Tripodi [/bib_ref] of 113 patients with NAFLD, factor VIII (one of the most potent drivers of thrombin generation) increased from steatosis (99 U/dL, (71-150)) to metabolic cirrhosis (157 U/dL (64-232)), p < 0.001. Conversely, protein C (one of the most potent drivers of thrombin downregulation) decreased from steatosis (103 U/dL, (77-228)) to metabolic cirrhosis (77 U/dL (17-146)), p < 0.001. Consequently, the ratio of factor VIII/protein C, also taken as an index of hypercoagulability, increased from steatosis (0.96, (0.36-1.60)) or NASH to metabolic cirrhosis (2.05, (0.81-12.1)), p < 0.001 and was correlated with the ETP-TM ratio (rho = 0.543, p < 0.001). Antithrombin was reduced only in patients with metabolic cirrhosis (patients vs. controls, 78 U/dL, vs. 109 U/dL (74-140)), p < 0.001).
## Thrombin generation in nafld
Thrombin generation expressed as the ETP-TM ratio increased from controls (0.57 (0.11-0.89)) to steatosis or NASH (0.72 (0.33-0.86)) and metabolic cirrhosis (0.80 (0.57-0.95)), (p < 0.001), the latter being comparable to that observed for alcoholic/viral cirrhosis (0.80 (0.57-0.95) vs. 0.80 (0.44-0.96)) taken as positive controls [bib_ref] Procoagulant imbalance in patients with non-alcoholic fatty liver disease, Tripodi [/bib_ref]. These results provide evidence that a state of hypercoagulability does exist in patients with NAFLD and that it progresses with the severity of the disease, being relatively low in steatosis and dramatically high in metabolic cirrhosis where it is similar to that observed in a group of patients with viral/alcoholic cirrhosis taken as positive controls. Similar results were obtained upon investigating the same cohort of patients by means of an additional procedure for thrombin generation, in which TM was substituted with Protac ® , a snake venom, non-physiological activator, which shares the same protein C-activating properties as TM [bib_ref] Procoagulant imbalance in patients with non-alcoholic fatty liver disease, Tripodi [/bib_ref]. Interestingly, these values were also comparable in a subgroup of patients with less severe cirrhosis (Child A) of both etiologies.
Additional evidence that strengthens the relationship between the ETP-TM ratio and hypercoagulability in NAFLD is shown in [fig_ref] Figure 4: Correlation between coagulation parameters and the ETP-TM ratio in patients wit NAFLD... [/fig_ref]. Patients with NAFLD showed a significant inverse correlation of the ETP-TM ratio vs. antithrombin (rho = −0.50, p < 0.01) or protein C (rho = −0.56, p < 0.01), and a direct correlation of the ETP-TM ratio vs. factor VIII (rho = 0.27, p < 0.01) or vs. the ratio of factor VIII/protein C (rho = 0.54, p < 0.01). Furthermore, the ETP-TM ratio was directly correlated with the body mass index (rho = 0.26, p < 0.01), which is considered to be a generic risk factor for hypercoagulability. Interestingly, the above correlations were not evident in the control group (see [fig_ref] Figure 4: Correlation between coagulation parameters and the ETP-TM ratio in patients wit NAFLD... [/fig_ref] , suggesting that the ETP-TM ratio may be considered to be a reliable candidate laboratory tool for assessing hypercoagulability in patients with NAFLD. Additional evidence that supports the concept that hypercoagulability may play a role as a possible contributor to the risk of cardiovascular disease in NAFLD is the observation that patients with a ETP-TM ratio higher than the median value of controls had a significantly high risk of metabolic syndrome, high media-intima thickness, fibrosis, steatosis, or lobular inflammation [bib_ref] Procoagulant imbalance in patients with non-alcoholic fatty liver disease, Tripodi [/bib_ref] , all typical features of NAFLD. Finally, additional contributors supporting the concept of hypercoagulability is the occurrence in NAFLD of increased NETs levels [bib_ref] Neutrophil extracellular traps promote inflammation and development of hepatocellular carcinoma in nonalcoholic..., Van Der Windt [/bib_ref] and microvesicles [bib_ref] Extracellular vesicles in non-alcoholic and alcoholic fatty liver diseases, Eguchi [/bib_ref] , which, through dissemination into the circulation of procoagulant activity, may trigger coagulation and, consequently, thrombin generation. Inflammation, typically observed in the more severe form of NAFLD, namely NASH, may also trigger coagulation and thrombin generation through the release of proinflammatory mediators (e.g., IL-1, IL-6, and TNFα) [bib_ref] Chronic Inflammation in Non-Alcoholic Steatohepatitis: Molecular Mechanisms and Therapeutic Strategies, Luci [/bib_ref]. Although the relationship between microvesicles, NETs, or inflammation and thrombin generation in NAFLD has not yet been directly explored, analogy with other diseases such as type II diabetes [bib_ref] Hypercoagulability in patients with type 2 diabetes mellitus detected by a thrombin..., Tripodi [/bib_ref] and othersand the lack of correlation between the ETP-TM ratio and the individual components of coagulation observed in heathy controls (see [fig_ref] Figure 4: Correlation between coagulation parameters and the ETP-TM ratio in patients wit NAFLD... [/fig_ref] support the concept of a causal relationship in patients with NAFLD. 0.26, p < 0.01), which is considered to be a generic risk factor for hypercoagulability Interestingly, the above correlations were not evident in the control group (see [fig_ref] Figure 4: Correlation between coagulation parameters and the ETP-TM ratio in patients wit NAFLD... [/fig_ref] suggesting that the ETP-TM ratio may be considered to be a reliable candidate laboratory tool for assessing hypercoagulability in patients with NAFLD. Additional evidence tha supports the concept that hypercoagulability may play a role as a possible contributor to the risk of cardiovascular disease in NAFLD is the observation that patients with a ETP TM ratio higher than the median value of controls had a significantly high risk o metabolic syndrome, high media-intima thickness, fibrosis, steatosis, or lobula inflammation [bib_ref] Procoagulant imbalance in patients with non-alcoholic fatty liver disease, Tripodi [/bib_ref] , all typical features of NAFLD. Finally, additional contributor supporting the concept of hypercoagulability is the occurrence in NAFLD of increased NETs levels [bib_ref] Neutrophil extracellular traps promote inflammation and development of hepatocellular carcinoma in nonalcoholic..., Van Der Windt [/bib_ref] and microvesicles [bib_ref] Extracellular vesicles in non-alcoholic and alcoholic fatty liver diseases, Eguchi [/bib_ref] , which, through dissemination into the circulation of procoagulant activity, may trigger coagulation and, consequently, thrombin generation. Inflammation, typically observed in the more severe form of NAFLD, namely NASH, may also trigger coagulation and thrombin generation through the release o proinflammatory mediators (e.g., IL-1, IL-6, and TNFα) [bib_ref] Chronic Inflammation in Non-Alcoholic Steatohepatitis: Molecular Mechanisms and Therapeutic Strategies, Luci [/bib_ref]. Although the relationship between microvesicles, NETs, or inflammation and thrombin generation in NAFLD ha not yet been directly explored, analogy with other diseases such as type II diabetes [12 and othersand the lack of correlation between the ETP-TM ratio and the individua components of coagulation observed in heathy controls (see [fig_ref] Figure 4: Correlation between coagulation parameters and the ETP-TM ratio in patients wit NAFLD... [/fig_ref] support the concep of a causal relationship in patients with NAFLD. More recently, Potze et al. [bib_ref] Preserved hemostatic status in patients with non-alcoholic fatty liver disease, Potze [/bib_ref] investigated a cohort of patients with NAFLD and concluded that hemostasis (i.e., platelets, coagulation, and fibrinolysis) was rebalanced and that there were no biochemical signs of hypercoagulability as measured by thrombin generation, despite elevated factor VIII, von Willebrand factor (another index of procoagulant imbalance), and reduced protein C reported in their cohort.
There are various explanations for the apparent contrasting conclusions reported by the two studies. First, the sample size was considerably different (n = 113 (ours) vs. n = 68 (Potze et al. study)). Therefore, it is possible that the latter study was underpowered to show statistical significance. In fact, the ETP-TM ratio in the Potze study (they called it TM-SR) tended to increase in patients with NASH cirrhosis as compared with controls, without reaching statistical significance. Second, Potze et al. [bib_ref] Preserved hemostatic status in patients with non-alcoholic fatty liver disease, Potze [/bib_ref] reported unusually high factor VIII in their population of controls (median, 144 U/dL vs. expected 100 U/dL). Therefore, it is possible that the selected control population possessed a state of hypercoagulability, as shown by high factor VIII that could have masked that observed for NAFLD. Third, the thrombin generation procedures in the two studies were not necessarily comparable. For example, Potze et al. [bib_ref] Preserved hemostatic status in patients with non-alcoholic fatty liver disease, Potze [/bib_ref] used tissue factor and negatively charged phospholipids at concentrations that were four times higher than those used in our study. The concentrations of tissue factor and negatively charged phospholipids are among the most important determinants of the sensitivity of thrombin generation parameters to detect hypo-or hypercoagulability [bib_ref] Thrombin generation assessed as endogenous thrombin potential in patients with hyper-or hypo-coagulability...., Chantarangkul [/bib_ref]. Finally, Potze et al. [bib_ref] Preserved hemostatic status in patients with non-alcoholic fatty liver disease, Potze [/bib_ref] did not report on the type/concentrations of TM, another important determinant of the ETP-TM ratio. All in all, the above circumstances may well explain the contrasting results obtained in the two studies.
In conclusion, the experimental results of our study as well as the growing body of epidemiological observations, support the concept that NAFLD is associated with a state of plasma hypercoagulability, which alone or (more likely) in combination with other circumstantial risk factors may contribute to the progression of the disease from steatosis to the most severe form of metabolic cirrhosis, atherosclerosis, and the increased risk of cardiovascular diseases, which are typically observed in these patients.
## Hypercoagulability and anticoagulation in nafld
Although data in humans have consistently demonstrated the presence of a procoagulant imbalance in NAFLD, to date, few observations are available on the effect of anticoagulant therapy on hypercoagulability and, more generally, on the clinical outcome of fatty liver. On the contrary, interesting results have been observed in preclinical studies with animal models. Dabigatran, a direct thrombin inhibitor, has shown a protective effect on hepatic fibrin deposition, inflammation, and hepatocellular damage in mice treated with high-fat diets [bib_ref] Thrombin inhibition with dabigatran protects against high-fat diet-induced fatty liver disease in..., Kopec [/bib_ref]. Furthermore, the activity of dabigatran has been related to a reduction in high-fat diet-induced weight gain [bib_ref] Thrombin promotes diet-induced obesity through fibrin-driven inflammation, Kopec [/bib_ref]. Interestingly, possible interactions among anticoagulant activity, weight gain/obesity, and NAFLD were recently reported in humans by Wen et al. [bib_ref] Required warfarin dose and time in therapeutic range in patients with diagnosed..., Wen [/bib_ref]. Lower average warfarin daily dosage and shorter time in the therapeutic range were found in patients diagnosed with NAFLD/NASH [bib_ref] Required warfarin dose and time in therapeutic range in patients with diagnosed..., Wen [/bib_ref]. Notwithstanding, this anticoagulation in patients with NAFLD has been poorly investigated.
In addition to vitamin K antagonists (VKA), other drugs such as low molecular weight heparin (LMWH) or direct oral anticoagulants (DOAC), widely used in the general population to prevent venous thromboembolism and/or systemic embolism/ischemic stroke in patients with atrial fibrillation, could be potentially used in patients with NAFLD as a therapeutic strategy aimed to slow down the progression to liver fibrosis, typically observed in these patients [bib_ref] Hepatic and portal vein thrombosis in cirrhosis: Possible role in development of..., Wanless [/bib_ref] [bib_ref] Parenchymal extinction: Coagulation and hepatic fibrogenesis, Anstee [/bib_ref] [bib_ref] Association between thrombotic risk factors and extent of fibrosis in patients with..., Assy [/bib_ref]. For example, experimental observations in animal models have shown that coagulation modulated murine hepatic fibrogenesis and several clinical and preclinical studies have demonstrated a link between fibrogenesis and hyperactivation of hemostasis [bib_ref] Coagulation status modulates murine hepatic fibrogenesis: Implications for the development of novel..., Anstee [/bib_ref]. Specifically, it has been shown that the progression of liver fibrosis is mediated by hypercoagulability, which, in turn, can modulate various aspects of organ fibrogenesis through thrombin and the family of protease activated receptors (PAR-1/2), which are potent direct effectors of stellate cell activation and fibrosis. The above processes can be slowed down in animal models by the administration of LMWH, VKA, or DOAC, leading to downregulation of thrombin production (reviewed in [bib_ref] Hypercoagulability in cirrhosis: Causes and consequences, Tripodi [/bib_ref]. Taken together, these observations suggest that anticoagulation could be useful for preventing/treating cardiovascular diseases in patients with NAFLD and also for preventing/treating progression to liver fibrosis.
## Hypercoagulability and statins in nafld
Statins are widely used as cardioprotective drugs in patients with heightened cardiovascular risk. Furthermore, recent nationwide cohort and nested case-control studies have shown that statin use in the general population was associated with reduced venous thromboembolism recurrence [bib_ref] Statin use and venous thromboembolism recurrence: A combined nationwide cohort and nested..., Schmidt [/bib_ref]. However, although considered to be relatively safe by guidelines [bib_ref] Statins: An Under-Appreciated Asset for the Prevention and the Treatment of NAFLD..., Athyros [/bib_ref] [bib_ref] Antidiabetic Drugs and Statins in Nonalcoholic Fatty Liver Disease, Kothari [/bib_ref] [bib_ref] Statins for non-alcoholic fatty liver disease and non-alcoholic steatohepatitis, Eslami [/bib_ref] [bib_ref] Effects of Statin Use on the Development and Progression of Nonalcoholic Fatty..., Lee [/bib_ref] , statins are not widely used to treat NAFLD. Indeed, there are animal studies and post hoc small clinical trials in humans that have shown some beneficial effects of statins in improving liver histology and, most importantly, risk reduction of cardiovascular diseases (reviewed in [bib_ref] Antidiabetic Drugs and Statins in Nonalcoholic Fatty Liver Disease, Kothari [/bib_ref]. A recent relatively large nationwide nested case-control study investigated the effect of statins on the development and progression of NAFLD and concluded that statin use decreased the risk of NAFLD occurrence and the risk of liver fibrosis once NAFLD had developed [bib_ref] Effects of Statin Use on the Development and Progression of Nonalcoholic Fatty..., Lee [/bib_ref]. Interestingly, statins seem to mediate their cardioprotective beneficial effect in patients with familial hypercholesterolemia, at least in part, by downregulating thrombin production [bib_ref] Statins decrease thrombin generation in patients with hypercholesterolemia, Tripodi [/bib_ref]. The beneficial effect of statins on thrombin generation is probably mediated through a reduction in procoagulant factors such as factor VIII [bib_ref] Statin therapy and levels of hemostatic factors in a healthy population: The..., Adams [/bib_ref] or other unknown mechanisms.
## The value of measuring thrombin generation in nafld
All in all, monitoring thrombin generation in patients with NAFLD could be a new laboratory tool for gaining insights into the pathophysiology of the disease and for helping to improve its treatment. Clinical trials with prospective design are, however, needed to establish conclusively any cause-effect relationship. In this respect, the above observations suggest that thrombin generation, expressed as the ETP-TM ratio, may be considered to be a reliable candidate laboratory tool for assessing, by means of a single procedure, the extent of hypercoagulability in patients with NAFLD. Specifically, thrombin generation expressed as the ETP-TM ratio could be helpful not only to define the state of hypercoagulability in this category of patients, but also as a reliable index parameter to assist organizing clinical trials needed to establish the treatment of choice (e.g., anticoagulation, use of statins, etc.) that may help reduce hypercoagulability and, hence, the sequel of adverse events known to burden NAFLD, such as the high risk of cardiovascular diseases, and progression to atherosclerosis and liver fibrosis. In this respect, based on pathophysiological considerations, in the future, anticoagulation and/or statin use could be attractive therapeutic strategies in patients with NAFLD.
## Concluding remarks
NAFLD is the most common chronic liver disease in Western countries, and it is anticipated that it could become even more prevalent in the general population in parallel with an increase in the prevalence of metabolic diseases that are closely associated with NAFLD (e.g., obesity, type II diabetes, dyslipidemia, and arterial hypertension). Therefore, NAFLD is a formidable burden for health systems across the world. In addition to liver impairment, NAFLD is associated with cardiovascular diseases (these being the most important causes of death in this condition), including atrial fibrillation, atherosclerosis progression, and venous thromboembolism that are all plausibly associated with hypercoagulability, which may, in turn, develop as a consequence of an imbalance of pro-vs. anticoagulants and the presence of such molecular species as procoagulant microvesicles, NETs, and inflammation that are associated with NAFLD. In addition, all these conditions may require anticoagulant therapy. The assessment of hypercoagulability by means of thrombin generation, a global procedure that mimics the coagulation process occurring in vivo much better than any other coagulation test, may be considered to be a candidate laboratory tool for assessing, with a single procedure, the balance of coagulation in NAFLD. In addition to defining the state of hypercoagulability, the assessment of thrombin generation could also be used to investigate patients with NAFLD enrolled in clinical trials, with the aim of determining the best approach (therapeutic and/or lifestyle change) for minimizing hypercoagulability and the risk of atherosclerosis progression and venous thrombosis in this condition.
Author Contributions: Conceptualization, data curation, and writing, A.T.; data curation, R.L., M.P., V.L.M. and A.L.F.; revision of the manuscript, R.L., M.P., V.L.M., F.P. and A.L.F. All authors have read and agreed to the published version of the manuscript.
[fig] Figure 2: Schematic representation of the thrombin generation curve (thrombogram) with relevant parameters. [/fig]
[fig] Figure 3: Schematic representation of the activation of protein C on the membrane of endothelial cells and the mechanism of action of the complex of activated protein C/protein S in inhibiting factor Va and factor VIIIa. Instrumental to the mechanism are two endothelial receptors, thrombomodulin (TM) that binds thrombin (Th) and endothelial protein C receptor (EPCR) that binds plasma protein C (PC). The proximity of the above receptors localizes the conversion of the substrate (PC) by the enzyme (Th) into activated PC (APC). APC in complex with its plasmatic cofactor protein S (PS) eventually inhibits factors Va and VIIIa, thus, downregulating thrombin production. [/fig]
[fig] Figure 4: Correlation between coagulation parameters and the ETP-TM ratio in patients wit NAFLD and health subjects. The ETP-TM ratio represents the ratio of endogenous thrombi potential (ETP) measured in the presence/absence of thrombomodulin (TM). Adapted wit permission from ref.[41]. Copyright 2014 Elsevier. [/fig]
|
Elevated peripheral expression of neuregulin-1 (NRG1) mRNA isoforms in clozapine-treated schizophrenia patients
Differential expression of neuregulin-1 (NRG1) mRNA isoforms and proteins has been reported in schizophrenia, primarily in post-mortem brain tissue. In this study, we examined 12 NRG1 SNPs, eight NRG1 mRNA isoforms (type I, type I (Ig2) , type II, type III, type IV, EGFα, EGFβ, pan-NRG1) in whole blood, and NRG1-β1 protein in serum of clozapinetreated schizophrenia patients (N = 71) and healthy controls (N = 57). In addition, using cultured peripheral blood mononuclear cells (PBMC) from 15 healthy individuals, we examined the effect of clozapine on NRG1 mRNA isoform and protein expression. We found elevated levels of NRG1 mRNA, specifically the EGFα (P = 0.0175), EGFβ (P = 0.002) and type I (Ig2) (P = 0.023) containing transcripts, but lower NRG1-β1 serum protein levels (P = 0.019) in schizophrenia patients compared to healthy controls. However, adjusting for smoking status attenuated the difference in NRG1-β1 serum levels (P = 0.050). Examination of clinical factors showed NRG1 EGFα (P = 0.02) and EGFβ (P = 0.02) isoform expression was negatively correlated with age of onset. However, we found limited evidence that NRG1 mRNA isoform or protein expression was associated with current chlorpromazine equivalent dose or clozapine plasma levels, the latter corroborated by our PBMC clozapine exposure experiment. Our SNP analysis found no robust expression quantitative trait loci. Our results represent the first comprehensive investigation of NRG1 isoforms and protein expression in the blood of clozapine-treated schizophrenia patients and suggest levels of some NRG1 transcripts are upregulated in those with schizophrenia.
# Introduction
Neuregulin-1 (NRG1) is vital for neurodevelopment and plasticity 1 , making it an appealing gene to examine in schizophrenia. This appeal has been weakened by genome-wide association study results that have failed to identify it as a top schizophrenia ''risk'' gene 2 ; questioning the relevance of NRG1 in schizophrenia [bib_ref] Neuregulin-1 and schizophrenia in the genome-wide association study era, Mostaid [/bib_ref]. However, the relevance of any gene to schizophrenia should not be determined exclusively on whether sequence variations within the gene meet genome-wide significance but rather on the compendium of knowledge available for that gene. A recent meta-analysis 4 and systematic review [bib_ref] Neuregulin-1 and schizophrenia in the genome-wide association study era, Mostaid [/bib_ref] have showed a number of NRG1 genetic variants as well as mRNA and protein levels associated with schizophrenia in specific populations or in certain contexts, which could, in part, be attributed to the complex and highly interactive nature of the NRG-ErbB signaling pathway [bib_ref] Neuregulin 1 in neural development, synaptic plasticity and schizophrenia, Mei [/bib_ref] [bib_ref] Neuregulin 1-erbB signaling and the molecular/cellular basis of schizophrenia, Corfas [/bib_ref]. Nevertheless, the bulk of the evidence to date suggests NRG1 remains an important target for schizophrenia research.
Post-mortem human brain studies in schizophrenia have shown differential expression of NRG1 mRNA and protein in various brain regions, most notably in dorsolateral prefrontal cortex and hippocampus [bib_ref] Expression analysis of neuregulin-1 in the dorsolateral prefrontal cortex in schizophrenia, Hashimoto [/bib_ref] [bib_ref] Elevated neuregulin-1 and ErbB4 protein in the prefrontal cortex of schizophrenic patients, Chong [/bib_ref] [bib_ref] Immunohistochemical evidence for impaired neuregulin-1 signaling in the prefrontal cortex in schizophrenia..., Bertram [/bib_ref] [bib_ref] Neuregulin 1 transcripts are differentially expressed in schizophrenia and regulated by 5'..., Law [/bib_ref] , although other studies of both regions have been negative [bib_ref] Altered neuregulin 1-erbB4 signaling contributes to NMDA receptor hypofunction in schizophrenia, Hahn [/bib_ref] [bib_ref] Levels of neuregulin 1 and 3 proteins in Brodmann's area 46 from..., Boer [/bib_ref] [bib_ref] Gene expression of neuregulin-1 isoforms in different brain regions of elderly schizophrenia..., Parlapani [/bib_ref] [bib_ref] A 5' promoter region SNP in NRG1 is associated with schizophrenia risk..., Nicodemus [/bib_ref] [bib_ref] Schizophrenia-associated HapICE haplotype is associated with increased NRG1 type III expression and..., Weickert [/bib_ref]. Similar evidence of differential gene and protein expression in the peripheral tissue of schizophrenia patients is also available. NRG1 mRNA expression, specifically type II β3 and NRG1 type III isoforms, have also been shown to be increased in peripheral leukocytes in Portuguese schizophrenia patients [bib_ref] Support for involvement of neuregulin 1 in schizophrenia pathophysiology, Petryshen [/bib_ref] and pan-NRG1 was shown to be decreased in Chinese schizophrenia patients compared to healthy controls [bib_ref] Explorative study on the expression of neuregulin-1 gene in peripheral blood of..., Zhang [/bib_ref]. Furthermore, the only two protein studies of NRG1 in peripheral samples found decreased plasma NRG1-β1 [bib_ref] Decreased plasma levels of neureglin-1 in drug naive patients and chronic patients..., Wang [/bib_ref] and serum Ig-NRG1 levels [bib_ref] Measurement and comparison of serum neuregulin 1 immunoreactivity in control subjects and..., Shibuya [/bib_ref] in people with schizophrenia relative to healthy controls. Collectively, these studies suggest NRG1 may be dysregulated in brain and blood at both the mRNA and protein level in schizophrenia and flag peripheral blood as a potential surrogate for brain NRG1 dysregulation [bib_ref] Comparison of peripheral and central schizophrenia biomarker profiles, Harris [/bib_ref]. However, the number of peripheral blood studies is limited and the influence specific clinical subgroups (e.g., treatmentresistant), genetic variation, medication, lifestyle (e.g., smoking), and/or symptom severity may have on NRG1 mRNA and protein expression is not clear.
The aim of this study was to address these gaps in the current literature by investigating peripheral mRNA and protein levels of NRG1 in schizophrenia, as peripheral measures have the potential to serve as biomarkers in the clinical setting. We particularly focused on patients being treated with clozapine. Clozapine is the drug of choice for a subgroup of schizophrenia patients who do not respond to other antipsychotics, referred to as treatment-resistant schizophrenia [bib_ref] Treatment-resistant schizophrenia: Treatment Response and Resistance in Psychosis (TRRIP) Working Group Consensus..., Howes [/bib_ref]. Herein, we examined in whole blood, several NRG1 mRNA isoforms, and NRG1-β1 protein levels in serum within those with schizophrenia compared to healthy controls. We also explore how these expression levels relate to symptom severity, age of onset, duration of illness, and NRG1 genetic variation as well as examine clozapine's effect on NRG1 mRNA and protein expression in peripheral blood mononuclear cells (PBMCs) from healthy control subjects.
# Materials and methods
## Participants
## Clinical samples
Seventy-one individuals with schizophrenia were recruited from inpatient and outpatient clinics located in Melbourne, Australia. Inclusion criteria included: (1) diagnosis of schizophrenia, (2) currently prescribed and taking clozapine, and (3) aged between 18-65 years. Fiftyseven unrelated healthy controls matched for age and sex with similar socio-economic backgrounds were recruited from the general community. Controls with a first-degree family history of psychiatric illness, neurological disease, head injury, seizures, prior or current use of antipsychotic medication, impaired thyroid function and/or substance abuse/dependence were excluded from the study. Participant characteristics are shown in [fig_ref] Table 1: Demographic data and clinical characteristics of participants [/fig_ref].
All participants were administered the Mini International Neuropsychiatric Interview (MINI) [bib_ref] The Mini-International Neuropsychiatric Interview (M.I.N.I.): the development and validation of a structured..., Sheehan [/bib_ref] to confirm the diagnosis of schizophrenia and to rule out current or past psychiatric illness in healthy controls. Clinical symptoms were assessed using the Positive and Negative Syndrome Scale (PANSS) [bib_ref] The positive and negative syndrome scale (PANSS) for schizophrenia, Kay [/bib_ref] and scored in accordance with the consensus five-factor (i.e., positive, negative, depressed, [bib_ref] Searching for a consensus five-factor model of the Positive and Negative Syndrome..., Wallwork [/bib_ref]. Tobacco, alcohol, and illicit drug use in the past 3-months was collected using a substance use questionnaire. Blood was collected after overnight fasting and processed according to standardized blood collection and processing protocol (see Supplementary Methods for more details). Clozapine plasma level was measured and current chlorpromazine equivalent dosage (except clozapine) was calculated in all patients by following standard guidelines [bib_ref] Chlorpromazine equivalent doses for the newer atypical antipsychotics, Woods [/bib_ref]. The study was approved by the Melbourne Health Human Research Ethics Committee (MHREC ID 2012.069), and all participants provided written informed consent prior to participation.
## In vitro clozapine exposure samples
Fresh frozen human PBMCs were obtained from 15 healthy donors (eight males and seven females) of Caucasian ethnicity with a mean age of 35 (sd = 13.5; range 20-54 years) from STEMCELL™ Technologies Inc. (Vancouver, British Columbia, Canada). One-third (n = 5) of the PBMC donors were current smokers. All the donors tested negative for HIV-1, HIV-2, Hepatitis B, and Hepatitis C. Sample size calculations showed 15 samples were sufficient to detect a large effect (Cohen's d = 0.80) between exposed and unexposed conditions at α = 0.05 and power = 0.80.
PBMCs were isolated from peripheral blood and were supplied as vials of 100 million cells. PBMCs were seeded at a concentration of 2 million cells per well (1 × 10 6 cells/ mL) in triplicate in six-well plates and incubated in RPMI-1640 medium (Sigma-Aldrich; St. Louis, Missouri, USA) supplemented with L-glutamine (0.3 g/L) and sodium bicarbonate (2 g/L), penicillin (100units/mL), streptomycin (100 µg/mL), 10% fetal bovine serum for 24 h. PBMCs were then exposed to clozapine (Sigma-Aldrich, St. Louis, Missouri, USA) at a concentration of 1.2 µM (control wells exposed to vehicle only, see Supplementary Methods for details) and incubated at 37°C in 5% CO 2 . Absolute ethanol was used to dissolve clozapine and media was used for dilution. The concentration of clozapine used was determined from the mean plasma concentration of clozapine found in the first 48 recruited clinical samples (1.2 µM or 384 ng/mL). Toxicity assays (CytoTox 96® Non-Radioactive Cytotoxicity Assay; Promega Corporation, Madison, Wisconsin, USA) were performed at baseline, 24 h and 7-day time points to measure the production of lactate dehydrogenase within the media (see [fig_ref] Figure 1: Normalized relative quantities [/fig_ref] for more details).
## Snp selection, dna extraction, and genotyping
Fourteen NRG1 single-nucleotide polymorphisms (SNPs) were selected based on their previously reported associations with schizophrenia (for review see refs. [bib_ref] Neuregulin-1 and schizophrenia in the genome-wide association study era, Mostaid [/bib_ref] [bib_ref] Meta-analysis reveals associations between genetic variation in the 5' and 3' regions..., Mostaid [/bib_ref] along with 60 unlinked ancestry-informative markers [fig_ref] Table 1: Demographic data and clinical characteristics of participants [/fig_ref] representing the three HapMap phase III populations (Northern/Western European, Han Chinese, and Yoruba in Nigeria) [bib_ref] Using ancestryinformative markers to define populations and detect population stratification, Enoch [/bib_ref]. DNA extraction and quantification were performed using standard procedures described in detail in the Supplementary Methods. SNPs were genotyped at the Australian Genome Research Facility (Brisbane, Australia) with the Sequenom Mas-sARRAY MALDI-TOF genotyping system using Sequenom iPLEX Gold chemistries according to manufacturer's instructions (Sequenom, Inc., San Diego, CA). Two (rs113317778 and rs6150532) of the 14 NRG1 SNPs assessed failed (0% call rate) but call rates for all remaining SNPs including the 60 ancestry markers were > 96% [fig_ref] Table 1: Demographic data and clinical characteristics of participants [/fig_ref].
## Rna extraction and gene expression analysis
RNA extraction and quantification for both clinical and in vitro samples were performed using PureLink RNA Mini Kit (ThermoFisher scientific, Waltham, MA, USA) per the standard manufacturer's instructions. Total RNA from both clinical and in vitro samples was reverse transcribed to cDNA using SuperScript® IV First-Strand Synthesis System (Invitrogen, Foster city, CA, USA) using random hexamers. cDNA (10.25 ng) was used as a template for quantitative reverse transcriptase (RT-qPCR) using master-mix and gene specific validated Taqman assays from Applied Biosystems, Foster City, California, USA. Custom designed primer and probe combinations were used for NRG1 isoforms (type I (Ig2) , type II and type IV) previously investigated 9,14,28 , while inventoried assays (TaqMan®, aInvitrogen, USA) were used for all other NRG1 isoforms (type III, Pan-NRG1, type I, EGFα and EGFβ) and four reference genes (beta-actin, ACTB; ubiquitin C, UBC; glyceraldehyde-3-phosphate dehydrogenase, GAPDH; and TATA box-binding protein, TBP). NRG1 mRNA isoforms were selected based on reported associations in previous gene expression experiments using post-mortem brain tissue or peripheral blood from schizophrenia patients [bib_ref] Neuregulin 1 transcripts are differentially expressed in schizophrenia and regulated by 5'..., Law [/bib_ref] [bib_ref] Schizophrenia-associated HapICE haplotype is associated with increased NRG1 type III expression and..., Weickert [/bib_ref] [bib_ref] Support for involvement of neuregulin 1 in schizophrenia pathophysiology, Petryshen [/bib_ref] [bib_ref] Dysbindin-1 and NRG-1 gene expression in immortalized lymphocytes from patients with schizophrenia, Yamamori [/bib_ref]. See Supplementary and [fig_ref] Figure 2: Distribution of age of onset at first diagnosis a and the correlation... [/fig_ref] for a list and genomic locations of each of the NRG1 probes and primers.
Gene expression levels were determined in duplicate using FAM-MGB TaqMan® gene expression probes (Invitrogen, Foster city, CA, USA) in 192 × 24 Dynamic Arrays IFC in Fluidigm® BioMark™ HD system (South San Francisco, CA, USA) at the Monash Health Translation Precinct Medical Genomics Facility (Hudson Institute of Medical Research, Clayton, VIC, Australia). In addition, no reverse transcriptase controls and no template controls were included to rule out genomic DNA contamination and reagent contamination, respectively. Adhering to minimum information for publication of RT-qPCR (MIQE) guidelines [bib_ref] The MIQE guidelines: minimum information for publication of quantitative real-time PCR experiments, Bustin [/bib_ref] , normalized relative quantities (NRQ, i.e., 2 -ΔCt where ΔC t = Ct (candidate gene) -Ct (geometric mean of reference genes) ) of each NRG1 mRNA isoform was calculated using the geometric mean expression of two reference genes (ACTB and UBC) that did not differ between groups in either the clinical or in vitro cohorts, with the exception of the 24-hour in vitro time point for which no reference gene was stable. GAPDH and TBP were not used as reference genes because their expression differed significantly by group in both the clinical and in vitro cohorts (Supplementary Figs. S3-S5).
## Protein quantification clinical samples
Human NRG1-β1 ELISA kits (Catalog number: EHNRG1; ThermoFisher Scientific™, Life Technologies®, Waltham, MA, USA) were used to measure NRG1-β1 levels in serum according to the manufacturer's protocol (see Supplementary Methods for details). In brief, 100 ul of serum or standard NRG1-β1 (0 pg/mL-20,000 pg/mL) was added to the wells in duplicate. The ELISA kits have a sensitivity of 50 pg/mL. Absorbance was measured on a SpectraMax® M3 multi-mode microplate reader (Molecular Devices, LLC; Sunnyvale, CA, USA) at 450 nm and 550 nm wavelength. The 550 nm values were subtracted from the 450 nm values to correct for optical imperfections in the microplate. A standard curve was generated for each assay by plotting mean absorbance for each standard concentration vs. the corresponding NRG1-β1 concentration. The standard curve (r 2 ≥ 0.99) was generated with a four-parameter logistic curve fit. The concentration of NRG1-β1 in the serum samples was obtained by interpolating the absorbance values using the standard curve in GraphPad Prism 6.
## In vitro samples
The same NRG1-β1 ELISA kit used for the clinical samples was also used for the in vitro samples. Prior to ELISA, protein lysates were prepared from both the 24hour clozapine exposed and control cells using RIPA buffer (Sigma-Aldrich®, Saint Louis, Missouri, USA). Due to limited baseline quantity of cells, protein lysates at 7 days were not available. The amount of total protein was quantified from each cell lysate using the Thermo-Scientific™ Pierce™ BCA Protein Assay Kit (Thermo-Fisher Scientific, MA, USA). Absorbance was measured at 562 nm using SpectraMax® M3 microplate reader. A standard curve (r 2 ≥ 0.99) was generated by plotting the absorbance value at 562 nm for each bovine serum albumin (BSA) standard vs its concentration (µg/mL). The total protein concentration of each unknown sample was determined using the standard curve. Five microgram of total cell lysate samples were mixed with appropriate amount 1x assay diluent B to be used in the ELISA system. One-hundred microliters of total cell lysate samples (0.05 µg/µL) or standard NRG1-β1 (0 pg/mL-20,000 pg/ mL) was added to the wells in duplicate. Assay diluent B was used to prepare standards and served as the zero standards (0 pg/mL).
# Statistical analysis
Two-tailed tests were used for all statistical analyses. Quantile-quantile (Q-Q) plots and the Shapiro-Wilk test were used to assess normality of variable distributions. Student's t-tests were used to test differences for continuous variables between schizophrenia patients and healthy controls, while chi-squared (χ 2 ) tests were used for categorical variables. The Benjamini and Hochberg (B-H) step-up procedure 32 was used to adjust for multiple comparisons for all analyses. Effect sizes were calculated using the Hedges' g method 33 .
# Nrg1 isoform/protein analysis
Prior to analysis, the normalized relative quantity data for all the NRG1 isoforms and the NRG1-β1 data were checked for normality using Q-Q plots and as required were log10 transformed for subsequent analysis. The log-transformed values were compared among groups using a general linear model (GLM), with the group as a fixed factor and age, gender, RNA integrity number ((RIN) (isoforms only)), and current smoking status as covariates. Despite differences in alcohol use between the schizophrenia and control groups, alcohol was not included as a covariate because it had no effect on NRG1 isoform or protein expression (see . For protein analysis, we used the generalized linear model, as NRG1-β1 levels were not normally distributed (see . Outliers were identified using the Grubbs' test for outliers and removed from further analysis.
Within the schizophrenia group, Pearson correlations between NRG1 isoform/protein levels and symptom severity, age of onset, illness duration, current chlorpromazine equivalent dose, and clozapine plasma levels were assessed. In addition, NRG1 isoform/protein levels between participants in positive symptom remission and non-remission were assessed using a t-test. Positive symptom remission was defined as a score of ≤ 3 on four PANSS items (delusions, hallucinations, grandiosity, and unusual thought content) [bib_ref] Searching for a consensus five-factor model of the Positive and Negative Syndrome..., Wallwork [/bib_ref].
SNP and haplotype analysis NRG1 SNPs were mapped using the GRCh38/hg19 human genome reference assembly. Linkage disequilibrium (LD) between SNPs was examined in Haploview and haplotype blocks determined using the solid spine method [bib_ref] Haploview: analysis and visualization of LD and haplotype maps, Barrett [/bib_ref]. For each individual, haplotypes were determined based on the best posterior probability procedure implemented in PLINK 1.07 [bib_ref] PLINK: a tool set for whole-genome association and population-based linkage analyses, Purcell [/bib_ref]. GLMs were used to explore cis-regulatory effects of NRG1 SNPs and haplotypes on isoforms and protein expression. Each GLM included genotype/haplotype, case status, genotype/haplotype x case status as well as other relevant covariates (age, gender, RIN). Significant genotype/haplotype x case status interactions were analyzed post hoc by case status stratification analyses.
## In vitro clozapine exposure analysis
Linear mixed models were used to determine the differences in gene expression over two-time points. In this model, the difference in transcript levels was the outcome variable and was adjusted for age, gender, and RIN. Due to non-normal distributions, Wilcoxon matched pair t-test was used to measure the difference in gene expression between clozapine exposed and unexposed cells at each time point.
# Results
## Nrg1 mrna expression
Among the eight NRG1 mRNA isoforms interrogated, four of them (type I, type II, pan-NRG1, and type IV) were not detectable in more than 60% of the full cohort and so were removed from further analysis. Rates of nondetection were evenly distributed between cases and controls, with the exception of NRG1 type II, which had a greater non-detect rate in controls (P = 0.00016, Supplementary . Among the remaining four NRG1 isoforms, levels of EGFα, EGFβ, and type I (Ig2) mRNA were significantly elevated and type III did not differ in schizophrenia patients compared to healthy controls after adjustment for covariates and correction for multiple testing [fig_ref] Figure 1: Normalized relative quantities [/fig_ref]. Importantly, gene expression levels of NRG1 isoforms were not correlated with clozapine plasma , which was further corroborated by our in vitro analysis that showed no difference in mRNA levels of detectable isoforms (EGFα, EGFβ, and type II) in clozapine exposed compared to unexposed PBMCs . Furthermore, within the patients with schizophrenia significant negative correlations between age of onset and NRG1 EGFα (r = −0.37, P raw = 0.002, P B-H = 0.02) and EGFβ (r = −0.36, P raw = 0.001, P B-H = 0.02) expression were detected [fig_ref] Figure 2: Distribution of age of onset at first diagnosis a and the correlation... [/fig_ref]. No significant correlations were observed between NRG1 isoforms and duration of illness after adjustment for multiple testing, although a trend-level negative correlation was found between NRG1 type III expression and duration of illness (r = −0.36, P raw = 0.027, P B-H = 0.167).
## Nrg1 protein expression
In contrast to the increase in NRG1 mRNAs, we found that NRG1-β1 protein levels were lower in schizophrenia patients compared to healthy controls (P = 0.019) but after adjustment for smoking status, this finding was attenuated (P = 0.050; . Current smokers had lower NRG1-β1 protein levels compared to non-smokers in the full cohort (P = 0.033, and schizophrenia participants were more likely to be current smokers as compared to controls (46.5% vs. 21.1%, P = 0.002, [fig_ref] Table 1: Demographic data and clinical characteristics of participants [/fig_ref]. However, clozapine plasma levels were not associated with NRG1-β1 expression (r = −0.023, P = 0.85), and there was no difference in NRG1-β1 protein levels between clozapine exposed and unexposed cells (exposed: median 2.31 log10 pg/mL, unexposed: median 2.2 log10 pg/mL; P = 0.191; . We found no association between NRG1-β1 protein levels and chlorpromazine equivalent antipsychotic exposure, age of onset, or illness duration .
## Nrg1 isoforms/protein expression and symptomatology
Significant negative correlations between NRG1 mRNA isoform EGFα expression and depression severity score (r = −0.241, P raw = 0.045, P B-H = 0.270) as well as type III expression and positive symptom severity score (r = −0.377, P raw = 0.020, P B-H = 0.120) were observed but did not survive correction for multiple comparisons (Supplementary . An exploratory examination of schizophrenia patients in positive symptom remission vs. non-remission revealed no statistically significant differences in levels of any of the NRG1 isoforms or NRG1-β1 serum protein after correction for multiple comparisons, although a trend (P raw = 0.013, P B-H = 0.065) toward greater NRG1 type III expression in remitters vs. non-remitters was observed .
Genotype and haplotype effects on NRG1 isoforms/protein expression LD analysis revealed two haplotype blocks among the 12 successfully genotyped SNPs . Block 1 contained four of the Icelandic schizophrenia-risk haplotype (Hap ICE ) SNPs (rs73235619, rs35753505, rs62510682, rs6994992) along with two other SNPs (rs4281084 & rs7014762) in the 5'-region and Block 2 included four SNPs (rs3924999, rs2439272, rs2954041, rs74942016) in the 3'-region of NRG1. Examination of these haplotypes, as well as each SNP, independently revealed several nominal NRG1 isoforms and protein expression quantitative trait loci but none survived correction for multiple comparisons .
# Discussion
Among the four detectable NRG1 isoforms in whole blood three (EGFα, EGFβ, and type I (Ig2) ) were elevated and one (type III) did not differ between clozapine-treated schizophrenia patients and healthy controls. Importantly, we could not attribute these overall increases in NRG1 mRNA levels to demographic characteristics and did not find a correlation with clozapine blood levels, suggesting that elevated NRG1 mRNA in whole blood may not be a direct consequence of age, sex, smoking, or exposure to clozapine; the latter supported by our in vitro experiments. However, age of illness onset was negatively correlated with expression of NRG1 EGFα and EGFβ containing trancripts, suggesting increased expression of these isoforms are assocated with an earlier age of illness onset.
To our knowledge, we are the first to report elevated levels of NRG1 EGFα, EGFβ, and type I (Ig2) in schizophrenia, specifically in those with treatment-resistant schizophrenia. Previous peripheral expression studies have not measured these isoforms [bib_ref] Support for involvement of neuregulin 1 in schizophrenia pathophysiology, Petryshen [/bib_ref] [bib_ref] Explorative study on the expression of neuregulin-1 gene in peripheral blood of..., Zhang [/bib_ref] [bib_ref] Dysbindin-1 and NRG-1 gene expression in immortalized lymphocytes from patients with schizophrenia, Yamamori [/bib_ref] , although a previous study examining peripheral expression of two other NRG1 isoforms (ndf43a and ndf43b) covered by the NRG1 EGFβ probe reported no difference between schizophrenia and control participants [bib_ref] Support for involvement of neuregulin 1 in schizophrenia pathophysiology, Petryshen [/bib_ref]. Furthermore, one post-mortem study, which specifically measured the EGF domain containing mRNAs reported no difference in NRG1 EGFβ levels and was unable to reliably detect NRG1 EGFα or type I (Ig2) in the dorsolateral prefrontal cortex of schizophrenia and control participants [bib_ref] Schizophrenia-associated HapICE haplotype is associated with increased NRG1 type III expression and..., Weickert [/bib_ref]. This suggests our findings, if extended, may contribute to a unique NRG1 mRNA alteration (signature) in the blood of individuals with schizophrenia or more specifically treatment-resistant schizophrenia. However, in the current study we were unable to compare individuals with and without treatment-resistant schizophrenia and as such the ability of these NRG1 isoforms to identify treatment-resistant schizophrenia patients remains to be confirmed.
We also found a novel and robust negative correlation between age of onset and expression of EGFα and EGFβ isoform levels, suggesting elevated levels of these isoforms were more frequently detected in those with an earlier age of illness onset. Interestingly, a previous post-mortem brain study [bib_ref] Schizophrenia-associated HapICE haplotype is associated with increased NRG1 type III expression and..., Weickert [/bib_ref] , showed that brain abundant NRG1 type III expression was negatively correlated with age of onset. These findings across two different cohorts and from two different cell populations, suggest a relationship between age of onset and NRG1 gene expression in both brain and blood such that higher gene expression of NRG1 may accelerate or serve to precipitate transition to full blown symptoms. While, these studies suggest that the distinct NRG1 isoforms may monitor clinically meaningful events in blood as compared to brain, and suggest the possibility that blood measures of NRG1 could serve as surrogate markers for NRG1 in brain. Future longitudinal studies are needed to explore whether elevated NRG1 gene expression is a precipitating factor and/or a consequence of an earlier age of onset.
Our analyses showed no difference in the expression of NRG1 type III between schizophrenia patients and controls, which do not concur with a previous study that reported increased expression of NRG1 type III in peripheral leukocytes of schizophrenia patients [bib_ref] Support for involvement of neuregulin 1 in schizophrenia pathophysiology, Petryshen [/bib_ref]. NRG1 type III is the most abundant of all NRG1 isoforms in the human brain and increased expression of this isoform was found to be associated with genetic variation in the NRG1 Hap ICE region [bib_ref] Neuregulin 1 transcripts are differentially expressed in schizophrenia and regulated by 5'..., Law [/bib_ref] [bib_ref] Schizophrenia-associated HapICE haplotype is associated with increased NRG1 type III expression and..., Weickert [/bib_ref]. However, we were unable to replicate the increase NRG1 type III in whole blood. We did however, find trend-level negative correlations between NRG1 type III expression and duration of illness and positive symptom severity as well as elevated expression in remitters, providing preliminary evidence that downregulation of this isoform in blood may occur with disease progression but this in turn is associated with greater positive symptom severity and lower likelihood of achieving positive symptom remission. To our knowledge no other blood-based study of NRG1 type III expression has examined these associations and as such it is not clear if they are unique to treatment-resistant schizophrenia or are generalizable to all those with a schizophrenia diagnosis.
We could not detect NRG1 type IV mRNA in any sample and pan-NRG1, type I, and type II mRNAs were not detectable in greater than 60% of our cohort. Our failure to detect NRG1 type IV and type I are in alignment with a previous study that failed to detect these isoforms in immortalized lymphocytes [bib_ref] Dysbindin-1 and NRG-1 gene expression in immortalized lymphocytes from patients with schizophrenia, Yamamori [/bib_ref]. However, pan-NRG1 was shown to be decreased in peripheral lymphocytes [bib_ref] Explorative study on the expression of neuregulin-1 gene in peripheral blood of..., Zhang [/bib_ref] and type II β3 increased in peripheral leukocytes [bib_ref] Support for involvement of neuregulin 1 in schizophrenia pathophysiology, Petryshen [/bib_ref] , suggesting detection of these isoforms in the periphery may depend on the cell populations examined. We did observe, however, a significantly lower frequency of type II non-detects among our schizophrenia group compared to controls (55% vs. 86%, P < 0.01, , indicating that there may be elevated type II expression in schizophrenia. Still, our low detection of pan-NRG1 was unexpected given that the probes for this transcript targeted both the Ig1 and Ig2 regions of the NRG1 gene, which all isoforms we measured contain (see Supplementary [fig_ref] Figure 2: Distribution of age of onset at first diagnosis a and the correlation... [/fig_ref] for the regions of amplification for each NRG1 isofrom). However, for all the isoforms with low detection the probes we used targeted the Ig1 region, suggesting this region of NRG1 may be downregulated in whole blood and resulted in lower amplification. Although, these probes and primers have been succesfully employed by our group in postmortem human brain, to our knowledge this is the first time they have been used in whole blood. Future studies using whole blood should consider alternative probes for these NRG1 isoforms.
In contrast to our mRNA findings, NRG1-β1 protein abundance was lower in clozapine-treated schizophrenia patients relative to controls. This finding did not appear to be related to demographic characteristics, genetic variation, or clozapine dose or clozapine blood levels (confirmed by our in vitro clozapine exposure experiment) but was attenuated after adjustment for smoking status. To our knowledge, previous studies measuring peripheral or central NRG1 protein levels have not accounted for smoking status as a potential confound. Our results suggest smokers, regardless of case status, have lower peripheral NRG1-β1 protein levels than non-smokers. Given that smoking prevalence rates are known to be significantly higher among individuals with schizophrenia compared to the general population [bib_ref] A meta-analysis of worldwide studies demonstrates an association between schizophrenia and tobacco..., De Leon [/bib_ref] , it is possible that previously reported differences in NRG1 protein levels between schizophrenia and control participants may have also been confounded by smoking. We are aware of two previous studies that have examined peripheral NRG1 protein levels in schizophrenia. The first reported lower Ig-NRG1 levels in serum 38 and the second reported lower NRG1-β1 protein in plasma from schizophrenia patients [bib_ref] Decreased plasma levels of neureglin-1 in drug naive patients and chronic patients..., Wang [/bib_ref]. However, neither study adjusted for smoking status in their analyses. It is not clear if smoking status would have similar effects on brain NRG1 protein levels reported in post-mortem studies [bib_ref] Immunohistochemical evidence for impaired neuregulin-1 signaling in the prefrontal cortex in schizophrenia..., Bertram [/bib_ref] [bib_ref] Levels of neuregulin 1 and 3 proteins in Brodmann's area 46 from..., Boer [/bib_ref] [bib_ref] Decreased neuregulin 1 C-terminal fragment in Brodmann's area 6 of patients with..., Barakat [/bib_ref] , as none of these studies examined this potential effect and it remains uncertain whether NRG1 protein levels in the brain concur with levels in the periphery. Nevertheless, our findings provide reasonable evidence for inclusion of smoking as a potential confound when measuring and interpreting NRG1 protein levels in groups with known differences in smoking prevalence and support future pre-clinical experiments assessing the effect of smoking on NRG1 protein levels in blood and brain.
Several limitations should be noted. First, the size of the cohort only allowed for detection of moderate to large differences in mRNA and protein abundances between groups. Second, the study employed a cross-sectional design, inhibiting examination of temporal expression patterns and how these patterns map on to clinical trajectories. Third, measurement of mRNA and protein expression occurred in whole blood and serum, respectively. Although both of these tissues are clinically accessible and commonly used in biomarker research, it is not fully clear how well our findings will generalize to other peripheral (e.g., plasma, lymphocytes) or central (e.g., brain) tissues despite some support for their applicability in schizophrenia [bib_ref] Comparison of peripheral and central schizophrenia biomarker profiles, Harris [/bib_ref]. In addition, the generalizability of our findings beyond those with treatmentresistant schizophrenia is not clear. Future studies comparing NRG1 gene and protein expression between treatment-resistant and non-resistance schizophrenia patients, including treatment-naive patients, are warranted. Finally, our in vitro clozapine exposure experiments examined a single clozapine concentration (1.2 µM) guided by pilot data from our study population. A previous study used higher clozapine concentrations (2 µM) for three weeks in cultured post-mortem human fetal brain tissue and showed an upregulation of NRG1 protein [bib_ref] Upregulation of NRG-1 and VAMP-1 in human brain aggregates exposed to clozapine, Chana [/bib_ref]. As such, future work with PBMCs should examine multiple concentrations that reflect the range of clozapine blood levels observed in the clinic. Future in vitro clozapine experiments with PBMCs should also screen a greater number of genes to identify more suitable NRG1-β1 protein expression between a schizophrenia and controls (schizophrenia: 2.55 ± 0.067 pg/mL, controls: 2.78 ± 0.077 pg/mL; P unadjusted = 0.019, P adjusted for smoking = 0.050, Hedges' g = 0.42) and b current smokers and non-smokers (smokers: 2.49 ± 0.083 pg/mL, non-smokers: 2.73 ± 0.064 pg/mL; t = −2.153, df = 118, P = 0.033, g = 0.43). Error bars represent mean ± s.e.m. *P < 0.05 references, particularly genes that are stable during acute clozapine exposure.
Despite these limitations, the current study represents the first comprehensive investigation of NRG1 isoforms and protein expression in whole blood of clozapinetreated schizophrenia patients. In general, our results support the notion posed by previous peripheral blood and post-mortem brain studies that NRG1 transcription is dysregulated in schizophrenia and perhaps more specifically treatment-resistant schizophrenia. However, we have also expanded on this by showing NRG1 mRNA isoforms EGFα, EGFβ, and type I (Ig2) are elevated in whole blood of clozapine-treated schizophrenia patients. Our findings further suggest that NRG1 expression is associated with age of onset, particularly NRG1 EGFα and EGFβ isoforms and that NRG1 type III expression may vary by disease progression. As such our results suggest that NRG1 overexpression may not be restricted to the brain of those with schizophrenia, and blood-based NRG1 transcription may serve, in part, as a suitable biomarker for schizophrenia and perhaps treatment-resistant schizophrenia.
[fig] Figure 1: Normalized relative quantities (NRQ) of NRG1 mRNA isoforms. a NRG1 EGFα (schizophrenia: 1.30 ± 0.12, controls: 0.93 ± 0.22; F 1,125 = 7.56, P = 0.0175, Hedges' g = 0.33); b NRG1 type I (Ig2) (schizophrenia: 2.23 ± 0.21, controls: 1.61 ± 0.22; F 1,103 = 6.261, P = 0.023, g = 0.70); c NRG1 EGFβ (schizophrenia: 10.95 ± 0.99, controls: 7.30 ± 2.08; F 1,122 = 13.14, P = 0.002, g = 0.66); and d NRG1 type III (schizophrenia: 12.22 ± 2.23, controls: 11.3 ± 2.80; F 1,59 = 4.23E −10 , P = 1.0, g = 0.02). Error bars represent mean ± s.e.m. Benjamini-Hochberg adjusted P-values are shown. *P < 0.05 and **P < 0.01 levels or chlorpromazine equivalent antipsychotic exposure (excluding clozapine) (Supplementary [/fig]
[fig] Figure 2: Distribution of age of onset at first diagnosis a and the correlation between age of onset and NRG1 EGFα (r = −0.37, P raw = 0.002, P B-H = 0.02) b and NRG1 EGFβ (r = −0.36, P raw = 0.001, P B-H = 0.02) c expression. NRG1 isoform expression is represented as the standardized residual from a linear regression model after adjusting for the significant effect of age on expression. Dotted lines represented the 95% confidence intervals for the linear regression line (solid black line) [/fig]
[table] Table 1: Demographic data and clinical characteristics of participants [/table]
|
Localizing evoked and induced responses to faces using magnetoencephalography
A rich pattern of responses in frequency, time and space are known to be generated in the visual cortex in response to faces. Recently, a number of studies have used magnetoencephalography (MEG) to try to record these responses non-invasivelyin many cases using source analysis techniques based on the beamforming method. Here we sought both to characterize best practice for measuring face-specific responses using MEG beamforming, and to determine whether the results produced by the beamformer match evidence from other modalities. We measured activity to visual presentation of face stimuli and phase-scrambled control stimuli, and performed source analyses of both induced and evoked responses using Synthetic Aperture Magnetometry. We localized the gamma-band response to bilateral lateral occipital cortex, and both the gamma-band response and the M170evoked response to the right fusiform gyrus. Differences in the gamma-band response between faces and scrambled stimuli were confined to the frequency range 50-90 Hz; gamma-band activity at higher frequencies did not differ between the two stimulus categories. We additionally identified a component of the M220-evoked responselocalized to the parieto-occipital sulcuswhich was enhanced for scrambled vs. unscrambled faces. These findings help to establish that MEG beamforming can localize facespecific responses in time, frequency and space with good accuracy (when validated against established findings from functional magnetic resonance imaging and intracranial recordings), as well as contributing to the establishment of best methodological practice for the use of the beamformer method to measure face-specific responses.
# Introduction
The cortical basis of face perception has been a topic of interest for a number of decades. Early work based on primate neurophysiology provided evidence that there might be identifiable cortical regions devoted specifically to the processes of face perception [bib_ref] Visual receptive fields of neurons in inferotemporal cortex of the monkey, Gross [/bib_ref] [bib_ref] Visual properties of neurons in inferotemporal cortex of the Macaque, Gross [/bib_ref]. The introduction of functional magnetic resonance imaging (fMRI) allowed the identification of an extended network of face processing regions, including the fusiform gyrus, lateral occipital cortex and the superior temporal sulcus (for a review see [bib_ref] The distributed human neural system for face perception, Haxby [/bib_ref]. The relatively poor temporal resolution of the fMRI blood oxygen-level dependent (BOLD) signal makes the technique ill-suited to measuring the temporal dynamics of cortical processing, however. This is unfortunate as intracranial electroencephalography (EEG) in humans has revealed a rich pattern of both evoked and induced responses to faces [bib_ref] Electrophysiological studies of human face perception. II: response properties of face-specific potentials..., Mccarthy [/bib_ref] [bib_ref] The many faces of the gamma band response to complex visual stimuli, Lachaux [/bib_ref] [bib_ref] Spatio temporal dynamics of face recognition, Barbeau [/bib_ref] [bib_ref] Decoding face information in time, frequency and space from direct intracranial recordings..., Tsuchiya [/bib_ref] [bib_ref] Neural "ignition": enhanced activation linked to perceptual awareness in human ventral stream..., Fisch [/bib_ref] [bib_ref] Selective attention modulates face-specific induced gamma oscillations recorded from ventral occipitotemporal cortex, Engell [/bib_ref] [bib_ref] Category-specific visual responses: an intracranial study comparing gamma, beta, alpha, and ERP..., Vidal [/bib_ref] [bib_ref] Antagonistic relationship between gamma power and visual evoked potentials revealed in human..., Privman [/bib_ref] [bib_ref] Exemplar selectivity reflects perceptual similarities in the human fusiform cortex, Davidesco [/bib_ref].
Magnetoencephalograpy (MEG) currently offers the best potential for the non-invasive study of face processing with high temporal resolution. Of particular interest are the gamma-band responsewhich has been shown to involved in the processing of facial emotion [bib_ref] Neural dynamics for facial threat processing as revealed by gamma band synchronization..., Luo [/bib_ref] [bib_ref] Visual awareness, emotion, and gamma band synchronization, Luo [/bib_ref] [bib_ref] Early gamma-band activity as a function of threat processing in the extrastriate..., Maratos [/bib_ref] and appears to be reduced in prosopagnosia [bib_ref] The role of gamma-band activity in the representation of faces: reduced activity..., Dobel [/bib_ref] and in other conditions which are characterized by face processing deficits such as autism [bib_ref] Gamma activation in young people with autism spectrum disorders and typically-developing controls..., Wright [/bib_ref] and schizophrenia [bib_ref] Deficits in high-(>60 Hz) gamma-band oscillations during visual processing in schizophrenia, Gr€ Utzner [/bib_ref] and the M170 event-related field (ERF)which is known to be face-specific [bib_ref] Face-specific responses from the human inferior occipito-temporal cortex, Sams [/bib_ref] [bib_ref] Cognitive response profile of the human fusiform face area as determined by..., Halgren [/bib_ref] [bib_ref] The selectivity of the occipitotemporal M170 for faces, Liu [/bib_ref] and may also be altered in prosopagnosia [bib_ref] Normal and abnormal face selectivity of the M170 response in developmental prosopagnosics, Harris [/bib_ref] [bib_ref] Early lefthemispheric dysfunction of face processing in congenital prosopagnosia: an MEG study, Dobel [/bib_ref] [bib_ref] Early (N170/ M170) face-sensitivity despite right lateral occipital brain damage in acquired..., Prieto [/bib_ref] [bib_ref] Investigating the features of the M170 in congenital prosopagnosia, Rivolta [/bib_ref].
Many of these previous studies have made use of beamforming, a well-established technique for MEG source localization that has been used extensively to localize visually evoked and induced responses (see for instance [bib_ref] Induced visual illusions and gamma oscillations in human primary visual cortex, Adjamian [/bib_ref] [bib_ref] Localizing human visual gamma-band activity in frequency, time and space, Hoogenboom [/bib_ref] [bib_ref] Visual gamma oscillations and evoked responses: variability, repeatability and structural MRI correlates, Muthukumaraswamy [/bib_ref]. However, the choice of various aspects of the beamformer analysisin particular the choice of time and frequency windows of interest and the choice of relevant statistical comparisonscan have a critical impact on the results found using the technique. Yet we are not aware of any general attempt to characterize best practice for measuring face-specific responses using MEG beamforming, nor of any attempt to determine whether the results produced by the beamformer match evidence from other modalities such as intracranial recordings or fMRI.
Here we show that the MEG beamformer is able to localize responses to faces in time, frequency and space in a manner that is consistent with findings from other modalities. In doing so we demonstrate that, particularly for the gamma-band response, statistical comparison against a control stimulus (phase-scrambled faces) and precise choice of frequency bandwidth are critical to accurately finding face-specific responses.
# Materials and methods
## Participants
The participants were 20 volunteers (age range: 21-43 years, mean: 29 years) with normal or corrected-to-normal vision (based on selfreport). The participant sample contained equal numbers of males and females. Each participant gave written consent to take part in the study in accordance with The Code of Ethics of the World Medical Association (Declaration of Helsinki). All procedures were approved by the ethics committee of the School of Psychology, Cardiff University.
## Stimuli
Face stimuli were 30 images each showing a front profile view of one of 15 females and 15 males, taken from the ECVP Utrecht face set (available at http://pics.psych.stir.ac.uk/zips/utrecht.zip). Images were 576 9 768 pixels in size and converted to 8-bit greyscale prior to use.
Control stimuli were created by phase scrambling each of the 30 images. Phase scrambling was achieved by performing a two-dimensional (2D) Fourier transform of each of the images, removing the phase information and replacing it with the corresponding phases taken from a white noise image of equal size to the original stimulus, and then performing the inverse Fourier transform to create the scrambled image. As phase scrambling can produce pixel intensities outside of the displayable range, pixel intensities of the scrambled images were linearly normalized (by subtraction of the minimum intensity and division by the sum of the minimum and maximum intensities) to within the displayable range. An identical normalization (that is using the minimum and maximum values taken from the matching scrambled image) was then applied to each of the original face stimuli to ensure that the low-level stimulus properties (i.e. the 2D Fourier amplitude spectrum, and the mean and standard deviation of the pixel intensities) were matched between each face stimulus and its scrambled counterpart. A new set of scrambled stimuli were generated for each participant.
## Data acquisition
Whole-head MEG recordings were made in 2.5-s epochs centred around the onset of each stimulus using a 275-channel CTF radial gradiometer system at a sampling rate of 1200 Hz. An additional 29 reference channels were recorded for noise cancellation purposes, and the primary sensors were analysed as synthetic third-order gradiometers [bib_ref] Signal processing in magnetoencephalography, Vrba [/bib_ref]. Two of the 275 channels were turned off due to excessive sensor noise, and a third sensor was turned off for the same reason during an annual system service which occurred part way through the study.
All stimuli were presented centrally on a mean grey background using a gamma-corrected Mitsubishu Diamond Pro 2070 CRT monitor with a screen resolution of 1024 9 768 pixels and a refresh rate of 100 Hz. The monitor was viewed from a distance of 2.1 m, with stimulus images subtending 8.3°9 6.1°of visual angle. In each trial participants viewed a white fixation cross for 1 s followed by one of the 60 stimulus images for 1 s, then a further period of fixation of random duration selected at each trial from a uniform distribution between 600 and 900 ms. Each stimulus was presented eight times during the experiment, leading to a total of 480 trials. Participants performed a change detection task in which, at the start of a pseudorandomly selected 10% of trials, the fixation cross turned red. The cross remained red until participants responded with a right index finger button press, with the trial 'proper' then beginning after a 500-ms delay. The experimental paradigm was implemented in Matlab (The Mathworks, Natick, MA, USA) using the Psychophysics Toolbox [bib_ref] The psychophysics toolbox, Brainard [/bib_ref] [bib_ref] The VideoToolbox software for visual psychophysics: transforming numbers into movies, Pelli [/bib_ref] [bib_ref] What's new in Psychtoolbox-3? Perception, Kleiner [/bib_ref].
At the start and end of each session each participant's head position was localized by means of three fiducial coils attached to specific anatomical locations on the scalp (nasion and left and right pre-auricular). The maximal head displacement between the start and end of a session in any participant was 1.2 cm. Fiduciary locations were verified afterwards using high-resolution digital photographs. These locations were then located on previously acquired anatomical MRIs of each participant for the purposes of source analysis.
# Data analysis
Data were manually inspected offline, and trials containing artefacts related to excessive muscle or head movements were excluded from analysis.
For evoked (but not induced) analysis data were bandpass filtered at 1-30 Hz (all filtering in the study was performed using thirdorder bi-directional IIR Butterworth filters, with the exception of filtering for source analysis where fourth-order filters were used). Evoked time windows of interest (TWOIs) were identified by averaging data across all trials for each participant to create an ERF timeseries for each sensor, then averaging these timeseries across participants to create a grand average ERF timeseries. These grand average timeseries were then baseline corrected against the 150 ms prior to stimulus onset, and the global field power (GFP) was calculated at each timepoint by calculating the root mean square (RMS) of the amplitude across sensors. TWOIs were then defined based on the local minima of the GFP timeseries.
For induced analysis, time-frequency spectrograms were calculated for each sensor (excluding the sensor which was turned off part way through the studysee Data acquisition section above) by bandpassing the data using filters of 4 Hz bandwidth centred on 2-160 Hz in 1 Hz steps and then using the Hilbert transform to compute the analytical signal at each frequency. Time-frequency data were than averaged across all trials for each participant. The magnitude of the response at each time-frequency point was calculated as the percentage amplitude change relative to the mean amplitude at that frequency during the time period À1 to 0 s relative to stimulus onset. To determine time-frequency windows of interest (TFWOIs), we used Wilcoxon signed-rank tests to identify time 9 frequency 9 sensor points for which the median magnitude across participants differed from zero. To correct for multiple comparisons we adjusted the alpha level of these tests to produce a false discovery rate of 0.05 [bib_ref] Thresholding of statistical maps in functional neuroimaging using the false discovery rate, Genovese [/bib_ref]. TFWOIs were then determined based on contiguous time-frequency regions that were significant over a substantial number of sensors (see Results).
Source analysis was performed using the Synthetic Aperture Magnetometry (SAM) beamformer algorithm, which is described in detail elsewhere [bib_ref] Functional neuroimaging by synthetic aperture magnetometry (SAM), Robinson [/bib_ref] [bib_ref] Signal processing in magnetoencephalography, Vrba [/bib_ref] [bib_ref] A new approach to neuroimaging with magnetoencephalography, Hillebrand [/bib_ref]. SAM operates by constructing an adaptive spatial filter based on a combination of a forward model and the data covariance matrix. Each participant had a previously acquired structural MRI, and a multiple local-spheres forward model [bib_ref] A sensor-weighted overlapping-sphere head model and exhaustive head model comparison for MEG, Huang [/bib_ref] was derived by fitting spheres to the brain surface extracted by the Oxford FMRIB Software Library's Brain Extraction Tool [bib_ref] Fast robust automated brain extraction, Smith [/bib_ref]. Data covariance matrices were calculated from the full dataset for each participant after bandpass filtering at either 1-30 Hz (in the case of the evoked analysis) or at specific frequencies of interest (for the induced analyses). For evoked analyses we followed the method suggested by [bib_ref] Localization of event-related activity by SAM(erf), Robinson [/bib_ref] and [bib_ref] Spatiotemporal mapping of cortical activity accompanying voluntary movements using an event-related beamforming..., Cheyne [/bib_ref] and calculated data covariance from unaveraged data. Volumetric images of source power were then derived at 4 mm isotropic resolution by projecting each participant's raw data through the corresponding spatial filters.
For each analysis, individual t-statistical images were calculated at each voxel for each participant using Welch's t-test for unequal variances:
[formula] t ¼ ðX f À X s Þ=ðS 2 f þ S 2 s Þ 1=2 [/formula]
, where X f and X s are the response measures to faces and scrambled stimuli, respectively, and S f and S s are the standard errors of those measures. For evoked analyses X was the mean power of the evoked time series within the time window of interest and S was estimated using a jackknife procedure (due to power being calculated after averaging over trials). For induced analyses X was the mean of the bandpass-filtered analytical time signal (generated using the Hilbert transform) averaged across trials within the time window of interest, and S was estimated from the standard deviation of the measure. All response measures, X, were baseline corrected (against the 150 ms prior to stimulus onset for evoked responses, and for time windows ending 200 ms prior to stimulus onset and of equal length to the time window of interest for induced responses).
To perform group-level analyses, individual-level t-statistic images were then normalized using FLIRT (www.fmrib.ox.ac.uk/ analysis/research/flirt) into MNI template space using an affine transform. A mask was used to exclude voxels outside of the brain.
The locations of any group-level differences in the response to faces and scrambled stimuli were then found using permutation testing [bib_ref] Nonparametric permutation tests for functional neuroimaging: a primer with examples, Nichols [/bib_ref] [bib_ref] Group imaging of taskrelated changes in cortical synchronisation using nonparametric permutation testing, Singh [/bib_ref]. For each analysis, group-level t-statistics were calculated for each voxel using participant-level t-statistics as the measure. Due to the focal nature of sources found in the evoked analysis, variance smoothing was applied using a Gaussian kernel (r = 12 mm) as this substantially improved the statistical power of the analysis. The distribution of group-level t-statistics was then estimated at each voxel by randomly flipping the signs of the individual-level test statistic (equivalent to randomly swapping the condition labels between face and scrambled stimuli) and recalculating the group-level t-statistic. Five thousand such permutations were generated and used to determine voxelwise P-values for the group-level difference between the response to faces and scrambled stimuli. Corrections for multiple comparisons were then performed by thresholding using the maximum test statistic.
For virtual sensor analyses, locations of interest were identified for each analysis by finding peaks in individual participants' t-statistical images. Sensor-level data were then spatially filtered using the beamformer weights at the location of interest to create a single timeseries per trial per participant per analysis. Evoked and induced analyses were then carried out in the manner outlined above for sensor-level data, but separately for face and scrambled trials. As SAM is a scalar beamformerwith source orientation set by a non-linear searchthe polarity of evoked timeseries is arbitrary (due to source direction and the sign of source moment being interchangeable) and so the polarities of all evoked timeseries in virtual sensor analyses were set such that the maximal deflection was positive within the TWOI.
# Results
## Induced responses
To perform source-level analysis of the induced responses, we first found TFWOIs in the sensor-level data. Rather than calculate TFW-OIs based on time-frequencies at which the responses to face and scrambled stimuli most differedwhich would potentially create a bias in favour of finding differences in the same time-frequency windows at the source level (and could constitute an instance of the statistically invalid practise of 'double dipping')we instead calculated the time-frequency response at each sensor for each participant across all trials. At each time 9 frequency 9 sensor point we tested against a median of zero across participants using a Wilcoxon signed-rank test (corrected for multiple comparisons using a false discovery rate of 0.05), to determine at which time-frequency points the response to the stimuli differed from zero independently of condition. [fig_ref] Figure 1: Plot of the number of sensors for which the response to stimuli [/fig_ref] shows the number of significant sensors at each time 9 frequency point. Based on this figure we determined three TFWOIs: 0-6 Hz, 100-600 ms; 15-40 Hz, 150-500 ms; 55-120 Hz, 100-400 ms. We refer to these time windows as: delta/ theta, beta and gamma, respectively.
We then used the SAM beamformer to produce volumetric images of source power for each condition (face and phase scrambled). Group-level analyses of within-participant differences between the two conditions revealed no significant difference in the delta/ theta or beta TFWOIs. In contrast, significant differences were found for the gamma TFWOI with enhanced response amplitudes found for faces relative to scrambled stimuli in a region of right lateral occipital cortex. The most significant voxel occurred on the In many previous studies in which MEG beamforming has been used to localize the gamma-band response to faces, control stimuli were not used and instead contrasts were performed between the response to faces vs. the pre-stimulus baseline. To compare our findings with this prior work, we repeated our gamma-band analysis using the same parameters as above, but contrasting the response to face stimuli with the response present in the stimulus baseline period rather than with the response to scrambled stimuli. Highly significant increases in gamma power to faces relative to baseline were found widely throughout occipital cortex , with the most significant voxels being found bilaterally around the occipital pole (Talairach coordinates ÀL: À11.0, À105.4, À3.0; R: 19.1, À103.4, 5.0). In contrast, no peak was present in the statistical parametric map around the middle occipital gyrus (or any other nearby location) as was found in the face vs. scrambled stimulus comparison. Instead, given their localization of the strongest effect to striate and/or neighbouring areas of extrastriate cortex, the findings from the face vs. baseline comparison suggest strongly that the largest gamma differences are generated by differences in image contrast between stimulus and baseline and not due to the presence (or absence) of a face. This has implications for previous studies of the gamma-band response to faces based on beamforming (see Discussion).
We were interested to explore the time course and frequency characteristics of the response found in the face vs. scrambled stimulus contrast, and so performed a virtual sensor analysis. For each participant we inspected the t-statistical image of the difference between conditions, and found the largest local maxima in the right lateral occipito-temporal region (two participants were excluded from this analysis due to absence of a local maximum in the area of interest). We then used the beamfomer weights to generate virtual sensor timeseries at this location. [fig_ref] Figure 4: Group average spectrograms depicting the virtual sensor responses to face [/fig_ref] demonstrates that, while both the face and the scrambled stimuli induce a broadband gamma response extending from around 50 Hz up to at least 150 Hz, there is little difference between conditions for frequencies above 80 Hz. This is confirmed in , which shows the group average amplitude spectrum across the time period 100-400 ms for both conditions. Significant differences between the two spectra were confined to the range of approximately 50-90 Hza narrower frequency bandwidth than used in our initial gamma TFWOI.
The time course of the gamma response was similar for both conditions , with a rapid rise in amplitude soon after stimulus onset, reaching a peak at around 220 ms for scrambled stimuli and 245 ms for face stimuli, followed by a decrease to a sustained response with a lower amplitude. The increased amplitude to faces vs. scrambled stimuli was present within the first 100 ms, but was largest around the response peak, although a moderate amplitude enhancement to faces remained throughout the stimulus presentation.
These findings suggest that our gamma TFWOI might not have been optimal to find differences between conditions. Thus, we reran our group-level volumetric analysis using a new gamma TFWOI: 50-90 Hz, 100-450 ms (see [fig_ref] Figure 6: Group-level t-statistical image [/fig_ref]. As before, significant differences were present in right lateral occipital cortex, but this time also extended more anteriorly across the ventral surface of the cortex, extending to the fusiform gyrus (Talaraich coordinates: 35.1, À65.3, À13) around the posterior part of the fusiform face area (FFA). Significant differences were also found in left lateral occipital cortex (Talaraich coordinates: À41.2, À87.3, À7.0). Thus, by optimizing our TFWOI we were able to increase the statistical sensitivity of our analysis and uncover regions of significant difference between conditions that were not present in our initial analysis. . Group-level t-statistical image (thresholded at P < 0.05 FWE) of the difference in response to faces vs. scrambled stimuli in the gamma TFWOI (55-120 Hz, 100-400 ms), overlaid on the template brain. . Group-level t-statistical image (thresholded at P < 0.001 FWE) of the difference in response to faces vs. baseline (À500 to À200 ms) in the gamma TFWOI (55-120 Hz, 100-400 ms), overlaid on the template brain.
To determine whether our use of a different frequency bandwidth to calculate the beamfomer weights affected the virtual sensor analysis, we re-ran the analysis using the new weightsthis time using local maxima from both the right and the left lateral occipito-temporal cortices (one participant was excluded due to an absence of local maxima in the corresponding areas, while four further participants contributed only one virtual sensor: two for the right hemisphere, and two for the left). The results shown in demonstrate that the time-frequency characteristics of the response to both faces and scrambled stimuli (and the difference between them) were not substantially different from those found in the initial virtual sensor analysis (shown in [fig_ref] Figure 4: Group average spectrograms depicting the virtual sensor responses to face [/fig_ref]. The figure also demonstrates that the time-frequency characteristics of the left hemisphere response were highly similar to those found in the right hemisphere, but with a weaker response amplitude in the left hemisphere.
## Evoked responses
To define TWOIs for the analysis of evoked responses, we calculated the omnibus (i.e. averaged across all trials) GFP of the group average sensor-level ERFs. Based on local minima in the omnibus GFP , we defined three TWOIs: 83-116 ms (M100), 116-181 ms (M170) and 181-300 ms (M220).
As with the induced responses, we then performed group-level volumetric analyses of differences in source power between the face and scrambled conditions. Significant differences between conditions were found in the M170 and M220 time windows, but not in the M100 window.
For the M170 analysis (see a single area was found in which the faces produced enhanced source power relative to scrambled stimuli. This was situated in a posterior region of ventro-temporal cortex, close to the region of fusiform gyrus for which we found a gamma amplitude enhancement for faces, but shifted somewhat anteriorally and superiorally so that the most significant voxel occurred in the adjacent white matter (Talairach coordinates: + 37.1, À55.2, À1.0). To explore the time course of this effect we performed a virtual sensor analysis of this response by finding local maxima in the approximate location of this effect in individual-level images of between-condition differences (two participants were excluded from this analysis due to absence of a local maximum in the area of interest). Consistent with our attribution of this effect to . (A) Group mean (AE SE) amplitude spectra of the response from 100 to 400 ms following stimulus onset for both faces and scrambled stimuli. Spectra were smoothed with a uniform kernel of width 9 Hz to reduce spectral noise. (B) Group mean (AE SE) amplitude of the gamma (55-120 Hz) response against time for both faces and scrambled stimuli. In both plots the dark bars above the x-axis depict frequencies/times at which the responses to face and scrambled stimuli were significantly different (Wilcoxon signed-rank test, P < 0.01 FDR). the M170 ERF, the major difference between conditions was a prominent peak around 140 ms to face stimuli, which was almost entirely absent when the stimuli were scrambled [fig_ref] Figure 1: Plot of the number of sensors for which the response to stimuli [/fig_ref].
## A b
For the M220 analysis regions were found in which source power to faces was decreased relative to that produced by scrambled stimuli. The most significant difference was found in a region of the parieto-occipital sulcus (Talairach coordinates: + 3.0, À77.3, + 41.0) extending rightwards into the cuneus, with an additional smaller region occurring on the medial surface of the lingual gyrus. A significant region was also found in the dorsal part of the right cerebellum, but as we consider it unlikely that the cerebellum would show differential activation to the two conditions in our study we consider that this was either a false positive finding or the mislocalization of activity stemming from the nearby ventral surface of the cortex. We again performed a virtual sensor analysis, this time based on the locations of local minima around the region of the parieto-occipital sulcus in individual-level images of differential source amplitude (one participant was excluded due to the absence of a local minimum in this region). [fig_ref] Figure 1: Plot of the number of sensors for which the response to stimuli [/fig_ref] illustrates the presence of a substantial peak around 210 ms (consistent with its identification as part of the M220 component), which was present for scrambled stimuli but largely attenuated for faces.
# Discussion
In this study we have explored the capabilities of MEG in conjunction with the SAM beamformer to localize face-specific cortical activity in frequency, time and space. The purpose of the study was both to demonstrate best practice for the use of the beamforming method to measure face-specific responses and to test the accuracy of these results against face-specific responses measured in other modalities. We will now discuss our results with respect to these aims, as well as the broader theoretical implications of our findings.
## Accuracy of source localization
One fundamental question mark over the use of MEG for localization is that source reconstruction is ill-posed: there is no unique solution to the MEG inverse problem, meaning that MEG source images cannot be guaranteed to accurately reflect the true state of cortical activity. In the case of face processing we have good prior knowledge of the expected sources of activity: a large body of fMRI studies have collectively mapped the cortical preference for faces and implicated two main regions: an area of the lateral occipital cortex, OFA, and an area of the fusiform gyrus, FFA (for review see [bib_ref] The distributed human neural system for face perception, Haxby [/bib_ref]. Moreover, a general tendency for face-specific activity to be right lateralized within these regions has also been described (e.g. [bib_ref] The fusiform face area: a module in human extrastriate cortex specialized for..., Kanwisher [/bib_ref] [bib_ref] TMS evidence for the involvement of the right occipital face area in..., Pitcher [/bib_ref]. Therefore, the fact that in this study we have found gamma-band activity in broadly similar cortical areas and with a general degree of right lateralization at the group level (albeit that with some refinement to our time-window frequency of interest we also found activity in a region probably corresponding to left OFA) provides confidence that the beamformer reconstructed the spatial distribution of the gammaband response with a reasonable degree of accuracy.
One further area found to be involved in face processing in fMRI studies, the superior temporal sulcus (STS), was not found to be active in this study. However, we note that the STS has been proposed to be involved in changeable aspects of faces, such as expression or gaze direction [bib_ref] The distributed human neural system for face perception, Haxby [/bib_ref] , and thus might not have been expected to be strongly active in the current experiment in which static images with mostly neutral expressions were used. Consistent with this, previous MEG beamformer studies have localized gamma-band activity to STS when dynamic face stimuli have been used [bib_ref] MEG demonstrates a supra-additive response to facial and vocal emotion in the..., Hagan [/bib_ref] [bib_ref] Neural responses to rigidly moving faces displaying shifts in social attention investigated..., Lee [/bib_ref].
One factor that we did not account for in our source localization was facial emotion in the stimuli used: the face database we used contained both neutral and happy faces and we did not distinguish between these emotions when selecting stimuli for use in the study. Thus, our data are from the presentation of a mixture of neutral and happy faces. However, we note that a meta-analysis of fMRI studies [bib_ref] Functional atlas of emotional faces processing: a voxel-based meta-analysis of 105 functional..., Fusar-Poli [/bib_ref] did not find substantial differences in localization of activity in the visual cortex to different facial emotions, and we therefore think it unlikely that our source localizations would have substantially differed if exclusively neutral or happy faces were used. We do note though that gamma-band differences in the response to different facial emotions have been reported [bib_ref] Neural dynamics for facial threat processing as revealed by gamma band synchronization..., Luo [/bib_ref] [bib_ref] Visual awareness, emotion, and gamma band synchronization, Luo [/bib_ref] [bib_ref] Early gamma-band activity as a function of threat processing in the extrastriate..., Maratos [/bib_ref] albeit in activity generated in different areas of visual cortex than those reported here (see below: Comparison with previous studies -Gamma). Future work in which the effects of facial emotion are explicitly tested using the methodology outlined in this study will be necessary to test whether such effects were present in our data.
## Gamma-band response
As further evidence of the accuracy of our beamfomer reconstruction the time frequency characteristics of the gamma-band response found in our virtual sensor analyses (namely a broadband response in the gamma frequency range peaking around 240 ms; see Figs 4 and 7) is highly consistent with that found in studies of intracranial electrophysiology in humans [bib_ref] Decoding face information in time, frequency and space from direct intracranial recordings..., Tsuchiya [/bib_ref] [bib_ref] Category-specific visual responses: an intracranial study comparing gamma, beta, alpha, and ERP..., Vidal [/bib_ref] [bib_ref] Antagonistic relationship between gamma power and visual evoked potentials revealed in human..., Privman [/bib_ref] [bib_ref] Exemplar selectivity reflects perceptual similarities in the human fusiform cortex, Davidesco [/bib_ref]. One interesting new finding in this study, however, is that not all of the faceinduced gamma-band response is truly face-specific. Only the response within a narrower frequency range from around 50 to 90 Hz differed between faces and scrambled stimuli, with higher frequency components being common to both. One potential caveat to this finding is that the power in MEG signals is known to scale in inverse proportion to (an exponent of) the frequency of the response. This could lead to data covariance calculations (and hence the beamformer source reconstruction) being dominated by signals at frequencies towards the lower end of the signal bandpass. Thus, our findings of differences only at the lower end of our gamma frequency bandwidth could be a result of this bias. To test this we re-ran our face vs. scrambled stimuli comparison at 100-150 Hz. We did not replicate any significant effects that were found in the 50-90 Hz bandwidth (or even find substantial, but non-significant, differences). Thus, we conclude that differences between stimuli were genuinely restricted to the 50-90 Hz frequency range.
Interestingly, studies in primates of the visual gamma response to grating stimuli have demonstrated a difference between high-frequency/broadband gamma, which appears to be generated by (or at least closely coupled to) local spiking activity, and low-frequency/ narrowband gamma, which reflects a coherent oscillation within local field potentials across extended regions of cortex [bib_ref] Stimulus selectivity and spatial coherence of gamma components of the local field..., Jia [/bib_ref] [bib_ref] Different origins of gamma rhythm and high-gamma activity in macaque visual cortex, Ray [/bib_ref]. Similarly, a recent MEG study has demonstrated the presence of separate high-and low-frequency components of the gamma-band that are differentially modulated by visual attention [bib_ref] Spatial attention increases high-frequency gamma synchronisation in human medial visual cortex, Koelewijn [/bib_ref].
The gamma response shown in appears to demonstrate both these components: in addition to the relatively narrowband response present for faces around 70 Hz, there is a broadband response in frequencies above 60 Hz which is of lower amplitude but common to both stimulus types. This would imply that the differences in response to faces and scrambled stimuli at our virtual sensor sites were not signalled by differences in mean firing rates, but were instead due to faces showing a greater propensity to induce a coherent gamma oscillation in local field potentials. This is consistent with theories that oscillations in the gamma frequency range play a critical role in the perception of coherent structure in an image through 'feature binding' and/or act as a mechanism to synchronize neuronal spiking to facilitate activation of downstream neurons (see [bib_ref] Human gamma-frequency oscillations associated with attention and memory, Jensen [/bib_ref] for more detailed discussion of these ideas). This finding should help to refine existing models of the face-specific gamma-band response [bib_ref] Modeling the electrical field created by mass neural activity, Privman [/bib_ref].
## Beta-and delta/theta-band responses
In this study we have also explored the presence of face-specific responses in the beta frequency range (15-40 Hz) and in a lower frequency range (0-6 Hz), which we have dubbed delta/theta. We did not find any differences between faces and scrambled stimuli at these frequencies. [bib_ref] Category-specific visual responses: an intracranial study comparing gamma, beta, alpha, and ERP..., Vidal [/bib_ref] have demonstrated response selectivity between a group of visual stimulus categories, including faces, in the range 8-24 Hz at intracranial recording sites in bilateral occipito-temporal cortex, and thus we may have expected to find face-specific responses in our beta-band analysis. In contrast, [bib_ref] The many faces of the gamma band response to complex visual stimuli, Lachaux [/bib_ref] failed to find beta-band modulations to upright vs. inverted faces, while [bib_ref] Neural "ignition": enhanced activation linked to perceptual awareness in human ventral stream..., Fisch [/bib_ref] found that perception of backward masked faces was reflected in gamma-but not beta-band amplitude. Thus, the presence of differences in the beta frequency range appears to be sensitive to study design. As we are not aware of any intracranial studies that have used the current methodology (i.e. comparison of faces with matched control stimuli) we cannot independently validate our finding of a lack of effect in the betaband (or in the range 0-6 Hz).
## Evoked responses
In addition to analysis of the induced response, we also analysed the evoked response to faces in the form of three previously described ERFs, the M100, M170 and M220. The M170, and its event-related potential (ERP) counterpart the N170, is a well-known evoked response which is enhanced to faces relative to other stimuli. Source localization methods have implicated all three of the face responsive areas mentioned above -FFA [bib_ref] Face-specific responses from the human inferior occipito-temporal cortex, Sams [/bib_ref] [bib_ref] MEG/EEG sources of the 170-ms response to faces are co-localized in the..., Deffke [/bib_ref] [bib_ref] Population-level inferences for distributed MEG source localization under multiple constraints: application to..., Henson [/bib_ref] , OFA [bib_ref] Cognitive response profile of the human fusiform face area as determined by..., Halgren [/bib_ref] [bib_ref] Event-related brain potential evidence for a response of inferior temporal cortex to..., Schweinberger [/bib_ref] and STS [bib_ref] Source analysis of the N170 to faces and objects, Itier [/bib_ref] as potential sources of the M/N170. Here we localized the M170 to a region of ventral temporal cortex that was overlapping with gammaband activity and which we attributed to FFA. The location of our peak difference for the M170 was within 13 mm of that reported in an intracranial study , suggesting that the beamfomer localization was reasonably accurate (although see below: Comparison with previous studies -M170).
The latency of the M170 peakaround 140 mswas earlier than that reported in other MEG studies of face processing, which report latencies of around 160 ms [bib_ref] Face-specific responses from the human inferior occipito-temporal cortex, Sams [/bib_ref] [bib_ref] Cognitive response profile of the human fusiform face area as determined by..., Halgren [/bib_ref] [bib_ref] The selectivity of the occipitotemporal M170 for faces, Liu [/bib_ref] [bib_ref] MEG/EEG sources of the 170-ms response to faces are co-localized in the..., Deffke [/bib_ref] , and substantially earlier than that found in intracranial recordings, which typically report latencies of around 200 ms [bib_ref] Category-sensitive excitatory and inhibitory processes in human extrastriate cortex, Allison [/bib_ref] [bib_ref] Electrophysiological studies of human face perception. II: response properties of face-specific potentials..., Mccarthy [/bib_ref] [bib_ref] Selective attention modulates face-specific induced gamma oscillations recorded from ventral occipitotemporal cortex, Engell [/bib_ref] ; although see [bib_ref] Spatio temporal dynamics of face recognition, Barbeau [/bib_ref] for latencies more consistent with the EEG/MEG data). However, the morphology of our response, dominated by a single large deflection present only for faces, was consistent with these prior studies. Thus, while the differences in latency between this study and others remain to be explained (and in the case of intracranial recordings may be due to medication effects in those undergoing a craniotomy), we are confident that the beamformer accurately reconstructed the time course of the evoked response.
We did not find any face-specific responses within the M100 time window. This stands in contrast to previous findings of an enhanced M100 response to faces relative to other stimulus categories [bib_ref] Stages of processing in face perception: an MEG study, Liu [/bib_ref]. However, we note that in a study of the ERP counterpartthe P100which also used phase scrambled control stimuli, between-category response differences were found to be eliminated when differences in low-level image properties were taken into account . This again illustrates the importance of using matched control stimuli.
We did find one response that was enhanced to the scrambled stimuli relative to the faces: a parieto-occipital source of the M220 ERF. This finding was surprising: while the M220 is a response to faces that has been described in the MEG literature [bib_ref] Inversion and contrast-reversal effects on face processing assessed by MEG, Itier [/bib_ref] , the response was previously localized to ventral occipital rather than parieto-occipital sites. Moreover, we are not aware of any previous evidence that an evoked response with a preference for scrambled over unscrambled faces exists at this site, or any predictions that it should exist. This makes the finding difficult to interpret and means that future studies will be necessary to clarify this effect. However, we note that our source localization suggests an origin in the dorsal visual stream and therefore we speculate that the effect may be due to differences in the spatial distribution of information between the two stimulus types. In particular, while global image energy was matched across stimulus classes, the local distribution is likely to have differed substantially between stimulus types. Image energy in the face stimuli is concentrated around the face outline and key feature such as the eyes and hairline, and therefore has predictable spatial structure, but is much more dispersed in the scrambled stimuli and lacking in structure. Therefore, spatial processing of the stimuli in the dorsal stream is likely to be substantially different and might be expected to evoke different neural responses.
## Comparison with previous studies -gamma
The use of the SAM beamformer to localize the gamma-band response to static images of faces has also been explored in other studies. [bib_ref] A magnetoencephalographic study of face processing: M170, gamma-band oscillations and source localization, Gao [/bib_ref] were able to localize activity to a similar area of right occipito-temporal cortex found in this study. However, they did not find left OFA activity shown here, but instead found some gamma activity in earlier visual areas, particular when only the inner components of the faces were present. A number of studies have looked at the response to static emotional faces [bib_ref] Neural dynamics for facial threat processing as revealed by gamma band synchronization..., Luo [/bib_ref] [bib_ref] Visual awareness, emotion, and gamma band synchronization, Luo [/bib_ref] [bib_ref] Early gamma-band activity as a function of threat processing in the extrastriate..., Maratos [/bib_ref] and found the strongest activity in relatively early, medial areas of visual cortex.
A comparison of the current study with this prior research points to two important methodological advances used here. First, in previous studies, statistical comparisons have been between activity during face presentation and the baseline period (or in the case of [bib_ref] Early gamma-band activity as a function of threat processing in the extrastriate..., Maratos [/bib_ref] between activity during presentation of different emotional expressions). Such comparisons do not match the low-level image properties of the stimuli and therefore differences found may reflect responses to low-level image features in addition to (or instead of) face-specific responses. This would explain why these previous studies found activity in early visual areas that are known to be sensitive to low-level image features but are generally not thought to be specific to face processing. Indeed, when we performed a comparison of the response to face trials vs. baseline using our current data, we found gamma activity widespread throughout occipital cortex with the most significant differences occurring around striate cortex and neighbouring areas . Thus, it is highly recommended that in future MEG studies researchers should aim to contrast face stimuli with control images that are as closely matched for low-level stimulus properties as possible.
Secondly, it can be seen that some of these previous studies have tended to use frequency bandwidths that, while consistent with previous studies of the visual gamma response using grating stimuli [bib_ref] Visual gamma oscillations and evoked responses: variability, repeatability and structural MRI correlates, Muthukumaraswamy [/bib_ref] , were not, on the basis of the current study, optimal for finding that part of the gamma response sensitive to the presence of faces. In particular, the 30-50 Hz bandwidth employed by [bib_ref] Neural dynamics for facial threat processing as revealed by gamma band synchronization..., Luo [/bib_ref] [bib_ref] Visual awareness, emotion, and gamma band synchronization, Luo [/bib_ref] would have filtered out much of the face-specific response found here.
Likewise, a comparison between Figs 2 and 6 demonstrates that while it is possible to measure face-specific effects using relatively broadband filtering encompassing the whole gamma frequency range, greater statistical sensitivity can be achieved by restricting analysis to that part of the gamma band for which the response is truly face-specific (approximately 50-90 Hz based on the evidence found here).
## Comparison with previous studies -m170
Unlike the gamma response, the M170 ERF seems to be highly specific to faces and is largely attenuated for non-face stimuli. Thus, it seems likely that the M170 can be well localized without the use of a matched control stimulus. However, we note that the beamfomer method used here is incompatible with the use of permutation testing of a single condition: a single state source analysis can only produce positive values of source power, while the one-sample permutation test requires participant-level measures to be able to take both signs (due to the use of sign-flipping to compute the null distribution of the test statistic). For this reason, it is necessary to contrast the response to faces with a second condition, and a control stimulus matched with faces for low-level image properties may be beneficial in this instance, as ERFs preceding the M170 should theoretically be well matched across conditions .
We were able to find three previous studies of the M170 which used beamforming and in which the coordinates of the M170 source were reported [bib_ref] Inversion and contrast-reversal effects on face processing assessed by MEG, Itier [/bib_ref] [bib_ref] Attention inhibition of early cortical activation to fearful faces, Bayle [/bib_ref] [bib_ref] A magnetoencephalographic study of face processing: M170, gamma-band oscillations and source localization, Gao [/bib_ref]. Along both the x-(posterior-anterior) and the y-(left-right) axes our M170 source was within 4 mm of the group mean coordinates from these previous studies. However, along the z-axis our source location was approximately 16 mm superior to the group mean. Likewise, our source was 13 mm superior to the centroid of the face-specific N200 (the intracranial analogue of the M170) found by [bib_ref] Electrophysiological studies of human face perception. I: potentials generated in occipitotemporal cortex..., Allison [/bib_ref] , despite our source location matching the centroid to within 1 mm along the other two axes. This suggests that there may have been a superior bias in our source localization.
This bias could have been due to the beamformer incorrectly combining two correlated sources, one in FFA and another in a superior part of temporal cortex (such as STS perhaps), to a single source in an intermediate location. However, it is not clear why such a bias would not also have been present in previous beamforming studies.
Alternatively, it is possible that differences between our source localization and those found in previous studies represent stochastic variation around the true source location. Consistent with this, when averaged across all four studies (the current study and the three prior studies) the mean source location of the M170 given by the beamformer was within 3.5 mm of the centroid found by [bib_ref] Electrophysiological studies of human face perception. I: potentials generated in occipitotemporal cortex..., Allison [/bib_ref] whereas source locations from individual studies were all > 11 mm from the centroidsuggesting convergence across experiments towards the intracranial findings.
# Conclusions
We have demonstrated the validity of MEG, combined with the SAM beamfomer, to measure and localize both induced and evoked responses to faces. Our results, particularly with respect to the gamma-band response and M170, are strongly consistent with data from fMRI on the location of cortical areas involved in face processing and data from intracranial recordings in humans on the temporal and time-frequency aspects of this response. We have also elucidated some important methodological points with respect to beamformer imaging of faces: namely that contrasts with appropriate control stimuli and careful selection of frequency bands of interest are necessary to maximize statistical power of the beamfomer. Incidental to this, we have produced the novel finding that face-specific responses are specific to a sub-band (50-90 Hz) of the broader gamma-band response.
Previously, combined data on the spatial, temporal and frequency characteristics of the face-specific response have come from intracranial EEG studies. While we do not dispute that such studies must remain the 'gold standard' in this field (not least because they provide necessary 'ground truth' data with which to validate MEG source reconstruction) we note that, as an experimental technique, MEG is much more suited to high-throughput experimental testing and/or testing of selectively defined populations (for instance in those with conditions such as prosopagnosia or autism). Furthermore, modern MEG systems have the advantage of whole-head coverage, meaning that they can be used to study extended brain networks in a manner that generally cannot be achieved using intracranial methods. For this reason, we believe that the potential for MEG to measure cortical responses to faces, as demonstrated in this and previous studies, offers exciting new avenues for research into the cortical basis of face perception.
[fig] Figure 1: Plot of the number of sensors for which the response to stimuli (averaged across all trials of both conditions) significantly differed from zero for each time 9 frequency point. The colour scale depicts number of significant sensors. middle occipital gyrus (Talaraich coordinates: + 43.2, À71.3, + 1.0), close to the proposed location of the occipital face area (OFA; see table 1 of Pitcher et al., 2011, for a summary of Talairach coordinates of right OFA from 14 fMRI studiesthe coordinates of OFA reported here are within 12 mm of the mean location across these studies). [/fig]
[fig] Figure 4: Group average spectrograms depicting the virtual sensor responses to face (left panel) and scrambled stimuli (middle panel), and the difference between the two responses (right panel). The colour scale depicts amplitude in percentage relative to baseline (À1 to 0 s). [/fig]
[fig] Figure 6: Group-level t-statistical image (thresholded at P < 0.05 FWE) of the difference in response to faces vs. scrambled stimuli in the refined gamma TFWOI (50-90 Hz, 100-450 ms), overlaid on the template brain.A BFig. 7.(A) Group average spectrograms depicting the right hemisphere virtual sensor responses to face (left panel) and scrambled stimuli (middle panel), and the difference between the two responses (right panel), using weights from the 50-90 Hz gamma analysis. (B) As for A but for the left hemisphere virtual sensors. [/fig]
[fig] Figure 8, Figure 9: Global field power of the grand average ERF across all trials, and for face and scrambled trials separately. (A) Group-level t-statistical image (thresholded at P < 0.05 FWE) of the difference in response to faces vs. scrambled stimuli in the M170 analysis, overlaid on the template brain. (B) As for A but for the M220 analysis. [/fig]
|
Transient Eye Shortening During Reading Text With Inverted Contrast: Effects of Refractive Error and Letter Size
Citation: Swiatczak B, Schaeffel F. Transient eye shortening during reading text with inverted contrast: Effects of refractive error and letter size. Transl Vis Sci Technol. 2022;11(4):17, https://doi.org/10.1167/tvst.11.4.17Purpose: Myopes have a reduced ability to elicit transient axial eye shortening after imposed positive defocus, which may be due to changes in the biochemical signaling cascade controlling choroidal thickness. We have investigated whether reading with inverted text contrast can still elicit transient axial eye shortening in myopes, like it has been shown in emmetropes.Methods:Changes in axial length before and after reading were measured with the Lenstar LS-900. Text with inverted contrast was read from a large screen at 2 m distance (angular subtense 35.9°, screen luminance matched in all conditions to 86 ± 7 cd/m 2 ) for 30 minutes. Moreover, we investigated the effects of letter sizes. Two text sizes were tested: "small" text (letter height 13.75 arcmin) and "large" text (letter height 34.39 arcmin).Results:Reading text with inverted contrast induced eye shortening (-10.2 ± 9.5 μm) in myopic eyes (n = 11; refraction -3.5 ± 1.9 diopters [D]), showing that an inhibitory signal was still generated by the retina as in emmetropes. In 15 subjects (refraction +1.7 to -4.2 D) we found that small text does not elicit significant differences in axial length (P = 0.09). However, with large text, changes in axial length were clearly different for the both contrast polarities (standard contrast, +1.7 ± 9.0 μm; inverted contrast, -9.7 ± 8.9 μm; P = 0.0017).Conclusions:Although positive defocus may not be an effective intervention to inhibit further eye growth in myopes, other visual stimuli can still trigger choroidal thickening and possibly generate signals to decrease myopia progression.Translational Relevance: Our results have shown the optimized text features, which may have a positive impact on myopia control.
# Introduction
There is extensive experimental evidence that the feedback mechanism of emmetropization (the developmental matching of focal length and eye length) uses retinal image defocus as an error signal when optimizing refractive state (review 1 ). However, central questions remain unresolved: if there is a closed loop feedback mechanism in operation, why does it not block the development of myopia? And why do children become myopic even though there is no consistent defocus on their retina? It is clear also that other factors, in addition to retinal image defocus, feed into the emmetropization mechanism.
Examples include limited exposure to outdoor lighting as well as extensive near work indoors; both are well-known risk factors of myopia development (recent review 2 ). However, it remains unexplained why the retina stops triggering emmetropization under these conditions. There are two main options:
(1) the visual information necessary for the retina to emmetropize the eye is altered or insufficient, or (2) the myopic retina undergoes functional changes early in myopia development which compromises correct retinal growth control, despite appropriate visual input.
(1) In favor of the first assumption is that emmetropization is affected by restrictions of the spectral composition of light (i.e., tree shrews 3 ; rhesus monkeys 4 ; chickens [bib_ref] Effect of duration, and temporal modulation, of monochromatic light on emmetropization in..., Lin [/bib_ref] , to lower average illuminances (i.e., human [bib_ref] Time spent outdoors in childhood is associated with reduced risk of myopia..., Lingham [/bib_ref] ; rhesus monkeys [bib_ref] Effects of low intensity ambient lighting on refractive development in infant rhesus..., She [/bib_ref] [bib_ref] The development of and recovery from form-deprivation myopia in infant rhesus monkeys..., She [/bib_ref] ; mice [bib_ref] Ambient light regulates retinal dopamine signaling and myopia susceptibility, Landis [/bib_ref]. It was also proposed that spatial features of the visual environment, like the "greenness of the residential area" 10 or changes in the spatial frequency spectra can affect emmetropization. [bib_ref] The spatial frequency content of urban and indoor environments as a potential..., Flitcroft [/bib_ref] It was also proposed that printed text represents an imbalance in ON/OFF pathway activation which was found to change choroidal thickness and may therefore affect emmetropization. [bib_ref] Reading and myopia: contrast polarity matters, Aleman [/bib_ref] In favor of the second assumption is the observation that none of the known interventions to control myopia could fully stop its progression and that imposed positive defocus no longer elicits axial eye shortening in myopic subjects. [bib_ref] Emmetropic, but not myopic human eyes distinguish positive defocus from calculated blur, Swiatczak [/bib_ref] There may be functional changes in the myopic retina that cannot be recovered by optimizing visual input or by pharmacological means. It is clear, however, that optically corrected myopes have otherwise normal visual function, at least when their myopia is moderate. [bib_ref] Effect of age and refractive error on quick contrast sensitivity function in..., Li [/bib_ref] The nature and mechanism of the presumed changes in the retina remain therefore obscure.
A major progress in myopia research was the finding that retinal responses to different visual stimuli can be studied in short-term experiments by measuring miniature changes in choroidal thickness (by optical coherence tomography) or axial length (by low coherence interferometry). [bib_ref] Human optical axial length and defocus, Read [/bib_ref] [bib_ref] The time course of the onset and recovery of axial length changes..., Delshad [/bib_ref] [bib_ref] Effect of optical defocus on choroidal thickness in healthy adults with presbyopia, Chiang [/bib_ref] [bib_ref] Effect of retinal image defocus on the thickness of the human choroid, Chiang [/bib_ref] [bib_ref] Astigmatic defocus leads to short-term changes in human choroidal thickness, Hoseini-Yazdi [/bib_ref] A basic assumption, made by several authors, is that these short-term changes have predictive power for future myopia development. [bib_ref] Human optical axial length and defocus, Read [/bib_ref] [bib_ref] Effect of retinal image defocus on the thickness of the human choroid, Chiang [/bib_ref] [bib_ref] Hyperopic defocus and diurnal changes in human choroid and axial length, Chakraborty [/bib_ref] It remains unknown how long the effects of visual stimulation persist and whether they remain stable over time. The hypothesis can probably be best supported by long-term studies in children, [bib_ref] Longitudinal changes in choroidal thickness and eye growth in childhood, Read [/bib_ref] [bib_ref] Longitudinal changes in choroidal and retinal thicknesses in children with myopic shift, Jin [/bib_ref] although it may be difficult to finally determine what is cause or consequence. [bib_ref] Emmetropic, but not myopic human eyes distinguish positive defocus from calculated blur, Swiatczak [/bib_ref] [bib_ref] Choroidal thickness and ametropia in children: a longitudinal study, Fontaine [/bib_ref] [bib_ref] Optical defocus rapidly changes choroidal thickness in schoolchildren, Wang [/bib_ref] [bib_ref] Effects of single and repeated intravitreal applications of atropine on choroidal thickness..., Mathis [/bib_ref] To learn more about what might have changed in the myopic retina, we asked young adult subjects to read text with standard contrast (dark text on bright background) or with inverted contrast (bright text on dark background) with matched luminance. The procedure was previously shown to trigger bidirectional changes in choroidal thickness, 12 which show up as small changes in measured axial length. Two questions were studied: [bib_ref] Homeostasis of eye growth and the question of myopia, Wallman [/bib_ref] Knowing that the choroid in our myopic cohort was not responsive to imposed positive defocus in our previous study, [bib_ref] Emmetropic, but not myopic human eyes distinguish positive defocus from calculated blur, Swiatczak [/bib_ref] can reading with inverted text contrast elicit axial eye shortening in myopic subjects?Are the effects dependent on letter size?
# Methods
## Subjects
In total, 21 young subjects participated in the experiment. The first part of the study (question 1) included 11 myopic subjects (2 males; aged 24.4 ± 2.6 years; average refraction, -3.4 ± 1.4 diopters [D]; range, -1.25 to -6.00 D). The second part (questions 2) included 15 subjects (2 males; aged 25.7 ± 3.1 years): 8 emmetropic, not needing any distance correction (average refraction, 0.0 ± 0.3 D; range, +0.25 to -0.50), 5 myopic (average refraction, -3.4 ± 0.8 D; range, -2.50 to -4.25 D), and 2 hyperopic (average refraction, +1.4 ± 0.5 D; range, +1.0 to +1.7 D). Myopic and hyperopic subjects wore their habitual corrections during the experiments. One emmetropic female subject could not take a part in the second appointment (reading large text) owing to restrictions related to the coronavirus disease-2019 pandemic. All subjects were in good general health and had no history of previous ocular pathologies, other than moderate refractive errors. They had normal far and near visual acuity and no accommodative anomalies. None of the subjects had astigmatism or anisometropia of more than 1 D, amblyopia, or squint. Subjects signed a consent form before participating in the experiments. The study was conducted in accordance with the tenets of the Declaration of Helsinki and approved by the Swiss Research Ethics Committee (EKNZ, reference 2020-01576).
## Experimental protocol question 1: can reading with inverted text contrast still elicit axial eye shortening in myopic subjects?
We have previously found that myopes, different from emmetropes, develop longer eyes when they watch a movie with imposed positive defocus. [bib_ref] Emmetropic, but not myopic human eyes distinguish positive defocus from calculated blur, Swiatczak [/bib_ref] Eleven myopic subjects were asked to watch a movie binocularly with +2.5 D of optical defocus, as previously described, [bib_ref] Emmetropic, but not myopic human eyes distinguish positive defocus from calculated blur, Swiatczak [/bib_ref] and subsequently to read "large" text on the screen with inverted contrast (bright letters on the darker background) wearing their habitual corrections (Calibri Body font, sentence case, 34.39 arcmin). Experimental stimuli were presented on a large TV screen at 2 m distance (LG OLED65C9, 65 inch, 4K resolution 3840 × 2160 pixels, subtending 35.9°of the visual field), with an average screen luminance of 86 ± 7 cd/m 2 . The distance was selected to stimulate a large retinal area in the visual field (35.9°). The visual angle was calculated as an arctangent of ratio of letter height to screen distance in millimeters. The room was illuminated only by the screen with no other . Schematic illustration of the study protocol for two parts of the experiment. Question 1 included one visit, where myopic subjects read inverted contrast "large" text and watched a movie with imposed positive defocus. Question 2 included two visits for separate testing the effects of "small" and "large" text, presented at standard and inverted contrast polarity. Before each task, a washout period of 10 minutes was imposed to eliminate possible influences of previous activities. Before and after each viewing task, axial length was measured. The order of the visits and viewing tasks was randomized. The spectacle symbol denotes conditions where positive defocus was imposed.
light sources in the room. Before each task, subjects underwent 10 minutes of a washout period where they had to look at an empty gray screen before the reading task, with the screen luminance matching the average luminance of text displays, or they had to watch a movie in focus before imposing optical defocus. Axial length was measured with the Lenstar LS-900 before and after 30 minutes of movie watching or text reading. The sequence of movie watching, and text reading was randomized and is illustrated in . After the reading period, subjects were asked about a content of the text that they had read to confirm that attention was paid to the screen content and visual stimulation had lasted for the entire period.
## Question 2: are the effects of text contrast polarity on axial length dependent on text size?
A second experiment was done to find out whether letter sizes were important to elicit the bidirectional changes in axial length during reading of text with different contrast polarities. Moreover, it was of interest whether changes in myopes were different from those in emmetropes. In all cases, reading with one contrast polarity served as a control for reading with the other contrast polarity, separated by washout periods of looking at a gray screen with matched luminance.
The effects of reading small or large text were studied on 2 separate days. Using the same screen and setup as in the experiments above, 15 subjects were asked to read text binocularly for 30 minutes (Calibri Body font, sentence case) with standard contrast (dark letters on the brighter background) or with inverted contrast (bright letters on the darker background). As in the experiments above, to determine the baseline axial length, subjects were asked to look at an empty gray screen with matched luminance for 10 minutes before each reading task. Presentation sequences of different text sizes and contrast polarities were randomized . Two text sizes were presented (1) small text with a capital letter height of 8 mm on the screen, subtending 0.23°(13.75 arcmin) in the visual field and (2) large text with capital letter height of 20 mm high, subtending 0.57°(34.39 arcmin) [fig_ref] Figure 2: Experimental set-up to study the effect of reading of text with different... [/fig_ref]. [bib_ref] The problem of size and distance, Joynson [/bib_ref] Screen luminance was carefully matched for the four types of texts (86 ± 7 cd/m 2 ) by digitally adjusting the pixel brightness of the text displays (140.5 ± 4.8 px) and by matching the number of dark and bright pixels for the two text sizes with the same type of contrast (standard contrast, ratio dark to bright pixels 1:5-1:6; inverted contrast, ratio dark to bright pixels 5:1-6:1). The relative strength of the ON or OFF stimulation for the different text types was analyzed with the software that was published by Aleman et al. [bib_ref] Reading and myopia: contrast polarity matters, Aleman [/bib_ref] and confirmed that both small and large standard contrast texts overstimulated OFF-center receptive fields, whereas both small and large inverted contrast texts overstimulated ON-center receptive fields. 12
## Measurements of axial length and refractive errors
Ocular biometry was measured with a commercial low coherence interferometer, Lenstar (Lenstar LS 900 with autopositioning system; Haag-Streit, Koeniz, Switzerland). [bib_ref] Eye elongation during accommodation in humans: differences between emmetropes and myopes, Drexler [/bib_ref] Axial length measurements of the right eyes were taken from each subject before and immediately (<1 minute) after each experimental session. The Lenstar LS 900 provides only 2 digits behind the decimal and is measuring in millimeters. This means that the effect sizes are outside the scale and can be resolved only by repeating measurements and taking averages. Averages of five repeated measurements were accepted for further analysis when the standard deviation was less than 10 μm. In our study, average standard deviation was 7 μm. All experiments were performed in the morning between 9:00 AM and 12:00 PM, at the same time for each individual subject to decrease the risk of effects of diurnal factors.
Refractive errors (denoted as spherical equivalents) were measured without cycloplegia using a commercially available infrared photorefractor (plusoptiX A12R binocular autorefractor, PlusOptix, Nuremberg, Germany) before the experiment. 28
# Data analysis
Data analysis was performed by using a freely available software for statistical computing, "R" (version R 4.0.1; R Core Team, R Foundation for Statistical Computing, Vienna, Austria). Pairwise comparisons were calculated to evaluate effect of reading standard and inverted contrast text on change in axial length. The relationship between axial length changes after reading texts with different contrast polarity and refractive errors was estimated by Pearson's correlation coefficient. Axial length data from individual subjects are presented as the averages of five repeated measurements ± their standard deviation. Averaged data from all subjects are presented with standard errors. P values of less than 0.05 were considered statistically significant.
# Results
## Question 1: can reading with inverted text contrast still elicit axial eye shortening in myopic subjects?
To find out whether the retina in myopic subjects was generally unable to trigger eye shortening, or whether this deficiency was limited to imposed positive defocus, myopic subjects were recruited to either watch movies with imposed positive defocus or read large text with inverted contrast. Interestingly, the eyes of the myopic subjects responded differently under both conditions. They became longer with the positive lenses, but shorter with text with inverted contrast [fig_ref] Figure 3: Myopic subjects responded to imposed positive defocus with elongation of the eye... [/fig_ref]. Ten of 11 myopic subjects responded with eye shortening to text reading with inverted contrast, whereas positive defocus had opposite or no effect on axial lengths in those subjects. Averaged over all subjects, there was a highly significant difference in axial length change after reading text with inverted contrast or watching movies with positive defocus (-10.2 ± 9.5 μm vs. 8.9 ± 9.6 μm, respectively; P < 0.001) [fig_ref] Figure 3: Myopic subjects responded to imposed positive defocus with elongation of the eye... [/fig_ref]. The result suggests that functional differences in the retina of emmetropes and myopes were limited to the responses to positive defocus.
## Question 2: are the effects of text contrast polarity on axial length dependent on text size?
Although some subjects also showed a transient increase in axial length when reading the small text with standard contrast (6 of 15), and a decrease in axial length when reading the small text with inverted contrast (8 of 15), the average effect over 15 subjects showed similar trend as reported by Aleman et al., [bib_ref] Reading and myopia: contrast polarity matters, Aleman [/bib_ref] but did not reach significance (+2.9 ± 7.9 μm vs. -5.0 ± 12.3 μm, respectively; P = 0.09). However, large text with inverted contrast generated consistent effects across the subjects and causes shorter axial lengths (12 of 14 subjects). Only six subjects developed longer eyes when they read large text with standard contrast. In average, the differences in axial lengths were significant between the two contrast polarities (standard contrast: +1.7 ± 9.0 μm vs. inverted contrast: -9.7 ± 8.9 μm; n = 14; P = 0.0017), in line with previous findings by Aleman et al. [bib_ref] Reading and myopia: contrast polarity matters, Aleman [/bib_ref] Analyz-ing significances in individual subjects, 12 of the 14 showed decreased axial lengths after reading inverted contrast large text. Moreover, in 8 of the 14 subjects, the difference in axial length change between both text contrast polarities was found to be significant. Subjects who did respond to the contrast polarity of the text did not differ in age, refractive error, or gender from nonresponding subjects. It remains unexplained why some subjects responded more to the small text and others to the large text. [fig_ref] Figure 5: Effects of text contrast polarity were not dependent on refractive errors [/fig_ref] shows that the effects of reading with inverted text contrast on axial length did not depend on refractive error. All subjects responded similarly, no matter whether they were hyperopic, emmetropic or myopic. [fig_ref] Figure 5: Effects of text contrast polarity were not dependent on refractive errors [/fig_ref] also shows that eye shortening could only be elicited by large text with inverted contrast [fig_ref] Figure 5: Effects of text contrast polarity were not dependent on refractive errors [/fig_ref].
# Discussion
The most striking finding of this study was that the myopic retina responded like the emmetropic retina when text was read with inverted contrast, triggering the development of shorter eyes. In contrast, when positive lenses were applied, the myopic retina triggered axial eye elongation and the emmetropic retina induced axial eye shortening, which confirmed results from the previous study. [bib_ref] Emmetropic, but not myopic human eyes distinguish positive defocus from calculated blur, Swiatczak [/bib_ref] There seems to be a difference in retinal responses in myopes when positive defocus is imposed. It is, however, difficult to determine what it might be, without further experiments. The problem is that the retinal tricks to detect the sign of defocus have still to be uncovered. At present we only know that a similar trick may also be used to guide accommodation. Del Aguila-Carrasco et al. [bib_ref] Accommodation responds to optical vergence and not defocus blur alone, Aguila-Carrasco [/bib_ref] have carefully studied visual cues controlling accommodation and concluded that "accommodation responds to optical vergence and not defocus blur alone."
## Speculations about the nature of "the" inhibitory signal for axial eye growth
A striking observation in the chicken model was that the retina can distinguish positive from negative defocus, even if the retinal images are continuously and heavily out of focus, containing no high spatial frequencies. Two laboratories placed chickens individ-ually in the center of a large drum, covered inside with pictures, so that only one viewing distance was possible for the chickens. [bib_ref] Homeostasis of eye growth and the question of myopia, Wallman [/bib_ref] [bib_ref] Long-term changes in retinal contrast sensitivity in chicks from frosted occluders and..., Diether [/bib_ref] Lenses were placed in front of the eyes so that similar amounts of defocus were presented to the retina, although with a different sign. In the experiment by Diether and Schaeffel, [bib_ref] Long-term changes in retinal contrast sensitivity in chicks from frosted occluders and..., Diether [/bib_ref] chickens were cyclopleged in addition so that they could not refocus their retinal images. Even under these conditions, eyes started to compensate appropriately for the imposed positive and negative defocus by changing their axial lengths in different directions. Apparently, growth-inhibiting signals were generated, even in severely low-pass filtered retinal images that would normally induce deprivation myopia, except for the case when low-pass filtering resulted from positive defocus. Similarly, the human retina can generate a growth-inhibiting signal with continuously low-pass filtered retinal images, combined with a positive defocus. [bib_ref] Emmetropic, but not myopic human eyes distinguish positive defocus from calculated blur, Swiatczak [/bib_ref] Emmetropic subjects developed shorter eyes when they watched movies with continuous positive defocus but myopes did not. We speculate that lack of the (yet undefined) inhibitory signal in the myopic retina is responsible for myopia development in general and not the low-pass filtering itself. Interestingly, hyperopia induced by positive lenses in animal models is based on a different biochemical signaling cascade, because it cannot be blocked by atropine ( 31 and our own unpublished observations), has different time kinetics, [bib_ref] Temporal properties of compensation for positive and negative spectacle lenses in chicks, Zhu [/bib_ref] and activates a different set of genes in marmosets. [bib_ref] Gene expression in response to optical defocus of opposite signs reveals bidirectional..., Tkatchenko [/bib_ref] Furthermore, the genetic networks responsible for a visual acuity do not match the ones controlling emmetropization in mice, [bib_ref] Genetic network regulating visual acuity makes limited contribution to visually guided eye..., Tkatchenko [/bib_ref] in line with our conclusion that high spatial frequencies and high visual acuity are not required for emmetropization. To interfere with myopia development, we believe that this inhibitory signal requires more studies since it may be the key element responsible for myopia development. This cannot be done by studying the signaling cascade for deprivation myopia or myopia-induced by negative lenses.
## Effects of letter size
We found that contrast polarity of small text (letter size 0.23°of visual angle -comparable with the text on a smartphone) had, averaged over 15 subjects, less effect on axial length than contrast polarity of large text (letter size 0.57°-comparable with the text in a textbook). Interestingly, the large text was in the range of best readability as described by [bib_ref] Does print size matter for reading? A review of findings from vision..., Legge [/bib_ref] and the small text was in a range where reading speed starts to decline. That large text is more effective is in line with the view that not foveal, but parafoveal, receptive ON /OFF fields are involved. It has recently been shown by Panorgias et al. [bib_ref] Retinal responses to simulated optical blur using a novel dead leaves ERG..., Panorgias [/bib_ref] that the retinal area between 6°and 12°eccentricity were most responsive to blur in electroretinogram recordings with the phase-reversing dead leave paradigm. Assuming that line thickness is one-fifth of the letter height (as in a Snellen chart), receptive field sizes of the assumed underlying ON and OFF cells should be in the range of 0.1°to 0.2°, resulting in spatial frequencies around 5 cyc/deg. Also, Charman [bib_ref] Near vision, lags of accommodation and myopia, Charman [/bib_ref] concluded that text sizes in regular textbooks require a resolution of 5 cyc/deg to be read. Furthermore, it has been shown by Sanz Diez et al. [bib_ref] Effect of spatial filtering on accommodation, Sanz Diez [/bib_ref] that accommodation is driven by information in the low spatial frequency range approximately 5 cyc/deg. Swiatczak and Schaeffel 13 had found that emmetropization uses visual cues in the low spatial frequency range between 1 and 10 cyc/deg. The question arises as to why emmetropization should use low spatial frequencies for a highly precise task like controlling the growth of the eye. A possible reason is that the depth of focus decreases with increasing spatial frequencies. If only high spatial frequencies were to be analyzed, contrast at these spatial frequencies would decrease sharply already with small amounts of defocus and the dioptric range of possible regulation would become narrow. In contrast, with low spatial frequencies, contrast declines continuously over a wide range of defocus, providing a continuous signal for emmetropization. [bib_ref] Accommodation, refractive error and eye growth in chickens, Schaeffel [/bib_ref] Perhaps for this reason also accommodation is controlled by low spatial frequency information at less than 10 cyc/deg.
Evidence for axial eye shortening in response to a reading text with inverted contrast polarity was previously found by Aleman et al., [bib_ref] Reading and myopia: contrast polarity matters, Aleman [/bib_ref] although, in their study choroidal thickness was measured rather than axial length. However, because thicker choroids cause a shorter axial length, both measures should at least partially reflect the same intraocular changes. A difference to the work by Aleman et al. [bib_ref] Reading and myopia: contrast polarity matters, Aleman [/bib_ref] was that bidirectional changes in choroidal thickness were found after 60 minutes of reading, whereas in the current study only contrast-inverted text elicited significant changes in axial length after 30 minutes. Possible explanations include (1) choroidal thinning was too small to be reliably detected by low coherence interferometry, (2) reading tasks lasted too short time or, that, (3) in the current study, screen luminance was higher (82 cd/m 2 ) than in the previous experiments (35 cd/m 2 ). Different luminance may partially explain the different outcomes, although it was shown that looking at an empty screen at three different screen luminance [bib_ref] Does print size matter for reading? A review of findings from vision..., Legge [/bib_ref] , and 62 cd/m 2 ) did not influence on axial length change. [bib_ref] Reading and myopia: contrast polarity matters, Aleman [/bib_ref] In line with the previous study is that "choroidal thickening with ON stimulation was not correlated with refractive errors," we also found that the effects of reading text with inverted contrast were not dependent of refractive error. Recently, Hoseini-Yazdi et al. found that reading text with standard contrast, stimulating predominantly the OFF pathways, caused choroidal thinning that was additive to the choroidal thinning induced by accommodation (Hoseini-Yazdi H RS, et al. ARVO Abstract. 2021;2021). [bib_ref] Retinal OFF pathway activation leads to greater accommodation-induced choroidal thinning, Hoseini-Yazdi [/bib_ref] Again, there was no correlation noted with refractive error, in line with our findings and findings by Aleman et al. [bib_ref] Reading and myopia: contrast polarity matters, Aleman [/bib_ref] Also, recently, Hogue and Taylor found that the individual sensitivity to ON or OFF stimuli is variable and, interestingly, it was related to axial length as well ;2021:ARVO Abstract). [bib_ref] Axial length is associated with individual differences in ON-and OFF-pattern detection, Hogue [/bib_ref] As in other studies, it should be kept in mind that effect sizes in our study of around 10 μm are not clinically relevant for vision (note that this corresponds to changes in refractive state of approximately 0.027 D [bib_ref] A 3-year randomized clinical trial of MiSight lenses for myopia control, Chamberlain [/bib_ref] [bib_ref] Comparison of ocular component growth curves among refractive error groups in children, Jones [/bib_ref] , but that the study was designed to learn about the output of the retina, taking choroidal thickness changes as a measure.
These results suggest that the functional changes in the myopic retina are limited to the detection of positive defocus, whereas other visual stimuli can still shorten axial length and probably inhibit myopia development. Extrapolating our results to myopia correction paradigms, it becomes clearer why undercorrection may have inconsistent effects on myopia progression. However, other visual stimuli may still be effective to inhibit myopia.
[fig] Figure 2: Experimental set-up to study the effect of reading of text with different contrast polarities on axial length. The large 65" screen was placed at 2 m distance so that its angular subtense was 35.9°in the visual field. [/fig]
[fig] Figure 3: Myopic subjects responded to imposed positive defocus with elongation of the eye or displayed no changes (pink bars). However, their eyes became still shorter when they read text with inverted text contrast (dark gray bars), like in emmetropes. Left: comparisons between reading with inverted contrast and wearing positive lenses in individual myopic subjects. Right: average data, showing significantly opposite responses to both stimuli, inverted text contrast and positive defocus (p < 0.001). [/fig]
[fig] Figure 4: (A) Transient axial length changes in 15 subjects reading "small" text for 30 minutes. Light bars denote standard text contrast, and dark bars are inverted text contrast. Although some subjects displayed bidirectional changes in axial lengths depending on contrast polarities (2, 13, and 15), the average effect was not significant (see plot of the right). (B) With "large" text, most subjects developed shorter eyes (no data for subject 10), resulting in significant differences in axial length changes (plot on the right). Refractions refer to the spherical equivalents. * P < 0.05, ** P < 0.01. D, diopters. [/fig]
[fig] Figure 5: Effects of text contrast polarity were not dependent on refractive errors. Only inverted text contrast (C and D) caused, on average, a decrease in axial length, which was significant only for the large text fonts ("large, inverted text", D). There were no correlations with refractive state. [/fig]
|
A medical peer-delivered intervention comprising brief motivational interviewing via instant-messaging interaction to reduce drug misuse among youth in Hong Kong: A protocol for a randomised controlled trial
Aims: Youth are frequently exposed to drugs, and most youth who misuse drugs are reluctant to seek help from services due to the worry of others being judgmental, lacking expertise, exposing their personal information, or informing their parents. Considering these concerns, we propose to evaluate the effectiveness of a medical peer-delivered intervention comprising brief motivational interviewing via instant-messaging interaction in reducing drug misuse among youth in Hong Kong.Methods:A two-group single-blind, randomised controlled trial will be conducted. Multiple approaches, including online and face-to-face methods, will be used to recruit the participants. The participants, aged 25 years or younger and reporting any drugs that they have taken within the past 30 days, will be recruited and randomised to receive either brief motivational interviewing via interactive instant-messaging (the intervention) or general health textmessages (comparator). The primary outcome will be the change in the participants' reductions in self-reported drug consumption at 12 months compared to that at baseline. The secondary outcomes will be the changes in the drug-abusing participants' reductions in self-reported drug consumption at 6 months, the changes in the drugquitting participants' 6-and 12-month contemplation stages and relapse risk compared to that at baseline, 30 days' self-reported drug abstinence at 6 and 12 months, and the treatment needs and motivation at 6 and 12 months compared to that at baseline. The effectiveness of the proposed intervention will be examined with adjusted regression models, with adjustment for baseline characteristics and the use of an intention-to-treat approach.Discussion: This proposed study will be the first randomised controlled trial to assess the effectiveness of a medical peer-delivered interactive intervention to reduce drug misuse among youth in Hong Kong. The proposed intervention has the potential to increase the help-seeking behaviour and intention to quit among youth who misuse drugs.© The Author(s) 2021. This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article' s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article'
# Introduction
Drug misuse is a worldwide problem, especially among youth. In Hong Kong, there is easy access to a wide range of illicit drugs from multiple sources, which has led to an increase in drug misuse among youth. According to the Central Registry of Drug Abuse Report in Hong Kong, 87.3 % of people with a drug use disorder reported that they first used drugs at 25 years of age or younger. Statistics from the Central Registry of Drug Abuse show that the number of students who claimed to have used drugs increased by 23 % between 2014/15 and 2017/18 and that 81 % of those had begun using drugs before 15 years of age. In addition, official figures revealed that 58 % of students who misuse drugs only at their own home or a friend's home, which was a substantial increase from the 43 % recorded in 2008. These statistics show the alarming amount and nature of drug misuse and hidden drug-taking among youth in Hong Kong.
An early age at first drug use is strongly associated with the risk of developing a substance-use disorder later in life. Youth who persistently misuse substances often have an array of problems, such as academic difficulties, physical and/or mental health-related problems, poor peer relationships, a tendency to exhibit risky sexual behaviour, an increased incidence of blood-borne diseases, and involvement with the juvenile justice system. In addition, there are harmful consequences for family members of those who misuse drugs and for their communities and wider society. As a result, youth drug misuse in Hong Kong has become a major problem that cannot be overlooked.
A combination of medical treatment and counselling is the most common treatment for people with drug misuse. However, of the drug-misusing youth surveyed, more than half reported that they had never attempted to abstain from using drugs, only 12.2 % had attempted to seek help, and more than 75 % reported they would not accept medical treatment because they did not believe that they were addicted. Many non-pharmaceutical interventions have previously been applied, but a metaanalysis found that these interventions were ineffective at helping youth stop misusing drugs. A primary reason for this failure was that the patient education methods used may have made drug-misusing youth resistant to accepting interventions. Previous studies have confirmed that recovery from drug misuse is influenced by users' intentions to change, which are directly determined by their perceptions of anti-drug education and their perceived self-efficacy to change. However, their attitudes, social influences, and demographic characteristics have more indirect effects on recovery from drug misuse. Another reason for the failure of non-pharmaceutical interventions may be that standard interventional content may fail to meet the personal needs of people who misuse drugs. In addition, the meta-analysis also showed that individual interventions comprising more than one session had a stronger effect, whereas brief interventions provided an insufficient number of sessions to support behaviour changes among youth who misuse drugs. Furthermore, qualitative interviews with youth drug misusers have shown that adolescents were reluctant to seek help from sources who they believed would be judgmental, lacked expertise, would reveal their personal information to others, or would inform their parents. These findings demonstrate that feasible, individualised, and effective non-pharmaceutical interventions must be developed to assist Hong Kong's drug-misusing youth to change their unhealthy behaviour.
## Conceptual framework
Our proposed intervention will be developed with the use and guidance of (1) medical peer-delivered counselling, (2) instant messaging, and (3) brief motivational interviewing.
## Medical peer-delivered counselling
The use of peer counsellors rather than general counsellors increases the appeal of addiction services to youth and young adult drug users. Because most young people are not eligible to hold positions of power in society, they are subject to authority. This power differential means that communication between adults and youth can be difficult, which manifests as a knowledge gap when attempts are made to engage youth in drugmisuse reduction initiatives. In contrast, the equality in power status between youths means that peer-delivered risk communication is more likely to be successful. In addition, the psychological and social factors associated with the initiation and continuation of drug misuse differ between young people and older adults. Worries about the stigma of drug use, being judged, the exposure of personal information, their parents being informed, not knowing what to expect from the treatment, and major lifestyle changes are the main barriers that prevent youth and young adults from seeking substance use treatment. Youth who misuse substances As a result, more youth misusing drugs may be helped to abstain from drugs. This proposed study will inform decisions on whether it is worthwhile to invest resources in large-scale implementation of such an intervention.
often experience an array of problems, including academic difficulties, health-related problems (including mental health problems), poor peer relationships, and risky sexual behaviour. Due to their similar age, peers can address the concerns of youths by drawing on their own experiences. Knowledge-sharing and empathetic listening from peers may be more effective at helping drug-misusing youth to overcome or reduce their fear of the unknown regarding the treatment programme. However, addiction counselling on substance use is more complex than counselling for smoking cessation or alcohol and requires more professionalism. Peers with a medical background can therefore best address youths' concerns, as they have the combined advantages of a similar age and a better ability to master related knowledge and counselling skills than non-medical peers.
## Instant messaging
Given the ubiquity of mobile technology among youth and the considerable logistical barriers to face-to-facebased care, instant messaging (IM) via mobile applications (apps) such as WhatsApp or WeChat is increasingly being used as a new strategy for health promotion and for enhancement of treatment compliance. A previous study reported that drug-abusing clients preferred to use online counselling because it provided a private and emotionally safe environment that was less confrontational than traditional forms of counselling. Therefore, interaction via smartphone apps encourages those who are reluctant to seek face-to-face help to obtain direct anti-drug misuse advice. IM also enables rapid, real-time interactions, thus delivering continual professional advice and personalised support to help drug misusers overcome drug-withdrawal symptoms.
## Brief motivational interviewing
Motivational interviewing (MI) was originally developed in the field of alcohol misuse and has been successfully used to treat other health-related behaviours. In contrast to most patient education methods, MI is a direct, client-centred counselling strategy that encourages clients to explore and resolve their ambivalence and promotes their confidence in their ability to change their behaviour. Brief MI has the advantage of saving manpower and, like MI, treats individuals as advocates to initiate and continue their own behavioural change. However, individuals may experience fluctuating motivations before changing their behaviour and thus may do so in a state of ambivalence. Brief MI therefore emphasises the use of shorter and simpler strategies, which include an opening strategy, asking a subject to describe a typical day, exploring the good things and the less good things about exhibiting the targeted behaviour, providing information, comparing the future and the present, exploring concerns, and helping with decision-making. Brief MI uses these strategies as part of a dynamic and interactive process to explore and resolve individuals' ambivalence and to develop discrepancies between individuals' core beliefs and the behaviour of not engaging in a desirable health-related lifestyle practice. This process enhances the confidence of individuals and their motivation to change their behaviour.
## Incorporating medical peer counselling, im, and brief mi into the promotion of drug misuse reduction among youth
A medical peer-delivered intervention via IM interaction may be more effective in helping to eliminate or reduce drug-abusing youths' fear of personal-information exposure, in increasing their knowledge of what to expect from treatment and their ability to change their entire lifestyle, and in reducing their worries about withdrawal symptoms. This type of intervention therefore has the potential to increase help-seeking behaviour and the intention to quit among youth who misuse drugs, who are typically reluctant to seek help from services that require real-name registration. Brief MI strategies may be used to guide youth who misuse drugs to attempt a stepby-step process of behaviour change, which may enable them, with assistance, to abstain from drugs.
As stated above, the proposed intervention may be feasible and effective in reducing drug misuse among youth. However, our literature search revealed a paucity of studies that explored the effectiveness of such interventions. This proposed study will therefore involve a randomised controlled trial to determine the effectiveness of a medical peer-delivered intervention that comprises brief MI via IM in reducing drug misuse among youth in Hong Kong.
# Methods
## Study design
This proposed study will be a two-arm, parallel-group, pragmatic randomised controlled trial (RCT) with a 1:1 allocation ratio. It will compare the reduction in drug misuse at 12 months in two groups of youths: those who are individually randomised to participate in the intervention group, who will receive medical peer-delivered, interactive brief MI via IM; and those in the control group, who will receive general health information via text messaging .
## Resource and setting
This proposed study will be conducted in association with the Medical Peer Addiction Counselling Quitline Service (MedPAC Quitline), supported by the Beat Drugs Fund, Narcotics Division, Security Bureau of Hong Kong. The team of the School of Nursing at the University of Hong Kong established this first peer-oriented drug misuse quitline in 2020. Trained youth and young adult counsellors with a medical background provide peer-oriented drug-misuse counselling via MedPAC Quitline to drug-abusing youth and young adults and actively refer those in need to further treatment and rehabilitation services.
## Participants and sampling
Callers to the MedPAC Quitline and those who are actively recruited from schools and communities will be eligible service targets for the MedPAC Quitline. Youth who misuse drugs and have contacted the Med-PAC Quitline will be eligible for inclusion in this proposed study if they (1) are 25 years of age or younger, (2) report having taken drugs within the past 30 days, Because no previous study has conducted a brief telephone-based MI intervention, the sample size for the proposed RCT has been calculated according to a previous RCT that involved a computer-based MI intervention. In this previous RCT, the mean number of drug-misuse consumptions per day decreased from 11.94 at baseline to 5.52 at 12 months for participants in the intervention group, and 9.22 at baseline to 8.61 at 12 months for those in the control group. The effect size was 0.64 for the MI intervention on drug-misuse reduction, with reference to the control group. Based on these results, to detect a two-sided significant difference between groups with a power of 80 % and a significance level of 5 %, which are commonly accepted metrics for behavioural studies, the required sample size for this proposed RCT has been estimated to be 40 participants in each group. Given an expected retention rate of approximately 60 % at the 12-month follow-up, the target number of participants will be at least 134 (67 per group).
## Recruitment
## Recruitment will begin in march 2021. the medpac
Quitline Service is promoted to Hong Kong residents through a variety of platforms, such as publicity and exhibitions in secondary schools and tertiary educational institutions, non-governmental organisations, and industries; community outreach activities; advertisements in newspapers, magazines, and on the Internet; the Med-PAC Quitline website and social media platforms; posters; and snowball approaches. At least 20,000 pamphlets, booklets, and other publicity materials as well as invitation letters for this study will be delivered to secondary schools and tertiary educational institutions, including middle schools, vocational training schools, colleges, and universities. Exhibitions, carnivals, and anti-drug games and competitions will be held in secondary schools and tertiary educational institutions that agree to participate in this project. Our research assistant will screen all potential subjects and will invite those who meet the inclusion criteria to participate in the study.
## Randomisation and blinding
A simple complete randomisation method will be adopted. The subjects will be randomly allocated to either the intervention or the control group. The random numbers used for group assignment will be generated using a personal computer by another research assistant, who will not be involved in the subject recruitment. The randomisation will be performed by a research assistant who will open a serially numbered, opaque, sealed envelope that contains a card indicating the randomly allocated group. Because the subjects will be aware of their allocation according to the content received, we will be unable to completely blind them. However, a single-blind approach will be adopted, in which all outcome assessors and data analysts will be blinded to the group assignment.
## Intervention
In addition to receiving the basic service provided by the MedPAC Quitline, the subjects in the intervention and control groups will receive brief MI via IM interaction and general health text-messages, respectively.
## Intervention group
The interaction communication using brief MI will be applied via the chat function of IM apps (e.g., WhatsApp or WeChat). The intervention will be performed for 6 months at a frequency of at least twice a week. To begin the conversation, trained peer counsellors will actively contact the participants and explore their lifestyles and their current barriers to and facilitators of change. The counsellors will do this by asking participants about their current lifestyles, drug-misuse habits, and levels of stress to establish a collaborative relationship with and obtain a general understanding of each participant (opening strategy). To further understand why a participant does not intend to reduce his/her drug misuse, the counsellor (interventionist) will review a typical day of the subject (a typical day) and ask about the advantages and disadvantages of not reducing drug misuse. Each participant will receive further information about the health benefits of reducing or quitting drug use, if he/she is interested in doing so (providing information). To elicit information on discrepancies, each participant will be encouraged to talk about his/her ideal future and to compare the past and the present (the future and the present). The interventionist will also listen to, explore, and summarise the concerns of the participant (exploring concerns). The interventionist will resolve the ambivalence of the participant by describing what others have done in a similar situation and presenting options, while not urging a participant to make a decision (helping with decisionmaking). Importantly, the participant's confidence about reducing or quitting his/her drug use will be enhanced throughout the process of receiving personalised health advice, as this advice will emphasise the health benefits of reducing or quitting and encourage the participant to be aware of resources that enable and facilitate such behavioural change. At the request of the participant, the interventionist will provide more comprehensive information and suggestions on how to reduce drug use, including the skills to overcome withdrawal symptoms.
## Control group
The subjects will receive general health information via text message, such as 'do physical exercise for at least 30 minutes per week to keep yourself healthy' , for 6 months at a frequency of twice a week.
## Measures
The demographics and baseline characteristics of participants will be collected during the baseline phone counselling. A structured questionnaire will be developed for this purpose by adopting or modifying international and/ or locally validated instruments. The questionnaire will gather information on a participant's (a) drug misuse profile, including the types of drugs used, his/her number of days of use, his/her daily drug consumption, and his/her history of quitting; (b) beliefs and values relating to drug use, using the Chinese Drug Involvement Scale; (c) drug misuse-related knowledge; (d) attitudes towards their drug abuse; (e) contemplation stage of drug-misuse abstinence, using the Contemplation Ladder Tool; (f ) treatment needs and motivation, using the Texas Christian University Motivation and Readiness For Treatment form; (g) self-efficacy to avoid drug use, using the Adolescent Relapse Coping Questionnaire;
(h) quality of life, using the EuroQol 5-Dimension 5-Level Questionnaire; (i) depression level, using the Center for Epidemiologic Studies Depression Scale; and (j) self-esteem, using the Rosenberg Self-Esteem Scale. The Chinese versions of the above scales have been evaluated to show acceptable validity and reliability.
All participants will receive follow-up telephone calls from our trained research assistant at 1 week, 1 month, and 3, 6, 9, and 12 months. During each follow-up telephone call, the participants will be asked to complete the structured questionnaire. Participants who are lost to one follow-up will be contacted again in the next follow-up.
## Outcomes
The primary outcome will be changes in the participants' reductions in self-reported drug consumption at 12 months compared to that at baseline. Secondary outcomes will be (1) changes in the participants' reductions in self-reported drug consumption at 6 months compared to that at baseline; (2) 30 days self-reported drug abstinence at 6 and 12 months; (3) changes in relapse risk among the quitters at 6 and 12 months compared to that at baseline; (4) changes in the participants' contemplation stages at 6 and 12 months compared to those at baseline; and (5) treatment needs and motivation at 6 and 12 months compared to those at baseline.
# Data analysis
Data analysis will be performed using the Statistics Package for the Social Sciences (SPSS) version 25.0 for Windows (SPSS Inc., Chicago, IL, USA). Analyses will be conducted on an intention-to-treat basis, with missingvalue substitutions based on the last observation carried forward. A two-sided statistical significance level of 0.05 will be adopted for all analyses. Descriptive statistics will be used to detail participants' demographic characteristics and drug misuse profiles. Frequencies and percentages or means and standard deviations will be used to present categorical data and continuous data, respectively. The differences between intervention and control groups, in terms of sociodemographic factors and outcomes, will be compared using a chi-square test for categorical variables and by an analysis of variance for continuous variables.
# Main analysis
Odds ratios, risk differences, and 95 % confidence intervals will be used to compare the primary and secondary outcomes between the two trial groups. A generalised estimating equation model-based analysis will be used to summarise the intervention effect on the primary and secondary outcomes with and without adjustment characteristics, referral to a treatment institution, and the interaction times.
# Exploratory analysis
A subgroup analysis of the participants in the intervention group will be conducted using different IM apps and data on referrals to treatment institutions to determine whether the relative effects of study treatments vary significantly among subgroups of IM app users. Propensity score matching, including demographics and types of abused drugs, will be used to further determine the true cause-and-effect relationship between the intervention and control groups, with restrictions of these subgroups. Missing outcomes will be assumed to be dependent on observed data (missing at random), and the observed data will be analysed with a multiple imputation procedure to impute the missing values, as a sensitivity analysis.
## Text mining of im app group conversations
All interaction content will be archived and anonymised to remove identifying personal information. Due to the large number of messages that will be collected from the IM apps, we will use automatic, computational textmining and visualisation of the dataset for the content analysis. First, using a lexicon of keywords derived from our qualitative analysis of the pilot RCT (unpublished), we will develop a heat-map visualisation to illustrate the prevalence of the discussion topics. Second, we will apply topic modelling to investigate emerging themes in our content dataset, using the Malletimplementation of the Latent Dirichlet Allocation topic modelling algorithm. Topic modelling algorithms use a text dataset (in this case, the content dataset) as input and then output a set of topics (and their associated keywords) in addition to estimates of the proportion of each topic.
# Discussion
Four caveats should be considered when the findings of this proposed study are interpreted. First, this pragmatic RCT will not guarantee 'full' intervention compliance by all participants. For instance, the participants in the intervention group will be encouraged to interact with the interventionists, but they will have the right to ignore the IMs. Thus, they will be able to choose to either respond to an interventionist or to remain silent throughout the intervention period. As a result, although the RCT will show the effectiveness of delivering such an interaction for drug-misuse cessation, we expect that there will be a variety of levels of participation in the interaction. Nevertheless, all IM conversations will be archived for further content analysis and compliance analysis. The actual effect of participating in an interaction will be investigated by filtering out participants who exhibit low participation. Second, we will only recruit drug-misusing youths in this RCT who have also enrolled in the MedPAC Quitline service, and thus the findings will not be generalisable to unassisted youths who misuse drugs. Future trials that customise the intervention for drug-misusing youths recruited from outside the service will be warranted. Third, recreational users, who are common among young people with substance use, will be also included in this trial. This could potentially limit the room for significant improvement. Considering recreational drug use is a common situation among the youth population, this group should not be ignored. To explore the effect of the intervention in the different participants with such use patterns, we will conduct post hoc analysis for subgroups of the pattern of use, dependence level, readiness to quit in the next 30 days, and previous quit history at baseline. Fourth, all measures will be self-reported. However, the validity of self-reported versus biochemical assessment of substance use may vary according to a variety of contextual factors. Evidence of the ability of existing biochemical methods to detect reductions in drug misuse remains scarce. Nevertheless, a previous study indicated that self-report may outperform biological assessment for the assessment of patterns of drug use over the past month, because biological assessment typically only captures substance use that has occurred in the past 5 days.
The findings from this study will be disseminated locally and internationally through manuscript publications in peer-reviewed journals and conference presentations at national and international platforms. The proposed intervention has the potential to increase the help-seeking behaviour and intention to quit of youth who misuse drugs and are reluctant to seek help from services that require real-name registration. With such assistance, more youth may abstain from drug misuse. This proposed study will be the first RCT to assess the effectiveness of a medical peer-delivered, IM-based, interactive intervention in reducing drug misuse among youth in Hong Kong. Thus, this study will inform decisions on whether it is worthwhile to invest resources in large-scale implementation of such an intervention to reduce drug misuse in Hong Kong's youth.
## Abbreviations
|
Effects of antidiabetic drugs on epicardial fat
cytes with pathophysiological properties similar to those of visceral fat, located in the space between the myocardial muscle and the pericardial sac. When compared with subcutaneous adipose tissue, visceral adipocytes show higher metabolic activity, lipolysis rates, increased insulin resistance along with more steroid hormone receptors. The epicardial adipose tissue interacts with numerous cardiovascular pathways via vasocrine and paracrine signalling comprised of pro-and anti-inflammatory cytokines excretion. Both the physiological differences be tween the two tissue types, as well as the fact that fat distribution and phenotype, rather than quantity, affect cardiovascular function and metabolic processes, establish epicardial fat as a biomarker for cardiovascular and metabolic syndrome. Numerous studies have underlined an association of altered epicardial fat morphology, type 2 diabetes mellitus (T2DM) and adverse cardiovascular events. In this review, we explore the prospect of using the epicardial adipose tissue as a therapeutic target in T2DM and describe the underlying mechanisms by which the antidiabetic drugs affect the pathophysiological processes induced from adipose tissue accumulation and possibly allow for more favourable cardiovascular outcomes though epicardial fat manipulation.AbstractEpicardial adipose tissue is defined as a deposit of adipo-Submit a Manuscript: http://www.f6publishing.com
# Introduction
Subcutaneous (SCAT) and visceral adipose tissue (VAT) are two extremely heterogenous tissue types, differentiated by anatomical, molecular, cellular, physiological and clinical characteristics [bib_ref] Subcutaneous and visceral adipose tissue: structural and functional differences, Ibrahim [/bib_ref]. Researchers have suggested that the variation of composition and function of the two tissue types is induced very early in the tissue developmental pathway, as a result of adipose stem cell distinction [bib_ref] Functional differences in visceral and subcutaneous fat pads originate from differences in..., Baglioni [/bib_ref]. VAT has an anatomically distinct distribution in the me sentery and omentum, when compared to SCAT that is mainly located in the femerogluteal area, back and abdominal wall [bib_ref] Subcutaneous and visceral adipose tissue: structural and functional differences, Ibrahim [/bib_ref]. As a result of the anatomical differences, vascularization and innervation vary between the tissues, with VAT having superior nerve and vascular networks, as well as draining into the portal system of veins. Based on the aforementioned anatomical link, the "portal theory" of metabolic inflammation states that free fatty acids and proinflammatory molecules from VAT, interact with the liver, promoting hepatocellular dysfunction in the form of insulin resistance and steatosis [bib_ref] Visceral fat and metabolic inflammation: the portal theory revisited, Item [/bib_ref]. The dissimilarity in cellular composition between SCAT and VAT is a result of divergent ratio of large to small adipocytes between the two tissues. Large, metabolically dysfunctional, adipocytes, predominate in VAT, while SCAT is mainly composed by small adipocytes with higher free fatty acids and triglycerides capacity and increased insulin sensitivity [bib_ref] The cell size and distribution of adipocytes from subcutaneous and visceral fat..., Fang [/bib_ref] [bib_ref] High fat diet consumption differentially affects adipose tissue inflammation and adipocyte size..., Poret [/bib_ref].
The signaling pathways activated in the two tissue types vary due to a shift in receptor distribution and adipokine synthesis [bib_ref] Subcutaneous and visceral adipose tissue: structural and functional differences, Ibrahim [/bib_ref]. Glucocorticoid and androgen receptors present with a higher density in VAT while oestrogen receptors are more active in SCAT. Adrenergic signaling patterns are distinct for the two cell populations, with VAT being more β3 and α2 adrenoreceptor sensitive [bib_ref] Effect of functional sympathetic nervous system impairment of the liver and abdominal..., Fountaine [/bib_ref]. The biologically active molecules produced by the adipose tissue, referred to as adipokines, are formed and released at different rates between VAT and SCAT. Adipokines are the basis of adipose tissue participating in and regulating endocrine and paracrine funtions [bib_ref] Functional and structural features of adipokine family, Raucci [/bib_ref]. The diversity of adipokines is directly linked to sympathetic excitation, metabolic regulation, including insulin sensitivity and appetite, inflammatory response and other homeostatic mechanisms. Some of the most prominent members of this family, as far as metabolic processes and cardiovascular function are examined, are: leptin, adiponectin, interleukin6 (IL6), plasminogen activator inhibitor1 (PAI1) and tumor necrosis factor alpha (TNFα) [bib_ref] Subcutaneous and visceral adipose tissue: structural and functional differences, Ibrahim [/bib_ref] [bib_ref] Functional and structural features of adipokine family, Raucci [/bib_ref]. Leptin levels are elevated in obese subjects, along with TNFα, IL6 and PAI1 that are proatherogenic and prodiabetic, in contrast to plasma adiponectin that protects against vascular damage and metabolic syndrome and is reduced, as it would be expected [bib_ref] Obesity, adipokines and neuroinflammation, Aguilar-Valles [/bib_ref]. The variety in cytokine profile, along with the anatomic and cellular diversity that differentiate SCAT and VAT clarify and support the physiological and metabolic properties excreted by each adipocyte group. VAT cells allow for increased insulinmediated glucose uptake and are more insulinresistant and lipolysisprone that those of SCAT. In contrast, the latter, exhibit a greater capacity for postprandial free fatty acid and triglyceride [bib_ref] Subcutaneous and visceral adipose tissue: structural and functional differences, Ibrahim [/bib_ref]. Taking into consideration the pivotal role of VAT in metabolic impairment, as such is supported by its aforementioned properties, it is comprehensible that studying the metabolic properties of visceral adiposity and mainly, organ-specific depositions, such as epicardial fat, has been incremental in the process of stratification of cardiometabolic risk factors. In this review, we aim to compare the morphology of epicardial fat deposits between nondiabetic individuals and subjects with type 2 diabetes mellitus (T2DM). Moreover, we will discuss the affect excreted by the antidiabetic substances in epicardial VAT, while contemplating on its clinical utility, as estimated by means of cardiovascular risk reduction.
## Function and composition of
## Epicardial fat
Epicardial adipose tissue (EAT) is an adipocyte depot of VAT with anatomical continuity to the myocardial tissue, located under the visceral layer of the pericardium . It has been suggested that it can serve as a quantifiable and modifiable therapeutic target for cardiovascular adverse events, as it can be measured with noninvasive imaging techniques such as twodimensional echocardiography, computed tomography (CT) and magnetic resonance imaging (MRI) [bib_ref] Epicardial fat: definition, measurements and systematic review of main outcomes, Bertaso [/bib_ref]. Spatial imaging, as such is provided by MRI and CT scans, is preferable to that of twodimensional echocardiography technique, in order to accurately measure the thickness of EAT. Along with technical shortcomings, operator and subject related variability deem echocardiographic imaging a formidable solution solely because of the rapid and costeffective patient assessment it facilitates. Otherwise, MRI is considered to be the gold-standard method for EAT quantification and area placement, even though three dimensional image reconstruction by utilizing multidetectorrow CT is slightly superior in achieving the latter [bib_ref] Epicardial fat: definition, measurements and systematic review of main outcomes, Bertaso [/bib_ref] [bib_ref] Measurement of Epicardial Fat Thickness by Echocardiography Presents Challenges Arq, Gulgun [/bib_ref]. On a cellular level, the epicardial adipocytes are embryologically derived from the splachnopleuric meso derm, similarly to the mesenteric and omental adipocytes. EAT is characterized by high cellularity, defined by the concentration of adipocytes in this tissue being notably higher than that of other depots of adipose tissue [bib_ref] The role of epicardial and perivascular adipose tissue in the pathophysiology of..., Ouwens [/bib_ref].
EAT is a depot of white adipocytes, cells that specialize in energy storage, as opposed to brown adipocytes that are involved in energy expenditure [bib_ref] beige/brite: different adipose cells for different functions?, Giralt [/bib_ref]. EAT extends on an area exceeding 80% of the myo cardial total surface in an otherwise healthy individual, spreading heterogeneously, mostly accumulating on the lateral and anterior walls of the right atrium [bib_ref] Clinical importance of epicardial adipose tissue, Nagy [/bib_ref]. The physiological structure and composition of EAT variates depending on age, gender, body weight and ethnicity [bib_ref] Clinical importance of epicardial adipose tissue, Nagy [/bib_ref]. The properties of EAT and its contribution in physiological and pathophysiological pathways have been extensively described. Due to its spatial distribution, EAT acts as a mechanical and thermoprotective layer for the myocardial tissue and coronary arteries. Through endocrine and paracrine function, epicardial adipocytes ameliorate endothelial response of the coronaries and insulin sensitivity, while reducing oxidative stress of the cardiac tissue. Additionally, the small adipocytes of EAT are characterized by high rates of free fatty acids turnover, allowing for both energy supply and storage as demand shifts [bib_ref] Clinical importance of epicardial adipose tissue, Nagy [/bib_ref].
## Epicardial fat, type 2 diabetes
## Mellitus and coronary artery disease
The prevalence of type 2 diabetes mellitus (T2DM) has quadrupled in the last three decades according to International Diabetes Federation (IDF) reports. The epidemic escalation has been attributed to numerous factors including population aging as a result of improved healthcare, socioeconomic development, unhealthy diet regimes and sedentary lifestyle [bib_ref] Global aetiology and epidemiology of type 2 diabetes mellitus and its complications, Zheng [/bib_ref]. EAT has been associated with numerous pathophysio logical processes, such as coronary artery disease , even though the significance of such association has not been adequately supported by all relevant studies , electrophysiological abnormalities of the heart [bib_ref] Pericardial fat is associated with atrial fibrillation severity and ablation outcome, Wong [/bib_ref] [bib_ref] Pericardial fat is associated with prevalent atrial fibrillation: the Framingham Heart Study, Thanassoulis [/bib_ref] , cardio vascular disease in human immunodeficiency virus treated with antiretroviral therapy [bib_ref] Epicardial adipose tissue is an independent marker of cardiovascular risk in HIV-infected..., Guaraldi [/bib_ref] , amplified severity of non alcoholic fatty liver disease [bib_ref] Epicardial fat, cardiac geometry and cardiac function in patients with non-alcoholic fatty..., Petta [/bib_ref] [bib_ref] Increased severity of liver fat content and liver fibrosis in nonalcoholic fatty..., Brouha [/bib_ref] , metabolic syndrome and increased cardiovascular risk along with decline of renal function in individuals with T2DM . The pathophysiological pathways linking T2DM and EAT, support a multifactorial causative relationship be tween EAT attributes and structure such as volume and endocrine function and cardiovascular disease severity in the diabetic individual.
EAT deposition can be associated with coronary vas cular disease pathogenesis mainly by the dysregulation of cardiac metabolic processes and the disruption of the epicardial and myocardial structural integrity. Other mecha nisms that could be involved in the interaction between EAT and coronary vasculature are nerve damage and impaired cryoprotection of the heart [bib_ref] The role of epicardial adipose tissue in cardiac biology: classic concepts and..., Antonopoulos [/bib_ref] [bib_ref] Epicardial Fat in Nonalcoholic Fatty Liver Disease: Properties and Relationships With Metabolic..., Psychari [/bib_ref]. Furthermore, the epicardial adipocytes exhibit and arrhythmogenic potential, a theory suggested by many clinical trials ex ploring the causative relationship between EAT and atrial fibrillation [bib_ref] Pericardial fat is associated with atrial fibrillation severity and ablation outcome, Wong [/bib_ref] [fig_ref] Figure 1: Mechanisms involved in the crossplay between the heart and the epicardial adipocytes [/fig_ref].
## Epicardial fat and antidiabetic drugs
## Biguanides
Metformin is the most common first-line treatment choice for T2DM and a member of the biguanides drug class. Oral administration of the substance affects the liver and gut metabolic pathways in order for its hypoglycemic attributes to be put into effect [bib_ref] Mechanism of Metformin: A Tale of Two Sites, Song [/bib_ref]. Hepatic gluconeogenesis, glucose uptake, glycolysis and glucogen synthesis are some of the processes altered by metformin via AMPactivated protein kinase (AMPK)dependent and independent pathways [bib_ref] The mechanisms of action of metformin, Rena [/bib_ref]. At this point in time, there seem to be no randomized controlled trials designed for clarification of the effects excreted by metformin on the volume or function of EAT. Despite the fact that metformin has not been com pared with placebo, as of yet, studies conducted on sitagliptin and liraglutide as addon therapy to metformin monotherapy, combined with epicardial fat measurement, can be used as a preliminary source of data [bib_ref] Liraglutide causes large and rapid epicardial fat reduction, Iacobellis [/bib_ref].
Results from these trials confirm the inferiority of metformin monotherapy when compared to metformin/ sitagliptin and metformin/liraglutide for reduction of EAT volume. The findings can be either attributed to the synergy of two antidiabetic substances, affecting the EAT in a more effective manner than metformin alone, or to the complete lack of action of the biguanide class on the cardiac VAT deposits. The latter is supported by the results of the study performed by Iacobellis et al [bib_ref] Liraglutide causes large and rapid epicardial fat reduction, Iacobellis [/bib_ref] , that noted no EAT reduction in the metformin group during the 6mo follow up period. Conversely, metformin has been previously shown to have positive effects on VAT, inducing its reduction on diabetic subjects [bib_ref] Beneficial effects of metformin on energy metabolism and visceral fat volume through..., Tokubuchi [/bib_ref]. Furthermore, studies have confirmed a metformin-induced increase of plasma omentin1 levels, an adipokine produced by epicardial fat that ameliorates insulin sensitivity, inflammatory response and cardiovascular function [bib_ref] Adipose Tissue-Derived Omentin-1 Function and Regulation, Watanabe [/bib_ref]. Given the contradicting evidence concerning metformin, there is need for further research, as a definite conclusion on the manner by which biguanides interact with epicardial fat can only be provided by a randomized controlled trial with EAT measurement.
## Alpha-glucosidase inhibitors
Alphaglucosidase inhibitors (αGIs) are a class of anti diabetic drugs acting in the epithelium of the small intes tine mainly by delaying the digestion of carbohydrates through reversible and competitive inhibition of intestinal alphaglucosidases, consequently reducing glucose absorp tion and attenuating postprandial hyperglycemia [bib_ref] The mechanism of alpha-glucosidase inhibition in the management of diabetes, Bischoff [/bib_ref]. Some αGIs are acarbose, miglitol and voglibose. Similarly to the biguanide class of antidiabetic medication, there is a lack of data concerning the administration of αGIs and their effect on EAT mass, volume or metabolic activity.
## Thiazolidinediones
Thiazolidinediones (TZDs), also known as glitazones, are peroxisome proliferatoractivated receptor (PPAR) agonists with numerous actions, spanning from glycemic and lipid control to inflammatory signaling and cell cycle mediation [bib_ref] Thiazolidinedione drugs in the treatment of type 2 diabetes mellitus: past, present..., D A V I D S O N M A, M A T T I S O N D R , A Z O U L A Y L , K R E W S K I D [/bib_ref]. The phenomenon of glitazone treatment and subsequent increase in body weight that has been supported by the results of numerous studies appears to be tissue-specific, since the VAT depot of the subjects remains unaffected while there is a shift of excess energy storage towards the SCAT . Furthermore, pioglitazone treatment in T2DM or metabolic syndrome has been shown to attenuate the inflammatory signature of EAT by means of decreased expression of proinflammatory interleukins (IL) such as IL-1β, IL1Ra and IL10 [bib_ref] Inflammatory genes in epicardial fat contiguous with coronary atherosclerosis in the metabolic..., Sacks [/bib_ref]. In addition to the positive effect on the metabolic profile of EAT, pioglitazone can affect the epicardial fat depot directly. Nagai et al [bib_ref] Abstract 710: Pioglitazone Treatment Reduces Epicardial Fat in Patients with Type 2..., Nagai [/bib_ref] recruited 97 T2DM individuals that were divided into two groups according to baseline EAT thickness and underwent therapy with pioglitazone, along with EAT thickness mea surement, at the beginning and after a ninemonth follow up period. Pioglitazone reduced the EAT thickness in both groups, with more prominent results in the subjects that had a greater EAT depot at baseline. A different TZD, rosiglitazone, when administered to mice, induced the expression of brown adipose tissue specific proteins by the EAT, a tissue type normally presenting having a hormonal profile consistent with that of white adipose tissue [bib_ref] Early induction of a brown-like phenotype by rosiglitazone in the epicardial adipose..., Distel [/bib_ref]. Brown adipose tissue has been linked to high rates of lipid turnover and reduced body weight, while it is essential for thermogenesis and homeostasis, in contrast to white adipose tissue that serves as an energy reservoir for the body [bib_ref] Epicardial Fat: Physiological, Pathological, and Therapeutic Implications, Salazar [/bib_ref]. The data derived from the studies examining the effect of glitazones and EAT correlates with the established theory that TZDinduced weight gain is not concurrent with VAT deposition. Moreover, TZDs appear to have a favorable effect on EAT both by regulation of endocrine functions and mass reduction.
## Incretins
Glucagonlike peptide1 (GLP1) is an incretin hormone that delays gastric motility, supresses appetite, stimulates glucosedependent insulin and decreases glucagon se cretion. The enzyme dipeptidyl peptidase4 (DPP4) deactivates GLP1 interrupting all incretinstimulated signalling. DPP4 inhibitors (DPP4i) are one of the two categories of antidiabetic drugs acting on the incretin pathway, the other being GLP1 receptor agonists (GLP1 RA). DPP4i inhibit both GLP1 and glucose dependent insulinotropic polypeptide (GIP) degradation, thereby increasing plasma concentrations and stimulating the pancreatic βcell in order to better regulate glucose homeostasis.
## Dpp-4 inhibitors
The class of DPP4is includes sitagliptin, vildagliptin, saxa gliptin, linagliptin and alogliptin [bib_ref] DPP-4 inhibitors: impact on glycemic control and cardiovascular risk factors, Dicker [/bib_ref]. Sitagliptin is the only DPP4i whose effect on epicardial fat has been studied at this point in time. LimaMartínez et al formed a 24wk
interventional plan for 26 obese subjects with T2DM inade quately controlled on metformin monotherapy. Subjects meeting the inclusion criteria were introduced to a new regimen, receiving sitagliptin/metformin at a dosage of 50 mg/1000 mg respectively, twice a day. EAT deposits were reduced in size by approximately 15% (from 9.98 ± 2.63 to 8.10 ± 2.11 mm, P = 0.001) while the percentage of reduction in EAT was analogous to that of VAT (r = 0.456, P = 0.01). While the aforementioned study establishes a favourable effect of sitagliptin on the mass of epicardial VAT, there is definite need for further research, in order to establish the reduction of EAT as a class effect of DPP 4is .
## Glp-1 receptor agonists
GLP1 RAs utilize the "incretin effect", similarly to DPP4 inhibitors, so as to attenuate the diabetesinduced hyper glycemia. GLP1 RAs are divided into short and long acting compounds that activate the GLP1 receptor in a manner similar to that of the endogenous GLP1 [bib_ref] GLP-1 receptor agonists for individualized treatment of type 2 diabetes mellitus, Meier [/bib_ref]. Epicardial adipocytes have been shown to express GLP1 and 2 receptor genes by use of RNA sequencing, while the possible quantity and dispersion pattern of the receptors in vivo has not been described [bib_ref] Human Epicardial Fat Expresses Glucagon-Like Peptide 1 and 2 Receptors Genes, Iacobellis [/bib_ref]. Furthermore, GLP1 and GLP1 receptor signaling affect the differentiation and growth of adipocytes by regulation of fatty acid synthase activity [bib_ref] GLP-1/GLP-1R Signaling in Regulation of Adipocyte Differentiation and Lipogenesis, Chen [/bib_ref]. Even though, the effects of numerous GLP1 RAs have been studied in correlation to the metabolic regulation or mass reduction of visceral adipose tissue, the clinical trials concerning organspecific deposits are few [bib_ref] Liraglutide causes large and rapid epicardial fat reduction, Iacobellis [/bib_ref]. Current data on EAT remodeling by GLP1 RAs is derived by two studies, conducted with liraglutide and exenatide [bib_ref] Liraglutide causes large and rapid epicardial fat reduction, Iacobellis [/bib_ref] [bib_ref] Exenatide decreases liver fat content and epicardial adipose tissue in patients with..., Dutour [/bib_ref].
A trial designed by Iacobellis et al [bib_ref] Liraglutide causes large and rapid epicardial fat reduction, Iacobellis [/bib_ref] included 95 T2DM obese subjects with hemoglobin A1c ≤ 8% while being treated with metformin. The patients were randomized into two groups to either receive a combination of met formin/liraglutide, with the latter being administered once daily, in doses up to 1.8 mg, or stay on metformin monotherapy, up to 1000 mg administered twice daily for 6 mo. EAT thickness measurements were acquired by ultrasound imaging at baseline and at 3 and 6 mo. Subjects in the liraglutide group presented with a decline in EAT thickness, 29% and 36% reduction from baseline at 3 and 6 mo respectively. Given that there were no similar changes in the metformin group, the EAT mass reduction is considered to be an effect of the liraglutide treatment, or possibly a result of the synergy between the two antidiabetic substances. The study involving exenatide had a broader spec trum than that of liraglutide, examining the effect of the GLP1RA on numerous VAT depots including epicardial, myocardial, hepatic and pancreatic adipose pads. Measure ments of EAT thickness were performed by magnetic resonance imaging and spectroscopy at baseline and at 26 wk. A total of 44 obese individuals with uncontrolled T2DM, originally receiving oral therapy, were randomized to two groups, either receiving exenatide or other treatment chosen according to the local guidelines. EAT was reduced by approximately 8.8% after treatment with exenatide and by 1.2% on the patients receiving oral therapy, with the difference between the two being statistically significant (P = 0.003) [bib_ref] Exenatide decreases liver fat content and epicardial adipose tissue in patients with..., Dutour [/bib_ref]. Current research conducted on incretin treatment and ectopic adipose tissue deposition supports the theory that EAT reduction could be a class effect of GLP1RAs and possibly a mediator of their beneficial actions on cardiovascular disease in the diabetic and obese subjects.
## Sglt-2 inhibitors
Sodiumglucose cotransporter 2 (SGLT2) inhibitors are a novel class of antidiabetic substances that bind on the SGLT2 transporter in the proximal tubule of the kidney, facilitating glucose excretion via hindering reabsorption. SGLT2mediated reabsorption constitutes the main pathway by which the renal system maintains glucose homeostasis [bib_ref] SGLT2 inhibitors to control glycemia in type 2 diabetes mellitus: a new..., Jabbour [/bib_ref]. Administration of SGLT2 inhibitors in obese individuals with T2DM has been linked with ab dominal VAT size reduction [bib_ref] Sodium-glucose Co-transporter 2 Inhibitors Reduce the Abdominal Visceral Fat Area and May..., Tosaki [/bib_ref]. Additionally, the effects of SGLT2 inhibition on tissuespecific depots such as EAT have been clarified by studies performed on luseogliflozin, ipragliflozin, canagliflozin and dapagliflozin . EAT measurements following a 12wk period of luseogliflozin administration demonstrate that treatment with luseogliflozin can reduce EAT volume in combination with adipocyterelated inflammation and metabolic dysregulation on type 2 diabetic patients. Along with EAT, numerous parameters were modified after luseogliflozin therapy including body weight, fasting plasma glucose, insulin resistance and Creactive protein (CRP) levels. A positive correlation was established between CRP and EAT reduction (r = 0.493, P = 0.019), suggesting a concurrent effect of the SGLT2 inhibitor on both the adipose tissue mass and metabolic activity [bib_ref] Luseogliflozin reduces epicardial fat accumulation in patients with type 2 diabetes: a..., Bouchi [/bib_ref]. Similar results concerning both EAT and biomarkers reduction were acquired after ipragliflozin administration, in a study designed similarly to that conducted for luseogliflozin. The two models differed in the selection of the study population, with luseogliflozin treatment being applied to obese subjects while ipragliflozin was administered to nonobese T2DM individuals [bib_ref] Ipragliflozin Reduces Epicardial Fat Accumulation in Non-Obese Type 2 Diabetic Patients with..., Fukuda [/bib_ref]. Yagi et al [bib_ref] Canagliflozin reduces epicardial fat in patients with type 2 diabetes mellitus, Yagi [/bib_ref] studied the interaction of canagliflozin and EAT during a 6mo period of treatment. The sample consisted of type 2 diabetic individuals, each of which was administered 100 mg of canagliflozin once daily. During the followup period EAT was evaluated by echo cardiographic imaging while VAT and SCAT size fluctuation was monitored by use of impedance methods. The mean EAT thickness values were 9.3 mm and 7.3 mm at base line and at 6 mo, respectively, with the change observed being statistically significant (P < 0.001) while there was only a trend for VAT and SCAT reduction. Dapagliflozin and epicardial adiposity were examined through two different clinical trials, studying both the shift in metabolic activity and size of the adipocytes after treatment [bib_ref] Effects of dapagliflozin on human epicardial adipose tissue: modulation of insulin resistance,..., Díaz-Rodríguez [/bib_ref] [bib_ref] The effect of dapagliflozin treatment on epicardial adipose tissue volume, Sato [/bib_ref]. The metabolic profile of adipocytes promoted by dapagliflozin was assessed ex vivo on fat explants obtained from patients undergoing cardiac surgery on a trial designed by DíazRodríguez et al [bib_ref] Effects of dapagliflozin on human epicardial adipose tissue: modulation of insulin resistance,..., Díaz-Rodríguez [/bib_ref]. Glucose uptake, transporter expression and adipokine secretion patterns were altered as a result of dapagliflozin application, a change indicative of a positive metabolic reform of the tissue induced by SGLT2 inhibition. Simulta neously, Sato et al [bib_ref] The effect of dapagliflozin treatment on epicardial adipose tissue volume, Sato [/bib_ref] followed a more conventional ap proach, estimating the dapagliflozin-induced EAT volume reduction, by means of computed tomography imaging. Individuals receiving both dapagliflozin and other regimens for T2DM control were observed for 6 mo, with biomarker and EAT measurement at baseline and following comp letion of the study. While the two groups had similar EAT size measurement before the initiation of dapagliflozin therapy, the patients receiving the SGLT2 inhibitor pre sented with a greater reduction of epicardial VAT volume after treatment (16.4 ± 8.3 for the dapagliflozin vs 4.7 ± 8.8 cm 3 for the control group, P = 0.01), combined with lowered plasma levels of inflammatory adipokines. Numerous studies conducted on many members of the SGLT2 inhibitor class of antidiabetic substances support the conclusion that EAT undergoes a multifaceted remodelling after SGLT2 inhibition, a trend that could be considered a class effect. The interconnection established between SGLT2 inhibitors and a known factor of cardiovascular risk such as epicardial adiposity could elucidate the manner by which the members of this class are cardioprotective, while, providing grounds for further therapeutic targeting of EAT [fig_ref] Table 1: Antidiabetic drug and their effect on epicardial adipose tissue [/fig_ref].
# Conclusion
Epicardial adipose tissue exhibits a unique metabolic and pathophysiologic profile, as a result of its anatomical location and its cellular composition, rendering it an appealing therapeutic target for reducing cardiovascular risk and enabling endocrine homeostasis in the dys metabolic individual. The recent studies concerning the effect of the antidiabetic substances on the multifactorial cardiomyopathy of the diabetic patient and, by extension, on epicardial adiposity, have yielded interesting results that support the use of treatment for a targeted approach, in order to reduce the size and metabolic activity of ectopic adipose tissue clusters. Despite the capacity of certain treatment regimens, mostly newer agents like GLP1 agonists and SGLT2 inhibitors, in the manipulation of both structural and functional parameters of the epicardial adipose tissue, the clinical efficacy of this approach remains unsubstantiated for the time being. There is definite need for further research, in order to elucidate whether the targeting of epicardial adiposity facilitates the procurement of better outcomes for individuals with diabetes and cardiovascular disease, while, additionally, clarify the manner by which the antidiabetic substances can attain such results.
## Antidiabetic drug
Effect on epicardial adipose tissue
## Biguanides
No effect/Possible synergistic effect with DPP-4 and/or GLP-1 [bib_ref] Liraglutide causes large and rapid epicardial fat reduction, Iacobellis [/bib_ref]
## Alpha-glucosidase inhibitors
Lack of data concerning the effect of this class Thiazolidinediones
Decreased inflammatory cytokine release and thickness of EAT (pioglitazone) modulation of cellular hormonal profile (rosiglitazone) [bib_ref] Abstract 710: Pioglitazone Treatment Reduces Epicardial Fat in Patients with Type 2..., Nagai [/bib_ref] [bib_ref] Early induction of a brown-like phenotype by rosiglitazone in the epicardial adipose..., Distel [/bib_ref]
## Dipeptidyl peptidase-4 inhibitors
Reduction of EAT thickness (sitagliptin) Glucagon-like peptide-1 receptor agonists Reduction of EAT thickness (liraglutide and exenatide) [bib_ref] Liraglutide causes large and rapid epicardial fat reduction, Iacobellis [/bib_ref] [bib_ref] Exenatide decreases liver fat content and epicardial adipose tissue in patients with..., Dutour [/bib_ref]
## Sodium-glucose cotransporter 2 inhibitors
Reduction of EAT thickness (luseogliflozin, ipragliflozin, canagliflozin, dapagliflozin) and inflammation (luseogliflozin, ipragliflozin, dapagliflozin) [bib_ref] Luseogliflozin reduces epicardial fat accumulation in patients with type 2 diabetes: a..., Bouchi [/bib_ref] [bib_ref] Ipragliflozin Reduces Epicardial Fat Accumulation in Non-Obese Type 2 Diabetic Patients with..., Fukuda [/bib_ref] [bib_ref] Canagliflozin reduces epicardial fat in patients with type 2 diabetes mellitus, Yagi [/bib_ref] [bib_ref] Effects of dapagliflozin on human epicardial adipose tissue: modulation of insulin resistance,..., Díaz-Rodríguez [/bib_ref] [bib_ref] The effect of dapagliflozin treatment on epicardial adipose tissue volume, Sato [/bib_ref]
[fig] Figure 1: Mechanisms involved in the crossplay between the heart and the epicardial adipocytes. [/fig]
[table] Table 1: Antidiabetic drug and their effect on epicardial adipose tissue [/table]
|
Sarcopenia in Patients With Parkinson's Disease: A Systematic Review and Meta-Analysis
Background: Parkinson's disease (PD) and sarcopenia are two common diseases in aging people. To date, the prevalence of sarcopenia in PD patients and the relationship between clinical features and sarcopenia in PD patients are not clear. The aim of the study was to (1) assess the prevalence of sarcopenia in PD patients and (2) reveal the clinical features between PD patients with and without sarcopenia.Method: A systematic review was carried out through screening PubMed, EMBASE, and Cochrane database in May 2020. All study designs (case-control, cohort, and cross-sectional studies) were eligible for meta-analysis. Data of patients' characteristics, sarcopenia criteria, sarcopenia prevalence, and sarcopenia measures were retrieved. The primary outcome was estimated prevalence of sarcopenia by a pooled prevalence (%) and its 95% confidence interval (CI), using a random-effects model. The secondary outcome was the differences in clinical features between PD patients with and without sarcopenia by meta-analysis. Included articles were assessed for risk of bias. Potential sources of variation were investigated by using subgroup analyses and meta-regression.Result: Ten studies were included in the review. Among them, nine were cross-sectional studies, and one was a prospective cohort study. Age of participants with PD in the studies ranged from 51.1 to 80.7 years. The estimated prevalence of sarcopenia ranged from 6 to 55.5%. The random-effects pooled prevalence was 29% (95% CIs: 0.18-0.40). When only studies at low risk of bias were considered, pooled prevalence decreased to 17% (95% CIs: 0.02-0.33), with still high heterogeneity. The incidence of falls in PD patients with sarcopenia was higher than that in PD patients without sarcopenia. There was no difference in sex ratio between PD patients with and without sarcopenia.Conclusion: Sarcopenia seems to be common in patients with PD. Early assessment of sarcopenia should be implemented in PD to avoid fall and disability.
# Introduction
Parkinson's disease (PD) is a common neurodegenerative disorder that becomes increasingly prevalent with aging and results in dependency over time, despite the best treatment approaches. Sarcopenia was recognized as a muscle disease with low muscle mass and muscle function by WHO with a specific International Classification of Disease, Tenth Revision (ICD-10 code, M62.84) in 2016. Sarcopenia is commonly seen in elderly individuals with chronic diseases, including PD, which is an important determinant of quality of life (QoL), disability, and mortality in the elderly population.
Several sarcopenia definitions have been developed by different working groups or societies since the first term of sarcopenia was defined in 1988[the European Working Group on Sarcopenia in Older People (EWGSOP), the International Working Group on Sarcopenia (IWGS), the Society on Sarcopenia, Cachexia and Wasting Disorders (SCWD), the Foundation for the National Institutes of Health Biomarkers Consortium Sarcopenia Project (FNIH), , Newman, the decreased appendicular skeletal muscle mass index (ASMMI), SARC-F (a tool for screening sarcopenia risk), probable sarcopenia, early stage sarcopenia (ESS), and the Asia Working Group for Sarcopenia (AWGS)]. Among which, the most used definition was developed by EWGSOP. However, the European screening algorithm of sarcopenia developed by EWGSOP has been updated to the 2nd version in 2019since the 1st version developed in 2010. Until now, no worldwide consensus has yet been reached.
Sarcopenia can be assessed by the EWGSOP algorithm including the combination of a low muscle mass and a low handgrip strength (HS) or a low gait speed (GS). Muscle mass is measured by using traditional anthropometric measuresbioelectrical impedance analysis (BIA), whereas several studies used more precise methods, such as dualenergy X-ray absorptiometry (DEXA)and magnetic resonance imaging (MRI). The prevalence of sarcopenia in community-dwelling older adults was reported in the wide range of 1-50%. The prevalence of sarcopenia in PD was higher than that of the healthy older control group matched for age and sex. However, the prevalence of sarcopenia in PD is varied among different studies. For example, some studies reported the prevalence of sarcopenia in PD with a range of 6-31.4%according to the 1st version EWGSOP; other studies found the prevalence of sarcopenia as 40.4, 17.2, 58, 47.2, and 38% according to the decreased ASMMI, AWGS (21), SARC-F (15), probable sarcopenia, and ESS, respectively. This large discrepancy among the studies may be due to diagnostic criteria, muscle mass measurement techniques, different cut-off values for muscle mass indexes for the definition of sarcopenia, and as well as the characteristic of enrolled PD patients.
Some studies reported that sarcopenia was related to disease durationHoehn and Yahr stage (HY), Unified PD Rating Scale (UPDRS)-I (15), UPDRS-II (15), depression, and recognition. However, some other studies believed that sarcopenia had no correlations with disease duration, HY, UPDRS-I (14), UPDRS-II, depression, and cognition. In addition, some studies found that falls incidenceand female proportionin PD with sarcopenia were higher than those in PD without sarcopenia. However, some studies reported that there were no differences in falls incidenceand female proportionbetween PD patients with and without sarcopenia. Therefore, the main objective of the current study was acted in accordance with PICOS through systematical review and meta-analysis method: participants (P): patients with a diagnosis of PD from medical institutions or population; intervention (I): PD patients with sarcopenia; control (C): PD patients without sarcopenia; outcome (O): the prevalence of sarcopenia in PD patients, and the differences in clinical features between PD patients with and without sarcopenia; and study design (S): cohort study, case-control study, or cross-sectional study. ). The search was limited to English-language publications. The inclusion criteria were as follows: (a) type of population: enrolled subjects with a diagnosis of PD, (b) definition of sarcopenia: measured sarcopenia assessed using a formal operationalized measure, and (c) type of study: prospective cohort study, case-control study, or cross-sectional study. The exclusion criteria were as follows: (a) type of population: populations or patients with parkinsonian symptoms, but without formal PD diagnoses, (b) definition of sarcopenia: studies that defined sarcopenia according to a non-objective or non-standardized manner, and (c) type of study: reviews, editorials, case studies, and conference abstracts.
# Methods
## Search strategy and selection criteria
Two reviewers (Feng, Cai) evaluated each abstract for inclusion according to these criteria, and of selected abstracts, full publications were then obtained and reviewed in detail by the same two independent reviewers. Any differences were resolved by discussion and referred to a third researcher (Jiang) when required.
## Data extraction and quality assessment
After inclusion, the following information was extracted from each study to a data extraction form: study date, sample size, demographic characteristics of subjects, number of people in the sample with PD, the sarcopenia measure used, and findings regarding sarcopenia. Study authors were contacted when required to provide further information.
## Risk of bias assessment
Articles included in the study were assessed for risk of bias using two domains of the Quality in Prognosis Studies toolthat are relevant to observational studies (1. study participation and 2. outcome measurement). Appraisal of each domain provided a subjective assessment of risk of bias (ranked as low, moderate, or high). A summary of the areas considered in the assessment of each domain was included in.
## Data synthesis
Data analysis was performed using the Stata version 15.0 and the Review Manager 5.3 software. Heterogeneity between estimates was assessed using the I 2 statistic. An I 2 value above 75% indicated high heterogeneity. Meta-analysis was undertaken using a random-effects model (to account for heterogeneity). A pooled prevalence figure was calculated with 95% confidence interval (CI).
Potential influences on the prevalence estimates of sarcopenia were investigated using subgroup analyses and meta-regression. We then assessed the influence on the prevalence estimates of sarcopenia by the following variables that were identified a priori as potential sources on the prevalence estimates of sarcopenia: (1) sarcopenia criteria, (2) geographical location, (3) disease duration, (4) risk of bias, and (5) cognition. We classified studies as being either at low risk of bias (low risk of both participants and outcome measurement bias) or at moderate-tohigh risk of bias (moderate or high risk of either participants or outcome measurement bias). We compared the European studies with the rest of the studies and the Asian studies with the rest of the studies. A comparison between studies using EWGSOP and Non-EWGSOP was performed since sarcopenia criteria were varied among different studies. Furthermore, a comparison between the studies with PD patients with average disease duration of more than 5 years and the rest of the studies and studies with PD patients with average disease duration of <5 years and the rest of the studies was conducted since disease durations of PD patients were varied among different studies. A comparison between the studies with PD patients with Mini-Mental State Examination (MMSE) score more than or equal to 24 scores and the rest of the studies was conducted since cognition levels of participants were varied among studies. We ran five meta-regression models including these covariates (including sarcopenia criteria, geographical location, risk bias, disease duration, and cognition) separately using Stata version 15.0. If a sufficient number of studies were identified, evaluation for publication bias will be performed using Begg's funnel plot.
# Results
## Study selection and characteristics
A total of 459 potentially eligible articles were identified using our search strategy (250 articles from PubMed and 209 articles from EMBASE and Cochrane database). After the exclusion of 209 duplicated articles, 250 articles underwent title and abstract review. Two hundred nineteen articles were excluded at this stage since they were case reports, editorials, review articles, expert consensus, or not relevant studies. After the full-length article reviewing, 21 from 31 articles were excluded as they did not report sarcopenia assessment. Finally, nine cross-sectional studies and one prospective cohort study were included in the final analysis with a total of 2,397 participants. The literature review process is shown in. The characteristics and quality appraisal of the included studies are presented in.
The description of sarcopenia according to different criteria of included studies is presented in.
## Risk of bias
A summary of the risk of bias of the included articles is provided in. Four studies (40%) were considered to be at low risk of bias for both study participants and outcome measurement, and one study (10%)was considered to be at high risk of bias for both domains. Two studieswere considered to be at moderate risk for outcome measurement bias. Seven studieswere considered to be at low risk for outcome measurement bias that used clearly defined diagnostic criteria, reliable and validated instruments, and a similar method and setting of outcome measurement for all participants.
## Population
One prospective cohort study sampled patients (255 PD patients) from the general population (17), and six cross-sectional studies sampled patients (913 PD patients) from the clinical cohort (hospitals, outpatients, or nursing facilities), whereas three cross-sectional studies did not declare the type of sampled patients (288 PD patients)in.
## Geographical variation
In Europe, the pooled prevalence of sarcopenia in PD patients was 0.19 (0.04-0.34). In Asia, the pooled prevalence of sarcopenia in PD patients was 0.20 (0.19-0.36).
## Prevalence of sarcopenia in pd
Ten studies that provided the prevalence estimates of sarcopenia were included in the meta-analysis. Participants were not recruited from the same sampling frame, which could not lead to an overlap of study populations.
The overall random-effects pooled prevalence of sarcopenia was 0.29 (95% CI: 0.18-0.40) with a high level of heterogeneity (I 2 = 96.6%). When only studies at low risk of bias (on both domains of the Quality in Prognosis Studies tool) were selected, the pooled prevalence decreased to 0.17 (95% CI: 0.02-0.33), with still high heterogeneity (I 2 = 96.7%). Similar results were obtained from all sensitivity analyses.
The results of five meta-regression analyses of pooled estimates of subgroups based on sarcopenia criteria, geographical location, disease duration, risk of bias, and cognition are included in. There was little evidence of an effect of location (P = 0.934), disease duration (P = 0.635), and cognition (P = 0.326) on the prevalence of sarcopenia. However, there was an apparent higher effect of sarcopenia criteria (EWGSOP 19 vs. Non-EWGSOP 40%, P = 0.036) and risk of bias (low risk 17 vs. moderate/high risk 37%, P = 0.049) on the prevalence of sarcopenia.
## Comparison of falls incidence and sex between pd patients with and without sarcopenia
The incidence of falls in PD patients with sarcopenia was higher than that in PD patients without sarcopenia (P = 0.007, I 2 = 0). There was no sex ratio difference between PD patients with sarcopenia and without sarcopenia (female P = 0.48, I 2 = 86%; male P = 0.54, I 2 = 86%).
## Evaluation for publication bias
Evaluation for publication bias was used by Begg's funnel plot. No significant publication bias was found (P = 0.371).
# Discussion
To my best knowledge, this is the first systematic review of sarcopenia in PD. Despite its clinical importance, sarcopenia in PD patients has not been much explored in clinical practice. Only 10 papers in accordance with the inclusion and exclusion criteria were enrolled in the meta-analysis. Our study mainly found that the pooled prevalence of sarcopenia was 29% in PD, which was higher than the healthy older control group. There are several probable mechanisms to explain the high coexistence between the two conditions.
Firstly, sarcopenia and PD may share common neuroinflammation pathways. Elevated levels of circulating inflammatory mediators were detected in the early stages of both PD patients and patients with sarcopenia. Interleukin-6 has been reported to be associated with muscle loss and poor physical performance in patients with PD. Secondly, the changes in brain structure and network were considered to play a critical role in the pathophysiology of PD patients who have sarcopenia. Wu et al. found that the gray matter volume reductions in specific regions, such as the uncus and superior temporal gyrus, were significantly associated with an increased fat percentage of the thigh and the decreased superior temporal gyrus volume was associated with an increased fat percentage of the thigh in PD patients. An increased fat percentage of the thigh means fatty infiltration and represents core muscle loss in PD patients. Decreased volume of the default mode network causes the insufficient activity of the task-related network when performing a task, consequently resulting in poor motor function. Fractional anisotropy values were also decreased in the regions of the right parahippocampal gyrus and left occipital and right temporal white matter (WM) in PD patients with sarcopenia compared with PD patients without sarcopenia by using the Diffusion Tensor Imaging technique. Fractional anisotropy values reduction in the left cingulum and right anterior thalamic radiation in PD patients with sarcopenia exhibited the strongest correlations with decreased muscle mass, which represented WM alterations in the executive functional network in PD patients with sarcopenia. Moreover, decreased ASMMI was associated with reduced fractional anisotropy in the fronto-striato-thalamic circuits in PD patients with sarcopenia. Thirdly, besides affecting the central nervous system, the decrease in the numbers of motoneurons, i.e., mild motor neuron degeneration, might be another mechanism for neurogenic sarcopenia in PD since a low number of motor units were observed only in PD patients compared with controls. Neurogenic sarcopenia was considered as a subgroup of sarcopenia with a reduced number of motor units, which may indicate that sarcopenia and PD may have overlapping pathophysiological mechanisms for decreased numbers of motor neuron (4). Fourthly, sarcopenia may be influenced by the hormonal alterations in PD. Androgen plays an important role in the maintenance of muscle mass. Low plasma testosterone levels can cause or accelerate muscle-and age-related diseases, such as sarcopenia. Nevertheless, no study has been conducted to explore the relationship between testosterone and sarcopenia in PD. Future study is needed to elucidate it. Moreover, gastrointestinal infections, such as Helicobacter pylori and small intestinal bacterial overgrowth in PD, have potentially relevant effects on body weight and neurogastrointestinal hormones. However, leptin, ghrelin and GLP-1 may not play a major role in altered body composition in PD.
Lastly, several PD clinical features may affect the body composition and physical performance of people with PD. PD patients have lower levels of physical activity (in terms of amount and intensity) than healthy older adults. Malnutrition affects up to 24% of PD patients. Anorexia, nausea, constipation or delayed digestion, depression, and some pharmacologic treatment concur to reduce energy intake. However, the current meta-analysis cannot perform subgroup analysis to exclude the effect of anti-parkinsonism medication.
Fall was common in PD and associated with disease duration, freezing of gait, postural instability, non-motor symptoms, high levodopa equivalent daily dosage (LEDD), and greater number of medications. The frequency of fall significantly worsens the outcome of PD. Several studies have shown that reduced mobility, poor balance, and reduced leg muscle strength increased fall risk . These signs are clinical manifestations of sarcopenia. Our current meta-analysis also found that the frequency of falls in PD patients with sarcopenia was higher than that in PD patients without sarcopenia. Furthermore, PD patients with probable sarcopenia and falls have been reported to have higher HY staging and lower Schwab and England Activities of Daily Living scores. Therefore, sarcopenia and fall in PD may interact with each other and establish a vicious circle.
Some studies on corticospinal activity have been performed in PD patients and healthy controls during gait. One study found corticospinal control of human locomotion as a new determinant of age-related sarcopenia through comparison of corticomuscular coherence (CMC) between sarcopenic and nonsarcopenic older adults during gait, which may hint at a novel possible mechanism derived from corticospinal control of locomotion in age-related sarcopenia. Another study found that older healthy controls and PD participants had significantly decreased CMC and electromyography (EMG) power at low beta frequenciesHz) compared with young healthy controls, whereas there was no difference between the older healthy and PD groups. Additionally, one study found that PD participants had significantly decreased beta frequenciesfor tibialis anterior muscle compared with older healthy controls, whereas there was no difference in the magnitude of CMC between older and younger healthy controls. This might hint toward one of the potential features underpinning gait speed changes and risk of falling in PD patients with sarcopenia. However, the limitation of the latter two studies did not screen sarcopenia in PD patients. In the future, more studies should be worked on it.
Two studies found no sex difference in PD patients with or without sarcopenia. Wu et al. revealed that female sex was associated with core muscle loss in PD patients. One study found that higher fat infiltrations were detected in the psoas and thigh muscles in female PD patients compared with female healthy controls. Another study found that female sex was independently associated with sarcopenia. The current meta-analysis did not find a difference in both male and female sex ratios between PD with and without sarcopenia. However, we have to recognize the small sample size of these enrolled studies. Thus, it is necessary to conduct a study with a large sample size of PD patients and sarcopenia to address the role of sex.
Among enrolled studies, several studies found no difference in SMMI between PD patients and healthy controlsand among different stages of PD; PD patients with sarcopenia had higher UPDRS-III scores, lower fat-free mass index (FFMI), lower ASMMI, and lower GS (15) than PD patients without sarcopenia. Moreover, few studies reported that sarcopenia was associated with osteoporosis (16), tremor dominant/non-tremor dominant, nursing home placement, and Parkinson's Disease Questionnaire (PDQ)in PD patients. Some studies found that sarcopenia in PD had no relationship with UPDRS-IV (15), orthostatic hypotension (15), hallucinations, LEDD, REM sleep behavior disorder (RBD), hyposmia, motor physical therapy, and social support. Therefore, due to the small number of studies, insufficient information of PD patients, and different evaluation standards and statistical methods, some factors, such as GS, HY, depression, LEDD, UPDRS-I, and UPDRS-II, could not be made meta-analysis in this review.
# Limitation
There were some limitations of the current study. First, only a small number of papers were enrolled, which also indicated the scarcity of research on this area. Second, most of the studies were cross-sectional study, and the cross-sectional design was not truly appropriate for addressing a cause-effect relationship between sarcopenia and disease-related factors. Prospective studies are required to address this issue. Third, the participants of nine studies were from the hospital and similar medical institutions. Therefore, the prevalence rates might not be generalizable to the overall PD population, and we could not conclude the definitive prevalence of sarcopenia in patients with PD. Fourth, several studies excluded patients with poor cognitive function and using a wheelchair. As a result, the prevalence of sarcopenia might have been underestimated in these studies. Fifth, as the effect of no worldwide consensus sarcopenia criteria, geographical location, risk bias, and population, high heterogeneity was found among the included studies. Sixth, language selection (only articles in the English language were selected) was also the limitation of this meta-analysis.
# Conclusion
Sarcopenia seems to be common in patients with PD. However, the current evidence is not enough to conclude the definitive prevalence of sarcopenia in patients with PD. The progressive loss of function associated with sarcopenia may eventually boost the neurodegenerative process in PD. It is necessary to optimize the assessment of sarcopenia and still needed to improve the sarcopenia screening procedures in PD patients. The assessment, prevention, and, hopefully in the future, treatment of sarcopenia may help to improve the prognosis of patients with PD. |
Injection Site Reaction to Extended-Release Buprenorphine (Sublocade®) for Opioid Use Disorder Fourteen Days after Administration
# Introduction
In recent years, treatment options for Opioid Use Disorder (OUD) have grown to encompass a host of medications with various treatment formulations. Buprenorphine, a mainstay of treatment of OUD since approval from the U.S. Food and Drug Administration (FDA) in 2002, has benefited from a variety of treatment formulations including sublingual, buccal formulations and long-acting subcutaneous injection. [bib_ref] Buprenorphine: A review, Wakhlu [/bib_ref] In recent years, buprenorphine extended-release injections have grown popular, providing patients and clinicians with another tool to treat the debilitating symptoms of addiction.
Buprenorphine, while generally well tolerated in its older formulations, has had less research in its newer extended-release form. [bib_ref] A (relatively) new treatment for opioid dependence, Welsh [/bib_ref] The novelty of the extended-release formulation of buprenorphine explains the dearth of research on adverse events specifically related to its subcutaneous route. In this case study, a particular case of post-injection cellulitis is discussed and briefly potential causes explored. As extended-release formulations of buprenorphine continue to increase in use, cases studies like these will help identify more uncommon adverse events and help broaden practitioners' differentials.
## Case report
A 35-year-old female with severe opioid use disorder on buprenorphine injection (Sublocade®) therapy presented to the addiction psychiatry treatment center with reported injection site pain, erythema, swelling, and a rash of two days duration. Fourteen days prior to presentation, she had received her second buprenorphine injection in the right lower quadrant of her abdomen. She initially did not develop redness or pain, but while reporting withdrawal symptoms eleven days post-injection, she mentioned unintentionally scratching the site with an acrylic nail the previous day. She denied bleeding but experienced slight pain when this occurred. Of note, she had tolerated her initial injection well with only slight soreness and redness lasting two days. Between her first and second subcutaneous doses, supplemental sublingual buprenorphine/naloxone 8-2 mg daily had been prescribed to ease irritability and cravings.
On evaluation, she emphasized the scratch was unintentional. The pain was significant enough that she needed to wear loose clothing over the area. She denied fevers, chills, or expression of material from the site. Vitals were stable with a temperature of 36.4 degrees Celsius, blood pressure of 136/93 mmHg, heart rate of 85 beats/minute, respiratory rate of 17 breaths/minute, and blood oxygen saturation of 100%. Physical exam demonstrated a 3 x 5 cm indurated injection site with slight protuberance from her abdomen. Erythema extended into the surrounding skin involving a total area of 10 x 15 cm, including her striae gravidarum [fig_ref] Figure 1: Injection site cellulitis at the time of presentation [/fig_ref]. The site was tender and warm to the touch, but no disruptions in the skin barrier were noted. The induration and surrounding erythema were outlined with a marking pen, and due to concern for cellulitis, the patient was escorted directly to the emergency department. After a seven-day course of cephalexin and trimethoprim-sulfamethoxazole, the cellulitis successfully resolved [fig_ref] Figure 2: Resolved cellulitis seven days post-treatment [/fig_ref]. This patient's history was significant for buprenorphine/naloxone diversion, which served as the indication for her treatment to be converted to the injectable formulation. While accepting her history of diversion and its implications on her treatment options, she initially desired to discontinue the injections and resume sublingual treatment on follow-up. She was given the option of converting to methadone treatment, but she ultimately chose to continue the injections due to the significant functional benefit that this treatment modality awarded her. Although guidelines recommended subcutaneous buprenorphine to be dosed at 300 mg for the first two injections and 100 mg thereafter,this patient likely will continue to receive doses of 300 mg due to continued cravings and withdrawal symptoms.
# Discussion
Opioid use disorder is a public health emergency that has become a serious epidemic in the United States, specifically the high risk of overdose with opioids such as heroin, prescription opioids, and illicit fentanyl. [bib_ref] SAMHSA publishes TIP 63, focusing on medications for OUD treatment, Knopf [/bib_ref] Moderate to severe forms of OUD requires continuing sustained care for effective treatment and achieving long-term opioid abstinence. There are three medications currently approved by FDA used for OUD treatment that targets opioid receptors: methadone, buprenorphine, and long-acting naltrexone. These medications have been shown to reduce relapse to opioid use, overdose, and HIV and HCV transmission, as well as improve other health outcomes. [bib_ref] SAMHSA publishes TIP 63, focusing on medications for OUD treatment, Knopf [/bib_ref] [bib_ref] Long-acting buprenorphine injectables: Opportunity to improve opioid use disorder treatment among rural..., Lofwall [/bib_ref] [bib_ref] Trends in deaths involving heroin and synthetic opioids excluding methadone, and law..., O'donnell [/bib_ref] [bib_ref] An overview of systematic reviews of the effectiveness of opiate maintenance therapies:..., Amato [/bib_ref] Buprenorphine is a partial µ-opioid agonist used for the maintenance treatment of opioid use disorder. [bib_ref] A (relatively) new treatment for opioid dependence, Welsh [/bib_ref] Buprenorphine has potential merits, including a lower overdose risk and "low ceiling effect" for respiratory depression, fewer pharmaceutical interactions, and less risk of QTc-prolongation when compared to methadone.Buprenorphine is available in three formulations: sublingual films and tabs and buprenorphine extended-release injection (Sublocade®). [bib_ref] SAMHSA publishes TIP 63, focusing on medications for OUD treatment, Knopf [/bib_ref] [bib_ref] Trends in deaths involving heroin and synthetic opioids excluding methadone, and law..., O'donnell [/bib_ref] [bib_ref] An overview of systematic reviews of the effectiveness of opiate maintenance therapies:..., Amato [/bib_ref] The extended-release injectable formulation was approved by FDA in November 2017. This formulation is a monthly subcutaneous injection administered only by a healthcare provider. It is available in two doses of 300 mg/1.5 ml and 100/0.5 ml. Per current guidelines, the first two doses recommended are 300 mg each with the following monthly maintenance dose of 100 mg each. [bib_ref] SAMHSA publishes TIP 63, focusing on medications for OUD treatment, Knopf [/bib_ref] [bib_ref] Trends in deaths involving heroin and synthetic opioids excluding methadone, and law..., O'donnell [/bib_ref] [bib_ref] An overview of systematic reviews of the effectiveness of opiate maintenance therapies:..., Amato [/bib_ref] The monthly formulation of extended-release buprenorphine produces a sufficient steady-state blood level and may be more favorable for persons who have difficulty adhering to daily sublingual forms. [bib_ref] SAMHSA publishes TIP 63, focusing on medications for OUD treatment, Knopf [/bib_ref] [bib_ref] Long-acting buprenorphine injectables: Opportunity to improve opioid use disorder treatment among rural..., Lofwall [/bib_ref] [bib_ref] Depot buprenorphine injection in the management of opioid use disorder: From development..., Ling [/bib_ref] The most common adverse drug reactions of extended-release buprenorphine injection are injection site pain, injection site pruritus, constipation, sweating, nausea, vomiting, headache, fatigue, and sedation (more than 5%). [bib_ref] SAMHSA publishes TIP 63, focusing on medications for OUD treatment, Knopf [/bib_ref] [bib_ref] Prolonged-release buprenorphine formulations: Perspectives for clinical practice, Chappuy [/bib_ref] The majority of injection-site adverse reactions are mild to moderate in severity and include pain, hives, and erythema. [bib_ref] SAMHSA publishes TIP 63, focusing on medications for OUD treatment, Knopf [/bib_ref] These side effects are usually temporary and appear most commonly with the first injection and decrease in frequency with subsequent injections. [bib_ref] SAMHSA publishes TIP 63, focusing on medications for OUD treatment, Knopf [/bib_ref] [bib_ref] Efficacy and safety of a monthly buprenorphine depot injection for opioid use..., Haight [/bib_ref] Our case report was unique as an injection site reaction appeared in a patient 14 days after her second injection of extended-release buprenorphine, following unintentional scratching of the injection site with an acrylic nail while in the shower. Looking specifically at the Phase 3 trial data, only 1 of the 201 patients in the initial trial experienced injection site cellulitis, and 2 others of a total of 412 participants have experienced injection site cellulitis in the ongoing open-label long-term safety study. [bib_ref] Efficacy and safety of a monthly buprenorphine depot injection for opioid use..., Haight [/bib_ref] Possible explanations that may have predisposed our patient to an injection site reaction may have included improper injection of the medication, as adverse reactions are more likely to occur with intradermal or intramuscular injection.Both of these considerations were difficult to reconcile given the delayed onset from the injection. Our patients' acrylic nails may have acted as a nidus that seeded an infection over time as well.
It is important to consider that the cellulitis could have resulted from the patient attempting to remove the Sublocade®dose in an attempt to divert depot buprenorphine. The manufacturer suggested providers continue to monitor injection sites for evidence of tampering or attempts to remove the depot.In a clinical setting, removal of the depot can be performed through surgical excision under anesthesia within 14 BUPRENORPHINE (SUBLOCADE ® ) FOR OPIOID USE DISORDER continued.
days of injection. Given that the depot formulation is meant to decrease the risk of diversion, 10 this finding would be of considerable interest. However, the patient denied intentional tampering and it cannot be confirmed if this patient actually attempted to remove the medication.
[fig] Figure 1: Injection site cellulitis at the time of presentation. [/fig]
[fig] Figure 2: Resolved cellulitis seven days post-treatment. This is an open-access article distributed under the terms of the Creative Commons Attribution Non-Commercial No Derivatives (by-nc-nd) License. (CC-BY-NC-ND 4.0: https://creativecommons. org/licenses/by-nc-nd/4.0/) [/fig]
|
Expression and Purification of Integral Membrane Fatty Acid Desaturases
Fatty acid desaturase enzymes perform dehydrogenation reactions leading to the insertion of double bonds in fatty acids, and are divided into soluble and integral membrane classes. Crystal structures of soluble desaturases are available; however, membrane desaturases have defied decades of efforts due largely to the difficulty of generating recombinant desaturase proteins for crystallographic analysis. Mortierella alpina is an oleaginous fungus which possesses eight membrane desaturases involved in the synthesis of saturated, monounsaturated and polyunsaturated fatty acids. Here, we describe the successful expression, purification and enzymatic assay of three M. alpina desaturases (FADS15, FADS12, and FADS9-I). Estimated yields of desaturases with purity .95% are approximately 3.5% (Ca. 4.6 mg/L of culture) for FADS15, 2.3% (Ca. 2.5 mg/L of culture) for FADS12 and 10.7% (Ca. 37.5 mg/L of culture) for FADS9-I. Successful expression of high amounts of recombinant proteins represents a critical step towards the structural elucidation of membrane fatty acid desaturases.
# Introduction
Lipids are first synthesized as saturated fatty acids, and double bonds are introduced post-synthetically by oxygen-dependent enzymes known as fatty acid desaturases, in a process that is initiated by abstraction of hydrogen from a methylene group. Fatty acid desaturases are divided into soluble and integral membrane classes, which may have evolved independently [bib_ref] Stearoyl-acyl-carrier-protein desaturase from higher plants is structurally unrelated to the animal and..., Shanklin [/bib_ref]. The acyl-ACP desaturases are soluble enzymes found in the plastids of higher plants, whereas the more widespread class of integral membrane acyl-CoA desaturases is found in endomembrane systems in prokaryotes and eukaryotes [bib_ref] Desaturation and Related Modifications of Fatty Acids1, Shanklin [/bib_ref]. Fatty acid desaturases in each class are closely related homologs based on their amino acid sequences, and yet perform highly regio-and stereo-selective reactions on long-chain fatty acids composed of essentially equivalent methylene chains that lack distinguishing landmarks close to the site of desaturation. As pointed out by Nobel Laureate Dr. Konrad Bloch, this region-and stereo-specific removal of hydrogen ''would seem to approach the limits of the discriminatory power of enzymes'' [bib_ref] Enzymatic synthesis of monounsaturated fatty acids, Bloch [/bib_ref].
The membrane class of desaturases consists of enzymes with c5, c6, c9, c12 or v3 regio-selectivity. Structure determination would significantly improve our understanding of the structure-function relationships of this diverse class of proteins; however, there has been little progress, despite decades of efforts, due largely to the difficulty of generating recombinant membrane desaturase proteins for crystallographic analysis. Mammalian cells possess c5, c6 and c9, but lack c12 and v3 desaturases [bib_ref] Polyunsaturated fatty acid metabolism in prostate cancer, Berquin [/bib_ref] [bib_ref] Dietary fatgene interactions in cancer, Chen [/bib_ref]. Mortierella alpina belongs to the subphylum of Mucoromycotina [bib_ref] A higher-level phylogenetic classification of the Fungi, Hibbett [/bib_ref]. It is an oleaginous fungus that can produce lipids up to 50% of its dry weight. We have recently characterized the M. alpina genome [bib_ref] Genome Characterization of the Oleaginous Fungus Mortierella alpina, Wang [/bib_ref] which encodes one c5, two c6, three c9, one c12 and one v3 desaturase [fig_ref] Figure 1: Expression of recombinant FADS [/fig_ref]. Therefore, M. alpina has all known regioselective groups of membrane desaturases.
Membrane desaturases have been expressed in various hosts, such as Escherichia coli, Saccharomyces cerevisiae, Aspergillus oryzae, Mortierella alpina and cell-free system, for biochemical characterizations [bib_ref] Isolation of a Delta5-fatty acid desaturase gene from Mortierella alpina, Michaelson [/bib_ref] [bib_ref] Identification of Delta5-desaturase from Mortierella alpina by heterologous expression in Bakers' yeast..., Knutzon [/bib_ref] [bib_ref] Gene cloning and functional analysis of a second delta 6-fatty acid desaturase..., Sakuradani [/bib_ref] [bib_ref] Identification of Delta12-fatty acid desaturase from arachidonic acid-producing mortierella fungus by heterologous..., Sakuradani [/bib_ref] [bib_ref] A third fatty acid delta9-desaturase from Mortierella alpina with a different substrate..., Mackenzie [/bib_ref] [bib_ref] A novel fungal omega3-desaturase with wide substrate specificity from arachidonic acid-producing Mortierella..., Sakuradani [/bib_ref] [bib_ref] Functional characterization of a delta 6-desaturase gene from the black seabream (Acanthopagrus..., Kim [/bib_ref] [bib_ref] Isolation of a novel C18-Delta9 polyunsaturated fatty acid specific elongase gene from..., Li [/bib_ref] [bib_ref] Wheat germ cell-free translation, purification, and assembly of a functional human stearoyl-CoA..., Goren [/bib_ref]. However, none of these expression systems could achieve sufficient amount of desaturase proteins for crystal structure analysis. In the present study, we expressed M. alpina c9, c12 and v3 desaturases (FADS9-I, FADS12 and FADS15) in the methylotrophic yeast Pichia pastoris, purified the recombinant proteins and determined their enzymatic activities. High yield of these proteins paves the way for structural characterization of membrane desaturases.
# Materials and methods
## Plasmids, strains and growth conditions
Plasmids were amplified in DH5a bacteria grown in Luria-Bertani (LB) medium containing 100 mg/mL of ampicillin at 37uC with 250 rpm shaking, and purified using Qiagen Maxi Kit (Qiagen, CA). The expression vector pPinka-HC was purchased from Invitrogen (Carlsbad, CA). The P. pastoris strains used for protein expression were PichiaPink strain 1 (ade2), strain 2 (ade2, pep4), strain 3 (ade2, prb1) and strain 4 (ade2, pep4, prb1), purchased from Invitrogen. P. pastoris strains were cultured in Yeast Extract Peptone Dextrose (YPD, PichiaPink media kit Cat# A11156, Invitrogen) at 28uC with 250 rpm shaking.
Mortierella alpina (#32222, American Type Culture Collection, Manassas, Virginia, USA) was cultured as described previously [bib_ref] Genome Characterization of the Oleaginous Fungus Mortierella alpina, Wang [/bib_ref].
## Expression vector construction
M. alpina RNA extraction was performed using Trizol Reagent (Invitrogen, CA) according to the manufacturer's instructions. Total RNA was reverse transcribed with SuperScriptH III First-Strand Synthesis SuperMix (Invitrogen) following the manufacturer's instructions. Using both C-and N-terminal sequences as primers [fig_ref] Table 1: Purification of M [/fig_ref] , desaturase coding sequences were PCR amplified as follows: denaturation at 95uC for 30 sec, annealing at 55uC for 45 sec and extension at 72uC for 1 min for 35 cycles. The amplified products were cloned into a modified pET19 vector (Novagen) derivative containing a PreScission protease cleavage site (GE Healthcare) between the multiple cloning site and Nterminal His tag [bib_ref] Protein engineering of the quaternary sulfiredoxin.peroxiredoxin enzyme.substrate complex reveals the molecular basis..., Jonsson [/bib_ref] to construct pET19b-FADS15, pET19b-FADS12 and pET19b-FADS9-I). The desaturase genes, including the His-Tag and PreScission protease cleavage site, were then PCR amplified using primers SF1 and SR1-SR3 [fig_ref] Table 1: Purification of M [/fig_ref]. The PCR conditions used were the same as the first step for cDNAs. The PCR fragments were then purified and inserted into pPinka-HC to generate the expression vectors pPinka-HC-FADS15, pPinka-HC-FADS12 and pPinka-HC-FADS9-I. The presence of the inserts in the plasmids was confirmed by restriction digestion analysis and sequencing. The strategy used for constructing desaturase expression vectors is shown in [fig_ref] Figure 2: Fractionation of recombinant desaturases [/fig_ref].
## Pichiapink transformation
Desaturase expression vectors and pPinka-HC (negative control vector) were linearized with restriction enzyme Spe I and transformed into P. pastoris strains (PichiaPink strain 1, 2, 3 and 4) using the MicroPulser Electroporator (Bio-Rad Laboratories, Hercules, CA) according to the User Manual of PichiaPink Expression System (Invitrogen). P. pastoris were incubated with YPDS media (YPD with 1 M sorbitol) in the Gene Pulser Cuvettes at 28uC for 2 hr without shaking, spread onto PAD (Pichia Adenine Dropout) agar selection plates, and then incubated at 28uC for 4 days until distinct colonies were formed. Eight white colonies for each transformation were picked and plasmid integration in the yeast genome was confirmed by PCR.
Isolated clones were individually inoculated into 10 mL of BMGY medium (Buffered Glycerol-complex Medium, 1% yeast extract; 2% peptone; 100 mM potassium phosphate, pH 6.0; 1.34% YNB-Yeast Nitrogen Base; 0.0004% biotin; 1% glycerol) in 50 mL conical tubes. The cells were grown for 48 hr at 28uC with vigorous shaking at 250 rpm. Then, the cultures were centrifuged at 1,500 g for 5 min at room temperature, the cell pellets were resuspended in 2 mL of BMMY medium (Buffered Methanolcomplex Medium, 1% yeast extract; 2% peptone; 100 mM potassium phosphate, pH 6.0; 1.34% YNB; 0.0004% biotin; 0.5% methanol) and cultured at 28uC with shaking at 250 rpm to induce the expression. After continuous cultivation for 72 hr with daily addition of 0.5% methanol, cells were harvested by centrifuging for 10 min at 1500 g. Supernatant was transferred to a separate tube and both the supernatant and cell pellet were stored at 280uC until ready for assay. Supernatants and cell pellets were analyzed for protein expression by SDS-PAGE Coomassie blue staining and Western blot.
## Optimized protein expression condition
Individual colonies of P. pastoris-FADS15, FADS12 and FADS9-I were inoculated into 10 mL of BMGY medium in 50 mL conical tubes and cultured for 48 hr at 28uC at shaking speed of 250 rpm. Then, 2.5 mL of culture were inoculated into 50 mL of BMGY medium in 250-mL volume shaker flasks and grown at 28uC for 24 hr at 250 rpm. The cells were collected by centrifugation at 1500 g for 10 min, and resuspended in 10 mL induction medium (BMMY medium with 0.5% methanol) in a 100-mL shaker flask. The induction of protein expression was performed for 96 hr at 28uC with 250 rpm agitation and daily addition of 0.5% methanol. Samples were collected at 0, 6, 24, 48, 72 and 96 hr for measuring cell density at OD 600 , wet cell weight and total protein concentration, and for Western blot analysis of desaturase expression levels.
## Protein analyses
The cell pellets and supernatants were collected by centrifuging 100 mL cell culture at 1500 g for 10 min. Cell pellets were resuspended in 100 mL lysis buffer (20 mM Tris.Cl pH7.9, 1 mM EDTA, 5% Glycerol) with an equal volume of 0.5 mm Glass Beads (Biospec products, Inc.), and vortexed for 10 min at 4uC. Cell lysates were mixed with 46SDS sample buffer and heated for 5 min at 95uC. About 5 ml sample was loaded onto Mini-Protein Precast Gels (4-15%, Bio-Rad Laboratories, Cat #456-1086), and ran for 40 min at 150 V. Then, the SDS-PAGE gels were used for Coomassie blue stain, Invision His-Tag in-gel stain (Invitrogen) or Western blot.
For Western blot analysis, protein gels were transferred onto a nitrocellulose transfer membrane (Schleicher & Schuell GmbH, Germany) by electroblotting (100 V, 2 hr) using Mini Trans-Blot electrophoretic transfer cell (Bio-Rad Laboratories). The membrane was blocked with 3% BSA in TBST (150 mM NaCl, 10 mM Tris-Cl pH 7.5, 0.05% Tween20), and probed with mouse Penta?His antibody (Invitrogen) followed by HRP-conjugated goat anti-mouse IgG (GE Healthcare). Blots were then incubated with enhanced chemiluminescence reagent (ECL, GE healthcare) and analyzed using Fluorchem E (Cell Biosciences, Inc.).
The total protein concentration was determined with Pierce BCA protein assay kit (Thermo Scientific). The quantification of target protein on Coomassie blue stained gel was performed using known concentrations of BSA as standard, and analyzed with the AlphaView SA software (Cell Biosciences, Inc.).
## Cell fractionation
All purification procedures were performed at 4uC. Cells harvested from 800 mL of culture were suspended in 800 mL of lysis buffer. After addition of 0.5 mm glass beads to the cell suspension, P. pastoris cells were disrupted by vortexing at 4uC for 10 min. Cell lysis efficiency was usually more than 95% evaluated using a light microscope. Intact cells and cell debris were removed from the membrane suspension by low speed centrifugation (500 g, 10 min at 4uC). Then various centrifugation speeds and time (1,000 g for 10 min; 10,000 g for 10 min; 10,000 g for 20 min; 20,000 g for 10 min; 20,000 g for 20 min) were used to determine the best centrifugation conditions for collecting the membrane fraction.
## Protein solubilization
Fractions containing recombinant desaturases were solubilized in buffer, containing 20 mM Tris.Cl, pH 7.9, 500 mM NaCl, 10% glycerol, 0.1 mM EDTA, and different concentrations (0.5%, 1%, 2%) of various detergents (Tween 20, Tween 80, Nonidet P-40, DDM, Fos-Choline 12, Fos-Choline 16) at 4uC for different times (0.5, 1, 1.5, 2 hr and overnight). The insoluble materials were removed by centrifugation at 25,000 g for 30 min at 4uC.
## Protein affinity purification
Optimized culture and protein solubilization conditions were used for the subsequent purification process. His Mag Sepharose TM Ni affinity beads (GE Healthcare) were washed with binding buffer (20 mM Tris.Cl, pH 7.9, 500 mM NaCl, 10% glycerol, 0.1 mM EDTA, 0.5% Fos-Choline 16, 5 or 20 mM imidazole) and added to the solubilized fractions after detergent incubation. The bead-protein sample mixtures were incubated for 45 min at 4uC with end-over-end mixing. After washing three times with binding buffer containing 5 mM or 20 mM imidazole, desaturase enzymes were eluted with elution buffer (20 mM Tris.Cl, pH 7.9, 500 mM NaCl, 10% glycerol, 0.1 mM EDTA, 0.5% Fos-Choline 16, 500 mM imidazole). The purified FADS15, FADS12 and FADS9-I proteins were stored at 280uC in aliquots.
The quantity and quality of these purified enzymes were analyzed by SDS-PAGE, mass spectrometry and desaturase activity assay. Protein purity was presented in percentage, dividing the amount of a given desaturase protein quantified on gel by the amount of loaded protein quantified by the BCA protein assay.
# Fatty acid analysis
Approximately 20 mg of P. pastoris cell pellets were collected and used for each lipid extraction with the method of Bligh and Dyer [bib_ref] A rapid method of total lipid extraction and purification, Bligh [/bib_ref] under acidified conditions with pentadecanoic acid and heneicosanoic acid added as internal standards. The solvent from the extract was removed under a stream of nitrogen. Lipids were saponified in 1 mL of freshly prepared 5% ethanolic potassium hydroxide at 60uC for 1 hr under an argon atmosphere. After cooling, 1 mL of water was added to the samples and nonsaponifiable lipids were extracted into 3 mL of hexane. The aqueous layer was acidified with 220 mL of 6 M hydrochloric acid and the fatty acids extracted into 3 mL of hexane. After removing the hexane in a stream of nitrogen, fatty acids were converted to methyl esters by first treating with 1 mL of 0.5 M methanolic sodium hydroxide at 100uC for 5 min under argon followed by 1 mL of 14% methanolic boron trifluoride at 100uC for 5 min under argon [bib_ref] Rapid preparation of fatty acids esters from lipids for gas chromatographic analysis, Metcalfe [/bib_ref]. After cooling, the sample was mixed with 2 mL of hexane followed by 4 mL of saturated aqueous sodium chloride. After separating the phases, aliquots of the hexane layers were diluted 24-fold with hexane and then analyzed by GC/MS. One mL was injected in the splitless mode onto a 30 m6250 mm DB-WAXETR column (Agilent Technologies, Santa Clara, Califor- nia) with 0.25 mm film thickness. The temperature program was as follows: 100uC for 2 min, ramp to 200uC at 16uC per min, hold for one min, ramp to 220uC at 4uC per min, hold one min, ramp to 260uC at 10uC per min, and hold for 11 min. Helium was the carrier gas at a constant flow of 1.5 mL/min. The mass spectrometer was operated in positive-ion electron impact mode with interface temperature 260uC, source temperature 200uC, and filament emission 250 mA. Spectra were acquired from m/z 50 to 450 with a scan time of 0.433 s. Lower-boiling fatty acid methyl esters were quantified using the pentadecanoic acid internal standard, whereas higher-boiling methyl esters were quantified using the heneicosanoic acid internal standard. In vivo Desaturase Activity Analysis
Individual colonies of P. pastoris-FADS15, FADS12 and FADS9-I were cultured as described in the Recombinant protein expression section. Protein expression was induced for 72 hr with 0.5% methanol. Cell pellets were collected by centrifugation and stored at 280uC for fatty acid analysis.
In vitro Desaturase Activity Analysis 20 mL of the purified protein was added to 200 mL of yeast EGY49 cell homogenate, prepared by breaking cells with 0.5 mm glass beads in lysis buffer (20 mM Tris-HCl pH7.9, 1 mM EDTA, 5% Glycerol). The enzyme reactions were performed at 28uC for 3 h with shaking (250 rpm), and the assay mixture (220 mL) were stored at 280uC for fatty acid analysis.
# Results
## Sequence of m. alpina fatty acid desaturases
Recently, we sequenced the genome and EST of M. alpina ATCC #32222. The cDNA and protein sequences of FADS9-I, FADS12 and FADS15 were compared to published sequences from other strains of M. alpina [fig_ref] Figure 3: Solubilization of recombinant desaturases [/fig_ref]. The FADS12 and FADS9-I genes from M. alpina ATCC#32222 are 99.9% and 98.4% identical, respectively, to the corresponding genes from M. alpina 1s-4. The FADS12 and FADS9-I proteins from M. alpina ATCC#32222 are 100% and 99.6% identical, respectively, to these proteins from M. alpina 1s-4. The high similarity of FADS12 and FADS9-I genes between two strains indicates that these genes are highly conserved in M. alpina. Interestingly, the FADS15 gene is much less conserved at both DNA (93.1% identity) and protein (97.9%) levels.
## Expression of fatty acid desaturases in the pichiapink system
After several unsuccessful attempts to express recombinant M. alpina desaturases in bacteria, we tried to express FADS15, FADS12 and FADS9-I in the Pichia pastoris PichiaPink expression system (Invitrogen). Our data showed that PichiaPink strain 2(ade2, pep4) supported the highest level of expression for FADS15, 12 and 9-I. Interestingly, all recombinant desaturase proteins remained on the cell membrane despite the presence of afactor secretion signal, whereas EGFP (enhance green fluorescent protein) was successfully secreted into culture medium .
To determine potential toxicity of recombinant proteins, we first examined cell growth density, weight and total protein synthesis of the PichiaPink pPinka-HC-FADS clones. The recombinant PichiaPink pPinka-HC-FADS cells had growth characteristics similar to the control [fig_ref] Figure 1: Expression of recombinant FADS [/fig_ref]. A time course experiment showed that desaturase expression was detectable after 24 hr induction with 0.5% methanol and remained high for at least 72 hr post-induction [fig_ref] Figure 1: Expression of recombinant FADS [/fig_ref]. There were no significant differences in protein expression when cells were induced at different temperatures (16uC, 22uC, 28uC) or with a different concentration of methanol (0.5%, 1%). Therefore, we used an optimized procedure as described in the Materials and Methods for the expression of recombinant desaturase. Under this condition, expression levels of recombinant desaturase proteins reached approximately 130 mg/L of culture for FADS15, 110 mg/L for FADS12 and 350 mg/L for FADS9-I [fig_ref] Figure 1: Expression of recombinant FADS [/fig_ref] and [fig_ref] Table 1: Purification of M [/fig_ref]. FADS15 and FADS12 recombinant proteins overlapped with endogenous background proteins in Coomassie blue staining gels. The InVision TM His-tag In-Gel staining, however, showed clearly the expression of FADS15 and FADS12 over control [fig_ref] Figure 1: Expression of recombinant FADS [/fig_ref].
## Solubilizaton and purification of recombinant desaturases
In order to solubilize and purify the recombinant desaturases from cell membrane for in vitro enzymatic activity, we first tested conditions to enrich the cell membrane containing recombinant FADS15, FADS12 and FADS9-I. Different centrifugation speeds and times were examined for the separation of the membrane fractions containing target proteins. Efficient recovery of each recombinant desaturase produced in P. pastoris was achieved by centrifuging the cell homogenates at 500 g for 10 min to remove cell debris, then at 10, 000 g for 10 min to collect membranes [fig_ref] Figure 2: Fractionation of recombinant desaturases [/fig_ref].
Solubilization of membrane proteins requires the presence of detergents. Therefore, we tested the conditions for solubilization of the recombinant FADS15, FADS12 and FADS9-I from enriched cell membrane fractions using a panel of detergents: Tween-20, Tween-80, NP-40, n-Dodecyl-b-D-maltoside (DDM), Fos-Choline 12 or Fos-Choline 16. After treatment with 1% (w/v) of Fos-Choline 12 or Fos-Choline 16, FADS9-I and FADS12 were totally solubilized, and approximately 50% and 80% of FADS15 was solubilized with Fos-Choline 12 and Fos-Choline 16, respectively [fig_ref] Figure 3: Solubilization of recombinant desaturases [/fig_ref]. Tween-20, Tween-80, NP-40 and DDM had little effect on extracting these desaturase enzymes from the membrane. In addition, we noticed that FADS9-I protein degradation occurred during protein solubilization. This phenomenon was visible for proteins solubilized by both Fos-Choline 12 and 16. Thus, we investigated detergent incubation time during solubilization to optimize for the least protein degradation. Our results showed that the solubilization of FADS9-I protein reached its maximum level after incubation with detergent for 1.5 hr. Degradation of desaturase protein increased after more than 3 hr of incubation [fig_ref] Figure 3: Solubilization of recombinant desaturases [/fig_ref]. To maximize the ratio of intact vs. degraded proteins, we used 1.5 hr as our standard detergent incubation time for protein solubilization. We also compared the effect of detergent concentrations on protein solubilization efficiency and found that 0.5%, 1% or 2% of Fos-Choline 16 had similar effects. Taken together, our results indicate that all three recombinant desaturase enzymes can be solubilized efficiently from the cell membrane with 0.5% Fos-Choline 16 for 1.5 hr at 4uC. Solubilized FADS15, FADS12 and FADS9-I were affinitypurified on His Mag Sepharose Ni beads (GE healthcare) with aims of high purity or high yield. High purity (.95%) was achieved after one step purification using the His Mag Sepharose Ni beads with high stringency wash before elution [fig_ref] Figure 3: Solubilization of recombinant desaturases [/fig_ref] , S7). High yield (2-fold higher than that in the high purity process) was achieved with low stringency wash. Yield and quantity of each desaturase enzyme are summarized in [fig_ref] Table 1: Purification of M [/fig_ref]. Our estimated yields of desaturases with purity .95% are approximately 3.5% (Ca. 4.6 mg/L of culture) for FADS15, 2.3% (Ca. 2.5 mg/L of culture) for FADS12 and 10.7% (Ca. 37.5 mg/L of culture) for FADS9-I.
## In vivo and in vitro analyses of recombinant desaturase activities
To determine the functional activity of the recombinant M. alpina desaturase in vivo, PichiaPink cells were cultured and induced to express desaturases. Fatty acid methyl esters (FAME) analysis of cell pellets showed that expression of recombinant desaturases in PichiaPink cells altered their fatty acid contents compared to the control (typical samples data are shown in [fig_ref] Table 2: Calculated conversion rate of M [/fig_ref]. [fig_ref] Table 2: Calculated conversion rate of M [/fig_ref] shows the percentage increase of C16:1 D9 , C18:1 D9 , C18:2 D9,12 and C18:2 D9,12,15 compared to the negative control. The C16:1 D9 and C18:1 D9 were increased 40% and 20%, respectively, in PichiaPink cells expressing FADS9-I, suggesting that FAD9-I can insert the first double bond into both C16:0 and C18:0 with a preference for C16:0 as substrate. The C18:2 D9,12 content was 27% higher in cells expressing FADS12, suggesting that FADS12 can desaturate C18:1 D9 at the c12-position to produce C18:2 D9,12 . There was a 5% increase in C18:3 D9,12,15 in cells expressing FADS15, suggesting that FADS15 can desaturate C18:2 D9,12 at the c15-position to produce C18:3 D9,12,15 . These results suggest that the recombinant desaturases, FADS9-I, FADS12 and FADS15, were active in P. pastoris.
Saccharomyces cerevisiae is an excellent experimental system to study fatty acid desaturation, as it provides a eukaryotic endoplasmic reticulum, cytochrome b5 and NADH and lacks polyunsaturated fatty acid [bib_ref] Characterization of the Brassica napus extraplastidial linoleate desaturase by expression in Saccharomyces..., Reed [/bib_ref]. We used yeast EGY49 cell homogenate for our in vitro assay of recombinant desaturase activity. Our results showed that purified recombinant FADS12 converted C18:1 D9 to C18:2 D9,12 in vitro, and C18:2 D9,12 level was increased 116% compared to the control [fig_ref] Table 2: Calculated conversion rate of M [/fig_ref]. Activities of purified FADS9-I and FADS15 were relatively low in vitro.
We noticed that the size of FADS9-I was smaller on SDS-PAGE than its predicted molecular weight. To determine whether a cleavage had occurred on the N-terminus which has a His-tag and Precision protease (PP) cleave site, the purified recombinant proteins were digested with PP (GE Healthcare) and analyzed by Coomassie staining of SDS-PAGE gel. The reduction in molecular weight supported that all three desaturases had the His-tag and Precision protease cleave site which were removed by PP digestion. Therefore, FADS9-I might have been cleaved on its C-terminus. Molecular weight of the FADS9-I protein was determined by mass spectrometry. Result suggests that the cytochrome b5 domain was cleaved off from FADS9-I during protein solubilization and purification [fig_ref] Figure 4: Precision protease [/fig_ref].
# Discussion
No structural information is available to explain the regio-and substrate-selectivity of the membrane class of fatty acid desa-turases. Our initial goal was to express high amounts of recombinant desaturase as secreted proteins in the methylotrophic yeast Pichia pastoris. The pre-pro a-factor from Saccharomyces cerevisiae, the most commonly used signal sequence for targeting protein secretion, was used in our recombinant desaturase expression. However, the expressed FADS proteins remained membrane-bound, whereas recombinant EGFP (enhance green fluorescent protein) proteins were efficiently secreted in our expression system. This membrane association of FADS proteins necessitates an effective solubilization procedure for protein purification.
There is no reference that we can find in the literature about detergents used for solubilization of integral membrane desaturases. Detergents, such as Tween-20, Tween-80, NP-40, and DDM have been commonly used for solubilization of membrane proteins, and their efficiency could be explained by their polar head group structure . Our results showed that none of them had any effect on solubilization of our recombinant FADS. Instead, we found that FADS9-I and FADS12, and FADS15 were efficiently solubilized in detergent Fos-Choline 12 and 16. It is possible that the structural similarity of Fos-Choline 12 and 16 to fatty acids may contribute to their ability to solubilize membrane fatty acid desaturase. The Fos-Choline detergents have also been successfully used in other membrane protein studies [bib_ref] The structure of phospholamban pentamer reveals a channel-like architecture in membranes, Oxenoid [/bib_ref].
In vivo analysis showed high desaturase activity of FADS9-I. In contrast, the purified FADS9-I had relatively low desaturase activity. Some fatty acid desaturases possess a cytochrome b5 domain [fig_ref] Figure 1: Expression of recombinant FADS [/fig_ref] used for electron transfer, whereas others do not have such domain and rely on external source of cytochrome b5. The molecular weight of purified FADS9-I was smaller than the predicted, and mass spectrometry data indicated a spontaneous removal of the cytochrome b5 domain [fig_ref] Figure 4: Precision protease [/fig_ref]. Therefore, the loss of internal cytochrome b5 may explain the low FADS9-I activity in vitro. FADS12, on the other hand, uses external source of cytochrome b5, and thus, retains high activity.
Sequence alignment showed that the FADS15 from M. alpina ATCC#32222 has eight amino acids differences (2.1%) compared to M. alpina 1s-4, which may affect substrate specificity. Typically, v3 desaturases convert C18:2 D9,12 to C18:3 D9,12,15 . Expression of the FADS15 from M. alpina ATCC#32222 resulted in a moderate increase in C18:3 D9,12,15 . Adding various amounts of C18:2 D9,12 fatty acid into Pichia culture or EGY49 yeast culture did not significantly increase the level of C18:2 D9,12 and thus the production of C18:3 D9,12,15 . Meanwhile, the amount of C18:0 and C18:1 D9 was decreased, suggesting that the FADS15 may use C18:0 and C18:1 D9 as substrates as well.
Cloning, expression, purification and functional characterization of FADS9-I, FADS12 and FADS15 from M. alpina ATCC#32222 represent a critical step towards the structural elucidation of membrane fatty acid desaturases. Mortierella alpina is an oleaginous fungus which can produce lipids accounting for up to 50% of its dry weight in the form of triacylglycerols. It is used commercially for the production of arachidonic acid. Understanding the regio-and substrate-selectivity of membrane class fatty acid desaturases may also be useful for the genetic engineering of strains producing higher levels and different constituents of dietary fat. [fig_ref] Figure 2: Fractionation of recombinant desaturases [/fig_ref] Diagram of the cloning strategy for desaturase expression vectors. FADS coding sequences were PCR amplified using primers listed in [fig_ref] Table 1: Purification of M [/fig_ref]. PCR fragment were digested with indicated restriction enzymes, column purified and inserted into the pET-19b(PP) vector linearized with corresponding restriction enzymes. The FADS coding sequence plus His tag and Precision protease recognition sequence were PCR amplified and inserted into the pPinkalpha-HC vector. TRP2: TRP2 gene, AmpR: ampicillin resistance gene, pUC ori: oriental promoter of pUC, PAOX1:59AOX1 promoter region, a-factor: a-mating factor secretion signal, CYC1 TT: CCY1 transcription termination region, PADE2 HC: high-copy ADE2 promoter region,
## Supporting information
[fig] Figure 1: Expression of recombinant FADS. (A) Growth curve of the recombinant P.pastoris measured by cell density, wet weight and total protein concentration. (B) Kinetics of recombinant protein induction. Desaturase expression was determined by Western blotting using anti-His tag antibody. The normalized level of highest expression was set at one arbitrary unit. Three independent experiments were performed and bars represent standard deviations. (C) Quantification of the recombinant desaturase proteins by Coomassie blue staining after SDS-PAGE. Known concentrations of BSA were used as quantification standard. (D) InVision TM His-tag In-Gel Stain of recombinant FADS proteins. The arrow head indicates the addition of methanol for induction of recombinant protein expression. The triangles indicate the expressed recombinant proteins. M: protein marker, Cont: negative control which was PichiaPink TM harboring pPinka-HC, 15: FADS15, 12: FADS12, 9-I: FADS9-I. doi:10.1371/journal.pone.0058139.g001 [/fig]
[fig] Figure 2: Fractionation of recombinant desaturases. (A) InVision TM His-tag In-Gel Staining after SDS-PAGE analysis of FADS9-I membrane (top panel) and supernatant fraction (bottom panel), using different speeds of centrifugation. The triangles indicate the recombinant FADS9-I. (B) Coomassie blue Staining and InVision TM His-tag In-Gel Staining of recombinant desaturases after fractionation. T: total protein after grinded by glass beads, D: debris after centrifugation at 500 g for 10 min, M: membrane fraction after centrifugation at 10,000 g for 10 min, S: supernatant after the centrifugation. The triangles indicate the recombinant desaturase proteins. doi:10.1371/journal.pone.0058139.g002 [/fig]
[fig] Figure 3: Solubilization of recombinant desaturases. (A) Membrane fractions were suspended in 1% concentrations of various detergents and incubated at 4uC for 2 hr. Proteins were visualized by InVision TM His-tag In-Gel Staining (upper panel) and Western blot (lower panel). T20: Tween-20, T80: Tween-80, N40: NP-40, DDM: n-Dodecyl-b-D-maltoside, F12: Fos-Choline 12, F16: Fos-Choline 16,S: supernatant, P: pellet. (B) Membrane fractions of recombinant FADS9-I were suspended in 1% Fos-Choline 16 and incubated at 4uC for various time (0, 0.5, 1.5, 3, 12 hr). Aliquots were analyzed by InVision TM His-tag In-Gel Staining. S: supernatant, P: pellet. (C) One-step purification using His Mag Sepharose Ni beads under the high yield (upper panel) and high stringency conditions (lower panel). Proteins were analyzed by SDS-PAGE and Coomassie blue staining. M: protein marker, S: supernatant, F: flow through, E: eluate. doi:10.1371/journal.pone.0058139.g003 [/fig]
[fig] Figure 4: Precision protease (PP) digestion of recombinant desaturases. (A) Recombinant FADS proteins were digested with PP and visualized by Coomassie blue staining. Triangles indicate the recombinant desaturases before and after PP digestion. Arrow indicates Precision protease proteins. M: protein marker. (B) Diagram of recombinant desaturase structures with indicated molecular weight and cleavage sites. Long arrow indicates the Precision protease digestion site. Short arrow indicates the cleavage site before the Cyto b5 domain. doi:10.1371/journal.pone.0058139.g004 [/fig]
[fig] Figure S1 Fatty: acid desaturase identified in M. alpina ATCC#32222. (A) Fatty acid synthesis pathway. Enzymes involved in this pathway are indicated in red. Desaturase studied in this paper are highlighted in yellow. ACC: acetyl-CoA carboxylase, ELOVL: fatty acid elongase, FASN: fatty acid synthase, ACSL: acyl-CoA synthetase, FADS9: fatty acid delta 9 desaturase, FADS12: fatty acid delta 12 desaturase, FADS15: fatty acid delta 15 desaturase, FADS6: fatty acid delta 6 desaturase, FADS5: fatty acid delta 5 desaturase. (B) Diagram of desaturase structures. FADS: fatty acid desaturase domain. Cyto b5: cytochrome b5 domain. (TIF) [/fig]
[fig] Figure: S3 Alignment of FADS9-I coding sequences from different strains of M. alpina. Differences in nucleotide or amino acid are highlighted. (PDF) S4 Alignment of FADS12 coding sequences from different strains of M. alpina. Differences in nucleotide or amino acid are highlighted. (PDF) S5 Alignment of FADS15 coding sequences from different strains of M. alpina. Differences in nucleotide or amino acid are highlighted. (PDF) S6 Membrane association of recombinant desaturase proteins. PichiaPink cells were cultured for 24 hr, and induced with 0.5% methanol for 48 hr. Cell pellet and culture medium were analyzed by Western blot using anti-His tag antibody. Recombinant desaturase proteins were present exclusively in the cells fraction whereas EGFP protein was secreted into the culture medium. M: protein marker, Cont: PichiaPink cell harboring pPinkalpha-HC vector, EGFP: PichiaPink cell harboring pPinkalpha-HC-EGFP vector. C: cells, M: medium. (TIF) S7 Quantification of the recombinant desaturase proteins after one-step purification. Known concentrations of BSA were used as quantification standard. Proteins were analyzed by SDS-PAGE and Coomassie blue staining. Protein purity was calculated by dividing the amount of a given desaturase protein quantified on gel by the amount of loaded protein quantified by the Pierce BCA protein assay. M: protein marker, S: supernatant, F: flow through, E: eluate. (TIF) S8 Structures of detergents used in the experiments. (TIF) [/fig]
[table] Table 1: Purification of M. alpina desaturases. [/table]
[table] Table 2: Calculated conversion rate of M. alpina FADS9-I, FADS12 and FADS15. [/table]
[table] Table S1: Primers for PCR reaction. (DOC) [/table]
[table] Table S2: FAME analysis of recombinant desaturaseexpressing PichiaPink cells. (DOC) Conceived and designed the experiments: HC YQC. Performed the experiments: HC ZG MW. Analyzed the data: HC ZG MW WTL YQC. Contributed reagents/materials/analysis tools: HZ WC. Wrote the paper: HC ZG YQC. [/table]
|
The disease burden attributable to 18 occupational risks in China: an analysis for the global burden of disease study 2017
Background: China has more than 18% of the global population and over 770 million workers. However, the burden of disease attributable to occupational risks is unavailable in China. We aimed to estimate the burden of disease attributable to occupational exposures at provincial levels from 1990 to 2017. Methods: We estimated the summary exposure values (SEVs), deaths and disability-adjusted life years (DALYs) attributable to occupational risk factors in China from 1990 to 2017, based on Global Burden of Disease Study (GBD) 2017. There were 18 occupational risks, 22 related causes, and 35 risk-outcome pairs included in this study. Meanwhile, we compared age-standardized death rates attributable to occupational risk factors in provinces of China by socio-demographic index (SDI). Results: The SEVs of most occupational risks increased from 1990 to 2017. There were 323,833 (95% UI 283,780 -369,061) deaths and 14, 060,210 (12,022,974 -16,125,763) DALYs attributable to total occupational risks in China, which were 27.9 and 22.1% of corresponding global levels, respectively. For attributable deaths, major risks came from occupational particulate matter, gases, and fumes (PGFs), and for the attributable DALYs, from occupational injuries. The attributable burden was higher in males than in females. Compared with high SDI provinces, low SDI provinces, especially Western China, had higher death rates attributable to total occupational risks, occupational PGFs, and occupational injuries. Conclusion: Occupational risks contribute to a huge disease burden in China. The attributable burden is higher in males, and in less developed provinces of Western China, reflecting differences in risk exposure, socioeconomic conditions, and type of jobs. Our study highlights the need for further research and focused policy interventions on the health of workers especially for less developed provinces in China to reduce occupational health losses effectively.
# Introduction
China is the most populous country in the world with a population of 1.37 billion [1]. Over the past few decades, China has emerged as a global leader in manufacturing with growing competitiveness and increasing impact on the global economy. However, rapid economic growth also brings to the workplace a variety of risks that threaten the health of workers. There are more than 770 million workers in China, and more than 200 million workers are exposed to occupational hazards . It has become a priority for China to meet the challenges in the monitoring of the health of workers and in the improvement of occupational health services.
Occupational risks, as part of environmental hazards, contribute to the development of many diseases and injuries. By evaluating burden attributable to occupational risks, accurate and comprehensive data can be offered to policymakers to effectively prevent related health losses. Although attempts have been made to estimate the burden of air pollution at the national level, only a few studies estimated the burden of occupational carcinogens and injuries and they were limited to several provinces of China. Additionally, occupational exposure exhibits spatial and temporal heterogeneity and is closely related to socioeconomic levels in different regions. Therefore, a comprehensive study on the spatiotemporal trend of the burden of disease attributable to occupational risks is urgently needed in China.
In this paper, we evaluated the disease burden levels attributable to 18 occupational risks and their geographical heterogeneity by socio-demographic index (SDI) in China from 1990 to 2017, as part of the Global Burden of Disease Study 2017 (GBD 2017). We aimed to find out the key problems in occupational health so as to provide useful information for occupational protection strategies and interventions in China.
# Methods
## Overview
The comparative risk assessment (CRA) approach was developed to estimate levels and trends of sex-specific, risk-specific, and cause-specific mortality and disease burden of behavioral, environmental, occupational, and metabolic risks from 1990 to 2017 for 195 countries and territories in GBD 2017. The detailed framework and data analysis methods have been provided previously . In the CRA framework, the attributable burden was calculated as the reduction in the current disease burden if the past population exposure shifted to the counterfactual level of risk exposure. By using a consistent approach, CRA allows rankings and comparisons among deaths and DALYs attributable to various risk factors, providing further data guidance for policymakers. SDI as a combined indicator was estimated based on fertility among women, years of education and income per person. The SDI of China in 2017 was estimated in our previous article. Here we focused on accessible data of occupational risks from GBD 2017 to estimate the disease burden attributable to occupational exposure in China.
## Risk factors and related causes
Risk-outcome pairs satisfying the World Cancer Research Fund (WCRF) grades of convincing or probable evidence with biologically plausible associations were included in GBD 2017. There are 18 occupational risks, 22 related causes and 35 risk-outcome pairs included in this study. The occupational risk factors hierarchy and related causes were shown in. There are six risk categories including carcinogens, asthmagens, PGFs, noise, injuries, and ergonomic factors for occupational risks. The occupational carcinogens include 13 agents classified as Group 1 carcinogens by the International Agency for Research on Cancer (IARC). The exposure definitions, International Classification of Diseases (ICD) codes of related cases, and.
## Estimation of exposure
Data for occupational risk factors were collected from all accessible resources. The data included raw data on Chinese economic activity proportions, occupation proportions, employment to population ratio estimates, and fatal injury rates from the International Labour Organization, survey data including China National Population Census, China Intercensal Population Sample Survey of One-Percent, China International Social Survey Programme. The Spatio-temporal Gaussian process regression (ST-GPR) approach was used to integrate multiple data inputs and generate year-specific, and location-specific estimates. For each occupational risk, the theoretical minimum risk exposure level (TMREL) was assumed to no given risk exposure or the lowest levels of risk exposure without established riskoutcome. Education, geological information and the socio-demographic level were included in models as covariates. The estimates differed for (1) occupational carcinogens, occupational noise, and occupational particulates, (2) occupational ergonomic factors and occupational asthmagens, and (3) occupational injuries using the following equations:
[formula] E [/formula]
Where E r,p,l,y,s,a is the prevalence of exposure for risk factor r in province p at level l in year y, sex s, and age group a. P e,p,y is the proportion of economically active population in province p, economic activity e, and year y. EAP p,y,s,a is economically active population in province p, year y, sex s and age group a. Exposure rate r,l,e is the rate of exposure to risk factor r at level l in economic activity e. E r,p,y,s,a is the prevalence of exposure for risk r in province p, year y, sex s, and age group a. P occ,p,y is the proportion of economically active population in occupation occ in province p, and year y. Occupational fatal injuries p,y,s,a is the occupational fatal injuries counts in province p, year y, sex s and age group a. Injury rate e,p,y,s is the injury rate of economic activity e in province p, year y and sex s. Population p,y,s,a is the population in year y, province p, sex s and age group a. All occupational risk exposures were estimated for ages 15 and older. The estimates were further divided by the sum of all the estimates to be rescaled to sum as 1 across different categories.
## Relative risks and the population attributable fraction
Information from the cohort, pooled cohort, and casecontrol studies was obtained to determine the relative risk for each risk-outcome pair by systematic reviews in GBD 2017. The risk factors were categorized based on the measurement of exposure: dichotomous, polytomous, and continuous. The relative risks for each exposure category were listed in Tables S4 and S5.
The population attributable fraction (PAF) is the proportion of outcomes or causes in the population that are attributable to the associated risk factor. It is estimated independently by relative risks and calculated as the proportion of the decreased outcome among a given population if the past exposure levels of risk were reduced to the counterfactual level of the TMREL in a given year.
The equation for calculating PAFs of occupational risks with the exception of injuries:
[formula] PAF [/formula]
Where PAF r,c,p,y,s,a is the population attributable fraction for cause c due to risk factor r in province p, year y, sex s and age group a. RR r,c,s,a is the relative risk as a function of exposure level x (ranged from lowest exposure level (l) to highest exposure level (u)) for risk r, cause c, sex s, and age group a. P r,p,y,s,a (x) is the distribution of exposure for risk r, in province p, year y, sex s, and age group a. TMREL r,s,a is the theoretical minimum risk exposure level for risk factor r, sex s, and age group a.
The
Where PAF p,y,a,s is the population attributable fraction in province p, year y, sex s and age group a. Occupational fatal injuries p,y,s,a is the occupational fatal injuries counts in province p, year y, sex s and age group a. Fatal injuries p,y,s,a is the total fatal injuries counts in province p, year y, sex s and age group a, which were obtained from causes of death in GBD 2017. And the PAFs of multiple risks are aggregated by a mediation adjustment to calculate the excess attenuated risk. The PAFs for different causes were shown in .
## Estimation of attributable burden
For the given exposure risk-outcome pair, the attributable deaths were estimated as total deaths for the outcome multiplied by the PAF for the risk-outcome pair. The other three metrics of burden including years of life lost (YLLs), years lived with disability (YLDs), and DALYs (the sum of YLLs and YLDs) were also assessed in a similar way. The attributable burden was estimated by location, age, sex, and year. The standard population from WHO was used to calculate age-standardized deaths and DALYs per capita for each country.
## Summary exposure values
Summary exposure values (SEV) is an exposure metric of the risk-weighted prevalence of each risk. SEV standardizes the prevalence by relative risks of related causes to offer a concise comparable summary of risk exposure for different locations and years. The range of SEV is from 0 to 100%, where 0% indicates no given risk exposure in a population, and 100% means the total population is exposed to the maximum possible level for a given risk.
# Results
There were SEVs for 17 occupational risks with the exception of occupational injuries in GBD 2017. The leading exposure category for SEVs was ergonomic factors, followed by asthmagens, noise, and PGFs. All-age SEVs of 15 in 17 occupational risks increased from 1990 to 2017 in China. Age-standardized SEVs for three risks increased by more than 20% from 1990 to 2017: occupational exposure to benzene, trichloroethylene, and chromium. Conversely, SEVs for occupational ergonomic factors and asthmagens decreased by more than 20%.
As shown in, there were 323,833 (95% UI 283780-369,061) deaths attributable to total occupational risks in 2017, China, which accounted for 27.9% of global attributable deaths. The deaths attributable to PGFs, carcinogens, injuries, and asthmagens accounted for 57.8, 21.1, 20.8, 0.2% of deaths attributable to total occupational risks in 2017, respectively. From 1990 to 2017, age-standardized death rate attributable to injuries, PGFs and asthmagens declined by more than 60%, but the rate attributable to carcinogens increased by 16.8%.
As shown in, there were 14,060,210 (12,022, 974 -16,125,763) DALYs attributable to total occupational risks in 2017, China, which accounted for 22.1% . For different genders, we have not observed a wide difference in SEV between males and females . Deaths attributable to total occupational risks in males were over 2.0 times than in females in 2017. As shown in, the leading risk was occupational PGFs for deaths in both sexes, but for DALYs, the leading risk was occupational PGFs in females and occupational injuries in males. DALYs attributable to total occupational risks were higher in males than in females.
Age-standardized death rates attributable to total occupational risks, carcinogens, asthmagens, PGFs, and injuries in the provinces of China grouped by SDI in 2017 are shown in . Compared with high SDI provinces, low SDI provinces had higher death rates attributable to total occupational risks, occupational PGFs, and occupational injuries. Western China, especially for Yunnan and Tibet, had the highest death rates attributable to total occupational risks, while Liaoning and Heilongjiang in Eastern China had the highest death rates attributable to occupational carcinogens. The highest death rates attributable to occupational PGFs and injuries mainly concentrated in Western China. Age-standardized DALY rates attributable to total occupational risks by the provinces of China in 2017 showed similar results that Western China had higher occupational age-standardized DALY rates than other regions in China.
# Discussion
To our knowledge, this study presents the first comprehensive assessment so far of burden attributable to occupational risks at the provincial level of China in 2017. China has a huge labor force. The most up-todate estimation of the disease burden attributable to occupational exposures provides important insights into health losses related to occupational risks in China. Meanwhile, the standardized assessment of attributable burden to different risk factors enables direct comparison and priority ranking. The SEV provides a concise and comparable summary of risk exposure for different locations and years. Rapid industrialization in China has been accompanied by increased occupational carcinogen exposures. Our study showed SEVs of most occupational carcinogens increased in the past 28 years, especially for asbestos, trichloroethylene. Over the past half-century, the production of asbestos in the world has been transferred to developing countries. China produces about 16% of global total output and is the second-biggest consumer of asbestos in 2016. Trichloroethylene use has been increased with China's growing telecommunications, electronic, and microelectronics industries since the early 1990s.
China has the largest occupational disease burden in the world. We estimated that 0.3 million deaths and 14.1 million DALYs were attributable to total occupational risks in 2017, China, which might be attributed to the huge number of labor force. Occupational injuries were the leading risk for attributable DALYs in China. A previous report showed occupational injuries and disabilities were common among Chinese workers. In comparison, the leading risk for occupational burden was ergonomic factors in developed countries. Socioeconomic inequality is one of the main reasons for this difference. The type and income of occupations, age, and education levels are important influence factors of occupational injuries. The economic transition from agriculture to manufacturing industry causes the migration of great rural labor forces into factories, and migrant workers have become an important part of the population who are subject to occupational hazards. For example, migrant workers account for 80% of total workers in the construction industry. With poor educational backgrounds, migrant workers are only able to locate jobs with low wages, high-risk operation, and long working hours. They also have a lower sense of self-protection and less experience. Almost all of the occupational injuries and deaths are preventable. Therefore, more vocational training and safety education targeting migrant workers are necessary to prevent health loss in China.
Although the burden attributable to all occupational risks is higher in China, it has declined more than 50% over the past three decades. The Chinese government has made numerous efforts to protect the health of workers in China, including revision of the occupational law, development of new technologies and monitoring of occupational diseases. It is further supported by coincidence of increased SEVs for occupational noises, PGFs, benzene and formaldehyde and reduced burden attributable to these risks. However, the burden attributable to occupational carcinogens, especially for asbestos, trichloroethylene, chromium, PAHs, and DEE still shows marked increases. Notably, among the 13 occupational carcinogens included in this study, only cancers caused by arsenic, asbestos, benzene, and chromium are classified into occupational cancer in the Categories and Catalogs of Occupational Diseases in China. Thus, our results provided basis for policymakers to reevaluate the potential of cancers caused by other nine carcinogens as national legal occupational disease. The western provinces in China have the highest death and DALY rates attributable to occupational risks. The spatial inequality of disease burden for occupational hazards is similar to previous studies on environmental and occupational burden of diseases in Iran, which Age-standardized death rates (per 100,000) attributable to total occupational risks, occupational carcinogens, occupational asthmagens, occupational PGFs, and occupational injuries in the provinces of China grouped by SDI, 2017. SDI: socio-demographic index; PGFs: particulate matter, gases, and fumes might be attributed to the regional inequality of socioeconomic level and medical services of occupational diseases. Institutions certificated to diagnose the occupational disease concentrate in big cities and developed areas, with few institutions in less developed areas, especially for Western China. The imbalance between supply and demand is challenging the medical services of occupational diseases for workers from the western region.
A number of limitations exist in our estimates. The most important limitation is the lack of accessible data on the prevalence of occupational exposure in China. Scarce data are available for the specific exposure levels of occupational risks, compared with air pollution and hazards in drinking water. The exposure levels are estimated by the proportion of the population in specific types of work where the exposures are expected. The exposure assessment for occupational carcinogens was based on carcinogen exposure (CAREX) database, which included 32 million workers exposed to almost all known and suspected carcinogens in the 15 countries of the European Union in the early 1990s. However, workers in developing countries are likely to be exposed to higher levels of occupational hazards than developed countries. Employing CAREX data may underestimate the health effects in developing countries. Meanwhile, China is a top producer of cement, coal, iron, steel with rich mineral reserves and resources. Rapid industrialization in China in recent years also brings more categories of jobs and occupational hazards. More efforts are required to more accurately determine the proportion of the working population exposed to occupational hazards in China.
Due to the stringent causal criteria, many risk-outcome pairs were hampered to be included in GBD estimates, which leads to underestimation of the disease burden attributable to occupational risks, in comparison with other published studies. For example, benzene can cause not only leukemia but also myelodysplastic syndromes, multiple myeloma and non-Hodgkin lymphoma. Apart from lung cancer, arsenic exposure can also cause urothelial cancer, skin cancer, and. Meanwhile, disabilities in GBD methods are associated with health conditions confined to the most common sequelae of occupational exposure but not mental disorders. Furthermore, there are joint effects attributable to co-exposure of multiple occupational hazards. For example, exposure to benzene interacts with work stress can reduce birth weight in petrochemical workers. As most of the occupational risk-outcome pairs cannot be accessed by randomized controlled trials, evidence from epidemiological studies and toxicological studies deserves consideration in future estimates.
Finally, there are many uncertainties of applying the default parameters from the developed country to China. The relative risk is a critical parameter for the PAF and SEV estimations. It is derived from prospective observational studies and case-control studies by systematic reviews. However, most of the relative risks come from studies conducted in high-income countries with lower occupational exposure, rather than low-and middle-income countries, which may cause biased estimates of PAF and SEV for a given population. Moreover, the estimation of the TMREL is also a key step in the calculation of the PAF. A minor change of the TMREL can lead to relatively large changes in the PAF. For all occupational risks, the TMREL is defined as no corresponding occupational exposure or background level to a given risk. However, the background level of occupational hazards including carcinogens, asthmagens, and noises is often difficult to specify. Additional epidemiological studies in regions with both high and low occupational exposure should be encouraged and supported to estimate relative risks and TMREL more accurately in China.
A human capital approach is suggested as an alternative method to provide accurate estimates of the impact of chemical exposure on population health. It is a health economic method to assign monetary costs associated with adverse outcomes, including but not limited to traditional physical health such as cognitive deficits. A hybrid approach of the existing GBD method and the human capital approach should be encouraged to obtain more comprehensive and accurate estimates of the burden attributable to occupational risks.
Although the current method in GBD could not fully estimate the impacts of occupational risks on population health, the results were estimated from all accessible resources to produce relatively valid estimates up till now. Moreover, we take socioeconomic impact into our estimates to analyze differences among regions with different levels of development. Further work to increase the quality and availability of data are expected to devote adequate evidence for more valid estimates in China.
In conclusion, China is facing a tremendous disease burden attributable to occupational risk factors. Although the burden attributable to total occupational risks decreased between 1990 and 2017, the burden attributable to occupational carcinogens is rising greatly. The attributable burden is higher in males, and in less developed provinces of Western China. Our estimates will benefit policymakers to focus on preventing and reducing the health losses of workers in China. |
Medication-overuse headache: a widely recognized entity amidst ongoing debate
Medication overuse in primary headache disorders is a worldwide phenomenon and has a role in the chronification of headache disorders. The burden of disease on individuals and societies is significant due to high costs and comorbidities. In the Third Edition of the International Classification of Headache Disorders, medication-overuse headache is recognized as a separate secondary entity next to mostly primary headache disorders, although many clinicians see the disease as a sole complication of primary headache disorders. In this review, we explore the historical background of medicationoveruse headache, its epidemiology, phenomenology, pathophysiology and treatment options. The review explores relevant unanswered questions and summarizes the current debates in medication-overuse headache.
# Background
Overuse of symptomatic medication is a common problem in patients with primary headache syndromes [bib_ref] Medication-overuse headache: risk factors, pathophysiology and management, Diener [/bib_ref] [bib_ref] Clinical features, pathophysiology, and treatment of medication-overuse headache, Evers [/bib_ref]. Headache syndromes such as migraine or tension-type headache cause painful experiences and significant disability in patients. The use of analgesics is therefore justifiable when correctly utilized. For more than 50 years, clinicians have recognized and reported on headache chronification occurring during a period of frequent use of analgesics. The underlying consensus for the entity of medication-overuse headache (MOH) consists of a deterioration of a pre-existing headache syndrome whilst overusing one or several types of acute painkilling treatments. MOH is widely accepted and recognised in the neurological and headache community nowadays, although the entity keeps raising important questions. Debates on the pathophysiological mechanisms, definitions of overuse and the nosology of MOH are ongoing. This review presents the current state of literature and knowledge on MOH. It provides an overview of the history, clinical features, epidemiology of MOH, an update on the current understanding of the underlying neurobiological mechanisms and treatment, before discussing the key topics in the controversies surrounding MOH.
## Moh in historical perspective
The first descriptions of MOH date back to 1930s, when multiple authors started to associate prolongation of migraine with ergotamine-overuse [bib_ref] Chronic migraine and medication-overuse headache through the ages, Boes [/bib_ref] [bib_ref] The use of ergotamine tartrate in migraine, Lennox [/bib_ref] [bib_ref] Gynergen abuse in cases of migraine. Acta Psychi Search results, Silfverskiöld [/bib_ref] [bib_ref] Treatment of migraine, Graham [/bib_ref] [bib_ref] Characteristic headache resulting from prolonged use of ergot derivatives, Lippman [/bib_ref] [bib_ref] Ergotamine tolerance in patients with migraine, Friedman [/bib_ref]. Chronic headache following overuse of ergotamine was clearly defined by Peters and Horton in 1951 [bib_ref] Headache: with special reference to the excessive use of ergotamine preparations and..., Peters [/bib_ref]. They reported on 52 migraine patients who developed daily headache after daily use of ergotamine and noted improvement after the drug was stopped. The same authors published their withdrawal protocol in 1963 [bib_ref] Clinical manifestations of excessive use of ergotamine preparations and Management of Withdrawal..., Horton [/bib_ref]. The first ergotamine withdrawal protocols were proposed independently by [bib_ref] Chronic migraine and medication-overuse headache through the ages, Boes [/bib_ref] [bib_ref] Treatment of migraine, Graham [/bib_ref] [bib_ref] Characteristic headache resulting from prolonged use of ergot derivatives, Lippman [/bib_ref] [bib_ref] Ergotamine tolerance in patients with migraine, Friedman [/bib_ref]. In the 1970s, multiple authors wrote on the association between overuse of mixed analgesics, including those based on ergotamine, barbiturates and codeine, and headache progression [bib_ref] Headache and dependence on mixed analgesic preparations, Wörz [/bib_ref] [bib_ref] Abuse and paradoxical effects of analgesic drug mixtures, Worz [/bib_ref]. In 1982, Mathew et al. outlined that overuse of analgesics contributed to the transformation of episodic migraine (EM) into daily headaches and a few years later the same group introduced the term "transformed or evolutive migraine" to describe the entity [bib_ref] Chronic migraine and medication-overuse headache through the ages, Boes [/bib_ref] [bib_ref] Transformation of episodic migraine into daily headache: analysis of factors, Mathew [/bib_ref] [bib_ref] Transformed or Evolutive migraine, Mathew [/bib_ref].
The first edition of the International Classification of Headache Disorders (ICHD) was published in 1988 which introduced the term "drug-induced headache". It also introduced and specified the entities "ergotamine-induced headache", "analgesics abuse headache" and "other substances". This was based on the experience with overuse of analgesics and ergots only. After the introduction of triptans, it became clear that this class of drugs could also induce headache deterioration if used excessively [bib_ref] Headache after frequent use of serotonin agonists zolmitriptan and naratriptan, Limmroth [/bib_ref] [bib_ref] Features of medication overuse headache following overuse of different acute headache drugs, Limmroth [/bib_ref]. In 1994 Silberstein et al. proposed criteria for "transformed migraine", since transformation of EM to daily or almost-daily head pain (> 15 days/month) was associated with medication overuse [bib_ref] Chronic migraine and medication-overuse headache through the ages, Boes [/bib_ref] [bib_ref] Classification of daily and near-daily headaches: proposed revisions to the IHS criteria, Silberstein [/bib_ref].
The term "medication-overuse headache" was first introduced in the second edition of the ICHD in 2004. It also defined MOH subtypes induced by simple analgesics, combination-analgesics, ergots, triptans and opioids. The diagnostic criteria included a mandatory prerequisite that the headache syndrome resolved or reverted to the previous pattern within 2 months after discontinuation of the overused drug. This caused the entity of definite MOH to be diagnosed retrospectively and more difficult to handle in clinical practice. The criterion was changed in 2006 when a board of experts published revisions by consensus and introduced a broader concept of MOH, in which the diagnosis was based on headache frequency (equal to or greater than 15 days/month) and overuse of headache medication, but did not require the headache to improve after withdrawal [bib_ref] Headache classification Committee of the International Headache Society (IHS) the international classification..., Olesen [/bib_ref]. This criterion was omitted again in the latest and current Third Edition of the International Classification of Headache Disorders (ICHD-3) [22].
## Current definitions
In ICHD-3, chronic headache syndromes are defined by expert consensus as headache disorders that share characteristics with pre-existing headache syndromes, occur for a certain amount of time (at least 3 months in e.g. chronic tension-type headache (CTTH), chronic migraine (CM); or at least 1 year in e.g. chronic trigeminal autonomic cephalalgia (TAC)) and have an additional time-criterion (e.g. headache days per month in CTTH and CM, or the absence of remissions for more than 3 months in TAC's). MOH is found in ICHD-3 under subsection 8.2 as a chronic headache disorder secondary to a pre-existing headache syndrome. It is stipulated as a consequence of regular overuse of drugs for the acute treatment of headache. To establish the diagnosis, patients have to use symptomatic headache medication on more than 10 or more than 15 days per month, depending on the drug class, for more than 3 months. MOH has 8 subforms -MOH induced by ergotamine, triptans, analgesics including paracetamol, aspirin and other non-steroidal anti-inflammatory drugs (NSAIDs), opioids, combination analgesics, unspecified multiple drug classes and others [fig_ref] Table 1: International Classification of Headache Disorders Third [/fig_ref].
Although considered a general rule in the past, it is now well stated in the classification that MOH usually, but not invariably, resolves once the overuse is stopped [bib_ref] Chronic headaches: a clinician's experience of ICHD3 beta, Manzoni [/bib_ref]. As with all secondary headache syndromes in ICHD-3, there is no longer a necessary requirement of remission or substantial improvement of the underlying causative disorder for the diagnosis to be made. Therefore, when MOH is confirmed by using the medical history of the patient, a two-fold diagnosis is made: the first one entailing the primary headache syndrome that resulted in drug overuse, the second one MOH [bib_ref] Chronic headaches: a clinician's experience of ICHD3 beta, Manzoni [/bib_ref].
## Epidemiology
The prevalence of chronic headache is 4% to 5%, with an incidence of 3% per year [bib_ref] Medication-overuse headache (MOH), Tepper [/bib_ref] [bib_ref] Patterns of medication use by chronic and episodic headache sufferers in the..., Scher [/bib_ref]. The incidence of new-onset CM in patients with EM is around 2.5% per year [bib_ref] Medication-overuse headache (MOH), Tepper [/bib_ref] [bib_ref] Excessive acute migraine medication use and migraine progression, Bigal [/bib_ref]. Even higher incidence rates up to 14% were reported from a tertiary centre [bib_ref] Incidence and predictors for chronicity of headache in patients with episodic migraine, Katsarava [/bib_ref]. Prevalence rates for MOH in the general population level are situated between 1 and 2%, with a range between 0.5% and 7.2% [bib_ref] Definitions of medication-overuse headache in population-based studies and their implications on prevalence..., Westergaard [/bib_ref]. The highest prevalence has been shown in Russia (7.2%) [bib_ref] The prevalence of primary headache disorders in Russia: a countrywide survey, Ayzenberg [/bib_ref]. Knowledge about prevalence and socio-economic burden in lesser developed countries has been very limited for a long time, although studies have been published lately for prevalence in Africa (Zambia 7.1%; Ethiopia 0.7%), Latin America (Brazil 1.4%, Colombia 4.3%) and Asia (Korea 0.5%; China 0.6%) [bib_ref] Chronic daily headache in Korea: prevalence, clinical characteristics, medical consultation and management, Park [/bib_ref] [bib_ref] The prevalence and burden of primary headaches in China: a population-based door-to-door..., Yu [/bib_ref] [bib_ref] The prevalence of primary headache disorders in Ethiopia, Zebenigus [/bib_ref] [bib_ref] The epidemiology of primary headache disorders in Zambia: a population-based door-to-door survey, Mbewe [/bib_ref] [bib_ref] Chronic headache and comorbibities: a two-phase, population-based, cross-sectional study, Silva [/bib_ref]. MOH is estimated to affect around 63 million people worldwide [bib_ref] Medication-overuse headache: epidemiology, diagnosis and treatment, Kristoffersen [/bib_ref]. The prevalence of medication overuse is higher in studies from headache specialist centers, with numbers ranging from 30% to 50% of patients [bib_ref] Medication overuse headache, Cheung [/bib_ref] [bib_ref] Refractory migraine in a headache clinic population, Irimia [/bib_ref].
A systematic review of epidemiological studies found that MOH is most common among middle-aged adults from 30 to 50 years of age, and predominant in females in the majority of studies. The male to female ratio is around 1 to 3-4 [bib_ref] Definitions of medication-overuse headache in population-based studies and their implications on prevalence..., Westergaard [/bib_ref] [bib_ref] Medication-overuse headache: epidemiology, diagnosis and treatment, Kristoffersen [/bib_ref] [bib_ref] Chronic daily headache in Chinese elderly: prevalence, risk factors, and biannual follow-up, Wang [/bib_ref] [bib_ref] Prevalence of chronic migraine and medication overuse headache in Germany -the German..., Straube [/bib_ref] [bib_ref] Epidemiology of medication overuse headache in the general Swedish population, Jonsson [/bib_ref]. Among U.S. children and adolescents, the prevalence of CM was found to be 0.79% if medication overuse was excluded, and 1.75% if it was included [bib_ref] Medication overuse in children and adolescents, Gelfand [/bib_ref] [bib_ref] Prevalence and burden of chronic migraine in adolescents: results of the chronic..., Lipton [/bib_ref]. Prevalence of MOH was greater in girls than boys [bib_ref] Medication overuse in children and adolescents, Gelfand [/bib_ref]. Furthermore, between 21% and 52% of pediatric patients with chronic headache met the criteria for MOH [bib_ref] Prevalence and burden of chronic migraine in adolescents: results of the chronic..., Lipton [/bib_ref] [bib_ref] Chronic daily headache in children and adolescents, Wiendels [/bib_ref]. Worldwide, the prevalence of MOH in pediatric samples was 3.3%, 0.3%, 0.5% and 1.6% in Italy, Taiwan, Norway and Canada respectively [bib_ref] Headache prevalence and disability among Italian adolescents aged 11-15 years: a population..., Maffioletti [/bib_ref] [bib_ref] Chronic daily headache in adolescents, Wang [/bib_ref] [bib_ref] Prevalence and disability of headache among Norwegian adolescents: a cross-sectional school-based study, Krogh [/bib_ref] [bib_ref] Chronic daily headaches in pediatric neurology practice, Moore [/bib_ref]. In the elderly population, studies from multiple headache centers found that around 35% of patients older than 64 years were overusing medication [bib_ref] Prevalence of headache in an elderly population: attack frequency, disability, and use..., Prencipe [/bib_ref] [bib_ref] Headache characteristics and clinical features of elderly migraine patients, De Rijk [/bib_ref]. Reports on prevalence of MOH in specific populations and minorities have been published. In Europe, certain minorities or ethnic groups, such as first-generation migrants, show higher than expected rates of MOH.
Potential explanations for these findings include socioeconomic (e.g. use of healthcare), biological (e.g. genetic) or cultural reasons (e.g. language barriers) [bib_ref] Prevalence of chronic headache with and without medication overuse: associations with socioeconomic..., Westergaard [/bib_ref].
The burden of disease for MOH has been shown to be a worldwide problem. The disorder causes important negative social and economic effects in both rich and poor countries. Mean per-person annual costs were calculated at €3561 for medication overuse [bib_ref] The cost of headache disorders in Europe: the Eurolight project, Linde [/bib_ref]. Not only economic factors, but also psychological and physical disability of chronic headache and MOH needs to be considered. The global campaign "Lifting the Burden" has contributed to the acquisition of new data and to the promotion of accurate epidemiological methods all over the world [bib_ref] Lifting the burden: the global campaign to reduce the burden of headache..., Steiner [/bib_ref] [bib_ref] The impact of headache in Europe: principal results of the Eurolight project, Steiner [/bib_ref]. In the most recent issue of the Global Burden of Disease (GBD) in 2016, migraine became the second largest cause of disability, mainly because MOH was considered a sequela of migraine and tension-type headache [bib_ref] Global, regional, and national incidence, prevalence, and years lived with disability for..., Vos [/bib_ref].
## Risk factors
Medication overuse was found to be an important risk factor for chronification of primary headaches [bib_ref] Headache characteristics and chronification of migraine and tension-type headache: a population-based study, Ashina [/bib_ref]. A systematic review analysed twenty-nine studies and found differences in the risk of developing MOH and the type of used drug. The risk was lowest for triptans (relative risk (RR) 0.65) and ergotamine (RR 0.41) compared to combined analgesics. Triptans and ergotamine containing drugs were found more favorable when compared to opioids [bib_ref] Risk of medication overuse headache across classes of treatments for acute migraine, Thorlund [/bib_ref]. This is in line with Bigal et al. who reported that people using medication containing barbiturates or opiates had a two-fold higher risk of developing chronic headache than patients using single analgesics or triptans. In this study, NSAIDs were protective against developing chronic headache at low to moderate level of monthly headache days, but were associated with an increased risk of developing chronic headache in patients with a high level of monthly headache days (more than 10 days per month) [bib_ref] Excessive acute migraine medication use and migraine progression, Bigal [/bib_ref].
An important risk factor for the development of MOH is predisposition for migraine or tension-type headache as an underlying biological trait. Migraine is the most common pre-existing headache disorder complicated by MOH. Other pre-existing headache disorders can be complicated by MOH as well, such as tension-type headache or cluster headache [bib_ref] Acute migraine medications and evolution from episodic to chronic migraine: a longitudinal..., Bigal [/bib_ref]. Paemeleire et al. investigated the presence of MOH in patients suffering from cluster headache and found this complication only in patients also suffering with migraine or having at least a family history of migraine [bib_ref] Medication-overuse headache in patients with cluster headache, Paemeleire [/bib_ref]. In addition, the clinical experience shows that the majority of patients suffering from cluster headache do not complicate into MOH although overuse of sumatriptan injections can lead to increased frequency of cluster attacks [bib_ref] Medication overuse headache in patients with cluster headache, Paemeleire [/bib_ref]. Patients with other chronic pain disorders who overuse painkillers for non-cephalic pain conditions do not seem to acquire chronic headache, unless they have a pre-existing history of a primary headache disorder [bib_ref] Opiate use to control bowel motility may induce chronic daily headache in..., Wilkinson [/bib_ref].
In a large prospective population-based study, Hagen et al studied 25.596 patients who did not suffer from chronic daily headache at baseline but had MOH 11 years later (n = 201, 0.8%) [bib_ref] A 4-year follow-up of patients with medication-overuse headache previously included in a..., Hagen [/bib_ref]. In this study, the following risk factors were found to be associated with the development of MOH: regular use of tranquilizers (odds ratio (OR) 5.2, 95% confidence interval (CI) 3.0-9.0), combination of chronic musculoskeletal complaints, gastrointestinal complaints and Hospital Anxiety and Depression Scale (HADS) score > = 11, physical inactivity (defined as > = 3 h hard physical activity/week), and smoking (daily vs. never). Furthermore, migraine was a stronger risk factor for MOH than nonmigrainous headache. A strong association was found for a high-frequency headache defined as 7-14 days/months compared to absence of headache days. Non-modifiable risk factors for MOH were age younger than 50, female gender and low level of education. Interestingly, the authors found several risk factors for MOH (e.g. smoking, inactivity) that were not found to increase the risk for chronic daily headache without the overuse of analgesics. Therefore, the authors concluded that both entities might be pathogenetically distinct [bib_ref] A 4-year follow-up of patients with medication-overuse headache previously included in a..., Hagen [/bib_ref]. Lastly, Cevoli et al. detected a more than threefold increased risk of MOH if a family history of MOH or other substance abuse, such as drug or alcohol abuse, was present [bib_ref] Family history for chronic headache and drug overuse as a risk factor..., Cevoli [/bib_ref].
## Clinical features of moh
A comprehensive medical history, clinical examination and the use of internationally accepted criteria and guidelines are the required tools for the diagnosis of MOH. A confirmatory diagnostic test for MOH is currently not available. The headache phenotype of MOH may be indistinguishable from other forms of chronic daily headache. Moreover, the ICHD-3 criteria do not stipulate MOH-specific clinical features (such as headache characteristics or associated symptoms). Awareness for potential secondary headache syndromes is required and 'red flags' have to be searched for in order to the avoid a false-positive diagnosis of MOH in escalating headache disorders, some of which may require medical imaging or lumbar puncture. In practice, an in-depth enquiry of headache types, frequency and especially drug use is always mandatory, as overuse of ergotamine, triptans, NSAIDs, opioids, or analgesic combinations entail different prognostic properties [bib_ref] Medication-overuse headache (MOH), Tepper [/bib_ref] [bib_ref] The differential diagnosis of chronic daily headaches: an algorithm-based approach, Bigal [/bib_ref].
## Comorbidities
Comorbidity is the simultaneous existence of two or more different medical conditions. Comorbidities occur by chance, or by more than chance, suggesting a potential association, causality, common aetiological factors or common pathophysiological processes. In the field of MOH, these terms are often difficult to appoint although researchers have found multiple associations.
Psychiatric comorbidities in MOH are frequent and have been studied extensively since the earliest descriptions of patients with MOH [bib_ref] Longterm prognosis of analgesic withdrawal in patients with drug-induced headaches, Baumgartner [/bib_ref]. MOH and mood disorders such as anxiety and depression are thought to be comorbid disorders by more than chance [bib_ref] Transformation of episodic migraine into daily headache: analysis of factors, Mathew [/bib_ref] [bib_ref] Risk factors for medication-overuse headache: an 11-year follow-up study. The Nord-Trøndelag health..., Hagen [/bib_ref] [bib_ref] Depression and risk of transformation of episodic to chronic migraine, Ashina [/bib_ref] [bib_ref] Psychopathological comorbidities in medication-overuse headache: a multicentre clinical study, Sarchielli [/bib_ref]. In the BIMOH study, a prospective interventional study, Hospital Anxiety and Depression Scale (HADS) scores were collected in patients with MOH (before and after a brief intervention) and controls. MOH patients were found to show significantly higher HADS scores for anxiety [bib_ref] Disability, anxiety and depression in patients with medication-overuse headache in primary care..., Kristoffersen [/bib_ref]. In the "COMOESTAS" trial, using HADS, 40.0% of MOH patients fulfilled the criteria for depression and 57.7% for anxiety [bib_ref] Disability, anxiety and depression associated with medication-overuse headache can be considerably reduced..., Bendtsen [/bib_ref]. The "Eurolight" trial, a cross-sectional study in the adult population of ten European Union countries, found similar results. The association was even stronger compared to a group of patients with migraine without overuse [bib_ref] Headache, depression and anxiety: associations in the Eurolight project, Lampl [/bib_ref].
In the Sodium Valproate in Medication Overuse Headache Treatment (SAMOHA) study, a more extensive screening for psychopathological comorbidities was performed in MOH patients in comparison to patients with EM and healthy controls [bib_ref] Psychopathological comorbidities in medication-overuse headache: a multicentre clinical study, Sarchielli [/bib_ref]. The rate of moderate to severe anxiety in MOH was significantly higher compared to EM patients and healthy controls. Values on the Leeds Dependency Questionnaire were significantly higher in MOH patients compared to EM patients, which indicates a greater susceptibility to drug dependency. When looked at the number of psychiatric disorders, MOH patients were more likely to have multiple psychiatric comorbidities.
An association between clinically relevant obsessive-compulsive disorder (OCD) and MOH was demonstrated [bib_ref] Psychopathological comorbidities in medication-overuse headache: a multicentre clinical study, Sarchielli [/bib_ref]. Around 30% of MOH patients are estimated to show clinical features of subclinical OCD on neuropsychological evaluation. Subclinical OCD may be an additional risk factor for headache chronification [bib_ref] Psychopathological profile of patients with chronic migraine and medication overuse : study..., Curone [/bib_ref] [bib_ref] Obsessive-compulsive disorder and migraine with medication-overuse headache: research submission, Cupini [/bib_ref]. Also, MOH can be associated to substance-related disorder spectrum, moreover since MOH and dependence share common neurobiological pathways, although MOH patients do not share common personality characteristics with drug addicts [bib_ref] What is the role of dependence-related behavior in medication-overuse headache?, Radat [/bib_ref] [bib_ref] Differences in the personality profile of medication-overuse headache sufferers and drug addict..., Galli [/bib_ref].
For metabolic disorders, a couple of studies from North America on obesity found an increased risk of developing chronic headache, although in the European study by no such association found [bib_ref] Risk factors for medication-overuse headache: an 11-year follow-up study. The Nord-Trøndelag health..., Hagen [/bib_ref] [bib_ref] Factors associated with the onset and remission of chronic daily headache in..., Scher [/bib_ref] [bib_ref] Obesity and migraine: a population study, Bigal [/bib_ref]. In a Chinese cohort, an association between MOH and metabolic disturbances such as obesity and hypertension was shown in female patients [bib_ref] Metabolic syndrome in female migraine patients is associated with medication overuse headache:..., He [/bib_ref]. Recent data on smoking, physical inactivity and obesity provided by a Danish cross-sectional analysis confirmed an association between MOH and those metabolic derangements although causality could not be proven [bib_ref] Medication overuse, healthy lifestyle behaviour and stress in chronic headache: results from..., Westergaard [/bib_ref]. In children, the association between obesity and chronic headache has been shown in observational studies, but the link with medication overuse is unclear [bib_ref] Pakalnis A, Kring D (2012) Chronic daily headache, medication overuse, and obesity..., Hershey [/bib_ref]. Lastly, patients with chronic headache and MOH present a high prevalence of sleep complaints [bib_ref] Increased prevalence of sleep disorders in chronic headache: a case-control study, Sancisi [/bib_ref].
## Pathophysiology
A complete understanding of the pathophysiology of MOH currently does not exist [bib_ref] Pathophysiology of medication overuse headache-an update, Srikiatkhachorn [/bib_ref] [bib_ref] Medication overuse headache: pathophysiological insights from structural and functional brain MRI research, Schwedt [/bib_ref] [bib_ref] Neuroimaging findings in patients with medication overuse headache, Lai [/bib_ref]. Although the clinical aspects of MOH seem to be ambivalent, there is evidence for specific neurobiological aspects in MOH-models. Animal studies, genetic studies, structural and functional neuroimaging, and electrophysiological analyses have added to the current knowledge on the pathophysiology of MOH [fig_ref] Figure 1: Current understanding of the pathophysiology of medication-overuse headache [/fig_ref].
Animal studies have shown changes in multiple physiological processes in the central nervous system (CNS) after repetitive administration of analgesics. Chronic sumatriptan exposure produces long-lasting increased susceptibility to evoked cortical spreading depression (CSD) due to lower threshold [bib_ref] Cortical hyperexcitability and mechanism of medication-overuse headache, Supornsilpchai [/bib_ref] [bib_ref] Bongsebandhu-Phubhakdi S, Srikiatkhachorn A (2012) Pathophysiology of medication-overuse headache: implications from animal..., Green [/bib_ref]. Upregulation of vaso-active and pro-inflammatory mediators such as calcitonin gene-related peptide (CGRP), substance P, and nitric oxide synthase were found in trigeminal ganglia [bib_ref] Expression of calcitonin gene-related peptide, substance P and protein kinase C in..., Belanger [/bib_ref] [bib_ref] Triptan-induced enhancement of neuronal nitric oxide synthase in trigeminal ganglion dural afferents..., De Felice [/bib_ref]. An expansion of the receptive nociceptive field, a decreased nociceptive threshold and decreased noxious inhibitory control have been reported [bib_ref] Sustained morphine-induced sensitization and loss of diffuse noxious inhibitory controls (DNIC) in..., Okada-Ogawa [/bib_ref]. Furthermore, chronic exposure to analgesics was found to increase the excitability of neurons in the central nucleus of the amygdala, which may conceptualise the development of anxiety or depression in patients with MOH [bib_ref] Neural hyperactivity in the amygdala induced by chronic treatment of rats with..., Wanasuntronwong [/bib_ref]. The serotonergic modulating system is presumably affected by chronic analgesic use, resulting in neuronal hyperexcitability, enhanced CSD and trigeminal nociception, caused by increased expression of pro-nociceptive serotonin 2A (5HT-2A) receptor binding sites and a decrease in the production of serotonin in the CNS [bib_ref] Effect of chronic analgesic exposure on the central serotonin system: a possible..., Srikiatkhachorn [/bib_ref] [bib_ref] Effects of selective 5-HT1A receptor antagonists on regional serotonin synthesis in the..., Tohyama [/bib_ref] [bib_ref] Effects of acute or chronic administration of anti-migraine drugs sumatriptan and zolmitriptan..., Dobson [/bib_ref]. In analogy to the findings in animals, an upregulation of 5HT-2 receptors on platelet membranes during analgesic abuse and lower platelet levels of serotonin were found, probably caused by suppressed serotonin transport [bib_ref] Platelet serotonin in patients with analgesic-induced headache, Srikiatkhachorn [/bib_ref].
Genetic studies have been performed in MOH although high-quality evidence for genetic traits is currently lacking. In a recent systematic review, Cargnin et al. described candidate polymorphic variants in genes of the dopaminergic gene system (DRD4, DRD2, SLC6A3), and genes related to drug-dependence pathways (WSF1, BDNF, ACE, HDAC3). The authors concluded that these traits are potential risk factors for MOH susceptibility or determinants of monthly drug consumption [bib_ref] A systematic review and critical appraisal of gene polymorphism association studies in..., Cargnin [/bib_ref] [bib_ref] HDAC3 role in medication consumption in medication overuse headache patients: a pilot..., Pisanu [/bib_ref] [bib_ref] Role of 2 common variants of 5HT2A gene in medication overuse headache, Terrazzino [/bib_ref] [bib_ref] The wolframin His611Arg polymorphism influences medication overuse headache, Lorenzo [/bib_ref] [bib_ref] Functional polymorphisms in COMT and SLC6A4 genes influence the prognosis of patients..., Cargnin [/bib_ref] [bib_ref] A genetic association study of dopamine metabolism-related genes and chronic headache with..., Cevoli [/bib_ref] [bib_ref] Drug consumption in medication overuse headache is influenced by brain-derived neurotrophic factor..., Lorenzo [/bib_ref] [bib_ref] Cortical response to somatosensory stimulation in medication overuse headache patients is influenced..., Lorenzo [/bib_ref].
Research shows that central sensitization has a major role in the pathophysiology of MOH [bib_ref] Neuroimaging findings in patients with medication overuse headache, Lai [/bib_ref] [bib_ref] Central sensitization: implications for the diagnosis and treatment of pain, Woolf [/bib_ref]. Using somatosensory evoked potentials comparing cortical responses in MOH patients with responses in healthy volunteers and episodic migraineurs, hypersensitivity (a sign of central sensitization) and hyper-responsiveness of the cerebral cortex were shown in MOH patients as potential markers of altered functioning. The authors suggested that the somatosensory cortex in MOH patients is somehow "locked" in a kind of pre-ictal state [bib_ref] Abnormal cortical responses to somatosensory stimulation in medication-overuse headache, Coppola [/bib_ref] [bib_ref] Drug-induced changes in cortical inhibition in medication overuse headache, Curra [/bib_ref].
More recently, a cohort of MOH patients was followed during a 12 month period, evaluating central sensitization through pain-perception assessment. The authors found evidence of central sensitization at baseline, but most importantly, the study permitted to expose the slow progression towards normalization of sensory processing after detoxification during the extended follow-up window. This adds to the importance of detoxification and observation after withdrawal in order to prevent relapses [bib_ref] Modulation of central sensitisation by detoxification in MOH: results of a 12-month..., Munksgaard [/bib_ref].
Over the last decade, imaging studies have increased the knowledge of structural alterations and physiological events in MOH. Structural imaging studies performed by separate groups have found increased gray matter volume in following areas: periaqueductal gray (PAG) area, posterior cingulate cortex, hippocampus, thalamus, fusiform gyrus, cerebellum and ventral striatum [bib_ref] Medication overuse headache: pathophysiological insights from structural and functional brain MRI research, Schwedt [/bib_ref] [bib_ref] Neuroimaging findings in patients with medication overuse headache, Lai [/bib_ref] [bib_ref] Grey matter changes associated with medication-overuse headache: correlations with disease related disability..., Riederer [/bib_ref]. Less volume was found in the orbitofrontal cortex (OFC), anterior cingulate cortex, left middle occipital gyrus, insula and precuneus [bib_ref] Medication overuse headache: pathophysiological insights from structural and functional brain MRI research, Schwedt [/bib_ref] [bib_ref] Neuroimaging findings in patients with medication overuse headache, Lai [/bib_ref] [bib_ref] Grey matter changes associated with medication-overuse headache: correlations with disease related disability..., Riederer [/bib_ref]. These structures are involved in pain modulation and processing, cognition, affective behavior, addiction and awareness [bib_ref] Medication overuse headache: pathophysiological insights from structural and functional brain MRI research, Schwedt [/bib_ref]. A recent study described disturbances in white matter integrity in the insular cortex and in the parietal operculum [bib_ref] Pain modulation is affected differently in medication-overuse headache and chronic myofascial pain..., Michels [/bib_ref]. It has to be noted however that not all studies found the same morphologic differences in the brains of MOH patients with migraine, including those comparing scans before and after withdrawal [bib_ref] Medication overuse headache: pathophysiological insights from structural and functional brain MRI research, Schwedt [/bib_ref] [bib_ref] Gray matter decrease in patients with chronic tension type headache, Schmidt-Wilcke [/bib_ref] [bib_ref] Brain functional connectivity and morphology changes in medication-overuse headache: clue for dependence-related..., Chanraud [/bib_ref].
Functional imaging has shown altered functional connectivity in pain processing areas, the mesocorticolimbic 'reward' system, the salience network, the fronto-parietal attention network, the default network and memory processing networks [bib_ref] Medication overuse headache: pathophysiological insights from structural and functional brain MRI research, Schwedt [/bib_ref] [bib_ref] Pain modulation is affected differently in medication-overuse headache and chronic myofascial pain..., Michels [/bib_ref] [bib_ref] Brain functional connectivity and morphology changes in medication-overuse headache: clue for dependence-related..., Chanraud [/bib_ref] [bib_ref] Pain processing in medication overuse headache: a functional magnetic resonance imaging (FMRI)..., Ferraro [/bib_ref] [bib_ref] In medication-overuse headache, fMRI shows long-lasting dysfunction in midbrain areas, Ferraro [/bib_ref] [bib_ref] Nucleus accumbens functional connectivity discriminates medication-overuse headache, Torta [/bib_ref]. The mesocorticolimbic dopaminergic 'reward' system, characterized by structures such as the ventromedial prefrontal cortex (VMPFC), the nucleus accumbens and the substantia nigra/ventral tegmental area, seems to be affected in MOH, linking psychiatric characteristics such as dependence mechanisms and addictive components to the disorder [bib_ref] Medication overuse headache: pathophysiological insights from structural and functional brain MRI research, Schwedt [/bib_ref] [bib_ref] In medication-overuse headache, fMRI shows long-lasting dysfunction in midbrain areas, Ferraro [/bib_ref] [bib_ref] Nucleus accumbens functional connectivity discriminates medication-overuse headache, Torta [/bib_ref].
It is noteworthy that, in MOH, changes of functional connectivity and structure may be reversible in some but not all regions and sometimes normalize after treatment [bib_ref] Medication overuse headache: pathophysiological insights from structural and functional brain MRI research, Schwedt [/bib_ref]. By using [18F] fluorodeoxyglucose-Positron emission tomography (FDG-PET), it was detected how several pain processing regions in the brain were hypometabolic during medication overuse but recovered to normal metabolism after withdrawal. An exception to these findings was found in the OFC, a region linked with drug dependence and addiction. This region remained hypometabolic despite discontinuation of analgesics [bib_ref] Orbitofrontal cortex involvement in chronic analgesic-overuse headache evolving from episodic migraine, Fumal [/bib_ref]. Other groups have confirmed changes in this region of interest. Reduction of gray matter volume in the OFC was correlated with headache days at follow-up, hereby exhibiting predictive capability in terms of poor response to treatment [bib_ref] Grey matter changes associated with medication-overuse headache: correlations with disease related disability..., Riederer [/bib_ref] [bib_ref] Cortical alterations in medication-overuse headache, Riederer [/bib_ref]. Non-responders to withdrawal therapy seemed to have less gray matter in the OFC on their pre-detoxification scan and that there was a positive correlation of gray matter in the OFC with response to treatment [bib_ref] Decrease of gray matter volume in the midbrain is associated with treatment..., Riederer [/bib_ref]. Interestingly, in a longitudinal study, MOH patients with clinical improvement after withdrawal had a significant decrease of previously increased gray matter in the midbrain (PAG, nucleus cuneiformis), whereas patients without improvement did not [bib_ref] Decrease of gray matter volume in the midbrain is associated with treatment..., Riederer [/bib_ref]. Another group found that VMPFC dysfunction is reversible and might be attributable to headache, whereas dysfunction observed in the mid-brain dopaminergic areas (substantia nigra/ventral tegmental area) are probably long-lasting and related to drug overuse [bib_ref] Pain processing in medication overuse headache: a functional magnetic resonance imaging (FMRI)..., Ferraro [/bib_ref] [bib_ref] In medication-overuse headache, fMRI shows long-lasting dysfunction in midbrain areas, Ferraro [/bib_ref]. In conclusion, the evidence provided in multiple studies shows that medication overuse causes changes to the CNS in people with an underlying susceptibility for progression. Changes in pain processing networks, dependence networks, sensitization and receptor density in the CNS presumably explain the clinical characteristics of the disorder.
## Treatment
Education and prevention MOH is often considered to be a preventable condition [bib_ref] Medication overuse headache: awareness, detection and treatment, Rapoport [/bib_ref]. Instructing patients about the relationship between an excessive use of acute medications and headache progression is an important preventive measure. The results from multiple studies have shown that most MOH patients have little to no knowledge about excessive drug intake headache chronification. Many patients however received correct information, but often did not remember or had not fully understood the message [bib_ref] Advice alone vs. structured detoxification programmes for medication overuse headache: a prospective,..., Rossi [/bib_ref] [bib_ref] Sociodemographic differences in medication use, health-care contacts and sickness absence among individuals..., Jonsson [/bib_ref] [bib_ref] Patient satisfaction with a neurological specialist consultation for headache, Bekkelund [/bib_ref]. As in other patients with chronic pain conditions, MOH patients seem mainly focused on the side effects related to the acute pain medications, including gastrointestinal bleeding, kidney damage, and liver impairment. They are often surprised when they learn that the excessive use of acute pain drugs might increase headache frequency, leading to MOH [bib_ref] A critical evaluation on MOH current treatments, Negro [/bib_ref]. This is due to the fact that for many MOH patients the symptomatic medications are merely the drugs they need to relief their pain, and the only way that could bring relief to the impact on their lives [bib_ref] Holding on to the indispensable medication-a grounded theory on medication use from..., Jonsson [/bib_ref].
According to this evidence, developing information campaigns and strategies to target patients at risk, preferably before MOH onset, represents an essential objective in headache medicine. In German headache centers, a brochure on medication overuse was effective in preventing the development of MOH in people with migraine and frequent medication use [bib_ref] Prevention of medication overuse in patients with migraine, Fritsche [/bib_ref]. Primary care is the best setting for prevention and initial treatment of MOH, since most MOH patients consult their general practitioner (GP) for headache (80%) [bib_ref] Brief intervention for medication-overuse headache in primary care. The BIMOH study:a doubleblind..., Kristoffersen [/bib_ref]. GPs can play a key role in providing patient education about medication use and modifiable risk factors, such as stress, daily smoking, physical inactivity, and obesity [bib_ref] Medication overuse, healthy lifestyle behaviour and stress in chronic headache: results from..., Westergaard [/bib_ref]. GPs are also capable of prescribing first-line headache prophylaxis in episodic patients when required. MOH patients often bypass medical advice by using over-the-counter medication. A study recruited patients in pharmacies and found that only 14.5% were ever advised to limit intake frequency of acute headache treatments [bib_ref] Self-medication of regular headache: a community pharmacy-based survey, Mehuys [/bib_ref]. In a recent Swedish study investigating the knowledge of 326 pharmacists on headache treatment, only 8.6% demonstrated knowledge that overuse of all types of acute headache medications could lead to the development of MOH [bib_ref] Medication overuse headache: selfperceived and actual knowledge among pharmacy staff, Hedenrud [/bib_ref].
In 2016, the Danish national awareness campaign for MOH was conducted to reach the general public, GPs, and pharmacists. Online resources, print media, radio interviews, and a television broadcast were used to bring key messages such as overuse of pain medication can worsen headaches, pain medication should be used rationally, and medication overuse headache is treatable. The survey showed an increase in percentage of the public who knew about MOH [bib_ref] National awareness campaign to prevent medication-overuse headache in Denmark, Carlsen [/bib_ref].
## Withdrawal as the first phase of treatment
Despite the large controversies about whether medication overuse should be regarded as a cause or a consequence of headache chronification, to date, the worldwide consensus agrees that (ideally complete) withdrawal of acute painkilling drugs is the approach of choice for the acute management of MOH patients [bib_ref] Medication overuse headache: an ongoing debate, Louter [/bib_ref] [bib_ref] Detoxification for medication overuse headache is the primary task, Olesen [/bib_ref] [bib_ref] Probable medication-overuse headache: the effect of a 2-month drug-free period, Zeeberg [/bib_ref]. In a recent randomized controlled open-label trial, complete discontinuation of acute medications came forward as the most effective detoxification program compared to restricted drug intake [bib_ref] Complete detoxification is the most effective treatment of medication-overuse headache: a randomized..., Carlsen [/bib_ref]. Drug discontinuation is advised in most headache treatment guidelines, including guidelines for primary care [bib_ref] Treatment of medication overuse headache??Guideline of the EFNS headache panel, Evers [/bib_ref] [bib_ref] Diagnosis and management of headaches in young people and adults: NICE guideline, Kennis [/bib_ref] [bib_ref] Guideline for primary care management of headache in adults, Becker [/bib_ref]. The crucial therapeutic aspect of withdrawal is that, on the one hand, it is an occasion for the physician to help the patient decrease or stop the use of acute medication, while potentially initiating a new preventive therapy. It is an opportunity for the patient to reconsider his or her headache history, to discover the link with medication overuse and to be guided by the physician in the process of withdrawal [bib_ref] What future for treatment of chronic migraine with medication overuse?, Grazzi [/bib_ref]. Drug discontinuation is performed variously in different headache clinics. In terms of timing, no studies have investigated the abrupt interruption versus the progressive cessation of the overused drugs, but it is widely agreed that for triptans, ergots, combination analgesics, simple analgesics, and NSAIDs the abrupt withdrawal is the treatment of choice, since these medications do not cause severe withdrawal symptoms [bib_ref] Treatment of medication overuse headache??Guideline of the EFNS headache panel, Evers [/bib_ref]. On the opposite, a gradual drug reduction is the best option with barbiturates, benzodiazepines, and opioids [bib_ref] Treatment of medication overuse headache??Guideline of the EFNS headache panel, Evers [/bib_ref]. Withdrawal symptoms (e.g. headache, nausea, vomiting, arterial hypotension, tachycardia, sleep disturbances, etc) generally last for 2-10 days. Seizures or hallucinations are rare, even in patients who are barbiturate abusers. The withdrawal phase is shorter in subjects who excessively use triptans [bib_ref] Medication-overuse headache, Katsarava [/bib_ref].
Certain studies have demonstrated that simple information and advice may be enough to achieve headache improvement in many MOH patients [bib_ref] Short-term effectiveness of simple advice as a withdrawal strategy in simple and..., Rossi [/bib_ref] [bib_ref] Advice alone versus structured detoxification programmes for complicated medication overuse headache (MOH):..., Rossi [/bib_ref]. In the Brief Intervention for Medication-Overuse Headache (BIMOH) study, a sample of MOH patients received a brief intervention of education on medication overuse from their GPs. After 3 months, headache and medication days were reduced by 7.3 days/month, and chronic headache resolved in 50% of the cases [bib_ref] Brief intervention for medication-overuse headache in primary care. The BIMOH study:a doubleblind..., Kristoffersen [/bib_ref]. The effectiveness of this brief intervention was confirmed at 6 month-follow up: headache and medication days were reduced by 5.9 days/month, and chronic headache resolved in 63% [bib_ref] Brief intervention by general practitioners for medication-overuse headache, follow-up after 6 months:..., Kristoffersen [/bib_ref].
Deciding on the setting for withdrawal is a key point of MOH treatment. The choice between outpatient and inpatient withdrawal has to consider many factors, including the patient's motivation, the duration of the overuse, the type of overused drugs, possible previous detoxification failures and comorbidities. An outpatient detoxification can be the preferred setting for highly motivated patients, with a short duration of overuse of simple analgesics, and whose everyday life make an inpatient withdrawal unsuitable [bib_ref] In-patient versus out-patient withdrawal programmes for medication overuse headache: a 2-year randomized..., Créac'h C [/bib_ref]. Instead, inpatient withdrawal therapy is recommended for patients overusing more complex analgesics (such as opioids, tranquilizers, or barbiturates), long duration of overuse, previous failure to withdraw drugs as outpatients and in more complex clinical situations (e.g. psychiatric comorbidities) [bib_ref] Treatment of medication overuse headache??Guideline of the EFNS headache panel, Evers [/bib_ref]. No standardized therapeutic protocol for medication withdrawal is accepted worldwide. Different strategies are employed in clinics such as intravenous hydration, rescue medications such as IV aspirin and IV dihydroergotamine, symptomatic drugs other than those overused, and drugs for withdrawal symptoms including antiemetics (e.g. metoclopramide), clonidine, benzodiazepines, and corticosteroids [bib_ref] Medication-overuse headache: epidemiology, diagnosis and treatment, Kristoffersen [/bib_ref] [bib_ref] Intravenous aspirin (lysine acetylsalicylate) in the inpatient management of headache, Weatherall [/bib_ref] [bib_ref] Repetitive intravenous dihydroergotamine as therapy for intractable migraine, Raskin [/bib_ref] [bib_ref] Amelioration of ergotamine withdrawal symptoms with naproxen, Mathew [/bib_ref] [bib_ref] Prognostic factors for chronic headache, Probyn [/bib_ref] [bib_ref] Intravenous dihydroergotamine for inpatient management of refractory primary headaches, Nagy [/bib_ref]. Considering corticosteroids, there is low evidence for change in various headache outcome measures (i.e. use of rescue medication, days with severe or moderate headache, days without headache, headache days, and headache frequency) [bib_ref] Treatment of withdrawal headache in patients with medication overuse headache: a pilot..., Cevoli [/bib_ref] [bib_ref] The effectiveness of treatments for patients with medication overuse headache; a systematic..., De Goffau [/bib_ref]. Evidence in favor of inpatient withdrawal comes from an observational study showing statistically significant improvement of quality of life, depression and anxiety at 6-month follow-up [bib_ref] Quality of life, depression, and anxiety 6 months after inpatient withdrawal in..., Zebenholzer [/bib_ref]. Furthermore, it is recognized that a proper therapeutic approach to MOH requires a multistep and multidisciplinary program [bib_ref] Efficacy of multidisciplinary treatment in a tertiary referral headache Centre, Zeeberg [/bib_ref] [bib_ref] Detoxification in medicationoveruse headache, a retrospective controlled follow-up study: does care by..., Pijpers [/bib_ref]. The "COMOESTAS" consortium provided an expert consensus protocol in four centers from Europe and two centers in Latin American. The results show that after multiphasic and personalized treatment, two thirds of patients were no longer overusers and almost half reverted to an episodic headache syndrome over a six-month period [bib_ref] A consensus protocol for the management of medication-overuse headache: evaluation in a..., Tassorelli [/bib_ref].
## Prophylaxis
The initiation of preventative therapy is a fundamental therapeutic step to prevent episodic headache converting into a chronic condition. However, the question remains unresolved whether starting prophylactic treatment at the beginning of withdrawal or awaiting the effect of detoxification is the most effective approach. Certain authors recommend that in non-complicated MOH patients, the decision to start preventive treatment may be postponed for two to 3 months following withdrawal. On the contrary, patients who already have a high frequency of headaches before medication overuse and who have been previously treated with more than one preventive treatment, might need early prophylaxis [bib_ref] A narrative review on the management of medication overuse headache: the steep..., Rossi [/bib_ref]. Other clinicians believe that the detoxification can be effective without an immediate prophylaxis [bib_ref] A critical evaluation on MOH current treatments, Negro [/bib_ref]. To date, as confirmed in a recent meta-analysis of randomized controlled trials on the effect of prophylactic therapies (i.e. valproate, nabilone, onabotulinumtoxinA, topiramate, amitriptyline), there is not a preventative drug that has demonstrated superiority to other therapies in a qualitative, appropriately designed study [bib_ref] The effectiveness of treatments for patients with medication overuse headache; a systematic..., De Goffau [/bib_ref]. The results of randomized controlled trials with patients affected by chronic migraine and MOH suggest the use of onabo-tulinumtoxinA and topiramate without early discontinuation. However, the quality of the data is limited due to the fact that it is based on post hoc analysis [bib_ref] Treatment of medication-overuse headache: a systematic review, Chiang [/bib_ref]. A future role for monoclonal antibodies targeting the CGRP pathway is to be awaited [bib_ref] The journey from genetic predisposition to medication overuse headache to its acquisition..., Martelletti [/bib_ref]. Ultimately, the identification of proper prophylaxis should be driven by clinical history, comorbidity, contraindications and side effects of the possible drugs [bib_ref] A critical evaluation on MOH current treatments, Negro [/bib_ref].
## Treatment of comorbidities
Comorbidities have important implications for the management of MOH in daily clinical practice. Co-existence of mood disorders may lead to poorer adherence of headache treatment, leading to unsuccessful headache treatment.
Comorbid psychiatric disorders add to overall burden and reduced quality of life in headache patients and may lead to poorer outcomes after treatment. Therefore, screening patients for anxiety and depression, is important for clinical outcomes and for trials studying MOH. Lastly, attention for metabolic disturbances or unhealthy lifestyle behavioural aspects, such as obesity, smoking and inactivity, in daily practice is probably beneficial not only for general health but also for headache outcomes. As these are mostly modifiable factors, it is reasonable to discuss and treat these conditions accordingly.
## Prognosis
In general, overuse of acute treatment can lead to a poor prognosis of chronic headache and lower quality of life by itself [bib_ref] Prognostic factors for chronic headache: a systematic review, Probyn [/bib_ref]. The outcome for MOH patients withdrawing from their acute treatments has been reported in multiple studies. An accepted endpoint for good response to therapy is a ≥ 50% reduction from baseline headache frequency and/or headache index [bib_ref] Medication overuse headache: a critical review of end points in recent follow-up..., Hagen [/bib_ref]. Successful withdrawal has been found in around 50-70% of MOH patients after 1 year [bib_ref] Longterm prognosis of analgesic withdrawal in patients with drug-induced headaches, Baumgartner [/bib_ref] [bib_ref] Analgesic-induced chronic headache: long-term results of withdrawal therapy, Diener [/bib_ref] [bib_ref] Drug-induced headache: long-term results of stationary versus ambulatory withdrawal therapy, Suhr [/bib_ref] [bib_ref] Long-term outcome of patients with headache and drug abuse after inpatient withdrawal:..., Schnider [/bib_ref] [bib_ref] Drug-induced headache: longterm follow-up of withdrawal therapy and persistence of drug misuse, Fritsche [/bib_ref] [bib_ref] Behavioral and pharmacologic treatment of transformed migraine with analgesic overuse: outcome at..., Grazzi [/bib_ref] [bib_ref] Medication overuse headache: clinical features predicting treatment outcome at 1-year follow-up, Zidverc-Trajkovic [/bib_ref] [bib_ref] Transformed migraine and medication overuse in a tertiary headache Centre-clinical characteristics and..., Bigal [/bib_ref] [bib_ref] Medication-overuse headache and personality: a controlled study by means of the MMPI-2:..., Sances [/bib_ref]. Retaining full withdrawal after 1 year was found to be a good predictor for long-term success [bib_ref] Chronic headache with medication overuse: long-term prognosis after withdrawal therapy, Bøe [/bib_ref] [bib_ref] Long-term predictors of remission in patients treated for medication-overuse headache at a..., Zidverc-Trajkovic [/bib_ref]. In studies with long-term evaluations up to 6 years, relapsing rates between 40 and 50% were found [bib_ref] Drug-induced headache: long-term results of stationary versus ambulatory withdrawal therapy, Suhr [/bib_ref] [bib_ref] Long-term outcome of patients with headache and drug abuse after inpatient withdrawal:..., Schnider [/bib_ref] [bib_ref] A retrospective long-term analysis of the epidemiology and features of drug-induced headache, Evers [/bib_ref] [bib_ref] Long-term follow-up of patients treated for chronic headache with analgesic overuse, Pini [/bib_ref] [bib_ref] Are there predictive factors for long-term outcome after withdrawal in drug-induced chronic..., Tribl [/bib_ref] [bib_ref] Medication overuse headache: rates and predictors for relapse in a 4-year prospective..., Katsarava [/bib_ref] [bib_ref] Long-term evolution of chronic daily headache with medication overuse in the general..., Fontanillas [/bib_ref]. A successful withdrawal leads to a better response for prophylactic treatment, even in patients with little improvement in headache frequency [bib_ref] Discontinuation of medication overuse in headache patients: recovery of therapeutic responsiveness, Zeeberg [/bib_ref]. Multiple predictors of relapse have been documented. Patients with tension-type headache have a higher relapse risk . A longer duration of regular intake is a predictor for relapse [bib_ref] Long-term follow-up of patients treated for chronic headache with analgesic overuse, Pini [/bib_ref] [bib_ref] Ergotamine abuse. Do patients benefit from withdrawal?, Tfelt-Hansen [/bib_ref]. Patients who kept overusing medication in the long-term had a poor response to withdrawal therapy and had a higher frequency of chronic headache [bib_ref] Chronic headache with medication overuse: long-term prognosis after withdrawal therapy, Bøe [/bib_ref]. Risk factors for relapse in short-term (1 year) were: high number of acute treatments, smoking, alcohol consumption and return to overused drugs [bib_ref] Risk factors in medication-overuse headache: a 1-year follow-up study (care II protocol), Sances [/bib_ref]. Patients withdrawn from triptans have a lower risk for relapse, while combined drug therapy had a higher relapse rate [bib_ref] Drug-induced headache: long-term results of stationary versus ambulatory withdrawal therapy, Suhr [/bib_ref] [bib_ref] Rates and predictors for relapse in medication overuse headache: a 1-year prospective..., Katsarava [/bib_ref] [bib_ref] Medication overuse headache in Scandinavia, Jensen [/bib_ref]. Codeine containing drugs, low self-reported sleep quality and high self-reported bodily pain are probable predictors for poor outcome after 1 yr.
## Debates in moh
The idea of MOH is well-known and widespread in clinics worldwide. By using the operational criteria for MOH in the ICHD-3 classification, clinicians are able to diagnose MOH as early as first clinic visits in order to guide patients in cutting down the amount of frequently used analgesics. The evidence in favor of the disorder MOH is substantial since global research has gradually improved our knowledge on the complexity of the disorder. Consistent observations from population-based longitudinal studies by headache experts in expertise centers worldwide, have established the entity of MOH in a considerable amount of headache patients. Research on the pathophysiological mechanisms is steadily unravelling the different processes involved with analgesic overuse in headache syndromes. Agreement in findings from imaging studies for entity-specific alterations in the brain have been published, although the amount of data is still limited and needs further research [bib_ref] Pathophysiology of medication overuse headache-an update, Srikiatkhachorn [/bib_ref] [bib_ref] Medication overuse headache: pathophysiological insights from structural and functional brain MRI research, Schwedt [/bib_ref]. Furthermore, the results from neuroimaging suggest that neuroplasticity exists and that specific imaging findings can be predictive for the outcome after withdrawal. Lastly, the field of genetics in MOH is in development, reaching out to a more personalized approach for MOH [bib_ref] The journey from genetic predisposition to medication overuse headache to its acquisition..., Martelletti [/bib_ref].
However, it is important to raise awareness on the current limitations to the state of literature on MOH. Questions need to be put forward on how to analyse the phenomenon of deteriorating headaches with the use of analgesics. Mostly, the disorder is seen in patients with pre-existing headache disorders and therefore analysing it as a complication to these conditions is reasonable. Given the potential for the onset of chronic headache after regular intake of analgesics for other medical conditions, MOH can be conceptualized as a secondary headache disorder. But reminding ourselves to patients who experience increasing headache severity and frequency without drug overuse, the overuse of analgesics can be seen as a mere epiphenomenon to the primary headache disorder, a cycling disorder with good and bad phases, for which treatment of the headache syndrome without detoxification is required [bib_ref] Detoxification in medicationoveruse headache, a retrospective controlled follow-up study: does care by..., Pijpers [/bib_ref] [bib_ref] Fluctuations in episodic and chronic migraine status over the course of 1..., Serrano [/bib_ref]. The scientific community has not yet come to the end of this discussion. The lack of robust evidence from high-quality, well-designed and large randomized controlled clinical trials on MOH is important in this analysis [bib_ref] Medication overuse headache an entrenched idea in need of scrutiny, Scher [/bib_ref]. Withdrawal studies over the years have delivered evidence of moderate quality, mostly due to the lack of control groups, lack of randomization, difficulties in adequate blinding and often high dropout rates [bib_ref] Treatment of medication-overuse headache: a systematic review, Chiang [/bib_ref] [bib_ref] Medication overuse headache an entrenched idea in need of scrutiny, Scher [/bib_ref]. Furthermore, evidence in favour of starting prophylactic treatment in MOH comes from post-hoc analysis without adequate power [bib_ref] Detoxification for medication overuse headache is the primary task, Olesen [/bib_ref] [bib_ref] Treatment of medication-overuse headache: a systematic review, Chiang [/bib_ref]. One of the most critical aspects of MOH treatment concerns discontinuation of the symptomatic medication. This concept is installed by using observational data and specialist consensus, not on solid level of evidence from large and well powered randomized, blinded trials [bib_ref] Treatment of medication overuse headache??Guideline of the EFNS headache panel, Evers [/bib_ref] [bib_ref] Guideline for primary care management of headache in adults, Becker [/bib_ref]. Because of the huge burden of disease for patients, larger and high-quality interventional trials on the efficacy of treatments are needed [bib_ref] Treatment of medication-overuse headache: a systematic review, Chiang [/bib_ref] [bib_ref] Medication overuse headache an entrenched idea in need of scrutiny, Scher [/bib_ref]. This is complicated by a significant limitation. For an individual patient, the existence of MOH can neither be proven nor invalidated due to the lack of pathognomonic clinical aspects or a clinical useful biomarker, and therefore studies will still rely on consensus criteria.
The diagnostic criteria for MOH in the international classification remain fuel for debate, even after three editions and multiple decades. The discussion whether MOH has a rightful place in the classification as a secondary headache disorder is interesting and relevant. The current ICHD-3 criteria do not denominate MOH to be a 'transformed' version of a primary headache disorder, but instead describe a concurrent medical problem occurring with an underlying headache disorder. The diagnosis is neither a definitive claim on the cause of a progressive headache disorder either. It has a more moderate approach to the occurrence of medication overuse than previous clinical criteria. Furthermore, the current classification uses clinical features that do not touch upon on the underlying neurobiological processes and has rigid elements such as the cut-off of 15 days per month. These elements may need to be revisited when new evidence becomes available in the future [bib_ref] Chronic headaches: a clinician's experience of ICHD3 beta, Manzoni [/bib_ref].
Finally, different authors discussed the previous, current and possible future terminology used in the field of MOH. In terms of semantics, the term "medication-overuse headache" was challenged by Solomon et al. in 2011 [bib_ref] Medication adaptation headache, Solomon [/bib_ref]. MOH has a potential stigmatizing and (self-) blaming message to patients that can put pressure on a good patient-physician relationship. Appellations as "iatrogenic headache" and "medication-overtreatment headache" have a potential to lay blame on healthcare providers [bib_ref] Medication overuse headache an entrenched idea in need of scrutiny, Scher [/bib_ref]. On the other hand, terminology such as "medication misuse headache", "medication abuse headache", "drug abuse headache" also carry a clue of leaving patients responsible for the development of the situation. Hence Solomon et al analyzed possible mechanism-based definitions, for instance "medication-induced headache", "feed-forward headache", "drug-transformed (or amplified) headache" and suggested implicating the term "medication-adaptation headache" as the most appropriate [bib_ref] Medication adaptation headache, Solomon [/bib_ref].
To summarize, after analyzing the literature on MOH, it is our understanding that for clinicians in daily practice, the evaluation of the frequency and quantity of use of analgesics in patients with headache syndromes is a key component of medical assessment in headache disorders. Side effects of analgesic overuse must be actively evaluated and treated accordingly. However, MOH is not to be diagnosed rapidly without further intellectual perseverance, since a false-positive diagnosis of MOH can lead to misdiagnosis. Other secondary causes of headache may lead to headache progression in conjunction with medication overuse. We therefore state that a critical appraisal of the entity of MOH is required in every individual patient. A thorough clinical approach with exact history taking to detect temporal relationships, and clinical examination focused on neurological deficits, remain the foremost necessary assets for clinicians in the absence of accurate technical tools.
# Conclusion
Research in MOH is moving forward and is discovering the mechanisms underlying headache progression and medication overuse. Whether MOH is a definitive distinct entity, a complication in the pathophysiology of primary headache disorders or an epiphenomenon in the natural course of headache disorders is still up for debate. Since methodology is improving and world-wide collaborative efforts are being established, it is clear that high-quality research will help us to resolve multiple questions stated above. Ultimately, by making scientific progress, we are hopeful that new evidence will help clinicians to make the right choices for patients suffering highly disabling headaches and comorbidities.
[fig] Figure 1: Current understanding of the pathophysiology of medication-overuse headache (MOH). The knowledge on the pathophysiology of MOH involves conversion from and reversion to primary headache disorders, showing changes in physiological processes, functional connectivity and structural changes of the central nervous system, in patients with underlying genetic susceptibility. Abbreviations: MOH: medication-overuse headache; CNS: central nervous system [/fig]
[table] Table 1: International Classification of Headache Disorders Third [/table]
|
Evaluation of the feasibility of explainable computer-aided detection of cardiomegaly on chest radiographs using deep learning
We examined the feasibility of explainable computer-aided detection of cardiomegaly in routine clinical practice using segmentation-based methods. Overall, 793 retrospectively acquired posterioranterior (PA) chest X-ray images (CXRs) of 793 patients were used to train deep learning (DL) models for lung and heart segmentation. The training dataset included PA CXRs from two public datasets and in-house PA CXRs. Two fully automated segmentation-based methods using state-of-the-art DL models for lung and heart segmentation were developed. The diagnostic performance was assessed and the reliability of the automatic cardiothoracic ratio (CTR) calculation was determined using the mean absolute error and paired t-test. The effects of thoracic pathological conditions on performance were assessed using subgroup analysis. One thousand PA CXRs of 1000 patients (480 men, 520 women; mean age 63 ± 23 years) were included. The CTR values derived from the DL models and diagnostic performance exhibited excellent agreement with reference standards for the whole test dataset. Performance of segmentation-based methods differed based on thoracic conditions. When tested using CXRs with lesions obscuring heart borders, the performance was lower than that for other thoracic pathological findings. Thus, segmentation-based methods using DL could detect cardiomegaly; however, the feasibility of computer-aided detection of cardiomegaly without human intervention was limited.Cardiomegaly, also known as an enlarged heart, is a symptom of cardiac insufficiency that occurs in various diseases and conditions, including high blood pressure, coronary artery disease, heart valve disease, and pulmonary hypertension. Because of its noninvasive nature, minimal radiation exposure, and economic considerations, chest X-ray imaging (CXR) is one of the most widely used medical imaging tests for early cardiomegaly detection. Recent advances in deep learning (DL) technologies and large-scale CXR database constructions have enabled significant performance improvements in computer-aided detection of cardiomegaly to a level comparable to that of radiologists 1-12 . Methods based on binary classification of the entire CXR into cardiomegaly detection via image-level label-dependent learning have dominated in the literature on DL-based automated detection of cardiomegaly. However, classification-based methods have an intrinsic limitation, as the mechanism by which DL arrives at the conclusion (i.e., the condition of anomalies in heart structure) remains obscure. Conversely, segmentation-based methods extract boundaries of the lungs and heart on CXRs to automatically calculate the cardiothoracic ratio (CTR), which is a useful index of cardiac enlargement. These approaches are more intuitive and explainable than the classification-based methods and have demonstrated promising results for limited datasets 5-12 .between the methods was observed(Fig. 1). For both methods, the diagnostic performances of subgroups 2 and 3 were lower than those of other subgroups, most notably that of subgroup 3 (method 1: accuracy = 78% [95% CI 69, 87], sensitivity = 98% [95% CI 95, 100], specificity = 56% [95% CI 45, 67], AUC = 0.86 [95% CI 0.76, 0.98]; method 2: accuracy = 75% [95% CI 66, 84], sensitivity = 91% [95% CI 85, 97], specificity = 53% [95% CI 42, 64], AUC = 0.81 [95% CI 0.69, 0.93]).
www.nature.com/scientificreports/ The aim of this study was to determine the feasibility of explainable computer-aided detection of cardiomegaly on chest radiographs. Hence, we developed two fully automated segmentation-based methods using state-ofthe-art DL models for lung and heart segmentation, and evaluated their diagnostic performance and reliability for CTR measurements using chest radiographs of normal patients and those of patients with diverse thoracic pathological conditions commonly encountered in routine clinical practice. According to the experimental results, CTRs derived from deep learning models and diagnostic performance exhibited excellent agreement with reference standards (method 1: the area under the receiver operating characteristic curve [AUC] = 0.96, p = 0.655; method 2: AUC = 0.95, p = 0.917). Performance of segmentation-based methods differed depending on the thoracic pathological conditions. When tested using chest x-ray images with lesions obscuring heart borders, the performance was lower than that for other thoracic pathological findings (method 1: AUC = 0.86, p = 0.003; method 2: AUC = 0.81, p = 0.001).
# Results
Study participants. In total, 1000 posterior-anterior (PA) CXRs of 1,000 patients (mean age ± standard deviation , 63 ± 23 years; age range 10-98 years; sex, 480 men and 520 women) were included in our study .
Diagnostic performance of cardiomegaly detection. The diagnostic performance of cardiomegaly detection is summarized in [fig_ref] Table 2: Diagnostic performance of detection of cardiomegaly [/fig_ref]. When the entire test dataset was considered, both segmentation-based methods exhibited similar overall performance (method 1: accuracy = 91% (95% confidence interval Reliability of CTR calculation. We observed a high spatial overlap between the manual and automatic segmentation results for both segmentation-based methods. On validation data, the average Dice scores for the lungs and heart segmentation tasks by the segmentation-based method 1 were 0.968 and 0.955, respectively. Similarly, the average Dice scores of the segmentation-based method 2 were 0.971 and 0.957 for the lung and heart segmentation tasks, respectively. When considering all test data, no significant differences were observed between the automatic calculation of CTRs and reference standards of both methods (method 1: p = 0.655; method 2: p = 0.917). However, individual methods exhibited different behaviors depending on thoracic pathological conditions. In method 1, the p-values of subgroups 1, 2, and 3 were < 0.05 (subgroup 1: p = 0.007; subgroup 2: p = 0.043; subgroup 3: p = 0.003). Similarly, in method 2, the p-values of subgroups 1 and 3 were < 0.05 (subgroup 1: p < 0.001; subgroup 3: p = 0.002). www.nature.com/scientificreports/ The overall mean absolute errors (MAEs) of methods 1 and 2 for the entire test dataset were 0.023 and 0.024, respectively. Similar to the diagnostic performance, the largest MAE for each method occurred in subgroup 3 (MAE ± SD, method 1: 0.058 ± 0.074; method 2: 0.059 ± 0.066). For subgroup 5, in which hard cases were excluded, the corresponding MAEs were the smallest in each method (MAE ± SD, method 1: 0.019 ± 0.024; method 2: 0.019 ± 0.023). The results are summarized in and performance examples are presented in Figs. 2 and 3.
# Discussion
Segmentation-based approaches to detect cardiomegaly are designed to address cardiomegaly detection tasks on chest radiographs via segmentation of the lungs and heart and subsequent calculation of the CTR. Here, we developed two fully automated segmentation-based methods using different state-of-the-art DL models for semantic segmentation (method 1: Dice = 0.968 and 0.955 for the lungs and heart segmentation tasks, respectively; method 2: Dice = 0.971 and 0.957 for the lungs and heart segmentation tasks, respectively). We observed that the CTR values derived from the DL models and diagnostic performance exhibited excellent agreement with reference standards for the entire test dataset (method 1: AUC = 0.96 [95% CI 0.94, 0.97], p = 0.655; method 2: AUC = 0.95 [95% CI 0.94, 0.97], p = 0.917). Both segmentation-based methods exhibited noticeable performance differences depending on thoracic pathological conditions. For instance, in subgroup 3, the mean difference between the automatic CTR calculations and reference standards was statistically significant at the 5% significance level in both methods (method 1: p = 0.003; method 2: p < 0.001). Consequently, the corresponding AUCs (method 1: AUC = 0.86 [95% CI 0.76, 0.98]; method 2: AUC = 0.81 [95% CI 0.69, 0.93]) were the lowest. In contrast, in subgroup 5, which did not contain any hard cases, the diagnostic performance and reliability of CTR calculations were consistent with those obtained with the entire test dataset (method 1: AUC = 0.97 [95% CI 0.95, 0.98], p = 0.071; method 2: AUC = 0.97 [95% CI 0.96, 0.99], p = 0.084). Notably, although the paired t-test results for subgroup 1 revealed significant differences between the automatic CTR calculations and reference standards for both methods, the corresponding AUCs reflected the opposite results (method 1: AUC = 0.96 [95% CI 0.92, 0.99]; method 2: AUC = 0.98 [95% CI 0.96, 1.00]). The main reason for this was that the test data in subgroup 1 were imbalanced, as only 80% of CXRs (98 out of 123) were negative. In addition, as both DL models tended to underestimate the CTRs for negative CXRs with pneumothorax (PX), the number of true negative cases increased, leading to an increase in the overall detection accuracy for subgroup 1. This also accounted for the high specificity and low sensitivity observed in subgroup 1. For subgroup 2, no significant differences were observed between the automatic CTR calculation and reference standard (p = 0.130), but the diagnostic performance was relatively low. This suggested that the CTRs were not underestimated nor overestimated relative to the reference standard, but the errors were large.
The overall diagnostic performance of our methods exceeded that reported by Chamveha et al.(accuracy: 68.45%, sensitivity: 75.0%, specificity: 69.5%). The AUCs when hard cases were excluded (e.g., subgroup 5, method 1: AUC = 0.97 [95% CI 0.95, 0.98]; method 2: AUC = 0.97 [95% CI 0.96, 0.99]) were comparable to those reported by Sogancioglu et al. [bib_ref] Cardiomegaly detection on chest radiograph: Segmentation versus classification, Sogancioglu [/bib_ref] (0.98), in which CXRs with ambiguous heart borders were excluded. The corresponding MAEs of both methods for subgroup 5 (method 1: 0.019, method 2: 0.019) were also comparable to their results (0.0135). Li et al. [bib_ref] Automatic cardiothoracic ratio calculation with deep learning, Li [/bib_ref] reported that paired differences between their DL model and reference standards for CTR measurements (p = 0.55) using 500 CXRs were not statistically significant, in line with our observations. They reported poor performance of their DL model for pleural effusion (PE) and fat pad of the pericardium, which belonged to subgroups 2 and 3 in our study, respectively. Gupte et al.trained the segmentation deep learning model with 1952 CXRs obtained from a public dataset released by the National Institute of Health 13 and two private hospitals. Interestingly, they achieved a sensitivity of 96% and a specificity of 81% using the held-out test data, and a sensitivity of 87% and a specificity of 97% using the out-of-source dataset. Saiviroonporn et al. [bib_ref] Cardiothoracic ratio measurement using artificial intelligence: Observer and method validation studies, Saiviroonporn [/bib_ref] clinically evaluated a deep learning-based automatic CTR measurement on a large dataset (n = 7517). They . Reliability of the cardiothoracic ratio calculation. CI confidence interval, RS reference standard, MAE mean absolute error, SD standard deviation. www.nature.com/scientificreports/ www.nature.com/scientificreports/ reported that the diagnostic performance based on the automated calculation of CTRs using the DL model can provide an excellent outcome (AUC = 0.96). However, similar to our findings, they observed that it should be improved on heart diameter calculation, which is difficult to be performed because its pixel value is low, and its edges are fused with the lung borders or the thoracic spine [bib_ref] Artificial intelligence-based diagnosis of cardiac and related diseases, Arsalan [/bib_ref].
The methodology used in this study is distinct to segmentation-based methods used in previous work in several ways, which makes our study novel. First, segmentation of anatomical structures in chest radiographs has been extensively investigated, but only a few studies have evaluated lung boundary detection algorithms in lungs with structural deformities [bib_ref] A review on lung boundary detection in chest X-rays, Candemir [/bib_ref]. Thus, in an effort to process more comprehensively a wide variety of lung shapes, we manually annotated the masks of the lungs and heart of CXRs from 270 patients with PX or manifested tuberculosis (TB), which were subsequently employed as training data. Second, previous studies have only performed statistical analyses on the entire test data as a whole to evaluate the performance of their solutions. In contrast, in our work, we analyzed the overall performance of segmentation-based methods for the entire test data and their detailed behaviors depending on various thoracic pathological conditions, particularly those recognized as hard samples for automatic CTR measurements. To the best of our knowledge, this is the first study to explore the impact of individual thoracic pathologies on the effectiveness of DL-based automatic CTR measurements and application in diagnosis. Further, we harnessed state-of-the-art semantic segmentation DL models, which may be more efficient at segmenting the lungs and heart compared to the standard 2D U-Net architectures with different backbone networks used previously [bib_ref] Automatic cardiothoracic ratio calculation with deep learning, Li [/bib_ref] [bib_ref] Cardiomegaly detection on chest radiograph: Segmentation versus classification, Sogancioglu [/bib_ref] [bib_ref] Cardiothoracic ratio measurement using artificial intelligence: Observer and method validation studies, Saiviroonporn [/bib_ref]. This study had several limitations. First, we only considered the CXRs from patients with PX or TB to more accurately analyze abnormal lung anatomy. In general, DL models can recognize more patterns with the availability of more training data [bib_ref] Do we need more training data?, Zhu [/bib_ref]. Therefore, training using a wide variety of CXRs with ambiguous lung and cardiac silhouettes because of a disease, accidents, or postsurgical alternations would enhance the generalization capability of DL models. Second, the accurate segmentation of the lungs and heart is a prerequisite for the automatic CTR calculation. Our findings were based on results obtained from two state-of-the-art DL models for segmentation. Therefore, future studies should validate whether similar results can be obtained using other stateof-the-art DL models. Third, we used training data from different institutions, but the test data were retrieved from a single institution, which may have affected the reliability of our findings.
In conclusion, segmentation-based methods using DL detected cardiomegaly with an acceptable level of performance and better interpretability. For patients with certain thoracic pathologies, such as PE or lesions obscuring the heart border, CTR calculations may be inaccurate and generate more false positive errors. Our findings suggested that the feasibility of explainable computer-aided detection of cardiomegaly without complete human intervention is limited and, thus, careful attention should be paid to patients with certain thoracic lesions.
# Methods
This study was approved by the institutional review board of the Keimyung University Dongsan Medical Center, with a waiver for written informed consent (DSMC-2021-02-021). In addition, we confirm that all methods were performed in accordance with the relevant guidelines and regulations.
Dataset for training DL models. In total, 408 unique PA CXRs between January 2016 and December 2019 were extracted from the picture archiving and communication system repository of our hospital (mean age ± SD, 50 ± 11 years; age range 18-95 years; 184 men and 224 women). To enable processing of a variety of lung shapes, 270 PA CXRs were randomly selected from patients with PX and the remaining 138 PA CXRs were obtained from patients with TB. Digital Imaging and Communication in Medicine images were converted into lossless 24-bit gray-scale JPEG format while maintaining their original resolution and default window level settings as stored in the Digital Imaging and Communication in Medicine metadata. All the images were deidentified before analysis. Masks of the lungs and heart in CXRs were manually segmented by a board-certified radiologist (M.L.).
To prevent overfitting and enhance the generalization capacity of the DL models for lung and heart segmentation, CXRs obtained from two publicly available datasets were used in this study . The Japanese Society of Radiological Technology dataset [bib_ref] Development of a digital image database for chest radiographs with and without..., Shiraishi [/bib_ref] contained 247 PA CXRs, of which 154 had lung nodules (100 malignant cases, 54 benign cases) and 93 had no lung nodules. All CXRs had a resolution of 2048 × 2048 pixels and a gray-scale color depth of 12 bit. The dataset provided reference boundaries for the lungs, heart, and clavicle. The Montgomery dataset 17 contained 138 PA CXRs, including 80 normal patients and 58 patients with TB. The resolution of CXRs was 4020 × 4892 or 4892 × 4020 pixels with a 12-bit gray-scale color depth. As the Montgomery dataset only provided pixel-wise lung mask annotations, annotations for heart masks were performed by a board-certified radiologist (M.L.). In total, 793 PA CXRs of 793 patients in different age groups were used for training DL models for lung and heart segmentation [fig_ref] Figure 4: Flow chart for datasets [/fig_ref].
The segmentation performance was evaluated using the Dice score that was calculated using the following formula: (2 × TP)/((TP + FP) + (TP + FN)), where TP, FP, and FN indicated the number of true positive, false positive, and false negative pixels, respectively.
Test dataset. Additional 1000 PA CXRs were collected between January 2016 and December 2019 from our picture archiving and communication system repository to evaluate the diagnostic performance and reliability of CTR measurements of segmentation-based methods. The test dataset was carefully curated to encompass diverse cases including the CXRs with deformed lungs and/or the silhouette sign to investigate the feasibility of explainable computer-aided detection of cardiomegaly in routine clinical practice. The inclusion criteria were as follows: (1) PA CXRs and (2) no overlap of patients. Each CXR in the test dataset was reviewed and annotated by two board-certified radiologists (M.L. and S.K.). For cases of disagreement in findings or CTR measurements, www.nature.com/scientificreports/ consensus was achieved by discussion between the two radiologists. Consequently, the annotation information included the existence of any thoracic pathologies and the corresponding CTR value. Details of the breakdown of the test dataset are presented in . In addition to the performance analysis for the whole test dataset, a subgroup analysis was performed to determine whether specific thoracic pathological conditions influenced the performance of segmentation-based methods. The test dataset was grouped into the following five overlapping subgroups: subgroups 1 (patients with PX), 2 (patients with PE), 3 (patients with lesions obscuring the heart border [e.g., right middle robe pneumonia, left lingular segment pneumonia, atelectasis, among others]), 4 (patients with lesions obscuring the diaphragm border excluding PE [e.g., right lower lobe consolidation, left lower lobe consolidation, cancer, among others]), and 5 (patients with no findings, cardiomegaly only, or other thoracic pathological findings that did not belong to the other subgroups). Interestingly, subgroups 1-4 included various hard cases, such as lungs with a deformed appearance or ambiguous cardiac silhouette.
Deep neural network architecture for semantic segmentation. Two fully automated segmentation-based methods were developed and evaluated using different state-of-the-art DL models to minimize the derivation of biased conclusions. Method 1 used two separate XLSor 18 models for segmenting the lungs and heart, respectively. The XLSor model consisted of three functional components: a deep convolutional neural network, a recurrent attention module, and segmentation layers. The deep convolutional neural network employed the ImageNet pre-trained ResNet-101 19 as a backbone, and replaced the last two down-sampling layers with dilated convolution operations. The output feature map of the convolutional neural network was fed into the recurrent attention module, in which long-range contextual information was collected from all pixels to enhance pixel-wise representation capability. Finally, the segmentation layers applied multiple transpose convolution operations to the output of the attention module to generate the final segmentation masks.
The DL model of method 2 was built upon the standard U-Net 20 architecture for multi-class semantic segmentation (e.g., lungs, heart, and background). Similar to the XLSor model, the DL model of method 2 applied self-attention modules to improve discriminative feature representation ability. The channel attention block of the attention module extracted the inter-channel relationships of the input feature map. The spatial channel block encoded the relative importance of each spatial position of the input feature map [bib_ref] Automatic lung segmentation on chest X-rays using self-attention deep neural network, Kim [/bib_ref]. Because of the self-contained nature of the attention module, it could be located in the U-Net architecture at any point and in any number. The current implementation applied the attention modules at shallow layers of both contracting and expanding paths of the U-Net, which was determined empirically. The code for DL models of both methods is accessible in Github (https:// github. com/ mkmk0 612/ segme ntati on-based-cardi omega ly-detec tion).
Training and validation details. Of the total data, 80% were used as training data and the remaining 20% were used as validation data for semantic segmentation. This split was conducted for three datasets, respectively, and the resulting sets were recombined for training and validation. Because of the variation in the image www.nature.com/scientificreports/ intensity of CXRs in the dataset, histogram equalization was applied to reduce source-dependent variance and increase the levels of contrast before feeding CXRs into the DL models for both training and testing. Furthermore, because of the variation in the image resolution of CXRs, all CXR images were rescaled into 512 × 512 pixels, which enabled retention of sufficient visual details to delineate boundaries of the lungs and heart. Since the conventional data augmentation schemes based on affine transformations (e.g., shifting, flipping, and zooming) did not improve performance in pilot trials, data augmentation was not applied. Following prediction of lung and heart masks, image post-processing (e.g., small object removal and hole filling) was performed to further improve segmentation, followed by automatic calculation of CTR. DL models were implemented using the PyTorch framework (https:// www. pytor ch. org/) with a CUDA backend. The entire networks of both DL models were trained end-to-end using the stochastic gradient descent optimizer with a mini-batch size of 4. For the DL model of method 1, the initial learning rate was 0.02 and was updated using a ploy learning rate policy [bib_ref] XLSor: A robust and accurate lung segmentor on chest X-rays using criss-cross..., Tang [/bib_ref]. For the DL model of method 2, the base learning rate was set to 0.01 and subsequently decreased by a factor of 10 when the validation set accuracy stopped improving. For both DL models, the mean square error loss function was employed, and the number of iterations for training was set to 40,000 on two graphic cards (GTX Titan XP; NVIDIA, Santa Clara, CA, USA). Early stopping was applied to avoid overfitting.
Statistical analysis. . Diagnostic performance was evaluated in terms of accuracy, sensitivity, specificity, and AUC; the cutoff value of the CTR regarded as cardiomegaly was set to 0.5 in accordance with the regular diagnostic practice. Reliability of automatic CTR calculation was determined using MAE and the paired t-test . Details of the test dataset to evaluate the performance of segmentation-based methods using deep learning models for detection of cardiomegaly. Data in parentheses indicate the percentage of radiographs with respect to the total number of radiographs. NF no finding, CM cardiomegaly, PX pneumothorax, PE pleural effusion, L1 lesions obscuring the border of heart, L2 lesions obscuring the border of diaphragm except PE, OT other common thoracic pathological findings. www.nature.com/scientificreports/ Publisher's note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.
[fig] Figure 1: AUCs for the detection of cardiomegaly according to thoracic pathological conditions. (a) Segmentation-based method 1; (b) segmentation-based method 2. Best viewed in color. AUC area under the receiver operating characteristic curve. Scientific Reports | (2021) 11:16885 | https://doi.org/10.1038/s41598-021-96433-1 [/fig]
[fig] Figure 2, Figure 3: Radiographs of correctly classified examples. CTR values and three lines were generated by segmentation-based methods using deep learning models for the detection of cardiomegaly. The red lines represent the maximum transverse diameter of the left or right side of the heart, respectively; the yellow and blue lines represent the transverse thoracic diameter and the midline of the spine, respectively. For each case, the left image was generated by segmentation-based method 1, and the right image was generated by segmentation-based method 2.(a) PA CXR of a patient with cardiomegaly (reference standard [RS] = 0.522); (b) PA CXR of a patient with cardiomegaly and pleural effusion (RS = 0.611); (c) PA CXR of a patient with pneumothorax, pleural effusion, and left lung collapse, (RS = 0.466); (d) PA CXR of a patient with no finding (RS = 0.41). CTR cardiothoracic, PA CXR posterior-anterior chest X-ray. Radiographs of selected misclassified examples. CTR values and three lines were generated by segmentation-based methods using the deep learning models for the detection of cardiomegaly. The red lines represent the maximum transverse diameter of the left or right side of the heart, respectively; the yellow and blue lines represent the transverse thoracic diameter and the midline of the spine, respectively. For each case, the left and right images were generated by segmentation-based methods 1 and 2, respectively. (a) PA CXR of a patient with cardiomegaly, pleural effusion, consolidations in both lungs, and lesions obscuring the border of the heart (reference standard [RS] = 0.693); (b) PA CXR of a patient with cardiomegaly and linear atelectasis in the left lower lung (RS = 0.522); (c) PA CXR of a patient with cardiomegaly (RS = 0.581); (d) PA CXR of a patient with cardiomegaly and pleural effusion (RS = 0.639). CTR cardiothoracic, PA CXR posterior-anterior chest X-ray. Scientific Reports | (2021) 11:16885 | https://doi.org/10.1038/s41598-021-96433-1 [/fig]
[fig] Figure 4: Flow chart for datasets. The final training and validation sets were used to train the DL models for lung and heart segmentation. The final test dataset was used to evaluate the segmentation-based methods using DL models. PACS picture archiving and communications system, PX pneumothorax, JSRT Japanese Society of Radiological Technology, TB tuberculosis, DL deep learning.Scientific Reports | (2021) 11:16885 | https://doi.org/10.1038/s41598-021-96433-1 [/fig]
[table] Table 2: Diagnostic performance of detection of cardiomegaly. Data in parentheses indicate 95% confidence intervals. TP true positive, FN false negative, TN true negative, FN false negative, AUC area under the receiver operating characteristic curve. [/table]
|
Accessibility of pediatric inpatient services in Japan
In Japan, all citizens are covered by the national insurance system. Children's medical expenses are subsidized by local government co-payments. This removed most economic barriers to visiting medical facilities, geographical obstacles to pediatric medical services remain, including distance to medical facilities and transportation time. However, information on geographic accessibility of pediatric inpatient services is scarce. In this study, I calculated the proportion of children resident in areas accessible to pediatric inpatient service providers within 30 and 60 minutes by automobile. Calculations were based on addresses of hospitals that met criteria for high reimbursement for secondary and tertiary pediatric inpatient services, data for residential blocks, and data for the average velocity of an automobile. In total, 88.0% of children lived within 30 minutes of these hospitals and 95.2% of children lived within 60 minutes. The percentage of children with such access was higher in regions with high population density (e.g., Kanto and Kinki) compared with regions with low population density (e.g., Hokkaido, Tohoku, and Shikoku). Furthermore, regions with high population density also had high rates of children that lived within reach of hospitals with at least five full-time pediatricians.OPEN ACCESSCitation: Ehara A (2018) Accessibility of pediatric inpatient services in Japan. PLoS ONE 13(8): e0201443. https://doi.org/10.
# Introduction
In Japan, all citizens are covered by the national insurance system [bib_ref] Japanese universal health coverage: evolution, achievements, and challenges, Ikegami [/bib_ref]. Local governments subsidize co-payment of medical expenses for children. Although these initiatives removed most economic barriers to visiting medical facilities, geographical obstacles to pediatric medical services remain, including distance to medical institutions and transportation time. Japanese citizens have free access to medical facilities. This system allows patients to choose any clinic or hospital. At night and during holidays, however, most patients are not able to visit their preferred clinics or hospitals because these facilities do not always provide medical services. Therefore, the geographic accessibility of the nearest hospitals must also be analyzed. To provide effective pediatric services, the Japanese government promoted concentration of medical resources into "regional pediatrics centers," which provide pediatric inpatient services and primary care at night and during holidays. The number of hospitals with pediatric departments decreased from 4,120 in 1990 to 2,656 in 2014. However, the accessibility of these facilities has not been comprehensively analyzed. Previously, I calculated the linear distance from each residential block in which children lived to the nearest hospital that met high reimbursement criteria for secondary and tertiary pediatric inpatient services, to show an index of accessibility [bib_ref] Residence of children and the nearest hospital compatible with facility standard for..., Ehara [/bib_ref]. Determining the time and distance involved in visiting medical facilities is necessary to discuss accessibility of such facilities. The Japanese government has reported average automobile velocities according to the type and width of roads in urban and rural areas. This makes it possible to calculate the average time required to visit the closest pediatric inpatient service provider using geographic information systems. In some hospitals, pediatric departments only provide pediatric outpatient services; it is unknown how many hospitals provide inpatient services for children. However, hospitals can receive high reimbursement for secondary and tertiary pediatric inpatient services if they meet health authority criteria for the allocation of full-time pediatricians and nurses [bib_ref] Fees for medical practice reimbursed by health insurance, Igakutushinsya [/bib_ref]. There are five high reimbursement criteria for secondary pediatric inpatient services. Hospitals with at least one full-time pediatrician are able to receive medical fees based on one of these criteria. The criteria met by hospitals with pediatric departments provide an indicator of medical resources for pediatric inpatient care, including the number of pediatricians and nurses. Japan has no clear distinction between medical facilities for secondary and tertiary inpatient care; however, hospitals with many pediatricians are likely to provide tertiary medical services. Lists of these hospitals are available on regional health authority web pages.
Although the address of each child in Japan is unknown, the Population Census reports the number of children in 217,186 residential blocks. When children need pediatric inpatient services, most guardians transport children to hospital by automobile, including private cars, taxis, and ambulances. In this study, I calculated the proportion of children resident in areas accessible to pediatric inpatient service providers within 30 and 60 minutes, using the address of hospitals that met the high reimbursement criteria, data for residential blocks, and data for the average velocity of an automobile.
# Materials and methods
The special hospital fee for secondary and tertiary pediatric inpatient services (Syouni-Nyuin-Iryou-Kanriryou in Japanese) is higher than the ordinary hospital fee, and is grouped in five criteria according to the allocation of full-time pediatricians and nurses [fig_ref] Table 1: High reimbursement criteria for secondary and tertiary pediatric inpatient services [/fig_ref]. This study used addresses for 803 hospitals that met these criteria, as listed on regional health authority web pages (at January 17, 2017, S1 .
Japan is divided into 217,186 residential blocks under the Act on Indication of Residential Address. The mean area of these blocks is 1.72 km 2 and the median is 0.18 km 2 . The location and child population aged under 15 years for each residential block were drawn from the 2010 Population Census (latest data as at January 17, 2017, S2 .
The latitude and longitude of 803 hospitals and 217,186 residential blocks were determined using ArcGIS version 10.4 software (Esri, Redlands, CA, USA). Residential blocks from which hospitals were reachable by automobile within 30 and 60 minutes were calculated with Dijkstra's algorithm using ArcGIS Online (Esri, Redlands, CA, USA) from February 5-22, 2017 [bib_ref] A note on two problems in connection with graphs, Dijkstra [/bib_ref]. Software presets for road data and the average velocity of an automobile (according to the type and width of roads in urban and rural areas) were used to calculate transportation time to hospitals (Table 2). ArcGIS Online (http://www.arcgis.com/index.html) offers a fee-based service (Service area analysis) that identifies areas that can reach a facility within a certain time. This service does not provide raw data of road networks; however, the latest road information used in this service is available through purchase from ESRI Japan as "ArcGIS Geo Suite Road Network" (https://www.esrij.com/products/data-content-geosuite-douromo/). Japan is divided into eight regions [fig_ref] Fig 1: Japan's eight regions [/fig_ref]. In 2014, Japan's population density was 341 persons/km 2 [fig_ref] Table 3: Total population and land area of eight Japanese regions [/fig_ref]. Kanto (including Tokyo and Yokohama: 1,320 persons/km 2 ) and Kinki (including Osaka, Kyoto, and Kobe: 759 persons/km 2 ) have high population density, whereas the other six regions have low density (Hokkaido, 69 persons/km 2 ; Tohoku, 135 persons/km 2 ; Shikoku, 206 persons/km 2 ; Chugoku, 233 persons/km 2 ; Chubu, 321 persons/ km 2 ; Kyushu and Okinawa, 325 persons/km 2 ). Hokkaido accounts for 21.0% of Japan's total area and Tohoku for 17.9%, but only 4.2% of the population lives in Hokkaido and 7.1% live in Tohoku. I compared the proportions of children younger than 15 years resident in areas accessible to hospitals within 30 and 60 minutes among these eight regions.
This study only used previously published data obtained from the Japanese government; therefore, ethical approval from the Medical Research Ethics Committee of Hiroshima International University was not required. Access to pediatric inpatient services
# Results
In total, 88.0% of children in Japan lived in areas accessible to hospitals that met one of the high reimbursement criteria for pediatric inpatient services within 30 minutes. These areas represented 25.7% of Japan's total land area (S3 . The proportion of children who could reach hospital within 30 minutes was higher in regions with high population density (e.g., Kanto and Kinki) compared with regions with low population density [fig_ref] Table 4: Percentage of land area and child population in areas accessible to hospital... [/fig_ref] , [fig_ref] Fig 2: Population density and percentage of child population in areas accessible to hospitals... [/fig_ref].
In addition, 75.5% of the child population lived in areas that could reach facilities in which five or more full-time pediatricians worked (criteria 1-3) within 30 minutes (equivalent to 13.3% of the total land area, S3 . The proportion of children with access to these hospitals was also higher in regions with high population density [fig_ref] Table 4: Percentage of land area and child population in areas accessible to hospital... [/fig_ref] , [fig_ref] Fig 2: Population density and percentage of child population in areas accessible to hospitals... [/fig_ref].
In addition, 95.2% of children lived in areas accessible to hospitals that met one of the high reimbursement criteria within 60 minutes; 90.5% could access hospitals with pediatric departments that met criteria 1-3 within this time (S4 [fig_ref] Table 5 ,: Figs 4 and 5). [/fig_ref] , [fig_ref] Fig 4: Population density and percentage of child population in areas accessible to hospitals... [/fig_ref]. The percentage of children resident in areas accessible to hospitals that met one of the criteria within 60 minutes was high in regions with high population density, such as Kanto (99.2%) and Kinki (97.3%), and was relatively high in other regions (from 85.3% in Shikoku to 95.9% in Chubu). However, lower proportions of children lived in areas accessible to facilities that met criteria 1-3 within 60 minutes in regions with low population density (e.g., Hokkaido, 77.9%; Tohoku; 69.5%; Shikoku, 79.8%) compared with other regions (
# Discussion
Previously, I calculated the linear distance from each residential block to the nearest hospital that met high reimbursement criteria for secondary and tertiary pediatric inpatient services [bib_ref] Residence of children and the nearest hospital compatible with facility standard for..., Ehara [/bib_ref]. Hospitals with at least one full-time pediatrician are able to receive medical fees based on one of five high reimbursement criteria. In Japan, children are able to visit secondary and tertiary medical centers without a referral letter, and there is no clear distinction between secondary and tertiary facilities. However, as provision of tertiary care requires more pediatricians, hospitals that meet criteria 1-3 may mainly provide tertiary services, whereas those meeting 4-5 may mainly provide secondary services. Predicting the time required to transport sick children to these facilities required data for road mileage and velocity of automobiles.
Japanese citizens have free access to medical facilities, which allow patients to choose any clinic or hospital. At night and during holidays, however, most patients are not able to visit their preferred clinic or hospital because these facilities do not always provide medical services. Therefore, in this study, I calculated the time required to visit the nearest hospital providing pediatric inpatient services from each residential block, using data for the average velocity of an automobile (according to the type and width of roads in urban and rural areas). This showed that 88.0% of children in Japan are able to access hospitals that meet one of the high reimbursement criteria for pediatric inpatient services within 30 minutes, and 95.2% within 60 minutes. Options for transporting children who need inpatient care in major cities and rural areas in Japan include private cars, taxis, and ambulances; it is unlikely that other forms of transport such as walking, bicycles, or trains are used in such circumstances. According to a recent Patient Survey, 92,000 children aged under 15 years were discharged from Japanese hospitals in September 2014 . In addition, 471,000 children were transported by ambulance between January 1, 2014 and December 31, 2014. This suggests many children in Access to pediatric inpatient services Japan who need inpatient treatment access hospital by automobile (including ambulances), meaning this analysis based on automobile velocity is appropriate. Most children lived in areas from which pediatric inpatient services were accessible within 60 minutes. Although the number of hospitals with pediatric departments has decreased (from 4,120 in 1990 to 2,656 in 2014), hospital accessibility for children within 60 minutes is likely to be maintained. This secure access is not limited to pediatrics, as 97.8% of women in Japan can access their nearest regional perinatal center within 60 minutes [bib_ref] A nationwide study of accessibility to hospitals providing delivery care service with..., Ishikawa [/bib_ref]. The proportion of children who could access hospitals meeting one of the high reimbursement criteria within 60 minutes did not differ markedly among the eight regions. However, the proportion of children who could access hospitals meeting criteria 1-3 (five or more fulltime pediatricians) and estimated to provide tertiary services was lower in regions with low population density compared with regions with high population density. About 80% of the total person-days for pediatric inpatient care were estimated to be in hospitals that received medical fees based on high reimbursement criteria [bib_ref] Estimation of the total person-days of children receiving inpatient care in Japan, Ehara [/bib_ref]. This indicates that this study appropriately considered children's access to the nearest hospital that provided secondary and tertiary pediatric inpatient care. In Japan, a worker's death due to stroke or ischemic heart disease can be considered an industrial accident from overwork if they worked more than 252 hours/month. Concentration of medical resources into regional pediatrics centers to help decrease pediatricians' work hours is essential to prevent similar issues related to overwork.
Most sick children in Japan are treated by pediatricians rather than by general physicians. In 2014, 29,878 physicians had a main or minor specialty in pediatrics, whereas only 179 physicians were general practitioners and family physicians. In addition, most hospitals with pediatric departments provide pediatric emergency services. Hospitals with few pediatricians are unlikely to be able to sustain 24-hour, 365-day secondary and tertiary pediatric medical services without the help of physicians with other specialties. In 2014, Japan had only 2.35 pediatricians per 10,000 population, meaning employing sufficient pediatricians to provide 24-hour, 365-day services in each municipality may be challenging, especially in underpopulated areas [bib_ref] Presence of pediatricians and population of city, town and village in Japan, Ehara [/bib_ref]. Therefore, establishing secondary or tertiary pediatric centers in each municipality may not be practical. It is necessary to transport children with severe conditions to tertiary centers rather than treating them in small-scale pediatric departments, as the fatality rate of children treated by pediatric critical medicine specialists is lower than that for other specialties [bib_ref] It is essential to concentrate children with severe condition into pediatric intensive..., Takei [/bib_ref]. It is also essential to concentrate human resources (including pediatricians and nurses) in pediatric centers operated for children in areas that cover several cities, towns, and villages. However, it is important to construct appropriate transport systems.
Providing 24-hour, 365-day pediatric services may become a political promise, but provision of such services requires employment of a large number of pediatricians and nurses. Politicians should refrain from requiring hospital administrators to provide such services unless provision is made to employ sufficient medical staff, including pediatricians and nurses. Despite the decrease in number of hospitals with pediatric departments, this study demonstrated that most children can access medical facilities that provide secondary or tertiary pediatric services within 60 minutes by automobile.
# Limitations
1. Japanese citizens have free access to medical facilities, which allows patients to choose any clinic or hospital. However, the accessibility of medical facilities other than the nearest hospital was not analyzed. 2. As only data for the average velocity of an automobile were used, effects of factors such as traffic jams and climate worsening were not considered. However, the percentage of children resident in areas from which secondary or tertiary pediatric inpatient services could be accessed within 30/60 minutes did not differ markedly across regions. Therefore, it is likely that most children could reach such hospitals within 60 minutes even if the transportation velocity was reduced.
3. Each high reimbursement criterion for secondary and tertiary pediatric inpatient services only shows the minimum number of full-time pediatricians; the exact number is unknown.
4. I calculated transportation time to the nearest hospital from each residential block. Transportation time from other places was unknown. However, children's homes are likely to be located near most facilities (e.g., nursery schools, kindergartens, and schools). Therefore, the approximation of transport time in this study may not cause serious issues.
# Conclusions
To provide effective pediatric services, the Japanese government promoted concentration of medical resources into "regional pediatrics centers," which provide secondary or tertiary pediatric inpatient services and primary care at night and during holidays. Around 95.2% of children in Japan can reach such facilities within 60 minutes by automobile, although this proportion is lower in regions with low population density.
Supporting information S1 [fig_ref] Table: Latitude and longitude of 803 hospitals that met one of the high... [/fig_ref] Residential blocks accessible to a hospital meeting criteria 1-3 and any one by automobile within 60 minutes.
(XLSX)
[fig] Fig 1: Japan's eight regions. A, Kanto; B, Kinki; C, Kyushu and Okinawa; D, Chubu; E, Chugoku; F, Shikoku G, Tohoku; H, Hokkaido. Kanto and Kinki have high population density. Hokkaido, Tohoku, Shikoku, Chugoku, Chubu, and Kyushu and Okinawa have low population density. Map of Japan reprinted from Global Map Japan (public domain, open-access resources) under a CC BY license, with permission from the Geospatial Information Authority of Japan. https://doi.org/10.1371/journal.pone.0201443.g001 [/fig]
[fig] Fig 2: Population density and percentage of child population in areas accessible to hospitals meeting criteria 1-3 (red) and any one (blue) within 30 minutes by automobile. A, Kanto; B, Kinki; C, Kyushu and Okinawa; D, Chubu; E, Chugoku; F, Shikoku G, Tohoku; H, Hokkaido. https://doi.org/10.1371/journal.pone.0201443.g002 [/fig]
[fig] Fig 3: Areas accessible to pediatric inpatient service providers meeting criteria 1-3 (red) and 4-5 (blue) within 30 minutes by automobile.https://doi.org/10.1371/journal.pone.0201443.g003 [/fig]
[fig] Fig 4: Population density and percentage of child population in areas accessible to hospitals meeting criteria 1-3 (red) and any one (blue) within 60 minutes by automobile. A, Kanto; B, Kinki; C, Kyushu and Okinawa; D, Chubu; E, Chugoku; F, Shikoku G, Tohoku; H, Hokkaido. [/fig]
[fig] Fig 5: Areas accessible by automobile to pediatric inpatient service providers meeting criteria 1-3 (red) and 4-5 (blue) within 60 minutes.https://doi.org/10.1371/journal.pone.0201443.g005 [/fig]
[table] Table 1: High reimbursement criteria for secondary and tertiary pediatric inpatient services.Table 2. Mean velocity of an automobile according to the type and width of roads (km/h).https://doi.org/10.1371/journal.pone.0201443.t002 [/table]
[table] Table 3: Total population and land area of eight Japanese regions. [/table]
[table] Table 5 ,: Figs 4 and 5). [/table]
[table] Table 4: Percentage of land area and child population in areas accessible to hospital by automobile within 30 minutes. [/table]
[table] Table 5: Percentage of land area and child population in areas accessible to hospital by automobile within 60 minutes. [/table]
[table] Table: Latitude and longitude of 803 hospitals that met one of the high reimbursement criteria for pediatric inpatient services. (XLSX) S2 Table. Location and child population aged under 15 years for each residential block. (XLSX) S3 Table. Residential blocks accessible to a hospital meeting criteria 1-3 and any one by automobile within 30 minutes. (XLSX) [/table]
|
The positive effect of mother-performed infant massage on infantile eczema and maternal mental state: A randomized controlled trial
# Introduction
Eczema, also known as atopic eczema or atopic dermatitis, is a chronic relapsing inflammatory dermatosis characterized by pruritus, xerosis and a close association with immunoglobulin E (IgE)-mediated sensitization to aeroallergens and foods [bib_ref] Atopic dermatitis, Peters [/bib_ref]. The incidence of eczema in children has reached 15-20% [bib_ref] Atopic dermatitis, Ständer [/bib_ref] , with the highest incidence in infants between the ages of 3 and 6 months old [bib_ref] Atopic dermatitis, Ständer [/bib_ref]. A substantial portion of cases with eczema can go into complete remission by 2 years of age, while there are about 40% of cases with a prolonged duration and the highest risk for the atopic march [bib_ref] Clinical phenotypes and endophenotypes of atopic dermatitis: where are we, and where..., Bieber [/bib_ref]. Moreover, infantile eczema obviously impairs infantile quality of life and potentially influence growth. Current treatments for eczema aim to relieve symptoms since there is no cure for it [bib_ref] Johansson M, Benderix Y, Svensson I. Mothers' and fathers' lived experiences of..., Wu [/bib_ref]. General measures include the application of emollients and topical agents and avoidance of infections and trigger factors [bib_ref] Atopic dermatitis, Ständer [/bib_ref].
It is noteworthy that eczema in infants also negatively influences maternal mood. Postpartum mothers are especially susceptible to depressive and anxious episodes (8). It is reported that 8.5% of postpartum mothers experienced anxiety disorder [bib_ref] Anxiety disorders in postpartum women: a systematic review and meta-analysis, Goodman [/bib_ref] and 13% or higher experienced depression in their first postpartum year [bib_ref] Screening for and treating postpartum depression and psychosis: a cost-effectiveness analysis, Fisher [/bib_ref] [bib_ref] A systematic review and meta-regression of the prevalence and incidence of perinatal..., Woody [/bib_ref]. Moreover, the mothers of infants with eczema are more susceptible to the psychological distress such as frustration, depression and anxiety, the level of which is often correlated to the eczema severity.
Infant massage, as a common traditional practice, is widely used all over the world for both preterm and full-term infants nowadays [bib_ref] Massage for promoting growth and development of preterm and/or low birth-weight infants, Vickers [/bib_ref]. A series of studies have revealed its benefits to infants, such as enhancing growth and development, improving sleep and increasing interactions with parents [bib_ref] The efficacy of massage on short and long term outcomes in preterm..., Abdallah [/bib_ref] [bib_ref] Massage therapy research review, Field [/bib_ref] [bib_ref] Review and critical analysis of massage studies for term and preterm infants, Juneau [/bib_ref]. Previous trials also demonstrated that massage could relieve eczema symptoms in young children [bib_ref] Evaluation of massage with essential oils on childhood atopic eczema, Anderson [/bib_ref] [bib_ref] Massage therapy for skin conditions in young children, Field [/bib_ref]. In China, infant massage is often applied on certain meridians and acupoints based on the theory of traditional Chinese medicine (TCM), which effectively relieves eczema in infants and toddlers [bib_ref] 21. He YH, Kang J. Efficacy of Tuina on 120 cases of..., Du [/bib_ref] [bib_ref] Observation on short-and long-term effects of Tuina on eczema among infants and..., He [/bib_ref]. As we all know, infant massage can be performed not only by professionals in clinical setting, but also by parents at home. Systematic reviews recommend that parents can perform massage on lowrisk infants for promoting mental and physical health due to its cost-effectiveness and no evidence of any risk [bib_ref] Massage for promoting growth and development of preterm and/or low birth-weight infants, Vickers [/bib_ref]. Moreover, a series of studies indicated that mother-performed infant massage (MPIM) improved maternal depression, stress or negative mood during postpartum period [bib_ref] Effect of preterm infant massage by the mother on the mood of..., Lotfalipour [/bib_ref] [bib_ref] Blended infant massage-parenting enhancement program on recovering substance-abusing mothers' parenting stress, self-esteem,..., Oswalt [/bib_ref]. However, few trials have simultaneously observed the outcomes of both mothers and infants following MPIM. So far, there is also no trial to investigate whether MPIM can improve infantile eczema and whether MPIM influences the growth of infants with eczema.
Therefore, this trial was designed to observe the potential influence of MPIM on infantile eczema, growth and maternal mental state, which may be a beneficial health-care method for mother-infant dyads.
# Methods
## Trial design
The study used a prospective block-controlled randomized design shown in . Given convenient and practical implement, participants were recruited through publicity in Dingshan street community (Nanjing in Jiangsu province, China). This study was conducted in the health service center of Dingshan street community and at home, respectively based on the intervention protocol between April 2020 and March 2021. The optimal sample size of 87 (29 per group) was calculated by using PASS software (version 15.0, NCSS, USA) based on the following assumptions: β = 0.10, α = 0.05, σ1 = 1.8, σ2 = 0.4, µ1 = 1.6, and µ2 = 0.3, with 90% power, a two-sided alpha of 5%, and an estimated 20% loss to follow-up [bib_ref] The improvement of infantile atopic dermatitis during the maintenance period: a multicenter,..., Wang [/bib_ref]. Thirtyone healthy full-term infants and 66 full-term infants with acute eczema were enrolled in this study. Thirty-one healthy full-term infants were enrolled in the healthy control group (HC group, n = 31) for the comparison of infants' growth and mothers' mental state with eczema groups. By using a random number list produced by Excel software (version 2016, Microsoft, US), 66 infants with eczema were randomly divided into eczema control group (EC group, n = 33) and eczema with MPIM group (EM group, n = 33) in a 1:1 ratio by one investigator, who didn't participate in the following intervention or assessment. Participants were not blinded due to the nature of the intervention. The investigators responsible for the assessment of the infantile growth and eczema severity were blinded. However, the investigators responsible for the instruction of routine care and MPIM as well as regular supervision were not blinded due to the nature of the intervention.
## Inclusion criteria
For the infants with eczema:(1) full-term infant at or under 12 months old; (2) diagnosed as infantile eczema based on the clinical criteria for pediatric atopic dermatitis (28); (3) with informed consent and voluntary compliance with the study arrangement from the infant's mother.
For the healthy infants: (1) same as (1) and (3) for the infants with eczema above; (2) infants without any history of obvious visceral and functional diseases; (3) infants without eczema and other atopic diseases.
## Exclusion criteria
For the infants with eczema: (1) preterm infants or infant over 1 year old; (2) infant with any visceral disease or dysfunction except for eczema; (3) infant with infection, or the history of topical and systemic application of corticosteroids, .
/fpubh. .
## Figure
Flowchart of this study. MPIM, mother-performed infant massage; HC group, healthy control group; EC group, eczema control group; EM group, eczema MPIM group.
antihistamines, antibiotics agents, traditional Chinese herbs or other specific intervention during the last 2 weeks; (4) infants with severe eczema, i.e., eczema area and severity index (EASI) score > 21 (2); (5) infant with obvious eczema or skin lesion on the back where massage is applied; (6) mother with diagnosed severe mental disorders or the scores of Zung's self-rating anxiety scale (SAS) and self-rating depression scale (SDS)≥70; (7) mother unable to follow study protocol. For the healthy infants: same as (1), (3), (5), (6) and (7) for the infants with eczema above.
## Interventions
This study was a 5-month-long intervention study, which included the first 2-month intervention with regular supervision and the latter 3-month intervention without any supervision (shown in . This kind of design aimed to have a knowledge of the feasibility and maternal adherence of MPIM, which is very important for promoting its application as a homebased healthcare method in the community. Data in the three groups were collected at the baseline and at the end of 2and 5-month intervention to evaluate and analyze the outcome difference among groups. One investigator communicated with the mothers in the EC and EM group to enhance the adherence to the respective instruction by wechat communication once a week and by face to face once a month during the first 2month intervention. During the latter 3-month intervention, the mothers in the EC and EM group were advised to record the adherence of respective intervention without any contact from investigators until the last assessment.
## Protocol in the hc group
Infants and mothers in the HC group received none of any specific intervention, except for the assessment at the three timepoints.
. /fpubh. .
## Figure
The protocol of the study.tif.
## Protocol in the ec group
The mothers in the EC group only received the instruction of routine care from one investigator for about half an hour in a small group of 3-5 mothers according to their convenience after the baseline observation. The tips of routine care for infantile eczema included the avoidance of skin irritants and allergens, avoidance of allergenic food in maternal and infantile diet, breast-feeding as much as possible and the application of emollient. For avoiding the confounder caused by using different emollients in this study, Pigeon skin lotion (Pigeon, Shanghai, CNH) was recommended in this study since it is popularly used on infants in China. It was advised to apply the skin lotion on the whole body of the infant except for the scalp at least twice daily, once during day time and once either after bathing or before sleep (if no bathing that day) in the evening. The used amount of the skin lotion was not fixed but determined by the body surface area and dry condition of skin in individual infant, which aimed to achieve the lubricant effect. In addition, let the infant lie on his/her stomach safely and comfortably for 10 min after the application of skin lotion before bedtime, for controlling the massaging position and time in EM group. The mothers in the EC group were instructed about infant massage by investigators after this study observation ended based on the voluntary rule.
## Protocol in the em group
The mothers in the EM group were also instructed with the same routine care as the EC group and infant massage. The mothers were instructed with infant massage by the investigator for about 1 h in a small group of 3-5 mothers according to their convenience after the baseline observation, following the instruction about the same routine care as the EC group. The whole procedure of massage was determined according to the clinical application and textbook [bib_ref] Effect of chiropractic therapy on allergic constitution and intestinal flora in children, Xiong [/bib_ref]. The whole procedure took about 10 min as follows: Let the infant lie safely and comfortably on his/her stomach; the mother washed her hands and applied some skin lotion (Pigeon, Shanghai, CNH) to lubricate her hands and infant's back; firstly, the mother stroked the infant's back longitudinally along the middle line with finger bellies from shoulder level down to the sacrum level for 10 times; secondly, horizontally rubbed the back swiftly and gently with a palm for 20 times respectively at upper, middle, lower back levels; thirdly, applied traditional back-pinching manipulation (BP) for 6 repeats with the finger bellies; lastly, kneaded topdown on both sides of the back slowly and gently with a palm for 5 repeats. The details of one-repeat BP were as follows: pinched the skin located on the middle sacrum with the finger bellies and lifted to twist and move forward swiftly upward to the should level. The intensity of pinching and twisting would be increased gradually to avoid intolerable discomfort. For infants beyond 6 months, BP was performed for 9 repeats. After the instruction and learning, the mothers performed the whole procedure of infant massage on their infants for about 10 min in the presence of the investigator at the health care center. Afterwards, the mothers performed infant massage once daily at home 6 times per week. It was recommended to massage infants before bedtime in the evening as previous reports due to its beneficial effect on infantile sleep pattern [bib_ref] Blended infant massage-parenting enhancement program on recovering substance-abusing mothers' parenting stress, self-esteem,..., Oswalt [/bib_ref]. The investigator also communicated with the mothers on massage practice during wechat communication once a week and by face to face once a month. The mothers were also provided with massage video.
## Outcome observation
Weight, length, head circumference and BMI in infants
The weight, length and head circumference in infants were measured to assess infant's physical growth at the baseline, at the end of the 2-and 5-month intervention by one investigator, who was blinded from group division. The weight and length were measured by using an intelligent physical examination instrument (WS-RTG-ID, Wuhan Computer Software Development Co. LTD, China). Body mass index (BMI) was also calculated by using the following formula: BMI= weight (kg) /length (m) 2 . Head circumference was measured by a measuring tape (Guoshi measuring tape Co., LTD, China).
## Eczema severity
Eczema severity was assessed by using the eczema area and severity index (EASI) at the baseline and the end of the 2-month intervention, which was conducted by one blinded investigator. EASI is a simple, reliable and easily understood system, which can be used by practitioners and investigators as a baseline evaluation and to track changes of eczema condition over time . EASI assesses the key signs of eczema including redness, thickness, excoriation, lichenification and the percentage of skin involving four areas (the head and neck, the trunk, the upper and lower extremities) . The scores are summed to achieve a total score ranging from 0 to 72. An EASI score ≤7 is considered as mild, 8-21 moderate, 22-50 severe, 51-72 very severe [bib_ref] Atopic dermatitis, Ständer [/bib_ref].
## Quality of life in infants with eczema
The quality of life (QOL) in infants with eczema was evaluated by their mothers using the infants' dermatitis quality of life index (IDQOL) at the baseline and the end of the 2month intervention. IDQOL is an easy and sensitive method with good reproducibility for parents to assess QOL impairment in infants with eczema (35). IDQOL contains 11 questions about current dermatitis severity, symptoms such as itching and scratching, mood, sleep, play, family activities, mealtimes, treatments, dressing and bathing. Ten questions present with the scores ranging from 0 to 3 and 1 question from 0 to 4 (35). In addition, the sleep condition in infants was also evaluated by the total scores of 2 questions involving sleep in IDQOL.
## Mental state in mothers
Maternal anxiety and depression levels were evaluated, respectively by using SAS and SDS to investigate their mental state. Mothers in the three groups completed SAS and SDS questionnaires at the baseline and the end of the 2-month intervention. SAS and SDS were designed to quantify the severity of anxiety and depression symptoms, which is widely used as a common and effective self-assessment method (36-39). SDS is often used for perinatal women in China [bib_ref] Discrimination and structural validity evaluation of Zung self-rating depression scale for pregnant..., Chen [/bib_ref] [bib_ref] The interaction between estradiol change and the serotonin transporter gene (5-HTTLPR) polymorphism..., Hu [/bib_ref]. Both of SAS and SDS are 20-item self-reported assessment scales. The total score multiplied by 1.25 provides a standard score. The severity of symptoms was determined based on the standard scores of SAS and SDS as follows: < 50 (normal), 50-59 (mild), 60-69 (moderate), and ≥70 (severe). The score below 50 also indicates the levels of depression and anxiety mood [bib_ref] Prenatal anxiety and obstetric decisions among pregnant women in Wuhan and Chongqing..., Liu [/bib_ref] [bib_ref] Prevalence and risk factors of depression symptoms among Chinese seafarers during the..., Qin [/bib_ref].
## Adverse event
During the study, the mothers and investigators observed whether there were any adverse events in the infants after intervention, including any local impairment of skin, continuous crying, any abnormal change in sleep, eating and bowel movement.
## Observation on the relapse of eczema
Since infantile eczema is a chronic relapsing inflammatory dermatosis (1), this study also observed the relapse condition of eczema in infants at the end of the 5-month intervention comparing with eczema condition at the end of the 2-month intervention. Persistent zero score of EASI from the end of 2month to 5-month interventions was considered as complete remission. Eczema recurred after complete remission at the end of 2-month intervention was considered as a relapse. Persistent eczema without complete remission from the end of 2-month to 5-month interventions was considered as noncomplete-remission. The general condition of infantile eczema was evaluated by a blinded investigator.
# Statistical analysis
The one-way ANOVA and χ2 test were performed to compare baseline data among groups. The repeated-measures ANOVA was performed to compare infantile growth data among groups. The student's t-test was used to compare the scores of EASI, IDQOL and sleep in infant-mother dyads preand post-intervention among groups. The rank sum test was used to analyze the infantile eczema condition and the level of maternal anxiety and depression among groups. The SAS score and SDS score among groups were analyzed by oneway ANOVA and t-test for the comparison of pre-and postintervention. Finally, we used spearman's rank correlation to assess the correlation between EASI score and the scores of
# Results
Demographical data of infant-mother dyads Thirty-one dyads and 63 dyads completed the study. The demographical profile of these infant-mother dyads is shown in . The age and the gender ratio in infants were not significantly different among the three groups (F = 0.464, P = 0.630; χ 2 = 0.663, P = 0.718). There were no significant differences in the age, education background, primiparity and breast-feeding condition among the mothers in the three groups (F = 2.637, P = 0.077; χ 2 = 1.274, 0.141 and 1.005, P = 0.693, 0.932 and 1.000).
## Weight, length, head circumference and bmi in infants
## Improvement on infantile eczema
The severity of infantile eczema was evaluated by EASI and IDQOL scores as shown in . At the baseline, the scores of EASI and IDQOL in infants were not significantly differenct between the EC and EM groups (t = 1.389, P = 0.172; t= 0.349, P =0.728). There were significantly lower scores of EASI (t = 8.749 and 15.460; both P < 0.001) and IDQOL (t = 5.981 and 8.132; both P < 0.001) after the 2-month intervention in both groups. Moreover, the EM group had significantly lower scores of EASI and IDQOL than the EC group (t = 3.953 and 3.797; both P < 0.001). Meanwhile, as shown in , the EM group had significantly lower scores involving sleep state following MPIM than the EC group (t = 3.969, P < 0.001). In addition, there was a significantly positive correlation of EASI scores with IDQOL scores before (γ = 0.348, P = 0.005, and after 2month intervention (γ = 0.709, P < 0.001, . There was also a significant correlation of EASI scores with sleep scores after 2-month intervention (γ = 0.559, P < 0.001, , but not at the baseline (γ = 0.171, P =0.180, . At the end of 5-month intervention, the EM group had significantly more cases of complete remission and fewer relapse cases than the EC group (Z = 3.124, P = 0.002, .
## Improvement on mental state in mothers
The mental state in mothers was evaluated by SAS and SDS scores as shown in . Compared with the HC group, the mothers in the EC and EM groups showed significantly higher scores of SDS (P = 0.032, P = 0.010) and non-significantly higher scores of SAS (P = 0.158, P = 0.236). At the end of 2month intervention, there were significantly lower scores of SAS and SDS in the EC group (t = 2.087, P = 0.046; t = 2.695, P = 0.011) and EM group (t = 6.066 and 7.972, both P < 0.001). The EM group had significantly lower scores of SAS and SDS than the EC group (P = 0.003, P = 0.002). Intriguingly, the scores of SAS and SDS in EM group were significantly lower at the end of 2-month intervention than those in the HC group (P = 0.031, P = 0.039). The scores of SAS and SDS in the HC group didn't show significant change (t = 0.220, P = 0.827; t = 0.000 P = 1.000). Moreover, compared with the EC group, the ratio of cases with mild-to-moderate depression to normal cases lowered significantly (Z = 2.349, P = 0.019), but the ratio of cases with mild-to-moderate anxiety to normal cases didn't change significantly (Z = 1.016, P = 0.310, . Meanwhile, the correlation of maternal scores of SAS and SDS with infantile EASI scores showed significantly positive before (γ = 0.294, P < 0.019; γ = 0.267, P < 0.035; Figures 4C,D) and after 2month intervention (γ = 0.495, P < 0.001; γ = 0.524, P < 0.001, . In addition, as shown in , maternal sleep state got improved remarkably in the EC and EM groups after 2-month intervention (t = 2.160, P = 0.039; t = 4.923, P < 0.001,). MPIM further improved maternal sleep state (t = 2.468, P = 0.018).
## Safety observation of mpim
During the study, none of any obviously adverse events was observed and reported following MPIM or routine care except for temporal crying during BP only on the first several days in 4 cases in the EM group. Data are presented as the mean ± standard deviation or n/n. HC, healthy control; EC, eczema control; EM, eczema with mother-performed infant massage. , P B = . , P C = . , P D = .
).
## Adherence in routine care or mpim
This study was also designed to observe maternal adherence condition during the latter 3-month intervention without any supervision from investigators. The investigator asked about (1) whether the mothers persisted in following instructed routine care in the EC and EM groups; (2) how often MPIM was conducted in the EM group. It was found that routine care was followed by all mothers in both groups. Nineteen mothers performed MPIM at least 6 days a week, 12 mothers 1 or 2 days a week, and only 1 mother discontinued for no special reason.
# Discussion
This study aimed to observe the potentially positive effect of MPIM on infantile eczema, growth and the mental state in mothers. This study indicated that infantile eczema impaired infantile quality of life and negatively influenced maternal mental state. Infantile eczema improved over time after mothers followed the instructions about the routine care for infantile eczema, along with improved depressive and anxious mood in mothers. More importantly, this study demonstrated that MPIM further enhanced eczema remission and decreased its relapse rate, together with further improved mental state in mothers. However, the growth in infants with eczema was not affected by MPIM. Eczema is typically the first allergic manifestation to appear [bib_ref] Intestinal microbiota in infants at high risk for allergy: effects of prebiotics..., Wopereis [/bib_ref]. Precipitating or aggravating factors of eczema include food allergens, environmental allergens or irritants, climatic condition, stress and genetic predisposition, although the exact cause of eczema is not clear. Infantile emollient is inexpensive, widely available, and used extensively for relieving eczema (46), which can improve the function of skin barrier and reduce itch and irritation. About 35% to 40% of children with moderate to severe eczema have food allergy, and eczema can be improved significantly by eliminating the causative food from their diet (46). Previous studies demonstrated that breastfeeding was associated with lower incidences of allergic diseases, eczema included [bib_ref] Breastfeeding and asthma and allergies: a systematic review and meta-analysis, Lodge [/bib_ref] [bib_ref] Breast-milk characteristics protecting against allergy, Minniti [/bib_ref]. Breast-feeding is proved to prevent allergy due to the immune mediators and oligosaccharides in maternal milk, which facilitates balanced gut microbiota to induce tolerance [bib_ref] Intestinal microbiota in infants at high risk for allergy: effects of prebiotics..., Wopereis [/bib_ref]. In this study, infantile eczema improved after the mothers followed these instructions about the routine care for infantile eczema. Itching, the cardinal symptom of eczema, obviously impairs infantile QOL, therefore, IDQOL is widely used in conjunction with EASI for assessing clinical severity of eczema. This study also demonstrated that QOL of infants with eczema was impaired by eczema and improved along with relieved eczema after the mothers followed the instruction of routine care.
As we all know, postpartum mothers are susceptible to depression and anxiety episodes (9-12), whose mood is negatively influenced by infant eczema. This study showed that the respective prevalence of anxiety and depression among these postpartum mothers was 1/31 and 3/31 in HC group, 3/31 and 5/31 in EC group, 2/32 and 7/32 in EM group at baseline, which are consistent with previous reports [bib_ref] Screening for and treating postpartum depression and psychosis: a cost-effectiveness analysis, Fisher [/bib_ref] [bib_ref] A systematic review and meta-regression of the prevalence and incidence of perinatal..., Woody [/bib_ref]. The mothers in the EC and EM groups had significantly higher levels of depression and anxiety mood than those in the HC group at baseline. Moreover, the mothers in the EC and EM groups had higher rates of depression symptoms than those in the HC group .
/fpubh. .
## Figure
Correlation between EASI score in infant and indexes in infant-mother dyads at baseline and after -month intervention. Baseline/ After -month intervention: (A,E) EASI score and IDQOL score in infants; (B,F) EASI score and sleep score in infants; (C,G) EASI score and SAS score in mothers; (D,H) EASI score and SDS score in mothers. Data are presented as n (%). * * P < 0.01. at baseline, but not higher rates for anxiety symptom. This result indicated that infantile eczema might increase the susceptibility of mothers to depression, which is consistent with previous study. Two previous trials demonstrated that mother-performed massage could relieve eczema symptoms in young children [bib_ref] Evaluation of massage with essential oils on childhood atopic eczema, Anderson [/bib_ref] [bib_ref] Massage therapy for skin conditions in young children, Field [/bib_ref]. One trial showed that the children with eczema were massaged with skin oil on the whole body (except for head and face) by the therapist once a week and by mothers every day for 8 weeks, which decreased night-time disturbance score [bib_ref] Evaluation of massage with essential oils on childhood atopic eczema, Anderson [/bib_ref]. The other one demonstrated that mother-performed massage on the whole body (except for head) for 1 month after the first instruction by therapist could relieve eczema symptoms in children and reduce the anxiety of mothers and children. In China, infant massage is applied based on the theory of TCM, which is performed on the specific areas to relieve eczema in infants and toddlers [bib_ref] 21. He YH, Kang J. Efficacy of Tuina on 120 cases of..., Du [/bib_ref] [bib_ref] Observation on short-and long-term effects of Tuina on eczema among infants and..., He [/bib_ref]. So far, there is no trial to investigate whether mother-performed massage can relieve infantile eczema. It is well-known that the development of eczema is associated with Th2-skewed inflammation, which is closely related with intestinal dysbiosis [bib_ref] Association between the intestinal microbiota and allergic sensitization, eczema, and asthma: a..., Macia [/bib_ref]. Our previous experiments demonstrated that BP, the major manipulation in MPIM, attenuated Th2-skewed inflammation and regulate the intestinal dysbiosis in immature rats with allergic airway inflammation [bib_ref] Effect of chiropractic method on intestinal flora and pulmonary inflammation in asthmatic..., Xiong [/bib_ref]. In TCM, the back is the location where Du vessel and Bladder meridian run, which can be stimulated to regulate visceral function and relieve allergic symptoms [bib_ref] Effect of chiropractic therapy on allergic constitution and intestinal flora in children, Xiong [/bib_ref]. Based on our previous trials and animal experiments, this study was designed to massage the back [bib_ref] Effect of chiropractic therapy on allergic constitution and intestinal flora in children, Xiong [/bib_ref]. This study demonstrated that MPIM further enhanced eczema remission and decreased its relapse rate, along with improved infantile QOL. It is worth mentioning that MPIM remarkedly improved the sleep state of infants with eczema and their mothers, which is consistent with previous studies [bib_ref] Massage therapy by mothers enhances the adjustment of circadian rhythms to the..., Ferber [/bib_ref] [bib_ref] Mothers massaging their newborns with lotion versus no lotion enhances mothers' and..., Field [/bib_ref] [bib_ref] Massage-based bedtime routine: impact on sleep and mood in infants and mothers, Mindell [/bib_ref]. Our previously trial showed mother-performed massage improved the depression and anxiety state of the mothers of asthmatic children. This study indicated that MPIM also significantly reduced the levels of maternal depression and anxiety mood, including obvious depression symptom, which consist with previous trials . Further investigation in this study revealed the positive correlation between the levels of depression and anxiety and the severity of infant eczema. Therefore, on one side, the sleep state of infants got better with eczema improvement, which might beneficially influence the mothers sleep and mood. On the other hand, skin-to-skin contact during MPIM might trigger oxytocin (OT) production and release, which contributed to anti-stress effect and improving sleep state in mothers as previous reports [bib_ref] Natural variations in maternal and paternal care are associated with systematic changes..., Feldman [/bib_ref] [bib_ref] Postpartum maternal oxytocin release by newborns: effects of infant hand massage and..., Matthiesen [/bib_ref]. OT is synthesized and released from the magnocellular neurons of the paraventricular (PVN) and supraoptic nuclei (SON) of the hypothalamus [bib_ref] Neuromodulation by oxytocin and vasopressin in the central nervous system as a..., Stoop [/bib_ref]. OT has positive central effects on psychological adjustment and maternal behaviors during postpartum period [bib_ref] The effects of intranasal oxytocin administration on sensitive caregiving in mothers with..., Mah [/bib_ref] [bib_ref] Female oxytocin-deficient mice display enhanced anxiety-related behavior, Mantella [/bib_ref] [bib_ref] Oxytocin and depression in the perinatal period-a systematic review, Moura [/bib_ref] [bib_ref] Evidence that oxytocin exerts anxiolytic effects via oxytocin receptor expressed in serotonergic..., Yoshida [/bib_ref]. It is also believed that tactile contact between mother and child during massage could reduce the levels of stress-related hormones (cortisol and norepinephrine) in children and mothers and led to relieved eczema and maternal anxiety [bib_ref] Massage therapy for skin conditions in young children, Field [/bib_ref] [bib_ref] Massage reduces anxiety in child and adolescent psychiatric patients, Field [/bib_ref]. The placebo effect could also play a role due to a beneficial expectation [bib_ref] Evaluation of massage with essential oils on childhood atopic eczema, Anderson [/bib_ref].
Few trials investigated the outcomes of both infants and mothers following MPIM. To our knowledge, only one trial did investigate the physical status of preterm infants and the psychological state in the mothers following MPIM, which led to greater weight, motor development, and larger bicep and thigh circumference in infants as well as increased maternal attachment and decreased anxiety compared to the control group. However, our study showed that the growth of the infants with eczema was not affected by eczema during 5-month observation in this study. MPIM didn't significantly enhance the infantile growth although it improved infantile eczema. Previous reports also demonstrated inconsistent results about the effect of MPIM on enhancing infantile growth in preterm infants and healthy infants. Gonzalez [bib_ref] Weight gain in preterm infants following parent-administered Vimala massage: a randomized controlled..., Gonzalez [/bib_ref] and Zhang [bib_ref] Massage intervention for preterm infants by their mothers: a randomized controlled trial, Zhang [/bib_ref] reported that MPIM could enhance the growth of preterm infants while Abedi [bib_ref] The effect of tactile-kinesthetic stimulation on growth indices of healthy neonates, Abedi [/bib_ref] reported that MPIM didn't enhance the growth of healthy neonates. In this study, most of infants had mild eczema and thus their growth might not be affected by eczema, which might explain the result.
In addition, the correlations of EASI score with infantile scores of IDQOL and sleep and maternal scores of SAS and Previous study showed that parents preferred to learn and practice infant massage on their own babies either in a class, in a hospital or at home under the investigators' supervision and instruction . However, for the feasibility and practicality, the adherence of MPIM at home without any specific supervision of investigators should be investigated, which is the minor aim of this study. This study demonstrated good adherence of MPIM without persistent supervision from investigators, which indicated that MPIM is feasible and convenient to implement at home after the relative training in the community background.
# Limitations
There are some limitations in this study. Firstly, for the convenient implementation in this pilot study, it was designed to enroll participants from one community, which might influence the real intervention effects. Secondly, this study was designed not to supervise the implementation of MPIM during the latter 3 months, aiming to observe the feasibility and adherence of MPIM. Therefore, the outcomes of infant-mother dyads at the end of 5-month intervention might be affected by the various performing frequency of MPIM. Thirdly, due to the small sample size, this study didn't analyze the potentially different effects caused by various frequency of MPIM at the end of 5-month intervention. Fourthly, this pilot study only enrolled infants and their mothers to observe the potential effect on maternal mental state. It can also extend to fathers of eczema infants who also experience worse mental state during postpartum. Fifthly, due to the study feature, mothers could not be blinded and they also assessed their own mental state and infantile quality of life, which might bring certain placebo effect. In the future, multi-centered randomized controlled trials with larger size are warranted to further investigate the potential benefits of parent-performed infant massage on the outcomes of infant-parent dyads for a prolonged time.
# Conclusion
In conclusion, this study demonstrated for the first time that MPIM enhanced the remission of infantile eczema, reduced the relapse rate and improved maternal depression and anxiety mood. Given its safety, cost-effectiveness and feasibility, MPIM may be recommended as a routine home healthcare method for infants with eczema in the community background.
# Data availability statement
The original contributions presented in the study are included in the article/supplementary material, further inquiries can be directed to the corresponding author.
# Ethics statement
The studies involving human participants were reviewed and approved by Ethics Committee of Jiangsu Provincial Hospital of Intergrated Chinese and Western Medicine. Number of ethics approval: 2020LWKY010. Written informed consent to participate in this study was provided by the participants' legal guardian/next of kin.
# Author contributions
YX contributed to the conception and design of the study. LY and JL participated in the design of the manuscript and collected data. LL and YX drafted the manuscript. LL and SZ analyzed data. All authors contributed to manuscript revision, read, and approved the submitted version.
# Funding
This study was supported by National Natural Science Foundation of China (Nos. 81973970 and 81774446).
. /fpubh. .
. /fpubh. .
[fig] FIGURE: Comparison of infantile growth indexes at three timepoints. (A) Weight; (B) Length; (C) BMI; (D) Head circumference. No significant di erence in all indexes among the three groups (P A = . [/fig]
[table] TABLE Demographical data: of infant-mother dyads. [/table]
[table] TABLE Comparison of: infantile eczema condition. [/table]
|
Evaluation of anti-cancer and anti-covid-19 properties of cationic pentapeptide Glu-Gln-Arg-Pro-Arg, from rice bran protein and its d-isomer analogs through molecular docking simulations
A B S T R A C TBioactive peptides derived from food proteins are becoming increasingly popular due to the growing awareness of their health-promoting properties. The structure and mechanism of anti-cancer action of pentapeptide Glu-Gln-Arg-Pro-Arg (EQRPR) derived from a rice bran protein are not known. Theoretical and experimental methods were employed to fill this gap. The conformation analysis of the EQRPR pentapeptide was performed first and the obtained lowest energy conformer was optimized. The experimental structural data obtained by FTIR and CD spectroscopies agree well with the theoretical results. D-isomer introduced one-by-one to each position and all D-isomers of the peptide were also examined for its possible anti-proteolytic and activity enhancement properties. The molecular docking revealed avid binding of the pentapeptide to the integrins α 5 β 1 and α IIb β 3 , with K d values of 90 nM and 180 nM, respectively. Moreover, the EQRPR and its D-isomers showed strong binding affinities to apo-and holo-forms of M pro , spike glycoprotein, ACE2, and dACE2. The predicted results indicate that the pentapeptide may significantly inhibit SARS-CoV-2 infection. Thus, the peptide has the potential to be the leading molecule in the drug discovery process as having multifunctional with diverse biological activities.
# Introduction
Rice bran is rich in proteins, lipids, dietary fibers, vitamins and minerals. Components of rice bran show significant biological activities that make it a promising raw material for various medical applications [bib_ref] Health benefits of rice bran -a review, Nagendra Prasad Mn [/bib_ref] [bib_ref] Rice bran: a novel functional ingredient, Sharif [/bib_ref] [bib_ref] Functional properties of dietary fibre prepared from defatted rice bran, Abdul-Hamid [/bib_ref]. Particularly, proteins and small molecular weight peptides derived from rice bran exhibit a wide variety of biological activities, including anti-cancer, anti-inflammatory, anti-diabetic and angiotensin-converting enzyme (ACE) inhibitory properties and may be valuable against certain chronic diseases [bib_ref] Preparation and functional properties of rice bran protein isolate, Wang [/bib_ref] [bib_ref] Rice bran and its main components: potential role in the management of..., Cicero [/bib_ref] [bib_ref] An overview on antioxidant peptides from rice bran proteins: extraction, identification, and..., Zaky [/bib_ref] [bib_ref] Fractionation and characterization of antioxidant peptides from rice bran protein hydrolysates stimulated..., Phongthai [/bib_ref] [bib_ref] Ultrasonication of milky stage rice milk with bioactive peptides from rice bran:..., Ngamsuk [/bib_ref]. Because of having high activity and low toxicity properties, some of these peptides are in clinical trials or are considered to have the great potential to be drugs [bib_ref] Anti-hypertensive effects of peptides derived from rice bran protein, Shobako [/bib_ref] [bib_ref] The anti-cancer activity and potential clinical application of rice bran extracts and..., Yu [/bib_ref].
Significant advances in cancer therapy by short peptides lead to a new class of peptides, namely anti-cancer peptides (ACPs) [bib_ref] Design and modification of anticancer peptides, Hu [/bib_ref]. ACPs are classified into three groups (inhibitor activity, necrosis activity and apoptosis activity) with respect to their biological actions [bib_ref] Design and modification of anticancer peptides, Hu [/bib_ref] [bib_ref] Anti-cancer peptides: classification, mechanism of action, reconstruction and modification: anticancer peptides, Xie [/bib_ref]. The peptides that adhere to the cell surface are linked to the inhibitory activity. For anti-cancer activity, some of ACPs interact with integrins to inhibit migration and metastatic processes of the cancer cells [bib_ref] Design and modification of anticancer peptides, Hu [/bib_ref] [bib_ref] RGD and other recognition sequences for integrins, Ruoslahti [/bib_ref]. Direct interaction of ACPs with cancer cell membrane that destabilizes and/or disrupts the cell membrane stimulates necrotic processes [bib_ref] Studies on anticancer activities of antimicrobial peptides, Hoskin [/bib_ref] [bib_ref] Alpha-helical cationic anticancer peptides: a promising candidate for novel anticancer drugs, Huang [/bib_ref] [bib_ref] Studies on mechanism of action of anticancer peptides by modulation of hydrophobicity..., Huang [/bib_ref]. For apoptosis activity, it is considered that some class of ACPs is internalized inside the cancer cell and then permeates and swells the mitochondria membrane. The result of this action may result in the release Cyt c promoting apoptosis in cancer cells [bib_ref] A proapoptotic peptide for the treatment of solid tumors, Mai [/bib_ref] [bib_ref] Cytochrome c release and apoptosis induced by mitochondrial targeting of nuclear orphan..., Li [/bib_ref]. ACPs inhibit tumor angiogenesis and participate in immune regulation to fight cancer cells by inducing the production of cytokines [bib_ref] Anticancer peptide: physicochemical property, functional aspect and trend in clinical application, Chiangjong [/bib_ref]. The cationic pentapeptide Glu-Gln-Arg-Pro-Arg (EQRPR) derived from rice bran by enzyme hydrolysis has been characterized to be an anti-cancer agent. The peptide shows inhibitory effects on cell proliferation and apoptosis activity in many cancer types [bib_ref] Human cancer cell proliferation inhibition by a pentapeptide isolated and characterized from..., Kannan [/bib_ref] [bib_ref] Rice bran derived pentapeptide-induced apoptosis in human breast cancer cell models (MCF-7..., Li [/bib_ref].
Structural information about the cationic pentapeptide EQRPR as well as the mechanism of action in anti-cancer activity is absent. The most ACPs and other bioactive peptides are cationic with alpha-helical structures, anti-cancer activity of which higher compared to that of beta-pleated, beta-sheet and random coil APCs [bib_ref] Anti-cancer peptides: classification, mechanism of action, reconstruction and modification: anticancer peptides, Xie [/bib_ref]. From sequence consideration, the EQRPR is not likely to have an alpha-helical conformation. Proline (Pro) residue located in the N-terminus can initiate alpha-helix formation. However, in positions distant from the N-terminus, as in EQRPR, Pro residue destabilizes alpha-helical conformation [bib_ref] Positional preference of proline in α-helices, Kim [/bib_ref] [bib_ref] Proline in α-helix: stability and conformation studied by dynamics simulation, Yun [/bib_ref].
We have employed theoretical and experimental methods to elucidate the conformational features of the cationic pentapeptide EQRPR. Advanced molecular modeling indicates that the peptide mainly forms beta-structure and beta-turn conformations in the lowest energy state. Assuming that the peptide may function in the excited states, conformations having energies higher than the lowest energy were also analyzed. Fourier transform infrared (FTIR) and circular dichroism (CD) spectroscopies agree well with the theoretical analysis. Molecular docking was performed to test binding of the cationic pentapeptide EQRPR to integrins α 5 β 1 and α IIb β 3 that may implicate the peptide in anti-proliferation activity for a particular anti-cancer function. High binding affinities of the peptide to the integrins suggest an anti-cancer activity. Docking studies of the peptide with integrin α IIb β 3 and ACE 2 receptor demonstrate that the peptide has a potential against SARS-CoV-2 infection. Altogether, the cationic pentapeptide EQRPR shows multifunctional properties that demand further investigation of this peptide both experimentally and theoretically. D-amino acid substitution in the peptide may increase proteolytic stability and/or biological activity compared to that of the native peptide [bib_ref] Improved protease stability of the antimicrobial peptide pin2 substituted with d-Amino acids, Carmona [/bib_ref] [bib_ref] The unexpected advantages of using Damino acids for peptide self-assembly into nanostructured..., Melchionna [/bib_ref] [bib_ref] Roles of D-amino acids on the bioactivity of host defense peptides, Li [/bib_ref] [bib_ref] Inspiration from the mirror: D-amino acid containing peptides in biomedical approaches, Feng [/bib_ref] [bib_ref] D-amino acids in nature, agriculture and biomedicine, Front, Grishin [/bib_ref]. Therefore, D-isomer modified peptides, D-amino acid substituted one-by-one at each position and all D-isomers were also examined.
# Materials and methods
The peptide EQRPR and its sequential D-isomer substituted and all Disomer peptides were custom synthesized by Synpeptide Co., Ltd, Shanghai, China. Common chemical reagents for the preparation of buffer were purchased from Sigma-Aldrich.
## Experimental studies
## Ftir spectroscopy
The FTIR spectra of peptide solutions were recorded using a BioATR cell of the spectrometer (VERTEX 70v, Bruker, Inc., Germany) with a liquid nitrogen-cooled mercury-cadmium telluride detector. All FTIR spectra were measured with 2 cm − 1 resolution in the spectral range from 400 cm − 1 to 4000 cm − 1 . For each sample, 256 scans were collected. Spectral contributions of water were removed from the sample spectra by spectral subtraction procedure to obtain a flat recording around 2130 cm − 1 associated only with water. After baseline correction spectral regions that cover Amide I and Amide II bands were used to analyze the secondary structure of the peptide solutions. Fitting the spectra to multiple Gaussian components was performed using OriginLab software (OriginLab Corp., Northampton, MA). Because of variations of intensity ratios of Amide I and Amide II bands, there will be some uncertainties in choosing a baseline for Amide I. Therefore, all peptide FTIR spectra (Amide I + Amide II) were subjected to Global analysis where only width and peak positions of Amide II band were global parameters. A certain restriction was applied in multiple Gaussian component fitting of Amide I bands. The band position and width values associated with various secondary structures were allowed to vary in a fixed range [bib_ref] Silk fibroin-based films enhance rhodamine 6G emission in the solid state: a..., Gasymov [/bib_ref] [bib_ref] A robust spectroscopic method for the determination of protein conformational composition -application..., Belton [/bib_ref]. Fixing the number of the peaks, their spectral positions and limiting the peak width range, significantly decreases the number of floating parameters and uncertainties. The application of this methodology in FTIR spectral fitting shows consistent results [bib_ref] Silk fibroin-based films enhance rhodamine 6G emission in the solid state: a..., Gasymov [/bib_ref] [bib_ref] A robust spectroscopic method for the determination of protein conformational composition -application..., Belton [/bib_ref].
The amide III bands of the peptides were further analyzed for the confirmation of the obtained results on the secondary structure. The amide III bands of the FTIR spectra of proteins and peptides cover a spectral region of 1200-1350 cm − 1. The six Gaussian components were sufficient to fit the Amide III spectral regions of the peptides. As in the case of Amide I, band positions and width values were allowed to vary in a fixed range. One component outside of secondary structure elements was added to cover the long wavenumber tail of the spectra.
## Circular dichroism (cd) spectroscopy
The CD spectra of the peptide solutions were recorded using a Chirascan V100 (Applied Photophysics, UK) circular dichroism spectrometer. Far-UV CD measurements were recorded from 190 to 260 nm with a step of 0.5 nm and bandwidth of 1 nm. A path length of a quartz cell was 0.2 mm.
Secondary structure analyses of the peptide solutions from far-UV CD spectra were performed using the DICHROWEB server (http://dichr oweb.cryst.bbk.ac.uk/html/home.shtml) [bib_ref] DICHROWEB, an online server for protein secondary structure analyses from circular dichroism..., Whitmore [/bib_ref] [bib_ref] Protein secondary structure analyses from circular dichroism spectroscopy: methods and reference databases, Whitmore [/bib_ref]. The CDSSTR method with SMP 180 library dataset was used for the calculation of secondary structure content [bib_ref] Analysis of protein circular dichroism spectra for secondary structure using a simple..., Compton [/bib_ref] [bib_ref] Variable selection method improves the prediction of protein secondary structure from circular..., Manavalan [/bib_ref] NRMSD values in the structure analysis of the peptide solutions were less than 0.02.
## Computational studies
## Conformational analysis and structure optimization
The method of molecular mechanics was applied to study the spatial structure and conformational properties of the cationic pentapeptide EQRPR and its sequentially substituted single D-isomer and all D-isomer analogs. shows the amino acid sequences of all L-pentapeptide and its analogs. The application of the method of molecular mechanics permits the determination of a set of possible stable spatial forms of the peptide and its analogs.
The molecular geometry of the L-pentapeptide and its various Disomers were investigated by conformational analysis by using the program written by Ref. [bib_ref] A program for the semiemprical calculation of the conformations of molecular-complexes on..., Maksumov [/bib_ref] , benefiting from the Ramachandran maps [bib_ref] Conformation of polypeptides and proteins, Ramachandran [/bib_ref].
The lowest energy conformations of the pentapeptide and its D-isomers obtained after the conformational analyses were optimized using the Amber force field molecular mechanics method [bib_ref] A second generation force field for the simulation of proteins, nucleic acids,..., Cornell [/bib_ref] available in the Gaussian16 program package. Moreover, for docking studies, the 310 helix, alpha helix, antiparallel beta, parallel beta, beta-sheet, left-hand alpha, and pi helix secondary structures of the pentapeptide EQRPR were formed by altering the dihedral angles. The energies of the obtained structures were obtained by PM3MM calculations.
## Molecular docking
For docking studies the crystal structure of DNA, ACE-2, COVID-19 main protease apo-and holo-forms, spike glycoprotein, α 5 β 1 and α ııb β 3 Amino acid sequence of the anticancer pentapeptide and its single and all Disomer substituted analogs. [DGlu 1 ]-EQRPR DGlu -Gln-Arg-Pro -Arg 3
[DGln 2 ]-EQRPR Glu -DGln -Arg-Pro-Arg 4.
[DArg 3 ]-EQRPR Glu -Gln -DArg -Pro -Arg 5.
[DPro 4 ]-EQRPR Glu-Gln-Arg -DPro -Arg 6.
[DArg 5 ]-EQRPR Glu-Gln-Arg -Pro -DArg 7.
[DX [bib_ref] Health benefits of rice bran -a review, Nagendra Prasad Mn [/bib_ref] [bib_ref] Rice bran: a novel functional ingredient, Sharif [/bib_ref] [bib_ref] Functional properties of dietary fibre prepared from defatted rice bran, Abdul-Hamid [/bib_ref] [bib_ref] Preparation and functional properties of rice bran protein isolate, Wang [/bib_ref] [bib_ref] Rice bran and its main components: potential role in the management of..., Cicero [/bib_ref] ]-EQRPR DGlu -DGln -DArg -DPro -Darg integrins and HSA were obtained from the protein data bank (PDB IDs: 1bna; 6m0j; 6m03; 6lu7; 6vxx; 4wk0; 3zdx; 5z0b, respectively) [bib_ref] Structure of a B-DNA dodecamer: conformation and dynamics, Drew [/bib_ref] [bib_ref] Structure of the SARS-CoV-2 spike receptor-binding domain bound to the ACE2 receptor, Lan [/bib_ref] [bib_ref] The Crystal Structure of COVID-19 Main Protease in Apo Form, Zhang [/bib_ref] [bib_ref] Structure of Mpro from SARS-CoV-2 and discovery of its inhibitors, Jin [/bib_ref] [bib_ref] Structure, function, and antigenicity of the SARS-CoV-2 spike glycoprotein, Walls [/bib_ref] [bib_ref] Metal ion and ligand binding of Integrin α5β1, Xia [/bib_ref] [bib_ref] Complete integrin headpiece opening in eight steps, Zhu [/bib_ref]. Molecular docking simulations were performed by the Autodock Vina program [bib_ref] AutoDock Vina, Improving the speed and accuracy of docking with a new..., Trott [/bib_ref] and binding affinities were calculated. The binding free energies of the most stable ligand-DNA and ligand-protein systems, determined by molecular docking analysis, were calculated by the programs developed by Ref. [bib_ref] FarPPI: a webserver for accurate prediction of protein-ligand binding structures for small-molecule..., Wang [/bib_ref] and the ACFIS 2.0 [bib_ref] ACFIS: a web server for fragment-based drug discovery, Hao [/bib_ref] [bib_ref] Computational discovery of picomolar Q o site inhibitors of cytochrome bc 1..., Hao [/bib_ref] [bib_ref] OpenGrowth: an automated and rational algorithm for finding new protein ligands, Chéron [/bib_ref] [bib_ref] PADFrag: a database built for the exploration of bioactive fragment space for..., Yang [/bib_ref].
The active sites of receptors were screened by using the CAVER program [bib_ref] CAVER Analyst 2.0: analysis and visualization of channels and tunnels in protein..., Jurcik [/bib_ref]. [fig_ref] Glu 1 -: Gln 2 -Arg 3 -Pro 4 -Arg 5 2 [/fig_ref]. The molecular structure of the most stable conformer of EQRPR obtained by optimization with the AMBER force field molecular mechanics method (a) and its wireframe representation (b).
## The toxicological and physicochemical properties prediction of eqrpr pentapeptide
The toxicological risk and physicochemical properties of the EQRPR pentapeptide were obtained using the OSIRIS. The ADMET (absorption, distribution, metabolism, excretion and toxicity) properties of the pentapeptide, from its 2D structure, were predicted using admetSAR [bib_ref] AdmetSAR: a comprehensive source and free tool for assessment of chemical ADMET..., Cheng [/bib_ref].
# Results and discussion
## Theoretical study on the tertiary structure of the pentapeptide eqrpr
The structure of the most stable conformer of EQRPR obtained by optimization by AMBER force field molecular mechanics method embedded in GAUSSIAN 16 program package is shown in [fig_ref] Glu 1 -: Gln 2 -Arg 3 -Pro 4 -Arg 5 2 [/fig_ref]. The dihedral angles of the optimized structures of the EQRPR pentapeptide and various D-isomers are tabulated in (in Supplementary Data). The comparison of the molecular structure of the most stable conformer (I) to those of the molecular models of EQRPR peptide in 310helix (II), alpha-helix (III), antiparallel-beta (IV), beta-sheet (V), lefthand-alpha (VI), parallel-beta, and pi-helix (VII) structures are given in [fig_ref] Glu 1 -: Gln 2 -Arg 3 -Pro 4 -Arg 5 2 [/fig_ref] (in Supplementary Data).
The energies, obtained by PM3MM calculations of the optimized structures of all L-and various D-analogs of EQRPR, together with its non-optimized various conformations were given in , comparatively. As seen in , among all L analogs of EQRPR, the antiparallel-beta plate sheet conformation (− 132.288 kcal/mol) is the second-lowest energy conformation after the most stable conformer (− 183.268 kcal/mol) obtained by optimization with the AMBER force field. Moreover, the D-isomer substitution of the first amino acid gives the lowest energy (− 202.311 kcal/mol), and the D-isomer substitution of the 4th amino acid gives the second lowest energy (− 183.971 kcal/mol), in comparison to all optimized isomers of EQRPR pentapeptide (see . The results revealed that the substitution of the first or fourth amino acid residue with its D-isomer increases the stabilization of the pentapeptide.
## Ftir spectroscopy of the peptide eqrpr
FTIR spectroscopy gives quantitative information regarding the secondary structures of proteins and peptides. IR absorption bands do not depend on the L/D isomeric state of the constituent amino acids of peptides. Therefore, FTIR spectroscopy is best suited to estimate the secondary content of the peptide EQRPR and its D-isomer analogs.
Amide I band components cover the 1700 cm − 1 -1600 cm − 1 region and arise primarily from the C--O stretching vibration with a minor contribution of the NH in-plane bend, the CCN deformation and the outof-phase CN stretching vibration [bib_ref] Fourier transform infrared spectroscopic analysis of protein secondary structures, Kong [/bib_ref] [bib_ref] Infrared spectroscopy of proteins, Barth [/bib_ref]. Amide I vibration exhibits a strong signal by protein and peptides and is sensitive to the secondary structure. In FTIR spectroscopy Amide I band is widely used to estimate secondary structure elements. The frequency of this vibration band depends on the geometry of hydrogen bonding involving the backbone amide group [bib_ref] Fourier transform infrared spectroscopic analysis of protein secondary structures, Kong [/bib_ref] [bib_ref] Infrared spectroscopy of proteins, Barth [/bib_ref]. In the fitting procedure, difficulties arise from the fact that Amide I components responsible for alpha-helix, beta-sheet, beta-turn and random coil conformations are overlapping. The optimal peak parameters established in the work [bib_ref] A robust spectroscopic method for the determination of protein conformational composition -application..., Belton [/bib_ref] were used in the fitting procedure of the Amide I bands of the peptides. This method was successfully tested for proteins having various beta-sheet content.
The FTIR spectra of the peptides and their best fitting curves using the methodology presented in Ref. [bib_ref] A robust spectroscopic method for the determination of protein conformational composition -application..., Belton [/bib_ref] are shown in . Usually, the amide I band of the peptides and proteins are used to calculate their secondary structures. To eliminate possible errors that may be resulted from the uncertainties in baseline levels of amide I, amide I and amide II bands were analyzed together. All FTIR spectra show adequate fit to the multiple Gaussian components with restricted parameters (see the method for details). The secondary structure fractions derived from the fitting procedure are demonstrated in . The common feature for all peptides is almost the lack of alpha-helix structure. This result is consistent with the position of Pro in the peptides, as mentioned above. The peptides show from 36% to 47% secondary structure related to beta-structures. The beta-sheet aggregates (1611 cm − 1 and 1696 cm − 1 ) indicate the possibility of forming the sheet structure from several peptide molecules, particularly in high concentrations. Indeed, to get high-quality FTIR spectra, a high concentration of the peptide (about 0.64 mM) were used. Calculated values for beta-turn structure vary within 28%-39% interval. It is undeniable that the single pentapeptide molecule may not contain the secondary structures derived from the fitting procedure. Instead, data indicate that each peptide molecule samples structures with different secondary structures.
Amide III bands that cover the 1200 cm − 1 -1350 cm − 1 region were also examined to estimate the secondary structure elements of the peptide. Amide III band arises from N-H in-plane bending coupled with C-N stretching. The C-H and N-H deformation vibrations also contribute to the Amide III band. The Amide III vibration exhibits a weak signal and, therefore, is not widely used for the analysis of secondary structure elements of peptides and proteins. Because of being a weak signal, baseline uncertainties in the region of 1200 cm − 1 -1350 cm − 1 influence estimation of the secondary structure. However, in the Amide III region, the overlap of the bands responsible for secondary structure elements is much less compared to that of Amide I. Therefore, the inspections of the Amide III bands together with Amide I bands are valuable to validate the estimation of secondary structure content of proteins and polypeptides.
The FTIR spectra of the peptide in the Amide III region and their best fitting curves to Gaussian components are shown in . The assignment of the peak frequencies and the secondary structure content of the peptides calculated from the fitting procedure are shown in . These data corroborate the result of the Amide I analysis to show that the peptides do not assume alpha-helix conformation and mostly adopt beta-sheet conformation. However, as indicated above the quantitative values for secondary structure content estimated by Amide III bands are less accurate than from Amide I bands.
The secondary structure of the peptides was also investigated by CD spectroscopy.
## Cd spectroscopy of the peptide eqrpr
Far-UV CD spectroscopy is a powerful tool to determine the secondary structure of peptides and proteins [bib_ref] Estimation of protein secondary structure from circular dichroism spectra: comparison of CONTIN,..., Sreerama [/bib_ref]. Far-UV CD spectra of basic secondary structure components (beta-sheet, alpha-helix and random coil) are very distinct. However, the CD spectra of peptides and proteins are different for Land D-isomers. In fact, CD spectra of D-isomers have ellipticities equivalent to that of matching L-isomer but of opposite sign. Therefore, the CD spectrum of each peptide with single D-isomer amino acid will contain a CD band that is a mirror-image of the corresponding L-isomer. The CD spectra of all peptides considered in this The binding affinities (ΔG) of optimized all L-and various D-analogs of EQRPR pentapeptide, docked into the target DNA or proteins in kcal/mol. study are shown in . Although FTIR spectroscopy reveals a similar secondary structure for peptides, CD spectra of the peptides, predominantly determined from the electronic transition in the amide groups, exhibit very different patterns . D-isomers, all-and sequentially substituted single D-isomer, are the origin of the differences in CD spectra. The contribution of D-isomer amino acid to the far-UV CD spectra depends on its position in the sequence. The contribution of the "mirror-image type" spectral component coming from the single
[formula] 1BNA-DNA − 6.2 ¡6.8 − 6.4 − 6.4 − 6.4 − 6.1 − 6.6 3zdx-integrin ¡9.2 − 8.5 − 8.5 − 9.1 − 8.0 − 8.6 − 8.9 4wk0-integrin − 9.6 ¡10.3 − 9.8 − 8.7 − 9.7 − 9.6 − 7.4 O.K. [/formula]
## Fig. 2. a)
Normalized amide I and amide II bands of FTIR spectra of the peptide EQRPR and its D-isomer analogs. Thick black and thin white lines represent experimental and best fitting spectra, respectively. The best fitting spectra are results of eight Gaussian component fitting (See Methods for details). Two Gaussian components around 1730 cm − 1 and 1545 cm − 1 that represent the contribution from side-chains and amid II, respectively, were not considered in the calculation of the secondary structures of the peptides. B) and C) Secondary structure content of the peptide EQRPR and its D-isomer analogs. A and B are data derived from FTIR and CD spectroscopies, respectively.
## Table 3
The binding affinities and the molecular interactions between EQRPR and target DNA, and proteins, obtained by molecular docking simulations a .
Ligand/Target DNA (1BNA)
[formula] α 5 β 1 (4WK0) α IIB β 3 (3ZDX) apo-form (6M03) holo- form spike glycoprotein (6VXX) ACE2 Truncated ACE2 dACE2 Albumin (6LU7) (6M0J) (5Z0B) EQRPR (all L-) DG2 PHE21 LEU23 THR24 THR26 THR(B)912 PHE40 ASN599 TYR454 ARG114 DG4 SER22 ASP24 LEU141 HIS41 ARG(C)1091 TRP69 LYS600 ARG482 LEU115 DA5 VAL23 PHE25 ASN142 TYR54 GLU(A)1092 ALA348 ASN601 GLU483 TYR138 DA6 ARG106 ALA95 SER144 PHE140 GLU(B)1092 ASP350 SER602 ILE484 GLU141 DT7 PHE168 SER96 CYS145 ASN142 GLY(B)1093 ASP382 PHE603 VAL485 HIS146 DT20 GLU171 SER99 HIS163 SER144 ARG(B)1107 LEU391 LYS676 ASN599 LYS190 DC21 SER234 PHE171 GLN189 CYS145 GLN(B)1113 ASN394 ARG678 ASN601 LYS519 DG22 VAL235 SER173 HIS163 ASN(A)1119 HIS401 ASN682 GLY605 DC23 LYS269 TYR237 GLU166 ASN(B)1119 PHE684 SER680 TYR287 SER238 GLN189 TYR781 THR803 LEU355 ARG261 SER804 SER417 HIS291 ARG420 VAL293 ILE360 SER420 LEU421 ARG422 Binding affinity (kcal/mol) − 6.2 − 9.6 − 9.2 − 6.6 − 7.5 − 8.0 − 7.6 − 6.7 − 8.0 − 8.4 [/formula]
a Truncated ACE2 target is the optimized truncated protein that lacks 356 amino acids from the N-terminal of ACE2 and has 449 amino acids. D-isomer in peptides is very difficult to estimate. However, the CD spectrum of the peptide with all D-isomer can be used for calculation by taking the mirror-image of the CD spectrum. The mirror-image of mirror-imaged spectra of D-isomer will produce an L-isomer type CD spectrum. The CD spectra of the peptide EQRPR and its all D-isomer analog are demonstrated in . It is evident that the mirror-image of the CD spectrum of all D-isomer analog produces a spectrum very similar to that of all L-isomer peptide. CDSSTR derived curves of the peptides are consistent with the corresponding experimental spectrum. The results of the calculation are shown in . The lack of the alpha-helix fraction obtained from CD spectra is very consistent with the results of FTIR analysis. Thus, secondary structure content obtained by analysis of CD spectra shows that the peptide mainly samples beta-sheet, beta-turn and random coil conformations. Despite the qualitative agreement between the results of the two methods, there are some quantitative discrepancies in the estimation of secondary structure fractions. Estimation of secondary structure content from CD spectra is based on the use of various protein libraries. The accuracy of the estimate is high if the protein library contains data of proteins that are structurally closely related to examined one. Currently, the short peptides are not well represented in the protein library data. Therefore, it is reasonable to speculate that the secondary structure estimate from CD data is affected by lack of the data for short peptides in the protein library. The most cationic anti-cancer peptides possess alpha-helix structure. Therefore, the secondary structure of the peptides was investigated in 2,2,2-trifluoroethanol (TFE) to reveal a propensity for alpha-helix formation. It is well established that TFE induces alpha-helix formation in peptides and proteins [bib_ref] Mechanism of helix induction by trifluoroethanol: a framework for extrapolating the helix-forming..., Luo [/bib_ref] [bib_ref] Quantitative determination of helical propensities from trifluoroethanol titration curves, Jasanoff [/bib_ref]. TFE titration of the peptide has been used to determine the hidden propensity for alpha-helix formation of small peptides [bib_ref] Mechanism of helix induction by trifluoroethanol: a framework for extrapolating the helix-forming..., Luo [/bib_ref]. In all cases, alpha-helix formations in peptides are saturated at 50% (v/v) TFE. Far-UV CD spectra of the peptide in the buffer with and without 50% TFE are shown in . Contrary to the most expected transition induced by TFE, the difference CD spectrum reveals beta-sheet formation for a small portion of the peptide. Thus, theoretical and experimental data indicate a lack of alpha-helix structure in the peptide EQRPR. Moreover, this peptide does not have any hidden propensity for alpha-helix formation.
## In silico molecular docking analysis of eqrpr as anticancer agent
To evaluate the anticancer activity of EQRPR pentapeptide, molecular docking studies were performed and the interactions of EQRPR ligand with DNA and target proteins α 5 β 1 and α IIb β 3 integrins were revealed.
For the docking simulations of EQRPR pentapeptide with DNA, the crystal structure of DNA was obtained from the protein data bank (PDB ID: 1BNA) [bib_ref] Structure of a B-DNA dodecamer: conformation and dynamics, Drew [/bib_ref] , and, was prepared for the docking by removing water molecules in DNA and adding polar hydrogens to it. The optimized structure of EQRPR was adapted for the docking. The partial charges of the EQRPR molecule were calculated using the Geistenger method. The active site of DNA was defined within the grid size of 40Å × 40Å × 40 Å.
The EQRPR pentapeptide is found to interact with DG2, DG4, DA5, DA6, DC21, DG22, DC23 and DT20 residues of DNA via the intermolecular hydrogen bonds. These hydrogen bonds are between DG2 and Glu (2.59 Å); DG4 and Glu (two hydrogen bonds; 2.43 ve 2.82 Å); DG4 and Gln (1.99 Å); DA5 and Gln (3.48 Å); DA6 and Arg 5 (3.40 Å); DC21 and Arg 5 (2.54 Å); DG22 and Glu (3.08 and 2.78 Å); DG22 and Gln (2.53 Å); DG22 and Pro (3.40 Å); DC23 and Gln (2.01 Å); DT20 and Arg 5 (2.09 ve 2.74 Å). Moreover, there is an attractive charge interaction between DA5 and Glu residue of the pentapeptide, at 4.74 Å length, and an unfavorable negative-negative interaction between DT7 and Arg5 residue at 3.98 Å length. The result shows that the binding affinity (ΔG) is − 6.2 kcal/mol (See [fig_ref] Figure 4: Molecular docking results of EQRPR bound to DNA [/fig_ref]. The docking simulations of D-isomer analogs of the EQRPR pentapeptide into DNA were also investigated, and it was revealed that the pentapeptide, all residues of which were D-amino acids, showed the highest binding affinity to DNA (− 6.8 kcal/mol) (see .
The anti-proliferation effect of the EQRPR pentapeptide for the anticancer function was revealed by molecular docking studies of the pentapeptide into target proteins α5β1 and αIIbβ3 integrins. After the α5β1 integrin (PDB ID: 4WK0) was prepared for molecular docking, the binding probability of EQRPR pentapeptide in the two different active sites was examined. The molecular model of α5β1 integrin was shown in . The docking results were also evaluated by sorting the docked conformations of EQRPR according to their predicated binding free energy. The best predicted binding affinity for DNA was found for the pentapeptide (− 6.8 kcal/ mol), all of which are D isomers. The best predicted binding affinities for DNA and for α 5 β 1 integrin, with − 6.8 kcal/mol and − 10.3 kcal/mol, respectively, were found for the pentapeptide, which has all D isomers. However, the EQRPR pentapeptide that has all L-isomers is found to have the best binding affinity for α IIB β 3 integrin (− 9.2 kcal/mol).
The high binding affinity between the investigated molecule and the integrins prompted us to investigate the relationship between binding energy and conformation of the pentapeptide. shows the binding affinities and binding free energies, calculated by Poisson Boltzmann ΔG (PB), and generalized Born ΔG (GB) approximations of EQPRP pentapeptide in various conformations. As can be seen from , the pentapeptide in the 310 helix conformation has the highest binding affinity (− 10.7 kcal/mol) to the α 5 β 1 integrin (in the active site 2), followed by parallel beta (− 9.6 kcal/mol) and optimized structure (− 9.6 kcal/mol). Despite the highest affinity to the integrin α IIB β 3 , the peptide does not assume the 310 helix structure that has high unfavorable energy (see . On the other hand, the binding affinity of the pentapeptide in the optimized structure has the highest binding affinity to α IIB β 3 integrin (− 9.2 kcal/mol) followed by the antiparallel beta structure of the EQRPR (-9.1 kcal/mol).
## In silico molecular docking analysis of eqrpr as anticovid-19 agent
In this study, we performed docking-based virtual screening on three SARS-CoV-2 targets (the proteases M pro and the spike glycoprotein) using investigated pentapeptide to suggest as a potential compound that may act as antivirals. SARS-CoV-2 that causes COVID-19 infection uses angiotensin-converting enzyme 2 (ACE2) for entry into target cells. For this reason, the interaction between the investigated pentapeptide and ACE2 enzyme was also investigated.
The crystal structures of spike glycoprotein (PDB ID: 6VXX), apoform of COVID-19 M pro (PDB ID: 6M03) and holo-form of COVID-19 M pro (PDB ID: 6LU7), ACE2 (PDB ID: 6M0J) were obtained from the protein database [bib_ref] Structure of the SARS-CoV-2 spike receptor-binding domain bound to the ACE2 receptor, Lan [/bib_ref] [bib_ref] The Crystal Structure of COVID-19 Main Protease in Apo Form, Zhang [/bib_ref] [bib_ref] Structure of Mpro from SARS-CoV-2 and discovery of its inhibitors, Jin [/bib_ref] [bib_ref] Structure, function, and antigenicity of the SARS-CoV-2 spike glycoprotein, Walls [/bib_ref]. The docking for molecules was adapted by removing the water molecule from the receptors and adding polar hydrogens. The active sites of the target proteins were defined in the grid size of 40 Å × 40 Å × 40 Å.
It was determined that the investigated pentapeptide bound to the spike glycoprotein, most strongly; the binding affinities of all L-and all D-isomers analogs of EQRPR for spike glycoprotein are − 8.0 and − 7.9 kcal/mol, respectively. This is followed by ACE2: The binding affinities of all L-and all D-isomer analogs of EQRPR for ACE2 are − 7.6 and − 7.5 kcal/mol, respectively. The binding affinities of L-and all D-isomer analogs of EQRPR for the holo-and apo-forms of COVID-19 M pro are found to be − 7.5 and − 7.5 kcal/mol, and − 6.6 and − 6.9 kcal/mol, respectively. Figs. 6 and 7 shows the molecular docking results of EQRPR to target proteins (Apo and holo forms of M pro , Spike glycoprotein and ACE2). gives the binding affinities (kcal/mol) of the optimized structures of all L-and all D-isomers of EQRPR with the target proteins and DNA.
In the study conducted by Onabajo et al., in 2020, it was proposed that a truncated isoform of ACE2, which was designated as deltaACE2 (dACE2) acts as an interferon-stimulated gene. Onabajo et al. proposed that the dACE2, which lacks 356 amino acids from the N-terminal of ACE2, was non-functional in binding the SARS-CoV-2 spike protein and as a carboxypeptidase. In the study of Onabajo et al. [bib_ref] Interferons and viruses induce a novel truncated ACE2 isoform and not the..., Onabajo [/bib_ref] , not ACE2, but dACE2 has been reported to be an interferon-stimulated gene. Since the interferon-induced variability in ACE2 expression levels is thought important for susceptibility to COVID-19, dACE2 has importance. For this reason, in this study, the interaction between EQRPR and dACE2 is also investigated by docking simulations. The dACE2 is a novel inducible isoform of ACE2 [bib_ref] Interferons and viruses induce a novel truncated ACE2 isoform and not the..., Onabajo [/bib_ref] , so its crystal structure is not available in the database. But its structure for docking purposes is prepared as described [bib_ref] Interferons and viruses induce a novel truncated ACE2 isoform and not the..., Onabajo [/bib_ref]. The full-length ACE2 protein, with 805 amino acids was taken from the study conducted by Ref.and by using CABS-flex 2.0 online server, the first 356 amino acids from the N-terminal were taken out. Since, by Onabajo et al. it was proposed that the remaining part, which lacks 356 amino acids from the N-terminal of ACE2, was non-functional in binding the SARS-CoV-2 spike protein and as a carboxypeptidase, we also investigated this truncated ACE2. The remaining truncated protein (which lacks 356 amino acids from the N-terminal of ACE2), with 449 amino acids, was then optimized and prepared for docking with EQRPR. The docking results are shown in [fig_ref] Figure 8: The 3D docked views of the most stable conformer of EQRPR in... [/fig_ref]. Binding affinity is found to be O.K. reduced to − 6.7 kcal/mol. It must be noted that the binding affinity of EQRPR to ACE2 was − 7.6 kcal/mol (See [fig_ref] Figure 7: The 3D docked views of the most stable conformer of EQRPR in... [/fig_ref]. Then dACE2 was prepared, by adding ten amino acids (MREAGWDKGG) to the N terminal of the remained truncated ACE2, which lacks 356 amino acids from its N terminal. In other words, to prepare dACE2, the first 356 amino acids of ACE2 were replaced by 10 amino acids in the MREAGWDKGG sequence and then optimized by AMBER. The docking of the optimized dACE2 with EQRPR pentapeptide resulted in − 8.0 kcal/mol binding affinity (K d = 1.4 μM). The molecular docking of EQRPR with dACE2 and the interaction diagrams are shown in [fig_ref] Figure 8: The 3D docked views of the most stable conformer of EQRPR in... [/fig_ref]. This preliminary result reveals the antiCOVID-19 activity of the EQRPR molecule.
As a summary, the binding affinities and interactions of EQRPR ligand-target complexes were tabulated in .
One of the difficulties in the biomedical application of the natural or derived short peptides is a rapid renal clearance in part owing to the low molecular weight. Since the peptides do not show a renal tubular reabsorption property, they have a high clearance rate and short halflife [bib_ref] Strategic approaches to optimizing peptide ADME properties, Di [/bib_ref]. Several approaches have been used to increase the half-life of the peptide. However, binding to the plasma proteins is a well-known mechanism to retard renal clearance. To decide what approach is the most practical for biomedical application, in silico evaluation was performed for binding of the EQRPR to the human serum albumin.
## Molecular docking analysis of eqrpr with plasma-derived human serum albumin (hsa)
The crystal structure of plasma-derived HSA was taken from the Protein Data Bank (PDB ID: 5Z0B) and, was confirmed to the docking by removing water molecules and adding polar hydrogens in it. The active sites of HSA were searched by the CAVER program [bib_ref] CAVER Analyst 2.0: analysis and visualization of channels and tunnels in protein..., Jurcik [/bib_ref]. As a result of the docking simulations of EQRPR pentapeptide with HSA, binding affinity is found as − 8.4 kcal/mol (K d = 0.7 μM). The Albumin-EQRPR interaction is given in [fig_ref] Figure 9: The 3D docked view of the most stable conformer of EQRPR in... [/fig_ref]. The molecular interaction between EQRPR and HSA (5Z0B) is given in . High-affinity binding to HSA suggests that as the first approach EQRPR can be used in vitro without any modification and expect to get similar results obtained in vivo. A webserver "SuperCYPsPred" predicts that the modifications of the peptide by cytochrome P450 enzymes are unlikely [bib_ref] SuperCYPsPred-a web server for the prediction of cytochrome activity, Banerjee [/bib_ref].
## Back to covid-19 and cancer
Possible biomedical application of the peptide EQRPR without any modification makes it more valuable for COVID-19 treatment. Evidencebased data suggest neuronal dissemination of SARS-CoV-2 invasion into the central nervous system (CNS) [bib_ref] Neurological consequences of COVID-19: what have we learned and where do we..., Jarrahi [/bib_ref]. SARS-CoV-2 RNA was detected in the cerebrospinal fluid (CSF) in the patient diagnosed with viral meningitis [bib_ref] A first case of meningitis/encephalitis associated with SARS-Coronavirus-2, Moriguchi [/bib_ref]. It was suggested that COVID-19 may enter the CNS via retrograde neuronal diffusion and then, infection swiftly spread to other brain regions by trans-neuronal path [bib_ref] SARS-CoV-2: olfaction, brain infection, and the urgent need for clinical samples allowing..., Butowt [/bib_ref]. Consistent with that SARS-CoV particles were observed in CNS neurons and brain samples from patients diagnosed with SARS [bib_ref] Detection of severe acute respiratory syndrome coronavirus in the brain: potential role..., Xu [/bib_ref]. ACE2 was identified in both neurons and glial cells [bib_ref] Angiotensin II regulates ACE and ACE2 in neurons through p38 mitogen-activated protein..., Xiao [/bib_ref]. As in the lung, ACE2 in the brain also provides viral entry for SARS-CoV-2. ADMET profile predicts that the peptide EQRPR having a molecular weight less than 1000 Da is permeable to the blood-brain barrier (BBB). Data imply that the peptide EQRPR may also effectively work in the brain against the viral infection.
Integrins play a significant role in tumor invasion and progression in glioblastomas [bib_ref] Interest of integrins targeting in glioblastoma according to tumor heterogeneity and cancer..., Malric [/bib_ref]. Therefore bioavailability of the peptide EQRPR in the brain environment could be beneficial for brain cancer treatment.
Despite the significant success in chemo-and radio-therapy, rare subpopulations of cancer cells that show chemo-, radio-and immunoresistance capacity are major obstacles to effectively treat cancer patients. This population of cancer cells identified as cancer stem cells is believed to be the source of relapses after intensive therapies [bib_ref] Side population and cancer stem cells: therapeutic implications, Moserle [/bib_ref] [bib_ref] The critical, clinical role of interferon-beta in regulating cancer stem cell properties..., Doherty [/bib_ref]. Varies approaches have been developed to increase the susceptibility of cancer stem cells to various therapies. Interferons with strong antiviral properties have also been effectively used in cancer treatment. One of the many effects of interferon actions is to force cancer stem cells to differentiate. Loss of stem cell like properties upon differentiation increases the effectiveness of radio-and chemotherapy. The peptide EQRPR has immense potential to be effective with combinatorial use with interferon. As indicated above interferon also stimulate the production of dACE2 receptors on epithelial cells including cancer cells. Increased expression of dACE2 was identified in squamous tumors of the respiratory, gastrointestinal and urogenital tracts [bib_ref] Interferons and viruses induce a novel truncated ACE2 isoform and not the..., Onabajo [/bib_ref]. Effective interaction of the peptide EQRPR with dACE2 could be very beneficial to use it in combination with interferon in cancer treatments.
Altogether, comprehensive in silico analysis reveals that the peptide has the potential to be the leading molecule in the drug discovery process for multiple therapeutic applications involving multiple organs.
## Drug-likeness prediction studies of eqrpr pentapeptide
Considering the potential use of the studied pentapeptide as a drug, it is important to estimate the Absorption, Distribution, Metabolism, Excretion and Toxicity (ADMET) properties of EQRPR in humans. The drug-likeness properties of EQRPR pentapeptide were calculated using OSIRISand the results were summarized in . The toxicity risk predictor shows that the compound does not present risks of mutagenicity, tumorigenicity, irritant or reproductive effects. A significant determinant of the pharmacokinetic behavior of a drug is lipophilicity. LogP {logP value = log (C octanol /C water )}) is a measure of the lipophilicity of a compound. The EQRPR pentapeptide shows low lipophilicity, which reflects a good aqueous solubility.
The silico module admetSAR [bib_ref] AdmetSAR: a comprehensive source and free tool for assessment of chemical ADMET..., Cheng [/bib_ref] has been used to characterize the various ADMET parameters for EQRPR pentapeptide and the results are given in . The expected probabilities are given between 0 and 1 as a numerical value.
The blood brain barrier (BBB) is the brain's microvascular endothelial cell layer which plays a very significant role in the separation of blood from the brain. While most central nervous system (CNS) targeting drugs require high penetration, BBB penetration should be reduced for non-CNS drugs to prevent unintended effects. The BBB permeability of EQRPR pentapeptide is found to be positive with 0.9448 probability value, which has importance for acting as an antiCOVID-19 agent.
The cytochrome P450 (CYP) enzyme is the most effective enzyme in drug metabolism. The probability value of the CYP3A4 substrate for EQRPR is estimated as 0.6724, which means that it can be metabolized in the liver, as Cytochrome P450s is an essential enzyme system for liver drug metabolism. In addition, when we look at the estimation of outflow by P-glycoprotein (P-gp), EQRPR is a P -glycoprotein's substrate with a probability of 0.7122 and its inhibitor with a probability of 0.6470.
Thyroid hormone receptors regulate gene expression by binding to DNA through (HREs) [bib_ref] Thyroid hormone receptors and resistance to thyroid hormone disorders, Ortiga-Carvalho [/bib_ref]. When activated by estrogen, the Estrogen Receptor (ER) can settle in the nucleus and bind to DNA (i.e., it is a DNA binding transcription factor) to regulate the activity of different genes. The primary function of the androgen receptor is to control gene expression and activate the DNA-binding transcription factor by binding to some of the androgenic hormones in the cytoplasm, including testosterone and dihydrotestosterone [bib_ref] Regulation of androgen action, Roy [/bib_ref]. The probability values of Estrogen receptor binding, Androgen receptor binding and Thyroid hormone receptor for EQRPR were found to be 0.6846, 0.6182 and 0.5400 respectively.
# Conclusions
The FTIR and CD experimental results and theoretical conformational analysis presented here establish that the cationic pentapeptide EQRPR derived from rice bran and its D-isomer analogs mostly adopt beta-sheet conformations. Molecular docking studies provided the multiple ways by which the peptide may act as anti-cancer as well as an anti-viral agent. The results reveal strong interaction between the pentapeptides and integrins (α 5 β 1 and α IIb β 3 ) important for anti-cancer activity.
The present study demonstrates the mechanism of binding of the pentapeptides to apo-and holo-forms of M pro , spike glycoprotein, ACE2 and dACE2. Data indicate that the peptide can use multiple ways to prevent SARSCoV-2 (COVID- [bib_ref] Anticancer peptide: physicochemical property, functional aspect and trend in clinical application, Chiangjong [/bib_ref] infection. The docking studies show the effectiveness of the use of the pentapeptides in combination with interferons. Furthermore, the present study highlights the importance of Disomer substitutions in the peptides to control selected functions.
The important conclusion from this study is that the examined pentapeptide is not predicted. Our recent experimental data support some of our findings. The peptides have been found incorporated into the phospholipid monolayer (DPPC) and the membranes of human lung cancer cells. The results of the studies will be published elsewhere.
# Funding
This study is supported by TUBITAK-ANAS international project (Project numbers TUBITAK 118F445 and ANAS PH05-01).
## Declaration of competing interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
[fig] Glu 1 -: Gln 2 -Arg 3 -Pro 4 -Arg 5 2. [/fig]
[fig] Figure 3S: The circles 1 and 2 shown in Fig. 3S have roughly indicated the two active sites of α5β1. The docking simulations of EQRPR to the α5β1 integrin were performed for both active sites. The most effective binding was found in the active site 2 of α5β1 integrin with − 9.6 kcal/mol binding energy. Fig. 5 a-c shows the 3D docked view of EQRPR in theFig. 3. A) Far-UV CD spectra of the peptide EQRPR and derived curves calculated by CDSSTR method. Positive and negative sign dashed black curves represent the peptides with all D-isomer and its mirror-image, respectively. Solid black line is CD spectrum of all L-isomer peptide. CDSSTR derived curves of the peptides conform well to the respective experimental spectra. B) Secondary structure transition in the peptide EQRPR induced by TFE. Far-UV CD spectra of the peptide in buffer pH 7.3 (open cirles), in 50% (v/ v) TFE (solid circles) and their difference (open triangles). The minimum observed around 218 nm in the difference spectrum indicates betasheet formation. active site 2 of α5β1 integrin. The interaction diagrams of EQRPR-α5β1 integrin complex are also indicated in the figure. The molecular docking of α IIB β 3 integrin (PDB ID: 3ZDX) with EQRPR resulted in a binding affinity of − 9.2 kcal/mol. The 3D docked view of EQRPR in α IIB β 3 integrin and the estimated interactions are shown in Fig. 5 d-f. The effect of the D-isomer substitutions on the binding free energy of pentapeptide-DNA and pentapeptide-integrin interaction complexes was evaluated by docking simulations using the most stable conformers of all-L, all-D and D-amino acid substituted one-by-one at each position conformers. The results are tabulated in [/fig]
[fig] Figure 4: Molecular docking results of EQRPR bound to DNA (a), the doted lines present the interactions of EQRPR with the nucleic acids of DNA (b).O.K.Gasymov et al. [/fig]
[fig] Figure 5: The 3D docked views of the most stable conformer of EQRPR in active site 2 of α 5 β 1 integrin (a-c) and in α IIB β 3 integrin (d-f). The ligand interaction diagrams of receptor ligand complexes are show. O.K. Gasymov et al. [/fig]
[fig] Figure 6: The 3D docked views of the most stable conformer of EQRPR in Apo (a) and Holo (b) forms of M pro . The ligand interaction diagrams of receptor-ligand complexes are shown. [/fig]
[fig] Figure 7: The 3D docked views of the most stable conformer of EQRPR in the Spike glycoprotein (a), in ACE2 (b). The ligand interaction diagrams of receptor-ligand complexes are shown.O.K.Gasymov et al. [/fig]
[fig] Figure 8: The 3D docked views of the most stable conformer of EQRPR in optimized truncated ACE2, which lacks 356 amino acids from the N-terminal of ACE2 (a) and in dACE2 (b). The ligand interaction diagrams of receptor-ligand complexes are shown. [/fig]
[fig] Figure 9: The 3D docked view of the most stable conformer of EQRPR in albumin. The ligand interaction diagrams of receptor ligand complex are show. O.K. Gasymov et al. [/fig]
|
Durable Response of Human Epidermal Growth Factor Receptor-2-Positive Gastric Adenosquamous Carcinoma to Trastuzumab-Based Chemotherapy
# Introduction
Adenosquamous carcinoma (ASC) of the stomach is a rare malignancy, accounting for <1% of all gastric carcinomas [bib_ref] p53, p16 and RB expression in adenosquamous and squamous cell carcinomas of..., Lee [/bib_ref] [bib_ref] A clinicopathological and immunohistochemical study of gastric cancer with squamous cell carcinoma..., Saito [/bib_ref] [bib_ref] The clinicopathologic and prognostic analysis of adenosquamous and squamous cell carcinoma of..., Quan [/bib_ref] [bib_ref] Squamous cell carcinoma and adenoacanthoma of the stomach. A clinicopathologic study, Boswell [/bib_ref] , and is characterized by a mixture of two components: adenocarcinoma and squamous cell carcinoma. In the adenocarcinoma component, the differentiated type is more frequently observed than the poorly differentiated type [bib_ref] p53, p16 and RB expression in adenosquamous and squamous cell carcinomas of..., Lee [/bib_ref] [bib_ref] A clinicopathological and immunohistochemical study of gastric cancer with squamous cell carcinoma..., Saito [/bib_ref] [bib_ref] Adenosquamous carcinoma of the stomach. A clinicopathologic analysis of 28 cases, Mori [/bib_ref] , and the clinicopathological features of this tumor are dependent on the histological type of the adenocarcinoma component [bib_ref] Adenosquamous carcinoma of the stomach. A clinicopathologic analysis of 28 cases, Mori [/bib_ref]. On the other hand, lymph node and liver metastases are frequently observed in ASC, and these metastases also tend to be found in advanced stages at diagnosis [bib_ref] p53, p16 and RB expression in adenosquamous and squamous cell carcinomas of..., Lee [/bib_ref] [bib_ref] A clinicopathological and immunohistochemical study of gastric cancer with squamous cell carcinoma..., Saito [/bib_ref] [bib_ref] The clinicopathologic and prognostic analysis of adenosquamous and squamous cell carcinoma of..., Quan [/bib_ref] [bib_ref] Squamous cell carcinoma and adenoacanthoma of the stomach. A clinicopathologic study, Boswell [/bib_ref] [bib_ref] Adenosquamous carcinoma of the stomach. A clinicopathologic analysis of 28 cases, Mori [/bib_ref] [bib_ref] Adenosquamous carcinoma of the remnant stomach: report of a case, Mori [/bib_ref] [bib_ref] Early gastric cancer of adenosquamous carcinoma type: report of a case and..., Yoshida [/bib_ref] ; hence, this malignancy seems to a have a poorer prognosis than conventional gastric adenocarcinoma [bib_ref] p53, p16 and RB expression in adenosquamous and squamous cell carcinomas of..., Lee [/bib_ref] [bib_ref] Adenosquamous carcinoma of the stomach. A clinicopathologic analysis of 28 cases, Mori [/bib_ref].
Human epidermal growth factor receptor-2 (HER2), a membrane-associated receptor with an intracellular tyrosine kinase domain, dimerizes with other HER family members and mediates signal transduction pathways linked to cell growth and survival. HER2 is overexpressed in approximately 20% of gastric adenocarcinoma cases and is considered a key therapeutic target for HER2-positive gastric adenocarcinoma. In the Trastuzumab for Gastric Cancer (ToGA) trial, trastuzumab, a monoclonal antibody targeting HER2, significantly improved overall survival when combined with chemotherapy in the treatment of advanced HER2-positive gastric adenocarcinoma [bib_ref] ToGA Trial Investigators: Trastuzumab in combination with chemotherapy versus chemotherapy alone for..., Bang [/bib_ref]. However, to date, there are no reports regarding immunoreactivity for HER2 and its role as a therapeutic target in gastric ASC.
Here, we report a patient with gastric ASC with HER2 overexpression who achieved a dramatic durable response with trastuzumab-based chemotherapy, presumably because of strong HER2 immunoreactivity in the tumor. The findings of the present report warrant further investigation to clarify the role of HER2 in this rare tumor.
## Case report
A 66-year-old man was referred to our hospital with a diagnosis of gastric adenocarcinoma. Upper gastrointestinal endoscopy revealed type 3 advanced gastric cancer of the lesser curvature of the upper to lower gastric body , and histological analysis of biopsy samples showed poorly-differentiated adenocarcinoma. Serum concentrations of carcinoembryonic antigen (CEA) and carbohydrate antigen 19-9 (CA19-9) were highly elevated (86.8 ng/ml and 6,970 U/ml, respectively). An abdominal CT scan revealed thickening of the gastric wall and enlargement of the abdominal lymph nodes, but no evidence of distant metastasis. A total gastrectomy with lymph node dissection (D2) was performed with a curative intent. In the postoperative period, the patient developed a pancreatic fistula with a subphrenic abscess, which was subsequently drained.
A macroscopic examination revealed a type 1 protruded lesion occupying the lesser curvature of the gastric body, measuring 87 × 82 mm at its largest dimensions, which invaded beyond the serosal layer. A pathological diagnosis of the resected ASC specimens showed that it was composed of a mixture of adenocarcinoma and squamous cell carcinoma components . Both components were moderately differentiated; the adenocarcinoma component showed tubular structures and the squamous cell carcinoma component showed partly differentiated keratinizing cells. The individual components constituted nests adjacent to one another. Venous and lymphatic invasion were identified (v3 and ly2, respectively). Metastatic lesions were identified in 6 of 49 dissected lymph nodes, in which only an adenocarcinoma component was observed. According to the 14th edition of the Japanese Classification of Gastric Carcinoma, the cancer was graded as T4aN2M0 (stage IIIB). Immunohistochemical staining for the membrane-bound HER2 protein using a polyclonal antibody (HercepTest TM ; Dako A/S, Glostrup, Denmark) showed that both the squamous carcinoma and adenocarcinoma components of the primary tumor were strongly positive (score, 3+) for HER2 . A high level of HER2 gene amplification in both components was also detected by dual-color chromogenic in situ hybridization (INFORM HER2 Dual ISH; Ventana Medical Systems Inc., Oro Valley, Ariz., USA) . About 2.5 months after curative surgery, an abdominal CT scan showed multiple metastatic lesions in the bilateral lobes of the liver . The serum concentrations of CEA and CA19-9 increased to 669.6 ng/ml and 7,380 U/ml, respectively. Systemic treatment with capecitabine (2,000 mg/m 2 , days 1-14), cisplatin (80 mg/m 2 , day 1), and trastuzumab (8 mg/kg dose-loading followed by 6 mg/kg, day 1) was initiated every 21 days. After 3 treatment cycles, all metastatic liver lesions had regressed and a partial response (PR) was confirmed after 6 cycles . After confirmation of a PR, the patient was maintained on combination chemotherapy with capecitabine and trastuzumab. The levels of all tumor markers were within normal limits 9 months after chemotherapy initiation. PET and CT scans performed 13 months after chemotherapy initiation revealed no uptake in the metastatic liver lesions . Thereafter, maintenance therapy with trastuzumab alone was started and the PR was maintained for 16 months after initiation of chemotherapy.
# Discussion
To our knowledge, this is the first report of HER2-positive ASC successfully treated with trastuzumab-based combination chemotherapy. Several hypotheses for the origin of a squamous cell carcinoma component in ASC have been proposed: (1) squamous metaplasia of a pre-existing adenocarcinoma [1, 2, 5, 10]; (2) oncogenic transformation of metaplastic or ectopic squamous cells [bib_ref] Squamous cell carcinoma and adenoacanthoma of the stomach. A clinicopathologic study, Boswell [/bib_ref] , and (3) stem cell differentiation into both adenocarcinoma and squamous cell carcinoma [bib_ref] Primary adenosquamous carcinoma of the stomach. A case report and review, Straus [/bib_ref] [bib_ref] Adenosquamous carcinoma of the stomach: histological, histochemical and ultrastructural observations, Mingazzini [/bib_ref]. Accumulating evidence supports the first hypothesis. First, although the occurrence of metaplastic or ectopic squamous epithelium is extremely rare, neither can explain the coexistence of an adenocarcinoma component. Second, ASC is most often diagnosed at an advanced stage of disease [bib_ref] Adenosquamous carcinoma of the stomach. Histogenetic and ultrastructural studies, Mori [/bib_ref] , and the squamous cell carcinoma component is often located in a deeper layer than the adenocarcinoma component [bib_ref] A clinicopathological and immunohistochemical study of gastric cancer with squamous cell carcinoma..., Saito [/bib_ref]. In a study by Mori et al. [bib_ref] Adenosquamous carcinoma of the stomach. Histogenetic and ultrastructural studies, Mori [/bib_ref] , ASC was exclusively found in 9 (0.9%) of 976 patients with advanced disease, whereas it was not found among 1,028 patients with early-stage disease. Third, independent ultrastructural studies revealed the existence of a single malignant cell containing both secretory vesicles and tonofibrils in ASC [bib_ref] Adenosquamous carcinoma of the stomach. Histogenetic and ultrastructural studies, Mori [/bib_ref] [bib_ref] Adenosquamous carcinoma of the stomach. Histochemical and ultrastructural study, Grigolato [/bib_ref]. Moreover, independent immunohistochemical studies demonstrated that both components expressed identical levels of the p53 tumor suppressor gene [bib_ref] p53, p16 and RB expression in adenosquamous and squamous cell carcinomas of..., Lee [/bib_ref] [bib_ref] A clinicopathological and immunohistochemical study of gastric cancer with squamous cell carcinoma..., Saito [/bib_ref]. In our patient, there was no normal squamous epithelium in the surrounding mucosa and both components positively stained for HER2, supporting this hypothesis.
Previous studies indicated that both components can metastasize into a variety of patterns. Lymph node metastasis frequently occurs with an adenocarcinoma pattern [bib_ref] p53, p16 and RB expression in adenosquamous and squamous cell carcinomas of..., Lee [/bib_ref] [bib_ref] Early gastric cancer of adenosquamous carcinoma type: report of a case and..., Yoshida [/bib_ref] and occasionally with both components [bib_ref] p53, p16 and RB expression in adenosquamous and squamous cell carcinomas of..., Lee [/bib_ref] [bib_ref] Adenosquamous carcinoma of the stomach. Histogenetic and ultrastructural studies, Mori [/bib_ref] [bib_ref] Adenosquamous carcinoma of the stomach. Histochemical and ultrastructural study, Grigolato [/bib_ref] [bib_ref] Primary gastric adenosquamous carcinoma in a Caucasian woman: a case report, Faria [/bib_ref]. Lee et al. [bib_ref] p53, p16 and RB expression in adenosquamous and squamous cell carcinomas of..., Lee [/bib_ref] reported that metastatic lymph nodes were composed of only the adenocarcinoma components in 9 (64.3%) of 14 cases, whereas both components and only the squamous cell carcinoma component were found in 3 and only 1 patient, respectively [bib_ref] p53, p16 and RB expression in adenosquamous and squamous cell carcinomas of..., Lee [/bib_ref]. Consistent with this observation, metastatic lymph nodes in our patient consisted of only the adenocarcinoma component. Recently, Saito et al. [bib_ref] A clinicopathological and immunohistochemical study of gastric cancer with squamous cell carcinoma..., Saito [/bib_ref] reported that the Ki-67 index in the adenocarcinoma component was higher than that in the squamous cell carcinoma component in all 8 patients examined, further supporting these findings. Regarding other metastatic sites, some previous studies reported that the adenocarcinoma component was predominant [bib_ref] Adenosquamous carcinoma of the remnant stomach: report of a case, Mori [/bib_ref] [bib_ref] Early gastric cancer of adenosquamous carcinoma type: report of a case and..., Yoshida [/bib_ref] ; however, others found both components [bib_ref] Adenosquamous carcinoma of the stomach. A clinicopathologic analysis of 28 cases, Mori [/bib_ref] [bib_ref] Primary gastric adenosquamous carcinoma in a Caucasian woman: a case report, Faria [/bib_ref] [bib_ref] Weekly paclitaxel therapy is effective for gastric adenosquamous carcinoma: a case report, Tohma [/bib_ref]. Mori et al. [bib_ref] Adenosquamous carcinoma of the stomach. A clinicopathologic analysis of 28 cases, Mori [/bib_ref] found that both components existed in almost all metastatic lesions at autopsy in 9 patients. In the present case, the tumor was strongly positive for HER2 in both components; therefore, it was not of therapeutic significance to examine which component type had metastasized to the liver.
ASC is often diagnosed as advanced or recurrent disease; therefore, the majority of patients may be candidates for systemic chemotherapy. In several case reports, systemic chemotherapy, such as S-1 [bib_ref] A patient with gastric adenosquamous carcinoma with intraperitoneal free cancer cells who..., Ebi [/bib_ref] , paclitaxel [bib_ref] Weekly paclitaxel therapy is effective for gastric adenosquamous carcinoma: a case report, Tohma [/bib_ref] [bib_ref] Gastric adenosquamous carcinoma producing granulocyte-colony stimulating factor, Endo [/bib_ref] , or irinotecan plus cisplatin [bib_ref] A case of gastric adenosquamous carcinoma successfully treated with second-line chemotherapy (CPT-11..., Ishiguro [/bib_ref] , has been reported as an effective treatment for advanced ASC; however, no standard chemotherapy has been established for ASC because of its rarity. We selected trastuzumab-based chemotherapy due to the fact that this tumor was clinicopathologically similar to classical adenocarcinoma, even though the clinical prognosis is different between these two types of tumors [bib_ref] Adenosquamous carcinoma of the stomach. A clinicopathologic analysis of 28 cases, Mori [/bib_ref]. The durable response achieved in our patient can be attributed to the strong immunoreactivity of HER2 and high-level HER2 amplification in both the adenocarcinoma and squamous cell carcinoma components. Considering that the differentiated type is more common in the adenocarcinoma component [bib_ref] p53, p16 and RB expression in adenosquamous and squamous cell carcinomas of..., Lee [/bib_ref] [bib_ref] A clinicopathological and immunohistochemical study of gastric cancer with squamous cell carcinoma..., Saito [/bib_ref] [bib_ref] Adenosquamous carcinoma of the stomach. A clinicopathologic analysis of 28 cases, Mori [/bib_ref] , and that HER2 overexpression is more frequently observed in the intestinal type of gastric adenocarcinoma [bib_ref] Assessment of a HER2 scoring system for gastric cancer: results from a..., Hofmann [/bib_ref] , further studies to evaluate the rate of HER2-positivity and its clinical significance in ASC are warranted.
In summary, we reported a patient with HER2-positive ASC who achieved a durable PR after administration of trastuzumab-based combination chemotherapy. Our results indicate that trastuzumab is an effective treatment for ASC as well as gastric adenocarcinoma; in addition, they indicate that HER2 status should be examined in ASC.
# Disclosure statement
The authors declare that they have no potential conflicts of interest. An endoscopic examination revealed an ulcerated protruding tumor that occupied the lesser curvature of the upper to lower gastric body. This tumor was diagnosed as advanced type 3 gastric cancer. The pathological diagnosis of the resected specimens was adenosquamous cell carcinoma with a mixture of both adenocarcinoma (a) and squamous cell carcinoma components (b). c, d Immunohistochemical analysis showed strong positivity for HER2 (Herceptest TM ) in both components. e, f Dual-color chromogenic in situ hybridization revealed high amplification of the HER2 gene (INFORM HER2 Dual ISH; red, chromosome 17 centromere; black, HER2) in both components. . a An abdominal CT scan revealed multiple liver metastases. b After 3 cycles of chemotherapy with capecitabine, cisplatin, and trastuzumab, a PR was observed. c After 6 treatment cycles, a PR was confirmed with no evidence of new lesions. d Thirteen months after initial chemotherapy, a PET-CT scan showed no increase in the metastatic hepatic lesions.
© 2014 S. Karger AG, Basel www.karger.com/cro |
Consensus on Perioperative Transesophageal Echocardiography of the Brazilian Society of Anesthesiology and the Department of Cardiovascular Image of the Brazilian Society of Cardiology
Through the Life Cycle of Intraoperative Transesophageal Echocardiography (ETTI/SBA) the Brazilian Society of Anesthesiology, together with the Department of Cardiovascular Image of the Brazilian Society of Cardiology (DIC/SBC), created a task force to standardize the use of intraoperative transesophageal echocardiography by Brazilian anesthesiologists and echocardiographers based on scientific evidence from
# Introduction
Since its introduction in clinical practice in the late 1980s, transesophageal echocardiogram (TEE) has become one of the main diagnostic modalities in cardiology, as it guides anesthetic/surgical procedures and reduces morbidity and mortality in cardiac surgeries. [bib_ref] Guidelines for performing a comprehensive transesophageal echocardiographic examination: recommendations from the American..., Hahn [/bib_ref] Due to the great proximity of the esophagus to the heart, to the absence of bones or lung tissue, and the use of high frequency transducers, it is possible to obtain high quality images. [bib_ref] Guidelines for performing a comprehensive transesophageal echocardiographic examination: recommendations from the American..., Hahn [/bib_ref] The first guideline on perioperative TEE was published in 1999 by the Society of Cardiovascular Anesthesiologists/American Society of Echocardiography (SCA/ASE), which defined the nomenclature and the 20 cross-sections for basic TEE. [bib_ref] ASE/SCA Guidelines for performing a comprehensive intraoperative multiplane transesophageal echocardiography examination: recommendations..., Shanewise [/bib_ref] In Brazil, we have the Brazilian Society of Cardiology (SBC) guidelines on the use of TEE. [bib_ref] Sociedade Brasileira de Cardiologia. Diretrizes das indicações da ecocardiografia, Barbosa [/bib_ref] The levels of evidence and indications for using TEE in cardiac and non-cardiac surgeries are shown in [fig_ref] Table 1: Level of evidence ---intraoperative transesophageal echocardiography [/fig_ref].
The SCA/ASE and SBC guidelines define professionals qualified to use echocardiography as a diagnostic method or as hemodynamic monitoring according to their basic and advanced knowledge criteria. [bib_ref] Guidelines for performing a comprehensive transesophageal echocardiographic examination: recommendations from the American..., Hahn [/bib_ref] [bib_ref] Sociedade Brasileira de Cardiologia. Diretrizes das indicações da ecocardiografia, Barbosa [/bib_ref] In Brazil, the area of perioperative echocardiography activity is being defined by the Brazilian Society of Anesthesiology (SBA) and SBC. As a first step in the standardization of this qualification and in order to promote continuing education to its members, in the last five years the SBA has taught the course of intraoperative echocardiography (ETI/SBA), divided into two modules, basic (Module I) and advanced (Module II).Thus, the SBA and DIC/SBC consensus on intraoperative TEE aims to standardize intraoperative echocardiography for Brazilian anesthesiologists and echocardiographers based on the scientific evidence of ASE/SCA and SBC.
## Equipment
TEE probe was developed to improve images for which the transthoracic technique had limitations, such as for obese and emphysematous patients and those with thoracic anomalies. [bib_ref] Guidelines for performing a comprehensive transesophageal echocardiographic examination: recommendations from the American..., Hahn [/bib_ref] The TEE ultrasound wave passes only through the esophagus and pericardium to form the heart images. In this way, images with higher resolution and greater number of anatomical sections are obtained . In addition, Large non-cardiac surgery in high-risk patients IIa the TEE transducer may be attached to a specific part of the esophagus or stomach, which allows a more detailed analysis of the cardiac anatomy. [bib_ref] Guidelines for performing a comprehensive transesophageal echocardiographic examination: recommendations from the American..., Hahn [/bib_ref] [bib_ref] Basic physics of transesophageal echocardiography, Rengasamy [/bib_ref] The current TEE transducers operate at a frequency of 3.5---7 MHz, reaching up to 20 MHz. [bib_ref] Basic physics of transesophageal echocardiography, Rengasamy [/bib_ref] Most adult TEE probes have two knobs on the handle. One for anteflex and retroflex movements and the other for lateral movement, clockwise and counterclockwise. [bib_ref] Guidelines for performing a comprehensive transesophageal echocardiographic examination: recommendations from the American..., Hahn [/bib_ref] Multiplane transducers have ultrasonic beam angle control, which can vary from zero degrees to 180 - . All these controls, associated with the probe removal and introduction into the esophagus, allow the visualization of several echocardiographic cross-sections [fig_ref] Figure 2: Probe and transducer handling movements to acquire echocardiographic images [/fig_ref]. [bib_ref] Guidelines for performing a comprehensive transesophageal echocardiographic examination: recommendations from the American..., Hahn [/bib_ref] [bib_ref] Intraoperative monitoring with transesophageal echocardiography in cardiac surgery, Júnior [/bib_ref] Adult TEE probe measures about 100 cm and the diameter varies from 9 to 12 mm, 1 to 2 mm thicker at the tip. For using an adult TEE probe, the patient should weigh at least 20 kg. [bib_ref] Physics of echocardiography, Bulwer [/bib_ref] Complications TEE-related complications may be separated into two groups: (1) Direct esophageal, stomach and/or airway trauma; (2) TEE indirect effects [fig_ref] Table 2: Transesophageal echocardiography complications [/fig_ref]. [bib_ref] Guidelines for performing a comprehensive transesophageal echocardiographic examination: recommendations from the American..., Hahn [/bib_ref] In group 1, complications include esophageal bleeding, burning, dysphagia, and laryngeal discomfort. [bib_ref] Intraoperative monitoring with transesophageal echocardiography in cardiac surgery, Júnior [/bib_ref] Most of these complications occur during the catheter passage. In patients undergoing cardiopulmonary bypass, the probe passage should occur prior to heparinization and withdrawal should occur only after reversal with protamine and with activated clotting time (ACT) less than 120 s. [bib_ref] Guidelines for performing a comprehensive transesophageal echocardiographic examination: recommendations from the American..., Hahn [/bib_ref] In a study of 10,000 TEE exams, there was one case of hypopharyngeal perforation (0.01%), two cases of cervical esophageal perforation (0.02%), and no case of gastric perforation. The incidence of morbidity and mortality is 0.2% and zero, respectively. [bib_ref] Impacto da ecocardiografia transesofágica intraoperatória na cirurgia cardíaca. Análise retrospectiva duma série..., Silva [/bib_ref] The most common complications associated with the intraoperative use of TEE are: odynophagia (0.1%), dental injury (0.03%), orotracheal tube misplacement (0.03%), upper gastrointestinal bleeding (0.03%), and bacteremia (0---17%). Although the incidence of bacteremia is high, there is no correlation with the development of infectious endocarditis. [bib_ref] Impacto da ecocardiografia transesofágica intraoperatória na cirurgia cardíaca. Análise retrospectiva duma série..., Silva [/bib_ref] Proximity of transducer with left atrium and pulmonary vein s Transesophageal probe Upper-esophageal crosssection Mid-esophageal crosssection Transgastric crosssection Deep transgastric crosssection (45-50 cm) Anatomical relationships for transesophageal probe, esophagus, and heart. Document downloaded from http://www.elsevier.es, day 05/01/2018. This copy is for personal use. Any transmission of this document by any media or format is strictly prohibited. Document downloaded from http://www.elsevier.es, day 05/01/2018. This copy is for personal use. Any transmission of this document by any media or format is strictly prohibited. In group 2, we have complications indirectly related to TEE, which include: hemodynamic, pulmonary changes, airway manipulations, and distraction during patient care. [bib_ref] Impacto da ecocardiografia transesofágica intraoperatória na cirurgia cardíaca. Análise retrospectiva duma série..., Silva [/bib_ref] It is important to leave all anesthesia workstation and monitor alarms turned on, as the TEE device may be positioned in a way that makes it difficult to see all the monitors at once. Furthermore, during examination, the examiner may become inattentive with the patient and attempt to do some echocardiographic imaging or make a diagnosis. Initially, echocardiographic examination with a second anesthesiologist at site is important so that, while one anesthesiologist performs the echocardiographic exam, the other helps in the intensive control of the patient. [bib_ref] Guidelines for performing a comprehensive transesophageal echocardiographic examination: recommendations from the American..., Hahn [/bib_ref] [bib_ref] Guidelines for the use of echocardiography as a monitor for therapeutic intervention..., Porter [/bib_ref] Probe passing technique Probe introduction should be performed with the patient under anesthesia after tracheal intubation. [bib_ref] Guidelines for performing a comprehensive transesophageal echocardiographic examination: recommendations from the American..., Hahn [/bib_ref] Probe should be lubricated with proper TEE gel (usually lidocaine gel or intimate lubricant), and the stomach may be previously emptied to enhance the images. [bib_ref] Guidelines for performing a comprehensive transesophageal echocardiographic examination: recommendations from the American..., Hahn [/bib_ref] Probe passing is often a challenge. The technique consists of proper lubrication of the probe, passing it through the posterior oropharynx and elevating the mandible with the left hand. Passing it through the upper esophageal sphincter is the procedure critical time in which the most serious complications may occur. The probe should progress toward the esophagus without resistance. [bib_ref] Intraoperative monitoring with transesophageal echocardiography in cardiac surgery, Júnior [/bib_ref] In case of resistance, this usually occurs because the probe's tip is lodged in the pyriform sinus, epiglottic vallecula, posterior part of the tongue, or esophageal diverticula. The probe should never be forced against resistance, as it may lead to complications such as perforation and bleeding. [bib_ref] Guidelines for performing a comprehensive transesophageal echocardiographic examination: recommendations from the American..., Hahn [/bib_ref] Another technique that can be used is the probe introduction with the help of a laryngoscope. In this case, care should be taken with the hemodynamic stimulus triggered by a second laryngoscopy. [bib_ref] Guidelines for performing a comprehensive transesophageal echocardiographic examination: recommendations from the American..., Hahn [/bib_ref] The TEE examination allows intra-and extracardiac anatomical analysis of the great vessels at the base, analysis of congenital heart defects, and qualitative and quantitative analysis of Doppler flow. 1,3 A complete intraoperative examination not only characterizes the patient's hemodynamic profile but may lead to changes in the surgical approach in up to 25% of the exams, in which changes such as the presence of a patent foramen ovale (PFO), left atrial (LA) thrombi, or atheroma plaques in the ascending aorta (Asc Ao). [bib_ref] Impacto da ecocardiografia transesofágica intraoperatória na cirurgia cardíaca. Análise retrospectiva duma série..., Silva [/bib_ref] The examinations must be recorded on a digital media for further analysis and medical report. Thus, it is possible to analyze the patient's evolution both intraoperatively and postoperatively. 6
## Management of tee probe
The correct use of the multiplane TEE probe functions provides adequate acquisition of cardiac images during intraoperative examination. In addition, its proper handling reduces the incidence of stomach and esophageal complications.
Moving the probe in the caudal and cranial direction produces changes in the images in the inferior and superior directions of the heart, respectively (high esophagus: about 20---25 cm; medium esophagus: about 30---40 cm; transgastric (TG): about 40---45 cm; deep TG: about 45---50 cm). Changes to the right or left of the heart can be obtained by moving the probe clockwise or counterclockwise. The best alignment of the images can be obtained with the anterior or posterior movement of the probe, using anteflexion or retroflexion with the help of the larger probe handle. [bib_ref] Guidelines for performing a comprehensive transesophageal echocardiographic examination: recommendations from the American..., Hahn [/bib_ref] Multiplane TEE provides fine adjustments in the inclination angle of the image plane and consequently allows for more precise anatomical analysis. The angle may range from 0 - to 180 - [fig_ref] Figure 2: Probe and transducer handling movements to acquire echocardiographic images [/fig_ref]. 1
## Comprehensive echocardiographic examination
The technique description for performing the comprehensive transesophageal echocardiographic examination will follow the ASE/ACS model. [bib_ref] Guidelines for performing a comprehensive transesophageal echocardiographic examination: recommendations from the American..., Hahn [/bib_ref] [bib_ref] ASE/SCA Guidelines for performing a comprehensive intraoperative multiplane transesophageal echocardiography examination: recommendations..., Shanewise [/bib_ref] The nomenclature of the mitral valve (MV) cusps will follow the classification of Carpentier et al. [bib_ref] The ''physio-ring'': an advanced concept in mitral valve annuloplasty, Carpentier [/bib_ref] The cross-sectional images established for this complete analysis are as follows.
## Mid-esophageal five-chamber cross-section
Advance the probe up to 30 cm from the incisors in the midesophagus. Rotate the transducer angle up to 10 - and look for the left ventricular outflow tract (LVOT) and aortic valve (AV) and flex the probe anteriorly. It is termed five chambers because we visualize the LA, right atrium (RA), left ventricle (LV), right ventricle (RV), and LVOT with part of the AV. The evaluation of regional LV function is impaired by the apex shortening [fig_ref] Figure 3: A [/fig_ref].
## Mid-esophageal four-chamber cross-section
Advance the probe from the previous position (fivechamber) up to 30---35 cm. We can describe the LV, RV, LA, RA, interatrial septum (IAS), MV, and tricuspid valve (TV). The true (not shortened) apex may be exposed by retroflexing the probe. It is one of the most used views for diagnosis [fig_ref] Figure 3: A [/fig_ref].
## Mid-esophageal mitral commissural cross-section
From the previous position (four-chamber), rotate the probe up to 45---60 - . The MV has a typical appearance in this image (segments P 1 ---A 2 ---P 3 ), the section passes through the MV commissural axis. Papillary muscles and chordae tendineae are identified. Small manipulations of the probe in this image may provide anatomical details and a more complete analysis of the MV [fig_ref] Figure 3: A [/fig_ref].
## Mid-esophageal two-chamber cross-section
From the commissural view, rotate the angle between 60 - and 90 - . LA, left atrial appendage (LAA), LV, and MV are identified. The anterior and inferior walls are exposed and both ventricular and mitral valve function can be evaluated. Coronary venous sinus (CS) is seen on the short axis, just above the LV basal inferior wall [fig_ref] Figure 3: A [/fig_ref].
## Mid-esophageal long axis cross-section
In the two-chamber window, rotate the angle to 120 - . LA, LV, LVOT, AV, proximal Asc Ao, CS, and MV (P 2 and A 2 segments) are visualized. Adjustments are made to maximize and accurately measure the LVOT diameter. The LV anteroseptal and inferolateral movements can be investigated [fig_ref] Figure 4: A [/fig_ref].
## Mid-esophageal aortic valve long-axis cross-section
The probe is retracted a few centimeters from the position of the long axis in the mid-esophagus. Maintaining an angle of 120---140 - , the LVOT, AV, and proximal Asc Ao are aligned and, from this point, AV can be evaluated and the sinotubular and annular diameters of Ao measured. Finding protruding atherosclerotic plaques is also a utility of this window.
## Mid-esophageal ascending aorta long axis cross-section
From the AV long axis view, the probe is removed, turned counterclockwise (90---110 - ). The walls of Asc Ao are inspected at various depths depending on the pathology. Dissection and graft sutures are examples of evaluations in this window [fig_ref] Figure 4: A [/fig_ref]. The Asc Ao distal portion and the proximal aortic arch are not usually visualized by this technique; they require complementation with ultrasoundepiaortic guidance for artery cannulation [fig_ref] Figure 4: A [/fig_ref].
## Upper esophageal ascending aorta short axis cross-section
From the AV long axis view (120---140 - ), the probe is withdrawn and, with 90 - counterclockwise rotation, the images of Asc Ao short axis and superior vena cava (SVC) are obtained. The main pulmonary artery can be seen with bifurcation (probe rotation to the left) and the right pulmonary artery branch can be seen to a large extent (probe rotation to the right). It is a good window to check pulmonary artery catheter placement [fig_ref] Figure 4: A [/fig_ref].
## Mid-esophageal right pulmonary vein cross-section
From the previous position (Asc Ao short axis) retract the probe between 0 - and 60 - and rotate it clockwise. The right upper pulmonary vein (RUPV) and the right lower pulmonary vein (RLPV) are visualized. The RUPV is shown with a flow parallel to the US beam and can be evaluated on Doppler mode. Occasionally, a right middle lobar pulmonary vein can be viewed entering the LA between the RUPV and RLPV orifices. The evaluation of pulmonary veins is of special interest in congenital diseases.
## Mid-esophageal aortic valve short axis cross-section
From the previously described view (right pulmonary vein) return to the Ao in the screen center with a counterclockwise rotation of the probe. Advance to AV commissural leaflets at an approximate angle of 45 - . The general morphology of AV (number of valves, presence of calcifications, motion) is observed, besides determining if there is presence of aortic stenosis by planimetry. Interatrial septum (IAS) bulging continuously in the cardiac cycle by high pressures can also be observed in this window [fig_ref] Figure 5: A [/fig_ref]. From this position, the discreet withdrawal of the probe or its anteflexion also depicts the LAA.
## Mid-esophageal right ventricle inflow-outflow cross-section
From the AV short axis, advance the probe and rotate the transducer angle between 50 - and 70 - until visualizing the TV, RV inflow tract (RVIT), RV outflow tract (RVOT), and proximal pulmonary artery. In addition, RA, LA, IAS, RV, and pulmonary valve (PV) are all observed. It is superior when compared to the visualization of four chambers in the mid-esophagus plane for flow analysis by TV on Doppler mode. This window is also useful in congenital heart diseases and in the correct placement of pulmonary artery catheter [fig_ref] Figure 5: A [/fig_ref].
## Mid-esophageal modified bicaval cross-section
From the RVIT and RVOT window, with a clockwise rotation between 50 - and 70 - , TV is centralized and LA, IAS, RA and inferior vena cava (IVC) are visualized. It may be an apical window to analyze eccentric regurgitant jets by TV on Doppler mode [fig_ref] Figure 5: A [/fig_ref].
## Mid-esophageal bicaval cross-section
From the window previously described (modified bicaval) raise the angular rotation to 90---110 - and rotate the probe clockwise. Structures to be analyzed include LA, IAS, RA, SVC, IVC, and right atrial appendage (RAA). This visibility position is especially important for analysis of PFO and IAS defects and to detect air within the atria, as well as to assist in the difficult passage of the pulmonary artery catheter into the RV [fig_ref] Figure 5: A [/fig_ref].
## Mid-esophageal right and left pulmonary veins cross-section
In the bicaval position in the mid-esophagus (90---110 - ) continue to rotate the probe clockwise until the right upper pulmonary vein (RUPV) and right lower pulmonary vein (RLPV) are visualized. When rotating the probe counterclockwise, the RUPV and RLPV can be visualized at the right end of the screen, where they are aligned parallel to the US sound beam. It is ideal for flow analysis on Doppler mode [fig_ref] Figure 6: A [/fig_ref].
## Mid-esophageal left atrial appendage cross-section
This cross-sectional view is obtained in the mid-esophagus with an angle between 90 - and 110 - , turning the probe clockwise. The upper left pulmonary vein (ULPV) is often visualized. Due to the complex and variable anatomy of the LAA, its evaluation must be done in different sections. Color Doppler and pulsatile Doppler are useful modalities of evaluation, particularly regarding the LAA contractile function [fig_ref] Figure 6: A [/fig_ref].
## Transgastric basal short axis cross-section
From mid-esophagus, the probe is advanced to the stomach and maintains the zero angulation. During the probe introduction, CS and TV are often visualized. Once in the stomach, this cross-section is obtained with a slight anteflexion of the probe. Its characteristic image is the MV on its short axis (fish mouth), with the anterior cusp on the left and the posterior cusp on the right. MV morphology and LV size and function can be assessed. In patients with mitral regurgitation, the use of color Doppler can be useful to characterize its regurgitant orifice [fig_ref] Figure 6: A [/fig_ref].
## Transgastric mid-papillary short axis cross-section
From TG basal short axis cross-section, the probe should return to the neutral position, or be minimally advanced, maintaining the zero angulation. This cross-section is extremely useful in assessing and monitoring the volume and size of regional and global LV function, as well as assessing territories irrigated by the right coronary (RC), anterior descending (AD), and circumflex (Cx) arteries. In it, the anterior, inferior, inferolateral, anterolateral, and septal walls can be identified, as well as the papillary anterolateral and posteromedial muscles [fig_ref] Figure 6: A [/fig_ref].
## Transgastric apical short axis cross-section
Maintaining contact with the gastric wall, the probe is slightly advanced and/or retroflexed from the TG midpapillary short axis, favoring the visualization of the apical segments of the LV and RV (turning the probe clockwise). This cross-section may be difficult to obtain due to the possible loss of contact between the probe and the gastric wall caused by retroflexion of the probe [fig_ref] Figure 7: A [/fig_ref].
## Transgastric right ventricular cross-section
This image is obtained at the same level as the TG basal short axis cross-section, only turning the probe clockwise. TV is visualized on its short axis. The use of color Doppler in this cross-section may help to characterize the TV regurgitant orifice [fig_ref] Figure 7: A [/fig_ref].
## Transgastric right ventricular inflow-outflow cross-section
This image is orthogonal to that described previously (simply rotate the transducer by 90 - ). The anterior and posterior TV cusps and left and right PV valves are usually visualized. This cross-section can be used to align the jets in PV with the US beam for use in the continuous and pulsatile Doppler modes [fig_ref] Figure 7: A [/fig_ref].
## Deep transgastric five-chamber cross-section
This image is obtained by advancing the probe in the stomach, maintaining contact with the gastric wall (probe anteflexion). It is the ideal incidence for Doppler analysis of AV, LVOT, and often MV, as blood flow is parallel to the US beam [fig_ref] Figure 7: A [/fig_ref].
## Transgastric two-chamber cross-section
This cross-section is obtained from TG mid-papillary short axis, changing the angle to 90 - . It allows evaluation of the LV anterior and inferior walls, as well as the entire mitral valve apparatus (MV, papillary muscles, chordae tendineae). It is also possible to visualize the LA and the LAA [fig_ref] Figure 8: A [/fig_ref].
## Transgastric long axis cross-section
This cross-section is obtained by advancing the angle to 120---150 - from TG two-chamber. Portions of the inferolateral and anterior septal walls, LVOT, AV, and proximal Ao are visualized. Because LVOT and AV are parallel to the US beam, a Doppler analysis is possible [fig_ref] Figure 8: A [/fig_ref].
## Descending aorta long and short axes cross-section
Because the descending aorta (Desc Ao) is adjacent to the esophagus and stomach, the acquisition of images is simple. From TG mid-papillary short axis cross-section, the probe should be rotated at about 180 - , with slight adjustments until the abdominal Ao (below the diaphragm) is visualized, at the celiac trunk level; the probe is drawn and the transducer angle is maintained at zero degree and the short axis is obtained; the long axis is obtained with the transducer angle at 90 - . Image quality can be improved by decreasing the depth and adjusting the gain. Because there are no internal anatomical structures in Desc Ao, the description of findings location is usually based on the distance from the finding to the incisor teeth. Another important factor in the Desc Ao evaluation is the hemiazygos vein presence, which drains the posterior left thorax, often visualized in the most distal field of the image, joining the upper thorax to the azygos vein, which drains the right thorax. The azygos vein, being normally parallel to Ao and its walls being contiguous, can often be erroneously identified as a dissecting aortic blade. Color and pulsatile Doppler analysis easily differentiates the arterial from the venous flow [fig_ref] Figure 8: A [/fig_ref] and D).
## Upper esophageal aortic arch long axis cross-section
This image is obtained from the Desc Ao short thoracic axis; the probe is withdrawn until Ao becomes elongated and the left subclavian artery emergence is visualized. This position indicates the end of the distal aortic arch. The probe is turned clockwise and it is possible to visualize the mean aortic arch and the left innominate vein. Because the left source bronchus interposes between the esophagus and Ao, it is usually not possible to visualize the proximal aortic arch and the distal Asc Ao [fig_ref] Figure 9: A [/fig_ref].
## Upper esophageal aortic arch short axis cross-section
From the cross-section described above, the transducer angle is increased to about 70---90 - to obtain the long axis. The pulmonary artery trunk and PV can be visualized on the long axis; it is possible to obtain Doppler measurements. Due to Ao curvature, the right brachiocephalic trunk and left common carotid artery can often be identified to the right on the monitor [fig_ref] Figure 9: A [/fig_ref].
# Mitral valve introduction
Among the cardiac valves, MV has the most favorable anatomic characteristics for transesophageal examination, as the near field is close to the transducer, has the LA as an ''acoustic window'', and has no cardiac structure that can be calcified and/or generation of acoustic shadow between the transducer and its structure .
During the MV intraoperative period, TEE is a fundamental tool, as it allows the lesion identification and a detailed description of its mechanism and severity, thus helping the surgical decision making. 11
## Anatomy and nomenclature
When studying the MV anatomy and function, the most correct is that it is viewed as a valve complex composed of distinct anatomical structures that work in a coordinated way for its correct functioning. [bib_ref] Anatomy of the mitral valve apparatus Role of 2D and 3D echocardiography, Dal-Bianco [/bib_ref] The mitral valve complex, or mitral valve apparatus, is composed of mitral annulus, cusps, chordae tendineae, papillary muscles, and the left ventricular musculature. 12
## Mitral annulus
The mitral annulus is a connective tissue structure with a three-dimensional saddle-shaped complex structure. It is anteriorly related to the aortic valve apparatus, joins the atrium and left ventricle, and receives along its perimeter the insertion of the valvular cusps. [bib_ref] Anatomy of the mitral valve apparatus Role of 2D and 3D echocardiography, Dal-Bianco [/bib_ref] Mitral annulus has two main axes, one anteroposterior (upper and shorter) and one commissural (lower and longer). These axes may also be referred to as anteroposterior and anterolateral---posteromedial [bib_ref] Intraoperative transesophageal echocardiography for surgical repair of mitral regurgitation, Sidebotham [/bib_ref] [fig_ref] Figure 10: Mitral annulus assessment by three-dimensional echocardiography [/fig_ref].
Under the ventricular contraction action, the mitral annulus posterior region undergoes a significant reduction of its total area, with reduction up to 25%. The mitral annulus anterior region folds during systole and further decreases the anteroposterior diameter. 13
## Leaflets (cusps)
The normal mitral valve complex is composed of two anterior and posterior leaflets. [bib_ref] Anatomy of the mitral valve apparatus Role of 2D and 3D echocardiography, Dal-Bianco [/bib_ref] [bib_ref] Intraoperative transesophageal echocardiography for surgical repair of mitral regurgitation, Sidebotham [/bib_ref] The anterior leaflet is inserted into the anterior annulus, in a fibrous skeleton region adjacent to the aortic valve apparatus, called mitral-aortic intervalvular fibrosa (MAIVF). It occupies approximately 1/3 of the annular perimeter and 2/3 of its area. The posterior leaflet inserts into the posterior annulus and occupies 2/3 of its perimeter, but corresponds to only 1/3 of the annular area. [bib_ref] Intraoperative transesophageal echocardiography for surgical repair of mitral regurgitation, Sidebotham [/bib_ref] [bib_ref] Mitral valve repair: an echocardiographic review: Part 1, Gething [/bib_ref] The leaflets are in a curved coaptation line along the intercommissural axis (anterolateral---posteromedial); in normal situations, there are approximately 1 cm of tissue overlap. [bib_ref] Anatomy of the mitral valve apparatus Role of 2D and 3D echocardiography, Dal-Bianco [/bib_ref] MV leaflets were subdivided into distinct regions in order to improve communication among medical team members. Among the proposed schemes, we will use Carpentier nomenclature, 10 which is also used by ASE and SCA. [bib_ref] Guidelines for performing a comprehensive transesophageal echocardiographic examination: recommendations from the American..., Hahn [/bib_ref] According to Carpentier nomenclature, MV is subdivided into eight regions, from anterior to posterior, through small grooves or indentations present in the posterior leaflet. The posterior leaflet is divided into three segments, which are numbered from 1 to 3, with P1 being the most anterior and P3 the most posterior. [bib_ref] The ''physio-ring'': an advanced concept in mitral valve annuloplasty, Carpentier [/bib_ref] The anterior leaflet usually has no real anatomical grooves; however, didactically, it is subdivided in the same way as the posterior one: a more
## Papillary muscles, chordae tendineae, and left ventricle
The two papillary muscles (anterolateral and posteromedial) support the leaflets, located parallel to the ventricular musculature. [bib_ref] Anatomy of the mitral valve apparatus Role of 2D and 3D echocardiography, Dal-Bianco [/bib_ref] The anterolateral muscle generally arises from the middle portion of the anterolateral LV wall, receives vascularization from the AD and Cx artery branches. The posteromedial papillary muscle arises from the middle portion of the inferior wall, is exclusively vascularized by the RC artery branch, which makes it more vulnerable to ischemic insult. [bib_ref] Anatomy of the mitral valve apparatus Role of 2D and 3D echocardiography, Dal-Bianco [/bib_ref] The chordae tendineae connect both the papillary muscles and ventricular musculature to the MV leaflets. The anterolateral papillary muscle supports, through its cords, the A1/P1 segments and the anterolateral portion of the A2/P2 segments; the A2/P2 posteromedial portion and the A3/P3 segments are supported by the posteromedial papillary 12 [fig_ref] Figure 2: Probe and transducer handling movements to acquire echocardiographic images [/fig_ref].
Chordae tendineae are classically divided into first, second and third-order chordae. First-order chordae are connected to the tip of the leaflets, having as a consequence of its rupture the systolic eversion of leaflets, echocardiographically known as flail. [bib_ref] Anatomy of the mitral valve apparatus Role of 2D and 3D echocardiography, Dal-Bianco [/bib_ref] [bib_ref] Intraoperative transesophageal echocardiography for surgical repair of mitral regurgitation, Sidebotham [/bib_ref] Second-order chordae are connected to the base of leaflets and are known as structural chordae; it is possible to identify two to four chordae thicker than the others. [bib_ref] Anatomy of the mitral valve apparatus Role of 2D and 3D echocardiography, Dal-Bianco [/bib_ref] [bib_ref] Intraoperative transesophageal echocardiography for surgical repair of mitral regurgitation, Sidebotham [/bib_ref] Third-order chordae usually connect the posterior MV leaflet to the left ventricle wall, having recognized importance in architecture maintenance and ventricular performance 12 [fig_ref] Figure 2: Probe and transducer handling movements to acquire echocardiographic images [/fig_ref].
## Intraoperative mitral valve echocardiographic examination
In order to include and standardize the intraoperative use of new technologies that emerged in the last decade, mainly real-time three-dimensional (3D) TEE, in addition to the inclusion of new imaging plans, the 1999 ASE/SCA guidelines were reviewed and republished in 2013. 1,2,15 Such patterns of image acquisition and training were also adopted by the SBA for teaching and training.Didactically, we will describe the MV sequential evaluation method through the ASE/ACS methodology, 1 trying to identify its different segments according to their clinical importance. It is known that its accuracy is variable and highly dependent on the examiner's experience, as demonstrated by Mahmood et al. [bib_ref] Echocardiographic anatomy of the mitral valve: a critical appraisal of 2-dimensional imaging..., Mahmood [/bib_ref] On the other hand, the ability to localize it spatially in the mitral apparatus anatomy using two-dimensional (2D) examination still requires fundamental training.
## Mid-esophageal five-chamber cross-section
With the probe insertion at a depth of approximately 30 cm, with multiplane angle rotation about 10 - , AV, LVOT, LV (except its apex), and A1---A2 and P1---P2 segments are visualized [fig_ref] Figure 3: A [/fig_ref].
## Mid-esophageal four-chamber cross-section
The probe is inserted at a depth of approximately 35 cm, with multiplane angle rotation between 10 - and 20 - , until MV can be clearly visualized. The segments A3---A2 and P2---P1 are evident, as well as the TV septal and posterior leaflets [fig_ref] Figure 3: A [/fig_ref].
## Mid-esophageal commissural cross-section
With the probe in the four-chamber position, the multiplanar angle is advanced between 50 - and 70 - and the commissural plane will be visualized. The segments P1, A2, and P3 will be visible from right to left, in addition to the papillary anterolateral and posteromedial muscles. The probe is rotated to the right and the image plane will pass through the entire extension of the anterior leaflet (right commissural: A1---A2---A3), as well as the entire extension of the posterior leaflet when rotated to the left (left commissural: P1---P2---P3, [fig_ref] Figure 3: A [/fig_ref].
## Mid-esophageal two-chamber cross-section
From the commissural plane, the multiplane angle is advanced between 80 - and 100 - and the image plane called two-chamber will appear. From right to left, the MV segments will be A1/A2/A3 and P3 (Figs. 3D and 4A).
## Transgastric basal short axis cross-section
The transducer is advanced to the stomach, with the multiplane angle between zero and 20 - , and a MV image is obtained with a fish mouth opening and closing. The anterior leaflet appears to the left and the posterior leaflet to the right; the posteromedial commissure appears near the transducer and the anterolateral commissure more distal to the transducer [fig_ref] Figure 6: A [/fig_ref].
## Three-dimensional frontal cross-section (face view)
MV apparatus 3D evaluation is useful for defining and locating the pathology, describing the pathophysiological mechanism and its severity, as well as facilitating communication with the interventional surgeon or cardiologist. [bib_ref] Guidelines for performing a comprehensive transesophageal echocardiographic examination: recommendations from the American..., Hahn [/bib_ref] With the evolution of surgical and percutaneous techniques of mitral apparatus repair there was a need for real-time high quality images, which is now possible thanks to the technological evolution of the matrix array and image manipulation softwares.The complete 3D examination will be treated in a specific topic, but we will describe two more representative and useful intraoperative imaging plans, which may be in realtime or in multi-beat acquisitions.
## Left atrium view or surgeon's view
From the LA view, it is possible to identify all the valve segmentation, which facilitates the topographic description of a pathology and communication with the surgical team. By convention, AV is positioned at 12 o'clock in the image and LAA at 9 o'clock [fig_ref] Figure 3: A [/fig_ref].
## Left ventricle view
In the left ventricular aspect, the imaging plane should again be oriented with LVOT and AV positioned at 12 o'clock, with posterior leaflet at the bottom of the image, anterior leaflet at the top, anterolateral commissure to the right, and posteromedial commissure to the left [fig_ref] Figure 4: A [/fig_ref].
## 3d color doppler echocardiography
By means of multi-beat acquisition, it is possible to obtain volumetric images of the regurgitant jet and its relation with the MV structures and 3D evaluation of its various components, which allows to demarcate its exact location and quantitative evaluation in specific software [bib_ref] 3-Dimensional echocardiography and its role in preoperative mitral valve evaluation, Andrawes [/bib_ref] [fig_ref] Figure 5: A [/fig_ref].
## Three-dimensional quantitative evaluation of mitral valve
Besides the ability to generate real time 3D images, numerous software were developed with the ability to generate a quantitative analysis model from specific points marked on the 3D image of the valve apparatus. Among these, the most widely used and studied software are Mitral Valve Quantification ---MVQ (Phillips Healthcare ® , Inc., Andover, MA) and 4D MV---Assessment Software (TomTec Imaging Systems GmbH ® , Munich, Germany). [bib_ref] 3-Dimensional echocardiography and its role in preoperative mitral valve evaluation, Andrawes [/bib_ref] Although reference values and clinical usefulness are still in the validation stage, important knowledge has been accumulated regarding mitral valve apparatus remodeling in different pathophysiological states. 18
## Intraoperative evaluation of mitral valve
The evaluation objectives before cardiopulmonary bypass are to define the mechanism, locate the lesion, estimate its severity, and identify associated pathologies, such as pulmonary arterial hypertension, ventricular dysfunction, and tricuspid regurgitation. [bib_ref] Intraoperative transesophageal echocardiography for surgical repair of mitral regurgitation, Sidebotham [/bib_ref] It should be noted, however, that general anesthesia significantly modifies the hemodynamic conditions and often decreases the severity of regurgitant valve lesions. 13
## Evaluation of annular morphology
Different clinical situations and pathologies alter the annular dimensions. By definition, mitral annulus should be measured in the mid-esophagus, long axis at the end of systole. The measurement is made from the insertion of posterior leaflet to the AV base. The upper limit of normality is 35 mm, values greater than 40 mm indicate marked dilatation. [bib_ref] Intraoperative transesophageal echocardiography for surgical repair of mitral regurgitation, Sidebotham [/bib_ref] The intercommissural diameter is often also measured, but its reference values are less clear in the literature [bib_ref] Intraoperative transesophageal echocardiography for surgical repair of mitral regurgitation, Sidebotham [/bib_ref] [fig_ref] Figure 10: Mitral annulus assessment by three-dimensional echocardiography [/fig_ref]. Leaflets evaluation MV diseases can be classified according to the leaflets motion, according to Carpentier's classification. [bib_ref] The ''physio-ring'': an advanced concept in mitral valve annuloplasty, Carpentier [/bib_ref] [bib_ref] Echocardiography in mitral regurgitation with relevance to valve surgery, Shah [/bib_ref] Diseases are classified as type I if presenting with normal leaflet motion, such as mitral regurgitation due to endocarditis leading to leaflet perforation, congenital clefts or isolated annular dilatation, in case of atrial fibrillation [bib_ref] Echocardiography in mitral regurgitation with relevance to valve surgery, Shah [/bib_ref] ; type II if presenting with excessive leaflet motion, such as prolapse caused by fibroelastic deficiency or Barlow's disease [bib_ref] Echocardiography in mitral regurgitation with relevance to valve surgery, Shah [/bib_ref] ; and type III when presenting with a pathophysiological mechanism that causes restricted leaflet motion and may be subdivided into IIIa, IIIb, and IIIc. The IIIa subtype corresponds to restriction caused by shortening and fusion of the subvalvar apparatus, such as in rheumatic heart disease. The IIIb and IIIc subtypes represent restriction due to leaflet tethering, as in functional mitral regurgitation, with IIIb corresponding to symmetric and IIIc to asymmetric tethering 19 [fig_ref] Figure 6: A [/fig_ref] and C).
Mitral regurgitation caused by type II pathologies in general produce regurgitant jets in the opposite direction of the lesion; however, central jets may occur in case of both leaflets involvement. [bib_ref] Echocardiography in mitral regurgitation with relevance to valve surgery, Shah [/bib_ref] Type III pathologies, with restricted leaflet motion, usually produce regurgitant jets in the same direction as the leaflets, may be central in case of symmetrical involvement. [bib_ref] Echocardiography in mitral regurgitation with relevance to valve surgery, Shah [/bib_ref] Systolic anterior motion Systolic anterior motion (SAM) of MV has been reported after valve repair, with an incidence up to 16% in patients with myxomatous disease. [bib_ref] Echocardiographic predictors of left ventricular outflow tract obstruction and systolic anterior motion..., Maslow [/bib_ref] It consists of the anterior displacement of the coaptation point and subvalvular tissue toward LVOT during systole, causes varying degrees of dynamic obstruction 20 [fig_ref] Figure 7: A [/fig_ref].
Not all SAM causes clinically relevant obstruction. It is traditionally diagnosed in the presence of a gradient across the LVOT and mitral insufficiency. [bib_ref] Echocardiographic predictors of left ventricular outflow tract obstruction and systolic anterior motion..., Maslow [/bib_ref] The presence of mitral tissue across the LVOT is often possible to be verified, but without any gradient, a situation clarified by 3D/4D TEE. [bib_ref] Mitral valve repair: an echocardiographic review: Part 1, Gething [/bib_ref] The ability to echocardiographically predict which patients are at greatest risk for SAM development after mitral repair is a key task of intraoperative examination. In the last two decades, some criteria have been defined and validated. The most relevant ones are the following: - Distance between coaptation point and septum
## Left ventricular outflow tract, aortic valve and aorta
AV is a component of the aortic root, which by definition extends from the basal aortic valve annulus to the sinotubular junction. With the association of the LVOT interval triangles, we will have the so-called aortic valve complex 23 [fig_ref] Figure 8: A [/fig_ref]. Pathologies can occur not only in AV but also involve any components of this complex, which reinforces the importance of a detailed, hemodynamic, and anatomical evaluation of this region. [bib_ref] Guidelines for performing a comprehensive transesophageal echocardiographic examination: recommendations from the American..., Hahn [/bib_ref] With the advent of new technologies for treating aortic complex pathologies, there was a great advance in the understanding of its anatomy; 2D and 3DD TEE stand out in the detailed morphological evaluation of these anatomical structures. 24
## Anatomy
The aortic valvar complex is a continuation of the LVOT and is located to the right and posterior to the RVOT, with its posterior margin wedged between MV orifice and muscular IVS. [bib_ref] Anatomy of the aortic valvar complex and its implications for transcatheter implantation..., Piazza [/bib_ref] Its basal circumference is called aortic annulus or basal ring [fig_ref] Figure 8: A [/fig_ref] and has an oval shape in most individuals. NC leaflet LC leaflet RC leaflet [fig_ref] Figure 9: A [/fig_ref] Mid-esophageal aortic valve short axis crosssection. NC, non-coronary; LC, left coronary; RC, right coronary.
Approximately two thirds of this basal ring, where the AV valves nadirs are inserted, are connected to the muscular IVS and the remaining third is in contact with the mitral valve anterior leaflet, represents the base of a fibrous triangle between the non-coronary valves and left coronary artery (''fima'').
The aortic root extends from the aortic valves basal insertion to the sinotubular junction, passes through the Valsalva sinuses, from where on the upper border of the left and right coronary sinuses the coronary arteries originate. Thus, AV is a semilunar valve with three valves, identified according to the presence or absence of a coronary artery, which originates from the corresponding sinus of Valsalva: left coronary valve, right coronary valve, and non-coronary valve [bib_ref] Anatomy of the aortic valvar complex and its implications for transcatheter implantation..., Piazza [/bib_ref] [fig_ref] Figure 9: A [/fig_ref].
## 2d and doppler imaging of aortic valve and left ventricular outflow tract
Its anatomical location close to the LA, which is in contact with the esophagus in its mid-plane, provides a precise anatomical image, both in short [fig_ref] Figure 5: A [/fig_ref] and long axes views [fig_ref] Figure 4: A [/fig_ref]. [bib_ref] Anatomy of the aortic valvar complex and its implications for transcatheter implantation..., Piazza [/bib_ref] The perpendicular contact of the US beams with these near field allows us to use the highest available frequencies, with consequent detailed appreciation of their morphology. [bib_ref] Basic physics of transesophageal echocardiography, Rengasamy [/bib_ref] The use of zoom as a resource to increase the detail for distance and/or diameters measurements can further improve accuracy and is recommended [fig_ref] Figure 2: Probe and transducer handling movements to acquire echocardiographic images [/fig_ref]. The Doppler bundles perpendicularity with these structures in the mid-esophagus echocardiographic sections makes it impossible to accurately assess the flow velocities. [bib_ref] Basic physics of transesophageal echocardiography, Rengasamy [/bib_ref] This limitation is overcome with TG long axis [fig_ref] Figure 8: A [/fig_ref] and deep TG sections [fig_ref] Figure 7: A [/fig_ref] , in which the parallel alignment with the beams allows an accurate assessement. [bib_ref] Basic physics of transesophageal echocardiography, Rengasamy [/bib_ref] [bib_ref] Assessment of perioperative hemodynamics, Skubas [/bib_ref] Mid-esophageal aortic valve short axis cross-section [fig_ref] Figure 9: A [/fig_ref]. This cross-section presents excellent spatial and temporal resolution, allowing a detailed view of the valve shape and function, which is important to evaluate the valvar dysfunction mechanisms. Color Doppler flowmetry is applied to assess aortic regurgitation and the size, mechanism, and position of the regurgitant orifice. [bib_ref] Recommendations for quantification of Doppler echocardiography: a report from the Doppler quantification..., Quiñones [/bib_ref] Probe removal or anteflexion can display the left coronary ostium image, as well as its introduction or retroflexion can provide an image of the LVOT short axis.
## Mid-esophageal long axis cross-section
The multiplane angle is rotated to 120 - , from the midesophagial four-chamber cross-section, and the basal portion of IVS, LVOT, aortic root (aortic ring, Valsalva sinus, and sinotubular junction), and Asc Ao proximal tubular portion appears on the right side of the image [fig_ref] Figure 2: Probe and transducer handling movements to acquire echocardiographic images [/fig_ref]. Tw o AV valves are displayed, the right coronary is always that most distal to the transducer because it is the most anterior. The other cusp present in this cross-section may be the left coronary, or most often the non-coronary, depending on the commissure position between them and the exact location of the imaging plate as it passes through the AV. Presence of calcification, thickening, motion degree, and valve opening, as well as its anatomical relationship with adjacent structures, such as coronary ostia and IVS, should be evaluated. The basal aortic ring diameter measurement is made in this cross-section using 2D TEE, preferably with zoom image, as the distance between the most ventricular nadir of the right coronary valve and the base of the fibrous triangle (''fima''), contralateral and orthogonal to the aortic root longitudinal axis, in the ventricular systole. This measure is usually smaller than that of the coronal axis (only obtained via 3D TEE or tomography) due to the basal ring oval shape. This evaluation accuracy is important for hemodynamic measures of systolic volume and cardiac output, as well as for predicting the size of aortic, percutaneous, or surgical prostheses placed in this region. [bib_ref] Assessment of perioperative hemodynamics, Skubas [/bib_ref] In addition, the diameters of the other components of the aortic root, as well as that of Asc Ao, should be measured. In case of aortic stenosis, where there is proximal flow convergence, the LVOT diameter (located approximately 5 mm before the aortic ring) should be used ---pulsed-Doppler sample volume is placed in the same site to calculate the systolic volume ejection. [bib_ref] Echocardiographic assessment of valve stenosis: EAE/ASE recommendations for clinical practice, Baumgartner [/bib_ref] In this cross-section, the perpendicularity of the US beam incidence in this region flow does not allow accurate velocity measurements with pulsatile and/or continuous Doppler. [bib_ref] Recommendations for quantification of Doppler echocardiography: a report from the Doppler quantification..., Quiñones [/bib_ref] In contrast, color Doppler measurements are useful and reliable, providing important information, such as: turbulent flow regions with LVOT obstructions; measurement of vena contracta regurgitant jet; and relation between regurgitant jet diameter and LVOT diameter. 28
## Transgastric long axis and deep transgastric cross-sections
These cross-sections are essential in the assessment, allowing accurate alignment of LVOT and AV flows, with reliable measurement of their velocities through continuous and pulsed Doppler modes, important for gradation of valvar and/or subvalvular stenoses, regurgitant jets, and measurement of LV ejected systolic volume [fig_ref] Figure 7: A [/fig_ref]. [bib_ref] Assessment of perioperative hemodynamics, Skubas [/bib_ref] [bib_ref] Recommendations for quantification of Doppler echocardiography: a report from the Doppler quantification..., Quiñones [/bib_ref] In this cross-section, evaluations of anatomical structures lose spatial resolution because they are distant from the focus and parallel to the US bundle, the complementary action of trangastric cross-sections with mid-oresophageal cross-sections is emphasized in the complete assessment of AV and LVOT and measurement of ejection systolic volume and cardiac output.
## 2d imaging and doppler of ascending aorta
The necessary cross-sections for Asc Ao complete evaluation are the following: mid-esophageal ascending aorta long axes and short-axes. As mentioned previously, due to the perpendicular incidence of US bundle, these cross-sections have optimal spatial resolution and serve for measurement of Valsalva sinuses, sinotubular junction, and ascending aorta proximal tubular portion. [bib_ref] Assessment of perioperative hemodynamics, Skubas [/bib_ref] Simultaneous orthogonal cross-sectios using 3D transducers are useful for certainty of correct spatial orientation. When measuring Ao diameter, it is particularly important to measure the largest diameter perpendicular to the vessel long axis in that crosssection. [bib_ref] Recommendations for chamber quantification: a report from the American Society of Echocardiography's..., Lang [/bib_ref] Measurements should be made using 2D image because of the risk of underestimation when using M mode, due to the systolic movement of the base of the heart in the apical direction (mean variation of 2 mm in the sinus of Valsalva diameter). [bib_ref] Two-dimensional echocardiographic aortic root dimensions in normal children and adults, Roman [/bib_ref] [bib_ref] Normal values of aortic root dimensions in healthy adults, Vriz [/bib_ref] Color Doppler of these regions is important in the identification of abnormal flow turbulences and characterization of pathologies, such as aortic dissection, intramural hematomas, and other acute aortic syndromes. 28
## 2d imaging and doppler of descending aorta
From TG plane to upper esophageal plane, Ao can be visualize by rotating the probe in the posterior direction and adjusting the image to a depth of 6 cm. From TG plane, the probe is withdrawn in small increments while the proximal abdominal Ao and its branches, such as the renal artery, are evaluated [bib_ref] A transesophageal echocardiography technique to locate the kidney and monitor renal perfusion, Bandyopadhyay [/bib_ref] [fig_ref] Figure 2: Probe and transducer handling movements to acquire echocardiographic images [/fig_ref] , as well as the thoracic descending Ao along its length, stopping for a better analysis if there is any clinically significant lesion [fig_ref] Figure 2: Probe and transducer handling movements to acquire echocardiographic images [/fig_ref]. From gastric plane to mid-esophagial plane, the transverse image of Ao is generated at zero degree. When reaching the upper esophagus, the aortic arch appears in longitudinal view, due to its anatomical position at that level [fig_ref] Figure 9: A [/fig_ref]. If in this plane the angle is turned to approximately 90 - , the left subclavian artery exit can be evaluate and, turning the probe to the left, part of its extension can be evaluated [fig_ref] Figure 9: A [/fig_ref]. With an anti-clockwise rotation, the two other brachiocephalic vessels can be evaluate. The innominate artery view is the most difficult because it is located in a blind spot where the trachea interposes.The use of X-plane, available in 3D transducers, allows the thoracic Ao evaluation in both longitudinal and transverse planes simultaneously, improves the time needed for evaluation, and generates an Kidney Art. Renal Aorta
## Figure 21
Renal artery cross-section.
## Atherosclerotic plaques
## Figure 22
Atheroma plaques in descending aorta short axis.
image with correct spatial orientation of plaques that may be present and need to be evaluated. [bib_ref] Rapid and accurate assessment of aortic arch atherosclerosis using simultaneous multi-plane imaging..., Ito [/bib_ref] Color Doppler image of thoracic Ao is used for abnormal flow evaluation, particularly in acute aortic syndromes. Pulmonary Doppler is used to identify holodiastolic reflux, an important component of qualitative gradding of aortic regurgitation [fig_ref] Figure 2: Probe and transducer handling movements to acquire echocardiographic images [/fig_ref]. [bib_ref] Recommendations for quantification of Doppler echocardiography: a report from the Doppler quantification..., Quiñones [/bib_ref] Due to the variable anatomical relationship between the esophagus and thoracic Ao, it is difficult to determine the anterior-posterior and right-left orientations on echocardiographic images of this region. The definition of its relation with adjacent anatomical structures, such as LA, pulmonary vein, pulmonary artery, and LV becomes a useful tool for this position definition. 1
## 3d imaging of aortic valve and aorta
The 3D echocardiographic image is acquired by capturing an anatomical image volume [fig_ref] Figure 2: Probe and transducer handling movements to acquire echocardiographic images [/fig_ref] , defined in its size by the operator, unlike the 2D echocardiography acquired by slices of these images. [bib_ref] EAE/ASE recommendations for image acquisition and display using three-dimensional echocardiography, Lang [/bib_ref] This volume contains a more accurate information on the region of interest, allows the collection of detailed information through an ''electronic dissection'' of the aquired volume. [bib_ref] EAE/ASE recommendations for image acquisition and display using three-dimensional echocardiography, Lang [/bib_ref] This captured volume allows the acquisition of two-dimensional image slices, with spatial orientation appropriate to the desired information, through a resource called multiplane reconstruction 37,38 [fig_ref] Figure 2: Probe and transducer handling movements to acquire echocardiographic images [/fig_ref]. However, by acquiring more ultrasound information at each capture, we have a loss in temporal resolution proportional to the size of our region of interest. With this kept in mind, we must always balance the temporal and spatial resolutions by acquiring the smallest amount of information required and/or in several acquisitions synchronized by the electrocardiogram (ECG). [bib_ref] 3D echocardiography: a review of the current status and future directions, Hung [/bib_ref] In addition, 3D transducers allow the acquisition of up to three simultaneous two-dimensional images of different planes, which are very useful in the evaluation of LVOT, AV, and Ao 39 [fig_ref] Figure 2: Probe and transducer handling movements to acquire echocardiographic images [/fig_ref]. Detection of complex plaques in Ao is of great clinical importance due to the association between them and the risk of embolization and mortality in patients undergoing cardiothoracic surgery. As described previously, the use of Ao simultaneous biplanar images allows to save time and gain precision in the evaluation of atherosclerotic plaques [fig_ref] Figure 2: Probe and transducer handling movements to acquire echocardiographic images [/fig_ref]. [bib_ref] Aortic arch atheroma in transient ischemic attack patients, Guidoux [/bib_ref] Once identified, these plaques should also be evaluated on 3D mode, which adds diagnostic quality and sensitivity to this investigation, reason why it is the method of choice for aortic atherosclerotic plaques evaluation. [bib_ref] Transoesophageal echocardiography of aortic atherosclerosis: the additive value of three-dimensional over two-dimensional..., Weissler-Snir [/bib_ref] All these technical advantages are applied for detailed evaluation of aortic syndromes, such as aortic dissection, in which the 3D exam provides additional information, particularly in the quantification of the inflow orifice, in addition to allowing a better morphological understanding when the intimal dissection occurs in a spiral. [bib_ref] Usefulness of realtime three-dimensional transoesophageal echocardiography in the assessment of chronic aortic..., Evangelista [/bib_ref] The increased information on diagnostic evaluation is repeated in therapeutic procedures, when it becomes a valuable intraoperative tool to improve the operator's work, during and after the positioning of Ao thoracic stents. [bib_ref] Three-dimensional intraoperative echographic monitoring for endovascular stent-graft repair in a patient with..., Catena [/bib_ref] With the exponential expansion of the percutaneous treatments of AV pathologies, there was a great interest in the precise anatomical definition of LVOT and AV, through the evaluation of the shape and measurements of diameters, area and planimetric perimeters. This definition influences influence Holodiastolic reflux the outcome of these procedures, as far as two-year mortality is concerned, [bib_ref] Predictors of one-year mortality after transcatheter aortic valve implantation for severe symptomatic..., Zahn [/bib_ref] [bib_ref] Two-year outcomes after transcatheter or surgical aortic-valve replacement, Kodali [/bib_ref] and could only be measured with 3D images [fig_ref] Figure 2: Probe and transducer handling movements to acquire echocardiographic images [/fig_ref] , since calculations using the mathematical formulas of 2D echocardiography were shown to be significantly inaccurate. [bib_ref] Three-dimensional imaging of the left ventricular outflow tract: impact on aortic valve..., Gaspar [/bib_ref] [bib_ref] Assessment of the aortic root using real-time 3D transesophageal echocardiography, Otani [/bib_ref] The identification that LVOT has an elliptical shape in most patients increased the importance of 3D TEE due to the inability of 2D echocardiography to perform these important measurements accurately. [bib_ref] Three-dimensional imaging of the left ventricular outflow tract: impact on aortic valve..., Gaspar [/bib_ref] The distant location from the transducer focus, the distal curvature, the thin thickness of the AV valves, and the artifacts resulting from the reverberation and acoustic shadow of the existing calcifications are some of the challenges for the acquisition of 3D echocardiographic imaging of the aortic root. [bib_ref] Three-dimensional imaging of the left ventricular outflow tract: impact on aortic valve..., Gaspar [/bib_ref] Color Doppler 3D should also be used to evaluate normal flow and abnormalities, allow two-dimensional cuts with perfect spatial orientation, and the odd analysis of all abnormal flow aspects, such as measurement of the contracted vein area, which is important to quantify these flows, without the need to apply geometric formulas that are imprecise for the situation. 37
## Pulmonary valve
Anatomy PV is a thoracic anterior structure located obliquely to the plane of the AV. [bib_ref] The feasibility of simultaneous orthogonal plane imaging with tilt for short-axis evaluation..., Dwarakanath [/bib_ref] It is a semilunar valve comprising three valves denominated by their positions relative to the AV (left, right, and anterior). [bib_ref] The feasibility of simultaneous orthogonal plane imaging with tilt for short-axis evaluation..., Dwarakanath [/bib_ref] Unlike PV, there is no continuity of PV with the cardiac fibrous skeleton or right atrioventricular valve. [bib_ref] The feasibility of simultaneous orthogonal plane imaging with tilt for short-axis evaluation..., Dwarakanath [/bib_ref] Because its valves are thin and poorly echogenic and because it is far from the transesophageal probe, PV evaluation using TEE is generally difficult. 1
## 2d and doppler modes
RVOT, PV, and pulmonary artery dimensions are best evaluated through the cross-sectional view of the mid-esophageal RV inflow and outflow tracts, in which the US beam is perpendicular to these structures [fig_ref] Figure 5: A [/fig_ref]. [bib_ref] The feasibility of simultaneous orthogonal plane imaging with tilt for short-axis evaluation..., Dwarakanath [/bib_ref] The anterior valve in this cross-section is the one furthest from the transducer, whereas the one closest to the AV may correspond to the right or left valve. In cases where these sectional view is of poor quality, TG cross-sections of the RV inflow and outflow tracts, upper esophageal aortic arch short axis [fig_ref] Figure 9: A [/fig_ref] , and mid-esophageal Asc Ao short and long axes [fig_ref] Figure 4: A [/fig_ref] and D) may be used to supplement information on structure, dimensions, and function of the pulmonary valve and artery. [bib_ref] Guidelines for performing a comprehensive transesophageal echocardiographic examination: recommendations from the American..., Hahn [/bib_ref] The presence of regurgitation or flow acceleration can be assessed by color Doppler mode in the mid-esophageal RV inflow and outflow tracts view.
## 3d pulmonary valve
PV 3D images can be acquired on upper esophageal aortic arch short axis or mid-esophageal aortic valve long axis cross-sections after a slight rotation of the probe to the left. [bib_ref] Guidelines for performing a comprehensive transesophageal echocardiographic examination: recommendations from the American..., Hahn [/bib_ref] [bib_ref] The feasibility of simultaneous orthogonal plane imaging with tilt for short-axis evaluation..., Dwarakanath [/bib_ref] By allowing the simultaneous view of two planes, modern 3D probes enabled the assessment of PV in its short axis. For this purpose, the cross-sectional view of the mid-esophageal RV inflow and outflow tracts is obtained and an orthogonal plane is positioned over the PV. [bib_ref] Guidelines for performing a comprehensive transesophageal echocardiographic examination: recommendations from the American..., Hahn [/bib_ref] Multiplanar image 3G image [fig_ref] Figure 2: Probe and transducer handling movements to acquire echocardiographic images [/fig_ref] Three-dimensional image of descending aorta atherosclerotic plaque.
## Left atrium and pulmonary veins anatomy
In relation to the rib cage, the LA is the most posterior chamber of the heart. It is located very close to the esophagus, separated only by the fibrous pericardium. [bib_ref] Anatomy of the left atrium for interventional echocardiography, Ho [/bib_ref] Trachea bifurcation, esophagus, and Desc Ao are immediately behind the LA posterior wall. This LA proximity to the esophagus is very advantageous for TEE, as LA is used as a window to obtain the mid-transesophageal views. [bib_ref] Guidelines for performing a comprehensive transesophageal echocardiographic examination: recommendations from the American..., Hahn [/bib_ref] [bib_ref] Anatomy of the left atrium for interventional echocardiography, Ho [/bib_ref] In these views, LA is always at the top of the screen, close to the US beam . [bib_ref] Guidelines for performing a comprehensive transesophageal echocardiographic examination: recommendations from the American..., Hahn [/bib_ref] In relation to RA, LA is more posterior and superior and is separated from RA by IAS. [bib_ref] Left atrial anatomy revisited, Ho [/bib_ref] LA walls can be described as superior, posterior, left lateral, septal (or medial), posterior---inferior, and anterior.LA anterior wall is behind the transverse sinus, that is, behind the Ao root. CS goes through the posterior---inferior wall of the LA. LA walls are muscular and its thickness may vary from 1 ± 0.5 mm. Abnormal wall thickening may indicate the presence of a mural thrombus or even endocarditis. Mitral annulus calcification may extend to the LA wall and make it thicker. [bib_ref] Guidelines for performing a comprehensive transesophageal echocardiographic examination: recommendations from the American..., Hahn [/bib_ref] [bib_ref] Left atrial anatomy revisited, Ho [/bib_ref] LAA is a blind bottom structure with an aperture to LA. In relation to LA, it is located lateral and superior and its tip is directed anteriorly, overlaps pulmonary artery trunk and left coronary artery. [bib_ref] Anatomy of the left atrium for interventional echocardiography, Ho [/bib_ref] LAA has an average diameter of 10---24 mm, but this value may vary. Compared to RAA, LAA has a narrower orifice that facilitates the formation of thrombi in situations of low blood flow and in non-sinus cardiac rhythms. [bib_ref] Anatomy of the left atrium for interventional echocardiography, Ho [/bib_ref] Between LAA and ULPV there is a triangular fold of the serous pericardium called coumadin ridge, which when prominent may be confused with thrombus or atrial mass [fig_ref] Figure 6: A [/fig_ref].
According to the blood flow, LA starts at the venoatrial junction and ends at the MV orifice. The four pulmonary veins enter the posterior part of the LA, the left veins are located more superior than the right veins. [bib_ref] Anatomy of the left atrium for interventional echocardiography, Ho [/bib_ref] There are two right (RUPV and RLPV) and two left pulmonary veins (LUPV and LLPV). The RUPV passes behind the RA junction with SVC and the RLPV passes behind the intercaval area. [bib_ref] Left atrial anatomy revisited, Ho [/bib_ref] The right pulmonary veins orifices are directly adjacent to the IAS plane. On the other hand, the left pulmonary veins are located between LAA and Desc Ao, ULPV is positioned posteriorly superior to the IAS, whereas LLPV is positioned posteriorly inferior. [bib_ref] Left atrial anatomy revisited, Ho [/bib_ref] [bib_ref] Intraoperative pulmonary vein examination by transesophageal echocardiography: an anatomic update and review..., Cartwright [/bib_ref] Physiology LA is not just a simple blood-carrying camera. It responds dynamically to its distention with the secretion of atrial natriuretic peptides and participates in the management of body fluids. [bib_ref] Left atrial function: physiology, assessment, and clinical implications, Blume [/bib_ref] LA also functions as a blood reservoir from the pulmonary veins during ventricular systole and isovolumetric relaxation. During diastole, LA works as blood conduit to LV. At the end of diastole, atrial systole occurs, contributing with 15---30% of the volume transferred to LV. [bib_ref] Mechanical function of the left atrium ---new insights based on analysis of..., Pagel [/bib_ref] Because LA is an LV continuum during diastole, changes in ventricular compliance affects the size and function of LA. [bib_ref] Mechanical function of the left atrium ---new insights based on analysis of..., Pagel [/bib_ref] [bib_ref] Compensatory changes in atrial volumes with normal aging: is atrial enlargement inevitable?, Thomas L [/bib_ref] The increase in LA is a predictor of cardiovascular adverse events, as well as the onset of atrial fibrillation, intracardiac thrombi formation, and cerebrovascular accidents [fig_ref] Figure 2: Probe and transducer handling movements to acquire echocardiographic images [/fig_ref]. [bib_ref] Left atrial size and the risk of stroke and death ---the Framingham..., Benjamin [/bib_ref] When a thromboembolic source is investigated, LAA is the first site to be evaluated; TEE has a sensitivity and specificity for thrombus diagnosis of 100% and 99%, respectively. However, because they are small and complex structures, they can be left undiagnosed in some circumstances. 59
## 2d and doppler imaging of left atrium and left atrial appendage
The most frequently used cross-sections for LA evaluation are [bib_ref] Guidelines for performing a comprehensive transesophageal echocardiographic examination: recommendations from the American..., Hahn [/bib_ref] - Mid-esophageal four-chamber [fig_ref] Figure 3: A [/fig_ref] ;
- Mid-esophageal commissural [fig_ref] Figure 3: A [/fig_ref] ;
- Mid-esophageal two-chamber [fig_ref] Figure 3: A [/fig_ref] ;
- Mid-esophageal left atrial appendage [fig_ref] Figure 6: A [/fig_ref] ;
- Mid-esophageal long axis [fig_ref] Figure 4: A [/fig_ref] ;
- Mid-esophageal aortic valve short axis [fig_ref] Figure 5: A [/fig_ref] ;
- Mid-esophageal bicaval [fig_ref] Figure 5: A [/fig_ref] ;
- Transgastric two-chamber [fig_ref] Figure 8: A [/fig_ref] ;
- Deep transgastric [fig_ref] Figure 7: A [/fig_ref].
The cross-sectional images that most easily evaluate the LA are the mid-esophageal ones. LA is the structure closest to the US bundle in the mid-esophageal views; it is always found at the top of the screen. TEE cross-sections that evaluate MV will necessarily evaluate LA. [bib_ref] Guidelines for performing a comprehensive transesophageal echocardiographic examination: recommendations from the American..., Hahn [/bib_ref] LA evaluation starts from the mid-esophageal fourchamber view [fig_ref] Figure 3: A [/fig_ref]. Rotate the angle, going through the other views to obtain 2D cross-sections. In two-chamber view [fig_ref] Figure 3: A [/fig_ref] , LAA image should be increased [fig_ref] Figure 6: A [/fig_ref] to better visualize the presence of thrombi, particularly in risk patients. In bicaval view, the relationship between LA and IAS and RA is evaluated, looking for PFO and IAS defects. In these cross-sections, the pulmonary veins will also be evaluated, as will be discussed below. After the esophageal evaluation, the transducer is advanced to the TG cross-sections. At the 90 - angle, LA and especially MV and its subvalvar apparatus are evaluated. In deep TG crosssection [fig_ref] Figure 7: A [/fig_ref] , LA can also be visualized, but this view is more used for transaortic flow evaluation; LA is a secondary evaluation because it is further away from the US beam, with a lower resolution compared to transesophageal cross-sections.
Due to the proximity between LA and esophageal transducer, LA cannot be evaluated in its entirety by a single view, which makes its complete evaluation and its diameter and volume measurements difficult on TEE. 1 LA area and volume may be underestimated. The linear measurements acquired on TEE in mid-esophageal aortic valve long and short axes views are the ones that best correlate with the TEE parasternal long axis cross-sectional measurement, which measures LA in the anterior-posterior sense. [bib_ref] Comparison of left atrial dimensions by transesophageal and transthoracic echocardiography, Block [/bib_ref] In longitudinal sense, the vertex should be measured from sector to Ao root, but these measurements have no normalized values. Septolateral measurements and LA volume can be obtained with mid-esophageal four-and two-chamber cross-sections. However, these measurements, although correlated with those obtained on TEE, have no normalized values. [bib_ref] Comparison of left atrial dimensions by transesophageal and transthoracic echocardiography, Block [/bib_ref] LAA evaluation may be initiated with the four-chamber cross-section, but because it is a lateral structure, the probe should be rotated counterclockwise and anteflex to bring LAA to the screen center. 1 Also in the two-chamber cross-section, LAA site can be zoomed in or the screen depth be reduced to evaluate LAA more accurately. An organized thrombus is echocardiographically defined as a well-circumscribed mass of uniform consistency and a texture different from that of the atrial wall [fig_ref] Figure 2: Probe and transducer handling movements to acquire echocardiographic images [/fig_ref]. 1 Spontaneous contrast is not well circumscribed, has a dynamic localization, and appears as a ''cigarette smoke'' ---this image is caused by the slowing down of blood flow [fig_ref] Figure 2: Probe and transducer handling movements to acquire echocardiographic images [/fig_ref]. [bib_ref] Spontaneous echocardiographic contrast in the descending aorta, Castello [/bib_ref] In high-risk patients for LAA thrombus (atrial fibrillation), the risk of thrombus formation should be assessed through the analysis of blood flow velocity. [bib_ref] Spontaneous echocardiographic contrast in the descending aorta, Castello [/bib_ref]
## 2d and doppler imaging of pulmonary veins
ULPV is the most easily found by TEE. When LAA is found in the 60 - mid-esophageal plane, the probe is gently removed and ULPV will appear above and posterolateral to LAA and to the coumadin ridge (anatomic variant that is occasionally found in the left atrium) [fig_ref] Figure 6: A [/fig_ref]. [bib_ref] Guidelines for performing a comprehensive transesophageal echocardiographic examination: recommendations from the American..., Hahn [/bib_ref] The LLPV is the most difficult pulmonary vein to obtain on TEE image. After the ULPV image is obtained, the angle is increased to 90 - . The left pulmonary veins will appear as an inverted V-shape. Colored Doppler is set to a velocity of 40 cm.s −1 and laminar blood flow is observed toward the TEE transducer. Another way to acquire the LLPV image is, after finding LAA, advance the probe and rotate it slightly clockwise and when LAA has disappeared, LLPV will be visible. [bib_ref] Guidelines for performing a comprehensive transesophageal echocardiographic examination: recommendations from the American..., Hahn [/bib_ref] RUPV is found with the mid-esophageal bicaval crosssection at 110---120 - , near the right pulmonary artery. [bib_ref] Doppler and hemodynamics perioperative twodimensional transesophageal echocardiography, Vegas [/bib_ref] The angulation is returned to 90 - , and the two right pulmonary veins in inverted Y-shape can be evaluated. The right pulmonary veins can be visualized at zero degree midesophageal by rotating the probe to the right so that the right side of LA lies in the central part of the screen. From this position, the angle is opened to 30 - and RUPV is located to the right of the screen and RLPV to the left of the screen, entering perpendicular to LA. [bib_ref] Doppler and hemodynamics perioperative twodimensional transesophageal echocardiography, Vegas [/bib_ref] To assess the blood flow pattern of pulmonary veins, apply pulsed Doppler with a velocity limit of 40 cm.s −1 . The normal flow pattern is three-phase with systolic (S1 and S2), diastolic (D), and reverse atrial (A) waves. Diastolic dysfunction, mitral diseases, and rhythm disturbances change this normal pattern. 1
## Right atrium, tricuspid valve and venous connections
RA is the heart cavity that receives systemic venous blood from IVC and SVC, as well as blood returning from the coronary arteries through CS. Its medial and posterior wall is the IAS, the structure that separates it from LA. Its floor is the TV, which opens into the right ventricle. [bib_ref] Echo didactics: echocardiographic assessment of atrial septal defects, Burch [/bib_ref] Seen from the right side, IAS has a characteristic structure, an oval fossa, which shows outward contour and central region constituted by a delicate blade. This blade most anterior portion may not be completely adhered to the oval fossa edge, the so called PFO. [bib_ref] The significance of patent foramen ovale. A current review of associated conditions..., Johansson [/bib_ref] Necropsy studies suggest that it is present in up to 27% of adults. Its in vivo diagnosis depends on dynamic evaluation with maneuvers that cause increased RA pressure concomitant with the injection of agitated saline solution. A higher prevalence has also been associated with the presence of IAS aneurysm. [bib_ref] The significance of patent foramen ovale. A current review of associated conditions..., Johansson [/bib_ref] The LA lower portion is separated from LV by a portion of fibrous tissue that continues with IVS, called fibrous septum. This is due to the different levels of implantation of the tricuspid and mitral valves. TV has more apical insertion, which results in the area known as atrioventricular septum.
CS outflow is located posteriorly and medially to the IVC ouflow next to the atrioventricular transition. In this region, remnants of venous valves may be found, with the Eustachian next to IVC and the Thebesian related to CS. [bib_ref] Echo didactics: echocardiographic assessment of atrial septal defects, Burch [/bib_ref] RA also has two important structures for cardiac automatism: sinus node and atrioventricular (AV) node. The sinus node is located near the SVC outflow, while the AV node is close to TV. TV consists of fibrous annulus, chordae tendineae, papillary muscles, and three leaflets. [bib_ref] Guidelines for performing a comprehensive transesophageal echocardiographic examination: recommendations from the American..., Hahn [/bib_ref] RAA is an atrial cavity projection, shaped like a ''glove finger'', covering the AV groove on the right [fig_ref] Figure 5: A [/fig_ref]. The inner RAA surface has parallel muscle ridges, which extend posteriorly, named pectinate muscles, ending in a transverse muscle band rather prominent named terminal ridge. 66
## 2d echocardiography and doppler assessment
Starting with the mid-esophageal four-chamber plane [fig_ref] Figure 3: A [/fig_ref] , the general aspect of RA can be assessed, its size relationship with the other cardiac chambers, as well as the mid-anterior and inferior portions of IAS, corresponding to the oval fossa and septum primum region. [bib_ref] Guidelines for performing a comprehensive transesophageal echocardiographic examination: recommendations from the American..., Hahn [/bib_ref] From this position, the transducer is lowered toward the TG planes, and the angle is maintained at zero degree, eventually a slight retroflection of the probe is performed, and a longitudinal image of the CS is acquired, a lower and posterior structure. [bib_ref] Guidelines for performing a comprehensive transesophageal echocardiographic examination: recommendations from the American..., Hahn [/bib_ref] Occasionally, CS can be assessed from the midesophageal bicaval plane [fig_ref] Figure 5: A [/fig_ref] , with a discreet advancement and turning the probe clockwise. Still in the bicaval plane [fig_ref] Figure 5: A [/fig_ref] , SVC, IVC, Eustachian valve, terminal crest, and RAA are identified. From this view, the oval fossa is also well defined, as well as a PFO eventual blade. A discreet movement of the probe toward the upper esophagus allows a more complete picture of SVC. At this point, rotate the probe clockwise and the right upper and lower pulmonary veins will be visualized. Thus, the suspicion of anomalous drainage of these veins is typically investigated from this point of view. [bib_ref] Guidelines for performing a comprehensive transesophageal echocardiographic examination: recommendations from the American..., Hahn [/bib_ref] By introducing the probe toward the TG planes from the bicaval plane and maintaining the angle at 90 - , IVC and Eustachian valve can evaluated. Moreover, this position may be useful to evaluate the image of hepatic veins and their possible inadvertent cannulation during cardiopulmonary bypass. 1
## Interatrial shunt diagnosis
In suspected cases of atrial septal defect, a complete examination should include a comprehensive assessment of IAS in 2D mode and color Doppler, as defects may occur in any location. 67
## Assessment of left ventricular size and function
The evaluation of LV size and function is an important component of any perioperative echocardiographic exam. [bib_ref] The association of the left ventricular ejection fraction, mortality, and cause of..., Curtis [/bib_ref] The degree of ventricular systolic dysfunction besides being a strong predictor of clinical outcome aids in the stratification of surgical risk and therapeutic interventions. [bib_ref] The association of the left ventricular ejection fraction, mortality, and cause of..., Curtis [/bib_ref] Echocardiography provides a global and segmental assessment of ventricular performance through analysis of systolic thickening, ventricular size and volume. [bib_ref] Transesophageal echocardiographic (TEE) evaluation of ventricular function, Skiles [/bib_ref] Qualitative and quantitative measures to estimate ventricular function can be made through 2D, 3D, application of Doppler and by measurements of myocardial velocity and deformation. 69
## Anatomy
LV is a thick wall cavity with a conical shape, which decreases in diameter from the base to the apex, appears as a circular structure in the transverse plane. Through this same plan the IVS myocardium follows the LV shape and is part of it from the anatomical and functional point of view. [bib_ref] EAE/ASE recommendations for acquisition and display using three-dimensional echocardiography, Lang [/bib_ref] In order to facilitate an accurate description of the location and severity of LV segmental changes, besides allowing a standardized communication of echocardiography with other cardiovascular imaging methods, LV is divided into 17 segments. [bib_ref] EAE/ASE recommendations for acquisition and display using three-dimensional echocardiography, Lang [/bib_ref] In this model, LV is divided into three levels: basal, at the mitral valve level (six segments); medial, at the papillary muscle level (six segments); and apical, after insertion of papillary nerves up to the end of the cavity (four segments), with the 17th segment located at the tip of LV. [bib_ref] Recommendations for chamber quantification: a report from the American Society of Echocardiography's..., Lang [/bib_ref] The TEE cross-sectional image recommended for measurement of LV diameters are the mid-esophageal two-chamber and long axis and the TG two-chamber. Measurement is done from the endocardium of the anterior wall to the endocardium of the inferior wall, between the basal and medial third of the ventricle. The proposed view to measure the left ventricular wall thickness is the TG mid-short axis 30 [fig_ref] Figure 3: A [/fig_ref].
## 2d and 3d imaging of left ventricle
The assessment of left ventricular function and structure with 2D echocardiography uses five main cross-sections from mid-esophageal and TG planes, which are: midesophageal four-chamber, mid-esophageal two-chamber, mid-esophageal longitudinal, TG mid-short axis, and TG longitudinal. 1,2 The most used cross-section for LV segmental changes monitoring is the TG mid-short axis, in which we can visualize the territories irrigated by the three main coronary arteries 2,6 [fig_ref] Figure 3: A [/fig_ref]. The analysis of LV segmental function is based on the visual qualitative evaluation of parietal motion and systolic thickening. Ideally, the function of each segment should be assessed through multiple incidences. [bib_ref] Current and evolving echocardiographic techniques for the quantitative evaluation of cardiac mechanics:..., Mor-Avi [/bib_ref] LV parietal motion follows this recommendation: normal segments, hypokinetic segments (mild to moderate thickening), akinetic segments (without thickening), and dyskinetic segments (paradoxical systolic motion). [bib_ref] ASE/SCA Guidelines for performing a comprehensive intraoperative multiplane transesophageal echocardiography examination: recommendations..., Shanewise [/bib_ref] Despite the absence of reference values and reproducibility below ideal, the quantitative evaluation of the regional LV strain magnitude is promising, mainly through the longitudinal strain during ventricular systole. [bib_ref] Current and evolving echocardiographic techniques for the quantitative evaluation of cardiac mechanics:..., Mor-Avi [/bib_ref] In order to improve the image quality for ventricular function correct quantification and interpretation, some technical considerations must be observed: correct image depth adjustment to include the entire LV; avoid apical region shortening through proper manipulation of the probe (anteflexion and retroflexion); correct identification of the end of systole and diastole (check the motion of mitral and aortic valves, the largest and smallest cavity size and ECG signal); correct Doppler alignment with the direction of blood flow; appropriate gain and focus adjustment to improve endocardium visualization; and use of second harmonic image. [bib_ref] New generation 3-dimensional echocardiography for left ventricular volumetric and functional measurements: comparison..., Nikitin [/bib_ref] Perhaps one of the most valuable perioperative contributions of 3D echocardiography is related to the quantification
## Quantification of global lv systolic function
One of the most routinely used methods in the operating room for global LV function quantification is qualitative or semi-quantitative, through which a visual estimation of the ventricular ejection fraction is made after the evaluation of multiple orthogonal cuts. This method has an acceptable correlation compared to quantitative measurements. [bib_ref] A visual approach for the accurate determination of echocardiographic left ventricular ejection..., Hope [/bib_ref] Quantitatively, the systolic ventricular function estimation is evaluated by parameters that measure the difference between the final diastolic and systolic values related to left ventricular cavity dimension and volumes. Among the quantitative methods, the most used in clinical practice have focused on measuring cardiac output, ejection fraction, shortening fraction, fractional area change, ventricular performance index (Tei index), [bib_ref] Value of a Dopplerderived index combining systolic and diastolic time intervals in..., Ye O [/bib_ref] [bib_ref] A randomized blinded study of the left ventricular myocardial performance index comparing..., Filho [/bib_ref] and on methods that evaluate the velocity and range of myocardial motion and deformation (tissue Doppler, strain and strain rate). [bib_ref] Sociedade Brasileira de Cardiologia. Diretrizes das indicações da ecocardiografia, Barbosa [/bib_ref] [bib_ref] Assessment of left ventricular global and segmental systolic function with transesophageal echocardiography, Odell [/bib_ref]
## Ventricular function assessment limitations
The patient's overall hemodynamic condition should be taken into consideration during ventricular function assessment because changes in blood volume and use of anesthetic drugs may affect systolic function by its effects on pre-and post-load. Moreover, other factors that may compromise the TEE function evaluation are related to LV tip shortening and exclusion in the evaluation of global and segmental contractility and difficulties in correctly aligning Doppler to blood flow, which interferes directly in the calculation of ventricular ejection rates. 1
## Right ventricle anatomy systolic evaluation indexes
RV is a tubular V-shaped or ''bagpipe'' chamber, with tricuspid ring and pulmonary ring forming this V tips. The free, septal, and apical walls delineate the anterior, posterior, and inferior margins. RV anatomical divisions are the inflow, body, and outflow regions. The free wall is subdivided into inferior, anterior, and lateral segments, based on echocardiographic incidences. RV irregular shape makes it difficult to evaluate systolic volume and function with simple uniplanar and geometric methods. Moreover, the RV trabecular interior also creates problems in defining the endocardial border. 1
## Echocardiographic evaluation
Systolic evaluation indexes:
- Geometric indexes: reflect the extent of contraction, such as fractional area variation, ejection fraction, and tricuspid annular plane systolic excursion (TAPSE) [fig_ref] Figure 2: Probe and transducer handling movements to acquire echocardiographic images [/fig_ref] ; - Velocity indexes: isovolemic acceleration [fig_ref] Figure 3: A [/fig_ref] ; - Hemodynamic indexes: right ventricle dp/dt; - Time interval indexes: such as myocardial performance index or Te i index [fig_ref] Figure 3: A [/fig_ref].
The main incidences for RV visualization with TEE are:
- Mid-esophageal four-chamber [fig_ref] Figure 3: A [/fig_ref] : to visualize the free wall, IVS, IAS, and TV (anterior and septal leaflets); - Mid-esophageal RV inflow and outflow [fig_ref] Figure 5: A [/fig_ref]
## Left ventricular diastolic performance
Diastole is no longer considered a passive LV filling period but an important complex period, which depends on adequate ventricular relaxation, compliance and systolic function, intrathoracic pressure, ventricular interaction, cardiac rhythm, and atrial function.In the general population, the asymptomatic prevalence of diastolic dysfunction is approximately 30% in individuals older than 45 years. In surgical patients over 65 years of age, the prevalence of diastolic dysfunction with normal ejection fraction rises to about 60%.In patients undergoing large vascular surgery, the isolated diastolic dysfunction is associated with an increase in 30-day cardiovascular events and long-term mortality. relationship between dysfunction severity and reduced survival In cardiac surgery, diastolic dysfunction is associated with difficult cardiopulmonary bypass weaning, increased need for inotropic support, and increased morbidity. [bib_ref] Perioperative assessment of diastolic dysfunction, Matyal [/bib_ref] In a study published in 2014, Nicoara et al. [bib_ref] Perioperative diastolic dysfunction: a comprehensive approach to assessment by transesophageal echocardiography, Nicoara [/bib_ref] correlated the degree of diastolic dysfunction with survival after surgical procedures and showed a relationship between dysfunction severity and reduced survival.
Diastolic dysfunction is a finding that appears concurrently with a number of cardiovascular diseases, ranging from arterial hypertension to infiltrative diseases such as amyloidosis. [bib_ref] Perioperative diastolic dysfunction: a comprehensive approach to assessment by transesophageal echocardiography, Nicoara [/bib_ref] Heart failure due to diastolic dysfunction is being increasingly diagnosed (about 50% of patients with congestive heart failure have diastolic dysfunction and normal ejection fraction). [bib_ref] Diastolic dysfunction and heart failure with a preserved ejection fraction: relevance in..., Maharaj [/bib_ref] Diastolic function assessment allows the anesthesiologist to detect increases in left ventricular end-diastolic pressure in the absence of a pulmonary artery catheter. [bib_ref] Prognostic implications of preoperative e/e ratio in patients with off-pump coronary artery..., Lee [/bib_ref] Diastolic dysfunction precedes systolic dysfunction in cases of acute ischemia and during the perioperative period; diastolic function assessment may help to guide therapy, with the use of vasodilator drugs rather than inotropic drugs. [bib_ref] Perioperative diastolic dysfunction: a comprehensive approach to assessment by transesophageal echocardiography, Nicoara [/bib_ref] Echocardiography is the safest modality with high sensitivity and specificity to evaluate diastolic function and identify other associated abnormalities.Assessment should be performed at mid-esophageal level, where the best alignment between mitral flow and US bundle is obtained.
Diastole is divided into four phases. The first phase begins with the AV closure and ends with the MV opening, lasting between 90 and 120 ms. As both VA and MV are closed, this period is called the isovolumetric relaxation time (IVR). When the left ventricular pressure falls below the left atrial pressure, the MV opens and the second phase of the diastole begins, that is, the rapid ventricular filling, which corresponds to 80% of the ventricular filling. LV pressure increases during rapid filling and the pressure gradient between LV and LA decreases progressively. This reduction in pressure gradient delays ventricular filling and diastasis occurs, a period of pressure equalization until the beginning of the atrial contraction and which accounts for 5% of the volume. This period increases when there is reduced LV relaxation and decreases when ventricular compliance is decreased. Finally, atrial contraction occurs, contributing with only 20% of the ventricular filling, but in cases of diastolic dysfunction, especially in elderly patients, it may account for 50% of the filling volume. [bib_ref] Perioperative assessment of diastolic dysfunction, Matyal [/bib_ref] If we correlate the above periods of mitral diastolic flow pattern with pulsatile Doppler, we have: (a) isovolumetric relaxation period, time between LVOT and mitral flow (this period increases when there is compromised relaxation and decreases when left atrial pressure increases); (b) fast ventricular filling phase, represented by the E wave; (c) deceleration time (DT), which represents the time required for the pressure to drop from the E wave peak to baseline 79 ; (d) atrial contraction phase, responsible for the diastole fourth phase. This phase is also termed late ventricular filling and is represented by the A wave. [bib_ref] Perioperative assessment of diastolic dysfunction, Matyal [/bib_ref] It should be borne in mind that mitral flow velocities recorded by Doppler are determined by the transmitral pressure gradient, which depends on several variables: rhythm, early filling loads, atrial contractility, MV disease, ventricular septal interactions, LV intrinsic lusitropic state, ventricular relaxation and compliance. [bib_ref] Perioperative assessment of diastolic dysfunction, Matyal [/bib_ref] In this sense, the maximal velocity of the mitral E wave is an indirect measure of left atrial pressure. E wave velocity correlates with the difference between left atrial and left ventricular pressures at the time of mitral opening. Thus, the higher the left atrial pressure (or the higher the preload) at the time of MV opening, the higher the E wave velocity. [bib_ref] Perioperative assessment of diastolic dysfunction, Matyal [/bib_ref] The relationship between the velocities of E and A waves must be greater than 1 [fig_ref] Figure 3: A [/fig_ref]. Normally, this relationship is expressed as E/A > 1. When E < A, it can be said that there is an impairment of left ventricular relaxation [fig_ref] Figure 3: A [/fig_ref]. On the other hand, E wave greater than double the size of A wave represents a restrictive pattern; that is, LV compliance is compromised [fig_ref] Figure 3: A [/fig_ref]. There may however be a time when LV has a diastolic dysfunction in transition; the mitral flow pattern passes from a predominance of relaxation change to a predominance of compliance change. In this case, the E/A relationship is similar to the normal pattern (E > A). This pattern, termed pseudonormal, represents a moderate stage of diastolic dysfunction, in which an earlier, near-normal transmitral pressure gradient is generated by the balance between the impaired LV relaxation and the gradually increasing LV pressures, as compliance decreases. [bib_ref] Perioperative assessment of diastolic dysfunction, Matyal [/bib_ref] In order to differentiate a normal from pseudonormal pattern, mitral annular tissue Doppler (mitral TDI) and pulmonary veins Doppler flow pattern are used [fig_ref] Figure 3: A [/fig_ref]. [bib_ref] Perioperative assessment of diastolic dysfunction, Matyal [/bib_ref] The analysis of mitral TDI helps to differentiate a normal pattern from a pseudonormal because the wave that coincides with the rapid ventricular filling (represented as E ′ or Ea) remains reduced with pseudo-normalization. This approach is based on the Doppler technique to measure mitral annular velocity in diastole. A small volume sample should be used and the gain and filter should be set downward. This velocity profile seems to be more dependent on left ventricular relaxation and less dependent on transmitral pressure gradient. Thus, the E ′ wave assessment is a measurement relatively insensitive to LV preload and may be useful intraoperatively when the loading conditions may vary considerably. [bib_ref] Perioperative assessment of diastolic dysfunction, Matyal [/bib_ref] As diastolic dysfunction progresses, the E ′ wave tends to decrease and the mitral E wave tends to increase, due to the compensatory increase in left atrial pressure that accompanies the impaired relaxation. Thus, an E/E ′ ratio <10 is considered normal, while an E/E ′ ratio >15 predicts a left ventricular filling pressure above 15 mmHg. [bib_ref] Perioperative assessment of diastolic dysfunction, Matyal [/bib_ref] The flow pattern of pulmonary veins also has a systolic and a diastolic component. The systolic component can be divided in two: a first time when the flow accompanies the atrial relaxation and a second time that accompanies the mitral annulus displacement toward the left ventricular apex. The diastolic component occurs when the MV opens. At the end of diastole, simultaneously with the atrial contraction, a reverse flow is observed, representing blood from the LA toward the pulmonary veins. This reverse flow assessment is also important for diastolic function evaluation: an increased retrograde flow reflects an increased LV end-diastolic pressure. [bib_ref] Perioperative assessment of diastolic dysfunction, Matyal [/bib_ref] [bib_ref] Recommendations for transoesophageal echocardiography: update 2010, Flachskampf [/bib_ref] [fig_ref] Table 3: Classification of diastolic dysfunction [/fig_ref] shows the patterns of diastolic function schematically. [bib_ref] Perioperative assessment of diastolic dysfunction, Matyal [/bib_ref] Another parameter that may be used in diastolic function assessment is the propagation velocity (Vp) of transmitral flow within LV in color M mode. It has the advantage of being relatively independent of preload. This parameter reflects the effectiveness of LV suction at the onset of diastole. Values below 50 cm.s −1 are consistent with impairment of ventricular relaxation [fig_ref] Figure 3: A [/fig_ref]. [bib_ref] Perioperative assessment of diastolic dysfunction, Matyal [/bib_ref] [bib_ref] Perioperative diastolic dysfunction: a comprehensive approach to assessment by transesophageal echocardiography, Nicoara [/bib_ref] Recent studies show that Vp lower than 40 cm.s −1 may be a predictor of required cardiovascular support after aortic valve replacement due to stenosis. In addition, Vp may be useful in estimating filling pressures, as an E/Vp ratio greater than 2.5 predicts a pulmonary capillary occlusion pressure greater than 15 mmHg. [bib_ref] Perioperative assessment of diastolic dysfunction, Matyal [/bib_ref] [bib_ref] Perioperative diastolic dysfunction: a comprehensive approach to assessment by transesophageal echocardiography, Nicoara [/bib_ref] It is important to emphasize that diastolic function evaluation is subject to functional changes in MV itself, such as stenosis and insufficiency. Because much of the diastole interpretation is based on transmitral flow, the diastolic function assessment is impaired in patients with mitral valvopathy.
Patients with arrhythmias should also have diastolic function evaluated carefully, as transmitral flow is affected by heart rate and rhythm. Sinus tachycardia and firstdegree atrioventricular blockade may lead to fusion of E and A waves, hampering the visualization of both, as well as the assessment of deceleration time. Several atrioventricular blockades can lead to different atrial filling waves and also to mitral regurgitation in beats without conduction. [bib_ref] Perioperative assessment of diastolic dysfunction, Matyal [/bib_ref] In cases of flutter and atrial fibrillation, A wave is non-existent, but deceleration time correlates well with left ventricular filling when systolic function is depressed. In these cases, E wave evaluation is also feasible and presents the same cut-off values of patients with sinus rhythm.Regarding myocardial ischemia, mitral annular tissue Doppler has been used as a reliable and early marker. Reduction of E ′ wave velocity and E ′ /A ratio, which precede abnormalities in myocardial segmental motion, is observed. [bib_ref] Perioperative assessment of diastolic dysfunction, Matyal [/bib_ref] Perioperatively, multiple factors can alter and decompensate diastolic function. As seen so far, tachycardia and arrhythmia, which occur in hypovolemia and anemia, and hydroelectrolytic changes may hinder left ventricular filling. Myocardial ischemia changes calcium output from cytosol and the decoupling of actin---myosin bridges, which leads to worse LV relaxation. Increased preload, afterload, myocardial wall tension, and non-synchronous contraction lead to late and incomplete relaxation. [bib_ref] Perioperative assessment of diastolic dysfunction, Matyal [/bib_ref] In this scenario, Matyal et al. [bib_ref] Perioperative assessment of diastolic dysfunction, Matyal [/bib_ref] suggest that the perioperative evaluation of diastolic function should be different in patients with impaired systolic function and normal systolic function. In patients with impaired systolic function, the assessment of diastolic function should begin with analysis of mitral E/A ratio [fig_ref] Figure 3: A [/fig_ref] and, in patients with preserved systolic function, diastolic evaluation should begin with the analysis of E/E1 ′ ratio [fig_ref] Figure 3: A [/fig_ref].
From the diagnosis of diastolic dysfunction and severity classification, hemodynamic targets can be established. Although future studies are required to confirm the exact accuracy of the therapeutic strategies used, it may be suggested that in patients with altered relaxation, increased left atrial pressure, and frequency control (preferably bradycardia) lead to improved ventricular filling, as they increase the pressure gradient between LA and LV, in addition to increase the filling time. [bib_ref] Perioperative assessment of diastolic dysfunction, Matyal [/bib_ref] In patients with altered compliance, water restriction and even judicious use of diuretics, in addition to an increased blood pressure to improve coronary perfusion pressure, may be appropriate choices.
[fig] Figure 2: Probe and transducer handling movements to acquire echocardiographic images. Adapted from Galhardo et al.6 [/fig]
[fig] Figure 3: A) Mid-esophageal five-chamber cross-section. (B) Mid-esophageal four-chamber cross-section. (C) Mid-esophageal commissural cross-section. (D) Mid-esophageal two-chamber cross-section. [/fig]
[fig] Figure 4: A) Mid-esophageal long axis cross-section. (B) Mid-esophageal ascending aorta long axis cross-section. (C) Midesophageal aortic valve long axis cross-section. (D) Mid-esophageal ascending aorta short axis. [/fig]
[fig] Figure 5: A) Mid-esophageal aortic valve short axis cross-section. (B) Mid-esophageal right ventricle inflow-outflow cross-section. (C) Mid-esophageal modified bicaval cross-section. (D) Mid-esophageal bicaval cross-section. [/fig]
[fig] Figure 6: A) Upper-esophageal right and left pulmonary veins cross-section. (B) Mid-esophageal left atrial appendage crosssection. (C) Transgastric basal short axis cross-section. (D) Transgastric mid-papillary short axis cross-section. [/fig]
[fig] Figure 7: A) Transgastric apical short axis cross-section. (B) Transgastric basal right ventricle cross-section. (C) Transgastric right ventricle inflow-outflow cross-section. (D) Deep transgastric cross-section. [/fig]
[fig] Figure 8: A) Transgastric long axis cross-section. (B) Transgastric long axis cross-section. (C) Descending aorta short axis crosssection. (D) Descending aorta long axis cross-section. [/fig]
[fig] Figure 9: A) Upper-esophageal aortic arch long axis cross-section. (B) Upper-esophageal aortic arch short axis cross-section. [/fig]
[fig] Figure 10: Mitral annulus assessment by three-dimensional echocardiography. (A) Anteroposterior diameter. (B) Anterolateral---posteromedial diameter. Ao, aortic ring; A, anterior; P, posterior; AL, anterolateral; PM, posteromedial.anterior region (A1), a middle region (A2), and a posterior region (A3). At the anterolateral and posteromedial ends, two regions are defined as anterolateral commissure and posteromedial commissure, respectively10 (Fig. 11). [/fig]
[fig] Figure 11 12, Figure 12: Carpentier nomenclature: segments #1 are anterolateral, #3 are posteromedial, and #2 correspond to the valve mid-portion. The anterior and posterior commissures are also visualized. Mitral subvalvar apparatus, showing the chordae tendineae distribution at each cusp.12 [/fig]
[fig] Figure 13: Mitral valve view in three-dimensional image. (A) Mitral valve in diastole. (B) Mitral valve in systole. [/fig]
[fig] Figure 14: Mitral valve view of left ventricle in three-dimensional image. (A) Mitral valve in diastole. (B) Mitral valve in systole. [/fig]
[fig] Figure 15: Three-dimensional color Doppler of mitral valve. [/fig]
[fig] Figure 16: Carpentier classification: (A) normal motion; (B) excessive motion; (C) restricted motion. [/fig]
[fig] Figure 17, Figure 18: Mitral Aortic complex. LV, left ventricle; LVOT, left ventricular outflow tract; RVOT, right ventricular outflow tract. [/fig]
[table] Table 1: Level of evidence ---intraoperative transesophageal echocardiography. [/table]
[table] Table 2: Transesophageal echocardiography complications. [/table]
[table] Table 3: Classification of diastolic dysfunction. DT E, deceleration time of mitral E wave; E/A, mitral E wave and mitral A wave ratio; E ′ mitral, tissue Doppler velocity in mitral annulus; PDPP, pulsed Doppler in pulmonary vein; IVRT, isovolumetric relaxation time. 36 Color M-mode Doppler in diastolic evaluation. LAP increased E/A ≥ 1 -< 2 E > 50 cm.s -1 E/A > 2 E < 150 cm.s -1 Figure 37 Flowchart of diastolic dysfunction evaluation with impaired left ventricular systolic function. E/A, mitral E wave and A wave ratio; E/E ′ , mitral E wave and tissue E ′ wave velocity ratio; S/D, pulmonary vein systolic and diastolic ratio; Ar, reverse pulmonary A wave; LAP, left atrial pressure. Ar-A > 30 ms Val E/A ≥ 0.5 SBP > 30 mmHg E/E' septal ≥ 15 E/E' lateral ≥ 12 E/E' mean ≥ 13 [/table]
|
Reduction in Pseudomonas aeruginosa sputum density during a cystic fibrosis pulmonary exacerbation does not predict clinical response
Background: Pulmonary exacerbations (PEx) are critical events in cystic fibrosis (CF), responsible for reduced quality of life and permanent loss of lung function. Approximately 1/4 of PEx are associated with failure to recover lung function and/or resolve symptoms. Developing tools to optimize PEx treatment is of paramount importance. Methods: We retrospectively audited all adults infected with Pseudomonas aeruginosa, experiencing PEx necessitating parenteral antibiotic therapy from 2006-2012 from our center. Quantitative analysis of sputum at admission, twice-weekly during hospitalization, and end of therapy were compared to baseline (most recent healthy) and follow-up (after PEx) samples. Change in P. aeruginosa burden from baseline was assessed for any and all morphotypes (ALL), as well as mucoid (MUC) and non-mucoid (NON) isolates specifically. PEx were identified as failures if >90% of baseline pulmonary function was not recovered. Results: Forty-six patients meeting the above inclusion and exclusion criteria experienced 144 PEx during this time (median 3, IQR 2-6). Patients were treated for a median 14 days . No increase in ALL, MUC or NON were detected at PEx, nor was there an association between change in sputum density and magnitude of lung function decline. PEx failures were observed in 30% of events. Reductions of at least 1-log and 2 log P. aeruginosa sputum density was observed in 57% and 46% (ALL), 73% and 55% (MUC) and 58% and 46% (NON) of PEx, respectively. Factors associated with greater reduction of P. aeruginosa sputum density included choice of β-lactam antibiotic, antibiotics with in vitro predicted activity and treatment duration. PEx associated with reductions in P. aeruginosa sputum density were not associated with a reduced risk of PEx failure. Conclusions: Enhanced killing of P. aeruginosa during PEx does not predict improved clinical outcomes. Studies accounting for the polymicrobial nature of CF respiratory disease and the heterogeneity of P. aeruginosa causing chronic infection may enable the identification of a more appropriate pathogen(s) based biomarker of PEx outcomes.
# Background
The archetypal CF pathogen, Pseudomonas aeruginosa infects 50-70% of patients [bib_ref] Respiratory Microbiology of Patients With Cystic Fibrosis in the United States, Razvi [/bib_ref]. Patients with chronic P. aeruginosa infection have increased rates of lung function decline, health care utilization, and reduced survival [bib_ref] Risk factors for Pseudomonas aeruginosa colonization in cystic fibrosis patients, Kerem [/bib_ref] [bib_ref] Acceleration of lung disease in children with cystic fibrosis after Pseudomonas aeruginosa..., Kosorok [/bib_ref] [bib_ref] Longitudinal development of mucoid Pseudomonas aeruginosa infection and lung disease progression in..., Li [/bib_ref]. Chronic P. aeruginosa infection is punctuated by frequent acute deteriorations termed pulmonary exacerbations (PEx). PEx are characterized by increased cough and sputum production, disproportionate shortness of breath and loss of lung function, as well as increased inflammation [bib_ref] Effect of aerosolized recombinant human DNase on exacerbations of respiratory symptoms and..., Fuchs [/bib_ref] [bib_ref] Report from the EuroCareCF Working Group on outcome parameters in clinical trials, Bilton [/bib_ref] [bib_ref] Exacerbations in cystic fibrosis. 1: Epidemiology and pathogenesis, Goss [/bib_ref]. PEx are critical events in CF, associated with reduced quality of life [bib_ref] Impact of recent pulmonary exacerbations on quality of life in patients with..., Britto [/bib_ref] , increased cost [bib_ref] The cost of medical care for patients with cystic fibrosis in a..., Lieu [/bib_ref] , permanent lung damage [bib_ref] Failure to recover to baseline pulmonary function after cystic fibrosis pulmonary exacerbation, Sanders [/bib_ref] [bib_ref] Return of FEV1 after pulmonary exacerbation in children with cystic fibrosis, Sanders [/bib_ref] and increased short-term mortality [bib_ref] Developing cystic fibrosis lung transplant referral criteria using predictors of 2-year mortality, Mayer-Hamblett [/bib_ref] [bib_ref] Survival effect of lung transplantation among patients with cystic fibrosis, Liou [/bib_ref]. So important are these events, they may now constitute primary end-points in CF therapeutic trials [bib_ref] Inhaled aztreonam lysine for chronic airway Pseudomonas aeruginosa in cystic fibrosis, Mccoy [/bib_ref].
Treatment of PEx usually consists of aggressive airway clearance, respite, nutritional support, and antimicrobial therapy directed against chronically infecting pathogens. Despite therapy, 25% of PEx fail to achieve successful outcomes as determined by lung function recovery, resolution of symptoms and preventing recurrences [bib_ref] Incidence and risk factors for pulmonary exacerbation treatment failures in patients with..., Parkins [/bib_ref]. Patients more likely to experience unsuccessful PEx outcomes are infected with MRSA, Burkholderia cepacia complex, MDR (multi-drug resistant) P. aeruginosa, and those with CF-related diabetes and/or CF-liver disease [bib_ref] Failure to recover to baseline pulmonary function after cystic fibrosis pulmonary exacerbation, Sanders [/bib_ref] [bib_ref] Incidence and risk factors for pulmonary exacerbation treatment failures in patients with..., Parkins [/bib_ref]. Furthermore, more severe PEx including those with greater initial lung function decline, and those with greater inflammation at initiation and completion of treatment are associated with worse outcomes.
Developing tools to predict PEx outcomes is key to their successful management. It has been previously determined that antibiogram directed therapy against a limited number of isolates of P. aeruginosa has only a weak association with PEx outcomes [bib_ref] Incidence and risk factors for pulmonary exacerbation treatment failures in patients with..., Parkins [/bib_ref]. As such, other biomarkers are increasingly being evaluated for their ability to predict treatment responses. A number of host specific factors are being assessed [bib_ref] Systematic review of blood biomarkers in cystic fibrosis pulmonary exacerbations, Shoki [/bib_ref] [bib_ref] Blood mRNA biomarkers for detection of treatment response in acute pulmonary exacerbations..., Nick [/bib_ref]. However, given the critical intervention in PEx is anti-bacterial, the use of bacterial derived biomarkers to follow treatment response deserves attention. While antibacterials have been shown to reduce the bacterial load during the treatment of PEx, how this correlates with clinical response has not been established [bib_ref] Reduction of sputum Pseudomonas aeruginosa density by antibiotics improves lung function in..., Regelmann [/bib_ref]. Herein we evaluate the use of semi-quantitative reporting of P. aeruginosa sputum density and correlated the response with clinical outcomes during PEx treatment.
# Methods
All CF patients chronically infected with P. aeruginosa attending the Calgary Adult CF Clinic from 2006-2012 experiencing PEx treated with parenteral antibiotics were considered for inclusion if they had semi-quantitative sputum cultures performed ≥3 times during treatment (baseline, initiation, early, end of therapy) and at followup. Parenteral antibiotics provided for reasons other than PEx were excluded. Patients were excluded if they had a baseline FEV 1 < 30% predicted, were infected with Burkholderia cenocepacia or Mycobacterium abscessus, or were listed for lung transplantation. Detailed review of clinical records were performed from prior to the PEx, through treatment and in follow-up. Pulmonary function was evaluated by spirometry. Data was prospectively collected and retrospectively analyzed. Patients provide prospective consent for studies correlating clinical outcomes with lower airways infection. This project has been approved by the Conjoint Health Research Ethics Board of the University of Calgary (E23087).
## Definitions
Baseline lung function was defined as the best forced expiratory volume in one second (FEV 1 ) percent predicted in the preceding 6 months. The Leeds criteria was used to define chronic P. aeruginosa infection [bib_ref] Evaluation of a new definition for chronic Pseudomonas aeruginosa infection in cystic..., Lee [/bib_ref]. PEx were diagnosed on clinical grounds in real time, however, charts were evaluated and scored according to Fuchs criteria retrospectively [bib_ref] Effect of aerosolized recombinant human DNase on exacerbations of respiratory symptoms and..., Fuchs [/bib_ref]. Parenteral therapies were provided as either an inpatient, outpatient via the home parenteral therapy program (HPTP) or a combination thereof. Treatment failures were defined a priori as failure to have recovered 90% of baseline FEV 1 at follow-up (FTR90) [bib_ref] Failure to recover to baseline pulmonary function after cystic fibrosis pulmonary exacerbation, Sanders [/bib_ref] [bib_ref] Return of FEV1 after pulmonary exacerbation in children with cystic fibrosis, Sanders [/bib_ref] [bib_ref] Incidence and risk factors for pulmonary exacerbation treatment failures in patients with..., Parkins [/bib_ref]. Severity of lung disease was categorized as mild (FEV 1 % predicted ≥70%), moderate (40-70%), and severe (<40%). Bacterial sputum density is reported as log10 CFU/ml. Bacteriologic responses to antipseudomonal therapies were evaluated as either a one or two-log drop in CFU/ml.
## Antibacterial therapies
Anti-Pseudomonal β-lactams (aztreonam, ceftazidime, cefepime, and meropenem) were dosed 2 g every 8 hours infused over thirty minutes, ciprofloxacin as 600 mg IV or 750 mg PO every 12 hours, and tobramycin as 10 mg/kg/day divided in 1-2 doses. When patients with chronic Staphylococcus aureus co-infection were treated with regimens lacking Staphylococcal activity such as aztreonam/ceftazidime, supplemental anti-Staphylococcal therapies were used and consisted of trimethoprim/ sulfamethoxazole (TMP/SMX) 320/1600 mg PO BID, linezolid 600 mg PO BID, rifampin 300 mg PO BID, or cloxacillin 2 g IV q4h, clindamycin 600 mg IV q8h, ceftriaxone 2 g IV q12h or tigecycline 50 mg q12h.
## Microbiologic investigations
Baseline sputum samples were from the most recent clinic where patients were not experiencing PEx. Twiceweekly sputum samples were collected through treatment. Admission samples were collected before starting therapy. Changes in bacterial load were evaluated at "EARLY" (3 +/− 2 days from starting) and "END" (within 3 days of completion). Follow-up samples were collected at the next clinic visit attended.
Sputum samples were mixed with equal volumes of dithiothreitol (DTT, Sputolysin, EMD Millipore, MA) and vortexed. Samples were serial diluted in DTT to 10 −5 and plated on Columbia Blood Agar (CBA), MacConkey Agar, (MAC), Chocolate Agar (CHOC), Burkholderia cepacia selective agar, and Mannitol Salt Agar. Pathogens were identified by standard criteria [bib_ref] Laboratory standards for processing microbiological samples from people with cystic fibrosis, Denton [/bib_ref]. Oral flora (OF) included organisms such as non-hemolytic Streptococci, Haemophilus parainfluenzae, Corynebacterium spp., Neisseria spp. (non-meningitidis) and Coagulase-negative Staphylococci grown on CBA or CHOC. Organisms were quantified and P. aeruginosa colonies classified as mucoid (MUC) or non-mucoid (NON) isolates based on growth on MAC [bib_ref] Twenty-five-year outbreak of Pseudomonas aeruginosa infecting individuals with cystic fibrosis: identification of..., Parkins [/bib_ref]. P. aeruginosa susceptibility testing was performed on each morphotype by Kirby-Bauer or E-test as per CLSI guidelines. Where >1 MUC/NON isolates existed susceptibility testing was performed on each, and the result represents the least susceptible. MDR was defined as resistance to all antibiotics in ≥2 classes and pan-drug resistance (PDR) all classes, save for colistin [bib_ref] Incidence and risk factors for pulmonary exacerbation treatment failures in patients with..., Parkins [/bib_ref] [bib_ref] Survival of lung transplant patients with cystic fibrosis harboring panresistant bacteria other..., Hadjiliadis [/bib_ref].
# Statistical analysis
Means with standard deviation (SD) or medians with inter-quartile ranges (IQR) were used to describe normal and non-normally distributed data, respectively. Categorical variables were compared using a two-sided Fisher's exact test. Odds ratios (OR) were calculated by dividing the proportion with a factor to those without and were reported with 95% confidence intervals (CI). Continuous variables were compared using Pearsons correlation coefficient. Statistical analyses were performed using Stata 12.0 (Stata Corp., College Station, TX, USA). A p value of ≤0.05 was considered significant.
# Results
## Patients and pex
Of a total eligible cohort of 176, 46 patients met inclusion criteria. These 46 patients experienced 144 PEx during the six-year study period. Median age of those included was 32.9 years (IQR 27.9-40). Patients were diagnosed with CF at a median age of 1.35 years (IQR 0.5-6). 56% were F508del homozygous and 83% carried ≥1 allele. 43 patients (93%) were pancreatic insufficient. Co-morbidities included CF-related diabetes in 18 (40%), CF-liver disease 8 (18%), osteopenia 60% and osteoporosis 9%. During the six years of study, 14 patients initially included had future events excluded because of progression of disease and three patients died, although none from CF.
Patients experienced a median 3 exacerbations (IQR 2-6). Time to next PEx was 207 days (IQR 93-458). Complete details were not available for all. Using Fuchs criteria to confirm the clinical diagnosis of PEx, patients had a median of 5 criteria (IQR 5-7), where 4 meet the definition. The most common criteria were increased cough (99%) and sputum (99%), increased dyspnea (98%), decreased FEV 1 by ≥10% (94%), weight loss or anorexia (77%), change in chest exam (72%), radiographic infiltrates (52%), lethargy (46%), fevers (37%), and hemoptysis (36%). Sinus/nasal discharge symptoms were uncommonly documented (2%). At the time of exacerbation, chronic medications included: azithromycin 56%, H2-blockers or protonpump inhibitors 37%, inhaled corticosteroids (ICS) 54%, long acting bronchodilators 85%, nebulized DNase 66%, tobramycin inhalation solution 54%, and aztreonam solution for inhalation 4%. With the exception of inhaled antibiotics, chronic medications were continued during PEx. 70 (48%) of events were exclusively managed in hospital, 42 (29%) were treated with HPTP, 32 (22%) started off in hospital but completed at home, and two (1%) were admitted to hospital from HPTP. Exacerbations were associated with a median 19.3% (IRQ 11-28.5) and 18.2% (9.9-29) drop in FEV 1 and FVC respectively.
Patients were treated for a median 14 days (IQR 13-16). Two anti-pseudomonal antibiotics were used in 98% of exacerbations, and three in 2%. These generally consisted of an anti-pseudomonal β-lactam and tobramycin (96%). Ceftazidime was used in 51%, meropenem 24%, cefepime 16%, and aztreonam 7%. Rarely ciprofloxacin was combined with tobramycin (2%). Tobramycin was used in once (14%) or twice-daily doses (81%), and parenteral ciprofloxacin was used in place of tobramycin in five (3%) exacerbations. Additional anti-Staphylococcal therapy was used in 37 (26%) PEx and consisted of TMP/SMX (38%), cloxacillin and TMP/SMX (19%), cloxacillin and rifampin (14%), cloxacillin (11%) and rifampin, clindamycin, clindamycin and ceftriaxone, tigecycline and linezolid each in 3%.
Patients presented with an exacerbation 68 days (IQR 32-111) after their most recent clinic visit. Mean duration from initiation of antibiotics to collection of EARLY sputum culture was 3.84 days (SD 1.75). Mean duration from the end of therapy to collection of END sputum sample was 0.72 days (SD 1.70). 40 days (IQR 27-61) after completion of PEx treatment, patients were seen in follow-up at a regular CF clinic.
## Pex outcomes
PEx treatment failures, as defined by failing to recover 90% of baseline FEV 1 , were observed in 42/140 (30%). Neither sign, symptom, nor total Fuchs' initial PEx diagnostic score were associated with increased failure risk (not shown). Stage of CF-lung disease did not influence failure (not shown). Antibiotic courses ≤10 days were more likely to be deemed failures; 8/16 (50%) vs 38/128 (30%), OR 1.68 (0.97-2.94), p = 0.15) but this did not reach statistical significance. Few prior and concomitant medications influenced PEx outcomes. In particular, prior tobramycin inhalation solution was associated with a trend to reduced risk of failure
## Bacteriologic response to antibiotics during pex
Median sputum density of OF at PEx was 10 6 CFU/ml (IQR 10 5 -10 6 ), S. aureus 10 6 (IQR 10 5 -10 6 ), P. aeruginosa ALL 10 6 (IQR 10 6 -10 7 ), MUC 10 6 (IQR 10 5 -10 6 ) and NON 10 6 (IQR 10 5 -10 7 ). Relative to baseline, admission PEx samples did not demonstrate an increase in P. aeruginosa burden [fig_ref] Figure 1: Change in P [/fig_ref]. Furthermore, there was no correlation between the extent of pulmonary function decline at PEx presentation and change in bacterial load (not shown). Treatment produced a progressive decline in P. aeruginosa in 60% of patients peaking at end of therapy, but returned to baseline at follow-up [fig_ref] Figure 1: Change in P [/fig_ref]. Stage of lung disease during PEx did not influence bacteriologic response (not shown).
Clinical outcomes were not predicted based on changes in sputum bacterial burden at completion of PEx treatment [fig_ref] Table 1: Bacteriologic response does not predict clinical outcomes of pulmonary exacerbations *Failure was... [/fig_ref]. Relative to other antibiotics ceftazidime showed a modest advantage, whereas cefepime was associated with reduced P. aeruginosa killing. Tobramycin dosing once or twice daily did not influence bacteriologic response (data not shown). An association between number of antibiotics with in vitro predicted activity and reduction of P. aeruginosa density was observed although this reached statistical significance only for regimens with activity against the NON-isolates [fig_ref] Table 3: Relation of in vitro predicted susceptibility to reduction in P [/fig_ref].
No significant impact of concurrent therapies was noted on reductions in P. aeruginosa sputum density. In particular, concurrent chronic azithromycin, DNase, nor hypertonic saline impacted P. aeruginosa density. A trend was observed in individuals receiving concurrent ICS at having a lower likelihood of reducing P. aeruginosa sputum burden (not shown). Venue for antibiotic delivery did not impact bacteriologic response (not shown). Treatments <14 days produced an inferior bacteriologic response, although this met statistical significance for only the NON-isolates [fig_ref] Table 2: Treatment factors associated with reduced P [/fig_ref].
MSSA was a chronic colonizer of patients in 31/146 (21.2%) of PEx, and two had a history of recent-prior MRSA transient colonization, though none on admission. . EARLY changes in bacterial load did not, however, predict failure risk for any P. aeruginosa morphotype (not shown).
# Discussion
PEx are increasingly recognized as a significant contributor to the progressive nature of CF lung disease [bib_ref] Effect of pulmonary exacerbations on long-term lung function decline in cystic fibrosis, Waters [/bib_ref]. Avoiding PEx is critical in this regard, and many CF therapies may exert their beneficial effects in manner [bib_ref] Effect of aerosolized recombinant human DNase on exacerbations of respiratory symptoms and..., Fuchs [/bib_ref] [bib_ref] Azithromycin in patients with cystic fibrosis chronically infected with Pseudomonas aeruginosa: a..., Saiman [/bib_ref] [bib_ref] Intermittent administration of inhaled tobramycin in patients with cystic fibrosis. Cystic Fibrosis..., Ramsey [/bib_ref] [bib_ref] Inhaled aztreonam lysine vs. inhaled tobramycin in cystic fibrosis: A comparative efficacy..., Assael [/bib_ref]. However, despite attempts to limit their occurrence, PEx remain common. Given that some PEx are associated with worse outcomes than others, strategies to identify and optimize the management of these events are critical.
Anti-pseudomonal antibiotics are an integral component of PEx management. Independent of enhanced airway clearance, anti-pseudomonal antibacterials have been shown to improve pulmonary function in a linear capacity relative to reducing P. aeruginosa burden [bib_ref] Reduction of sputum Pseudomonas aeruginosa density by antibiotics improves lung function in..., Regelmann [/bib_ref]. Few studies however, have tried to identify factors associated with enhanced P. aeruginosa killing. Smith et al., followed 75 patients chronically infected with P. aeruginosa through PEx managed with single or dual agent anti-pseudomonals [bib_ref] Sputum changes associated with therapy for endobronchial exacerbation in cystic fibrosis, Smith [/bib_ref]. While they observed a general reduction in sputum density of P. aeruginosa, reduced sputum DNA and protein levels, none of these factors correlated with absolute change in pulmonary function. Similarly, Mclaughlin et al., did not observe a correlation between burden of P. aeruginosa and absolute lung function recovery [bib_ref] Clinical and bacteriological responses to three antibiotic regimens for acute exacerbations of..., Mclaughlin [/bib_ref]. More recently, Deschaght et al., performed an observational study where they identified a correlation between improvement in lung function and reduction in P. aeruginosa density using quantitative PCR [bib_ref] Is the improvement of CF patients, hospitalized for pulmonary exacerbation, correlated to..., Deschaght [/bib_ref]. These studies, however, focused on aggregate analysis of lung function improvement and did not categorize individual events on the basis of success or failure, nor assess impact of individual morphotypes.
We sought to determine if bacteriologic killing as a measure of effectiveness of antibacterial therapies correlated with clinical outcomes in PEx. During antibacterial treatment of PEx, a progressive decline in P. aeruginosa burden was noted in approximately 60% of patients. This decline was predictable based on early changes in P. aeruginosa burden after only four days of treatment. Reduction in total P. aeruginosa burden, nor mucoid or nonmucoid isolates specifically did not, however, correlate with risk of failing to recover lung function. Enhanced killing of P. aeruginosa was observed in patients receiving antibacterials estimated to have more potent in vitro activity. This was particularly true when antibiotics had greater activity against non-mucoid isolates that were more readily reduced with antibiotics. Furthermore, we have observed like others, that in those few patients who transiently achieve clearance of P. aeruginosa during PEx treatments, no enhancement in clinical outcomes was afforded [bib_ref] Clinical and bacteriological responses to three antibiotic regimens for acute exacerbations of..., Mclaughlin [/bib_ref].
To complicate matters, studies have shown resolution of PEx in patients chronically infected with P. aeruginosa treated with regimens devoid of anti-pseudomonal activity suggesting response is more complicated than a one-host and one-pathogen model [bib_ref] The Streptococcus milleri group-an unrecognized cause of disease in cystic fibrosis: a..., Parkins [/bib_ref] [bib_ref] Is anti-Pseudomonas therapy warranted in acute respiratory exacerbations in children with cystic..., Beaudry [/bib_ref]. Accordingly, more recent studies have assessed the impact of antibacterials on other constituents of the CF respiratory microbiome, beyond classical CF pathogens. Using enhanced culturing to identify the Streptococcus milleri group (S. anginosus group, SMG), 40% of PEx were identified to have this organism emerge as numerically dominant at the onset of PEx, and resolution of symptoms correlated with reduction in sputum burden [bib_ref] The Streptococcus milleri group-an unrecognized cause of disease in cystic fibrosis: a..., Parkins [/bib_ref] [bib_ref] McKay agar enables routine quantification of the 'Streptococcus milleri' group in cystic..., Sibley [/bib_ref] [bib_ref] The relevance of the polymicrobial nature of airway infection in the acute..., Sibley [/bib_ref]. Using culture independent techniques, others have observed that with the exception of the Prevotella and Chrysiogenales, levels of other microbiome constituents do not correlate with improvements in lung function [bib_ref] Inflammation and airway microbiota during cystic fibrosis pulmonary exacerbations, Zemanick [/bib_ref] [bib_ref] Microbiota and metabolite profiling reveal specific alterations in bacterial community structure and..., Twomey [/bib_ref]. We did not detect an increase in either total P. aeruginosa load, or specifically MUC or NON morphotypes associated with an exacerbation. Nor was there an association between change in P. aeruginosa sputum density and extent of decline in pulmonary function. This data mirrors that of other recent studies that have challenged the long held belief that an increase in P. aeruginosa triggers PEx [bib_ref] Comparison of real time diagnostic chemistries to detect Pseudomonas aeruginosa in respiratory..., Fothergill [/bib_ref] [bib_ref] Molecular analysis of changes in Pseudomonas aeruginosa load during treatment of a..., Reid [/bib_ref] [bib_ref] Does bacterial density in cystic fibrosis sputum increase prior to pulmonary exacerbation?, Stressmann [/bib_ref]. Clearly, factors other than changes in macroscopically apparent P. aeruginosa morphotypes are involved. These factors may include: environmental exposures such as pollution [bib_ref] Impact of air pollution on cystic fibrosis pulmonary exacerbations: a case-crossover analysis, Goeminne [/bib_ref] and allergens [bib_ref] How the airway smooth muscle in cystic fibrosis reacts in proinflammatory conditions:..., Mccuaig [/bib_ref] ; host factors such as medication compliance [bib_ref] Adherence with tobramycin inhaled solution and health care utilization, Briesacher [/bib_ref] and co-morbidities [bib_ref] Incidence and risk factors for pulmonary exacerbation treatment failures in patients with..., Parkins [/bib_ref] [bib_ref] Predictors of pulmonary exacerbations in patients with cystic fibrosis infected with multi-resistant..., Block [/bib_ref] [bib_ref] Factors associated with a shorter time until the next pulmonary exacerbation in..., Sequeiros [/bib_ref] pathogen factors such as respiratory viruses [bib_ref] The role of respiratory viruses in adult patients with cystic fibrosis receiving..., Etherington [/bib_ref] , changes in the constituents of the lower respiratory microbiome [bib_ref] McKay agar enables routine quantification of the 'Streptococcus milleri' group in cystic..., Sibley [/bib_ref] [bib_ref] A polymicrobial perspective of pulmonary infections exposes an enigmatic pathogen in cystic..., Sibley [/bib_ref] and specific P. aeruginosa sub-populations [bib_ref] Phenotypic heterogeneity of Pseudomonas aeruginosa populations in a cystic fibrosis patient, Workentine [/bib_ref] [bib_ref] Pseudomonas aeruginosa Population Diversity and Turnover in Cystic Fibrosis Chronic Infections, Mowat [/bib_ref]. Furthermore, it is likely that there are multiple convergent pathways leading to a PEx, and these events vary in their manifestations and potential outcomes.
Many contentious issues exist in PEx management strategies. Data in these areas are difficult to interpret as outcomes have traditionally focused on absolute improvement in pulmonary function, as opposed to a more appropriate model of proportional recovery [bib_ref] A polymicrobial perspective of pulmonary infections exposes an enigmatic pathogen in cystic..., Sibley [/bib_ref] [bib_ref] Phenotypic heterogeneity of Pseudomonas aeruginosa populations in a cystic fibrosis patient, Workentine [/bib_ref] Whereas some, but not all studies have suggested HPTP to be inferior, we noted similar outcomes in risk of failures and reductions in P. aeruginosa burden [bib_ref] A polymicrobial perspective of pulmonary infections exposes an enigmatic pathogen in cystic..., Sibley [/bib_ref] [bib_ref] Phenotypic heterogeneity of Pseudomonas aeruginosa populations in a cystic fibrosis patient, Workentine [/bib_ref]. Likewise length of therapy for PEx has been debated [bib_ref] Duration of intravenous antibiotic therapy in people with cystic fibrosis, Plummer [/bib_ref]. Herein we noted a trend towards increased risk of failure to recover lung function with therapies ≤10 days, but no benefit with extending therapies >14 days. We also observed treatments <14 days were inferior in reducing P. aeruginosa sputum density.
The effectiveness of different antibiotics in achieving good PEx outcomes has been previously assessed. Herein we did observe differential abilities of antibiotics to reduce P. aeruginosa burden. In particular, cefepime was inferior at reducing P. aeruginosa sputum burden, although no impact on PEx outcome was noted. Interestingly, metaanalyses have demonstrated patients treated with cefepime relative to an alternate β-lactam with a similar spectrum of activity may be associated with an increased risk of mortality [bib_ref] Efficacy and safety of cefepime: a systematic review and meta-analysis, Yahav [/bib_ref] [bib_ref] Empirical antibiotic monotherapy for febrile neutropenia: systematic review and meta-analysis of randomized..., Paul [/bib_ref]. While commonly used in CF, the only clinical studies evaluating cefepime in CF have been pharmacokinetic based [bib_ref] Cefepime pharmacokinetics in cystic fibrosis, Hamelin [/bib_ref] [bib_ref] Pharmacokinetics of cefepime in cystic fibrosis patients, Huls [/bib_ref]. Based on these data, further studies of cefepime usage in CF are advisable.
ICS are common agents in CF, with approximately one third of patients receiving them. However, their utility has not clearly been established and the general consensus is that they are over used and may even be harmful [bib_ref] Inhaled corticosteroids for cystic fibrosis, Balfour-Lynn [/bib_ref]. Indeed, in a randomized controlled trial to evaluate the effects of discontinuation of ICS in patients stably maintained on ICS showed equivalent outcomes [bib_ref] Multicenter randomized controlled trial of withdrawal of inhaled corticosteroids in cystic fibrosis, Balfour-Lynn [/bib_ref]. ICS use has been associated with increased risk of PEx [bib_ref] Predictors of pulmonary exacerbations in patients with cystic fibrosis infected with multi-resistant..., Block [/bib_ref] [bib_ref] Multiple antibiotic-resistant Pseudomonas aeruginosa and lung function decline in patients with cystic..., Ren [/bib_ref]. Our results now suggest an increased risk of PEx failure with ICS use, a trend towards reduced sputum density of P. aeruginosa during antibacterial treatment and may warrant further caution with these agents in CF.
Herein we have attempted to expand upon other studies assessing PEx outcomes and changes in P. aeruginosa sputum density. To do so we have utilized a proportional lung A B C Correlation in change in CFU from admission sputum sample Early in to exacerbation therapy (Y axis) to End of therapy (X axis). A). ALL P. aeruginosa morphotypes, B). Mucoid only isolates, C). Non-mucoid Isolates.
function recovery model to overcome the deficiency whereby other studies using absolute improvement fail to account for the extent of initial decline and therefore are poor predictors of success or failure [bib_ref] Incidence and risk factors for pulmonary exacerbation treatment failures in patients with..., Parkins [/bib_ref]. While this work relies on clinician diagnosed PEx events, we have retrospectively confirmed they met diagnostic criteria using an established PEx diagnostic algorithm [bib_ref] Effect of aerosolized recombinant human DNase on exacerbations of respiratory symptoms and..., Fuchs [/bib_ref]. Furthermore, we have attempted to evaluate outcomes based on individual morphotypes of P. aeruginosa reported routinely by clinical laboratories [bib_ref] Laboratory standards for processing microbiological samples from people with cystic fibrosis, Denton [/bib_ref]. Indeed, quantitative analysis of P. aeruginosa using qPCR would not distinguish between mucoid and non-mucoid isolates. As infection with mucoid isolates of P. aeruginosa is associated with reduced success of eradication treatment in patients with new infection [bib_ref] Pseudomonas aeruginosa in vitro Phenotypes Distinguish Cystic Fibrosis Infection Stages and Outcomes, Mayer-Hamblett [/bib_ref] [bib_ref] Pseudomonas aeruginosa Phenotypes Associated with Eradication Failure in Children with Cystic Fibrosis, Mayer-Hamblett [/bib_ref] and a worse prognosis for patients with chronic lung infection [bib_ref] Acceleration of lung disease in children with cystic fibrosis after Pseudomonas aeruginosa..., Kosorok [/bib_ref] [bib_ref] Mucoid Pseudomonas aeruginosa is a marker of poor survival in cystic fibrosis, Henry [/bib_ref] , it might be expected that a differential response against mucoid isolates might impact PEx outcomes. This appears not to be the case.
Several limitations of this work deserve consideration. Because diagnosis of PEx were collected and analyzed retrospectively, the data is not as robust as prospectively collected data. Symptoms or signs not documented in clinical records may inadvertently be documented as absent. This may be somewhat mitigated, as these findings were collected at a single CF clinic over a short time period, where practice patterns were relatively uniform. Our study excluded PEx not requiring parenteral antibiotics. We observed a significantly higher rate of PEx failures than in other works [bib_ref] Failure to recover to baseline pulmonary function after cystic fibrosis pulmonary exacerbation, Sanders [/bib_ref] [bib_ref] Return of FEV1 after pulmonary exacerbation in children with cystic fibrosis, Sanders [/bib_ref] [bib_ref] Incidence and risk factors for pulmonary exacerbation treatment failures in patients with..., Parkins [/bib_ref]. While similar criteria were used in the selection of only parenteral antibiotic treated PEx, far fewer patients were treated with parenteral antibiotics in this study. This likely reflects institutional practice variation in the management of PEx and availability of local resources. Parenteral antibiotics were reserved for the most severe events as evidenced by the significant decline in lung function on admission. This is further supported by the relatively low level of P. aeruginosa drug resistance observed herein relative to other works [bib_ref] Incidence and risk factors for pulmonary exacerbation treatment failures in patients with..., Parkins [/bib_ref] [bib_ref] Susceptibility testing of Pseudomonas aeruginosa isolates and clinical response to parenteral antibiotic..., Smith [/bib_ref]. Whether more aggressive management of exacerbations would result in improved outcomes is of interest.
# Conclusions
In this work we identified that while the majority of patients treated with parenteral anti-pseudomonals experienced a temporary reduction in sputum burden, approximately 40% did not. This bacteriologic response could be predicted on the basis of sputum samples collected early into antibiotic treatment. Antibacterial treatments with a greater number of agents with in vitro predicted activity against all morphotypes, and in particular non-mucoid isolates was more likely to result in enhanced P. aeruginosa killing during PEx. However, patients who experienced enhanced killing of P. aeruginosa after completion of therapy were not afforded a reduced risk of PEx failure. Accordingly, bacteriologic response during PEx treatment cannot be used to predict success. Competing interests MDP has received research funding from Gilead sciences and participated in advisory boards for Gilead, Novartis and Roche. HRR has participated in advisory boards for Novartis and Vertex. JCL, RS and MGS report no conflicts.
Author contributions JCL was responsible for data collection, analysis and initial drafting of the manuscript. RS was responsible for data analysis, and manuscript revisions. MGS and HRR were responsible data maintenance, participated in study design and analysis, and assisted in manuscript preparation and revisions. MDP conceived the study, oversaw data collection and analysis and manuscript preparation and revisions. All authors read and approved the final manuscript.
[fig] Figure 1: Change in P. aeruginosa CFU/ml during PEx. A). Change in CFU from baseline sputum to admission at time of PEx. B). Change in CFU at EARLY time point (~day 4) of antibiotic treatment relative to admission. C). Change in CFU at the END of antibiotic treatment for PEx relative to admission. D). Change in CFU at follow-up clinic visit relative to admission. Black bars represent all P. aeruginosa isolates, grey bars only MUC, and white bars with black dots NON. MSSA was more likely to demonstrate a significant bacteriologic response with 26/30 (81%), 20/30 (67%), 14/30 (46%) experiencing at least a 2-log, 4-log or 6-log drop in MSSA, respectively at the end of therapy. Bacteriologic responses of MSSA and OF during PEx did not predict clinical success (Table 1). During fourteen PEx, new pathogens were apparent during twice-weekly samples, different from admission. These included Aspergillus (4), Stenotrophomonas maltophilia (4), Group C Streptococcus (3), and one each of MSSA, MRSA and H. influenzae. None of these resulted in changed antimicrobial management (with the exception of TMP/SMX for MRSA) and none manifested in chronic infection. Exacerbations associated with the emergence of new pathogens did not impact risk of PEx failure [5/42 (12%) vs 9/108 (8%), OR 1.41 (CI 0.5-3.98), p = 0.54]. Early bacteriologic response and correlation to end of therapy Change in CFU of P. aeruginosa EARLY into the antibacterial treatment correlated with change in load by the END of the treatment for ALL (Prob > F <0.0001, r 2 = 0.27), MUC (Prob > F =0.0003, r 2 = 0.189) and NON-MUC (Prob > F <0.0001, r 2 = 0.30) [/fig]
[fig] Figure 3, Figure 2: Percentage of PEx achieving a 1 or 2-log reduction in P. aeruginosa sputum density at end of antibacterial therapy as a function of the β-lactam antibiotic used. Percentage of PEx achieving a 1 or 2-log reduction in P. aeruginosa sputum density at end of antibacterial therapy as a function of the number of empirically provided antibiotics with in vitro predicted activity from admission sputum sample. A). Antibiotics with activity against Mucoid isolates. B). Antibiotics with activity against Non-Mucoid isolates. [/fig]
[table] Table 1: Bacteriologic response does not predict clinical outcomes of pulmonary exacerbations *Failure was defined as failing to achieve 90% of baseline percent predicted FEV 1 at time of follow-up. **Risk of failure to achieve outcome in those PEx achieving the log drop in bacterial counts relative to those events where a drop was not observed. ALL = all P. aeruginosa morphotypes, MUC = mucoid, NON = non-mucoid isolates, OR = odds ration, CI = confidence intervals. [/table]
[table] Table 2: Treatment factors associated with reduced P. aeruginosa bacterial burden at END of PEx therapies *Other antibiotics including aztreonam, meropenem, ciprofloxacin and dosing of tobramycin did not show significance and have not been included. CAZ = Ceftazidime, FEP = Cefepime, ABx = antibiotics. [/table]
[table] Table 3: Relation of in vitro predicted susceptibility to reduction in P. aeruginosa bacterial burden at END of PEx therapies [/table]
|
Cost-Effective Control of Plant Disease When Epidemiological Knowledge Is Incomplete: Modelling Bahia Bark Scaling of Citrus
## Symbol
Description
[formula] Value/Definition t Time since initial planting - S(t) Number of susceptible plants - E(t) [/formula]
Number of exposed plants -
[formula] I(t) Number of infected plants - R(t) Number of removed plants - A(t) [/formula]
Number of asymptomatic plants
[formula] A(t) = S(t) + E(t) E 0 [/formula]
Percentage of plants exposed at t = 0 varied (default 1% or 4%) φ i Rate at which i th host becomes infected β j∈Ω I K(d ji ; α) Ω S Set of susceptible hosts -Ω E Set of exposed hosts -Ω I Set of infectious hosts - Profit (up to a scale factor) P = Y − σV
[formula] K(d ij ; α) Dispersal kernel (2πα 2 ) −1 exp (−d/α) d ijDistance [/formula]
|
Oxidative stress is associated with the number of components of metabolic syndrome: LIPGENE study
Previous evidence supports the important role that oxidative stress (OxS) plays in metabolic syndrome (MetS)-related manifestations. We determined the relationship between the number of MetS components and the degree of OxS in MetS patients. In this comparative cross-sectional study from the LIPGENE cohort, a total of 91 MetS patients (43 men and 48 women; aged between 45 and 68 years) were divided into four groups based on the number of MetS components: subjects with 2, 3, 4 and 5 MetS components (n ¼ 20, 31, 28 and 12, respectively). We measured ischemic reactive hyperemia (IRH), plasma levels of soluble vascular cell adhesion molecule-1 (sVCAM-1), total nitrite, lipid peroxidation products (LPO), hydrogen peroxide (H 2 O 2 ), superoxide dismutase (SOD) and glutathione peroxidase (GPx) plasma activities. sVCAM-1, H 2 O 2 and LPO levels were lower in subjects with 2 or 3 MetS components than subjects with 4 or 5 MetS components. IRH and total nitrite levels were higher in subjects with 2 or 3 MetS components than subjects with 4 or 5 MetS components. SOD and GPx activities were lower in subjects with 2 MetS components than subjects with 4 or 5 MetS components. Waist circumference, weight, age, homeostatic model assessment-b, triglycerides (TGs), high-density lipoprotein and sVCAM-1 levels were significantly correlated with SOD activity. MetS subjects with more MetS components may have a higher OxS level. Furthermore, association between SOD activity and MetS components may indicate that this variable could be the most relevant OxS biomarker in patients suffering from MetS and could be used as a predictive tool to determine the degree of the underlying OxS in MetS.
# Introduction
Metabolic syndrome (MetS) is characterized by different combinations of three or more features such as hyperglycemia, hypertriglyceridemia, low level of high-density lipoprotein (HDL) cholesterol (HDL-C), hypertension and abdominal obesity, 1 as defined by the criteria of the Third Report of the National Cholesterol Education Program Adult Treatment Panel III. The incidence of MetS has reached epidemic proportions due to complex interactions between genetic and environmental factors, as well as prevailing sedentary lifestyles and unhealthy dietary habits. Consequently, MetS is considered a major risk factor for diabetes, cardiovascular and kidney diseases in industrialized societies. [bib_ref] Prevalence of obesity, diabetes, and obesity-related health risk factors, Mokdad [/bib_ref] Although it is generally accepted that the main pathogenic mechanism underlying the metabolic changes in patients with MetS relies on insulin resistance, there is evidence to indicate that a state of chronic low-level inflammation and oxidative stress (OxS) demonstrates a close link to MetS. [bib_ref] Is oxidative stress the pathogenic mechanism underlying insulin resistance, diabetes, and cardiovascular..., Ceriello [/bib_ref] [bib_ref] Increased oxidative stress in obesity and its impact on metabolic syndrome, Furukawa [/bib_ref] [bib_ref] Increased oxidative stress in normal-weight postmenopausal women with metabolic syndrome compared with..., Kim [/bib_ref] [bib_ref] Relationships between inflammation, adiponectin, and oxidative stress in metabolic syndrome, Chen [/bib_ref] OxS is widely accepted as playing a key mediatory role in the development and progression of multiple pathophysiological conditions, such as endothelial dysfunction, hypertension and atherosclerotic cardiovascular disease. [bib_ref] Metabolic syndrome, endothelial function and lifestyle modification, Aizawa [/bib_ref] This process is due to the production of reactive oxygen species and the impairment of antioxidant enzymatic defenses such as superoxide dismutase (SOD) or glutathione peroxidase (GPx), as well as the deposition of advanced glycation end products and oxidized low-density lipoproteins in the vascular wall. [bib_ref] Increased oxidative stress in normal-weight postmenopausal women with metabolic syndrome compared with..., Kim [/bib_ref] [bib_ref] Inadequate cytoplasmic antioxidant enzymes response contributes to the oxidative stress in human..., Chaves [/bib_ref] [bib_ref] Impact of the components of metabolic syndrome on oxidative stress and enzymatic..., Abdilla [/bib_ref] Moreover, OxS is also present in insulin resistance and concomitant inflammation. Interestingly, in association with obesity (with several MetS components), patients with insulin resistance and hyperglycemia have been shown to have a propensity toward aggressive cardiovascular disease, proinflammatory changes and a reduced nitric oxide (NO) bioavailability. [bib_ref] Impaired NO-dependent vasodilation in patients with Type II (non-insulindependent) diabetes mellitus is..., Van Etten [/bib_ref] Similarly, MetS patients exhibit activation of biochemical pathways leading to an increased delivery of reactive oxygen species, altered antioxidant protection and increased lipid peroxidation (LPO). [bib_ref] Oxidative stress-induced risk factors associated with the metabolic syndrome: a unifying hypothesis, Grattagliano [/bib_ref] Nevertheless, there is still a lack of data in this field. It is unclear whether the accumulation of factors related with MetS increases the degree of underlying OxS.
Following these assumptions, the purpose of this study was to investigate whether the degree of OxS is influenced by the number of components of MetS. We analyzed the activity of the main antioxidant enzymes (SOD and GPx) as well as OxS biomarkers (LPO, hydrogen peroxide (H 2 O 2 ) and NO) in a cross-sectional study of MetS patients of four groups divided according to the number of MetS components: subjects with 2, 3, 4 and 5 MetS components.
# Materials and methods
## Subjects and design
This study was conducted within the framework of the LIPGENE study ('Diet, genomics and the metabolic syndrome: an integrated nutrition, agro-food, social and economic analysis'), a Framework VI Integrated Project funded by the European Union. [bib_ref] Effects of dietary fat modification on insulin sensitivity and on other risk..., Tierney [/bib_ref] Subject eligibility was determined using a modified version of the National Cholesterol Education Program criteria for MetS, according to the published criteria.Subjects were required to meet at least two of the following five criteria: waist circumference 4102 cm (men) or 488 cm (women); fasting glucose 5.5-7.0 mmol l À1 ; TGs X1.5 mmol l À1 ; HDL-C o1.0 mmol l À1 (men) or o1.3 mmol l À1 (women); blood pressure (BP) X130/85 mm Hg or treatment of previously diagnosed hypertension. We used a subgroup of pre-intervention data for 91 subjects (43 men and 48 women), which conformed to the LIPGENE inclusion and exclusion criteria (Supplementary [fig_ref] Table 1: Blood pressure, biochemical and anthropometric characteristics and metabolic assessment of the study... [/fig_ref]. [bib_ref] Dietary fat differentially influences regulatory endothelial function during the postprandial state in..., Perez-Martinez [/bib_ref] All participants provided written informed consent and underwent a comprehensive medical history, physical examination and clinical chemistry analysis before enrollment. Participants displayed no signs of cardiac dysfunction or hepatic, renal and thyroid diseases and were requested to maintain their regular physical activity and lifestyle. Participants were also asked to record in a diary any event that could affect the outcome of the study, such as stress, change in smoking habits and alcohol consumption or intake of foods not included in the experimental design. None of the participants showed evidence of high alcohol consumption or a family history of early-onset cardiovascular disease, nor were any active smokers. All patients were free from cardiovascular complications at the time of the enrollment. The study was carried out in the Lipid and Atherosclerosis Unit at the Reina Sofia University Hospital, from February 2005 to April 2006. The experimental protocol was approved by the local ethics committee according to the Helsinki Declaration. The study was registered with the US National Library of Medicine Clinical Trials registry (NCT00429195).
## Anthropometric measurements
After recording clinical histories and conducting physical examinations, we obtained the following anthropometric measurements for each individual: weight, height, body mass index and waist circumference. Weight was measured while the subject was wearing light indoor clothing, without shoes and after voiding. Height was obtained with a stadiometer graduated in millimeters. The subject was barefoot with the back and head in contact with the stadiometer in the Frankfurt horizontal plane. Body mass index was calculated by dividing weight (kg) by height squared (m 2 ). Waist circumference (cm) was measured to the nearest 0.5 cm with a tape measure at the umbilical scar level. A non-stretchable tape measure was used to measure waist circumference. The measurement was taken directly on the skin with the subject in a standing position with the abdomen relaxed, the arms at the sides and the feet together. We used the homeostatic model assessment index for insulin resistance (HOMA IR : fasting insulin (mU l À1 ) Â fasting glucose (mmol l À1 )/22.5) and HOMA b-cell function as the index of insulin secretory function derived from fasting plasma glucose and insulin concentrations, calculated as 20 Â fasting insulin (mU l À1 )/fasting glucose (mmol l À1 ) À3.5. [bib_ref] Homeostasis model assessment: insulin resistance and beta-cell function from fasting plasma glucose..., Matthews [/bib_ref] Insulin sensitivity was estimated by a quantitative insulin sensitivity check index (QUICKI) (1/[log insulin (mU l À1 ) þ log baseline glucose (mg dl À1 )]). [bib_ref] Indexes of insulin resistance and secretion in obese children and adolescents: a..., Conwell [/bib_ref] BP was measured using an automatic BP device. In accordance with the European Society of Hypertension Guidelines,BP measurement was obtained with an appropriately sized cuff positioned at the heart level and after the patient had been relaxed for at least 5 min. The same arm was used for each measurement, and the average of two measurements was used for data processing.
## Biochemical determinations
Plasma samples. Blood was collected in tubes containing ethylene diaminetetraacetic acid to yield a final concentration of 0.1% ethylene diaminetetraacetic acid. Plasma was separated from red cells by centrifugation at 1500 g for 15 min at 4 1C within 1 h of extraction. Plasma was immediately aliquoted and stored at À80 1C until analysis.
Biochemical analysis. The lipid variables were analyzed with a modular autoanalyzer (DDPPII Hitachi; Roche, Basel, Switzerland) with the use of Boehringer-Mannheim reagents. TGs in plasma were assayed by means of enzymatic procedures. [bib_ref] Quantitative determination of serum triglycerides by the use of enzymes, Bucolo [/bib_ref] HDL-C was measured by analyzing the supernatant obtained following precipitation of a plasma aliquot in dextran sulfate-Mg 2 þ , as described by Warnick et al. [bib_ref] Dextran sulfate-Mg 2 þ precipitation procedure for quantitation of high-density-lipoprotein cholesterol, Warnick [/bib_ref] Plasma glucose concentrations were measured with an Architect-CG16000 analyzer (Abbott Diagnostics, Tokyo, Japan) by the hexokinase method. Plasma insulin concentrations were measured by chemoluminescence with an Architect-I2000 analyzer (Abbott Diagnostics, Tokyo, Japan). High-sensitivity C-reactive protein concentrations were measured according to Rifai et al. [bib_ref] Clinical efficacy of an automated highsensitivity C-reactive protein assay, Rifai [/bib_ref] The estimated glomerular filtration rate was calculated using the Chronic Kidney Disease Epidemiology Collaboration equation. [bib_ref] A new equation to estimate glomerular filtration rate, Levey [/bib_ref]
## Study of endothelial function using laser doppler
The Laser-Doppler linear Periflux 5000 (Perimed S.A., Stockholm, Sweden) was used to measure ischemic reactive hyperemia (IRH).
We found an inter-study variability of 8.85% and an intra-study variability of 8.7%.
Briefly, capillary flow of the second finger of the dominant arm of the patient was assessed for 1 min before (t0) and after applying 4 min (td) of ischemia to the arm using a sphygmomanometer. The IRH was obtained via IRH ¼ (AUC td ÀAUC t0 ) Â 100 AUC t0 .
## Svcam-1 levels
Plasma concentrations of soluble vascular cell adhesion molecule-1 (sVCAM-1) were determined in duplicate with commercially available ELISA kits (R&D Systems, Minneapolis, MN, USA). Each assay was calibrated with sVCAM-1 standard curves. Absorbance was evaluated in an EIA plate reader (DTX 880 Multimode Detector; Beckman Coulter, Brea, CA, USA) at a wavelength of 450 nm. The minimum detectable level of sVCAM-1 was 0.17 ng ml À1 .
## Determination of oxs biomarkers
LPO is a mechanism of cellular and molecular injury. Plasmatic levels of LPO were determined using the Bioxytech LPO-586 Kit (OXIS International, Portland, OR, USA), [bib_ref] Reactions of N-methyl-2-phenylindole with malondialdehyde and 4-hydroxyalkenals. Mechanistic aspects of the colorimetric..., Erdelmeier [/bib_ref] an intra-assay coefficient of variation of 5.8% and an inter-assay coefficient of variation of 7.2%. The kit uses a chromatogenic reagent that reacts with malondialdehyde þ 4-hydroxyalkenals. The absorbance was evaluated in a spectrophotometer (UV-1603; Shimadzu, Kyoto, Japan) at a wavelength of 586 nm.
Plasmatic levels of H 2 O 2 were determined using the Bioxytech H2O2-560 Assay (OXIS International), an intra-assay coefficient of variation of 5.2% and an inter-assay coefficient of variation of 8.1%. The reaction was monitored at 560 nm.
Nitric oxide (NO) is a free gas produced endogenously by a variety of mammalian cells. This molecule induces vasodilatation and inhibits platelet aggregation and adhesion to the vascular endothelium. Total nitrite (nitrite and nitrate) was used as an indicator of NO production and was assayed using the Griess method, [bib_ref] Anticoagulants and other preanalytical factors interfere in plasma nitrate/nitrite quantification by the..., Ricart-Jane [/bib_ref] with an intra-assay coefficient of variation of 6.1%, and an inter-assay coefficient of variation of 7.7%. The reaction was monitored at 540 nm.
## Antioxidant enzyme activities
Total SOD (E.C: 1.15.1.1) activity was determined by colorimetric assay in plasma at wavelength of 525 nm, according to the method described by McCord and Fridovich et al., [bib_ref] Spectrophotometric assay of superoxide dismutase activity based on the activated autoxidation of..., Nebot [/bib_ref] with an intra-assay coefficient of variation of 7.3%, and an inter-assay coefficient of variation of 8.7%. GPx (E.C.: 1.11.1.9) activity was evaluated in plasma by the Flohe and Gunzler method, [bib_ref] Assays of glutathione peroxidase, Flohe [/bib_ref] [bib_ref] Coordinated upregulation of a series of endogenous antioxidants and phase 2 enzymes..., Zhu [/bib_ref] with an intra-assay coefficient of variation of 6.5% and an inter-assay coefficient of variation of 7.2%. The GPx assay is based on the oxidation of nicotinamide adenine dinucleotide phosphate to NAD þ , catalyzed by a limiting concentration of glutathione reductase, with maximum absorbance at 340 nm. The absorbance was evaluated in a Shimadzu UV-1603 spectrophotometer.
# Statistical analysis
The Statistical Package for the Social Sciences (SPSS 18.0 for Windows, Chicago, IL, USA) was used for statistical comparisons. The Kolmogorov-Smirnov test did not reveal a significant departure from the norms in the distribution of variance values. To evaluate data variation, Student's t-test and a univariate analysis of variance were performed with gender and waist circumference (Supplementary [fig_ref] Table 2: Correlation between plasma SOD activity and other factors a [/fig_ref] included as covariates. Bonferroni's test was used when post hoc analysis was required (Supplementary [fig_ref] Table 3: Correlations of plasma SOD activity with different factors by multiple regression analysis... [/fig_ref]. Pearson's linear correlation coefficient was calculated, and a multiple linear regression analysis was performed. Differences were considered to be significant when Po0.05. All data presented in text and tables are expressed as the means±s.e.
# Results
## Biochemical and anthropometric characteristics and metabolic assessment
The anthropometric, biochemical, BP and metabolic parameters of the subjects with a different number of MetS components are shown in [fig_ref] Table 1: Blood pressure, biochemical and anthropometric characteristics and metabolic assessment of the study... [/fig_ref]. Plasma TG, glucose and insulin levels and baseline HOMA IR (P ¼ 0.009) were higher in the subjects with 4 and 5 MetS components compared with the subjects with 2 and 3 MetS components (all Po0.05; [fig_ref] Table 1: Blood pressure, biochemical and anthropometric characteristics and metabolic assessment of the study... [/fig_ref]. Baseline HOMAb and QUICKI were lower in the subjects with 4 or 5 MetS components compared with the subjects with 2 or 3 MetS components (all Po0.05; [fig_ref] Table 1: Blood pressure, biochemical and anthropometric characteristics and metabolic assessment of the study... [/fig_ref]. High-sensitivity C-reactive protein levels were higher and estimated glomerular filtration rate was lower in the subjects with 5 MetS components than other subjects (all Po0.05). In addition, we found that body mass index, waist circumference and diastolic blood pressure were lower and HDL-C was higher in the subjects with 2 MetS components compared with the other subjects (all Po0.05; [fig_ref] Table 1: Blood pressure, biochemical and anthropometric characteristics and metabolic assessment of the study... [/fig_ref]. We did not find significant differences with respect to age (mean age: 58 years old; range: 45-68 years old) or systolic blood pressure (SBP) between the four groups of subject with MetS [fig_ref] Table 1: Blood pressure, biochemical and anthropometric characteristics and metabolic assessment of the study... [/fig_ref].
## Study of endothelial function
IRH decreased in the following order: 2 and 3 MetS components44 components45 components (all Po0.05) [fig_ref] Figure 1: Ischemic reactive hyperemia [/fig_ref]. In addition, we observed lower plasma sVCAM-1 levels in the subjects with 2-3 MetS components than the subjects with 4-5 MetS components (all Po0.05; [fig_ref] Figure 1: Ischemic reactive hyperemia [/fig_ref]. Total plasma nitrite levels were higher in the subjects with 2-3 MetS components than the subjects with 5 MetS components (all Po0.05; [fig_ref] Figure 1: Ischemic reactive hyperemia [/fig_ref].
## Biomarkers of oxs
Plasma H 2 O 2 and LPO levels were higher in the subjects with 5 MetS components compared with the subjects with 2 and 3 MetS components (all [fig_ref] Figure 2: Hydrogen peroxide [/fig_ref].
## Antioxidant enzyme activities
Plasma SOD activity was lower in the subjects with 2 MetS components compared with the subjects with 3, 4 and 5 MetS components (all Po0.05; [fig_ref] Figure 3: Superoxide dismutase [/fig_ref].
We found less plasma GPx activity in the subjects with 2 and 3 MetS components than the subjects with 4 and 5 MetS components (all Po0.05; [fig_ref] Figure 3: Superoxide dismutase [/fig_ref].
## Correlation and regression analysis
We observed significant correlations of plasma SOD activity with weight (r ¼ 0.299, P ¼ 0.004), waist circumference (r ¼ 0.323, P ¼ 0.002), TG level (r ¼ 0.238, P ¼ 0.023), HDL level (r ¼À0.357, P ¼ 0.001), age (r ¼À0.214, P ¼ 0.042), VCAM-1 level (r ¼ 0.292, P ¼ 0.005) and HOMAb (r ¼À0.244, P ¼ 0.028) [fig_ref] Table 2: Correlation between plasma SOD activity and other factors a [/fig_ref].
A multiple linear regression was performed. Waist circumference, gender, age, BP, IRH, NO, TG, glucose, sVCAM-1, LPO and HDL levels were included [fig_ref] Table 3: Correlations of plasma SOD activity with different factors by multiple regression analysis... [/fig_ref]. Only sVCAM-1 (b ¼ 0.410; P ¼ 0.017), waist circumference (b ¼ 0.391; P ¼ 0.029), TG (b ¼ 0.344; P ¼ 0.038) and HOMAb (b ¼ 0.406; P ¼ 0.036) were significant independent predictors of SOD activity [fig_ref] Table 3: Correlations of plasma SOD activity with different factors by multiple regression analysis... [/fig_ref].
# Discussion
MetS comprises a cluster of cardiovascular risk factors (low HDL-C, elevated BP, fasting glucose and TG, and abdominal obesity) 1 leading to accelerated atherosclerosis and an increased risk of type 2 diabetes. Additionally, MetS is associated with major cardiovascular events and a high mortality rate. [bib_ref] Cardiovascular morbidity and mortality associated with the metabolic syndrome, Isomaa [/bib_ref] [bib_ref] The metabolic syndrome and 11-year risk of incident cardiovascular disease in the..., Mcneill [/bib_ref] Although it is generally accepted that the main pathogenic mechanism underlying the development of metabolic changes in patients with MetS relies on insulin resistance, a large body of evidence supports the concept that increased OxS and a state of chronic low-level inflammation may have important roles in MetS-related manifestations, including atherosclerosis, inflammation, endothelial dysfunction, hypertension and type 2 diabetes mellitus. [bib_ref] Is oxidative stress the pathogenic mechanism underlying insulin resistance, diabetes, and cardiovascular..., Ceriello [/bib_ref] [bib_ref] Influence of metabolic syndrome on biomarkers of oxidative stress and inflammation in..., Van Guilder [/bib_ref] [bib_ref] Evaluation of oxidative stress and inflammation in obese adults with metabolic syndrome, Skalicky [/bib_ref] These data indicate that OxS could be an early event in the pathology of these chronic diseases rather than merely a consequence. However, the relationship between the number of MetS components and OxS in subjects with MetS has seldom been studied. [bib_ref] Relationship between metabolic syndrome components and oxidative stress in elderly community-dwelling Mexicans, Sánchez-Rodríguez [/bib_ref] With regard to the influence of the number of MetS components on OxS, our results showed that the higher the number of components, the greater the degree of OxS, which leads to increased plasma SOD and GPx activities, plasma H 2 O 2 , LPO and sVCAM-1 levels, and decreased IRH and total plasma nitrite levels. Fujita et al. [bib_ref] Systemic oxidative stress is associated with visceral fat accumulation and the metabolic..., Fujita [/bib_ref] demonstrated that systemic OxS increases in subjects with MetS and that this stress is closely linked to the accumulation of visceral fat and an increase in other anthropometric variables. Thus, we observed a higher body mass index, waist circumference, diastolic blood pressure, HOMA IR and TG, glucose and insulin levels, and lower HDL-C levels and HOMAb and QUICKI indexes when OxS increases.
It has also been suggested that increased OxS originates from mitochondrial dysfunction in patients with MetS. [bib_ref] Metabolic syndrome and mitochondrial function: molecular replacement and antioxidant supplements to prevent..., Nicolson [/bib_ref] [bib_ref] Reactive oxygen species have a causal role in multiple forms of insulin..., Houstis [/bib_ref] In response to OxS and to prevent oxidative damage, cells attempt to strengthen their antioxidant arsenal as the first line of defense. SOD, GPx and catalase (CAT) are considered primary antioxidant enzymes because they are involved in the direct elimination of reactive oxygen species. One major contributor to oxidative damage is H 2 O 2 , which is converted from the superoxide that originated from the mitochondria. CAT and SOD ameliorate the damaging effects of H 2 O 2 and superoxides by converting these compounds into the less damaging, benign molecules, oxygen and water. GPx decomposes H 2 O 2 and/or ROOH, thereby neutralizing their toxicity. Any changes in one of these systems may upset the equilibrium and result in cellular damage. [bib_ref] Increased oxidative/nitrosative stress and decreased antioxidant enzyme activities in prostate cancer, Arsova-Sarafinovska [/bib_ref] Previous studies have shown that the expression of SOD is upregulated by reactive oxygen species [bib_ref] Antioxidant enzymes and human diseases, Mates [/bib_ref] indicating that the increase in H 2 O 2 levels in the subjects with 4 or 5 MetS components could explain the greater increase in SOD and GPx activities in these groups of subjects. Moreover, we also found an increase in LPO levels in the subjects with 4 or 5 MetS components with respect to the other groups. LPO is an index of lipid damage, especially damage associated with OxS derived from free radicals. Increases in LPO levels have been observed in several human diseases including diabetes, arthritis and other inflammatory diseases, as well as in patients with MetS. [bib_ref] Reactive oxygen species have a causal role in multiple forms of insulin..., Houstis [/bib_ref] [bib_ref] Phagocytic NADPH oxidase overactivity underlies oxidative stress in metabolic syndrome, Fortuno [/bib_ref] OxS, through superoxide anion production, decreases NO availability. [bib_ref] Oxidized low density lipoprotein inhibits inducible nitric oxide synthase, GTP cyclohydrolase I..., Dulak [/bib_ref] It has also been described that impaired endothelial function is a key step in the pathogenesis of cardiovascular disease and that elevated levels of sVCAM-1 predict the incidental development of cardiovascular disease events. [bib_ref] Soluble intercellular adhesion molecule-1 is related to endothelial vasodilatory function in healthy..., Holmlund [/bib_ref] Thus, the improved bioavailability of NO (total nitrite) in subjects with 2 or 3 MetS components, as was also shown in our study, may be an indirect effect of a lower production of superoxide anions. Additionally, this observation coincides with higher IRH and lower plasma sVCAM-1 levels measured in these subjects.
Similarly, we also found correlations between weight, waist circumference, TG, HDL and sVCAM-1 levels, HOMAb index and SOD activity. These results could partially account for the increase in the production of superoxide anions, as seen in conditions that involve hypertriglyceridemia and obesity. [bib_ref] Increased superoxide production by mononuclear cells of patients with hypertriglyceridemia and diabetes, Hiramatsu [/bib_ref] We also found that the SOD activity is the most relevant OxS marker because several individual MetS components were correlated with this variable. Although the SOD activity may be directly or indirectly regulated by the redox balance, certain SOD activity is affected only by factors intrinsic to the individual, such as genetic factors. Ischemic reactive hyperemia (IRH) (a), soluble vascular cell adhesion molecule-1 (sVCAM-1) levels (b) and total nitrite levels (c) in plasma according to the number of metabolic syndrome (MetS) components. The data were analyzed using analysis of variance for repeated measurements. All values represent the means ± s.e. Bars with different superscript letters depict significant differences (Po0.05). NO, nitric oxide. Our study has some limitations. First, this cross-sectional study does not determine causality between the number of components of MetS with OxS. Second, this study examines the redox state by measuring plasma concentrations but does not evaluate the effects on various tissues involved. However, the findings obtained via plasma concentrations could reflect what is occurring in different body tissues. Another limitation of our study is that we did not include any subjects with 0 or 1 MetS criteria. This limitation is due to the characteristics of the LIPGENE study; additional studies including such criteria should thus be performed. Finally, the complexity of our study necessitated the use of a small sample population; it would be difficult to broaden the scope of the study to larger populations. However, it would be optimal if this study could be reproduced in large populations. Larger, prospective studies are needed to establish the relationship between the number of components of MetS and the degree of OxS.
In conclusion, MetS subjects with more MetS components may have a higher level of OxS. Furthermore, association between SOD activity and several MetS components may indicate that SOD activity could be the most relevant OxS biomarker in patients suffering from MetS. Measurement of SOD activity could be used as a predictive tool to determine the degree of underlying OxS in this disease.
These findings suggest that studying the redox state in early MetS patients may provide a starting point for understanding the pathways that contribute to both the development of MetS and its subsequent complications.
[fig] Figure 1: Ischemic reactive hyperemia (IRH) (a), soluble vascular cell adhesion molecule-1 (sVCAM-1) levels (b) and total nitrite levels (c) in plasma according to the number of metabolic syndrome (MetS) components. The data were analyzed using analysis of variance for repeated measurements. All values represent the means ± s.e. Bars with different superscript letters depict significant differences (Po0.05). NO, nitric oxide. [/fig]
[fig] Figure 2: Hydrogen peroxide (H 2 O 2 ) levels (a) and lipid peroxidation product (LPO) levels (b) in plasma according to the number of metabolic syndrome (MetS) components. The data were analyzed using analysis of variance for repeated measurements. All values represent the means±s.e. Bars with different superscript letters depict significant differences (Po0.05). [/fig]
[fig] Figure 3: Superoxide dismutase (SOD) activity (a) and glutathione peroxidase (GPx) activity (b) in plasma according to the number of metabolic syndrome (MetS) components. The data were analyzed using analysis of variance for repeated measurements. All values represent the means±s.e. Bars with different superscript letters depict significant differences (Po0.05). [/fig]
[table] Table 1: Blood pressure, biochemical and anthropometric characteristics and metabolic assessment of the study groups a,b [/table]
[table] Table 2: Correlation between plasma SOD activity and other factors a [/table]
[table] Table 3: Correlations of plasma SOD activity with different factors by multiple regression analysis a Abbreviations: SOD, superoxide dismutase; TG, triglycerides; VCAM-1, vascular cell adhesion molecule-1. a A multiple regression analysis was used to examine the correlations of plasma SOD activity with different factors of the study, SOD activity as dependent variable and waist circumference, TG and VCAM-1 levels as independent variables. [/table]
|
Genetic and environmental causes of variation in epigenetic aging across the lifespan
Background: DNA methylation-based biological age (DNAm age) is an important biomarker for adult health. Studies in specific age ranges have found widely varying results about its genetic and environmental causes of variation. However, these studies are not able to provide a comprehensive view of the causes of variation over the lifespan.Results:In order to investigate the genetic and environmental causes of DNAm age variation across the lifespan, we pooled genome-wide DNA methylation data for 4217 people aged 0-92 years from 1871 families. DNAm age was calculated using the Horvath epigenetic clock. We estimated familial correlations in DNAm age for monozygotic (MZ) twin, dizygotic (DZ) twin, sibling, parent-offspring, and spouse pairs by cohabitation status. Genetic and environmental variance components models were fitted and compared. We found that twin pair correlations were − 0.12 to 0.18 around birth, not different from zero (all P > 0.29). For all pairs of relatives, their correlations increased with time spent living together (all P < 0.02) at different rates (MZ > DZ and siblings > parent-offspring; P < 0.001) and decreased with time spent living apart (P = 0.02) at similar rates. These correlation patterns were best explained by cohabitationdependent shared environmental factors, the effects of which were 1.41 (95% confidence interval [CI] 1.16 to 1.66) times greater for MZ pairs than for DZ and sibling pairs, and the latter were 2.03 (95% CI 1.13 to 9.47) times greater than for parent-offspring pairs. Genetic factors explained 13% (95% CI − 10 to 35%) of variation (P = 0.27). Similar results were found for another two epigenetic clocks, suggesting that our observations are robust to how DNAm age is measured. In addition, results for the other clocks were consistent with there also being a role for prenatal environmental factors in determining their variation.Conclusions: Variation in DNAm age is mostly caused by environmental factors, including those shared to different extents by relatives while living together and whose effects persist into old age. The equal environment assumption of the classic twin study might not hold for epigenetic aging.
risk in adulthood [bib_ref] DNA methylation-based biomarkers and the epigenetic clock theory of ageing, Horvath [/bib_ref] [bib_ref] DNA methylation-based measures of biological aging, Dugué [/bib_ref] [bib_ref] Association of DNA methylation-based biological age with health risk factors and overall..., Dugué [/bib_ref] [bib_ref] DNA methylation-based biological aging and cancer risk and survival: pooled analysis of..., Dugué [/bib_ref]. DNAm age, therefore, is potentially an important biomarker for adult health.
Lifestyle factors, disease risk factors, and genetic variants have been reported to be associated with DNAm age [bib_ref] DNA methylation-based biomarkers and the epigenetic clock theory of ageing, Horvath [/bib_ref] [bib_ref] DNA methylation-based measures of biological aging, Dugué [/bib_ref] [bib_ref] Association of DNA methylation-based biological age with health risk factors and overall..., Dugué [/bib_ref] [bib_ref] Prenatal and early life influences on epigenetic age in children: a study..., Simpkin [/bib_ref] [bib_ref] Genetic variants near MLST8 and DHX57 affect the epigenetic age of the..., Lu [/bib_ref] [bib_ref] GWAS of epigenetic aging rates in blood reveals a critical role for..., Lu [/bib_ref] [bib_ref] DNA methylation-based biological age, genome-wide average DNA methylation, and conventional breast cancer..., Chen [/bib_ref]. Pedigree-based and single nucleotide polymorphism (SNP)-based studies have given widely varying estimates of the proportion of variation in DNAm age explained by genetic factors, ranging from 0 to 100% [bib_ref] Prenatal and early life influences on epigenetic age in children: a study..., Simpkin [/bib_ref] [bib_ref] Genetic variants near MLST8 and DHX57 affect the epigenetic age of the..., Lu [/bib_ref] [bib_ref] GWAS of epigenetic aging rates in blood reveals a critical role for..., Lu [/bib_ref] [bib_ref] DNA methylation age of human tissues and cell types, Horvath [/bib_ref] [bib_ref] A meta-analysis of genome-wide association studies of epigenetic age acceleration, Gibson [/bib_ref]. One possible reason for this is that these studies focused on specific age ranges only. There is also evidence that environmental factors shared within families explain a substantial proportion of variation in the middle age [bib_ref] Genetic and environmental causes of variation in the difference between biological age..., Li [/bib_ref]. Individual studies of specific age ranges are not able to provide a comprehensive view of the causes of variation over the lifespan.
We previously pooled DNA methylation data from a variety of twin and family studies in which participants were at different life stages, from birth to older age. We found evidence that variation in genome-wide average methylation is caused to a great extent by prenatal environmental factors, as well as by environmental factors shared by relatives (including spouse pairs) when they cohabit and that these effects can persist at least to some extent across the whole lifetime [bib_ref] Genome-wide average DNA methylation is determined in utero, Li [/bib_ref]. If specific age ranges were studied separately, these findings might not have been found.
We have now applied the same approach to investigate the genetic, shared environmental, and individualspecific environmental causes of variation in DNAm age across the lifespan.
# Results
## Sample characteristics
We analyzed genome-wide DNA methylation data from 10 studies (Additional file 1). The total sample included 4217 people aged 0-92 years from 1871 families, including monozygotic (MZ) twins, dizygotic (DZ) twins, siblings, parents, and spouses [fig_ref] Table 1: Sample characteristics by study EPIC the HumanMethylationEPIC array, 27K the HumanMethylation27 array,... [/fig_ref].
DNAm age was calculated using the Horvath epigenetic clock [bib_ref] DNA methylation age of human tissues and cell types, Horvath [/bib_ref] (https ://dnama ge.genet ics.ucla.edu/ new), as this clock is mostly applicable to our multi-tissue methylation data and study sample including newborns, children, and adults.
DNAm age was moderately to strongly correlated with chronological age within each dataset, with correlations ranging from 0.44 to 0.84 . The variance of DNAm age increased with chronological age, being small for newborns, greater for adolescents, and relatively constant with age for adults [fig_ref] Figure 2: Variance in age-adjusted DNAm age measured by the epigenetic clock by chronological... [/fig_ref]. A similar pattern was observed for the absolute deviation between DNAm age and chronological age [fig_ref] Table 1: Sample characteristics by study EPIC the HumanMethylationEPIC array, 27K the HumanMethylation27 array,... [/fig_ref]. Within each study, MZ and DZ pairs had similar absolute deviations and residuals in DNAm age adjusted for chronological age. [fig_ref] Table 2: Within-study familial correlations in DNAm age MZ monozygotic twin, DZ dizygotic twin,... [/fig_ref] shows the within-study familial correlation estimates. There was no difference in the correlation between MZ and DZ pairs for newborns or adults, but there was a difference (P < 0.001) for adolescents: 0.69 (95% confidence interval [CI] 0.63 to 0.74) for MZ pairs and 0.35 (95% CI 0.20 to 0.48) for DZ pairs. For MZ and DZ pairs combined, there was consistent evidence across datasets and tissues that the correlation was around − 0.12 to 0.18 at birth and 18 months, not different from zero (all P > 0. [bib_ref] Programs for pedigree analysis: MENDEL, FISHER, and dGENE, Lange [/bib_ref] , and about 0.3 to 0.5 for adults (different from zero in seven of eight datasets; all P < 0.01). Across all datasets, the results suggested that twin pair correlations increased with age from birth up until adulthood and were maintained to older age.
## Within-study familial correlations
The correlation for adolescent sibling pairs was 0.32 (95% CI 0.20 to 0.42), not different from that for adolescent DZ pairs (P = 0.89), but less than that for adolescent MZ pairs (P < 0.001). Middle-aged sibling pairs were correlated at 0.12 (95% CI 0.02 to 0.22), less than that for adolescent sibling pairs (P = 0.02). Parent-offspring pairs were correlated at 0.15 (95% CI 0.02 to 0.27), less than that for pairs of other types of first-degree relatives in the same study, e.g., DZ pairs and sibling pairs (both P < 0.04). The spouse-pair correlations were − 0.01 (95% CI − 0.25 to 0.24) and 0.12 (95% CI − 0.12 to 0.35).
From the sensitivity analysis, the familial correlation results were robust to the adjustment for blood cell composition (Additional file 1: [fig_ref] Table 1: Sample characteristics by study EPIC the HumanMethylationEPIC array, 27K the HumanMethylation27 array,... [/fig_ref].
## Familial correlations across the lifespan
From modeling the familial correlations for the different types of pairs as a function of their cohabitation status (Additional file 1: [fig_ref] Table 2: Within-study familial correlations in DNAm age MZ monozygotic twin, DZ dizygotic twin,... [/fig_ref] , the estimates of θ (see "Methods" section for definition) ranged from 0.76 to 1.20 across pairs, none different from 1 (all P > 0.1). We therefore fitted a model with θ = 1 for all pairs; the fit was not different from the model above (P = 0.69). Under the latter model, the familial correlations increased with time living together at different rates (P < 0.001) across pairs. The decreasing rates did not differ across pairs (P = 0.27). The correlations for DZ and sibling pairs were similar (P = 0.13), and when combined their correlation was different from that for parent-sibling pairs (P = 0.002) even though these pairs are all genetically first-degree relatives, and was smaller than that for the MZ pairs (P = 0.001).
We then fitted a model in which DZ and sibling pairs were combined and the decreasing rates were the same across all pairs. The goodness of fit of this model was not inferior to that of the model above (P = 0.14), and the model included fewer parameters. Under this model, the familial correlations for MZ, DZ and sibling, and parent-offspring pairs all increased with time living together (all P < 0.02) with different increasing rates (P < 0.001); most rapidly for MZ pairs (λ = 0.041, 95% CI 0.035 to 0.048), less rapidly for DZ and sibling pairs (λ = 0.026, 95% CI 0.020 to 0.031), and least rapidly for parent-offspring pairs (λ = 0.011, 95% CI 0.002 to 0.0021), and decreased with time living apart (P = 0.02); see [fig_ref] Figure 3: Familial correlations in DNAm age measured by the epigenetic clock for the... [/fig_ref].
## Causes of variation across the lifespan
Results from modeling the causes of variation across the lifespan are shown in [fig_ref] Figure 4: Proportion of variation in DNAm age measured by the epigenetic clock across... [/fig_ref] and Additional file 1: pairs. For all pairs, the proportion of variation explained by shared environmental factors increased with time living together (P < 0.001) and decreased at a slower rate with time living apart (P = 0.02).
Under the above cohabitation-dependent CE model, we further assumed that the variation is additionally caused by genetic factors whose effects are constant across the lifespan. Genetic factors were estimated to explain 13% (95% CI − 10 to 35%) of the variation (P = 0.27). That is, after taking into account the existence of non-genetic cohabitation-dependent effects, there was no evidence for a substantive role of genetic factors.
## Results for other dnam age measures
We also similarly studied two other DNAm age measures, a skin and blood clock developed by Horvath et al. [bib_ref] Epigenetic clock for skin and blood cells applied to Hutchinson Gilford Progeria..., Horvath [/bib_ref] and a blood clock developed by Han et al. [bib_ref] New targeted approaches for epigenetic age predictions, Han [/bib_ref] , which are also developed across tissues and/or ages. Overall, DNAm ages predicted by the two measures appeared to be more similar to chronological age than the DNAm age predicted by the Horvath epigenetic clock: within the same study, they had higher correlations with chronological age (Additional file 2: , Additional file 3: [fig_ref] Figure 2: Variance in age-adjusted DNAm age measured by the epigenetic clock by chronological... [/fig_ref] and lower absolute deviations from chronological age (Additional file 1: . For both measures, MZ and DZ pairs had similar absolute deviations and residuals in DNAm age adjusted for chronological age. Similar to the DNAm age predicted by the Horvath epigenetic clock, the variance of the DNAm ages predicted by the two measures increased with age in early life and remained relatively constant with age in adulthood (Additional file 4: Figure S3, Additional file 5: [fig_ref] Figure 4: Proportion of variation in DNAm age measured by the epigenetic clock across... [/fig_ref].
Additional file 1: shows the within-study familial correlation results for the two measures. For both measures, similar results to those for the Horvath epigenetic clock were observed: twin pair correlations increased with age from birth to adulthood and decreased with age in adulthood; no evidence that the twin-pair correlations differed by zygosity was observed across the lifespan, except in adolescence and at age 18 years. For both measures, newborn twins were
# Discussion
Our study provides novel insights into the causes of variation in DNAm age across the lifespan, which appear to be almost entirely environmental (i.e. non-genetic) factors. These include cohabitation-related environmental factors that are evident prior to adulthood, and whose effects persist across the whole of the lifespan. Two longitudinal studies have also found that DNAm age is largely set before adulthood [bib_ref] The trajectory of the blood DNA methylome ageing rate is largely set..., Kananen [/bib_ref]. Our data suggest that people in the same family are not correlated in DNAm age when they start cohabiting; the longer they live together, the more similar they become but at a rate that differs substantially depending on their relationship. This is likely due to the different types of relatives sharing environmental factors relevant to DNAm age to different degrees. When pairs of relatives live apart, they no longer share the cohabitation environment, and this is reflected by a slow dissipation of the effects of shared environmental factors across adulthood at a rate that appears to be similar for all pairs.
Our study is the first to provide a comprehensive view of the genetic and environmental causes of DNAm age variation across the lifespan. Focusing on limited age ranges or types of relatives might bias the interpretation for the causes. For example, if middle-aged (e.g., 40-70 years old) twins only (i.e., no siblings, parents or spouses) were studied, the higher MZ pair correlation compared with DZ pair correlation at that age range (see [fig_ref] Figure 3: Familial correlations in DNAm age measured by the epigenetic clock for the... [/fig_ref] might have been interpreted as being caused by genetic factors to some extent, as there are no data from other age ranges or types of relatives contributing to the interpretation. Without using data of various types of relatives whose ages cover the whole lifespan, the comprehensive view would have not been easily obtained.
For MZ pairs, some DNA methylation measures have been found to be similar at birth but divergent over the lifetime, a phenomenon called 'epigenetic drift' [bib_ref] Genome-wide average DNA methylation is determined in utero, Li [/bib_ref] [bib_ref] Epigenetic differences arise during the lifetime of monozygotic twins, Fraga [/bib_ref].
DNAm age, however, shows a different pattern; MZ pairs are not similar at birth (and neither are DZ pairs) but become more similar the longer they live together, and do so more rapidly than do DZ or other pairs of relatives. In adulthood, MZ pairs then appear to slowly become less similar in DNAm age the longer they live apart, at the same rate as for other pairs of relatives, but still maintain a substantial similarity even into late life. These observations suggest that DNAm age reflects biological aging processes beyond what is reflected by DNA methylation alone.
Our finding that environmental factors shared while cohabiting play a major role in determining the variation in DNAm age is also supported by the observation that the variance of DNAm age increased dramatically with age prior to adulthood and was relatively stable across adulthood [fig_ref] Figure 2: Variance in age-adjusted DNAm age measured by the epigenetic clock by chronological... [/fig_ref] , Additional file 4: Fgiure S3, Additional file 5: [fig_ref] Figure 4: Proportion of variation in DNAm age measured by the epigenetic clock across... [/fig_ref]. The latter has also been found by previous studies [bib_ref] The trajectory of the blood DNA methylome ageing rate is largely set..., Kananen [/bib_ref].
We investigated DNAm age based on other two pantissue/age clocks and found similar results to those for the Horvath clock. These results imply the role of cohabitation-related environmental factors in influencing the variation in these two clocks as well and suggest that our findings are robust to the way DNAm age is measured. These results of newborn MZ and DZ pairs were not differentially correlated in the two clocks implying the additional role of prenatal environmental factors in influencing the variation in these clocks, similar to what we found for the genome-wide average DNA methylation [bib_ref] Genome-wide average DNA methylation is determined in utero, Li [/bib_ref].
Given DNAm age has been found to be associated with the risks of death and various diseases in adulthood, identifying the environmental factors affecting DNAm age prior to adulthood might give novel insights into which, and how, early-life factors impact late-life health outcomes. This would have obvious implications for prevention and its timing. There is some evidence that DNAm age is associated with physical developmental characteristics, and exposures to stress and violence for children, although most studies had a moderate sample size [bib_ref] The epigenetic clock and physical development during childhood and adolescence: longitudinal analysis..., Simpkin [/bib_ref] [bib_ref] Accelerated DNA methylation age in adolescent girls: associations with elevated diurnal cortisol..., Davis [/bib_ref] [bib_ref] Exposure to violence accelerates epigenetic aging in children, Jovanovic [/bib_ref] [bib_ref] The epigenetic clock and pubertal, neuroendocrine, psychiatric, and cognitive outcomes in adolescents, Suarez [/bib_ref] [bib_ref] Faster ticking rate of the epigenetic clock is associated with faster pubertal..., Binder [/bib_ref]. Model details-AE model: variation was assumed to be caused by only A and E, and the effects of A are constant across the lifespan; cohabitation-dependent AE model: variation was assumed to be caused by only A and E, and the effects of A depend on cohabitation; cohabitation-dependent ACE model: variation was assumed to be caused by A, C and E, and the effects of A and C both depend on cohabitation; cohabitation-dependent CE model: variation was assumed to be caused by only C and E, and the effects of C depend on cohabitation
The classic twin design assumes that MZ and DZ pairs share environmental effects relevant to the trait of interest to exactly the same extent, i.e., the equal environment assumption. Our study shows that this assumption might not hold for DNAm age because there was strong evidence that MZ and DZ pairs share their pre-adult environmental effects to different extents. Furthermore, DZ and sibling pairs were more correlated than parent-offspring pairs, despite all being genetically first-degree relatives of one another; this is not consistent with the correlations predicted by additive genetic factors. Given there is no substantive evidence of genetic effects, our results are not consistent with gene-environment interaction either [bib_ref] Variance components models for gene-environment interaction in twin analysis, Purcell [/bib_ref] ; we found that models including genetic effects, no matter whether as constant or cohabitationdependent, were less consistent with the data compared with the cohabitation-CE model.
Previous twin and pedigree studies assumed the equal environment assumption holds perfectly and consequently reported the heritability of DNAm age to be ~ 40% in adolescence and middle age [bib_ref] GWAS of epigenetic aging rates in blood reveals a critical role for..., Lu [/bib_ref] [bib_ref] DNA methylation age of human tissues and cell types, Horvath [/bib_ref]. Note that under our cohabitation-dependent AE model (which makes the equal environment assumption), genetic factors would explain ~ 40% of variation in adolescence and middle age. This model, however, was not a good fit and was rejected in favor of models that included cohabitation-dependent environmental effects.
Studies have predicted that measured SNPs could explain 0-70% of variation in DNAm age measured from whole blood and brain tissue [bib_ref] Prenatal and early life influences on epigenetic age in children: a study..., Simpkin [/bib_ref] [bib_ref] Genetic variants near MLST8 and DHX57 affect the epigenetic age of the..., Lu [/bib_ref] [bib_ref] GWAS of epigenetic aging rates in blood reveals a critical role for..., Lu [/bib_ref]. Those analyses explicitly assumed, however, that all of the phenotypic covariance is due to genetic factors. In particular, one study predicted the SNP-based heritability of DNAm age based on mothers and children increased with the children's age, being zero when the children were around birth and 37% when the children were 15 years old [bib_ref] Prenatal and early life influences on epigenetic age in children: a study..., Simpkin [/bib_ref] in line with our data and the estimates under the cohabitation-dependent AE model that was rejected. Without relying on the equal environment assumption, we found that genetic factors explained at most a small, and not statistically significant, proportion (~ 10%) of variation. Therefore, studies using the equal environment assumption might have overestimated the influence of genetic factors on DNAm age variation.
Our study has several strengths. One strength is that we have included participants whose ages covered the whole lifespan, so we could provide insights into the genetic and environmental causes of DNAm age variation which are unable to be provided by studies focusing on specific ages only. The other strength is that we have substantial sample size, even within studies, so we can detect moderate correlations with good precision, and have the power to distinguish between different variance components models. Our findings should be interpreted with caution, given that they are from statistical modeling which alone cannot prove that a consistent model is a true representation of nature. All that can be said is whether or not the data 'are consistent with' a particular explanation. Nonetheless, statistical modeling is an attempt to identify the plausible and implausible explanations of data, and our results suggest that cohabitation environmental factors being shared by pairs of relatives to different extents are more plausible than genetic explanations.
# Conclusions
The variation in epigenetic aging across the lifespan is most consistent with having been caused, at least to a large extent, by environmental factors, including those shared to different extents by relatives while living together. The effects of the cohabitation environment increase with the time living together and persist into old age. The equal environment assumption of the classic twin study might not hold for epigenetic aging. Given the relationships between DNAm age and health outcomes, these findings highlight the importance and potential of pre-adulthood prevention related to environmental factors for adult diseases and biological aging.
# Methods
## Study sample
We analyzed genome-wide DNA methylation data from 10 studies, most of which were accessed through pub- [fig_ref] Table 1: Sample characteristics by study EPIC the HumanMethylationEPIC array, 27K the HumanMethylation27 array,... [/fig_ref] and Additional file 1).
## Data preprocessing
As several datasets on public repositories contained quality-controlled and preprocessed data only, we were unable to apply the same preprocessing methods across datasets. We used the study-specific data preprocessing methods to address study-specific technical variations. This design allows us to investigate true biological signals independent of any bias introduced from a unifying data preprocessing approach. In DNAm age calculation, we chose the 'Normalize Data' option of the online calculator to normalize each dataset to be comparable to the training data of this epigenetic clock.
## Dnam age and epigenetic age acceleration
We used the Horvath epigenetic clock [bib_ref] DNA methylation age of human tissues and cell types, Horvath [/bib_ref] to determine DNAm age (https ://dnama ge.genet ics.ucla.edu/new) because it was developed across tissues and ages, and the 353 methylation sites used by this clock are common to the three methylation arrays used by the 10 studies [fig_ref] Table 1: Sample characteristics by study EPIC the HumanMethylationEPIC array, 27K the HumanMethylation27 array,... [/fig_ref].
To adjust for the effects of chronological age on DNAm age, we studied epigenetic age acceleration, calculated as the residuals from a linear regression of DNAm age on chronological age. This calculation was done for each longitudinal measurement of the PETS 450K dataset and of the LSADT, for each generation of the BSGS, and for each age group of the DTR. For the PETS 27K dataset, DNAm age was standardized to have zero mean and unit variance for each type of biological sample, and the average standardized DNAm age across biological samples was used to calculate epigenetic age acceleration.
Sensitivity analyses were performed using only those studies in which DNA methylation was measured in blood to examine the robustness of results to cell composition. Naive CD8+ T cells, exhausted CD8+ T cells, plasmablasts, CD4+ T cells, natural killer cells, monocytes, and granulocytes estimated from the DNA methylation data [bib_ref] DNA methylation age of human tissues and cell types, Horvath [/bib_ref] [bib_ref] DNA methylation arrays as surrogate measures of cell mixture distribution, Houseman [/bib_ref] were additionally adjusted for in calculating epigenetic age acceleration.
We studied two other DNAm age measures which were developed across tissues and/or ages too, so they might be also applicable to our data. One is the skin and blood clock developed using multi-tissue methylation data of a sample aged 0-94 years [bib_ref] Epigenetic clock for skin and blood cells applied to Hutchinson Gilford Progeria..., Horvath [/bib_ref]. As some of the 391 methylation sites used by this clock were not included the PETS 450K and 27K datasets, these datasets were not included in its analysis. The other measure is developed by Han et al. [bib_ref] New targeted approaches for epigenetic age predictions, Han [/bib_ref] using a sample aged 1-101 years. As the measure is developed using HM450K array blood methylation data, non-blood or 27K datasets were not included in its analysis.
# Statistical analysis
Residuals of epigenetic age acceleration adjusted for sex were used in subsequent analyses. We used a multivariate normal model for pedigree analysis [bib_ref] Extensions to multivariate normal models for pedigree analysis. II. Modeling the effect..., Hopper [/bib_ref] [bib_ref] A multivariate normal model for pedigree and longitudinal data and the software..., Hopper [/bib_ref] and the program FISHER [bib_ref] Programs for pedigree analysis: MENDEL, FISHER, and dGENE, Lange [/bib_ref] to estimate correlations for different types of pairs (MZ, DZ, sibling, parent-offspring and spouse) and to fit variance components models. The likelihood ratio test was used to compare nested models. All P values were twosided, and P < 0.05 was considered significant.
According to the pattern in familial correlations by chronological age, and following previous theoretical and empirical studies [bib_ref] Genome-wide average DNA methylation is determined in utero, Li [/bib_ref] [bib_ref] Extensions to multivariate normal models for pedigree analysis. II. Modeling the effect..., Hopper [/bib_ref] [bib_ref] Cohabitation, convergence, and environmental covariances, Lange [/bib_ref] , the familial correlations across the lifespan were modeled as a function of the cohabitation status of the pair. The modeling was performed using the pooled data across all studies. Studyspecific variance in the residuals was used in analysis. For individuals i and j from the same family, their correlation was modeled as where 0 ≤ θ ≤ 2, and λ, υ ≥ 0.
Under this model, the correlation when the pairs start to live together is θ minus 1, and λ and υ reflect the increasing and decreasing rates at which the familial correlation increases with the length of cohabitation and decreases with the length of separation, respectively. The definitions of t and t 0 depend on the relationship between i and j: (1) for twin pairs, t = chronological age and t 0 = 18 years;
(2) for sibling pairs, t = chronological age of the younger sibling and t 0 = chronological age of the younger sibling when the older sibling was 18 years old; (3) for parentoffspring pairs, t = chronological age of the offspring and t 0 = 18 years; and (4) for spouse pairs, t = time in years since the pair married (assumed to be the average chronological age of the pair minus 24 years) and t 0 = time in years when the pair became separated (if known).
For individuals i and j from the same family, their covariance was modeled as where α, β A , β C , λ A , λ C , υ A , υ C ≥ 0, and the definitions of t and t 0 are the same as above.
We assumed that the variation of DNAm age can be caused by combinations of additive genetic factors (A), shared environmental factors (C), and individual-specific environmental factors (E). We assessed model fits using the Akaike information criterion (AIC) for the following models and assumptions:
## Ae model: variation is caused by only a and e;
the effects of A are constant across the lifespan; α = 2 × kinship coefficient, β A , β C , λ A , λ C , υ A , υ C = 0, and σ A 2 is free to be estimated. 2 Cohabitation-dependent AE model: variation is caused only by A and E; the effects of A depend on
[formula] ρ ij = θ − e − t if t ≤ t 0 (θ − e − t 0 )e −ν(t−t 0 ) , if t > t 0 COV ij = ασ 2 A + β A 1 − e − A t + β C (1 − e − C t ) if t ≤ t 0 ασ 2 A + β A 1 − e − A t 0 e −ν A (t−t 0 ) + β C (1 − e − C t 0 )e −ν C (t−t 0 ) , if t > t 0 [/formula]
[fig] Figure 2: Variance in age-adjusted DNAm age measured by the epigenetic clock by chronological age. PETS: Peri/postnatal Epigenetic Twins Study, including three datasets measured using the 27K array, 450K array, and EPIC array, respectively; BSGS: Brisbane System Genetics Study; E-Risk: Environmental Risk Longitudinal Twin Study; DTR: Danish Twin Registry; AMDTSS: Australian Mammographic Density Twins and Sisters Study; MuTHER: Multiple Tissue Human Expression Resource Study; OATS: Older Australian Twins Study; LSADT: Longitudinal Study of Aging Danish Twins; MCCS: Melbourne Collaborative Cohort Study [/fig]
[fig] Figure 3: Familial correlations in DNAm age measured by the epigenetic clock for the different types of pairs across the lifespan. Lines are the predicted familial correlations from modeling the familial correlation as a function of cohabitation status, and shadows are the corresponding 95% confidence intervals. MZ monozygotic twin, DZ dizygotic twin [/fig]
[fig] Figure 4: Proportion of variation in DNAm age measured by the epigenetic clock across the lifespan explained by genetic and environmental factors. Lines are the predicted proportions of variation explained by genetic and environmental factors from the variance components modeling, and shadows are the corresponding 95% confidence intervals. A: additive genetic factors; C: shared environmental factors; E: individual-specific environmental factors; MZ monozygotic twin, DZ Dizygotic twin. [/fig]
[table] Table 1: Sample characteristics by study EPIC the HumanMethylationEPIC array, 27K the HumanMethylation27 array, 450K the HumanMethylation450 array, MZ monozygotic twin, DZ dizygotic twin, N sample size, SD standard deviation a Studies-PETS Peri/postnatal Epigenetic Twins Study, including three datasets measured using the 27K array (using three biological samples), 450K array (at two points: at birth and age 18 months), and EPIC array, respectively; BSGS Brisbane System Genetics Study, E-Risk Environmental Risk Longitudinal Twin Study, DTR Danish Twin Registry, in two groups: younger and older adults, AMDTSS Australian Mammographic Density Twins and Sisters Study, MuTHER Multiple Tissue Human Expression Resource Study, OATS Older Australian Twins Study, LSADT Longitudinal Study of Aging Danish Twins, with samples collected at years 1997 and 2007, respectively, MCCS Melbourne Collaborative Cohort Study [/table]
[table] Table S3: Under this model, different pairs shared the effects of environmental factors to different extents. The effects for MZ pairs were 1.41 (95% CI 1.16 to 1.66) times those for DZ and sibling pairs, and the latter were 2.03 (95% CI 1.13 to 9.47) times those for parent-offspring Fig. 1 Correlation between chronological age and DNAm age measured by the epigenetic clock within each study. PETS: Peri/postnatal Epigenetic Twins Study, including three datasets measured using the 27K array, 450K array, and EPIC array, respectively; BSGS: Brisbane System Genetics Study; E-Risk: Environmental Risk Longitudinal Twin Study; DTR: Danish Twin Registry; AMDTSS: Australian Mammographic Density Twins and Sisters Study; MuTHER: Multiple Tissue Human Expression Resource Study; OATS: Older Australian Twins Study; LSADT: Longitudinal Study of Aging Danish Twins; MCCS: Melbourne Collaborative Cohort Study [/table]
[table] Table 2: Within-study familial correlations in DNAm age MZ monozygotic twin, DZ dizygotic twin, CI confidence interval a Studies-PETS Peri/postnatal Epigenetic Twins Study, including three datasets measured using the 27K array, 450K array (at two points: at birth and age 18 months), and EPIC array, respectively; BSGS Brisbane System Genetics Study; E-Risk Environmental Risk Longitudinal Twin Study, DTR Danish Twin Registry, in two groups: younger and older adults, AMDTSS Australian Mammographic Density Twins and Sisters Study, MuTHER Multiple Tissue Human Expression Resource Study, OATS Older Australian Twins Study, LSADT Longitudinal Study of Aging Danish Twins, with samples collected at years 1997 and 2007, respectively, MCCS Melbourne Collaborative Cohort Study [/table]
|
Practical aspects of protein co-evolution
Co-evolution is a fundamental aspect of Evolutionary Theory. At the molecular level, co-evolutionary linkages between protein families have been used as indicators of protein interactions and functional relationships from long ago. Due to the complexity of the problem and the amount of genomic data required for these approaches to achieve good performances, it took a relatively long time from the appearance of the first ideas and concepts to the quotidian application of these approaches and their incorporation to the standard toolboxes of bioinformaticians and molecular biologists. Today, these methodologies are mature (both in terms of performance and usability/implementation), and the genomic information that feeds them large enough to allow their general application. This review tries to summarize the current landscape of co-evolution-based methodologies, with a strong emphasis on describing interesting cases where their application to important biological systems, alone or in combination with other computational and experimental approaches, allowed getting new insight into these.
# Introduction
It is difficult to fully understand evolutionary phenomena without taking into account the important role played by co-evolution. Co-evolution, which can be defined as the interdependence between the evolutionary changes of two entities, plays an important role at all biological levels, from ecosystems to molecules. Co-evolution was first described at the species level. C. Darwin himself described the entangled evolution of orchids and their pollinators, in the sense that the length of the proboscis of the latest was related to the size of the orchid's corolla. In the first half of the XX century, other biologists continued studying this phenomenon and establishing its genetic basis [bib_ref] Genetics of natural populations. XIX. Origin of heterosis through natural selection in..., Dobzhansky [/bib_ref]. The term "co-evolution" was originally coined by P. Ehrlich, who studied this phenomenon at the species level [bib_ref] Butterflies and plants: a study in coevolution, Ehrlich [/bib_ref]. The definition of co-evolution as "reciprocal evolutionary change in interacting species" [bib_ref] Diversification through multitrait evolution in a coevolving interaction, Thompson [/bib_ref] is the most accepted one today. From these early works, the idea of "interaction" becomes intimately associated to co-evolution. Co-evolution takes place between related or interacting entities and that is actually the reason for its utility at the molecular level.
The study of co-evolution at the molecular level is much more recent . At this level, co-evolution is evident between protein residues, so that in many cases changes (mutations) in positions related by functional or structural (i.e., space closeness) reasons are correlated. The practical utility of this observation is the prediction of residue contacts in protein structures, using sequence information as the only input . Going up in the "molecular hierarchy," co-evolution is also evident between interacting and functionally related proteins. Many pairs of interacting proteins show entangled evolutionary histories. Such evolutionary entanglement can lead, in the extreme, to the disappearance of one of the proteins when the other is lost. This extreme phenomena is reflected in related patterns of presence/absence of the two proteins in a set of genomes, which is actually the basis of the "phylogenetic profiling" methodology for detecting interacting proteins . In other cases, the evolutionary entanglement of interacting proteins is reflected in similar evolutionary histories but without reaching the extreme of the co-disappearance of the proteins. Since protein evolutionary histories are represented by phylogenetic trees, a common way of inferring protein co-evolution is by quantifying the similarity of the phylogenetic trees of the corresponding families [bib_ref] Similarity of phylogenetic trees as indicator of protein-protein interaction, Pazos [/bib_ref]. Such idea was inspired by observations at the species level: it was described that the phylogenetic trees of interacting species (e.g., parasites and their hosts or predators and preys) were similar, reflecting a process of coadaptation between them. Back to the protein level, it was shown that there is a consistent relationship between tree similarity and interaction (physical or functional) of the corresponding proteins. That observation led to a large family of methodologies that predict protein interactions based on similarity of phylogenetic trees, using only sequence information as input.
Although co-evolution-based methodologies continue to be developed and improved, they reached a point at which they are mature enough to be used by the community and form part of the standard toolbox of computational methods used by Molecular Biologists. Not only because their performances, both in terms of accuracy and coverage, increased in the last years, but also because they are now implemented in usable software and web interfaces. The aim of this review is to provide an overview of the current landscape of the main co-evolution-based methodologies, including recent examples of their application to different biological systems.
## Co-evolutionary approaches
The evolutionary forces entangling interacting proteins very often restrict their sequence evolution to the point of being perceivable at a genomic level. In a time governed by the "omics" techniques, a family of computational methods aim to detect the marks left on the genome by co-evolving molecules as a symptom of interaction [bib_ref] Deciphering protein-protein interactions. Part II. Computational methods to predict protein and domain..., Shoemaker [/bib_ref] [bib_ref] Emerging methods in protein coevolution, Juan [/bib_ref]. The associations detected do not necessarily imply physical interaction, but can also reflect involvement in similar biological functions, such as the same protein complex, the same metabolic pathway or the same operon. In this section, we review the different computational methods of co-evolutionary basis, focusing on their application scope and potential limitations .
## Phylogenetic profiling
Methods based on phylogenetic profiles rely on the observation that functionally associated and potentially interacting proteins evolve in a codependent manner tending to be jointly inherited or eliminated. This extreme case of co-evolution between functionally related genes has been explained as a consequence of "reductive evolution," where the loss of one of the members of the cooperative interaction dismisses the evolutionary pressure to maintain its partner. Alternatively, the recruitment of a new protein requires the acquisition of its partner to form the new functional complex. As a consequence of this phenomenon, the patterns of presence/absence of the two interacting partners in a set of genomes would tend to be similar.
A phylogenetic profile summarizes that pattern of presence/absence of a given gene in a set of reference organisms. At first, the profiles were encoded as binary representations, where "1" denotes the presence of an ortholog gene in a given organism, and "0" its absence [bib_ref] Microbial genescapes: phyletic and functional patterns of ORF distribution among prokaryotes, Gaasterland [/bib_ref] [bib_ref] Detecting protein function and protein-protein interactions from genome sequences, Marcotte [/bib_ref] [bib_ref] Assigning protein functions by comparative genome analysis: protein pylogenetic profiles, Pellegrini [/bib_ref]. Changes in the information contained on the profiles lead to a number of variations of the original phylogenetic profiling approach. For instance, instead of being binary, the profiles can contain quantitative information, such as the similarity of the ortholog with that in a reference organism [bib_ref] Discovery of uncharacterized cellular systems by genome-wide analysis of functional linkages, Date [/bib_ref]. Other profiles successfully encoded phenotypic traits to predict functional linkages [bib_ref] Trait-to-gene: a computational method for predicting the function of uncharacterized genes, Levesque [/bib_ref] [bib_ref] Assigning functional linkages to proteins using phylogenetic profiles and continuous phenotypes, Gonzalez [/bib_ref]. On the other hand, although the phylogenetic profiles were originally designed to contain information at full-sequence level, profiles based on domain presence/absence successfully predicted domain interactions [bib_ref] A domain interaction map based on phylogenetic profiling, Pagel [/bib_ref] [bib_ref] Predicting protein function with hierarchical phylogenetic profiles: the Gene3D Phylo-Tuner method applied..., Ranea [/bib_ref].
Profile-profile similarity has been calculated using different metrics such as euclidean distance , mutual information [bib_ref] Discovery of uncharacterized cellular systems by genome-wide analysis of functional linkages, Date [/bib_ref] or Hamming distance [bib_ref] Identification of functional links between genes using phylogenetic profiles, Wu [/bib_ref].
Besides profile similarity, functional linkage has also been observed between pairs of anti-correlated profiles encoding for pairs of genes excluding each other [bib_ref] Systematic discovery of analogous enzymes in thiamin biosynthesis, Morett [/bib_ref]. In a similar way, higher order relationships described by logic operators have been explored in order to look for complementation and other functional relationships relating triplets of profiles [bib_ref] Use of logic relationships to decipher protein network organization, Bowers [/bib_ref]. Another interesting phenomenon evident in the phylogenetic profiles of some pairs of interacting proteins is the "disrupted co-occurrence" (the presence of A implies that of B but not the other way around). These cases can point to asymmetric protein relationships (A needs B but B does not need A) and as such provide additional functional information to static interactions [bib_ref] Asymmetric relationships between proteins shape genome evolution, Notebaart [/bib_ref] [bib_ref] Shared protein complex subunits contribute to explaining disrupted co-occurrence, Schneider [/bib_ref].
A comprehensive list of pre-calculated similarities between protein phylogenetic profiles can be found in resources such as STRING [bib_ref] STRING: a database of predicted functional associations between proteins, Mering [/bib_ref] , Prolinks [bib_ref] Prolinks: a database of protein functional linkages derived from coevolution, Bowers [/bib_ref] or ECID (Andres [bib_ref] EcID. A database for the inference of functional interactions in E. coli, Leon [/bib_ref].
From a practical perspective, one of the most critical issues on phylogenetic profiling methodologies is the selection of the reference set of organisms. The number of completely sequenced genomes continues growing. Nevertheless, the best predictions are not always obtained with profiles based on all the available genomic sequences [bib_ref] Phylogenetic profiles for the prediction of protein-protein interactions: how to select reference..., Sun [/bib_ref]. Indeed, the accumulation of close organisms, as well as the taxonomic bias in the sequenced genomes affect the profiles, decreasing their performance in interaction prediction. Moreover, depending on the type of functional relationship the prediction is aimed at, the optimal set of organisms might change. Profiles based on organisms belonging to the three super-kingdoms display better performances for detecting conserved interactions, whereas species in the same superkingdom are more accurate for pathways . More systematic approaches using 565 bacterial genomes confirmed that sub-samples of organisms can achieve better performances than the whole set of available genomes [bib_ref] Effect of reference genome selection on the performance of computational methods for..., Muley [/bib_ref]. In order to automatically select the reference set of organisms, recent studies use machine-learning algorithms trained with known sets of interactions to improve the accuracies of the arbitrary selection of organisms [bib_ref] Automatic selection of reference taxa for protein-protein interaction prediction with phylogenetic profiling, Simonsen [/bib_ref].
From an evolutionary perspective, the presence/absence of every gene in a phylogenetic profile is equally weighted, independently of the number of potential evolutionary events needed to explain it. The number of potential gene gains/losses might be informative in order to estimate the statistical likelihood of a similarity score. As a matter of chance, similarities based on multiple evolutionary events will be more reliable than those with the same score but based on a fewer number of events. The idea of combining phylogenetic profiles with phylogenetic trees to weight the gene co-presence or co-absence is exploited in different studies by using Markov models [bib_ref] Predicting functional gene links from phylogenetic-statistical analyses of whole genomes, Barker [/bib_ref] [bib_ref] Uncovering the coevolutionary network among prokaryotic genes, Cohen [/bib_ref] , kernel trees [bib_ref] A tree kernel to analyse phylogenetic profiles, Vert [/bib_ref] or explicit comparisons [bib_ref] Inferring functional linkages between proteins from evolutionary scenarios, Zhou [/bib_ref]. Another limitation arises from the fact that some gene clusters might strongly co-evolve in some parts of the evolutionary tree, while exhibiting a weak co-dependency in other organisms. This non-homogenous distribution of the co-evolution, referred as "local co-evolutionary" problem, has been subject of different studies, but remains as a computationally challenging task [bib_ref] Locally defined protein phylogenetic profiles reveal previously missed protein interactions and functional..., Kim [/bib_ref] [bib_ref] Discovering local patterns of coevolution: computational aspects and biological examples, Tuller [/bib_ref].
Besides the previously mentioned limitations, some technical issues have to be addressed when comparing phylogenetic profiles. This methodology requires complete and well-annotated genomes to be sure of the existence or absence of a given gene. Even in those cases, the orthology assignment is not trivial, FIGURE 1 | Extracting co-evolutionary linkages from genomic information to study protein interactions. Top panel: the large amount of available sequence and genomic information is used to construct phylogenetic profiles (patterns of presence/absence of the genes in a set of organisms) and phylogenetic trees at a genome-wide scale. Middle panel:
Co-evolving pairs of proteins (green and red) are detected by the similarity of their phylogenetic patterns and/or the similarity of their phylogenetic trees. Bottom panel: the co-evolutionary linkages obtained in this way contain a lot of information on the interactions and functional relationships for the proteins in the organisms of interest.
being particularly critical in eukaryotes, where the presence of multiple domains, pseudogenes or inactivated genes difficult the proper assignment. Furthermore, essential proteins or those specific of a given organism can not be addressed by this approach as they are encoded as flat profiles. In summary, this methodology displays optimal results when analyzing gene pairs with clear orthologs uniformly distributed on the tree of life and presenting a reasonable number of common gain/loss events.
## Similarity of phylogenetic trees
Phylogenetic profiles are based on genomic landmarks left by dramatic events affecting whole genes or genomic regions (genes gains and loses). However, that approach ignores subtle changes on the sequences of interacting proteins, which might be also reflecting co-evolution. Such coordinated sequence changes might shape the phylogenetic trees of interacting proteins increasing their similarities. The first observations of this phenomenon qualitatively described that the phylogenetic trees of some pairs of interacting families were more similar than expected [bib_ref] The coevolution of gene family trees, Fryxell [/bib_ref] [bib_ref] Species-specificity of the cohesin-dockerin interaction between Clostridium thermocellum and Clostridium cellulolyticum: prediction..., Pages [/bib_ref]. Despite not being quantified or assessed in an exhaustive way, the similarity between the phylogenetic trees of those protein families was interpreted as a symptom of protein co-evolution.
The first method to quantify tree similarities calculated the correlation of the distance matrices as descriptors of the phylogenetic trees. The algorithm was soon scaled up to predict protein interactions at a genome-wide scale based on similarities of automatically generated phylogenetic trees [bib_ref] Similarity of phylogenetic trees as indicator of protein-protein interaction, Pazos [/bib_ref]. This approach, generically termed mirrortree, uses a simple pipeline to evaluate the eventual interaction between a pair of proteins. On its initial implementation, for the two protein families for which co-evolution is to be evaluated, multiple sequence alignments are generated aligning all the orthologs present in a set of reference genomes. Phylogenetic trees for each of the protein families are generated from the multiple sequence alignments, frequently using fast and simple algorithms such as neighborjoining. Finally, tree similarities are estimated by calculating the correlation coefficient between equivalent inter-ortholog distances in the two alignments. Consequently, unambiguous correspondence between the sequences of the two alignments is required, in order to allow the distances in both trees to be compared. This problem is normally solved by selecting one single ortholog per organism, leading to a natural mapping between the leaves of both trees, given by the organisms. Alternative solutions try to match the equivalent orthologs under the hypothesis that the correct mapping maximizes the tree similarity [bib_ref] Exploiting the co-evolution of interacting proteins to discover interaction specificity, Ramani [/bib_ref] [bib_ref] TSEMA: interactive prediction of protein pairings between interacting families, Izarzugaza [/bib_ref] [bib_ref] Enhancing the prediction of protein pairings between interacting families using orthology information, Izarzugaza [/bib_ref] [bib_ref] Codep: maximizing coevolutionary interdependencies to discover interacting proteins, Tillier [/bib_ref] [bib_ref] Mirroring co-evolving trees in the light of their topologies, Hajirasouliha [/bib_ref]. Other modifications of the original mirrortree algorithm suggest that when cophenetic distances are extracted from the branch lengths of the phylogenetic trees the prediction performance becomes slightly improved [bib_ref] Assessing protein co-evolution in the context of the tree of life assists..., Pazos [/bib_ref].
That pipeline is now implemented in the Mirrortree server, which provides a user-friendly web interface to allow non-expert users to overcome most of the aforementioned tasks [bib_ref] Studying the co-evolution of protein families with the Mirrortree web server, Ochoa [/bib_ref] and to interactively and graphically inspect the tree similarity. The server combines a powerful and automatic pipeline for tree reconstruction with an interactive interface to explore tree similarities. In the simplest case, the user can provide single sequences as input, although more advance users can provide their own alignments or even trees, and tune the parameters of the workflow.
One of the main limitations of the original mirrortree algorithm is the large number of false positives produced as a consequence of the unspecific tree similarities. One of the possible reasons for such a large amount of highly correlated trees between unrelated proteins can be due to the background tree similarity occurred as a consequence of the speciation events. Since the proteins under study are both affected by the ongoing speciation process, we expect both trees to display certain basal similarity, independently of their eventual interaction. The correction of that unspecific similarity due to the underlying speciation process (shared by both trees and the tree of life) is addressed by different methods using different statistical corrections and different representations of that background similarity. The first attempts used the 16SrRNA tree as a representation of the speciation process and tried to subtract its phylogenetic distances directly from the distance matrices of the interacting candidates [bib_ref] Assessing protein co-evolution in the context of the tree of life assists..., Pazos [/bib_ref] [bib_ref] The inference of protein-protein interactions by co-evolutionary analysis is improved by excluding..., Sato [/bib_ref]. The corrected methodology, renamed tol-mirrortree, obtained higher performances than the original mirrortree. More successful examples of co-evolution detection on ligand-receptor interactions have been reported, this time applying a background speciation correction [bib_ref] Parallel evolution between aromatase and androgen receptor in the animal kingdom, Tiwary [/bib_ref]. However, these corrections are incomplete in the sense that they consider each value in the distance matrix as independent, which is not the case for phylogenetic trees. If we change a given distance on the tree, the lengths of all other paths involving the modified branch should also be changed to adapt to the new distance. Some sophisticated methods try to consider the distance dependency problem by aligning high-dimensional embeddings of the trees [bib_ref] Comparison of phylogenetic trees through alignment of embedded evolutionary distances, Choi [/bib_ref]. Instead of using canonical trees to remove unspecific similarities, other methods use the tendencies obtained from large collections of protein families as an evidence of the background similarity. One of the first attempts to take advantage of this contextual information introduced the partial correlation coefficient as a measure of similarity. This metric calculates the correlation between a pair of phylogenetic vectors, excluding the information of a third vector containing the background information. By using the variability of the phylogenetic data as third vector, the prediction false positive rate was drastically reduced [bib_ref] Partial correlation coefficient between distance matrices as a new indicator of proteinprotein..., Sato [/bib_ref]. ContextMirror, an alternative method that also uses contextual information to reduce the background similarity, goes one step further: the unspecific signal associated to a pair of phylogenetic trees can be removed by comparing them with many others . As a preliminary step, this method calculates the pairwise mirrortree correlation coefficients between all the proteins in a given organism. In the resulting matrix of tree similarities, a coevolutionary profile is defined as the vector of correlation coefficients of a given protein with the rest of the proteome. The correlation between coevolutionary profiles is calculated as an estimate of how similar are the co-evolutionary patterns of both with the rest of the proteome. Alternatively, partial correlation between coevolutionary profiles reports the correlation of a pair of coevolutionary profiles when a third coevolutionary profile is taken into consideration. ContextMirror amazingly reduces the number of false positives, producing performances comparable to some experimental techniques .
The similarity of the phylogenetic trees, likewise the similarity of the phylogenetic profiles, is greatly influenced by the reference set of organisms used to generate the trees. In practical terms, disregarding technical issues such as the computational power required for generating and comparing trees based on all available genomes, two factors have to be considered when selecting the reference set: the problem of redundant organisms and the type of interactions intended to detect. As a consequence of the "non-uniform" sequencing efforts, the trees generated with all the sequenced genomes available nowadays contain a large bias toward the organisms of interest, for instance containing dozens of strains of some model bacteria. The mirrortree algorithm, far from benefiting from the new information, is severely affected by such genomic redundancy [bib_ref] Selection of organisms for the co-evolution-based study of protein interactions, Herman [/bib_ref]. As a consequence, independent studies suggest that the interaction prediction is improved when the organism redundancy has been removed. Those studies also suggest that the redundancy problem is partially overcome by some of the methods that remove background similarities, such as the correlation of coevolutionary profiles and ContextMirror [bib_ref] Selection of organisms for the co-evolution-based study of protein interactions, Herman [/bib_ref] ; or tol-mirrortree [bib_ref] Effect of reference genome selection on the performance of computational methods for..., Muley [/bib_ref]. On the other hand, the type of interaction to be detected constrains the selection of organisms. Certain subsets of organisms seem to be more suitable for predicting certain types of interactions. This result makes sense in the light of the phylogenetic distribution of the organisms and the nature of the predicted interactions. Local tree similarities involving close homologs are more likely to be related with recent interactions, whereas global similarities of the phylogenetic trees may evidence a co-evolution occurring since ancestral species [bib_ref] Selection of organisms for the co-evolution-based study of protein interactions, Herman [/bib_ref]. Supporting evidence suggest that mirrortree predictions normalized by the level of conservation (evolutionary span) of the candidate interaction significantly improve the interaction predictions [bib_ref] Predicting protein-protein interaction by the mirrortree method: possibilities and limitations, Zhou [/bib_ref]. Dealing with this non-homogenous nature of the co-evolutionary signal is not trivial as it raises certain combinatorial problems when trying to evaluate the similarity locally in all possible subsets of tree clades. A particularly successful method, MatrixMatchMaker (MMM), approaches this problem by looking for the largest common submatrix compatible with the evolutionary distance matrices under comparison [bib_ref] The human protein coevolution network, Tillier [/bib_ref]. MMM changes the paradigm of phylogenetic tree comparison by reducing the problem to the minimal common submatrix. The evolutionary span of the protein interaction is no longer relevant as the method dynamically adapts to maximize tree similarity. As a desired side effect, the method tolerates matrices including paralogs, since these will most likely be excluded from the final similarity if wrongly assigned. Although a recent implementation reduces the computationally expensive task of optimizing matrix similarity [bib_ref] A new, fast algorithm for detecting protein coevolution using maximum compatible cliques, Rodionov [/bib_ref] , this algorithm still demands a significant amount of resources when working with large number of sequences.
Co-evolution has also been observed at the residue level, as pairs of individual protein positions which are close in 3D or related in some way and tend to mutate in a coordinated fashion (see [bib_ref] Emerging methods in protein coevolution, Juan [/bib_ref] , for a review). Consequently, a number of methodologies try to infer co-evolution between two proteins based on the "accumulation" of co-evolutionary signals between their corresponding residues [bib_ref] In silico two-hybrid system for the selection of physically interacting protein pairs, Pazos [/bib_ref] [bib_ref] Detecting coevolution in and among protein domains, Yeang [/bib_ref] [bib_ref] Accurate prediction of protein-protein interactions from sequence alignments using a Bayesian method, Burger [/bib_ref]. This evidence of co-evolution at sub-protein levels also led some authors to study whether restricting the assessment of co-evolution to certain subsets of protein residues might increase the performance of the methods or provide additional information on the interactions. In most cases, these restrictions were based on structural criteria (surfaces, structural domains, etc), when such information is available. For instance, by comparing the domains of the alpha and beta subunits of the mithocondrial F1-ATP synthetase, seven pairs of domains that are known to interact present higher correlations than the two non-interacting pairs [bib_ref] Coevolutionary analysis of domains in interacting proteins reveals insights into domain-domain interactions..., Jothi [/bib_ref]. As when comparing full sequence proteins, these predictions improve their performance when removing the background similarity of the phylogenetic trees. Indeed, the predictions are more accurate when the background removal is applied to the trees based on the most conserved residues, indicating that both signals are more easily disentangled on those regions [bib_ref] Predicting protein domain interactions from coevolution of conserved regions, Kann [/bib_ref]. The presence of regions that not necessarily share the evolutionary constraints of the whole protein has also been tested on protein interfaces with contradictory results. Studies suggest that residues in the interfaces of stable interactions evolve at a relatively slow rate, consequently affecting the eventual coevolutionary signal with their interacting partners. In contrast, residues involved in transient interactions would present a higher plasticity, leaving little or no co-evolutionary signal in the interaction interfaces [bib_ref] Structure, function, and evolution of transient and obligate protein-protein interactions, Mintseris [/bib_ref]. In both cases, the residues not present in the interface still contain enough coevolutionary signal to predict the interaction [bib_ref] Correlated evolution of interacting proteins: looking behind the mirrortree, Kann [/bib_ref]. These results have been interpreted as a clear symptom that the co-evolutionary signal is uniformly distributed along the protein sequence showing no improvement by limiting the study to either the protein surface or the interaction interface [bib_ref] Specificity in protein interactions and its relationship with sequence diversity and coevolution, Hakes [/bib_ref]. Others reported a stronger co-evolutionary signal on the interfaces (including a structural neighborhood) than in the same number of randomly selected residues outside the binding neighborhood [bib_ref] Correlated evolution of interacting proteins: looking behind the mirrortree, Kann [/bib_ref]. These analyses were based on limited and not necessarily overlapping sets of structures, so the true extent of their conclusions is hard to evaluate. On the other hand, phylogenetic trees based on residues predicted as accessible have been shown to be more informative for predicting physical protein interactions [bib_ref] Incorporating information on predicted solvent accessibility to the coevolution-based study of protein..., Ochoa [/bib_ref]. Structural information is necessary and critical in order to fully understand interactions at the molecular level, nevertheless the definition of a general recipe on how to incorporate it in co-evolution-based methods remains elusive.
## Examples of applications
In this section we describe some examples of recent applications of co-evolution-based approaches to different biological problems. In principle, these methodologies can be applied to any organism, and indeed different groups used them to predict interactions in species covering the whole range of taxonomical diversity, from bacteria , to fungi [bib_ref] Using coevolution to predict protein-protein interactions, Clark [/bib_ref] and human [bib_ref] A census of human soluble protein complexes, Havugimana [/bib_ref]. The successful application to eukaryotic organisms is more recent since in those the (automatic) generation of accurate multiple sequence alignments and trees, key for applying these methodologies, has some additional difficulties compared with prokarya (location of orthologs, multidomain proteins,. . .).
Strong co-evolutionary signals are found in pairs of families where one of them has to accommodate its evolutionary rate to that of the other, accelerated for some reason. For example, the nuclear-encoded components of the NADHubiquinone reductase complex show such accelerated evolutionary rate to accommodate the intrinsically fast evolution of their mitochondrial-encoded counterparts. This results in evolutionary entanglements that can be used to predict interactions between these two sets of proteins, interactions that were latter confirmed experimentally [bib_ref] Coevolution predicts direct interactions between mtDNA-encoded and nDNA-encoded subunits of oxidative phosphorylation..., Gershoni [/bib_ref]. Co-evolution was also found between the mitochondrial-encoded rRNAs of the mitochondrial ribosomes and the nuclear-encoded proteins of these organelles [bib_ref] Evidence for compensatory evolution of ribosomal proteins in response to rapid divergence..., Barreto [/bib_ref]. A similar strategy based tree-similarity was used to study the co-evolution between the nuclear-and chloroplast-encoded members of the RuBisCO protein complex [bib_ref] Codon usage and coevolution of the large and small subunits of ribulose-1,5-bisphosphate..., Pei [/bib_ref].
A link between to apparently independent processes such as redox homeostasis and cellular timekeeping was found based on the presence of co-evolutionary signals [bib_ref] Studying the co-evolution of protein families with the Mirrortree web server, Ochoa [/bib_ref] between pairs of proteins of these processes [bib_ref] Peroxiredoxins are conserved markers of circadian rhythms, Edgar [/bib_ref]. That was complemented with experimental observations on the oxidation/reduction cycles of peroxiredoxin being universal markers of circadian rhythms in bacteria, eukaryotes and archaea, despite the huge mechanistical differences of these processes in the three superkingdoms [bib_ref] Peroxiredoxins are conserved markers of circadian rhythms, Edgar [/bib_ref].
Recently, [bib_ref] A census of human soluble protein complexes, Havugimana [/bib_ref] generated a large catalog of human soluble protein complexes combining experimental mass spectrometry with computational inference of interactions using, among others, the MMM tree-similarity-based co-evolutionary approach [bib_ref] The human protein coevolution network, Tillier [/bib_ref]. That approach had been previously used alone to obtain a human coevolutionary network that was shown to reflect protein physical interactions [bib_ref] The human protein coevolution network, Tillier [/bib_ref] [bib_ref] Coevolution reveals a network of human proteins originating with multicellularity, Bezginov [/bib_ref]. In another interesting combination of experimental and computational approaches, Lu et al. filtered the intrinsically noisy Hi-C data on "contacts" between chromosome regions using coevolutionary information so as to obtain a reliable prediction of the target genes for distal regulatory elements (DRE) in human [bib_ref] Combining Hi-C data with phylogenetic correlation to predict the target genes of..., Lu [/bib_ref]. In this case, they applied phylogenetic profiling to the presence/absence patterns of DREs and genes. Gene phylogenetic profiling was also recently used to generate a network of relationships between human proteins useful, among other things, to locate disease-related modules [bib_ref] Human disease locus discovery and mapping to molecular pathways through phylogenetic profiling, Tabach [/bib_ref].
Co-evolution is also especially evident in systems were the interaction patterns have to maintain interactions while continue evolving to acquire new functions and/or avoid crosstalk with the ancestral systems. This is the case of signaling cascades, where a paralogous expansion has to rapidly diverge to avoid interference with the original system, and such change has to be compensated by the interacting partners so as to maintain a functional cascade. In this sense strong co-evolutionary signals were found, for example between members of the bacterial two-component signaling system [bib_ref] Adaptive mutations that prevent crosstalk enable the expansion of paralogous signaling protein..., Capra [/bib_ref]. Molecular systems related to sex are another prototypic case of rapidly evolving systems where co-evolution plays an important role, since they have to differentiate and acquire specificity quickly so to avoid crossfertilization, while maintaining the specific interactions at the same time. In Brassica campestris, sequencing of 14 alleles allowed to find co-evolution between the (male) SCR and the (female) S receptor kinase [bib_ref] Highly divergent sequences of the pollen self-incompatibility (S) gene in class-I S..., Watanabe [/bib_ref]. This system is involved in the pollen discrimination mechanism. Similarly, deep-sequencing was recently used to study the co-evolution between male and female fertilization proteins of abalone snails [bib_ref] Coevolution of interacting fertilization proteins, Clark [/bib_ref]. Similar cases were found in Yeast. For example found that the proliferating cell nuclear antigen (PCNA) co-evolves with its interaction partners across the whole fungi phylogeny, what contributes to generate hybrid incompatibility and promoting speciation.
Transcription factors were also shown to co-evolve with their DNA-binding sites so as to maintain interactions while continue diverging [bib_ref] Coevolution within a transcriptional network by compensatory trans and cis mutations, Kuo [/bib_ref] [bib_ref] Correlated evolution of transcription factors and their binding sites, Yang [/bib_ref].
# Conclusions
From the first anecdotic qualitative observations of tree similarity for some pairs of related protein families (e.g., insulins and their receptors [bib_ref] The coevolution of gene family trees, Fryxell [/bib_ref] a lot of efforts were devoted to better understand the phenomenon of protein-protein co-evolution and to find practical ways of taking advantage of it. The genomic revolution allowed the genome-wide generation of multiple sequence alignments and trees so as to study the extent of this phenomenon and statistically assess its relationship with protein interactions [bib_ref] Similarity of phylogenetic trees as indicator of protein-protein interaction, Pazos [/bib_ref]. From that point onwards, variations of the original idea and new methodologies were developed so as to achieve higher accuracies and coverages in protein interaction prediction (see [bib_ref] Emerging methods in protein coevolution, Juan [/bib_ref]. These methodological improvements, together with user-oriented implementations of these methods and usable web interfaces (e.g., [bib_ref] Studying the co-evolution of protein families with the Mirrortree web server, Ochoa [/bib_ref] took us to a point where we can say that these approaches are mature enough to form part of the toolboxes of bioinformaticians and molecular biologists. And, as such, they are currently being used, alone or in combination with other computational and experimental approaches, for getting insight into important biological systems. Even if we still have a long way ahead in terms of improving these methodologies, they reach the required performance for being applied in a quotidian basis.
Co-evolution-based approaches, together with other computational approaches which also use sequence and genomic information for inferring protein linkages, form a family of approaches termed "context-based methods" [bib_ref] Predicting biological networks from genomic data, Harrington [/bib_ref] [bib_ref] Challenges for the prediction of macromolecular interactions, Wass [/bib_ref] , which complement the classical homology-based methods in obtaining information on different aspects of the proteins from their raw sequences (Von [bib_ref] STRING: a database of predicted functional associations between proteins, Mering [/bib_ref].
Co-evolution is not yet a completely understood phenomenon. Getting insight into its ultimate causes will contribute to the improvement of the methodologies. For example, it is not totally clear whether the observed co-evolution between interacting proteins is due to a process of specific co-adaptation or to more unspecific causes which could be "pushing" the evolutionary rates of the two proteins in a similar magnitude .
What is clear by the discussed examples and others is that the ultimate reason for the observed co-evolution seems to be allowing the two (interacting/related) partners to evolve and change while maintaining the interaction. An alternative way to maintain the interaction is to stay conserved, and indeed that is the case for some interactions. But in most interactions the partners have to evolve for one reason or another. In some cases this evolution is mainly "neutral," for example intrinsic rapid evolution due to lack of repair mechanisms in the genomes of eukaryotic organelles of bacterial origin, such as the mitochondrial and chloroplast examples commented. In these cases, the nuclear-encoded interactors of these proteins have to accommodate their evolutionary rates, and such co-evolutionary signal can be detected. In other cases one of the interactors simply changes to acquire new functions and loses the previous ones (to avoid crosstalk), and consequently the partner has to change too so as to maintain a functional complex. Again, we find a parallelism here with co-evolution at the species level (see Introduction), since it is known that, at that level, co-evolution is also allowing species to change while maintaining ecological interactions such as mutualism [bib_ref] Diversification through multitrait evolution in a coevolving interaction, Thompson [/bib_ref].
ACKNOWLEDGMENTSWe thank Alfonso Valencia and David Juan (CNIO) and the members of the Computational System Biology Group (CNB-CSIC) for interesting discussions. This work was partially supported by project BIO2010-22109 from the Spanish Ministry of Science and Innovation. |
Progress and Prospects of Nanocellulose-Based Membranes for Desalination and Water Treatment
# Introduction
The growing global population together with the natural water resource depletion can increase the freshwater requirement seven-fold [bib_ref] Nanoparticles in Reverse Osmosis Membranes for Desalination: A State of the Art..., Saleem [/bib_ref]. During the upcoming thirty years, it is predicted that the total world population can increase by more than 40 percent and, moreover, the need for domestic, agricultural, and industrial water sources will be extended, predominantly in the developing countries where the water requirement is comparatively higher with respect to its population. The World Water Council reported that by 2030, there would be an extraordinary possibility that almost 3900 million people would reside in water scarce areas. The problem of lack of water is not just an issue of appropriate techniques, but also it is a social and educational issue depending on the country, together with global activities and on technical solutions. For overcoming the problem of water shortages and the greater requirement for clean water, there is a necessity for fresh water sources together with preservation of existing water resources by means of a proper technique for water treatment. Desalination has already proved to be an excellent technology capable of developing the natural hydrological cycle by enriching it with water from oceans and several brackish resources. Desalination is the process of removing excess salts and minerals from saline water to make it fresh water. In water desalination, various technologies have been employed such as thermal-based and membrane-based technologies [bib_ref] Nanoparticles in Reverse Osmosis Membranes for Desalination: A State of the Art..., Saleem [/bib_ref].
Agricultural discharge from saline-alkali lands and the waste generated from pharmaceutical, agri-food, oil and gas, chemical, and aquaculture based industries are the main sources of waste water salinity [bib_ref] Adaptation of a freshwater anammox population to high salinity wastewater, Kartal [/bib_ref] [bib_ref] Removal of selenite from artificial wastewater with high salinity by activated sludge..., Zhang [/bib_ref] and these sources determine the concentration and composition of the saline water. When the inorganic salt content ranges from 1 to . Description of different desalination technologies with the required pressures and membrane pore size [bib_ref] Preparation of an Ultrafiltration (UF) Membrane with Narrow and Uniform Pore Size..., Khan [/bib_ref] [bib_ref] Evaluation of the Pore Size Distribution of a Forward Osmosis Membrane in..., Fang [/bib_ref] [bib_ref] Extra-Thin Composite Nanofiltration Membranes Tuned by γ-Cyclodextrins Containing Amphipathic Cavities for Efficient..., Zhao [/bib_ref] [bib_ref] Desalination Characteristics of Cellulose Acetate FO Membrane Incorporated with ZIF-8 Nanoparticles, Li [/bib_ref] [bib_ref] Pervaporation from a Dense Membrane: Roles of Permeant-Membrane Interactions, Kelvin Effect, and..., Sharma [/bib_ref] [bib_ref] Membrane Distillation: Basics, Advances, and Applications, Shirzad Kebria [/bib_ref].
## Process
Reverse Osmosis In order to enhance the physiochemical properties of the membranes, various nanomaterials have been used in the past few years [bib_ref] Fabrication and Characterization of Polyamide Thin Film Nanocomposite (TFN) Nanofiltration Membrane Impregnated..., Rajaeian [/bib_ref] [bib_ref] Improvement in Desalination Performance of Thin Film Nanocomposite Nanofiltration Membrane Using Amine-Functionalized..., Zarrabi [/bib_ref] [bib_ref] SiO2 Modified Polyethyleneimine-Based Nanofiltration Membranes for Dye Removal from Aqueous and Organic..., Kebria [/bib_ref]. The nanomaterials can be coated, grafted, or implanted in different layers of the membrane to enhance the physiochemical and performance ability of the membrane. However, some nanomaterials have been reported to be toxic to the environment and harmful to human health [bib_ref] Antibacterial Activity of Graphite, Graphite Oxide, Graphene Oxide, and Reduced Graphene Oxide:..., Liu [/bib_ref] [bib_ref] Advances of Nanomaterials for Air Pollution Remediation and Their Impacts on the..., Saleem [/bib_ref] [bib_ref] Sustainable Use of Nanomaterials in Textiles and Their Environmental Impact, Saleem [/bib_ref] and this necessitated the development of a renewable, low-cost, and environmentally friendly nanomaterial for the desalination application. The different cellulose based nanomaterials such as cellulose nanocrystals(CNCs), cellulose nanofibrils (CNFs), and bacterial nanocelluloses (BNCs) [bib_ref] Future Perspectives of Nanocellulose-Based Membrane for Water Treatment, Tan [/bib_ref] [bib_ref] Cellulose Nanocrystals: Synthesis, Functional Properties, and Applications, George [/bib_ref] [bib_ref] Bacterial Nanocellulose from Agro-Industrial Wastes: Low-Cost and Enhanced Production by Komagataeibacter Saccharivorans..., Abol-Fotouh [/bib_ref] [bib_ref] Bacterial Nanocellulose and Its Application in Wastewater Treatment, Muhamad [/bib_ref] [bib_ref] Production and Characterization of Cellulose Nanofiber Slurries and Sheets for Biomedical Applications, Carter [/bib_ref] are ideal candidates for addressing the challenges in the field of water desalination. Most of the membranes used for water treatment application have been fabricated from cellulose or non-cellulose organic polymers such as polypropylene, polysulfone (PSF), polyvinylidene fluoride (PVDF), [bib_ref] The Role of MOFs in Thin-Film Nanocomposite (TFN) Membranes, Van Goethem [/bib_ref] and polyethersulfone (PES) [bib_ref] Thin Film Deposition Techniques for Polymeric Membranes-A Review, Mavukkandy [/bib_ref]. The nanocelluloses (NCs) are developed from renewable and sustainable substances that are hydrophilic, non-toxic, and less harmful to the environment [bib_ref] Cellulose Nanomaterials Review Structure, Properties and Nanocomposites, Moon [/bib_ref]. Cellulose, the most common natural polymer, has a wide applicationin different fields and has a significant impact on the planet in various ways. The advancement of this material in the field of instrumentation and processing technology allowed researchers from different fields to work on the nanoscale structure of cellulose. Ref.The possibility of converting cellulose into nanoscale materials using various methods has opened up a new arena to investigate the properties and possible applications in wastewater treatment [bib_ref] Nanocellulose Recovery from Domestic Wastewater, Espíndola [/bib_ref]. The nanocelluloses are chosen over micro-sized materials for successful desalination, adsorption, and pollutant removal due to their high surface area, nano dimensions, good strength, and non-toxic nature [bib_ref] Nanocellulose-Based Products for Sustainable Applications-Recent Trends and Possibilities, Reshmy [/bib_ref] [bib_ref] Adsorption of Pb (II) from Aqueous Solutions Using Nanocrystalline Cellulose/Sodium Alginate/K-Carrageenan Composite..., Shen [/bib_ref] [bib_ref] Fabrication of Polyethyleneimine-Functionalized Magnetic Cellulose Nanocrystals for the Adsorption of Diclofenac Sodium..., Zhu [/bib_ref]. The wastewater might be hypersaline and also include metals, organic wastes, emulsions, organic and inorganic dyes, and even harmful microorganisms. In re-. The number of publications (by year) in which the expressions "nanocellulose" and "nanocellulose + membrane" were seen in the title, keywords, or abstract in the last 10 years. These data were obtained from the database scopus.com .
To the best of our knowledge, there is no study reported on the application of membranes based on different nanocellulose forms such as cellulose nanocrystals, cellulose nanofibrils, and bacterial nanocellulose for water desalination applications such as nanofiltration, reverse osmosis, pervaporation, forward osmosis, and membrane distillation. We begin this paper by explaining the structure and properties of nanocellulose. Then, we discuss the different applications of cellulose nanocrystals in membrane desalination technologies such as nanofiltration, reverse osmosis, and pervaporation. The next section discusses the different applications of cellulose nanofibers in membrane desalination technologies such as ultrafiltration, nanofiltration, and reverse osmosis. Subsequently, the applications of bacterial nanocellulose have been detailed, along with the outlook of nanocellulose.
## Structure and properties of nanocellulose
Cellulose is a polysaccharide made up of 1,4-D-pyran-type dehydrated polydextrose [bib_ref] Cellulose: Fascinating Biopolymer and Sustainable Raw Material, Klemm [/bib_ref]. [fig_ref] Figure 2: Structural representation of cellulose polymer[58] [/fig_ref] presents the structural representation of cellulose polymer. The β-1,4-link- in which the expressions "nanocellulose" and "nanocellulose + membrane" were seen in the title, keywords, or abstract in the last 10 years. These data were obtained from the database scopus.com .
To the best of our knowledge, there is no study reported on the application of membranes based on different nanocellulose forms such as cellulose nanocrystals, cellulose nanofibrils, and bacterial nanocellulose for water desalination applications such as nanofiltration, reverse osmosis, pervaporation, forward osmosis, and membrane distillation. We begin this paper by explaining the structure and properties of nanocellulose. Then, we discuss the different applications of cellulose nanocrystals in membrane desalination technologies such as nanofiltration, reverse osmosis, and pervaporation. The next section discusses the different applications of cellulose nanofibers in membrane desalination technologies such as ultrafiltration, nanofiltration, and reverse osmosis. Subsequently, the applications of bacterial nanocellulose have been detailed, along with the outlook of nanocellulose.
## Structure and properties of nanocellulose
Cellulose is a polysaccharide made up of 1,4-D-pyran-type dehydrated polydextrose [bib_ref] Cellulose: Fascinating Biopolymer and Sustainable Raw Material, Klemm [/bib_ref]. [fig_ref] Figure 2: Structural representation of cellulose polymer[58] [/fig_ref] presents the structural representation of cellulose polymer. The β-1,4-linkages in the cellulose connect linear D-glucopyranose units through intermolecular hydrogen bonding and hydrophobic interactions, while individual polymer chains give rise to fibrous structure. Each cellulose chain has directional disorderness due to the presence of terminal groups. A hemiacetal group exists at the reducing end of the cellulose chain while a hydroxyl group hangs from the non-reducing end.. The number of publications (by year) in which the expressions "nanocellulose" and "nanocellulose + membrane" were seen in the title, keywords, or abstract in the last 10 years. These data were obtained from the database scopus.com (24 February 2022).
## Figure 2.
Structural representation of cellulose polymer [bib_ref] Lignocellulosic Biomass-Derived Nanocellulose Crystals as Fillers in Membranes for Water and Wastewater..., Sadare [/bib_ref].
No. of publications (related to "Nanocellulose")
No. of publications (related to "Nanocellulose for Membrane Application") Each molecule of cellulose contains three alcoholic hydroxyl groups, and these groups can be chemically altered by different chemical reactions (polymer grafting, esterification acetylation, oxidation, and many more) [bib_ref] Recent Approaches and Future Prospects of Bacterial Cellulose-Based Electroconductive Materials, Chen [/bib_ref]. Chemical alteration can take place in both heterogeneous and homogeneous environments; since cellulose's strong crystallinity indicates that it can only be dissolved in a few solvents [bib_ref] Fabrication of Cellulose Nanofiber-Based Ultrafiltration Membranes by Spray Coating Approach, Wang [/bib_ref] , many alterations have been carried out in a heterogeneous environment.
The sources of origin for cellulose are algae, tunicates, bacteria, and some sea animals that are also capable of producing cellulose in abundant quantity [bib_ref] Cellulose Nanocrystals: Synthesis, Functional Properties, and Applications, George [/bib_ref]. This indicates that cellulose is made up of a mixture of largely crystalline (well ordered) and amorphous (disordered) areas, and these areas, when chemically or mechanically treated, will give rise to different forms of nanocellulose, which will be discussed in this review study.
The term "nanocellulose" refers to cellulose that has been nanostructured. The properties which make nanocellulose incredible are that they are harmless, renewable, and degradable, and these make nanocellulose currently useful in several advanced materials [bib_ref] Nanocellulose Based Filtration Membrane in Industrial Waste Water Treatment: A Review, Liu [/bib_ref] [bib_ref] A Review of Nanocellulose as a New Material towards Environmental Sustainability, Dhali [/bib_ref] [bib_ref] Nanocellulose in Biomedicine: Current Status and Future Prospect, Lin [/bib_ref]. Nanocellulose has a crystalline or fibrous nature that can be chemically and physically treated [bib_ref] Biocompatible Elastomer of Waterborne Polyurethane Based on Castor Oil and Polyethylene Glycol..., Gao [/bib_ref] , and the different types of nanocellulose include cellulose nanocrystals, cellulose nanofibers, and bacterial nanocellulose, as presented in [fig_ref] Figure 3: Properties, structure, and synthesis of different types of nanocelluloses[62] [/fig_ref] [bib_ref] Nanocellulose, a Versatile Green Platform: From Biosources to Materials and Their Applications, Thomas [/bib_ref]. The BNCs are bacteria-produced nano-structured celluloses. All the different types of celluloses are classified on the basis of their morphology and sources of extraction. The main difference between BNCs and other nanocellulose materials is that BNCs are prepared using a bottom-up approach, which involves bacterial cultivation in an aqueous culture media, whereas CNCs and CNFs can be made using a top-down approach, which involves chemical and mechanical treatments or a combination of mechanical, chemical, and enzymatic processes. Due to the unique properties, nanocelluloses are being used in various waste treatment techniques [bib_ref] Potential of Nanocellulose for Wastewater Treatment, Reshmy [/bib_ref] [bib_ref] Nanocellulose-Based Composite Materials for Wastewater Treatment and Waste-Oil Remediation, Yuan [/bib_ref] , and one of the most important areas where nanocelluloses are in higher demand is in membrane desalination technology. The incorporation of nanocellulose within the existing membrane can lead to an improvement in the performance of the membranes. The inclusion of nanocelluloses during the fabrication of the membrane can result in good selective permeability, enhanced salt rejection, increased water flux, etc. [bib_ref] Organic Fouling Behaviour of Structurally and Chemically Different Forward Osmosis Membranes-A Study..., Mazlan [/bib_ref].
processes [bib_ref] Nanocellulose, a Versatile Green Platform: From Biosources to Materials and Their Applications, Thomas [/bib_ref]. Due to the unique properties, nanocelluloses are being used in va waste treatment techniques [bib_ref] Nanocellulose, a Versatile Green Platform: From Biosources to Materials and Their Applications, Thomas [/bib_ref] [bib_ref] Potential of Nanocellulose for Wastewater Treatment, Reshmy [/bib_ref] [bib_ref] Nanocellulose-Based Composite Materials for Wastewater Treatment and Waste-Oil Remediation, Yuan [/bib_ref] , and one of the most important areas where nan luloses are in higher demand is in membrane desalination technology. The incorpor of nanocellulose within the existing membrane can lead to an improvement in the pe mance of the membranes. The inclusion of nanocelluloses during the fabricatio the membrane can result in good selective permeability, enhanced salt rejection, incre water flux, etc..
## Sources for obtaining nanocellulose
In general, the cellulose must be extracted from a source prior to the production of nanocellulose. To date, cellulose has been attained from an extensive range of sources such as plants, bacteria, algae, and tunicates. The properties of CNC are dependent on the source, and therefore appropriate cellulose source material and suitable methods for extraction should be recognized depending on the final properties needed as well as the applications of nanocellulose.
Due to its environmental advantages and the multifunctional nature of nanocellulose, several scientists have begun to examine the capability for producing nanocellulose by using microbial hosts. Just as in algae-based cellulose, bacterial cellulose (BC) has high purity due to the fact that it has no other polymers or functional groups apart from alcohol. Further, nanocellulose generated from bacterial cellulose has improved crystallinity as compared to the plant-derived nanocellulose . Cellulose is mostly generated from agricultural residues or by-products (sugarcane bagasse, rice husk, and corn), non-wood lignocellulose (cotton, kenaf, ramie, jute, flax, and hemp), and wood of forest resources. These plant sources for cellulose are plentiful globally and are easily accessible, almost vast, and inexpensively sourced, leading to a lower production cost for cellulose nanocrystals. The cellulose fibers form plants are one of the most widely researched and employed major sources for the production of nanocellulose. Generally, the plant-based cellulose is chosen over bacterial cellulose and tunicate-based cellulose when very thin nanofibers are needed for specific applications. The commonly seen type of plant fiber employed for the preparation of cellulose is wood pulps because of its comparatively higher purity and durability of cellulose, as well as ductile groups and better physical characteristics relative to other plant-based cellulose sources.
Significant advancements towards ecological sustainability have been achieved by employing the forthcoming algae, the underground weapon of green energy. The green alga (Cladophora) is a good source for the preparation of nanocellulose production. The application of the green filament algae for the generation of nanocellulose is interesting, as it could support resolving the challenges related to the pollution of water in coastal areas. In a study carried out by Mihranyan and group [bib_ref] Crystallization Behavior of Poly(ε-Caprolactone)/Layered Double Hydroxide Nanocomposites, Yang [/bib_ref] , the team studies the extremely crystalline properties of Cladophora cellulose, and it was suggested that the extreme inertness of the cellulose reduced the exposure to chemical treatments relative to cellulose developed from conventional plant-based sources.
Furthermore, nanocellulose is also generated from tunicates (invertebrate animals in sea). Several researchers have recognized the application of sea squirts specifically as a current substitute for the preparation of nanocellulose. Nechyporchuk and group [bib_ref] Facile Fabrication of Superhydrophilic Membranes Consisted of Fibrous Tunicate Cellulose Nanocrystals for..., Cheng [/bib_ref] has prepared an extremely effective filter membrane to separate oil/water by con-solidating extremely hydrophilic tunicate CNC and cholesteric liquid crystal structure.
## Methods for obtaining nanocellulose
Before the production of nanocellulose, most cellulose should be isolated from untreated cellulose pulp that consists of lignin and hemicelluloses. Additional treating of the refined cellulose pulp is necessary for producing CNC/CNF. Generally, the majority CNFs and CNCs are generated by breaking down the cellulose fibers into nanosized fragments (by top-down approach), except for electrospun CNF and bacterial cellulose that employ the electrospinning method (bottom-up approach) and bacteria, respectively. In general, the manufacturing methods use different processes sequentially for producing different kinds of CNC/CNF which differ with respect to surface chemistry, crystallinity, and morphology [bib_ref] Production of Cellulose Nanofibrils: A Review of Recent Advances, Nechyporchuk [/bib_ref]. The processes include chemical pretreatment, biological treatment, and mechanical disintegration. The appropriate methods applied for producing the nanocellulose from purified cellulose are examined in brief in the following section.
In general, it was noted that mechanical treatment is the main treatment used for producing the CNFs, whereas it is also employed as the post-treatment as well as purification step to produce CNC. Mechanical disintegration is frequently employed for breaking the cellulose pulp into small particles. Nevertheless, effective mechanical disintegration of cellulosic fibers commonly needs simple fibril delamination of cellulosic fiber rather than simply fiber shredding. In a study performed by Park and team [bib_ref] Preparation and Characterization of Cellulose Nanofibrils with Varying Chemical Compositions, Park [/bib_ref] , the team studied the chemical composition in wood-based CNF and its impact on the defibrillation effectiveness in wet disk-milling, which enhanced defibrillation in the non-existence of hemicellulose as well as lignin in cellulose.
In a research work by Fan and group [bib_ref] High Strength Natural Fiber Composite: Defibrillation and Its Mechanisms of Nano Cellulose..., Fan [/bib_ref] , it was noted that the application of chemical agents is a major contributor to the defibrillation process during the fabrication of nanocellulose. A commonly seen method for extracting CNC from cellulose is acid hydrolysis. Different from mechanical disintegration, the chemical treatment method breaks the amorphous region in microfibrils, putting away the crystalline areas undamaged. The application of green solvents as both solvent as well as catalyst for cellulose hydrolysis has turned out to be a standard technique for the extraction of nanocellulose.
In nanocellulose production, the enzymatic hydrolysis is a frequently utilized method prior to mechanical disintegration of cellulose. Different enzymes have been employed to prepare nanocellulose: ligninases, xylanases, pectinases, and cellulases. A study by Nechyporchuk and team [bib_ref] Production of Cellulose Nanofibrils: A Review of Recent Advances, Nechyporchuk [/bib_ref] stated that the rheological characteristic of CNF suspension prepared from the enzymatic pretreatment demonstrated excellent flocculation ability relative to CNF developed from chemical pretreatment.
## Application of cellulose nanocrystals in desalination application
Depending on the source, bulk cellulose contains a mixture of highly ordered, crystalline regions, and some disordered (amorphous) regions in variable amounts [bib_ref] Carbon-13 NMR Distinction between Categories of Molecular Order and Disorder in Cellulose, Newman [/bib_ref] , as presented in [fig_ref] Figure 4: Schematic representation of crystalline region and amorphous region in cellulose and the... [/fig_ref]. ity relative to CNF developed from chemical pretreatment.
## Application of cellulose nanocrystals in desalination application
Depending on the source, bulk cellulose contains a mixture of highly ordered, crystalline regions, and some disordered (amorphous) regions in variable amounts [bib_ref] High Strength Natural Fiber Composite: Defibrillation and Its Mechanisms of Nano Cellulose..., Fan [/bib_ref] , as presented in [fig_ref] Figure 5: Schematic representation of crystalline region and amorphous region in cellulose and the... [/fig_ref]. The highly crystalline areas of the cellulose microfibrils could be extracted only when a combination of mechanical, enzymatic, and chemical process was imposed on the microfibrils. After these stages, the amorphous part of the microfibrils is removed, leaving only cellulose nanocrystals [bib_ref] The Potential of Cellulose Nanocrystals in Tissue Engineering Strategies, Domingues [/bib_ref]. CNCs are hard rod-like particles with a nearly perfect crystalline structure made of cellulose chain segments. Whiskers, nanoparticles, nanofibers, micro crystallites, and other terms have been used to describe these nanocrystals, but CNC is the most frequently accepted term [bib_ref] Cellulose nanocrystals: Chemistry, self-assembly, and applications, Habibi [/bib_ref]. These nanocrystals have higher specific strength, modulus, higher surface area, and distinctive liquid crystalline characteristics as compared to bulk cellulose, which contains more amorphous fractions. CNCs have a high surface-to-volume ratio and a lot of hydroxyl groups, thus, these materials are good for different surface functionalization. The type of interactions that a material exhibits with its atmosphere can be varied by putting any chemical functionality on its surface, and this property of CNCs diverts the researcher's attention. Surface modification can affect membrane performance, and the surface of CNC has many side hydroxyl groups, which allows for chemical modification, and this can affect membrane performance directly. Desalination membrane results can enhance dramatically when high density functional groups are chemically bonded with reactive hydroxyl groups [bib_ref] Cellulose Nanomaterials Review Structure, Properties and Nanocomposites, Moon [/bib_ref]. By implanting chemical modifiers, a highly cross-linked and more permissive and enhanced salt rejection has been observed as a result. Moreover, the improved surface hydrophilicity provided by the hydrophilic and highly functionalized nature of the CNC modified membrane together with its surface smoothness resulted in its higher organic fouling resistance and mechanical strength compared to commercial membranes. In this section, we have reviewed how the nanocellulose crystal inclusion, whether unmodified or modified, has improved the desalination efficiency of the membranes.
## Cellulose nanocrystals application in nanofiltration
Nanofiltration is a technique of removing undesirable substances and pollutants with micro or nanoscale dimensions using a semipermeable membrane. Pore sizes in nanofiltration membranes should be between 1 and 10 nm, and on comparing the pore diameters of membrane used in different technologies, it can be noted that nanofiltration membranes have smaller pore diameters than MF and UF membranes but larger pore diameters than RO membranes. Due to environmental and energy concerns, highly permeable nanofiltration membranes are in high demand for water purification (desalination) and soluble particle separation [bib_ref] Extra-Thin Composite Nanofiltration Membranes Tuned by γ-Cyclodextrins Containing Amphipathic Cavities for Efficient..., Zhao [/bib_ref] [bib_ref] Sustainable Nanofiltration Membranes Based on Biosourced Fully Recyclable Polyesters and Green Solvents, Hardian [/bib_ref] [bib_ref] Fabrication of CuO Nanoparticles Immobilized Nanofiltration Composite Membrane for Dye/Salt Fractionation: Performance..., Waheed [/bib_ref]. Researchers across different fields are working on fabrication and modification of nanofiltration membranes by using different approaches [bib_ref] Enhancement of Desalination Performance of Thin-Film Nanocomposite Membrane by Cellulose Nanofibers, Liu [/bib_ref] [bib_ref] A Cellulose-Based Nanofiltration Membrane with a Stable Three-Layer Structure for the Treatment..., Wang [/bib_ref] [bib_ref] Ultra-Low Pressure Cellulose-Based Nanofiltration Membrane Fabricated on Layer-by-Layer Assembly for Efficient Sodium..., Li [/bib_ref]. However, the chemicals utilized to manufacture NF membranes are not environmentally sustainable. Cellulose, on the other hand, is the most common naturally occurring polymer consisting of highly oriented microfibrils, and it is an ideal material for the preparation of biodegradable NF membranes because it is renewable, biocompatible, and sustainable [bib_ref] Lignin-Containing Cellulose Nanomaterials: A Promising New Nanomaterial for Numerous Applications, Ewulonu [/bib_ref] [bib_ref] Transparent and Conductive Cellulose Film by Controllably Growing Aluminum Doped Zinc Oxide..., Liu [/bib_ref] [bib_ref] Fabrication of Bacterial Cellulose/Polyaniline Nanocomposite Paper with Excellent Conductivity, Strength, and Flexibility, Fei [/bib_ref].
This section of the study features certain recent studies carried out for the preparation and testing of NF membranes developed using cellulose nanocrystal used in the nanofiltration application. Wang and group [bib_ref] Nanofiltration Membranes with Cellulose Nanocrystals as an Interlayer for Unprecedented Performance, Wang [/bib_ref] have developed an NF membrane and, when operating it at 0.6 MPa pressure, the fabricated membrane showed ultra-high permeation flux up to 204 L m −2 h −1 . Furthermore, the efficiency for salt rejection is more than 97% for Na 2 SO 4 , which they claim as the highest obtained result to date. The Donnan and steric effects of the CNC molecules were responsible for the rejection performance of nanofiltration membranes, resulting in the following rejection order: Na 2 SO 4 (97.7%) > MgSO 4 (86%) > MgCl 2 (15.5%) > CaCl 2 (11%) > NaCl (6.5%). The extremely higher water permeation flux facilitates nanofiltration at lower operational pressure, thereby making it to an energy-saving process.
Bai and group [bib_ref] Fabrication and Characterization of Thin-Film Composite (TFC) Nanofiltration Membranes Incorporated with Cellulose..., Bai [/bib_ref] A group fabricated a membrane by incorporating cellulose nanocrystals into a polyamide nanofiltration membrane (CNC-TFC-Ms), and it was found that the attachment of CNC molecules in the PA layer reduced the pore size. According to the pore size distribution data, the particle size distribution (PSD) curves became narrower, yet the permeate flux and rejection of NaCl both exhibited a positive deviation when compared to pristine membrane. Moreover, according to side potential assessment, the membrane surface negativity decreased, and the rejection rates of over 98.0% and 97.5% for Na 2 SO 4 and MgSO 4 have been reported by the group. This confirmed that the CNC modified NF membrane improved the rejection results for the mono and divalent salts. The CNC-TFC-Ms membrane's NaCl/Na 2 SO 4 selectivity reached to 60, indicating their distinct divalent/monovalent separation property.
In a study by Xu and group [bib_ref] Cellulose Nanocrystal/Silver (CNC/Ag) Thin-Film Nanocomposite Nanofiltration Membranes with Multifunctional Properties, Xu [/bib_ref] , the team modified PA layer of NF membrane by introducing CNC/silver composite to analyze the post fabrication performance of the membrane. The CNC/Ag modified thin film nanocomposite NF membrane showed some good results with respect to pure water permeability as well as rejection rate and, thus, 25.4 L m −2 h −1 bar −1 water permeability and Na 2 SO 4 (99.1%) salt rejection rate was obtained. In addition, the membrane also showed outstanding antifouling performance with humic acid's flux recovery ratio of 92.6% and antibacterial properties of 99.4% reduction in E. coli feasibility.
Bai and group [bib_ref] Incorporation of Cellulose Nanocrystals (CNCs) into the Polyamide Layer of Thin-Film Composite..., Bai [/bib_ref] demonstrated modification of membrane by introducing CNCs into NF membranes and comparing their permeability to the unmodified NF membrane. The team discovered that the modified membrane showed a 60% increase in permeability, and this result was obtained by introducing 0.020 wt.% of CNC into the PA layer. It was also observed that the salt rejection performance of the CNC-TFC membrane was outstanding for both mono and divalent salts Na 2 SO 4 and MgSO 4 with the results 98.7% and 98.8%, respectively. In addition, the salt rejection rate for monovalent ions was improved by increasing the concentration of CNCs during the modification of NF membrane. The CNC-TFC membranes showed good antifouling property due to the charge on the surface and increased hydrophilicity, as reported by the group. The antifouling ability of the CNCincorporated thin film composite membranes was examined by the filtration of 500 ppm humic acid. In the first stage, the membranes were compacted, employing distilled water for achieving a stabilized water flux. Subsequently, the feed solution was replaced with humic acid solution, and then the entire CNC-incorporated thin film composite membranes experienced a serious reduction in the water flux. The results of thin film nanocomposite membrane testing results confirmed that the resistance against humic acid fouling of the membranes was enhanced by the inclusion of CNCs into the PA layer. In the second stage, the membranes were filtered for four hours until a stabilized humic acid flux was accomplished, and subsequently, the membranes were carefully cleaned physically using the distilled water. In the third stage, the cleaned CNC-incorporated thin film composite membranes were filtered using distilled water, and the corresponding flux was recorded. Relative to the unmodified membrane, the higher flux recovery ratios (FRR) values of the thin film nanocomposite membranes indicated an improved cleaning efficiency with higher nanocellulose concentration. The flux decline ratio (FDR) and FRR results confirmed that the nanocomposite membranes performed better than the control composite membrane in both cleaning efficiency and humic acid fouling resistance.
In a different research work by Bai and group [bib_ref] High-Performance Nanofiltration Membranes with a Sandwiched Layer and a Surface Layer for..., Bai [/bib_ref] , the team studied TFC modified membranes with cellulose nanocrystal sandwiched layer and a polydopamine (PDA) layer to fabricate the microfiltration substrate [fig_ref] Figure 5: Schematic representation of crystalline region and amorphous region in cellulose and the... [/fig_ref]. It was found that the CNC interlayer increased membrane electronegativity along with increasing solute permeability and rejection. The modified membranes had a high removal rate for sodium lignosulfonate, congo red, rose Bengal, and alkaline lignin. The salt rejection performance can be explained by the results of the permeability and rejection for MgSO 4 and CaCl 2 , which was increased by 152% and 8.0% for modified membrane, respectively, when compared to the bare TFC membrane.
Membranes 2022, 12, x FOR PEER REVIEW 10 of 34 flux was accomplished, and subsequently, the membranes were carefully cleaned physically using the distilled water. In the third stage, the cleaned CNC-incorporated thin film composite membranes were filtered using distilled water, and the corresponding flux was recorded. Relative to the unmodified membrane, the higher flux recovery ratios (FRR) values of the thin film nanocomposite membranes indicated an improved cleaning efficiency with higher nanocellulose concentration. The flux decline ratio (FDR) and FRR results confirmed that the nanocomposite membranes performed better than the control composite membrane in both cleaning efficiency and humic acid fouling resistance.
In a different research work by [bib_ref] Incorporation of Cellulose Nanocrystals (CNCs) into the Polyamide Layer of Thin-Film Composite..., Bai [/bib_ref] , the team studied TFC modified membranes with cellulose nanocrystal sandwiched layer and a polydopamine (PDA) layer to fabricate the microfiltration substrate [fig_ref] Figure 7: Diagrammatic representation of the CNC-TFC-PDA membrane preparation [/fig_ref]. It was found that the CNC interlayer increased membrane electronegativity along with increasing solute permeability and rejection. The modified membranes had a high removal rate for sodium lignosulfonate, congo red, rose Bengal, and alkaline lignin. The salt rejection performance can be explained by the results of the permeability and rejection for MgSO4 and CaCl2, which was increased by 152% and 8.0% for modified membrane, respectively, when compared to the bare TFC membrane. Ref. [bib_ref] High-Performance Nanofiltration Membranes with a Sandwiched Layer and a Surface Layer for..., Bai [/bib_ref] fabricated polydopamine-modified cellulose nanocrystals incorporated on a thin-film nanocomposite membrane for the application of nanofiltration, showing ultrahigh pure water permeability of 128.4 LMH bar −1 , superior Congo red rejection of 99.91%, and remarkable salt permeation of 99.33%. Furthermore, on a Congo-red/NaCl mixture, a remarkable salt/dye selectivity of 98% was attained, revealing its enormous potential for successful separation of dye/salt mixtures and recovery of these vital components. Bai and group [bib_ref] High Performance Nanocomposite Nanofiltration Membranes with Polydopamine-Modified Cellulose Nanocrystals for Efficient Dye/Salt..., Yang [/bib_ref] studied an advanced cellulose nanocrystal incorporated TFN nanofiltration membrane, which showed enhanced water permeability as well as sodium chloride rejection.
All the findings noted from the above-mentioned studies will definitely have a big impact on the future development of high-performance NF membranes for water treat- Yang and group [bib_ref] High Performance Nanocomposite Nanofiltration Membranes with Polydopamine-Modified Cellulose Nanocrystals for Efficient Dye/Salt..., Yang [/bib_ref] fabricated polydopamine-modified cellulose nanocrystals incorporated on a thin-film nanocomposite membrane for the application of nanofiltration, showing ultrahigh pure water permeability of 128.4 LMH bar −1 , superior Congo red rejection of 99.91%, and remarkable salt permeation of 99.33%. Furthermore, on a Congored/NaCl mixture, a remarkable salt/dye selectivity of 98% was attained, revealing its enormous potential for successful separation of dye/salt mixtures and recovery of these vital components. Bai and group [bib_ref] Biodegradability and Compostability of Nanofibrillar Cellulose-Based Products, Vikman [/bib_ref] studied an advanced cellulose nanocrystal incorporated TFN nanofiltration membrane, which showed enhanced water permeability as well as sodium chloride rejection.
All the findings noted from the above-mentioned studies will definitely have a big impact on the future development of high-performance NF membranes for water treatment. The different modifications of the membranes such as the incorporation of pristine CNC, polydopamine-modified cellulose nanocrystals, CNC/silver composite, and cellulose nanocrystal sandwiched layer preparation can definitely improve the performance of membranes in the nanofiltration process.
## Cellulose nanocrystals application in reverse osmosis
Reverse osmosis (RO) is a purification technology that uses a semipermeable membrane to remove ions, hazardous chemicals, and micro and macro particles from contaminated water [fig_ref] Figure 6: A conventional reverse osmosis system [/fig_ref]. An applied external pressure is used in this technology to counteract the osmotic pressure (generated due to concentration difference). Reverse osmosis desalination systems are widely used around the world because this technology can produce freshwater at the lowest cost [bib_ref] Frontiers of Membrane Desalination Processes for Brackish Water Treatment: A Review, Honarparvar [/bib_ref]. The most significant part in RO technology is played by the membranes as the small molecules of water permeate through the membrane pores and generate pure water, while all the contaminants present in water are flushed and the discharged water is termed as reject water. As membranes are semipermeable, salt ions are not permitted to flow across the membrane, and they will be carried within the reject stream [bib_ref] Performance of Reverse Osmosis Based Desalination Process Using Spiral Wound Membrane: Sensitivity..., Aladwani [/bib_ref]. Nanocellulose based membranes could be employed for improving the reverse osmosis performance of membranes for desalination and considerably decreased the desalination cost of USD$5.4 million in 2019 to USD$0.5-1.2 million in Israel, Australia, China, Saudi Arabia, and United States [bib_ref] Innovative Nanostructured Membranes for Reverse Osmosis Water Desalination, Saleem [/bib_ref].
## Cellulose nanocrystals application in reverse osmosis
Reverse osmosis (RO) is a purification technology that uses a semipermeable membrane to remove ions, hazardous chemicals, and micro and macro particles from contaminated water [fig_ref] Figure 8: A conventional reverse osmosis system [/fig_ref]. An applied external pressure is used in this technology to counteract the osmotic pressure (generated due to concentration difference) [bib_ref] Biodegradability and Compostability of Nanofibrillar Cellulose-Based Products, Vikman [/bib_ref]. Reverse osmosis desalination systems are widely used around the world because this technology can produce freshwater at the lowest cost. The most significant part in RO technology is played by the membranes as the small molecules of water permeate through the membrane pores and generate pure water, while all the contaminants present in water are flushed and the discharged water is termed as reject water. As membranes are semipermeable, salt ions are not permitted to flow across the membrane, and they will be carried within the reject stream [bib_ref] Frontiers of Membrane Desalination Processes for Brackish Water Treatment: A Review, Honarparvar [/bib_ref]. Nanocellulose based membranes could be employed for improving the reverse osmosis performance of membranes for desalination and considerably decreased the desalination cost of US$5.4 million in 2019 to US$0.5-1.2 million in Israel, Australia, China, Saudi Arabia, and United States [bib_ref] Innovative Nanostructured Membranes for Reverse Osmosis Water Desalination, Saleem [/bib_ref]. This section of the study features certain recent studies carried out for the preparation and testing of reverse osmosis membranes developed using cellulose nanocrystal used in the reverse osmosis application. In a study by [bib_ref] Performance of Reverse Osmosis Based Desalination Process Using Spiral Wound Membrane: Sensitivity..., Aladwani [/bib_ref] , the team demonstrated the incorporation of carbonylated cellulose nanocrystal (CNCCR) with active layer (AL-CNCCR) and supporting layer (SL-CNCCR) of TFC membrane. It was noted that higher permeate flux and salt rejection was obtained in RO testing. The AL-CNCCR membrane showed an excellent pure water flux, and the reported value was 62.17 ± 3.12 LMH. On the other hand, the team reported a conflicting result for the flux for the SL-CNCCR membrane as a decreased value of pure water flux for SL-CNCCR membrane was noted, i.e., 45.02 ± 3.37 LMH. The contrasting result for the pure water flux was due to the difference in position of entrenchment of CNCCR in the TFC membrane. Both modified membranes showed increased value for salt rejection for sulphate ions. The AL-CNCCR rejection ratio increased to approximately 105.2%, and SL-CNCCR membranes showed 107.5% rejection ratio for sulphate ion. The lowest salt rejection value was shown by an unmodified one which was approximately 94.8 ± 0.7% and 89.7 ± 1.5% for Na2SO4 and MgSO4, respectively. Both modified membranes showed good stability under the harsh conditions along with a good water flux and increased permeation. A similar phenomenon as in the case of sulphate ions was also observed in the rejection of sodium chloride; the increased value for chloride ion rejection of SL-CNCCR membrane (134.4%) compared with AL-CNCCR membrane (108.7%) compared with pristine membrane is again due to the implantation location of nanocellulose. This section of the study features certain recent studies carried out for the preparation and testing of reverse osmosis membranes developed using cellulose nanocrystal used in the reverse osmosis application. In a study by Liu and group [bib_ref] The Role of Carboxylated Cellulose Nanocrystals Placement in the Performance of Thin-Film..., Liu [/bib_ref] , the team demonstrated the incorporation of carbonylated cellulose nanocrystal (CNC CR ) with active layer (AL-CNC CR ) and supporting layer (SL-CNC CR ) of TFC membrane. It was noted that higher permeate flux and salt rejection was obtained in RO testing. The AL-CNC CR membrane showed an excellent pure water flux, and the reported value was 62.17 ± 3.12 LMH. On the other hand, the team reported a conflicting result for the flux for the SL-CNC CR membrane as a decreased value of pure water flux for SL-CNC CR membrane was noted, i.e., 45.02 ± 3.37 LMH. The contrasting result for the pure water flux was due to the difference in position of entrenchment of CNC CR in the TFC membrane. Both modified membranes showed increased value for salt rejection for sulphate ions. The AL-CNC CR rejection ratio increased to approximately 105.2%, and SL-CNC CR membranes showed 107.5% rejection ratio for sulphate ion. The lowest salt rejection value was shown by an unmodified one which was approximately 94.8 ± 0.7% and 89.7 ± 1.5% for Na 2 SO 4 and MgSO 4 , respectively. Both modified membranes showed good stability under the harsh conditions along with a good water flux and increased permeation. A similar phenomenon as in the case of sulphate ions was also observed in the rejection of sodium chloride; the increased value for chloride ion rejection of SL-CNC CR membrane (134.4%) compared with AL-CNC CR membrane (108.7%) compared with pristine membrane is again due to the implantation location of nanocellulose.
Asempour and group [bib_ref] Synthesis and Characterization of Novel Cellulose Nanocrystals-Based Thin Film Nanocomposite Membranes for..., Asempour [/bib_ref] developed a new thin-film nanocomposite membrane with cellulose nanocrystals 0.1% w/v loaded in the polyamide active layer. The group observed the membrane performance with the synthetic brackish water (SBW) and the outcome showed that the fabricated membrane demonstrated two times increased water flux from 30 to 63 ± 10 L/m 2 ·h, without compromising the salt rejection, which is 97.8%. To evaluate the fouling properties, the team used BSA (Bovine Serum Albumin) in the feed solution with a stable concentration of 300 ppm, and good fouling resistance properties were observed for the modified membrane.
Researchers have pointed out that the main disadvantages with the RO filtration membrane are the low rejection rates (metal ion) [bib_ref] Hybrid Membrane System Design and Operation, Singh [/bib_ref] and an increase in fouling tendency [bib_ref] Development of Antifouling Reverse Osmosis Membranes for Water Treatment: A Review, Kang [/bib_ref]. Due to these reasons, the efficiency of the RO membranes reduces, and the scientists are continuously working to overcome these limitations. Hence, by considering the chemical structure of nanocellulose, many researchers implanted nanocellulose to the membrane just to overcome these issues and project the suitable path for the problem. Park and group [bib_ref] Cellulose Nanocrystal-Assembled Reverse Osmosis Membranes with High Rejection Performance and Excellent Antifouling, Park [/bib_ref] developed a poly(acryloyl hydrazide)-grafted cellulose nanocrystal (CNC-PAH) membrane that showed better water permeance, salt rejection, and organic fouling resistance as compared to the commercial membranes. Moreover, it showed higher boron capturing ability and rejection due to the abundance of hydroxyl and amine groups in it. Membrane fouling is influenced by the membrane surface's structure (roughness), hydrophilicity, and charge density. The reason for fouling is the roughness of surface area and the surface area roughness increases, fouling will increase simultaneously as foulants would have a larger surface area to adhere to [bib_ref] Influence of Active Layer and Support Layer Surface Structures on Organic Fouling..., Lu [/bib_ref] and obstructs hydrodynamic rinsing-induced flux recovery by producing foulant valley blockage [bib_ref] Characteristics and Influencing Factors of Organic Fouling in Forward Osmosis Operation for..., Ly [/bib_ref]. Hence, the enhanced antifouling performance and better flux recovery of the poly-acryloyl hydrazide modified membrane can be explained by its smoother surface.
Obtaining cellulose nanocrystals from natural waste made CNC synthesis more economically sound, and one of the reported wastes, i.e., sawdust, is a waste generated by the forestry and timber industries, accounting for more than 40% of all wastes [bib_ref] Nigerian Wood Waste: A Potential Resource for Economic Development, Akhator [/bib_ref]. Most of the sawdust waste is not transformed into valuable goods and is usually disposed of in landfills. Sawdust-derived cellulose nanocrystals embedded in polyamide were created by Adeniyi and group [bib_ref] Incorporation of Cellulose Nanocrystals (CNC) Derived from Sawdust into Polyamide Thin-Film Composite..., Adeniyi [/bib_ref] , The team discovered high rejection rate values for sodium chloride (98.3 ± 0.8%) and calcium chloride (97.1 ± 0.5%) at specific concentrations of sodium chloride (1500 ppm) and calcium chloride (2500 ppm) solutions. Additionally, in this study, the water flux was greatly increased, more than 23%. The membrane was noted to be extremely hydrophilic and thermally stable.
In a study by Smith and group [bib_ref] Functionalized Cellulose Nanocrystal Nanocomposite Membranes with Controlled Interfacial Transport for Improved Reverse..., Smith [/bib_ref] , the team fabricated 2,2,6,6-Tetramethylpiperidine-1-oxyl (TEMPO)-oxidized cellulose nanocrystals (TOCNs) modified RO thin-film nanocomposite membrane, and the best improvement was obtained by using the monomer dispersion method with 0.5 wt.% TEMPO-oxidized cellulose nanocrystals loading. The increased water flux and salt rejection data provided an evidence of improvement in the membrane performance when compared to the unmodified PA layer of pristine membrane. Water flux for TOCN modified RO membrane was 260% as compared with the unmodified one, and the salt rejection increased to 98.98 ± 0.41% from 97.53 ± 0.31%, as shown in [fig_ref] Figure 7: Diagrammatic representation of the CNC-TFC-PDA membrane preparation [/fig_ref]. The most important parameter that defines the formation of nanochannels and pore diameter is the strong amalgamation between TEMPO-oxidized CNCs and PA layer. Due to the presence of electronegative atoms, molecules can easily form hydrogen bonds [bib_ref] Evaluation of Hydrogen Bond Networks in Cellulose Iβ and II Crystals Using..., Hayakawa [/bib_ref] , and these hydrogen bonds provide strength and availability of nanochannels, whose dimensions can be altered by providing different conditions. Within this alteration, a significant number of hydrogen bonding reduced the nanochannel dimension, increasing the mem-brane's salt rejection ability and resulting in high salt rejection at all loading levels examined in this study. Therefore, from the analysis of different studies in this section, it was noted that the CNC incorporated membranes can significantly improve the performance of membranes in the reverse osmosis process. The different modifications of the membranes such as the incorporation of pristine CNC, 2,2,6,6-Tetramethylpiperidine-1-oxyl (TEMPO)-oxidized cellulose nanocrystals (TOCNs), carbonylated cellulose nanocrystal, and poly(acryloyl hydrazide)-grafted cellulose nanocrystal (CNC-PAH) membrane can definitely improve the performance of membranes in the reverse osmosis process.
## Cellulose nanocrystals in pervaporation
The basic membrane system for effectively separating a liquid mixture is pervaporation (PV). Since traditional methods are ineffective in separating mixtures, PV is a new technology that can help in a very efficient separation of liquid mixture [bib_ref] Evaluation of Hydrogen Bond Networks in Cellulose Iβ and II Crystals Using..., Hayakawa [/bib_ref]. In both the fundamental and applied aspects of PV, significant progress and exciting breakthroughs have been made in recent decades. In PV applications, different studies have focused on the modification of nanocomposite membranes fabricated with nanocellulose [bib_ref] Nanocomposites Based on Polyimide Thermoplastics and Magnesium Silicate Nanoparticles with Montmorillonite Structure, Golubeva [/bib_ref]. Different researchers have developed nanocomposite membranes with nanocellulose that show outstanding results in terms of selectivity, salt rejection, permeability, [bib_ref] Van Der Bruggen, B. 110th Anniversary: Cellulose Nanocrystals as Organic Nanofillers for..., Prihatiningtyas [/bib_ref] and also desirable and effective surface morphology that enhance the membrane efficiency and its mechanical properties [bib_ref] Modification of Films of Heat-Resistant Polyimides by Adding Hydrosilicate and Carbon Nanoparticles..., Gofman [/bib_ref].
Ref. [bib_ref] Nanocomposites Based on Polyimide Thermoplastics and Magnesium Silicate Nanoparticles with Montmorillonite Structure, Golubeva [/bib_ref] , by utilizing a solution casting method, fabricated cellulose triacetate/cellulose nanocrystals (3%) (CTA/CNCs) nanocomposite pervaporation membranes. The membrane was modified to a sponge-like structure which offered the evidence for incorporation. After running the modified pervaporation membrane, it was observed that the membrane showed an increment in the water flux with the factor of 3 and it changed from 2.16 kg·m −2 ·h −1 to 5.76 kg·m −2 ·h −1 . The casting blade height was reduced from 200 μm to 100 μm, resulting in a flow of 11.68 kg·m −2 ·h −1 , and a NaCl rejection was maintained at 99.9%. For 12 h of separation, the CNCs 3%-CTA PV membrane with a casting blade height of 100 μm performed well. Water and NaCl was separated very well with this newly created PV membrane. Furthermore, when compared to a pristine membrane, it had a significantly higher water flux and, hence, might be used for desalination.
The influence of different solvents on the morphology, desalination, and flux performance of cellulose triacetate/cellulose nanocrystals (CTA/CNCs) nanocomposite PV membranes was studied by [bib_ref] A Comparative Study on the Mechanical and Barrier Characteristics of Polyimide Nanocomposite..., Yudin [/bib_ref]. The DMSO-based membranes produced a nanocomposite membrane with uniformly distributed CNCs on the membrane surface and a matrix with self-assembled structure. This membrane demonstrated a high-water flux of 11.67 kg Therefore, from the analysis of different studies in this section, it was noted that the CNC incorporated membranes can significantly improve the performance of membranes in the reverse osmosis process. The different modifications of the membranes such as the incorporation of pristine CNC, 2,2,6,6-Tetramethylpiperidine-1-oxyl (TEMPO)-oxidized cellulose nanocrystals (TOCNs), carbonylated cellulose nanocrystal, and poly(acryloyl hydrazide)-grafted cellulose nanocrystal (CNC-PAH) membrane can definitely improve the performance of membranes in the reverse osmosis process.
## Cellulose nanocrystals in pervaporation
The basic membrane system for effectively separating a liquid mixture is pervaporation (PV). Since traditional methods are ineffective in separating mixtures, PV is a new technology that can help in a very efficient separation of liquid mixture. In both the fundamental and applied aspects of PV, significant progress and exciting breakthroughs have been made in recent decades [bib_ref] Nanocomposites Based on Polyimide Thermoplastics and Magnesium Silicate Nanoparticles with Montmorillonite Structure, Golubeva [/bib_ref]. In PV applications, different studies have focused on the modification of nanocomposite membranes fabricated with nanocellulose [bib_ref] Van Der Bruggen, B. 110th Anniversary: Cellulose Nanocrystals as Organic Nanofillers for..., Prihatiningtyas [/bib_ref]. Different researchers have developed nanocomposite membranes with nanocellulose that show outstanding results in terms of selectivity, salt rejection, permeability, [bib_ref] Modification of Films of Heat-Resistant Polyimides by Adding Hydrosilicate and Carbon Nanoparticles..., Gofman [/bib_ref] and also desirable and effective surface morphology that enhance the membrane efficiency and its mechanical properties [bib_ref] A Comparative Study on the Mechanical and Barrier Characteristics of Polyimide Nanocomposite..., Yudin [/bib_ref].
Prihatiningtyas and group [bib_ref] Van Der Bruggen, B. 110th Anniversary: Cellulose Nanocrystals as Organic Nanofillers for..., Prihatiningtyas [/bib_ref] , by utilizing a solution casting method, fabricated cellulose triacetate/cellulose nanocrystals (3%) (CTA/CNCs) nanocomposite pervaporation membranes. The membrane was modified to a sponge-like structure which offered the evidence for incorporation. After running the modified pervaporation membrane, it was observed that the membrane showed an increment in the water flux with the factor of 3 and it changed from 2.16 kg·m −2 ·h −1 to 5.76 kg·m −2 ·h −1 . The casting blade height was reduced from 200 µm to 100 µm, resulting in a flow of 11.68 kg·m −2 ·h −1 , and a NaCl rejection was maintained at 99.9%. For 12 h of separation, the CNCs 3%-CTA PV membrane with a casting blade height of 100 µm performed well. Water and NaCl was separated very well with this newly created PV membrane. Furthermore, when compared to a pristine membrane, it had a significantly higher water flux and, hence, might be used for desalination.
The influence of different solvents on the morphology, desalination, and flux performance of cellulose triacetate/cellulose nanocrystals (CTA/CNCs) nanocomposite PV membranes was studied by [bib_ref] Effect of Solvent on the Morphology and Performance of Cellulose Triacetate Membrane/Cellulose..., Prihatiningtyas [/bib_ref]. The DMSO-based membranes produced a nanocom-posite membrane with uniformly distributed CNCs on the membrane surface and a matrix with self-assembled structure. This membrane demonstrated a high-water flux of 11.67 kg m −2 h −1 and 99.9% NaCl rejection when the feed solution of 30 g L −1 of NaCl was used. Furthermore, for a highly saline feed (up to 90 g L −1 of NaCl), the produced nanocomposite membranes demonstrated good PV desalination efficacy.
Kamtsikakis and group [bib_ref] Cellulose Nanocrystals as a Tunable Nanomaterial for Pervaporation Membranes with Asymmetric Transport..., Kamtsikakis [/bib_ref] fabricated a different CNC incorporated nanocomposite membrane and explored the ethanol-recovery pervaporation performance for the mixture sample of ethanol-water. The fabricated membrane with CNCs integration was based on a polystyrene-poly(butadiene)-poly(styrene) (SBS) matrix and was further decorated with oleic acid moieties (OLA-CNCs). The team studied the effect of alteration on pervaporation performance and, as a result, it was discovered that the relative increase in water flux (172%) was significantly higher than the relative increase in ethanol flux (51%) when equated to the separate fluxes through the unmodified SBS. The addition of CNCs to the SBS resulted in a poorer ethanol purity in the permeate (26%) compared to the unmodified membranes (38%).
In a study by Prihatiningtyas and group [bib_ref] Ultra-High Flux Alkali-Treated Cellulose Triacetate/Cellulose Nanocrystal Nanocomposite Membrane for Pervaporation Desalination, Prihatiningtyas [/bib_ref] , the team developed a nanocomposite membrane with varying quantities of feed solution and tested its permeability and salt rejection without compromising its selectivity. The team noticed an increase in the CTA/CNCs membrane's water flux as compared to an unmodified membrane that produced flux of (3.6 kg m −2 h −1 ). The team used a 90 g/L NaCl hypersaline solution and ran the experiment for 30 min. They discovered that the water flux increased dramatically, reaching 107.5 kg m −2 h −1 with more than 99.8% salt rejection. The water flux reduced to 58.5 kg m −2 h −1 when 200 g/L NaCl was used as the feed solution, and the alkali-treated membrane showed stable performance when evaluated for 90 g/L and 200 g/L NaCl for 12 h.
Rahimi and Kashkoul group [bib_ref] Thin Film Nanocomposite Nanofiltration Membrane Incorporated with Cellulose Nanocrystals with Superior Anti-Organic..., Rahimi-Kashkouli [/bib_ref] incorporated a functionalized cellulose nanocrystal within the TFN's polyamide thin layer. Thin-film nanocomposite membranes were designated as TFN0.0, TFN-1, TFN-5, and TFN0.1, respectively, based on CNC (wt./v.%) concentrations of 0.01, 0.05, and 0.1. The water contact angle for TFN incorporated with 0.1 wt./v.% of CNC was reduced from 69 - for TFN0.0 to 45 - for TFN incorporated with 0.1 wt./v.% of CNC. TFN0.05 was the best performing membrane, with a water flux of 21.34 L/m 2 h. This was roughly 140% greater than TFN0.0 in an identical operational condition. In comparison to a pristine membrane, CNC also decreased the amount of concentration polarization in TFN0.05, which resulted in higher protein fouling resistance and a lower water flux drop.
Hence, from the examination of various studies in this section, it was noted that the CNC incorporated membranes can considerably improve the performance of membranes in the pervaporation process. [fig_ref] Table 2: The overview of the studies carried out using CNC-based membranes for the... [/fig_ref] presents the overview of the studies carried out using CNC-based membranes for the desalination and water treatment application. The different modifications of the membranes such as the incorporation of pristine CNC, functionalized cellulose nanocrystal, and cellulose triacetate/cellulose nanocrystals (CTA/CNCs) nanocomposite can enhance the performance of membranes in the pervaporation process.
## Cellulose nanofibrils (cnf): structure, properties, and desalination application
Cellulosic nanofibers have been intensively studied among the different nanostructures formed from cellulose due to their outstanding mechanical strength, flexibility, ability to functionalize and integrate, biodegradability, chirality, thermostability, and low thermal expansion. Young's modulus and tensile strength of CNFs are comparable to Kevlar fibers and are five times that of mild steel [bib_ref] Elastic Modulus of the Crystalline Regions of Cellulose Polymorphs, Nishino [/bib_ref]. In the axial direction, the CNFs have the same thermal expansion as quartz [bib_ref] Production of Green Nanomaterials, Tsuzuki [/bib_ref]. CNFs have been proposed as a potential matrix for a wide range of applications, including medical application, energy storage, adsorption, wastewater treatment, biomedical materials, nanofillers, filtration media antimicrobial activity, and membrane modification for desalination.
CNFs can form dense membrane with the help of hydrogen bonding during the water loss process of CNF solution. CNFs can interact with adjacent nanofibers hydroxyl group, and plenty of OH groups become interlocked into colloidal form in aq. solution and generate a CNF self-assembled membrane. The CNFs' integrated membranes have good strength qualities and lower permeability in comparison with high and low-density polyethylene membrane [bib_ref] Strength and Barrier Properties of MFC Films, Syverud [/bib_ref] of similar thickness, and hence these membranes can be used for water desalination application.
CNF's biocompatibility is a key feature that makes them ideal for use in filtration systems and membranes for water purification and desalination. CNFs are excellent biocompatible materials which makes them suitable for desalination application. Certain precautions are taken for the usage of materials in water filtration systems, such as it should not release and deteriorate harmful elements. The CNF structure's stability is enhanced by its crystalline character. When combined with other membrane materials having low crystallinity, CNF imparts stiffness and high strength to the nanocomposite membrane, thereby increasing water flux and antifouling in addition to the intrinsic properties [bib_ref] Mechanical Properties of Cellulose Nanofiber (CNF) Reinforced Polylactic Acid (PLA) Prepared by..., Jonoobi [/bib_ref] [bib_ref] The Effect of Cellulose Nanofibers on the Crystallinity and Nanostructure of Poly(Lactic..., Frone [/bib_ref] [bib_ref] Cellulose-Based Nanofibers Processing Techniques and Methods Based on Bottom-Up Approach-A Review, Kramar [/bib_ref] [bib_ref] Nanocomposite Desalination Membranes Made of Aromatic Polyamide with Cellulose Nanofibers: Synthesis, Performance,..., Cruz-Silva [/bib_ref]. In addition, CNFs are one of the best and cheapest options as they exhibit high-permeability, ultrahigh water flux, better hydrophilicity, and good salt rejection [bib_ref] Sub-10 Nm Wide Cellulose Nanofibers for Ultrathin Nanoporous Membranes with High Organic..., Zhang [/bib_ref] when used in membranes. The hydroxyl groups present on the CNF surface develop powerful hydrogen bonding, which after drying can convert them to extremely degradation resistant. The testing on the biodegradability (EN 14046) of cellulose nanofibril films and papers with cellulose nanofibrils demonstrated that the entire products of cellulose nanofibrils examined were noted to be biodegradable as per the necessities set in the standard EN 13432 [bib_ref] Biodegradability and Compostability of Nanofibrillar Cellulose-Based Products, Vikman [/bib_ref]. Some outstanding performance has been observed by integrating cellulose nanofibers in the membrane modification, which is discussed in the following section. There are numerous CNF-related experiments reported throughout the years, particularly with respect to the development of CNF incorporated membranes for filtration application. The application of CNFs in membrane production can be divided into numerous categories. The most widely described methods involve coating or infusing functional CNF onto/into a porous scaffold, while others involve including CNF into the composite suspension or using CNF as the major element for inorganics attachment.
## Cellulose nanofibers in ultrafiltration application
Ultrafiltration is a dedicated membrane-based filtration technology that improves the pressure-mediated suspension of pathogens as well as solid waste from waste solution. The product water obtained subsequent to the ultrafiltration process is extremely pure and free from any pathogen waste. The utilization of cellulose as well as its derivatives to make membranes for the application in ultrafiltration is well recognized at the industrial level. The usage of CNFs has been studied greater as compared to the CNCs to develop films for UF membranes. This section of the study features certain recent studies performed for the preparation and testing of ultrafiltration membranes developed using cellulose nanocrystal used in the ultrafiltration application.
In a study carried out by Hassan and group [bib_ref] Water Purification Ultrafiltration Membranes Using Nanofibers from Unbleached and Bleached Rice Straw, Hassan [/bib_ref] , utilized high-lignin unbleached neutral sulfite's cellulose nanofibers for the application in ultrafiltration. The team made a comparison of the developed membrane with the bleached UF membrane on the basis of water flux. During the 2 h of experiment, the team obtained the following data: water flux for bleached membrane was around~27 L/h/m 2 /MPa, while for unbleached one reached about~53 L/h/m 2 /MPa. Hence, the water flux efficiency was 96% higher in the case of the unbleached membrane, but when the experiment was run for 1 h, the improvement in flux as a result of using unbleached nanofibers instead of bleached ones was~120%. The water flux value for the bleached membrane was 4.4 L/h/m 2 /MPa after 30 min of passing the BSA solution (the first BSA cycle), and for the unbleached membrane it was 36.5 L/h/m 2 /MPa. Soyekwo and group [bib_ref] Cellulose Nanofiber Intermediary to Fabricate Highly-Permeable Ultrathin Nanofiltration Membranes for Fast Water..., Soyekwo [/bib_ref] opted for the approach of surface modification of ultrafine cellulose nanofiber (UCN) membranes via interfacial polymerization to make ultrathin polymeric nanofiltration membranes. In this study, the membrane was made up of an ultrathin selective layer interlaced with a cellulose nanofiber matrix. The developed membraned showed good water flux, salt rejection, and resistance to extreme conditions. The results for mono and di valent salt (rejection %) at pH 5.5 were as follows: MgCl 2 -89.7%, MgSO 4 -65.3%, NaCl-43.6%, and Na 2 SO 4 (39.1%). The team concluded that the membranes showed higher rejections towards divalent cations (Mg 2+ ) at pH less than 7 and higher rejections toward divalent anions (SO [bib_ref] Removal of selenite from artificial wastewater with high salinity by activated sludge..., Zhang [/bib_ref] 2− ) at pH greater than 7. The reason behind displaying lower rejections toward monovalent ions is that pH has an effect on salt flux, and in contrast to rejections, salt fluxes decrease as pH rises. Furthermore, the fouling ratios for reversible (F r ) and irreversible (F ir ) fouling were 28.1% and 16.7%, respectively. After cleaning, the recovered pure water flux was restored to 27.3 L m −2 h −1 bar −1 , equating to 83.3% flux recovery ratio with bovine serum albumin (BSA) mixed salt solution. Nanofiber-based UF membranes based on the thin-film nanofibrous composite format having a nanocomposite barrier layer consisting of cross-linked poly(ethylene glycol) matrix as well as ultra-fine CNFs were examined in a study by Wang and group [bib_ref] Metal-Coordinated Nanofiltration Membranes Constructed on Metal Ions Blended Support toward Enhanced Dye/Salt..., Fang [/bib_ref]. It was noted that the membrane developed was extremely hydrophilic and possessed superior antifouling characteristics, which were proved by long-term and short-term fouling experiments using bovine serum albumin (BSA) solution (1 g/L).
Therefore, from the analysis of different research works in this section, it was confirmed that the CNF-embedded membranes can considerably improve the performance of membranes in the ultrafiltration process. The different modified membranes such as the thin-film nanofibrous composite format having a nanocomposite barrier layer consisting of cross-linked poly(ethylene glycol) matrix as well as ultra-fine CNFs, surface modified ultrafine cellulose nanofiber (UCN) membranes, and high-lignin unbleached neutral sulfite cellulose nanofiber-based membranes can enhance the performance of membranes in the ultrafiltration process.
## Cellulose nanofibers in nanofiltration
Nanofiltration (NF) membranes are of greater interest in the desalination of brackish water and the purification of drinking water. Improvement in the separation performance as well as permeation still remain to be a great challenge in the latest innovative NF membranes. The nanofiltration process happening at lower pressure is in huge demand as it could save energy, and it could be highly conceivable in the case that the membrane could generate excellent permeation flux. Mohammed and group [bib_ref] Effect of Oxygen Plasma Treatment on the Nanofiltration Performance of Reduced Graphene..., Mohammed [/bib_ref] fabricated reduced graphene oxide/cellulose nanofibers to test the nanocellulose fabricated membrane's performance on various parameters of nanofiltration. The thick stacking of reduced graphene oxide frequently resulted in limited water flux, and that was the reason nanocellulose has been entrenched with the reduced graphene oxide. The obtained results are as follows: pure water permeance of 37.2 L m −2 h −1 bar −1 was obtained by maintaining the rejection above 90% for Acid Fuchsin (1. Moreover, it continues to be a great challenge for developing a superior-performance NF membrane having good antifouling properties as well as higher perm-selectivity. In a study reported by [fig_ref] Figure 2: Structural representation of cellulose polymer[58] [/fig_ref] , the team developed an advanced TFN membrane by in situ incorporation of zwitterionic nanocellulose at the time of interfacial polymerization process. The resulting membrane showed an enhanced water permeance and improved enhanced divalent salt rejection because of the enhanced surface hydrophilicity as well as higher surface area and the super hydrophilic interfacial nanochannels. Under the synergy of zwitterions and nanocellulose, therefore, the membrane developed demonstrated higher perm-selectivity together with beand antifouling properties, which would have promising potentials in separation fields.
Hence, from the examination of different research works in this section, it was proved that the CNF-embedded membranes can noticeably improve the nanofiltration performance of membranes. The different modified membranes such as those of zwitterionic nanocellulose modified membranes and reduced graphene oxide/cellulose nanofiber incorporated membranes could improve the performance of membranes in the ultrafiltration process.
## Cellulose nanofibers in reverse osmosis
As it was noted in the previous section, CNFs are noted to have promising application in ultrafiltration and nanofiltration based desalination technologies. From the analysis of different studies, it can be noted that the CNFs are also have having promising application in the reverse osmosis process. Ref. [bib_ref] Fabrication of CuO Nanoparticles Immobilized Nanofiltration Composite Membrane for Dye/Salt Fractionation: Performance..., Waheed [/bib_ref] demonstrated the application of cellu- [fig_ref] Figure 8: A conventional reverse osmosis system [/fig_ref]. Novel composite nanofiltration membrane using sulfated cellulose nanofibril (SCNF) for nanofiltration [bib_ref] Effect of Oxygen Plasma Treatment on the Nanofiltration Performance of Reduced Graphene..., Mohammed [/bib_ref].
Moreover, it continues to be a great challenge for developing a superior-performance NF membrane having good antifouling properties as well as higher perm-selectivity. In a study reported by [fig_ref] Figure 2: Structural representation of cellulose polymer[58] [/fig_ref] , the team developed an advanced TFN membrane by in situ incorporation of zwitterionic nanocellulose at the time of interfacial polymerization process. The resulting membrane showed an enhanced water permeance and improved enhanced divalent salt rejection because of the enhanced surface hydrophilicity as well as higher surface area and the super hydrophilic interfacial nanochannels. Under the synergy of zwitterions and nanocellulose, therefore, the membrane developed demonstrated higher perm-selectivity together with beand antifouling properties, which would have promising potentials in separation fields.
Hence, from the examination of different research works in this section, it was proved that the CNF-embedded membranes can noticeably improve the nanofiltration performance of membranes. The different modified membranes such as those of zwitterionic nanocellulose modified membranes and reduced graphene oxide/cellulose nanofiber incorporated membranes could improve the performance of membranes in the ultrafiltration process.
## Cellulose nanofibers in reverse osmosis
As it was noted in the previous section, CNFs are noted to have promising application in ultrafiltration and nanofiltration based desalination technologies. From the analysis of different studies, it can be noted that the CNFs are also have having promising application in the reverse osmosis process. Liu and group [bib_ref] Enhancement of Desalination Performance of Thin-Film Nanocomposite Membrane by Cellulose Nanofibers, Liu [/bib_ref] demonstrated the application of cellulose nanofibers as a cheap and biodegradable material to enhance the efficiency of membranes for RO application. The team fabricated the advanced membrane by interfacial polymerization with 2,2,6,6-tetramethylpiperidine-1-oxyl radical (TEMPO)-oxidized CNFs by integrating it into the polyamide layer of a thin film composite. The optimized loading of CNFs into the TFC membrane enhanced the permeance of the membrane by more than 50% (29.8 L m −2 h −1 at 1.5 MPa) while maintaining high NaCl rejection of 96.2%. The chlorine stability tests revealed that the fabricated membranes outperformed existing TFC membranes in terms of chlorine resistance.
Another modification was shown by Cruz-Silva and group [bib_ref] Nanocomposite Desalination Membranes Made of Aromatic Polyamide with Cellulose Nanofibers: Synthesis, Performance,..., Cruz-Silva [/bib_ref] , in which the team fabricated and examined the desalination performance of a RO membrane by coupling CNF with an aromatic PA layer. The team found that the membrane became stronger and stiffer as a result of the strong bonding between PA matrix and CNF, and due to this the matrix mobility decreased. Higher water permeability is dependent on the membrane's hydrophilicity; hence, the fabricated membrane showed an increase in water permeability, indicating the improved hydrophilicity. The CNF-PA membrane showed great hydrophilicity, which is an important parameter for the nanocomposite membrane's higher water permeability. The antifouling performance and chlorine resistance of the membranes were also improved. The following results were reported subsequent to the study: the CNF membrane showed a diffusion coefficient of 0.37 × 10 −5 cm 2 s −1 , which was 40% higher than the diffusion coefficient, 0.21 × 10 −5 cm 2 s −1 , of the pristine PA membrane. The antifouling performance with bovine serum albumin (BSA), water permeability, chlorine resistance, and salt rejection rate of the membranes were also investigated, and it was discovered that after 140 h, the water permeation rate decreased slightly (by 10% from its initial value), whereas the pristine PA membrane showed reduction in permeation rate of water by 40% from its initial value after 24 h. With active chlorine test at 20 ppm, it was found that due to the rapid deterioration of the PA membrane by chlorine, the pristine PA membrane rapidly dropped its salt rejection rate [bib_ref] Robust Water Desalination Membranes against Degradation Using High Loads of Carbon Nanotubes, Ortiz-Medina [/bib_ref] and dramatically increased its water absorption rate after 100 h was also reported previously [bib_ref] Surface Modification of Commercial Aromatic Polyamide Reverse Osmosis Membranes by Crosslinking Treatments, Wei [/bib_ref]. Moreover, CNF-PA membrane began to disintegrate at the same time with a lower pace compared to the PA membrane.
Hassan and group [bib_ref] Metallo-Terpyridine-Modified Cellulose Nanofiber Membranes for Papermaking Wastewater Purification, Hassan [/bib_ref] developed Cu-terpyridine-modified oxidized cellulose nanofibers membrane (OXCNF-Cu-Tpy) and compared its performance with the TEMPO-oxidized cellulose nanofibrous membrane (OXCNF). Depending on the pressure applied during filtration, chemically modified OXCNF with Cu-Tpy groups boosted membrane's pure water flux by 52 to 94% (0.5 and 1 MPa, respectively). Despite the fact that both OXCNF and OXCNF-Cu-Tpy membranes were highly efficient at removing suspended particles with the size range of (0.05-0.22 micron) from wastewater effluent, OXCNF-Cu-Tpy membrane showed a 30% higher flux rate than OXCNF membranes. Ref. [bib_ref] Fabrication of Cellulose Nanofiber-Based Ultrafiltration Membranes by Spray Coating Approach, Wang [/bib_ref] developed a nanofiber composite membrane with the help of corn stalks, by using it as the raw material. The team observed that the membrane was extremely successful at removing contaminants (diameter 1-100 nm) from waste water, and the retention rate of chromium (Cr(VI)) at pH = 11 was 80%, indicating that it can be utilized for short-term wastewater treatment and household water purification purposes.
The effect of cellulose nanofibers' incorporation on cellulose acetate membrane's morphology, water flux, and filtration performance was investigated by. Due to variations in the remixing process rate during membrane development, an increase in CNF content resulted in a sponge-like shape. Porosity and pure water flux (40 L·m −2 ·h −1 for pure CA to 880 L·m −2 h for CA/CNF) improved as CNF content increased, and retentate turbidity was 11% greater after the separation procedure.
Hence, from the examination of various studies in this section, it was noted that the CNF incorporated membranes can considerably improve the performance of membranes in the reverse osmosis process. [fig_ref] Table 3: The overview of the studies carried out using CNF-based membranes for the... [/fig_ref] presents the overview of the studies carried out using CNF-based membranes for the desalination and water treatment application. The different modified membranes, such as cellulose nanofibers, incorporated cellulose acetate membrane, 2,2,6,6-tetramethylpiperidine-1-oxyl radical (TEMPO)-oxidized CNF incorporated TFC membrane, Cu-terpyridine-modified cellulose nanofibers membrane, and TEMPO-oxidized cellulose nanofibrous membrane, and CNF modified TFC membranes can enhance the performance of membranes in the reverse osmosis process. There are also a lot of application for CNF-based membranes along with wastewater treatment.
## Bacterial nano cellulose (bnc): structure, property, wastewater, and desalination application
Bacterial nanocellulose can be synthesized using a two-step process called polymerization and crystallization. Glucose residues polymerize in the bacterial cytoplasm to form β-1,4 glucan linear chains, which are secreted extracellularly. The generated chains are crystallized into microfibrils, which are subsequently consolidated to form a very pure three-dimensional (3D) porous network of entangled nanoribbons with a width of 20-60 nm [bib_ref] Bacterial Cellulose Nanocomposites: An All-Nano Type of Material, Torres [/bib_ref]. This is because of its exceptional chemical and physical stabilities, greener synthetic approach, low manufacturing costs, hydrophilic nature, and good degradability. Due to the aforementioned reasons, bacterial nanocellulose is gaining increasing global interest, and hence scientists are focusing more studies on the BNC-based materials [bib_ref] A Natural in Situ Fabrication Method of Functional Bacterial Cellulose Using a..., Gao [/bib_ref] [bib_ref] Advances in Biomedical and Pharmaceutical Applications of Functional Bacterial Cellulose-Based Nanocomposites, Ullah [/bib_ref]. In many applications, the ultra-fine structure (as shown in of BNC outperforms plant cellulose in terms of stiffness [bib_ref] Bacterial Cellulose Nanocomposites: An All-Nano Type of Material, Torres [/bib_ref] ], water interactive capacity, rate of polymerization, and active surface area [bib_ref] Recent Approaches and Future Prospects of Bacterial Cellulose-Based Electroconductive Materials, Chen [/bib_ref].
The abundance of hydroxyl groups makes it easier for BNCs to functionalize or incorporate it with other reinforcing chemicals to give additional physical/chemical properties to the membranes [bib_ref] A Natural in Situ Fabrication Method of Functional Bacterial Cellulose Using a..., Gao [/bib_ref] , such as antibacterial properties [bib_ref] Potential of Biocellulose Carrier Impregnated with Essential Oils to, Junka [/bib_ref] [bib_ref] Nanocellulose Films with Multiple Functional Nanoparticles in Confined Spatial Distribution, Roig-Sanchez [/bib_ref]. As a result, BNC's application areas are constantly expanding, including bioprocessing, biomedical, and pharmaceutical applications, food industry [bib_ref] Multifunctional Adsorbent Based on Metal-Organic Framework Modified Bacterial Cellulose/Chitosan Composite Aerogel for..., Li [/bib_ref] , wastewater treatment [bib_ref] Multifunctional Adsorbent Based on Metal-Organic Framework Modified Bacterial Cellulose/Chitosan Composite Aerogel for..., Li [/bib_ref] , and many more. Additionally, when compared to cellulose derivatives from plants, the BNC's ability to be modified makes this material far superior. The bacteria strains are usually incubated in a nutrient-dense aqueous media and develop bacterial cellulose on the top layer (interface with air) as an exopolysaccharide. Here, the β-D-glucopyranose units are primarily present at the time of the development of cellulose molecules inside the bacteria cell [bib_ref] Nanocelluloses: Sources, Pretreatment, Isolations, Modification, and Its Application as the Drug Carriers, Lunardi [/bib_ref]. The elementary fibril is discharged across the cellulose surface pores, which were additionally arranged as well as crystallized into microfibrils with twisting ribbons shape succeeded by pellicle development. nm [bib_ref] Bacterial Cellulose Nanocomposites: An All-Nano Type of Material, Torres [/bib_ref]. This is because of its exceptional chemical and physical stabilities, greener synthetic approach, low manufacturing costs, hydrophilic nature, and good degradability. Due to the aforementioned reasons, bacterial nanocellulose is gaining increasing global interest, and hence scientists are focusing more studies on the BNC-based materials [bib_ref] A Natural in Situ Fabrication Method of Functional Bacterial Cellulose Using a..., Gao [/bib_ref] [bib_ref] Advances in Biomedical and Pharmaceutical Applications of Functional Bacterial Cellulose-Based Nanocomposites, Ullah [/bib_ref]. In many applications, the ultra-fine structure (as shown inof BNC outperforms plant cellulose in terms of stiffness, water interactive capacity, rate of polymerization, and active surface area [bib_ref] Lignocellulosic Biomass-Derived Nanocellulose Crystals as Fillers in Membranes for Water and Wastewater..., Sadare [/bib_ref] [bib_ref] Bacterial Cellulose Nanocomposites: An All-Nano Type of Material, Torres [/bib_ref]. The abundance of hydroxyl groups makes it easier for BNCs to functionalize or incorporate it with other reinforcing chemicals to give additional physical/chemical properties to the membranes [bib_ref] A Natural in Situ Fabrication Method of Functional Bacterial Cellulose Using a..., Gao [/bib_ref] , such as antibacterial properties [bib_ref] Potential of Biocellulose Carrier Impregnated with Essential Oils to, Junka [/bib_ref] [bib_ref] Nanocellulose Films with Multiple Functional Nanoparticles in Confined Spatial Distribution, Roig-Sanchez [/bib_ref]. As a result, BNC's application areas are constantly expanding, including bioprocessing, biomedical, and pharmaceutical applications, food industry [bib_ref] Multifunctional Adsorbent Based on Metal-Organic Framework Modified Bacterial Cellulose/Chitosan Composite Aerogel for..., Li [/bib_ref] , wastewater treatment [bib_ref] Multifunctional Adsorbent Based on Metal-Organic Framework Modified Bacterial Cellulose/Chitosan Composite Aerogel for..., Li [/bib_ref] , and many more. Additionally, when compared to cellulose derivatives from plants, the BNC's ability to be modified makes this material far superior. The bacteria strains are usually incubated in a nutrient-dense aqueous media and develop bacterial cellulose on the top layer (interface with air) as an exopolysaccharide. Here, the β-D-glucopyranose units are primarily present at the time of the development of cellulose molecules inside the bacteria cell [bib_ref] Nanocelluloses: Sources, Pretreatment, Isolations, Modification, and Its Application as the Drug Carriers, Lunardi [/bib_ref]. The elementary fibril is discharged across the cellulose surface pores, which were additionally arranged as well as crystallized into microfibrils with twisting ribbons shape succeeded by pellicle development. Accounting for all the properties and application, the majority of researchers in the field of water desalination consider BNCs as a promising material and are continuously involved in the fabrication of many diverse membranes for enhancing the desalination performance of membranes [bib_ref] Fabrication of Cellulose Nanofiber-Based Ultrafiltration Membranes by Spray Coating Approach, Wang [/bib_ref].is a cross-section of a pristine BNC membrane. BNCs are widely utilized in a variety of applications, but there has been little research on their use in membrane fabrication for desalination applications; therefore, we only mention a few bacterial nanocellulose based membrane applications in the following section. Accounting for all the properties and application, the majority of researchers in the field of water desalination consider BNCs as a promising material and are continuously involved in the fabrication of many diverse membranes for enhancing the desalination performance of membranes.is a cross-section of a pristine BNC membrane. BNCs are widely utilized in a variety of applications, but there has been little research on their use in membrane fabrication for desalination applications; therefore, we only mention a few bacterial nanocellulose based membrane applications in the following section.
Accounting for all the properties and application, the majority of researchers in the field of water desalination consider BNCs as a promising material and are continuously involved in the fabrication of many diverse membranes for enhancing the desalination performance of membranes [bib_ref] Fabrication of Cellulose Nanofiber-Based Ultrafiltration Membranes by Spray Coating Approach, Wang [/bib_ref].is a cross-section of a pristine BNC membrane. BNCs are widely utilized in a variety of applications, but there has been little research on their use in membrane fabrication for desalination applications; therefore, we only mention a few bacterial nanocellulose based membrane applications in the following section.
## Membrane distillation
Membrane distillation is a technology which separates out the water in vapor form from saline water with higher rejection factors. Only vapor molecules are allowed to pass through a porous hydrophobic membrane in MD, and this process will be thermally driven membrane-based separation process. The vapor pressure difference caused by the temperature gradient across the membrane surface is the driving force in the MD. A cross-section of a pristine BNC membrane. Reprinted with permission from Ref. [bib_ref] Photothermally Active Reduced Graphene Oxide/Bacterial Nanocellulose Composites as Biofouling-Resistant Ultrafiltration Membranes, Jiang [/bib_ref]. Copyright 2019 American Chemical Society.
## Membrane distillation
Membrane distillation is a technology which separates out the water in vapor form from saline water with higher rejection factors. Only vapor molecules are allowed to pass through a porous hydrophobic membrane in MD, and this process will be thermally driven membrane-based separation process. The vapor pressure difference caused by the temperature gradient across the membrane surface is the driving force in the MD process. Separation efficiency strongly depends on the volatility of the separating component and also the structure of the porous membrane. The bacterial nanocellulose has been proven to have good application in the development of membranes for the MD process. This section of the study presents certain recent studies performed for the preparation and testing of MD membranes developed using BNC in the MD application.
In a research work performed by Leitch and group [bib_ref] Bacterial Nanocellulose Aerogel Membranes: Novel High-Porosity Materials for Membrane Distillation, Leitch [/bib_ref] , the team fabricated a model high porosity membrane and used it to investigate the performance on distillation. In this study, the researchers used unsupported bacterial nanocellulose aerogel membranes in direct contact membrane distillation (DCMD). The authors claimed that the membrane had lower bulk thermal conductivity, a larger porosity, greater conductivity, and thinner fibers as compared to any other MD material previously reported. The developed membrane material demonstrated much higher intrinsic membrane permeability and thermal efficiency than symmetric PVDF phase inversion membranes with lower porosity. This was confirmed after performing modeling and tests. Further improvement in MD flux was noted from the development of macroporous fibrous membranes with aerogel-like porosity and heat conductivity in thinner-film shapes.
Furthermore, Wu and group [bib_ref] A Thermally Engineered Polydopamine and Bacterial Nanocellulose Bilayer Membrane for Photothermal Membrane..., Wu [/bib_ref] fabricated an advanced membrane by making a bilateral structure composed of polydopamine (PDA) particles and bacterial nanocellulose together for photothermal membrane distillation with bactericidal capability. The developed membraneshowed a high solar energy-to-collected water efficiency of 68%, permeate flux of 1.0 kg m −2 h −1 under sun irradiation, an improved membrane porosity, and high salt rejection (>99.9%). The team also reported reduced conductive heat transfer, which increased the thermal efficiency of the membrane, and contributed to a stable hydrophobicity. Furthermore, the membrane exhibited good interfacial photothermal disinfection property to destroy germs when exposed to sunlight, allowing for easy cleaning, and thereby extending the membrane's lifespan.
Hence, it was noted that the BNC incorporated membranes can considerably improve the performance of membranes in the membrane distillation process. The bacterial nanocellulose aerogel membranes and PDA/BNC membranes can enhance the performance of membranes in the MD process, and there are also a lot of application for BNC-based membranes along with wastewater treatment application.
bilateral structure composed of polydopamine (PDA) particles and bacterial nanocellulose together for photothermal membrane distillation with bactericidal capability. The developed membraneshowed a high solar energy-to-collected water efficiency of 68%, permeate flux of 1.0 kg m −2 h −1 under sun irradiation, an improved membrane porosity, and high salt rejection (>99.9%). The team also reported reduced conductive heat transfer, which increased the thermal efficiency of the membrane, and contributed to a stable hydrophobicity. Furthermore, the membrane exhibited good interfacial photothermal disinfection property to destroy germs when exposed to sunlight, allowing for easy cleaning, and thereby extending the membrane's lifespan.. Schematic representation of solar-driven photothermal membrane distillation (MD) system. Reprinted with permission from Ref. [bib_ref] A Thermally Engineered Polydopamine and Bacterial Nanocellulose Bilayer Membrane for Photothermal Membrane..., Wu [/bib_ref]. Copyright 2020 Elsevier Ltd. All rights reserved.
## Ultrafiltration
The majority characteristics of the UF process are the same as membrane-based filtration, as the UF technique is aligned with the usage of semipermeable membrane. This section of the study presents certain recent studies carried out for the preparation and testing of advanced membranes developed using BNC in the ultrafiltration application. Jiang and group [bib_ref] Photothermally Active Reduced Graphene Oxide/Bacterial Nanocellulose Composites as Biofouling-Resistant Ultrafiltration Membranes, Jiang [/bib_ref] used reduced graphene oxide and bacterial nanocellulose to fabricate an anti-biofouling UF membrane, as shown in. The team tested the membrane stability under various conditions, and it was found that rGO/BNC membrane exhibited outstanding stability under environmentally relevant pH conditions, dynamic mechanical agitation/sonication, and even at high pressure. Over a 5 h-long flux test, the team compared the water flux of fabricated membrane with the commercial one, and it was noted that there was a two-fold increment in the water flux (52.6 ± 2.5 L/m 2 h) for rGO/BNC membrane as compared with the water flux of commercial membrane (21.6 ± 0.8 L/m 2 h).
Organic dyes are the most commonly seen contaminants present in industrial wastewater, and these materials are the most difficult to remove in any waste water treatment facility. Xu and group [bib_ref] Catalytically Active Bacterial Nanocellulose-Based Ultrafiltration Membrane, Xu [/bib_ref] fabricated an ultrafiltration membrane for the removal of organic dye based on bacterial nanocellulose with graphene oxide (GO) (while it was growing, essentially trapping GO in the membrane to make it stable and durable) and palladium (Pd) nanoparticles. The fabricated membrane shows outstanding performance by removing methylene orange (99.3%). The membrane was also tested for mixture of 4-nitrophenol and rhodamine 6G. The membranes showed a stable flux (33.1 L m −2 h −1 ) under 58 psi, over longer duration. Gholami and group [bib_ref] A Robust and Scalable Polydopamine/Bacterial Nanocellulose Hybrid Membrane for Efficient Wastewater Treatment, Gholami Derami [/bib_ref] showed the membrane's efficiency in heavy metal (Pd and Cd) and organic dye removal by using in situ entrenchment of polydopamine and bacterial nanocellulose particles. The membrane was further tested at several pH levels (4-7) and found to be quite stable.
Many researchers also fabricated BNC-based membranes for treating wastewater to make it proficient in desalination application [bib_ref] A Robust and Scalable Polydopamine/Bacterial Nanocellulose Hybrid Membrane for Efficient Wastewater Treatment, Gholami Derami [/bib_ref] fabricated a membrane composed of bacterial nanocellulose and polydopamine materials. This membrane was tested for the removal of organic dyes (which are organic contaminants) and metal ions (lead and cadmium) present in the wastewater. It was found that the membrane was removing all the contaminants effectively.
(Pd) nanoparticles. The fabricated membrane shows outstanding performance by removing methylene orange (99.3%). The membrane was also tested for mixture of 4-nitrophenol and rhodamine 6G. The membranes showed a stable flux (33.1 L m −2 h −1 ) under 58 psi, over longer duration. [bib_ref] A Robust and Scalable Polydopamine/Bacterial Nanocellulose Hybrid Membrane for Efficient Wastewater Treatment, Gholami Derami [/bib_ref] showed the membrane's efficiency in heavy metal (Pd and Cd) and organic dye removal by using in situ entrenchment of polydopamine and bacterial nanocellulose particles. The membrane was further tested at several pH levels (4-7) and found to be quite stable.. Schematic representation of UF membrane synthesis by using BNC fibers. Reprinted/adapted with permission from [bib_ref] Photothermally Active Reduced Graphene Oxide/Bacterial Nanocellulose Composites as Biofouling-Resistant Ultrafiltration Membranes, Jiang [/bib_ref].
Many researchers also fabricated BNC-based membranes for treating wastewater to make it proficient in desalination application. [bib_ref] A Robust and Scalable Polydopamine/Bacterial Nanocellulose Hybrid Membrane for Efficient Wastewater Treatment, Gholami Derami [/bib_ref] fabricated a membrane composed of bacterial nanocellulose and polydopamine materials. This membrane was tested for the removal of organic dyes (which are organic contaminants) and metal ions (lead and cadmium) present in the wastewater. It was found that the membrane was removing all the contaminants effectively.
Therefore, from the analysis of different studies, it was noted that the BNC-added membranes can considerably improve the performance of membranes in the ultrafiltration process. The Pd/GO/BNC membranes (organic dye and heavy metal removal), bacterial nanocellulose/polydopamine membranes (removal of organic contaminants and. Schematic representation of UF membrane synthesis by using BNC fibers. Reprinted with permission from Ref. [bib_ref] Photothermally Active Reduced Graphene Oxide/Bacterial Nanocellulose Composites as Biofouling-Resistant Ultrafiltration Membranes, Jiang [/bib_ref]. Copyright 2019 American Chemical Society. Therefore, from the analysis of different studies, it was noted that the BNC-added membranes can considerably improve the performance of membranes in the ultrafiltration process. The Pd/GO/BNC membranes (organic dye and heavy metal removal), bacterial nanocellulose/polydopamine membranes (removal of organic contaminants and metal ions), and reduced graphene oxide/bacterial nanocellulose membranes can improve the performance of membranes in the ultrafiltration process, and there is also a huge application possibility for BNC-based membranes in desalination and wastewater treatment application, which need to be further explored. The novel design as well as the in situ inclusion of the membranes prepared in these studies presented a perception for obtaining advanced, ecofriendly, and fouling resistant membranes for water treatment application.
## Other application
For membrane-based water filtration application, Sijabat and group [bib_ref] Synthesis and Characterization of Bacterial Nanocellulose from Banana Peel for Water Filtration..., Sijabat [/bib_ref] developed nano cellulose derived from bacteria using Nangka banana peel media. The results confirmed the development of a potential membrane for an effective membrane filtration. The BNC colloidal solution was stable enough to be used in the manufacture of water filter catalytic membrane composites, according to the results obtained from electrophoretic light scattering (ELS) potential zeta absolute value (−11.39 mV). Moreover, a different study by the team fabricated an acetobacter xylinum-derived membrane for the application of PEG rejection. The team utilized β-chitin and deacetylated chitin sulfonates for solute separation. After performing the test, the team reported that all the membranes were performing exceptionally with respect to PEG50000 rejection, as each membrane contributed >85% rejection.
To avoid membrane fouling, pore blockage, and biofilm formation during the desalination process, a variety of chemically pretreatment processes are used these days, including chlorination, filtering, flocculation/sedimentation, and antiscalant dosage acidification [bib_ref] Water Treatment and Desalination, Abdel-Fatah [/bib_ref]. The difficulty that has put these approaches on the back foot is their lack of reusability and chemical unitability. To solve this problem, researchers are looking into ecologically friendly materials, one of which is without a doubt nanocellulose.
Through a simple paper production procedure, Mautner and group [bib_ref] Bacterial Nanocellulose Papers with High Porosity for Optimized Permeance and Rejection of..., Mautner [/bib_ref] developed a bacterial nanocellulose based porous filter paper dispersed in low surface tension liquids and water for the removal of nanosized particles. When compared to traditional nano papers made from aqueous dispersions, nano papers made from organic solution had 40 times higher permeance due to a reduced paper density and high porosity. Despite their increased porosity, nano sheets feature pore diameters of 15-20 nm, which are similar to BC nano papers generated from aqueous dispersions, allowing for the size-exclusion removal of pollutants the size of viruses at high permeance.
Thus, overall, it was noted that the bacterial nanocellulose can be used effectively for the desalination as well as wastewater treatment application. [fig_ref] Table 4: The overview of the studies carried out using BNC-based membranes for the... [/fig_ref] presents the overview of the studies carried out using BNC-based membranes for the desalination and water treatment application.
## Future perspectives
Nanocellulose has a lot of potential in the treatment of wastewater as well as in desalination. Nanocellulose is a one-of-a-kind solution in the field of wastewater treatment due to its mechanical strength, chemical stability, and resistance to environmental harshness. Researchers from all around the world have developed advanced membranes with promising results, but there has been little study on the application of BNCs and CNFs in the fabrication of desalination membrane. The majority of the study focuses on incorporating nanocellulose into the PA layer, but there is high potential to develop nanocellulose-based membranes by embedding this material in the support layer of the thin-film composite membranes, as well. The hydroxyl group of these nanocelluloses makes them appropriate candidates in the effort of membrane improvement. By considering this attribute of nanocellulose, there is a lot of scope for chemical alteration with naturally sound materials such as quantum dots and organic oxides.
Several recent studies have confirmed that the chemical modification of nanocellulosebased membranes is extremely significant for improving the membrane efficiency and the interactions between the contaminants and the membranes. The nanocellulose-based membranes with certain functionality can remove the required contaminants/ions present in water within a shorter time, thereby generating the fresh water efficiently. Therefore, the modification of membrane is anticipated to perform an important role in the forthcoming membrane development for desalination and water treatment applications.
It is recommended to have more studies based on developing membranes using the bacterial nanocellulose, cellulose nanocrystals, and cellulose nanofibers. Researchers should also consider cellulose extraction methods that are significantly more environmentally friendly, as well as the use of waste in this process.
# Conclusions
As a sustainable and environmental-friendly material, the nanocellulose-based membranes are anticipated to extend its use to more applications shortly. The different forms of nanocellulose (CNF, CNC, or BNC) are able to be converted into the membrane itself or integrated into membranes as additives.
The results obtained from different studies on nanocellulose crystal inclusion, whether unmodified or modified, on the membrane have improved the desalination efficiency of the membranes. Surface modification can affect membrane performance, and the surface of CNC has many side hydroxyl groups, which allows for chemical modification, and this can affect membrane performance directly. The CNCs, when incorporated into the TFC membranes, showed improvement in the nanofiltration efficiency of the membrane, and these membranes are noted to have a big impact on the future development of highperformance NF membranes for water treatment. Moreover, it was noted that CNCembedded TFC membranes showed higher permeate flux and salt rejection in RO testing. The CNC-based membranes also showed promising application in pervaporation. After running the CNC-modified pervaporation membrane, it was observed that the membrane showed an increment in the water flux with the factor of 3.
Moreover, the CNF-modified membranes are one of the best and cheapest options, as they exhibit high permeability, ultra-high water flux, better hydrophilicity, and good salt rejection. The surface modification of ultrafine cellulose nanofiber membranes by means of interfacial polymerization can develop ultrathin polymeric nanofiltration membranes of higher efficiency. The membrane becomes stronger and stiffer because of the strong bonding between PA matrix and CNF, and due to this, the matrix mobility decreases. CNF-PA membranes show great hydrophilicity, which is an important parameter for the nanocomposite membrane's higher water permeability. The antifouling performance and chlorine resistance of the membranes will also be improved. Furthermore, considering the properties of BNCs, majority researchers in the field of water desalination consider BNCs as a promising material in desalination and water treatment, and are continuously involved in the fabrication of many diverse membranes for enhancing the desalination performance of membranes.
[fig] Figure 1: of publications (related to "Nanocellulose")No. of publications (related to "Nanocellulose for Membrane Application") The number of publications (by year [/fig]
[fig] Figure 2: Structural representation of cellulose polymer[58]. [/fig]
[fig] Figure 3: Properties, structure, and synthesis of different types of nanocelluloses[62]. [/fig]
[fig] Figure 5: Schematic representation of crystalline region and amorphous region in cellulose and the formation of cellulose nanocrystals. Reprinted/adapted with permission from[77]. [/fig]
[fig] Figure 4: Schematic representation of crystalline region and amorphous region in cellulose and the formation of cellulose nanocrystals. Reprinted with permission from Ref.[78]. Copyright 2014 American Chemical Society. [/fig]
[fig] Figure 7: Diagrammatic representation of the CNC-TFC-PDA membrane preparation. Reprinted/adapted with permission from[91]. [/fig]
[fig] Membranes 2022 ,: 12, x FOR PEER REVIEW 11 of 34 [/fig]
[fig] Figure 8: A conventional reverse osmosis system. [/fig]
[fig] Figure 6: A conventional reverse osmosis system. [/fig]
[fig] Author: Contributions: Conceptualization, S.J.Z. and H.S.; methodology, A.S.; software, A.S.; validation, S.J.Z. and H.S.; formal analysis, H.S.; investigation, H.S.; resources, A.S.; data curation, A.S.; writing-original draft preparation, A.S.; writing-review and editing, H.S.; visualization, H.S.; supervision, S.J.Z.; project administration, H.S.; funding acquisition, S.J.Z. All authors have read and agreed to the published version of the manuscript. Funding: This research was funded by projects NPRP13S-0205-200263 and QUST-1-CAM-2022-441. Institutional Review Board Statement: Not applicable. Informed Consent Statement: Not applicable. Data Availability Statement: Not applicable. [/fig]
[table] Table 2: The overview of the studies carried out using CNC-based membranes for the desalination and water treatment application. [/table]
[table] Table 3: The overview of the studies carried out using CNF-based membranes for the desalination and water treatment application. [/table]
[table] Table 4: The overview of the studies carried out using BNC-based membranes for the desalination and water treatment application. [/table]
|
The spatial separation of processing and transport functions to the interior and periphery of the Golgi stack
It is unclear how the two principal functions of the Golgi complex, processing and transport, are spatially organized. Studying such spatial organization by optical imaging is challenging, partially due to the dense packing of stochastically oriented Golgi stacks. Using superresolution microscopy and markers such as Giantin, we developed a method to identify en face and side views of individual nocodazole-induced Golgi mini-stacks. Our imaging uncovered that Golgi enzymes preferentially localize to the cisternal interior, appearing as a central disk or inner-ring, whereas components of the trafficking machinery reside at the periphery of the stack, including the cisternal rim. Interestingly, conventional secretory cargos appeared at the cisternal interior during their intra-Golgi trafficking and transiently localized to the cisternal rim before exiting the Golgi. In contrast, bulky cargos were found only at the rim. Our study therefore directly demonstrates the spatial separation of processing and transport functions within the Golgi complex.
# Introduction
The Golgi complex is one of the most important processing and sorting stations along the secretory and endocytic pathway [bib_ref] Models for Golgi traffic: a critical assessment, Glick [/bib_ref] [bib_ref] Architecture of the mammalian Golgi, Klumperman [/bib_ref] [bib_ref] From endosomes to the trans-Golgi network, Lu [/bib_ref]. In mammalian cells, it consists of a network of laterally linked Golgi stacks. As the structural unit, a Golgi stack comprises 4-7 flattened cisternae and can be divided into cis, medial and trans-regions. The trans-Golgi region further develops into the trans-Golgi network (TGN). It is known that the cis-Golgi receives secretory cargos from the endoplasmic reticulum (ER) exit site (ERES) and ER Golgi intermediate compartment (ERGIC), while the trans-Golgi and TGN exchange materials with endosomes and the plasma membrane (PM). At the moment, we still don't understand how the Golgi becomes organized and works at the molecular and cellular level [bib_ref] Models for Golgi traffic: a critical assessment, Glick [/bib_ref]. One of the challenges in studying the Golgi is to spatiotemporally resolve residents and transiting cargos among individual cisternae of Golgi stacks, a task currently beyond the capabilities of even super-resolution and electron microscopy (EM).
It has been hypothesized that the two principal functions of the Golgi, processing and transport, are spatially organized for optimal efficiency [bib_ref] Transport through the Golgi apparatus by rapid partitioning within a two-phase membrane..., Patterson [/bib_ref]. However, such molecular organization across the Golgi stack has not been directly demonstrated. Previously, by utilizing nocodazole-induced Golgi mini-stacks, we developed a conventional microscopy based super-resolution method, named GLIM (Golgi localization by imaging center of fluorescence mass), to quantitatively map the axial position or localization quotient (LQ) of a Golgi protein with nanometer accuracy [bib_ref] Quantitative Localization of a Golgi Protein by Imaging Its Center of Fluorescence..., Tie [/bib_ref] [bib_ref] A novel imaging method for quantitative Golgi localization reveals differential intra-Golgi trafficking..., Tie [/bib_ref]. To understand the molecular organization of the Golgi ministack, the lateral localization, which refers to the distribution of a protein within Golgi cisternal membrane sheets, is also required. Although more structural details of the Golgi can be resolved with the The double-punctum appearances of Giantin, Golgin84 and GPP130 indicate side views of Golgi mini-stacks. In each merge, the intensity profile is generated along a thick line, represented by a dotted box, with the direction indicated by the arrow (the same scheme is used throughout this study). The dotted box schematically marked the start, end and width of the line. The direction arrow roughly follows the cis-to-trans Golgi axis using the cismost (GM130 in this case) and trans-most markers in each panel. Dotted pink lines connecting double-punctum are almost orthogonal to the cis-totrans Golgi axis. The intensity plot is normalized and color-coded as the corresponding merge image. (D-F) En face averaged images of Giantin, fluorescence protein (FP)-Golgin84 and GPP130-GFP. The corresponding radial mean intensity profile is shown at the right with distance from the center of fluorescence mass (normalized to the radius of Giantin) as the x-axis and radial mean intensity (normalized) as the y-axis. Both GFP and mCherry-tagged Golgin84 images were used for FP-Golgin84. n, the number of averaged Golgi mini-stacks. (G) GPP130 mostly localizes to the cisternal rim (arrows) of the native Golgi by EM. NRK cells transiently expressing GPP130-APEX2-GFP were subjected to APEX2-catalyzed reaction followed by EM. Note that cells were not subjected to nocodazole treatment. The EM thin section image displays the side view of a Golgi mini-stack. The electron density indicates the localization of GPP130 (arrows). (H) The histogram showing the distribution of diameters of Giantin-rings. (I, J) Giantin N and C-terminus colocalize at the cisternal rim. In (I), cells were co-stained using Giantin antibodies raised against its N and C-terminus. In (J), Giantin N-terminus was stained by an antibody and its C-terminus was revealed by exogenously expressed mScarlet-Giantin-C129. In the en face view, dotted arrow represents the line used to generate the line intensity profile (width = 1 pixel), while in the side view, the dotted box that is in the direction of the arrow and parallel to the Golgi cisterna represents the line for intensity profile. (K) The interior localization of MGAT2 within the Giantin-ring. Line intensity profiles of the en face and side views are acquired as those in (I) and (C) respectively. Scale bar, 500 nm. [fig_ref] Figure 1: Identifying the en face and side view of the Golgi mini-stack [/fig_ref] continued on next page
Identifying en face and side views of Golgi mini-stacks By assessing the super-resolution staining patterns of Giantin, GPP130 or Golgin84, we can conveniently identify en face and side view oriented Golgi mini-stacks, images of which should appear as a ring and double-punctum, respectively. It was discovered that some Golgi residents, such as MGAT2, localized to the interior of Giantin-rings [fig_ref] Figure 1: Identifying the en face and side view of the Golgi mini-stack [/fig_ref]. Consistent with this interpretation, side views of MGAT2 appeared as a short bar connecting the Giantin double-punctum [fig_ref] Figure 1: Identifying the en face and side view of the Golgi mini-stack [/fig_ref]. Under the EM, MGAT2-APEX2-GFP preferentially localized to the cisternal interior (next section). Therefore, there are at least two types of lateral localizations: rim and interior, as represented by Giantin and MGAT2.
Golgi trafficking components mainly localize to the periphery of a Golgi mini-stack
We systematically examined the lateral localization of Golgi residents using their en face and side views. Two types of residents were studied in this work -components of trafficking machinery, including those involved in the structure and organization of the Golgi, and enzymes involved in the post-translational modifications, particularly glycosyltransferases. Due to the lack of reagents to detect endogenous proteins, many residents were detected by the overexpression of their tagged fusions. Caution must be taken in the interpretation of our data as it has been documented that overexpression can change both the axial and lateral localization of Golgi residents [bib_ref] Dynamic transport of SNARE proteins in the Golgi apparatus, Cosson [/bib_ref]. We discovered that the lateral localization of trafficking machinery components shares common features according to their LQs.
## Eres, ergic and cis-golgi proteins (lq <0)
COPII coat subunits, including Sec13 and Sec23a, COPI coat subunits, including b and g-COP, KDEL receptor, GS27, ERGIC53, Arf4 and Arf5, displayed lumps or puncta around Giantin-rings in en face views and at one side of Giantin-double-punctum in side views , suggesting that they mainly localize to the rim of their corresponding cisternae and are mostly absent from the cisternal interior. Arf1, whose LQ is 0.75, is an exception here. Although its en face view demonstrated that it is in the cisternal interior, side view images uncovered that there were two pools: a cis/medial and a trans-Golgi/TGN pool, with a much reduced presence in between [fig_ref] Figure 3: continued on next page respectively [/fig_ref]. This observation is consistent with the notion that Arf1 functions in the cis-Golgi and TGN for the assembly of the COPI and clathrin coat, respectively [bib_ref] The small G proteins of the Arf family and their regulators, Gillingham [/bib_ref]. [bib_ref] Golgi structure in three dimensions: functional insights from the normal rat kidney..., Ladinsky [/bib_ref] and the role of clathrin coat in transporting these cargos to the endolysosome [bib_ref] The Di-leucine motif of vesicle-associated membrane protein 4 is required for its..., Peden [/bib_ref] [bib_ref] Recycling of furin from the plasma membrane. Functional importance of the cytoplasmic..., Teuchert [/bib_ref]. Our data are also consistent with the notion that the TGN comprises domains of distinct molecular compositions [bib_ref] GCC185 plays independent roles in Golgi structure maintenance and AP-1-mediated vesicle tethering, Brown [/bib_ref] [bib_ref] Mammalian GRIP domain proteins differ in their membrane binding properties and are..., Derby [/bib_ref].
In summary, our extensive super-resolution imaging data suggest that Golgi trafficking components mainly localize to the entire cis-cisternae, rim of medial and trans-cisternae and punctate or tubular profiles at non-stacked regions, which include the ERES, ERGIC and TGN.
## Glycosylation enzymes reside at the interior of a golgi stack
We studied components of Golgi post-translational modification machinerySupplementary file 1), including a GDP-fucose transporter, , and more than a dozen enzymes involved in N-glycosylation (Man1B1, MGAT1, ManII, MGAT2, GalT, SialT and MGAT4B), O-glycosylation (GALNT1, GALNT2 and POMGNT1), poly-N-acetyllactosamine synthesis (b4GalT3), glycosaminoglycan synthesis (b3GalT6 and b4GalT7) and sulfation (TPST1 and 2). Interestingly, their LQs were found to be in the range from 0.23 to 1.0 (Table 1), suggesting that Golgi enzymes mainly localize to the medial and trans-region of the Golgi, but not to the cis-Golgi and TGN. This observation is consistent with previous EM studies. For example, in plant cells, polysaccharides were mainly detected in the medial and trans-Golgi cisternae [bib_ref] Functional compartmentation of the Golgi apparatus of plant cells : immunocytochemical analysis..., Zhang [/bib_ref]. Similarly, in mammalian cells, the N-glycan modifying enzymes ManI, ManII and MGAT1 have been mapped to the medial and trans-region of the Golgi stack [bib_ref] Attachment of terminal N-acetylglucosamine to asparagine-linked oligosaccharides occurs in central cisternae of..., Dunphy [/bib_ref] [bib_ref] Overlapping distribution of two glycosyltransferases in the Golgi apparatus of HeLa cells, Nilsson [/bib_ref] [bib_ref] Mapping the distribution of Golgi enzymes involved in the construction of complex..., Rabouille [/bib_ref] [bib_ref] Cell type-dependent variations in the subcellular distribution of alpha-mannosidase I and II, Velasco [/bib_ref]. However, in contrast to our quantitative results, previous EM work has assigned GalT [bib_ref] Overlapping distribution of two glycosyltransferases in the Golgi apparatus of HeLa cells, Nilsson [/bib_ref] [bib_ref] Mapping the distribution of Golgi enzymes involved in the construction of complex..., Rabouille [/bib_ref] [bib_ref] Immunocytochemical localization of galactosyltransferase in HeLa cells: codistribution with thiamine pyrophosphatase in..., Roth [/bib_ref] and SialT [bib_ref] Mapping the distribution of Golgi enzymes involved in the construction of complex..., Rabouille [/bib_ref] [bib_ref] Demonstration of an extensive trans-tubular network continuous with the Golgi apparatus stack..., Roth [/bib_ref] to the TGN in addition to the trans-Golgi. Sub-Golgi localizations are not always consistently reported, which is likely due to two reasons. First, the cis, medial, trans-region and TGN are not rigorously defined and the assignment of Golgi regions can be subjective. Second, it has been documented that the sub-Golgi localization of enzymes can be cell-type dependent [bib_ref] Cell type-dependent variations in the subcellular distribution of alpha-mannosidase I and II, Velasco [/bib_ref].
In contrast to trafficking components, our Golgi enzymes and SLC35C1 localized within Giantinrings as a central disk in en face views [fig_ref] Figure 1: Identifying the en face and side view of the Golgi mini-stack [/fig_ref] ; , a significant amount of these proteins are expected to reside in the same cisternae. The lateral distribution pattern of MGAT2 and MGAT4B suggests that they should mainly localize to the interior of cisternae as a central disk and inner-ring, respectively, within the Giantinrim in the same cisternae [fig_ref] Figure 4: Golgi enzymes primarily localize to the interior of medial and trans-Golgi cisternae [/fig_ref]. Enzymes, such as b4GalT3 and ST6Gal1, which have similar LQs, were observed to localize to shared and distinct domains within Giantin-rings [fig_ref] Figure 4: Golgi enzymes primarily localize to the interior of medial and trans-Golgi cisternae [/fig_ref].
To substantiate our light microscopic data, we examined the localization of MGAT2-APEX2-GFP in the native Golgi by EM. 93% (n = 58) of Golgi stacks showed an enrichment of MGAT2 in the cisternal interior [fig_ref] Figure 4: Golgi enzymes primarily localize to the interior of medial and trans-Golgi cisternae [/fig_ref] ; [fig_ref] Figure 4: Golgi enzymes primarily localize to the interior of medial and trans-Golgi cisternae [/fig_ref] -figure supplement 3A-C), which is in contrast to the staining pattern observed for GPP130 [fig_ref] Figure 1: Identifying the en face and side view of the Golgi mini-stack [/fig_ref]. Noticeably, APEX2-generated electron density was also found in vesicles and budding profiles at the rim (arrow heads in [fig_ref] Figure 4: Golgi enzymes primarily localize to the interior of medial and trans-Golgi cisternae [/fig_ref]. However, we did not find MGAT2-AcGFP1 [fig_ref] Figure 1: Identifying the en face and side view of the Golgi mini-stack [/fig_ref] or MGAT2-APEX2-GFP (Figure 4-figure supplement 1U) signal outside Giantin-rings by fluorescence imaging of Golgi mini-stacks. Although the identity and destiny of these vesicles are currently unknown, our observations suggest that Golgi enzymes might be depleted from the rim either by retrieval to the interior or by sorting into membrane carriers. Together, our data demonstrate that Golgi enzymes mainly localize to the interior of medial and trans-cisternae as a concentric disk or inner-ring, while trafficking machinery components exhibit rim localization.
## A quantitative molecular map of the golgi mini-stack
To quantitatively describe the overall lateral distribution of Golgi proteins, we assume that a Golgi protein has a radial symmetry localization around the Golgi axis as a concentric disk or ring. The normalized radius of the ring or disk can be measured using the radial mean intensity profile of en face averaged images (see Materials and methods). A plot of the normalized radius versus LQ quantitatively summarizes our morphological observations of ring and disk distribution of various Golgi residents [fig_ref] Figure 4: Golgi enzymes primarily localize to the interior of medial and trans-Golgi cisternae [/fig_ref]. While medial and trans-Golgi trafficking machinery components are at the cisternal rim, Golgi enzymes all localize to the interior with Man1B1, ManII, MGAT4B and TPST2 appearing as concentric inner-rings and the rest as central disks. Interestingly, it also reveals that ciscisternae have smaller diameters than medial ones, consistent with many EM thin-section or tomographic 3D images [bib_ref] In situ structural analysis of Golgi intracisternal protein arrays, Engel [/bib_ref] [bib_ref] Nanoscale architecture of endoplasmic reticulum export sites and of Golgi membranes as..., Staehelin [/bib_ref] , though the biological significance of which remains to be further investigated.
## Imaging the organization of the native golgi complex
Having studied in detail the organization of Golgi mini-stacks, we attempted to resolve the organization of the native Golgi complex by the super-resolution microscopy. Giantin and Golgi enzymes were used to mark the rim and interior of stacked cisternae, respectively. In the less dense region, Giantin and GPP130 staining appeared as distinctive ring-or loop-patterns, with b4GalT3 and GM130 filling the interior [fig_ref] Figure 4: Golgi enzymes primarily localize to the interior of medial and trans-Golgi cisternae [/fig_ref] , similar to the nocodazole-induced mini-stack. b4GalT3 and GM130 positive membrane sheets likely correspond to stacked Golgi cisternae. In most cases, Giantin and GPP130 positive curvy lines did not correspond to side views or cross-sections of Golgi stacks. Instead, they corresponded to the rim of cisternae in oblique or en face views (arrows in [fig_ref] Figure 4: Golgi enzymes primarily localize to the interior of medial and trans-Golgi cisternae [/fig_ref]. In the more densely packed region, cisternae appeared to pile on top of each other, a configuration that requires much higher z-axis resolution to be resolved. Nonetheless, we
[formula] GM130 (n=147) GRASP65-GFP (n=23) GRASP55-GFP (n=41) GFP-Rab1a (n=34) ManII-SBP-GFP (n=40) GFP-ACBD3 (n=44) FP-Golgin84 (n=65) Man1B1-Myc (n=29) E3GalT6-Myc (n=26) MGAT4B- AcGFP1 (n=63) E4GalT7-Myc (n=35) MGAT2-Myc (n=44) Giantin (n=220) TPST2-GFP (n=21) POMGNT1-Myc (n=27) MGAT1-Myc (n=32) E4GalT3-Myc (n=30) ST6Gal1- Myc (n=38) TPST1-GFP (n=32) GPP130-GFP (n=28) SLC35C1-Myc (n=38) GALNT2 (n=23) GFP-GCC185 (n=20) GALNT1 (n=11) Rab6-GFP (n=24) Sec34-Myc (n=24) Arf1-GFP (n=39) GS15 (n=28) GS28( [/formula]
## The lateral localization of secretory cargos during their intra-golgi trafficking
To study the lateral localization of secretory cargos during their intra-Golgi trafficking, the retention using selective hooks (RUSH) system was adopted to synchronously release secretory cargos [bib_ref] Synchronization of secretory protein traffic in populations of cells, Boncompain [/bib_ref]. The RUSH reporter CD59, a GPI-anchored protein, was first detected in the interior of cis-Golgi cisternae after 10 min of chase . During its transition through the Golgi mini-stack, as evidenced in its LQ versus time plot , CD59 remained in the interior , although its total intensity in Golgi mini-stacks initially increased and subsequently decreased due to the export toward the PM. At the later stage of the chase, there were CD59 positive puncta and tubular profiles outside Giantin-rings, which were likely Golgi-derived exocytic transport carriers , arrows in 60 min). Similarly, in live-cell super-resolution imaging, RUSH reporter mCherry-GPI started to appear in the interior of the Golgin84-ring 6 min after chase; it remained there for >30 min before disappearing due to post-Golgi exocytic trafficking ; -video 1). Transmembrane RUSH reporters, E-cadherin, VSVG and CD8a-Furin, and a soluble secretory reporter, signal peptide fused GFP, followed similar lateral localization pattern during their intra-Golgi trafficking [fig_ref] Figure 1: Identifying the en face and side view of the Golgi mini-stack [/fig_ref]. Collectively, our data demonstrated that conventional secretory cargos partition to the interior of the cisternae during their Golgi transition.
The secretory cargo wave does not seem to grossly affect the interior distribution of Golgi enzymes, as evidenced by ST6Gal1 . By image quantification, >85% of ST6Gal1-moxGFP was found to remain in the interior during the Golgi transition of synchronized VSVG, although a small fluctuation (<4%) was noticed [fig_ref] Figure 2A -: D [/fig_ref]. Our finding is different from a previous EM study, in which the shift of Golgi enzymes from the rim to the interior was observed under a traffic wave [bib_ref] Golgi enzymes are enriched in perforated zones of golgi cisternae but are..., Kweon [/bib_ref]. A more systematic investigation is required to resolve this discrepancy.
## Bulky size prevents the localization of secretory cargos at the cisternal interior
Based on EM data, Rothman lab previously proposed that large secretory protein aggregates are segregated to the cisternal rim. To investigate if bulky cargos partition to the rim, we imaged the RUSH reporter GFP-collagenX, a soluble secretory protein that tends to form oligomers [bib_ref] Macromolecular organization of chicken type X collagen in vitro, Kwan [/bib_ref] , by Airyscan super-resolution microscopy. We observed that Golgitransiting GFP-collagenX appeared either diffuse or punctate . Assuming that Golgi- [fig_ref] Figure 4: Golgi enzymes primarily localize to the interior of medial and trans-Golgi cisternae [/fig_ref] continued stack images. (H) b4GalT3 and ST6Gal1 can localize to shared (arrows) and distinct domains within the cisternal interior. (I) MGAT2 localizes to the cisternal interior of the native Golgi by EM. NRK cells transiently expressing MGAT2-APEX2-GFP were subjected to APEX2-catalyzed reaction followed by EM. Note that cells were not subjected to nocodazole treatment. The thin section EM image displays the side view of a Golgi stack. MGAT2-APEX2 positive cisternal interior and budding profiles are indicated by arrows and arrow heads, respectively. (J) A quantitative molecular map of the Golgi mini-stack. The normalized radius of a Golgi protein is plotted against its corresponding LQ. Red open and closed circle denote ring and disk lateral localization pattern, respectively. n, the number of Golgi mini-stacks used to calculate normalized radius. (K,L) Identifying the rim and interior of native Golgi cisternae. Cells were not treated with nocodazole. In (K), the cisternal rim (arrows) and interior are labeled by Giantin and b4GalT3, respectively. In (L), Giantin and GPP130 positive curvy lines (arrows) represent cisternal rim and do not correspond to side views or cross sections of Golgi stacks. The boxed region in each image is enlarged in the upper right corner. Scale bars represent 500 nm unless specified otherwise. DOI: https://doi.org/10.7554/eLife.41301.011
The following figure supplements are available for figure 4: localized GFP-collagenX puncta were single multimeric aggregates, using GFP-tagged nucleoporin Nup133 as an in vivo GFP fluorescence standard, we estimated that Golgi-transiting GFP-collagenX puncta had 190 ± 20 copies (n = 77) [fig_ref] Figure 3: continued on next page respectively [/fig_ref]. The diffused collagenX is probably in a much lower oligomeric state. Throughout its intra-Golgi trafficking, collagenX, either in punctate or diffuse appearance, was excluded from the interior of Giantin-rings, where co-expressed mCherry-GPI clearly localized . Instead, it always resided at the rim, either colocalizing with Giantin or surrounding Giantin-rings as discrete puncta. At later stages, the puncta outside Giantin-rings were probably exocytic carriers targeting to the PM.
We also tested soluble and transmembrane secretory cargos, FM4-moxGFP and GFP-FM4-CD8a, whose aggregation states can be controlled by the small molecule -D/D solubilizer. These two cargos are similar to the ones used previously. NRK cells expressing either cargo were treated with D/D solubilizer at 20˚C for 2 hr to accumulate and arrest the de-aggregated chimera at the Golgi mini-stack. At 20˚C, cells were subsequently subjected to 2 hr of incubation in the presence or absence of D/D solubilizer to either de-aggregate or aggregate the cargo respectively (nocodazole was in the system throughout the procedure). Our previous work has established that secretory cargos such as VSVG are mostly arrested at the medial Golgi under 20˚C treatment [bib_ref] A novel imaging method for quantitative Golgi localization reveals differential intra-Golgi trafficking..., Tie [/bib_ref]. In some experiments, 10 min warm up at 37˚C was conducted before imaging. Using this protocol, the re-aggregated GFP-FM4-CD8a and FM4-moxGFP Golgi puncta upon D/D washout were estimated to have 830 ± 30 (n = 184) and 660 ± 50 (n = 127) copies, respectively [fig_ref] Figure 3: continued on next page respectively [/fig_ref]. We observed that, when in the de-aggregated state, both soluble and membrane FM4-chimeras localized to the interior of Giantin-rings ; -figure supplement 3E). Intriguingly, once aggregated, they partitioned to the rim as discrete puncta. Therefore, our light microscopic data indicated that large cargos preferentially partition to the cisternal rim, possibly due to their bulky sizes, while conventional or small cargos tend to locate to the interior.
# Discussion
It poses a great challenge to investigate the structure and organization of the Golgi complex by the light microscopy. We established a method to identify the cisternal rim and interior by taking advantage of rim-localized Golgi markers. In addition to quantitative axial localization using the LQ [bib_ref] A novel imaging method for quantitative Golgi localization reveals differential intra-Golgi trafficking..., Tie [/bib_ref] , we further showed the advantage of nocodazole-induced Golgi mini-stacks in elucidating the molecular organization of the Golgi complex. We analyzed dozens of Golgi residents representing diverse families of proteins for their lateral localizations. The distribution of enzymes is continued transition through the Golgi mini-stack. Cells transiently co-expressing RUSH reporter, SBP-mCherry-GPI, and GFP-Golgin84 were chased in biotin and imaged live under Airyscan super-resolution microscopy. The boxed region in the upper image, which was acquired before the chase, is selected to show the time series. Arrow heads indicate the interior localization. See also [fig_ref] Figure 1: Identifying the en face and side view of the Golgi mini-stack [/fig_ref] The partition of collagenX and mCherry-GPI to the cisternal rim and interior respectively during their intra-Golgi transport. Cells transiently co-expressing RUSH cargos, SBP-GFP-collagenX and SBP-mCherry-GPI were chased as in (A). Arrows and arrow heads indicate the cisternal rim and interior localization respectively. The intra-Golgi transport of collagenX and mCherry-GPI was demonstrated by LQ vs time plots measured from the same experiments in (E) and (F). Error bar, mean ± SEM. n, the number of Golgi mini-stacks used for the calculation. (G) GFP-FM4-CD8a partitions to the cisternal rim upon aggregation. NRK cells transiently expressing GFP-FM4-CD8a were subjected to a combination of D/D solubilizer treatment and wash out at either 20˚C or 37˚C, as indicated. First set of images is the negative control showing that aggregated GFP-FM4-CD8a was not exported from the ER. Aggregated GFP-FM4-CD8a partitioned to the rim (arrows), while non-aggregated form was still interior-localized (arrow heads). Scale bars represent 500 nm unless specified otherwise. DOI: https://doi.org/10.7554/eLife.41301.015
The following video and figure supplements are available for figure 5: restricted to the interior of the medial and trans-cisternae. In contrast, trafficking machinery components appear to complement Golgi enzymes by residing at the rim of medial and trans-cisternae, entire cis-Golgi cisternae and trans-Golgi/TGN. Previous EM studies on lateral localizations of trafficking machinery components, including COPI [bib_ref] Bidirectional transport by distinct populations of COPI-coated vesicles, Orci [/bib_ref] , giantin [bib_ref] The golgin tether giantin regulates the secretory pathway by controlling stack organization..., Koreishi [/bib_ref] , KDEL receptor [bib_ref] A resident Golgi protein is excluded from peri-Golgi vesicles in NRK cells, Cosson [/bib_ref] [bib_ref] Bidirectional transport by distinct populations of COPI-coated vesicles, Orci [/bib_ref] , GS27 [bib_ref] Dynamic transport of SNARE proteins in the Golgi apparatus, Cosson [/bib_ref] and GS15 [bib_ref] Dynamic transport of SNARE proteins in the Golgi apparatus, Cosson [/bib_ref] , Golgi enzymes, including Man1B1 [bib_ref] The dynamics of engineered resident proteins in the mammalian Golgi complex relies..., Rizzo [/bib_ref] , ManII [bib_ref] A resident Golgi protein is excluded from peri-Golgi vesicles in NRK cells, Cosson [/bib_ref] [bib_ref] Dynamic transport of SNARE proteins in the Golgi apparatus, Cosson [/bib_ref] [bib_ref] Exclusion of golgi residents from transport vesicles budding from Golgi cisternae in..., Orci [/bib_ref] , MGAT1 and GalT [bib_ref] Dynamic transport of SNARE proteins in the Golgi apparatus, Cosson [/bib_ref] , and Golgi-transiting cargos including VSVG [bib_ref] Small cargo proteins and large aggregates can traverse the Golgi by a..., Mironov [/bib_ref] and soluble aggregated FM4-fusion protein [bib_ref] Megavesicles implicated in the rapid transport of intracisternal aggregates across the Golgi..., Volchuk [/bib_ref] , which are summarized and compared in Supplementary file 1 and 2, are mostly consistent with our observations. Our qualitative and quantitative data sketch a Golgi mini-stack as spindle-shaped with medial-cisternae possessing a larger diameter than both cis-and trans-cisternae [fig_ref] Figure 4: Golgi enzymes primarily localize to the interior of medial and trans-Golgi cisternae [/fig_ref]. Our morphological description of the Golgi mini-stack, such as the spindle shape of the stack and organization of the TGN, bear similarities to the plant Golgi mini-stack observed by electron tomography [bib_ref] Nanoscale architecture of endoplasmic reticulum export sites and of Golgi membranes as..., Staehelin [/bib_ref] , probably due to the lack of microtubule cytoskeleton in plants, which is similar to nocodazole-treated mammalian cells. Our findings suggest the spatial partition of the processing and transport function to the interior and rim of the Golgi stack, as depicted by our model in [fig_ref] Figure 6: A schematic model summarizing the organization of a Golgi mini-stack [/fig_ref].
EM studies have revealed that cisternal rims are dilated with a width of~100 nm, while their stacked interiors are narrow and tightly spaced with a width of~20 nm [bib_ref] In situ structural analysis of Golgi intracisternal protein arrays, Engel [/bib_ref] [bib_ref] Nanoscale architecture of endoplasmic reticulum export sites and of Golgi membranes as..., Staehelin [/bib_ref]. Recently, zipper-like intracisternal and intercisternalare overlaid onto a simplified diagram of a Golgi mini-stack together with the ERES and ERGIC. The red circle represents the mean of the LQ with flanking black bars representing the SEM. The cisternal interior, including central disks and inner-rings, is shaded yellow while the periphery of the Golgi mini-stack, including the cisternal rim, is shaded blue. Within the plot, red circles representing Golgi enzymes (labeled orange at the x-axis) are overlaid onto the yellow-shaded interior region, while those of components of the transport machinery (labeled black at the x-axis) are outside the mini-stack to indicate their periphery localization. DOI: https://doi.org/10.7554/eLife.41301.020 protein arrays have been discovered at interior regions of medial and trans-cisternae in green alga through the cryo-electron tomography [bib_ref] In situ structural analysis of Golgi intracisternal protein arrays, Engel [/bib_ref]. It was proposed that these tightly packed protein arrays comprise Golgi enzymes. Our super-resolution and EM data from the Golgi mini-stack provide direct evidence supporting this hypothesis. These enzyme-arrays might organize as an 'enzyme matrix' to 1) stack cisternal membrane, 2) retain Golgi enzymes or accessory proteins and 3) exclude trafficking machinery components by a possible molecular crowding mechanism. Therefore, it seems that, collectively, Golgi enzymes determine and maintain the characteristic structure of the Golgi complex. Most secretory cargos in higher eukaryotes undergo glycosylation in the Golgi complex. Our finding that the cisternal interior and rim correspond to processing and transport domain, respectively, implies that secretory cargos must access interior domains of different cisternae and then reside there long enough for sequential glycosylation. This is indeed what we observed for conventional cargos, such as GPI-anchored proteins, Furin, E-cadherin, VSVG and secretory GFP. On the other hand, these cargos probably have a sufficiently short residence time in the cisternal rim, in which they are either retrieved and retained by the 'enzyme matrix' to the interior or packed into membrane carriers targeting to the PM at the trans-Golgi. It seems that the retention by the 'enzyme matrix' occurs by default and is independent of glycosylation because secretory GFP is preferentially found within the cisternal interior. However, this is not the case for bulky cargos, such as collagenX and aggregated GFP-FM4-CD8a and FM4-moxGFP, which localized only at the rim and were excluded from the interior. These observations suggest that bulky cargos might be incompatible with the crowded molecular environment of the tightly packed 'enzyme matrix' and/or the narrow luminal space at the interior, which can have a width of <10 nm [bib_ref] In situ structural analysis of Golgi intracisternal protein arrays, Engel [/bib_ref]. Rim partitioning of large secretory cargos has previously been noted by EM [bib_ref] Procollagen traverses the Golgi stack without leaving the lumen of cisternae: evidence..., Bonfanti [/bib_ref] [bib_ref] In situ structural analysis of Golgi intracisternal protein arrays, Engel [/bib_ref] [bib_ref] Megavesicles implicated in the rapid transport of intracisternal aggregates across the Golgi..., Volchuk [/bib_ref]. Here, we directly visualized by light microscopy the size-dependent lateral partitioning of secretory cargos within the Golgi stack.
[formula] Myc-Sec13 E-COP Arf4-GFP Sec23a-mCherry Arf5-GFP GS27 J-COP GFP-ERGIC53 KDEL receptor GM130 GRASP65-GFP GRASP55-GFP GFP-Rab1a GFP-ACBD3 GFP-Golgin84 Giantin Myc-Sec34 Arf1-GFP GS15 GPP130-GFP GFP-GCC185 GFP-Rab6 Arl1 Vti1a GFP-GGA1 Golgin245 GFP-Golgin97 CI-M6PR Syntaxin6 Vamp4-GFP Furin CLCB GGA2 0≤LQ<0.25 [/formula]
This study did not attempt to resolve different intra-Golgi trafficking models and our discoveries can be explained by both cisternal progression and stable compartment models or their modified variants. Nonetheless, our findings provide important insight into the structure and organization of the Golgi complex.
# Materials and methods
## Antibodies and small molecules
The following mouse monoclonal antibodies (mAbs) were purchased from BD Biosciences: GM130 C-terminus, Golgin245, GGA2, GS15, GS27, GS28, Syntaxin6 and Vti1a. The following mouse mAbs were from Santa Cruz: Myc, CLCB and gCOP. Rabbit polyclonal antibody (pAb) against Furin, mouse mAb against CI-M6PR and Alexa Fluor 594 conjugated streptavidin were from Thermo Fisher Scientific. The following antibodies were commercially available from respective vendors: mouse mAb against Flag-tag and bCOP (Sigma-Aldrich), rabbit mAb against the N-terminus of GM130 (Abcam), rabbit pAb against Giantin (BioLegend) and mouse mAb against KDEL receptor (Enzo Life Sciences). Mouse mAbs against GALNT1 and GALNT2 were from H. Clausen. Rabbit pAbs against Arl1 and Golgin97 were previously described [bib_ref] Interaction of Arl1-GTP with GRIP domains recruits autoantigens Golgin-97 and Golgin-245/ p230..., Lu [/bib_ref] [bib_ref] Regulation of Golgi structure and function by ARF-like protein 1 (Arl1), Lu [/bib_ref]. The following small molecules were commercially available: biotin (IBA), biotin phenol (Iris Biotech GmbH), nocodazole (Merck) and D/D solubilizer (Clontech).
## Cell lines
HeLa and normal rat kidney fibroblast (NRK) cells were from American Type Culture Collection. Cell were assumed to be authenticated by the supplier. The presence of mycoplasma contamination was monitored by Hoechst 33342 staining.
## Cell culture and transfection
HeLa and NRK cells were cultured in Dulbecco's Modified Eagle's Medium (DMEM) supplemented with 10% fetal bovine serum. Cell transfection was conducted using according to manufacturer's manual. In live-cell imaging, cells grown on a F35 mm glass-bottom Petri-dish (MatTek) were imaged in CO 2 -independent medium (Invitrogen) supplemented with 4 mM glutamine and 10% fetal bovine serum at 37˚C. Unless otherwise indicated, all cells used were HeLa and treated with 33 mM nocodazole to induce the formation of Golgi mini-stacks.
## Production of giantin c-terminal antibody
It was conducted as previously described [bib_ref] A ternary complex comprising transportin1, Rab8 and the ciliary targeting signal directs..., Madugula [/bib_ref] [bib_ref] Mammalian Mon2/Ysl2 regulates endosome-to-Golgi trafficking but possesses no guanine nucleotide exchange activity..., Mahajan [/bib_ref]. Briefly, His-Giantin(3131-3201) was purified in urea from bacteria and used as the antigen to raise the antiserum in rabbits (Genemed Synthesis Inc). Recombinant GST-Giantin(3131-3235) was purified from bacteria and subsequently used to purify the antibody from the anti-serum.
## Super-resolution fluorescence microscopy
The Airyscan super-resolution microscope system (Carl Zeiss) comprises a Zeiss LSM710 confocal microscope equipped with an oil objective lens (alpha Plan-Apochromat 100 Â, 1.46 NA), a motorized stage, a temperature control environment chamber and Airyscan module. Fluorophores were excited by three laser lines with wavelengths of 488, 561 and 640 nm and their respective emission bandwidths were 495-550 nm, 595-620 nm and long pass 645 nm. The microscope system was controlled by ZEN software (Carl Zeiss). Pixel size of images ranged from 40 to 54 nm. For 3D imaging, the z-step of image stacks was 170 nm. Image stacks were subjected to Airyscan processing and maximal intensity projection (MIP) in ZEN software. Image analysis was performed in Fiji (https:// imagej.net/Fiji). We exhausted our images for all Golgi mini-stacks that were visually identifiable as either en face or side views.
En face averaging of golgi mini-stack images and radial mean intensity profile acquisition
En face view images of Giantin-labeled Golgi mini-stacks were averaged in semi-automatic software tools that were developed using macros of Fiji. Mini-stack images were first cropped to square shape and subjected to background subtraction. To quantify the size of the Giantin-ring, we adopted the concept of the gyradius from physics. For pixel i in the Giantin-ring image, assuming that I i is its intensity and r i is its distance to the center of fluorescence mass, the gyradius of the Giantin-ring can be calculated as with all pixels of the image considered. The macro 'gyradius and intensity normalization' (see Source code 1) calculates the gyradius of Giantin in a set of multi-channel images and resizes the set of images so that the gyradius of Giantin is 100 pixels. The canvas of the image set is further expanded to 701 Â 701 pixel. Using the macro 'Golgi mini-stack alignment' (see Source code 2), Golgi marker images are aligned so that their centers of fluorescence mass are at (350, 350), the center of the image. The en face averaged Golgi mini-stack image is acquired by z-projection of these aligned images. The radial mean intensity profile is acquired using the macro 'Radial mean intensity profile' (see Source code 3). The mean intensity of all pixels within a circle around the center of the fluorescence mass is plotted against its radius (ranging from 1 to 350 pixels). The radius of a Golgi marker is defined by the half maximum position of its outer slope of the intensity plot and is normalized by the corresponding radius of Giantin. Detailed steps are described in Supplementary file 4.
## Measuring diameters of giantin-rings
To measure the diameter of a Giantin-ring, a line was first drawn across its center. In the resulting line intensity profile (Fiji), the diameter of the ring was defined as the distance between the two halfmaximum-intensity points at outer slopes.
## Immunofluorescence labeling and rush cargo trafficking assay
These were conducted as previously described [bib_ref] A novel imaging method for quantitative Golgi localization reveals differential intra-Golgi trafficking..., Tie [/bib_ref]. By default, tagged-proteins were transiently transfected while non-tagged proteins were native and immuno-stained by their antibodies.
## Fluorescence labeling of apex2-mediated biotinylation
Nocodazole-treated HeLa cells expressing MGAT2-APEX2-GFP were incubated with 500 mM biotin phenol for 30 min at 37˚C. Cells were subsequently transferred to ice and treated with 1 mM H 2 O 2 for 1 min with brief agitation. After extensive washing with PBS containing 10 mM sodium ascorbate (Sigma-Aldrich), 5 mM Trolox (Sigma-Aldrich), and 10 mM sodium azide (Sigma-Aldrich), cells were fixed and processed for immunofluorescence. Biotinylated proteins were labeled by Alexa Fluor 594 conjugated streptavidin.
## Apex2-em
EM was performed as previously described [bib_ref] Architecture of the caveolar coat complex, Ludwig [/bib_ref] with minor modifications. In brief, NRK cells transiently expressing GPP130-APEX2-GFP or MGAT2-APEX2-GFP were fixed with 2% glutaraldehyde in 0.1 M cacodylate buffer pH 7.4 (CB) containing 2 mM CaCl 2 for 1 hr on ice, rinsed three times in CB, and incubated in 0.5 mg/ml 3,3'-diaminobenzidine and 0.5 mM H 2 O 2 in CB for 5 min. Cells were washed several times in CB and post-fixed in 1% osmium tetroxide in CB containing plasmid-backbones/), which is called GFP in this study, and the copy number of FM4-moxGFP in the Golgi punctum was calculated as 10.9 Â I punctum / I Nup .
## Fluorescence-conjugation of giantin antibodies
Alexa Fluor 647 and Alexa Fluor 488 were covalently conjugated onto a commercial (BioLegend) (against the N-terminus) and our homemade (against the C-terminus) rabbit pAb against Giantin, respectively, using APEX antibody labeling kit (Invitrogen) according to the manufacturer's protocol. Data availability All data generated or analysed during this study are included in the manuscript and supporting files.
[fig] Figure 1: Identifying the en face and side view of the Golgi mini-stack. All cells are nocodazole-treated HeLa cells and all images are super-resolution images unless specified otherwise. By default, tagged-proteins were transiently transfected while non-tagged proteins were native and stained by their antibodies. (A) The staining patterns of Giantin, Golgin84 and GPP130 appear as concentric rings. (B) The schematic representation of different orientation views (en face, oblique and side) of a Golgi cisterna and the corresponding expected images of a rim-localized protein (colored as pink). (C) [/fig]
[fig] Figure 2A -: D; Figure 2-figure supplement 1A-E). cis-Golgi proteins (0 LQ < 0.25) GM130, GRASP55, GRASP65 and Rab1a mainly appeared as a central disk and bar in en face and side views, respectively (Figure 2E-I; Figure 2-figure supplement 1F,G). When they appeared as rings in en face views, there were usually some interior tubular or sheet connections (Figure 2E,H; Figure 2-figure supplement 1F). Both observations suggest that these proteins probably localize throughout cis-cisternae. Medial and trans-Golgi proteins (0.25 LQ < 1.0) ACBD3, Golgin84, Giantin, GS15, GS28, Sec34, GPP130 and GCC185, all displayed ring-pattern localizations (Figure 1A,C; Figure 3A-F;Figure 3-figure supplement 1A-D) [/fig]
[fig] Figure 2, Figure supplement 1, Figure 3: 4A-D; Figure 4-figure supplement 1A-T), except Man1B1, ManII, MGAT4B and TPST2, which mostly appear as an inner-ring concentric to the corresponding Giantin-ring (Figure 4E,F; Figure 4-figure supplement 2A-F). The disk and ring patterns were more obviously revealed after en face averaging (Figure 4B,D,F; Figure 4-figure supplement 1B,D,F,H,J,L,N,P,R,T; Figure 4-figure supplement 2B,D,F). Since MGAT2-Myc and MGAT4B-AcGFP1 had almost the same LQs as Giantin (mean values: 0.53 and 0.50 vs 0.57 Components of the ERES, ERGIC and cis-Golgi transport machinery mainly localize to the periphery of the Golgi mini-stack. (A-D, E and H) Typical en face and side view images of Golgi transport machinery components. (A-D) ERES, ERGIC and cis-Golgi proteins (LQ <0). (E and H) cis-Golgi proteins (0 LQ < 0.25). (F-I) En face averaged images and radial mean intensity profiles corresponding to (E) and (H). n, the number of averaged Golgi mini-stack images. Scale bar, 500 nm. DOI: https://doi.org/10.7554/eLife.41301.006 The following figure supplement is available for figure 2: Typical en face and side view images of Golgi transport machinery components. DOI: https://doi.org/10.7554/eLife.41301.007 Tie et al. eLife 2018;7:e41301. DOI: https://doi.org/10.7554/eLife.41301 Components of the medial, trans-Golgi and TGN transport machinery mainly localize to the periphery of the Golgi mini-stack. (A-H) Medial and trans-Golgi proteins (0.25 LQ < 1.0), except Arf1, localize to the cisternal rim. En face and side view images are shown. Corresponding en face averaged images and radial mean intensity profiles are shown in (B, D, F and H). n, the number of averaged Golgi mini-stack images. (I-L) trans-Golgi and TGN proteins (LQ !1.0) appear compact or scattered at one end of the mini-stack. Arrows in (I) indicate colocalization between CLCB and Vamp4-GFP. Scale bar, 500 nm. [/fig]
[fig] Figure supplement 1: En face and side view images of the medial and trans-Golgi SNAREs, including GS15 and GS28, showed their rim localization (A and C). DOI: https://doi.org/10.7554/eLife.41301.009 Figure supplement 2. The lateral localization of components of the trans-Golgi and TGN transport machinery in the Golgi mini-stack. DOI: https://doi.org/10.7554/eLife. [/fig]
[fig] Figure 6: A schematic model summarizing the organization of a Golgi mini-stack. LQs of various Golgi residents (see [/fig]
[bib_ref] Architecture of the caveolar coat complex, Ludwig [/bib_ref] |
Evaluation of Antioxidant, Antidiabetic and Anticholinesterase Activities of Smallanthus sonchifolius Landraces and Correlation with Their Phytochemical Profiles
The present study aimed to investigate the phytochemical profile of leaf methanol extracts of fourteen Smallanthus sonchifolius (yacon) landraces and their antioxidant, anticholinesterase and antidiabetic activities that could lead to the finding of more effective agents for the treatment and management of Alzheimer's disease and diabetes. For this purpose, antioxidant activity was assessed using different tests: ferric reducing ability power (FRAP), 2,2-diphenyl-1-picryl hydrazyl (DPPH), nitric oxide (˙NO) and superoxide (O2˙−) scavenging and lipid peroxidation inhibition assays. Anticholinesterase activity was investigated by quantifying the acetylcholinesterase (AChE) and butyrylcholinesterase (BChE) inhibitory activities, whereas antidiabetic activity was investigated by α-amylase and α-glucosidase inhibition tests. To understand the contribution of metabolites, phytochemical screening was also performed by high performance liquid chromatography-diode array detector (HPLC-DAD) system. Among all, methanol extract of PER09, PER04 and ECU44 landraces exhibited the highest relative OPEN ACCESS antioxidant capacity index (RACI). ECU44 was found to be rich in 4,5-di-O-caffeoylquinic acid (CQA) and 3,5-di-O-CQA and displayed a good α-amylase and α-glucosidase inhibition, showing the lowest IC50 values. Flavonoids, instead, seem to be involved in the AChE and BChE inhibition. The results of this study revealed that the bioactive compound content differences could be determinant for the medicinal properties of this plant especially for antioxidant and antidiabetic activities.
# Introduction
Medicinal plants, as source of remedies, are widely used as alternative therapeutic tools for the prevention or treatment of many diseases. Recently, great attention has been devoted to the use of natural compounds, due to their nutritional and pharmacological characteristics.
Smallanthus sonchifolius (Poepp. & Hendl.) H. Robinson, commonly called yacon, belongs to the Asteraceae family. It is a tuberous plant native to the Andes, but it has been introduced in Japan, New Zealand, Europe (primarily in Czech Republic) and United States [bib_ref] Smallanthus sonchifolius (Poeppig & Endlicher) H. Robinson): A new crop in the..., Fernández [/bib_ref]. Yacon is a perennial herb with large dark green leaves and two types of underground portions: edible tuberous roots, used by the plant for food storage, and fibrous roots used for vegetative reproduction. Each plant produces 4 to 20 edible tubers that can reach 20 kg [bib_ref] Genetic diversity between yacon landraces from different countries based on random amplified..., Milella [/bib_ref]. Andean indigenous people have always used yacon as an important crop: they have used it not only for its edible tuberous roots, but also as a vegetable and medicinal plant. Roots of yacon can contain high amounts (40%-70% dry weight) of fructooligosaccharides [bib_ref] Yacon, a new source of prebiotic oligosaccharides with a history of safe..., Ojansivu [/bib_ref] [bib_ref] Impact of yacon landraces cultivated in the czech republic and their ploidy..., Fernandez [/bib_ref] that are not metabolized in the human digestive tract and hence their consumption does not enhance the level of glucose in the blood. In addition, studies have reported that extracts of leaves reduce glycemia in the plasma of diabetic rats [bib_ref] Hypoglycemic effect of the water extract of Smallantus sonchifolius (yacon) leaves in..., Aybar [/bib_ref] and some constituents of yacon leaves inhibit the α-glucosidase enzyme involved in the diabetes [bib_ref] Anti-diabetes constituents in leaves of Smallanthus sonchifolius, Xiang [/bib_ref]. For this reason yacon is considered to be a food and a remedy with a high potential for diabetics. The low energy value of its tuberous roots makes yacon an ideal foodstuff for people suffering from obesity. Leaves are also a rich source of antioxidants [bib_ref] Smallanthus sonchifolia (Poepp. et Endl.) H. Robinson) chemical composition and use-A review, Lachman [/bib_ref] and several studies showed that this part of the plant possesses different biological effects, like inhibition of migration of polymorphonuclear leucocytes [bib_ref] Quantitative determination of enhydrin in leaf rinse extracts and in glandular trichomes..., Schorr [/bib_ref] , immunomodulation [bib_ref] Biological and chemical variability of maca and yacon, Lebeda [/bib_ref] , antioxidant, cytoprotection effects [bib_ref] Antioxidant activity of extracts from the leaves of Smallanthus sonchifolius, Valentova [/bib_ref] [bib_ref] Radical scavenging and anti-lipoperoxidative activities of Smallanthus sonchifolius leaf extracts, Valentova [/bib_ref] [bib_ref] Investigation of phenolic acids in yacon (Smallanthus sonchifolius) leaves and tubers, Simonovska [/bib_ref] and antimicrobical activity [bib_ref] Purification and identification of antimicrobial sesquiterpene lactones from yacon (Smallanthus sonchifolius) leaves, Lin [/bib_ref].
The search for the bioactive compounds from medicinal plants is always an alternative mean of finding new drugs. Phytochemical screening of S. sonchifolius will lead to the rationalization of the use of this plant in various diseases and could also lead to the finding of more effective agents that have effective treatment roles against specific diseases. Alzheimer's disease (AD) was firstly described in 1906 by Alois Alzheimer, a Bavarian neuropsychiatrist [bib_ref] Natural product inhibitors of acetylcholinesterase, Hostettmann [/bib_ref]. AD is a progressive neurodegenerative disorder of the elderly, characterized by widespread loss of central cholinergic function. The only symptomatic treatment proven effective to date is the use of cholinesterase (ChE) inhibitors to augment surviving cholinergic activity. ChE inhibitors act on the enzymes that hydrolyze acetylcholine (ACh) following synaptic release. In the healthy brain, acetylcholinesterase (AChE) predominates (80%) and butyrylcholinesterase (BChE) is considered to play a minor role in regulating brain ACh levels. Therefore both enzymes are likely to have involvement in regulating ACh levels and represent legitimate therapeutic targets to ameliorate the cholinergic deficit. Many plants have been studied by different approaches for the identification of new AChE inhibitors (AChE-Is) and different classes of plant-derived natural products have been considered as new AChE-Is potentially useful for AD treatment. Both non-alkaloids and alkaloid-derivative compounds [bib_ref] Acetylcholinesterase inhibitors from plants, Mukherjee [/bib_ref] have been studied.
The incidence of diabetes mellitus in the world is increasing at an alarming rate, affecting close to 5% of its population and it is strictly related to several other diseases [bib_ref] In vivo assessment of antidiabetic and antioxidant activities of rosemary (Rosmarinus officinalis)..., Bakirel [/bib_ref] [bib_ref] Relationship between proinflammatory and antioxidant proteins with the severity of cardiovascular disease..., García-Fontana [/bib_ref]. Diabetes mellitus is a complex disease characterized by gross derangement in carbohydrate, fat and protein metabolism, due to deficiency in insulin secretion and/or action [bib_ref] Hypoglycemic and hypolipidemic effects and antioxidant activity of fruit extracts from Lycium..., Luo [/bib_ref]. Mammalian α-amylase is a prominent enzyme in the pancreatic juice, breaking down large and insoluble starch molecules into absorbable molecules, ultimately maltose [bib_ref] Microbial α-amylases: A biotechnological perspective, Gupta [/bib_ref]. α-Glucosidase, on the other hand, anchored in the mucosal brush border of the small intestine, catalyzes the end step of digestion of starch and disaccharides that are abundant in human diet [bib_ref] α-Glucosidase inhibitor of Terminalia species, Anam [/bib_ref]. Inhibitors of α-amylase and α-glucosidase delay the breakdown of carbohydrate in the small intestine and decrease the postprandial blood glucose excursion levels in diabetic patients. The inhibition of these two prominent enzymes has been found as a useful and effective strategy to lower the levels of postprandial hyperglycemia [bib_ref] α-Amylase inhibitory effect of 3β-olean-12-en-3-yl (9Z)-hexadec-9-enoate isolated from Spondias mombin leaf, Fred-Jaiyesimi [/bib_ref].
Moreover oxidative stress has been implicated in various pathological conditions involving cardiovascular disease, cancer, neurological disorders (Alzheimer's disease and Parkinson's disease), diabetes, ischemia/reperfusion, other diseases and ageing. Convincing evidence for the association of oxidative/nitrosative stress and acute and chronic diseases lies on validated biomarkers of oxidative stress [bib_ref] Free radicals and antioxidants in normal physiological functions and human disease, Valko [/bib_ref]. For this reasons we have also investigated the antioxidant potential of yacon extracts by five different assays. In conclusion, the aim of this study was to investigate the biological properties, including antioxidant, antidiabetic and anticholinesterase activity, and the phytochemical profiles of leaf methanol extracts of fourteen yacon landraces.
# Results and discussion
## Extraction yields, total polyphenolic, tannin and flavonoid content
Dried foliar tissue of 14 landraces of yacon was extracted by maceration technique by solvents of crescent polarity: n-hexane (E), chloroform (C), chloroform:methanol 9:1 (CM) and methanol (M). Dried extract yield was calculated. Data showed the different extractive capacities of each solvent. Methanol was the solvent with the highest extractive capacity, the percentage of yield ranged from 4.5% to 9.8% followed by n-hexane (3.2% to 6.9%); in fact, the latter solvent allows to extract mainly the chlorophylls, abundant in the leaf tissue. The percentage of yield obtained by chloroform and chloroform:methanol mixture (9:1) ranged from 2.3% to 3.8% and from 0.7% to 3.0%, respectively.
The total polyphenolic content (TPC) of different extracts of 14 genotypes of yacon was determined by Folin-Ciocalteu assay. According to literature, methanol extracts show higher content of phenolic compounds than other extracts (data not shown); phenols are also the main responsible compounds of the antioxidant activity; for this reason, the leaf methanol extracts of fourteen yacon landraces have been further analyzed [bib_ref] Effects of extraction methods of phenolic compounds from Xanthium strumarium L. and..., Scherer [/bib_ref]. Results of methanol extracts were expressed as mg of gallic acid equivalent (GAE)/g of dried extract , by using a standard curve. Yacon landraces displayed quantitative differences in TPC, with a mean value of 74.47 mg GAE/g. The TPC ranged from 58.49 ± 1.03 to 91.07 ± 4.87 mg GAE/g of dried extract, in methanol extracts of PER07 (PER07-M) and PER09-M, respectively. Methanol extract of polyploid sample ECU44 showed a phenolic content similar to that of the mother plant (ECU41). . Results of total polyphenol content (TPC), total flavonoid content (TFC), total tannin content (TTC), ferric reducing power (FRAP), 2,2-diphenyl-1-picrylhydrazyl (DPPH) scavenging activity, lipid peroxidation (LOO˙) inhibition, nitric oxide (˙NO) and superoxide radical anion (O2˙−) scavenging activities of methanol extracts of investigated yacon landraces. Tannins and flavonoids are secondary metabolites that are widely distributed in the plant kingdom and that exhibit various biological properties.
Protein precipitation assays represent the operational property to measure the amount of both condensed and hydrolysable tannins. As reported by Hagerman and Butler [bib_ref] Protein precipitation method for the quantitative determination of tannins, Hagerman [/bib_ref] , bovine serum albumin (BSA) was used as protein in this study. Ferric chloride reacts with phenolic compounds in solution to form complexes with the general formula Fe-(OR)6 −3 where -OR − represents the ionized phenol. The maximum absorption wavelength of the complex is dependent upon the nature of the phenol and the solvent; the complex formed between tannins and ferric chloride in alkaline solution is violet (λmax 510 nm). The triethanolamine, necessary for the maintenance of high pH, was incorporated into the detergent sodium dodecyl suplhate (SDS) solution used to dissolve the tannin-protein complex.
Tannin content was expressed as mg tannic acid equivalent (TAE)/g of dried extract by using a standard curve and results are reported in PER10-M and PER11-M extracts of yacon showed the highest tannin content, 14.58 ± 0.47 and 14.28 ± 0.14 mg TAE/g, respectively. In addition, PER05-M (12.94 ± 0.16 mg TAE/g) and PER09-M (10.31 ± 0.80 mg TAE/g) exhibited a value higher than mean value (9.39 mg TAE/g). The lowest content was observed for PER04-M extract (6.38 ± 0.03 mg TAE/g), followed by PER08-M (6.45 ± 0.30 mg TAE/g). The polyploids landraces, ECU43-M (7.63 ± 0.15 mg TAE/g) and ECU44-M (8.13 ± 0.99 mg TAE/g), showed lower content than their mother plant, ECU41-M (9.33 ± 0.55 mg TAE/g).
The spectrophotometric assay based on aluminum complex formation is one of the most commonly used procedure for the total flavonoid determination.
Aluminum chloride forms acid stable complexes with the C-4 keto group and either the C-3 or C-5 hydroxyl group of flavones and flavonols. In addition, aluminum chloride forms acid labile complexes with the ortho-dihydroxyl groups in the A or B rings of flavonoids [bib_ref] Quantitative estimation of β-sitosterol, total phenolic and flavonoid compounds in the leaves..., Rajanandh [/bib_ref].
Total flavonoid content was expressed as mg of quercetin equivalent (QE)/g of dried extract . ECU44-M and PER02-M showed the highest content of flavonoids, 104.10 ± 2.45 and 102.37 ± 5.88 mg QE/g, respectively. Flavonoid content, higher than the mean value (841.52 mg QE/g of dried extract) was observed in PER01-M (99.59 ± 1.43 mg QE/g), PER05-M (98.43 ± 3.18 mg QE/g), PER07-M (93.00 ± 1.57 mg QE/g) and PER09-M (91.72 ± 3.46 mg QE/g). The lowest content was observed in the polyploid sample ECU43-M (57.17 ± 1.21 mg QE/g), more similar to its mother plant ECU41-M (62.43 ± 1.47 mg QE/g) and PER08-M (58.01 ± 0.48 mg QE/g).
In-vitro assays used in this study are most often colorimetric methods of choice for general determinations of polyphenols, tannins and flavonoids. Although the presence of other constituents in the crude extract could slightly interfere with their quantification, these methods are routinely used [bib_ref] Quantification of the total phenolic and flavonoid contents, and antibacterial activity of..., Edewor [/bib_ref] [bib_ref] Radial diffusion method for determining tannin in plant extracts, Hagerman [/bib_ref].
## In-vitro biological activity
## Reducing power by ferric reducing antioxidant power (frap) test
Generally, phenolic compounds are responsible for biological activities as antioxidant capacity and inhibition of enzymes involved in common diseases. Antioxidant activity should not be based on a single antioxidant test model and in practice several in vitro test procedures are carried out for evaluating antioxidant activities with the samples of interest [bib_ref] Prediction of the antioxidant activity of extra virgin olive oils produced in..., Condelli [/bib_ref]. The ferric reducing antioxidant power (FRAP) method is based on the reduction of the Fe 3+ complex of tripyridyltriazine (Fe(TPTZ) 3+ ) to the intensely blue colored Fe 2+ complex (Fe(TPTZ) 2+ ) by antioxidants. The method is simple and rapid; it was originally applied to plasma but has been extended to other biological fluids, foods, plant extracts, juices, etc.
Results ranged from 31.55 ± 0.96 to 66.80 ± 1.90 mg of Trolox Equivalent (TE)/g of dried extract. Extracts of PER09-M and PER04-M landraces showed the highest reducing power, 66.80 ± 1.90 and 65.32 ± 6.04 mg TE/g of dried extract, respectively. Sample PER01-M (31.55 ± 0.96 mg TE/g) exhibited the lowest value, followed by PER07-M (37.25 ± 4.59 mg TE/g), PER08-M (39.55 ± 3.46 mg TE/g) and ECU43-M (34.83 ± 3.79 mg TE/g). ECU44-M (49.79 ± 2.80 mg TE/g) showed higher reducing power than ECU41-M (45.97 ± 0.15 mg TE/g), its mother plant.
## Radical scavenging activity by 2,2-diphenyl-1-picrylhydrazyl (dpph) assay
In vitro antioxidant tests using free radical traps are relatively straightforward to perform. Among free radical scavenging methods, the one involving 2,2-diphenyl-1-picrylhydrazyl (DPPH) is rapid, simple, highly reproducible and inexpensive in comparison to other test models. The scavenging of the stable DPPH radical is widely used to evaluate the antioxidant activity of phenolic compounds extracted from fruits, vegetables, cereal grains, wine, etc. [bib_ref] Evaluation of free radical scavenging of dietary carotenoids by the stable radical..., Jimenez-Escrig [/bib_ref]. The method is based on the reduction of methanol DPPH solution in the presence of a hydrogen donating antioxidant, due to the formation of the non-radical form DPPH-H. The extracts were able to reduce the stable radical DPPH to the yellow colored diphenylpicrylhydrazine, in a concentration-dependent manner. Results were expressed as IC50 and all extracts showed a fair DPPH scavenging activity, lower than Trolox (used as standard) with an IC50 of 0.08 ± 0.002 mg/mL. The most active of the yacon landraces was PER04-M (IC50 = 2.08 ± 0.99 mg/mL) followed by PER09-M (IC50 = 2.89 ± 0.12 mg/mL). PER01-M (IC50 = 4.21 ± 0.35 mg/mL), PER08-M (IC50 = 4.19 ± 0.47 mg/mL) and PER07-M (IC50 = 4.39 ± 0.32 mg/mL) showed the lowest activity. Methanol extract of polyploids samples (ECU43-M and ECU44-M) and their mother plant (ECU41-M) displayed similar DPPH-scavenging activity.
## Nitric oxide (˙no) radical scavenging activity
˙NO is known to be a ubiquitous free radical, being distributed in tissues or organ systems and supposed to have a vital role in neuromodulation or as a neurotransmitter in the central nervous system CNS. In addition to reactive oxygen species, nitric oxide is also implicated in inflammation, cancer and other pathological conditions. In the in vitro assay performed, sodium nitroprusside is known to decompose in aqueous solution at physiological pH (7.2) producing ˙NO [bib_ref] Nitric oxide: No apoptosis or turning it on?, Brune [/bib_ref]. Under aerobic conditions, ˙NO reacts with oxygen to produce stable products (nitrate and nitrite), the quantities of which can be determined using Griess reagent [bib_ref] The nitric oxide-scavenging properties of Ginkgo biloba extract EGb 761, Marcocci [/bib_ref]. The Griess reaction is based on the two-step diazotization reaction in which acidified NO2 − produces a nitrosating agent, which reacts with sulfanilic acid to produce the diazonium ion. This ion is then coupled to N-(1-naphthyl) ethylenediamine to form a chromophoric azo-derivative, absorbing at 560 nm [bib_ref] Bauhinia forficata link authenticity using flavonoids profile: Relation with their biological properties, Ferreres [/bib_ref].
The ability to inhibit nitric oxide production was concentration-dependent and, as far as we know, not reported yet for yacon leaf extract. The IC50 was not reached with the tested concentrations, so the results were expressed as IC25 (mg/mL) and are reported in . Ascorbic acid was used as reference, showing a IC25 value of 0.06 ± 0.003 mg/mL. All extracts demonstrated to possess nitric oxide scavenging activity. Foliar extracts of ECU41-M (IC25 = 0.03 ± 0.002 mg/mL) and ECU44-M (IC25 = 0.04 ± 0.001 mg/mL) showed the highest scavenging activity; also PER09-M (IC25 = 0.05 ± 0.006 mg/mL) reported a good scavenging activity, higher than ascorbic acid. Sample PER01-M, with the highest IC25 value (0.27 ± 0.04 mg/mL), was the less active.
## Superoxide radical (o2˙−) scavenging activity
The production of reactive oxygen species, such as superoxide (O2˙−), by mitochondria is a major cause of cellular oxidative stress and it contributes to many pathological conditions [bib_ref] The mechanism of superoxide production by NADH:ubiquinone oxidoreductase (complex I) from bovine..., Kussmaul [/bib_ref]. Superoxide anion radical (O2˙−) is generated in aerobic cells due to electron leakage from the electron transport chain.
In this study, in vitro superoxide anions are generated by phenazine methosulfate-β-nicotinamide adenine dinucleotide (PMS-NADH) system. The superoxide scavenging capacity of methanol extracts was quantified by their ability to inhibit nitrotetrazolium blue chloride (NBT) reduction by superoxide.
Scavenging activity was, also in this case, concentration-dependent and the activity of ascorbic acid, used as reference, was compared with the samples by IC50 values. Ascorbic acid displayed the strongest activity, showing the lowest IC50 (0.28 ± 0.04 mg/mL). Methanol extract of ECU44-M exhibited the lowest IC50 value among investigated landraces (IC50 = 0.81 ± 0.05 mg/mL), so it was the most active against superoxide radical obtaining a comparable value to ascorbic acid, used as positive control. Also PER09-M (IC50 = 1.81 ± 0.08 mg/mL), PER04-M (IC50 = 1.85 ± 0.04 mg/mL) and PER02-M (IC50 = 2.17 ± 0.32 mg/mL) showed a good superoxide anion scavenging activity, with an IC50 lower than mean value (IC50 = 2.48 mg/mL). PER01-M was the weakest physiological radicals scavenger, both against superoxide radical (IC50 = 3.81 ± 0.29 mg/mL) and ˙NO.
## Lipid peroxidation (loo˙) inhibition test
Lipid peroxidation is thought to be an important factor in the pathophysiology of a number of diseases. Lipid peroxidation is probably the most extensively investigated free radical-induced process [bib_ref] Biomarkers of free radical damage applications in experimental animals and in humans...., De Zwart [/bib_ref]. In this study, lipid peroxidation inhibition was assessed by evaluating the formation of conjugated dienes, structures with a double-single-double bond (-C=C-C=C-) arrangement, absorbing UV light in the wavelength range 230-235 nm [bib_ref] The measurement and mechanism of lipid peroxidation in biological systems, Gutteridge [/bib_ref] [bib_ref] Lipid peroxidation: Its mechanism, measurement, and significance, Halliwell [/bib_ref] [bib_ref] Detection of conjugated dienes by second derivative ultraviolet spectrophotometry, Corongiu [/bib_ref].
This test was used in the antioxidant activity evaluation of the methanol extracts of yacon. Results were expressed as IC50 . Butylated hydroxytoluene (BHT) was used as reference, showing the highest inhibition of lipid peroxidation (IC50 = 0.01 ± 0.001 mg/mL). All extracts displayed IC50 values higher than the reference compound. Among all extracts, the most able to scavenge LOO˙ were PER01-M (IC50 = 0.51 ± 0.05 mg/mL) and PER11-M (IC50 = 0.52 ± 0.06 mg/mL); similar values were also observed with PER02-M (IC50 = 0.61 ± 0.03 mg/mL), PER09-M (IC50 = 0.64 ± 0.04 mg/mL) and PER07-M (IC50 = 0.67 ± 0.03 mg/mL), whereas the lowest capacity was found for PER08-M (IC50 = 1.26 ± 0.05 mg/mL), ECU43-M (IC50 = 1.20 ± 0.07 mg/mL) and ECU41-M (IC50 = 1.19 ± 0.13 mg/mL).
## Inhibition activity against α-glucosidase and α-amylase enzymes
Medicinal plants and herbal extracts containing glycosides, flavonoids, tannins, etc., have been reported to demonstrate antidiabetic [bib_ref] Antidiabetic potential of Barleria lupulina extract in rats, Suba [/bib_ref] and anticholinesterase activities [bib_ref] Antioxidant and anticholinesterase investigations of Rumex hastatus d. Don: Potential effectiveness in..., Ahmad [/bib_ref]. One therapeutic approach for treating diabetes is to decrease post-prandial hyperglycemia. This is done by hindering the absorption of glucose through inhibition of the carbohydrate hydrolyzing enzymes, α-amylase and α-glucosidase, in the digestive tract. In our investigation, the inhibitory activity of crude methanol extracts against both mentioned enzymes was carried out. Previous study reported only the α-glucosidase inhibitory activity of some isolated constituents from yacon leaves [bib_ref] Anti-diabetes constituents in leaves of Smallanthus sonchifolius, Xiang [/bib_ref].
Results were expressed as IC50. Acarbose was used as reference drug showing IC50 values lower than 0.20 mg/mL in both inhibition assays. Data obtained showed that methanol extracts of yacon leaves were stronger inhibitors of α-amylase than α-glucosidase, as presented in. Among yacon extracts, the most active was found to be ECU44-M landrace for both enzyme inhibitory activities (IC50 = 0.26 ± 0.02 mg/mL for α-amylase inhibition and 1.30 ± 0.04 mg/mL for α-glucosidase); ECU41-M and PER09-M extracts act as α-amylase inhibitors; additionally, PER11-M can be considered a moderate inhibitor of α-glucosidase; on the other hand, the less active against both enzymes were PER01-M and PER04-M extracts.
According to previous study [bib_ref] Anti-diabetes constituents in leaves of Smallanthus sonchifolius, Xiang [/bib_ref] , our investigation reported that methanol extracts moderately inhibited α-amylase and α-glucosidase and this result can, at least partially, explain the traditional use of yacon leaves for treating hyperglycemia.
## Inhibition activity against acetylcholinesterase and butyrylcholinesterase enzymes
For the first time yacon extracts were tested to evaluate their capacity to inhibit the cholinesterase enzymes. The crucial role of the cholinesterases in neural transmission makes them a primary target of a large number of cholinesterase-inhibiting drugs and this has relevance to the treatment of neurodegenerative disorders.
The method for the screening of ChE inhibitory activity from natural resources based on Ellman's reactions has been reported [bib_ref] A new and rapid colorimetric determination of acetylcholinesterase activity, Ellman [/bib_ref]. In 1959, Ellman introduced 5,5′-dithio-bis-(2-nitrobenzoic acid), also known as DTNB, as a versatile water-soluble compound for quantitating free sulfhydryl groups in solution. A solution of this compound produces a measurable yellow-colored product when it reacts with sulfhydryl. Consequently, Ellman's reagent is very useful as a sulfhydryl assay reagent because of its specificity for -SH groups at neutral pH, high molar extinction coefficient and short reaction time. DTNB reacts with a free sulfhydryl group to yield a mixed disulfide and 2-nitro-5-thiobenzoic acid (TNB). The target of DTNB in this reaction is the conjugated base (R-S-) of a free sulfhydryl group.
In the in vitro assays, AChE and BChE efficiently catalyze the hydrolysis of acetyl-and butyryl-thiocholine (AcSCh and BuSCh), sulfur analogs of their respective natural substrate, acetylcholine. Upon hydrolysis, these substrate analogs produce acetate (or butyrate) and thiocholine. Thiocholine in the presence of the highly reactive DTNB ion reacts to generate the yellow 5-thio-2-nitrobenzoate anion. The yellow color can be quantified at 405 nm [bib_ref] A new and rapid colorimetric determination of acetylcholinesterase activity, Ellman [/bib_ref].
Galantamine, a natural cholinesterase inhibitor used in therapeutics, was used as reference and its capacity to inhibit the two enzymes involved in Alzheimer' disease was measured: 93.5% ± 1.2% AChE inhibition and 68.3% ± 3.4% BChE inhibition, at only 0.01 mg/mL. Yacon landraces demonstrated a concentration-dependent inhibition, but unfortunately the weak activity of the extracts did not allow to reach, and consequently calculate, the IC50. In fact, results were expressed as % of inhibition at 2.5 mg/mL, and they showed an inhibition lower than 35% against both enzymes.
Among them, PER01-M showed the strongest capacity against both AChE and BChE, in this case similar values have been found: 30.87% ± 4.20% and 29.83% ± 2.68%, respectively. Most of the yacon landraces displayed the same effect against both cholinesterases, showing an inhibition lower than 25% at the tested concentration.
## High perfomance liquid chromatography-diode array detector analysis
Phenolic compounds have characteristic UV-Vis spectral properties that are unique for each of the different classes of phenolics, and may therefore be used for tentative identification of the class of a phenolic compound.
Phytochemical studies of yacon leaves showed the presence of several melampolide-type sesquiterpene lactones such as sonchifolin, uvedalin, enhydrin, fluctuanin [bib_ref] Hypoglycemic activity of leaf organic extracts from Smallanthus sonchifolius: Constituents of the..., Genta [/bib_ref] and phenolic compounds, mainly chlorogenic acid and other caffeic derivatives, have been identified in different extracts of yacon leaves together with some flavonoids [bib_ref] Smallanthus sonchifolia (Poepp. et Endl.) H. Robinson) chemical composition and use-A review, Lachman [/bib_ref] [bib_ref] Investigation of phenolic acids in yacon (Smallanthus sonchifolius) leaves and tubers, Simonovska [/bib_ref] [bib_ref] Determination of rutin and quercetin in extraction of leaves of yacon with..., Yang [/bib_ref] [bib_ref] Topical anti-infl ammatory activity of yacon leaf extracts, Oliveira [/bib_ref] [bib_ref] Antioxidant activities and qualiquantitative analysis of different Smallanthus sonchifolius ((Poepp. and Endl.)..., Russo [/bib_ref].
High perfomance liquid chromatography-diode array detector (HPLC-DAD) analysis of methanol extract of yacon leaves confirmed the presence of rutin and caffeic acid. Other compounds were identified by comparing with the results obtained with Cynara cardunculus extract analyzed under the same conditions [bib_ref] Antioxidative properties of cardoon (Cynara cardunculus L.) infusion against superoxide radical, hydroxyl..., Valentão [/bib_ref]. Methanol extract of C. cardunculus is characterized to be rich in caffeoylquinic acids and analytical parameters allowed to identify chlorogenic acid, 3,5-O-dicaffeoylquinic acid, 1,5-O-dicaffeoylquinic acid and 4,5-O-di-caffeoylquinic acid. The compound 1,5-O-dicaffeoylquinic acid was identified for the first time in yacon leaves.
Analyzing HPLC chromatograms, it is possible to observe a similar phytochemical profile; it was just reported a representative chromatogram [fig_ref] Figure 2: Cont. [/fig_ref]. Quantitative analysis demonstrated that all samples contained a high amount of phenolic acids. Results are reported in [fig_ref] Table 2: HPLC quantification of identified compounds in yacon landraces [/fig_ref]. [fig_ref] Table 2: HPLC quantification of identified compounds in yacon landraces [/fig_ref].
Among all methanol extracts, PER09-M contained the highest amount of secondary metabolites, both flavonoids and phenolic acids, and in particular, it was found to be rich in caffeic and chlorogenic acids. The ECU44 landraces showed large amounts of phenylpropanoid derivatives and, among all identified compounds, caffeoylquinic acids are abundant. Sample ECU41-M, together with its polyploids, ECU43-M and ECU44-M, demonstrated the highest content of 3,5-O-dicaffeoylquinic acid. PER01-M was the sample with the lowest content of secondary metabolites, in particular it showed a low content of phenylpropanoid acids, but together with PER02-M, contained a high amount of rutin and flavonoid derivatives. PER08-M showed the lowest content of rutin and flavonoids, according to the results obtained by colorimetric assay.
# Statistical analysis
To compare data obtained by different chemical method used to evaluate extract antioxidant activity, a new concept, relative antioxidant capacity index (RACI) [bib_ref] An integrated approach to evaluate food antioxidant capacity, Sun [/bib_ref] , was proposed. This concept provides a more comprehensive comparison when several samples are analyzed.
All methods used for antioxidant activity determination together with TPC were included in RACI calculation. Recently, TPC assay has been proposed for the measurement of total reducing capacity of samples [bib_ref] Vitis vinifera leaves towards bioactivity, Fernandes [/bib_ref].
Results of antioxidant activity expressed as IC25 or IC50 were converted in 1/IC25 and 1/IC50, before the RACI calculation. Data of relative antioxidant activity were represented as histograms [fig_ref] Figure 3: Relative antioxidant capacity index of investigated yacon landraces [/fig_ref] To understand the correlation among all studied variables, total polyphenols, flavonoid and tannin content and biological activity (reducing power, radical-scavenging activity, lipid peroxidation, inhibition of AChE, BChE, α-glucosidase and α-amylase), Pearson's correlation coefficient was calculated. The analysis was conducted using averaged values of each variable and results are reported in . Data expressed as IC50 and IC25, show a negative correlation. The highest correlation was observed between total polyphenol content and reducing power (r = 0.77). Polyphenols seem to be mostly involved in scavenging activity against DPPH (r = −0.58), superoxide radical (r = −0.39) and nitric oxide (r = −0.53) demonstrating the ability of these compounds to quench radicals and reduce the oxidative and nitrosative stress. Flavonoids were the main contributors to lipid peroxidation inhibition (r = −0.72). Pearson correlation between phenolic content and enzymatic inhibition was also calculated. Polyphenols showed to be moderately related to α-glucosidase (r = −0.51) and α-amylase (r = −0.41) enzyme inhibitions, as well flavonoids can contribute to AChE and BChE inhibition (r = 0.66 and r = 0.51, respectively), as previously reported [bib_ref] Physiological components of kiwifruits with in vitro antioxidant and acetylcholinesterase inhibitory activities, Lim [/bib_ref]. Among all investigated activities, tannin content demonstrated the highest correlation with α-glucosidase inhibition (r = −0.45). . Pearson correlation among total polyphenol content (TPC), total tannin content (TTC), total flavonoid content (TFC) and antioxidant activity (reducing power, 2,2-diphenyl-1-picryl hydrazyl (DPPH), ˙NO, O2˙− scavenging activity and lipid peroxidation inhibition) and enzymatic inhibition (AChE, BChE, α-amylase and α-glucosidase enzymes). In order to discriminate the yacon landraces on the basis of the chemical profile and biological activity, data were normalized and principal component analysis (PCA) was conducted on the correlation matrix. In this study, PCA was used for a better visualization of data sets obtained from the determinations of all studied variables as previously reported [bib_ref] Effect of dehydration process on mineral content, phenolic compounds and antioxidant activity..., Panceri [/bib_ref] [bib_ref] Analysis of phenolic compounds and antioxidant activity in wild blackberry fruits, Oszmiański [/bib_ref]
## Variables
## Experimental section
## Chemicals
Chloroform, n-hexane, methanol, hydrochloric acid and glacial acetic acid were purchased from Carlo Erba (Milano, Italy). Ethanol, potassium di-hydrogen phosphate, di-sodium tetraborate, iron (II) sulfate (FeSO4·7H2O) and N-(1-naphthyl) ethylenediamine dihydrochloride were obtained from Merck (Darmstadt, Germany). Folin-Ciocalteu reagent 2N, sodium carbonate, sodium acetate trihydrate, 2,4,6-tripyridyl-s-triazine (TPTZ), iron (III) chloride (FeCl3·6H2O), 2,2-diphenyl-1-picryl hydrazyl (DPPH) radical, and standards as 6-hydroxy-2,5,7,8-tetramethylchroman-2-carboxylic acid (Trolox) and gallic acid were purchased from Sigma-Aldrich (Milan, Italy). Butylated hydroxytoluene (BHT, 2,6-bis(1,1-dimethylethyl)-4-methylphenol), β-nicotinamide adenine dinucleotide reduced form (NADH), phenazine methosulfate (PMS), nitrotetrazolium blue chloride (NBT), 5,5′-dithio-bis(2-nitrobenzoic acid) (DTNB), sodium nitroprusside dehydrate (SNP), linoleic acid, L-ascorbic acid and sulphanilamide, acetylcholinesterase (AChE) from electric eel (type VI-s, lyophilized powder), acetylthiocholine iodide (ATCI), butyrylcholinesterase (BChE) from equine serum (lyophilized powder), S-butyrylthiocholine chloride (BTCC), α-glucosidase (type I from baker's yeast) and 4-nitrophenyl α-D-glucopyranoside (PNP-G) were purchased from Sigma (St. Louis, MO, USA). Trizma hydrochloride (Tris-HCl) and bovine serum albumin (BSA) were provided from Sigma-Aldrich (Steinheim, Germany).
Chlorogenic acid, caffeic acid, cynarin, kaempferol and rutin used as standards were purchased from Extrasynthese (Genay, France). HPLC-grade methanol and formic acid were acquired from Merck (Darmstadt, Germany). Water was deionized using a Milli-Q water purification system (Millipore, Bedford, MA, USA).
# Plant material
In this study, dried foliar tissues from fourteen landraces of Smallanthus sonchifolius (named PER01-PER11, ECU41, ECU43 and ECU44, as assigned in a previous studywere analyzed to investigate the phenolic content and biological properties. A voucher specimen for each landrace (numbered CZU-ECU41, ECU43, ECU44 and PERn-11, where n is the number of the landrace as named above) is deposited in herbarium of the Faculty of Tropical Agriculture, Czech University of Agriculture, Prague. The plant material has been acquired since 1993 from different parts of the world and selected for their morphological traits [bib_ref] Total phenolic content, RAPDs, AFLPs and morphological traits for the analysis of..., Milella [/bib_ref]. They are maintained and cultivated under the field conditions of trial plots at the Faculty of Tropical AgriSciences (FTA, former Institute of Tropics and Subtropics), Czech University of Life Sciences Prague (CULS).
## Preparation of plant extracts
Dried leaves of all yacon landraces were coarsely powered by mortar and pestle and extracted by maceration technique with frequent agitation for three times with the same solvent (ratio 1:8 w/v). Solvents with increasing polarity, n-hexane, chloroform (CHCl3), mixture CHCl3:methanol (MeOH) in ratio 9:1 and MeOH were used. The mixtures were filtered by cellulose filter paper and the combined liquids were evaporated to dryness under reduced pressure by using rotary evaporator. The extracts were kept in the dark at room temperature until use. The percentage of yield extracts was calculated as % yield = (Weight of dried extract/Initial weight of dried foliar tissue) × 100.
## Total polyphenol, tannin and flavonoid content
## Total polyphenol content
The total phenolic (TPC) content was determined by Folin-Ciocalteu reagent [bib_ref] Role of the cultivar in choosing clementine fruits with a high level..., Milella [/bib_ref]. Briefly, 75 μL of diluted extract and 425 μL of distilled water was added to 500 μL Folin-Ciocalteu reagent and 500 μL of Na2CO3 (10% w/v). The mixture was mixed and incubated for 1 h in the dark at room temperature. After incubation, the absorbance was measured at 723 nm using a UV-Vis spectrophotometer (DU 640 Spectrophotometer, Beckman, Brea, CA, USA). The total phenolic content was expressed as mg gallic acid equivalent (GAE)/gof extract.
## Total tannin content
To 250 μL of different concentrations of extract (10 mg/mL), 500 μL of bovine serum albumin solution in 0.2 mol/L acetic buffer, pH 5.0 with 0.17 mol/L NaCl (1 mg/mL) was added and mixed carefully [bib_ref] Protein precipitation method for the quantitative determination of tannins, Hagerman [/bib_ref]. After 15 min, the samples were centrifuged at 5000 g for 15 min. The supernatant was removed, and the pellet dissolved in 1 mL of aqueous solution containing 1% SDS and 4% triethanolamine. Then, 250 μL of 0.01 mol/L FeCl3 in 0.01 mol/L HCl was added. After 30 min the absorbance was recorded at 510 nm. Results were expressed as mg of tannic acid equivalent/g of extract (mg TAE/g of extract).
## Total flavonid content
An aliquot (150 μL) of extract at 10.00 mg/mL was added to 45 μL of 5% NaNO3 into microcentrifuge tube. After 5 min, 90 μL of 1% AlCl3 was added and at the 6th minute, 300 μL of 1 M NaOH solution was added and the total volume was made up to 1.5 mL with distilled water. The solution was mixed well and the absorbance was measured against reagent blank at 510 nm after 10 min of incubation at room temperature [bib_ref] Quantification of flavonoid and phenol content from Macrosolen parasiticus (L.) Danser, Lobo [/bib_ref]. Quercetin was used as standard to plot the calibration curve. The total flavonoid content was expressed as mg of quercetin equivalent/g of extract (mg QE/g of extract).
## Biological activity
## Reducing power
Reducing power of the extracts was determined by ferric reducing ability power test (FRAP) [bib_ref] Nutraceutical properties of Citrus clementina juices, Russo [/bib_ref]. The stock solution included 300 mM acetate buffer, pH 3.6, 10 mM TPTZ (2,4,6-tripyridyl-s-triazine) solution in 40 mM HCl, and 20 mM FeCl3·6H2O solution. The working solution was prepared by mixing acetate buffer, TPTZ and FeCl3·6H2O (10:1:1). Plant extracts (150 μL) were allowed to react with 2850 μL of the FRAP solution for 40 min at 37 °C. Readings of the colored product (ferrous tripyridyltriazine complex) were taken at 593 nm. Trolox was used as standard and results were expressed in mg of trolox equivalent (TE)/g of dried sample.
## Dpph scavenging activity
Radical-scavenging activity was evaluated by using DPPH test [bib_ref] Antioxidant and free radical-scavenging activity of constituents from two scorzonera species, Milella [/bib_ref]. Briefly, 80 μL of plant extracts were added to 1420 μL of DPPH (2,2-diphenyl-1-picrylhydrazyl) solution in a microcentrifuge tubes and left in the dark. The absorbance was monitored every 30 min until 90 min at 515 nm by using spectrophotometer (UV 640 Spectrophotometer). Results were expressed as IC50 and compared with the Trolox, used as standard.
## Nitric oxide (˙no) radical scavenging activity
The capacity to scavenge ˙NO was evaluated spectrophotometrically in a Multiskan Ascent plate reader (Thermo Electron Corporation, Shanghai, China), according to a previously described procedure [bib_ref] Bauhinia forficata link authenticity using flavonoids profile: Relation with their biological properties, Ferreres [/bib_ref] , with different concentration of extracts. ˙NO was generated in vitro from sodium nitroprussiate dehydrate (SNP) and measured by the Griess reaction. SNP solution (6 mg/mL) was prepared in phosphate buffer (KH2PO4 100 mM, pH 7.4) and mixed with the same volume (100 μL) of different concentrations of extracts, in a 96-wells plate. The mixture was further incubated at room temperature for 1 h under light. After that, 100 μL of Griess reagent (1:1 mixture (v/v) of 1% sulfanilamide and 0.1% N-(1-naphthyl) ethylenediamine in 2% H3PO4) was added and the mixture was further incubated for 10 min in the dark. The absorbance was read at 560 nm. Results were expressed as IC50 and ascorbic acid was used as positive control.
## Superoxide radical (o2˙−) scavenging activity
The effect of the extract on the superoxide radical-induced reduction of NBT was monitored spectrophotometrically in a Multiskan Ascent plate reader (Thermo Electron Corporation), in kinetic function, at 560 nm. Superoxide radicals were generated by the phenazine methosulfate-β-nicotinamide adenine dinucleotide (PMS-NADH) system, as previously reported [bib_ref] Bauhinia forficata link authenticity using flavonoids profile: Relation with their biological properties, Ferreres [/bib_ref]. Several dilutions of sample (50 μL) or phosphate buffer as control, NADH (50 μL) and NBT (150 μL) were put in the 96-well plate. The reaction was started by adding PMS (50 μL) to the mixture. The assay was conducted at room temperature at 560 nm for 2 min. For each extract, five different concentrations were tested. Results were expressed as IC50 and ascorbic acid was used as positive control.
## Lipid peroxidation inhibition (loo˙ radical)
Lipid peroxyl radical (LOO˙) was generated as proposed by Ferreres et al. [bib_ref] Phlorotannin extracts from fucales characterised by HPLC-DAD-ESI-MS n : Approaches to hyaluronidase..., Ferreres [/bib_ref] with slight modifications. The reaction mixture contained 250 μL of 5 mM linoleic acid, 150 μL of Tris-HCl (100 mM, pH 7.5), 50 μL of 4 mM FeSO4·7H2O and 50 μL of serial dilutions of extracts prepared in distilled H2O. Linoleic acid peroxidation was initiated by the addition of 50 μL of 5 mM ascorbic acid, and the mixture was immediately incubated for 1 h at 37 °C. After the incubation period, 3 mL of ethanol was added to each test tube. The mixtures were vortexed and the absorbance was immediately measured at 233 nm in a Helios α (Unicam) spectrophotometer, at room temperature. Results were expressed as IC50 and BHT was used as positive control.
3.5.6. α-Amylase and α-Glucosidase Inhibitory Activity Different concentration of each sample extract (100 μL) and 100 μL of 0.02 M sodium phosphate buffer (pH 6.9 with 0.006 M NaCl) containing α-amylase solution (0.5 mg/mL) were incubated at 25 °C for 10 min. After pre-incubation, 100 μL of a 1% starch solution in sodium phosphate buffer was added to each tube at timed intervals. The reaction mixtures were then incubated at 25 °C for 10 min. The reaction was stopped with 200 μL of dinitrosalicylic acid color reagent. The test tubes were then incubated in a boiling water bath for 5 min and cooled to room temperature. The reaction mixture was then diluted after adding 3 mL of distilled water, and absorbance was measured at 540 nm. The absorbance of blanks (enzyme solution was added during the boiling) and a control (buffer in place of sample extract) were recorded [bib_ref] Phenolic compounds, antioxidant activity and in vitro inhibitory potential against key enzymes..., Ranilla [/bib_ref]. Analyses were performed in triplicate and the final extract absorbance (540 nm) was obtained by subtracting its corresponding sample blank reading.
The effect on α-glucosidase was assessed in 96-well plates, using a procedure previously reported [bib_ref] In vitro studies to assess the antidiabetic, anti-cholinesterase and antioxidante potential of..., Vinholes [/bib_ref]. Briefly, each well contained 2.5 mM PNP-G (100 μL) in phosphate buffer pH 7.0 (150 μL) and methanol extract at different concentrations (50 μL). The reaction was initiated by the addition of 0.28 U/mL enzyme (20 μL) and the plates were incubated at 37 °C for 10 min. The absorbance at 400 nm was measured in a Multiskan Ascent plate reader (Thermo Electron Corporation). Results were expressed as IC50 calculated from three independent tests, performed in triplicate and acarbose was used as positive control in both assays.
## Acetylcholinesterase (ache) and butyrylcholinesterase (bche) inhibitory activity
The inhibition of AChE activity was determined based on Ellman's method, as previously reported [bib_ref] In vitro studies to assess the antidiabetic, anti-cholinesterase and antioxidante potential of..., Vinholes [/bib_ref]. In this assay, 25 μL of acetylthiocholine iodide (15 mM), 125 μL of DTNB (3 mM), 25 μL of buffer B (50mM Tris-HCl, pH 8 containing 0.1% BSA) and 50 μL of each test extract solution at the different concentrations were mixed. The mixture was monitored at 405 nm. The reaction was started by adding 25 μL of 0.44 U/mL AChE. The absorbance was measured at 405 nm and the rates of reactions were calculated by Ascent Software version 2.6 (Thermo Labsystems Oy, Vantaa, Finland). The BChE inhibition assay was performed in a similar way [bib_ref] Caffeic acid derivatives in the roots of yacon (Smallanthus sonchifolius), Takenaka [/bib_ref] using 25 μL of 15 mM butyrylthiocholine chloride as substrate and 0.1 U/mL of BChE as enzyme. Three independent assays were performed in triplicate at different concentrations.
## Hplc-dad qualitative and quantitative analyses
Redissolved methanol extract (50 mg/mL) was analyzed on an analytical HPLC-DAD unit (Gilson) using a Luna C18 column (250 × 4.6 mm, 5 μm particle size; Phenomenex, Macclesfield, UK). The mobile phase consisted of two solvents: water-formic acid (5%) (A) and methanol (B), starting with 5% B and using a gradient to obtain 15% B at 3 min, 25% B at 13 min, 30% B at 25 min, 35% B at 35 min, 45% B at 42 min, 55% B at 47 min, 75% B at 56 min and 100% B at 60 min. The flow rate was 0.9 mL/min. Spectral data from all peaks were collected in the range of 200-400 nm, and chromatograms were recorded at 320 nm for hydroxycinnamic acids and at 350 nm for flavonoids. The data were processed on Unipoint System software (Gilson Medical Electronics, Villiers le Bel, France).
Whenever available, reference standards of phenolics were used to substantiate the identification of the peaks. Phenolic compounds, which were not available as standard reference materials were tentatively identified according to the literature [bib_ref] Caffeic acid derivatives in the roots of yacon (Smallanthus sonchifolius), Takenaka [/bib_ref] and confirmed by matching their retention time and UV spectra using the same chromatographic condition of a previous work conducted by our team [bib_ref] Antioxidant compounds extracted from several plant materials, Seabra [/bib_ref].
Phenolic compounds quantification was achieved by the absorbance recorded in the chromatograms relative to external calibration standards.
Phenolic acid derivatives (320 nm) were quantified as caffeic acid, 1,5-di-O-caffeoylquinic, 3,5-di-O-dicaffeoylquinic acid and 4,5-di-O-caffeoylquinic acid were quantified as cynarin (1,3-O-dicaffeoylquinic acid, 320 nm), flavonoid derivatives were quantified as kaempferol (350 nm). The other compounds were quantified as themselves.
# Statistical analysis
A minimum of three independent experiments were carried out unless otherwise specified. Results are presented as mean ± standard deviation (Mean ± SD) by using Microsoft Excel. Value of p < 0.05 was considered statistically significant. Calibration curves of the standards were considered as linear if R 2 > 0.99. The comparison among all chemical methods used to investigate antioxidant activity was used to determine relative antioxidant capacity index (RACI). Pearson's correlation coefficient was used to determine the relation between the variables by using Microsoft Excel. For more detailed insight in the relations between the variables, results were submitted to multivariate principal component analysis (PCA) using XLSTAT Version 2015.1 (Addinsoft Inc., New York, NY, USA).
# Conclusions
This study supports the traditional use of S. sonchifolius to treat several ailments. Experimental studies of methanol extract of leaf tissue exhibited considerable antidiabetic and antioxidant activities. Significant differences between individual landraces were found and the results of this study showed that yacon could have a possible application in the pharmaceutical and nutraceutical fields, being a rich source of many bioactive compounds. The selection and propagation of specific landraces, supported by appropriate extraction procedure, would ensure greater concentration of active components and, consequently, a higher biological activity.
Our approach used some assays regarding reactive species and enzymes with biological significance (e.g., nitric oxide and superoxide anion radical scavenging activity and the cholinesterase inhibition), and results are reported here for the first time. Moreover, our HPLC protocol allowed the identification of 1,5-di-O-caffeoylquinic acid, which was not previously reported in yacon leaves.
The specific compounds responsible for yacon biological activities need to be found and further investigations for the most active compounds will be done in the near future.
[fig] Figure 1: (a) α-Amylase and α-glucosidase inhibition activity and (b) acetylcholinesterase (AChE) and butyrylcholinesterase (BChE) inhibition of tested yacon extracts. [/fig]
[fig] Figure 2: Cont. [/fig]
[fig] Figure 3: Relative antioxidant capacity index of investigated yacon landraces. [/fig]
[fig] Figure 4: Principal component analysis (PCA) among biological activity and chemical profile of investigated yacon landraces. [/fig]
[table] Table 2: HPLC quantification of identified compounds in yacon landraces. [/table]
|
Volumetric Assessment of the Frontal Sinus in Female Adolescents and Its Relationship with Craniofacial Morphology and Orthodontic Treatment: A Pilot Study
# Introduction
The paranasal sinuses, the hollow spaces in the craniofacial bones around the nose, are air-filled cavities within the frontal, ethmoidal, sphenoidal, and maxillary bones [bib_ref] Prenatal development of the maxillary sinus: A perspective for paranasal sinus surgery, Nuñez-Castruita [/bib_ref]. All the sinuses drain into the superior and lateral aspects of the nose [bib_ref] Normal anatomy and anatomic variants of the paranasal sinuses on computed tomography, Vaid [/bib_ref] , and the lining mucosa of the sinuses connects to the nasal cavity. Paranasal sinuses are complex anatomical structures, characterized by highly variable shapes, morphologies, and sizes.
The paranasal sinuses occupy a significant amount of space in the cranium and have been of interest in studies that seek to determine their function and the factors affecting their structure and size. The paranasal sinuses have essential functions in immune defense and air filtration processes carried out by the nose, and the walls of the sinus cavities are lightly coated with mucus, which keeps the tissue moist and healthy and also traps bacteria. Furthermore, Preuschoft et al. [bib_ref] Pneumatized spaces, sinuses and spongy bones in the skulls of primates, Preuschoft [/bib_ref] reported that the paranasal sinuses develop in response to the biomechanical requirements of the skull architecture. Thus, the magnitude and direction of the mastication forces, which are major contributors in mechanical stress induction, may be of great importance for paranasal sinus development.
The frontal sinus is the most complex of the paranasal sinuses owing to its location, anatomical variations, and multiple clinical presentations. The frontal sinuses are absent Participants who underwent conventional orthodontic treatment at the Yamada Orthodontic Office between January 2010 and December 2019 were recruited for this study. Informed consent was obtained from all the participants after a full explanation of the research purposes and procedures. The exclusion criteria were as follows: history of seasonal allergies; ear, nose, and throat-related diseases; hormonal disturbances; any deformity or disease in the craniofacial region; and previous orthodontic treatment. This study was approved by the Ethics Committee of Tokushima University Hospital (approval no. 3900).
The present study estimated the necessary sample size to meet the desired statistical constraints. The effect size was used for convenient statistical parametric and nonparametric tests. The effect size of the comparison among the three subgroups divided according to the maxillomandibular jaw-base relationship was considered medium (0.25). The statistical power (1-β) was calculated using G*Power software. Power analysis was based on one-way or two-way repeated measurement analysis of variance (r-ANOVA) with a medium effect size of 0.25, significance level (Type I error) of 0.05, and power level of 0.8. Power analysis was performed and the total estimated sample size was determined to be 42 when a two-way r-ANOVA was used.
For all the patients, 3D CT was performed before orthodontic treatment using a CT system (Alphard-3030, Asahi Roentgen Ind. Co., Ltd., Kyoto, Japan) with the following acquisition parameters: 60-110 kV; 3-15 mA; collimation, 0.6 mm; rotation time, 18 s; and reconstruction thickness, 0.39 mm. Imaging data were processed using Dolphin Imaging (Dolphin Imaging & Management Solutions, Verona, Italy) for orthodontic diagnosis and treatment planning. Using a series of CT DICOM data, a 3D model of the frontal sinus and volume-rendered images were extracted. CT was also performed after active orthodontic treatment. Furthermore, lateral cephalograms were obtained before and after orthodontic treatment using a cephalometric radiographic system (Hyper-X CM, Asahi Roentgen Ind. Co., Ltd., Kyoto, Japan). All cephalograms were obtained with the teeth in the intercuspal position. Briefly, the participant's head was fixed with ear rods and stabilized in a position such that the Frankfort horizontal plane was parallel to the floor.
Each lateral cephalogram was traced on acetate paper by one examiner (H.Y.). The tracings were computerized using a graphic digitizer (Dolphin Imaging, Dolphin Imaging & Management Solutions, Verona, Italy) by another examiner (M.S.) to obtain measurements of the craniofacial morphology. The accuracy of the tracing was confirmed by two orthodontic experts who joined this study as collaborators. All the investigators were blinded to the participants' general status. Before taking the measurements, the intra-examiner reliability of cephalometric analysis was determined using the intraclass correlation coefficient (ICC) on 20 randomly selected cephalograms that were traced and plotted with three arbitrary points (nasion, sella, and pogonion points) by the same examiner twice within an interval of one week. As a result, the ICC was 0.984, confirming a sufficient reliability of the selected measurements.
## Craniofacial morphology
Based on the cephalometric measurements, the participants were divided into the following three subgroups according to the ANB angle (angle between the nasion-A-point and nasion-B-point lines); skeletal Class III group (participants with ANB angle < 1.0 - ), skeletal Class I group (ANB angle ≥ 1.0 - but < 5.0 - ), and skeletal Class II group (ANB angle > 5.0 - ). From the lateral cephalogram, 13 angular and 10 linear measurements were analyzed for morphometric evaluation .
## Frontal sinus morphology
The 3D models constructed from CT Digital Imaging and Communications in Medicine (DICOM) data were analyzed using three-dimensional viewing software to automatically measure the maximum breadth, height, and depth of the frontal sinus [fig_ref] Figure 1: A representative image of the frontal sinus [/fig_ref]. The sinus
## Frontal sinus morphology
The 3D models constructed from CT Digital Imaging and Communications in Medicine (DICOM) data were analyzed using three-dimensional viewing software to automatically measure the maximum breadth, height, and depth of the frontal sinus [fig_ref] Figure 1: A representative image of the frontal sinus [/fig_ref]. The sinus volume was determined as an integral volume of the air cavity within the bony walls of the sinus in the frontal bone on the reformatted axial, sagittal, and coronal images. Volume-rendering images were used for the automatic calculation of the frontal sinus volume. The maximum width was measured between the most lateral points of the frontal sinus. The maximum height was measured between the baseline and the highest point of the frontal sinus. The maximum breadth was defined as the length between the most prominent point of the anterior and posterior parts of the frontal sinus. Furthermore, using a 3D frontal sinus model, morphological variations, including bilateral or unilateral, symmetrical or asymmetrical, fusion or separation, and presence or absence of changes in shape, were analyzed. According to the classification of previous studies [bib_ref] Identification using frontal sinus by three-dimensional reconstruction from computed tomography, Kim [/bib_ref] [bib_ref] Morphological analysis of three-dimensionally reconstructed frontal sinuses from Chinese Han population using..., Zhao [/bib_ref] , the shape of the frontal sinus in an anterior view was divided into three types: fan-shaped, quadrangular, and irregular. The term "separation" was defined if the frontal sinus was divided into two or more segments, while the term "fusion" denoted the frontal sinus connecting with two or more segments.
# Statistical analysis
Statistical analyses were performed using SPSS 27.0 (SPSS Inc., Chicago, IL, USA). The normality of each morphometric variable was assessed using the Shapiro-Wilk test. The average size and volume of the frontal sinus before orthodontic treatment were calculated for each subgroup according to the ANB angle. In addition, the differences in the sinus size and volume between the pretreatment and posttreatment stages were evaluated and compared among the three subgroups. Morphological features of the frontal sinus
# Statistical analysis
Statistical analyses were performed using SPSS 27.0 (SPSS Inc., Chicago, IL, USA). The normality of each morphometric variable was assessed using the Shapiro-Wilk test. The average size and volume of the frontal sinus before orthodontic treatment were calculated for each subgroup according to the ANB angle. In addition, the differences in the sinus size and volume between the pretreatment and posttreatment stages were evaluated and compared among the three subgroups. Morphological features of the frontal sinus were verified by subgroup analysis using the Fisher's exact test. For data with normal distribution, a general linear model analysis for repeated measures was performed to compare the three subgroups. Intergroup comparisons were performed using the paired t-test with the Bonferroni method as a post-hoc test. A linear single regression test was performed to detect the relationships between the morphological variables of the frontal sinus and the cephalometric measurement variable for each subgroup. Moreover, multiple regression analysis, including the morphological variables regardless of the subgroups, was performed to assess the relationships of the morphological variables as a response variable and the cephalometric measurement variables as explanatory variables. Probabilities below 0.05 as type I error (α) were considered statistically significant.
# Results
## Participants
The total sample size used in the present study was 53. The participants were divided into three subgroups: skeletal Class I, 20 females ranging from 11.9 to 17.3 years (mean age ± standard deviation [SD]: 13.9 ± 1.3 years); skeletal Class II, 20 females ranging from 11.7 to 16.4 years (mean age ± SD: 13.9 ± 1.4 years); and skeletal Class III, 13 females ranging from 10.3 to 15.7 years (mean age ± SD: 13.4 ± 1.6 years). There was no significant difference in age among the three subgroups (p = 0.53, one-way ANOVA). The duration of orthodontic treatment was 3.8 ± 1.1 years.
## Volumetric and geometric measurements of the frontal sinus and its relation with the skeletal pattern
Considering the morphological features of the frontal sinus, no patients exhibited agenesis of the frontal sinus. Of the 53 patients, most of the patients (98.1%) had a bilateral frontal sinus without any fusion [fig_ref] Table 2: Morphological features of the frontal sinus [/fig_ref]. Considering the shape of the frontal sinus, 42 patients (79.2%) exhibited symmetry, while 11 patients (20.8%) showed asymmetry. There were no statistically significant differences among the three subgroups (p = 0.43). Furthermore, 38 patients (71.7%) showed a fan-shaped sinus, followed by irregular (18.9%), and quadrangular sinuses (9.4%). The morphological shapes did not differ significantly among the three subgroups (p = 0.92). The morphological characteristics of the frontal sinus were not associated with the maxillomandibular jaw-base relationship (skeletal pattern), indicating no specific features of the frontal sinus (p = 0.37). Before orthodontic treatment, the breadth, height, and depth of the frontal sinus were 46.2 ± 12.5 mm (Mean ± SD), 29.8 ± 8.0 mm, and 22.8 ± 5.1 mm in the skeletal Class I; 44.8 ± 10.5 mm, 31.0 ± 6.2 mm, and 22.6 ± 4.9 mm in the skeletal Class II; and 46.5 ± 15.7 mm, 27.8 ± 8.2 mm, and 22.7 ± 6.2 mm in the skeletal Class III malocclusion, respectively [fig_ref] Table 3: The frontal sinus size and volume for each subgroup before and after... [/fig_ref]. No significant differences in the size of the frontal sinus were observed among the three subgroups. The volume of the frontal sinus was 4986.3 ± 2849.2 mm 2 , 5143.3 ± 2398.9 mm 2 , and 5418.9 ± 3221.9 mm 2 in the skeletal Class I, II, and III groups, respectively. No significant difference in the sinus volume was observed among the three subgroups (p > 0.63); however, the sinus volume in the skeletal Class III group tended to be larger than that in the remaining two subgroups.
By comparing pretreatment and posttreatment measurements, considerably large changes in the size and volume of the frontal sinus were observed following orthodontic treatment, regardless of the skeletal pattern [fig_ref] Table 3: The frontal sinus size and volume for each subgroup before and after... [/fig_ref]. Although there was no significant interaction among the three subgroups, the breadth, height, depth, and volume of the frontal sinus significantly (p < 0.001) increased during orthodontic treatment, regardless of the skeletal classification.
## Correlation between the craniofacial morphology and frontal sinus morphology
The standardized coefficients were calculated for four measurement values of the frontal sinus and 23 cephalometric variables using single regression analysis. Among the 92 correlations for all the participants, most coefficients had very weak or no correlations without statistical significance between pretreatment and posttreatment measurements [fig_ref] Table 4: Standardized coefficients between the measurement values for the frontal sinus in relation... [/fig_ref]. For the participants with skeletal Class I malocclusion, 2 of 92 correlations showed weak but significant (p < 0.05) correlations with sinus volume and FMA in the pretreatment measurements, and with sinus height and IMPA in the posttreatment measurements [fig_ref] Table 4: Standardized coefficients between the measurement values for the frontal sinus in relation... [/fig_ref]. For the participants with skeletal Class II, 4 of 92 correlations had significantly (p < 0.05, p < 0.01) negative correlations with sinus breadth and volume and palatal plane to FH, sinus width and L1-NB, and sinus height and overbite, and a significant positive correlation with sinus width and FMIA in the posttreatment measurements; however, no significant correlations were found for the pretreatment measurement variables [fig_ref] Table 4: Standardized coefficients between the measurement values for the frontal sinus in relation... [/fig_ref]. For the participants with skeletal Class III malocclusion, 11 of 92 correlations revealed weak or mild but significant (p < 0.05, p < 0.01) correlations with sinus breadth, width, and volume and FMA; sinus breadth, width, and volume and IMPA; sinus width, height, and volume and gonial angle; sinus width and Y-axis; and sinus width and facial angle at the pretreatment measurements [fig_ref] Table 4: Standardized coefficients between the measurement values for the frontal sinus in relation... [/fig_ref]. Furthermore, in the posttreatment measurements, significant (p < 0.05, p < 0.01) positive correlations were found between sinus breadth, width, height, and volume and Y-axis; sinus breadth, width, height, and volume and FMA; and sinus volume and gonial angle. Three correlations had significant (p < 0.05) negative correlations with sinus breadth, sinus volume, and IMPA; and sinus width and facial angle. Multiple regression analysis was applied to verify the relationships of the frontal sinus size and volume with the horizontal skeletal pattern as a qualitative variable, and the cephalometric measurement variable as a quantitative variable, in this study. However, no significant relationship was observed between the skeletal pattern and sinus size and volume. Thus, the difference in the skeletal pattern did not affect the sinus size and volume.
# Discussion
With technological advancements in CT, growing evidence suggests the critical role of paranasal sinuses in craniofacial growth and orthodontic treatment [bib_ref] Morphological analysis of three-dimensionally reconstructed frontal sinuses from Chinese Han population using..., Zhao [/bib_ref] [bib_ref] Volumetric analysis of the maxillary, sphenoid and frontal sinuses: A comparative computerized..., Cohen [/bib_ref] [bib_ref] Paranasal sinus development and implications for imaging, Goldman-Yassen [/bib_ref] [bib_ref] Controversy about the protective role of volume in the frontal sinus after..., Celiker [/bib_ref] [bib_ref] Correlation between the upper airway volume and the hyoid bone position, palatal..., Dastan [/bib_ref] [bib_ref] Clinical significance of maxillary sinus hypoplasia in dentistry: A CBCT study, Dedeoglu [/bib_ref] [bib_ref] Comparison of dimensions and volume of upper airway before and after mini-implant..., Li [/bib_ref] [bib_ref] Morpholetric characteristics of the sphenoid sinus and potential influencing factors: A retrospective..., Singh [/bib_ref]. In the present study, the three-dimensional size and volume of the frontal sinus were measured using CT images taken before orthodontic treatment in female adolescents. The average size and volume of the frontal sinus before orthodontic treatment were 45.8 ± 12.3 mm in breadth, 29.8 ± 7.3 mm in height, 22.7 ± 5.1 mm in depth, and 5151.6 ± 2711.4 mm 2 in volume. These values were nearly consistent with those of previous studies [bib_ref] Volumetric analysis of the maxillary, sphenoid and frontal sinuses: A comparative computerized..., Cohen [/bib_ref] [bib_ref] Volumetric evaluation of the paranasal sinuses in normal subjects using computer tomography..., Emirzeoglu [/bib_ref]. Furthermore, no age-related differences in the frontal sinus size and volume were found in the present study, similar to previous studies [bib_ref] Volumetric analysis of the maxillary, sphenoid and frontal sinuses: A comparative computerized..., Cohen [/bib_ref]. Using radiographic examination of the frontal sinus, Brown et al. [bib_ref] Enlargement of the frontal sinus, Brown [/bib_ref] reported that the main expansion of the sinus ceased at the age of 15.68 years in males and 13.72 years in females. On the other hand, growth ceases at approximately 20 years of age, when the shape and size of the frontal sinus become stable [bib_ref] Paranasal sinus development and implications for imaging, Goldman-Yassen [/bib_ref]. Considering the mean age of our patients was 13.8 ± 1.4 years before orthodontic treatment, the size and volume of the frontal sinus were assumed to already be at their largest.
Meanwhile, our results showed that the three-dimensional sizes and volume of the frontal sinus increased significantly during orthodontic treatment, while the morphological features of the sinus were not completely changed. As described above, since the mean age of our patients was 13.8 ± 1.4 years before orthodontic treatment, no or minimal changes in the frontal sinus size were found if they had not received orthodontic treatment. Kjaer et al. [bib_ref] Frontal sinus dimensions can differ significantly between individuals within a monozygotic twin..., Kjaer [/bib_ref] reported the importance of environmental factors on frontal sinus dimensions, indicating that medical treatment, including orthodontic and orthognathic treatment, may affect the size of the frontal sinus. This suggests that orthodontic treatment may contribute to the development of the frontal sinus as an environmental factor, even after pubertal growth of the frontal sinus is completed. However, since the increases in the size and volume of the frontal sinus were small in this study, these may have been caused by pubertal growth and not orthodontic treatment. Further studies involving control participants without an experience of orthodontic treatment are needed to examine whether orthodontic treatment affects the dimensions of the frontal sinus.
According to Rae and Koppe [bib_ref] Independence of biomechanical forces and craniofacial pneumatization in Cebus, Rae [/bib_ref] , possible functions of the paranasal sinuses include respiratory function, thermoregulation, and trauma protection to decrease the skull weight. Preuschoft et al. [bib_ref] Pneumatized spaces, sinuses and spongy bones in the skulls of primates, Preuschoft [/bib_ref] reported that the paranasal sinuses develop in response to the biomechanical requirements of the skull architecture. Thus, the magnitude and the direction of the forces of mastication, which are major contributors to mechanical stress induction, are of great importance. Furthermore, Said et al.investigated the relationship between anterior occlusion and frontal sinus size in adolescents using cephalograms, and reported that the frontal sinus size could be used as an indicator for harmonious anterior occlusion. This implies that a larger frontal sinus has favorable functions, such as serving as a shock absorber and in the transmission of occlusal forces. On the other hand, Benington et al. [bib_ref] Masseter muscle volume measured using ultrasonography and its relationship with facial morphology, Benington [/bib_ref] demonstrated that the group with the largest sinus size was the open bite group, which might be attributed to the reduced transmission of occlusal forces along the nasal pillars because of the lack of contact between the maxillary and mandibular incisors, and weaker muscles associated with the hyperdivergent morphology. Furthermore, recently, Celiker et al. [bib_ref] Controversy about the protective role of volume in the frontal sinus after..., Celiker [/bib_ref] investigated the relationship between the size of the frontal sinus and mortality in patients with cranial trauma, and suggested that the larger the sinus, the greater the risk of death resulting from trauma to the head. Our results showed very weak correlation between the three-dimensional size and volume of the frontal sinus and overbite, except a significant, but weak, negative correlation between the sinus height and overbite after the treatment in patients with skeletal Class II malocclusion. Based on the previous studies, a frontal sinus of proper size may have favorable functions for anterior occlusion in adolescents; however, an extraordinarily large sinus could contribute to one of the risk factors for death following severe head injury. Therefore, the decisive role of the frontal sinus remains controversial.
Previously, the association between frontal sinus size and volume and craniofacial morphology was evaluated using lateral cephalograms. Rossouw et al. [bib_ref] The frontal sinus and mandibular growth prediction, Rossouw [/bib_ref] found a correlation between the frontal sinus area on lateral cephalometric radiographs and maxillary length, mandibular length, symphysis width, and condylar length, indicating that the frontal sinus size may be a supplementary indicator for mandibular growth prediction. Benington et al. [bib_ref] Masseter muscle volume measured using ultrasonography and its relationship with facial morphology, Benington [/bib_ref] also demonstrated that anterior cranial base length, facial divergence, and inclination of the maxillary incisor in relation to the palate were statistically significant variables explaining frontal sinus size. Recently, Tehranchi et al. [bib_ref] Correlation between frontal sinus dimensions and cephalometric indices: A cross-sectional study, Tehranchi [/bib_ref] demonstrated that a larger frontal sinus size was associated with reduced inclination of the anterior cranial base, increased anterior facial height in males, and increased gonial angle in females. Furthermore, Yassaei et al. [bib_ref] Cephalometric association of mandibular size/length to the surface area and dimensions of..., Yassaei [/bib_ref] indicated that the dimensions and surface area of the frontal and maxillary sinuses in skeletal Class III malocclusion were greater than those in other groups. These variables (except for frontal sinus width) were significantly correlated with the anterior and posterior cranial bases and mandibular body length. In the present study, we evaluated the correlation between craniofacial morphology and frontal sinus morphology in Japanese female adolescents using 3D CT. No significant correlation was found among the 92 correlations for all the participants. For each group with different skeletal patterns, several cephalometric variables demonstrated weak or mild, but significant correlations with the frontal sinus size and volume. For the participants with a skeletal Class III jaw-base relationship, the values of the Y-axis and FMA were significantly positively correlated with the breadth, width, height, and volume of the frontal sinus, indicating that the longer the facial type, the greater the height of the frontal sinus. Furthermore, our results indicated a frontal sinus that was larger, but not significant, in patients with skeletal Class III malocclusion than in those with skeletal Class I and II malocclusions, which was consistent with the previous results [bib_ref] Cephalometric association of mandibular size/length to the surface area and dimensions of..., Yassaei [/bib_ref] [bib_ref] Correlations between the volume of the sinuses and the facial bones and..., Dah-Jouonzo [/bib_ref]. However, there was no significant difference in frontal sinus size among the participants with the same skeletal classification and different anterior occlusions. Further studies are needed to categorize treatment modalities and to perform more detailed statistical investigations. In addition, a larger number of participants should be included in future studies.
This study had some limitations. First, the study did not include any control samples, that is, participants who did not undergo orthodontic treatment. Our results showed small, but significant, increases in the size and volume of the frontal sinus during orthodontic treatment; however, without the control data, we could not distinguish whether these increases were affected by orthodontic treatment. It is difficult to record 3D CT images for individuals who are not undergoing treatment because of ethical concerns. Second, our participants were only females, and we could not estimate the sex-dependent differences in the dimensions of the frontal sinus. Furthermore, the effect of orthodontic treatment on the sinus dimensions may differ between male and female adolescents. Previous studies also did not report a sex-based difference in the size and volume of the frontal sinus [bib_ref] Paranasal sinus development and implications for imaging, Goldman-Yassen [/bib_ref] ; therefore, further studies involving both male and female participants are needed.
# Conclusions
In conclusion, the present study reported the average size and volume of the frontal sinuses of female adolescents before orthodontic treatment. During orthodontic treatment, the sizes and volume of the frontal sinus increased significantly regardless of the skeletal classification; however, since these changes were small, the increases in the size and volume of the frontal sinus may have been caused by pubertal growth and not orthodontic treatment. Further studies should be conducted to examine the causative effect of orthodontic treatment on the development of the frontal sinus.
[fig] .: Definitions of cephalometric measurement items Angular Measurement Items ( • ) SNA: Angle between the sella-nasion and nasion-A-point lines, indicating the anteroposterior position of the maxilla relative to the anterior cranial base SNB: Angle between the sella-nasion and nasion-B-point lines, indicating the anteroposterior position of the mandible relative to the anterior cranial base ANB: Angle between the nasion-A-point and nasion-B-point lines, indicating an anteroposterior relationship between the maxilla and mandible Facial angle: Angle between the nasion-pognion line and Frankfort horizontal plane, indicating chin prominence Y-axis: Angle between the sella-gnathion line and Frankfort horizontal plane, indicating the mandibular growth direction Gonial angle: Angle between the mandibular and ramal planes FMA: Angle between the mandibular plane and Frankfort horizontal plane, indicating divergence of the mandibular plane Occlusal plane to SN: Angle between the occlusal plane and sella-nasion line, indicating the inclination of the occlusal plane Palatal plane to FH: Angle between the palatal plane and Frankfort horizontal plane, indicating inclination of the palatal plane U1 to SN: Angle between the long axis of the maxillary central incisor and sella-nasion line, indicating labiolingual inclination of the upper incisor Interincisal angle: Angle between the long axes of the upper and lower incisors IMPA: Angle between the long axis of the mandibular incisor and the mandibular plane, indicating the labiolingual inclination of the mandibular central incisor relative to the mandibular plane FMIA: Angle between the long axis of the mandibular central incisor and Frankfort horizontal plane, indicating the labiolingual inclination of the mandibular incisor relative to the Frankfort horizontal plane Linear Measurement Items (mm) SN: Distance between the sella and the nasion, indicating the anteroposterior length of the anterior cranial base U1 to NA: Distance between the incisal edge of the maxillary central incisor and the line joining the A-point to nasion L1 to NB: Distance between the incisal edge of the mandibular central incisor and the line joining the B-point to the nasion Overjet: Anteroposterior distance between the maxillary and mandibular central incisal edges Overbite: Vertical dimension between the maxillary and mandibular central incisal edges Wits appraisal: Perpendicular distance between points A and B on the occlusal plane, indicating the degree of anteroposterior jaw disharmony N-Me: Distance between nasion and menton, indicating anterior facial height Ar-Go: Distance between the articulare and gonion, indicating the length of the mandibular ramus Ar-Me: Distance between the articulare and menton, indicating the effective mandibular length Go-Me: Distance between the gonion and menton, indicating the length of the mandibular corpus [/fig]
[fig] Figure 1: A representative image of the frontal sinus. [/fig]
[fig] Author: Contributions: M.S., H.Y. and E.T. conceived of and designed the study. M.S., H.Y. and M.H. performed the experiments. M.S., S.A. and E.T. analyzed the data. S.A. and E.T. prepared the manuscript. All authors have read and agreed to the published version of the manuscript. [/fig]
[table] Table 2: Morphological features of the frontal sinus. [/table]
[table] Table 3: The frontal sinus size and volume for each subgroup before and after orthodontic treatment. [/table]
[table] Table 4: Standardized coefficients between the measurement values for the frontal sinus in relation to the cephalometric measurement variables analyzed by single regression analysis. (a) All participants, (b) participants with skeletal Class I jaw-base relationship, (c) participants with skeletal Class II jaw-base relationship, and (d) participants with skeletal Class III jaw-base relationship. [/table]
|
Development and tracking of central patterns of subcutaneous fat of rural South African youth: Ellisras longitudinal study
Background: Individuals grow and accumulate central patterns of body fat into the diseases they will suffer from as older adults. The need to elicit the development and tracking of central patterns of body fat from younger age into adolescent remains to be explored.Method: Skinfolds measurements were done according to the standard procedures in the Ellisras Longitudinal Growth and Health Study. In total, 2,225 children--550 preschool and 1,675 primary school--aged 3-10 years (birth cohorts 1993 to 1986) were enrolled at baseline in 1996 and followed through out the eight-year periodic surveys. In 2003, 1,771 children--489 preschool and 1,282 primary school--were still in the study.Results:The development of triceps, biceps, suprailiac and suscapular skinfolds of Ellisras girls were significantly higher (p < 0.001 to 0.05) compared to boys over time. The tracking coefficient between the initial measurements and the subsequent measurements was higher for skinfolds (r about 0.63) than for skinfold ratios (r about 0.43). Longitudinal tracking coefficient measuring the association between the initial measurements and all the follow up measurements simultaneously was about 0.57.Conclusion:The accumulation of central patterns of body fat of Ellisras children starts in childhood and adolescence spurt with Ellisras girls acquiring more than boys over time. High significant tracking of skinfold thickness while the skinfold ratios show low and insignificant tracking over time. The magnitude of central patterns of body fat accumulation over time requires further investigation to clarify their association with risk factors for cardiovascular diseases.
# Background
Individuals grow and develop central patterns of body fat into the diseases they will suffer from as older adults. These accumulated central patterns of body fat in any of the four critical growth stages (1. the intra-uterine period, 2. infancy, 3. mid-childhood and 4. adolescence) may be an independent risk factor for the development of ele-vated blood pressure, clustering of various cardiovascular risk factors in metabolic syndrome, type 2 diabetes, abnormal vascular wall thickness, endothelial dysfunction of left ventricular hypertrophy, high lifetime risk of hypertension, coronary heart diseases, stroke, respiratory problem and some cancers later in their lifetime [bib_ref] Cardiovascular risks associated with obesity in children and adolescents, Ho [/bib_ref] [bib_ref] Critical period in human growth and their relationship to diseases of aging, Cameron [/bib_ref] [bib_ref] Development of fat tissue and body mass index from infancy to adulthood, Gasser [/bib_ref] [bib_ref] Critical period in childhood for the development of obesity, Dietz [/bib_ref] [bib_ref] Period of risk in childhood for the development of adult obesity-What do..., Dietz [/bib_ref]. Briefly, infancy, that is, from second postnatal month to two years, is characterized by high maternal investment, rapid growth particularly of neural tissue, the development of basic independent functional capacity and total dependence of the infant on the mother for survival. Childhood extends from the end of infancy to the start of adolescence growth. Dietz [bib_ref] Critical period in childhood for the development of obesity, Dietz [/bib_ref] characterized the childhood stage as the period of adipose rebound occurring between ages 5-7 years. Heavier adults tend to have an age of adiposity rebound earlier than 6 years of age while lean adults tend to initiate their adiposity rebound after 6 years of age [bib_ref] Adiposity in Czech children followed from 1 moth of age to adulthood:..., Prokopec [/bib_ref]. Adolescence (start at 10 years for girls and 12 years for boys) is initiated by the first appearance of the secondary sex characteristics or pubertal changes [bib_ref] Variation in the pattern pubertal changes in girls, Marshall [/bib_ref]. That is, the breast development in girls (at about 11 years of age), genitalia development (at about 11.5 years) in boys and pubic hair development in both sexes. In view of the rising public health problems of chronic diseases of lifestyle in adulthood, the need to elicit the development and tracking of central patterns of body fat from younger age and beyond is compelling.
Early prediction of obesity risk later in life is an important public health goal given the epidemic of obesity among children and adolescence [bib_ref] Overweight children and adolescent: description, epidemiology and demographics, Troiano [/bib_ref] [bib_ref] Sex differences in body fat of rural South African school children: Implications..., Mantsena [/bib_ref] [bib_ref] Body composition in children and adolescent born after Vitro fertilization or spontaneous..., Manon [/bib_ref]. Hence, tracking or stability of a characteristic is used mostly in relation to the risk factors of the chronic diseases [bib_ref] Development and tracking of central pattern of subcutaneous fat in adolescence and..., Van Lenthe [/bib_ref] [bib_ref] Mathematical and analytic aspect of tracking, Twisk [/bib_ref]. Early detection of these risk factors can lead to the possibility of early treatment. Quantification of the stability of a characteristic over time is important from a public health perspective in a longitudinal research. Its importance is evident in the effectiveness of lifestyle intervention to improve health. If the stability of a characteristic is very high (close to one), the level of this characteristic is usually hard to change, and therefore the interventions that focus on these characteristics are predestined to be ineffective [bib_ref] Review of AGAHLS and other observational Longitudinal Studies on Lifestyle and Health..., Koppes [/bib_ref]. Knowledge of the level of tracking of a characteristic further helps to answer the question whether or not lifestyle interventions should be given to the whole population or to a sub-sample.
The purpose of this study was to describe the development of two trunk and two extremity skinfolds of rural South African preschool children (mean age of 4.9 years at base line to11.5 years) and primary school children (mean age of 8.5 years at baseline to 14.9 years) over a period of eight years (Ellisras Longitudinal growth and Health Study (ELS)). Trunk-extremity skinfold ratios were constructed and their development were highlighted during the same period. In addition, the existence of tracking of skinfold and trunk-extremity skinfold ratios were investigated.
# Methods
## Sample
The ELS initially followed a cluster sampling method [bib_ref] Obesity: does it occur in African children in a rural community in..., Monyeki [/bib_ref] [bib_ref] Growth and nutritional status of rural South African children 3-10 years old:..., Monyeki [/bib_ref]. In brief, the study was undertaken at 22 schools (10 pre-school and 12 primary schools) randomly selected from 68 schools within the Ellisras area. Birth records were obtained from the school admission register through the assistance of principals in each school. Only those records that were verified against health clinic records were used to determine the ages of potential participants. Each of the 22 selected schools was assigned a grade with the expectation that most of the children in a particular age category (i.e. 3,4, ... 9,10) would be found in that grade.
For the purpose of this analysis, data collected in [bib_ref] Nutrition transition in Morocco, Benjelloun [/bib_ref] [bib_ref] Period of risk in childhood for the development of adult obesity-What do..., Dietz [/bib_ref] were included. A total of 2225 (550 preschool children mean age 4.4 years SD = 0.99 and 1675 primary school children mean age 8.0 years SD = 1.11) at baseline were followed throughout the periodic surveys. On average 1.05% of participants were permanently lost due to death and 11.47% subjects lost due to teenage pregnancy, illness, migration to urban areas and school drop-out were a temporary issue as the affected participants rejoined the study thereafter. A total of 1771 subjects (489 preschool children mean age 11.4 years SD = 0.96 and 1282 primary school children mean age 14.9 years SD = 1.11) were measured in November 2003.
## Anthropometric measurements
All children underwent skinfolds measurements (biceps, triceps, subscapular, suprailiac), suggested by the International Society for the Advancement of Kinanthropometry (ISAK). A Harpenden (John Bull) skinfold calipers with inter-jaw pressure of 10 g/mm 2 surface jaw face area for skinfolds measurements to the last completed 0.1 mm was used. For the indicators of central patterns of body fat the following ratios, contrasting subcutaneous fat on the trunk with fat on the extremities were used [bib_ref] Development and tracking of central pattern of subcutaneous fat in adolescence and..., Van Lenthe [/bib_ref] [bib_ref] Longitudinal analysis adolescent growth in height, fatness and fat pattering in rural..., Cameron [/bib_ref] :
## Maturity
The maturation assessment was included in the anthropometric survey of May 2001 and 2003 for all the children who were part of the ELS. The May 2003 assessment was included in the analysis. Breast development and genital/ pubic hair development stages were assessed by visual inspection using Tanner rating scale pictures ranging from 1 (no development) to 5 (matured stage). To reduce embarrassment, older children were provided with a separate private space to complete the self assessment. Once completed, self assessment of the Tanner scale was verified by visual inspection at the "skinfold measurements" station. In instances where the average breast score was between two breast stages, the breast stage was rounded ST-ratio Subscapular triceps SST tratio Subscapular subsca = = / /( p pular triceps SSTB ratio Subscapular Suprailiac biceps
[formula] + = + + ) ( [/formula]
) / t triceps subscapular suprailiac + + down because the higher breast stage has not been achieved. The palpation of the breast which is the superior method to assess breast development was not possible in this study. This was because it was conducted in a class room as it was included in the "skinfold measurements" station of the anthropometric survey. The qualitative Tanner score was converted into a quantitative variables (pubertal stage by Tanner Scale of both the sexual organ and the breast development): T1 = 0, T2 = 0, T3 = 1, T4 = 2, T5 = 3.
## Quality control
The survey was carried out over a three-week period by [bib_ref] Obesity: does it occur in African children in a rural community in..., Monyeki [/bib_ref] anthropometrists each year, who were required to undertake reliability testing as part of their training. This training was conducted by a level three criterion of ISAK following the guidelines of Norton and Olds. The absolute and relative values for intra-tester and inter tester technical error of measurements (% TEM) for all the skinfolds measurements ranged from 0.2 to 6 mm (0.4 to 6.8%) each year.
Maturational status was assessed by well trained field workers stationed at the "skinfold measurements" station. The intra-and inter-tester reliability conducted on 20 subjects (10 boys and 10 girls) who were not part of the survey was 100% in agreement on pubic hair and 92% on breast development.
## Ethics
The Ethics Committee of the University of Limpopo granted ethical approval prior to the survey and the parents or guardians provided informed consent.
# Statistical analysis
Descriptive statistics of the development of fat pattern variables were reported over time. Mann-Whitney U t-test was applied to test the significance differences between sexes.
Partial correlation coefficients controlled for maturation and age-were calculated to assess the association between the first fat pattern variables measurements and the follow-up measurements for boys and girls separately. Linear regression model was used to assess the relationship between fat pattern variables at the first measurements and the follow-up measurements, adjusted for age and maturation for boys and girls separately.
A longitudinal tracking (Generalized Estimating Equation (GEE) technique which measures the association between an indicator at the first period of measurements and the same indicator at all other periods of measurements was used with maturation and age being included in the model [bib_ref] Longitudinal data analysis for discrete and continuous outcomes, Zeger [/bib_ref] [bib_ref] Tracking of risk factors for coronary heart disease over a 14 year..., Twisk [/bib_ref] [bib_ref] Generalized estimating equations for correlated binary data using the odds ratio as..., Lipsitz [/bib_ref]. The statistical significance level was set at p < 0.05. All the statistical analyses were done using the Statistical Package for the Social Sciences (SPSS) and the STATA program.
# Results
To examine the effects caused by the subjects who were absent, we compared skinfold thickness with the paired follow up subjects during each period of measurement. There was no significant difference (p < 0.05) between subjects of the same age who were currently in the study and the drop-outs. Thus, dropout at this stage seems to have been random. Interestingly, to examine the effect of overlapping ages for skinfold thickness of preschool and primary school children we found a significant (p < 0.001) difference at a younger mean age (mean age 7.9 to 9.6 years) for boys while there was no distinct pattern of significant mean skinfold thickness difference for girls across the overlapping ages with the majority of overlapping ages showing no significant difference [fig_ref] Table 1: Differences in the descriptive statistics for overlapping mean ages# [/fig_ref]. [fig_ref] Figure 1: Development of median triceps skinfold of Ellisras rural children and NHANES III [/fig_ref] and 2 presents the development of the median triceps and subscapular skinfold thickness of Ellisras rural children compared to the reference population (National health and nutrition examination surveys III). The median triceps skinfold thickness of Ellisras rural children was significantly (p < 0.001 to 0.05) higher for girls than boys in both the preschool and primary school children through-out the time span [fig_ref] Figure 1: Development of median triceps skinfold of Ellisras rural children and NHANES III [/fig_ref]. The development of median suscapular skinfold of Ellisras primary school girls was significantly higher (p < 0.001 to 0.05) compared to boys while preschool girls showed a significantly (p < 0.001 to 0.05) higher median subscapular skinfold than boys between the mean ages of 6.9 years to 10.9 years . shows the development of median subscapular/triceps skinfold ratio of Ellisras rural children. Primary school boys showed a significant high (p < 0.001 to 0.05) skinfold ratios compared to girls while in the preschool children no clear pattern was observed .shows a specific tracking coefficient derived from partial correlation controlled for age and maturation between fat pattern variables values at the first measurement and the subsequent measurements. The tracking coefficient for triceps, biceps and subscapular skinfolds was high (r range from 0.20 to 0.82) and significant (p < 0.05 and 0.001) from mean ages 4.9 to 8.5 years for preschool children and 8.5 to 12.0 years for primary school children. For pre school children (aged 8.9 to 11.5 years) and primary school children (aged 12.5 to 14.9 years) the triceps, subscapular and biceps skinfolds tracking coefficient was low at times insignificant (r = -0.01 to 0.46). The tracking coefficient for the skinfold ratios was generally low and in most of the times insignificant (r = -0.01 to 0.58) for both preschool and primary school children ( [fig_ref] Table 3: shows regression coefficient, 95% confidence interval and p-value in the association of... [/fig_ref].
# Discussion
In this study the development and tracking of central patterns of subcutaneous fat of rural South African children aged 5 to 15 years was presented. A significant high subcutaneous fat (triceps, biceps, subscapular and suprailiac skinfold) was observed for girls compared to boys over time. Primary school boys exhibit high skinfold ratios than girls over time. To assess the stability of certain variables in time or assess the predictive value of variables which are measured in early life, the computation of tracking coefficients are considered to be critical in longitudinal epidemiological studies. Recommendations for interpreting tracking correlations are as follow: <0.3 = low, 0.3 to 0.6 = moderate and > 0.6 = high [bib_ref] Tracking of physical activity and physical fitness across the lifespan, Malina [/bib_ref] [bib_ref] Tracking of health-related fitness components in young ages 9 to 12 years, Marshall [/bib_ref]. Based on these recommendations, the results of this study suggest that skinfold measurements demonstrated a moderate tracking while skinfold ratios settled for a low tracking of both Ellisras rural preschool and primary school children.
Primary school children in the present study show low development of median triceps and subscapular values through out the age range compared to the reference population (National health and nutrition examination surveys III) [fig_ref] Figure 1: Development of median triceps skinfold of Ellisras rural children and NHANES III [/fig_ref]. There is no clear pattern between preschool children and reference population in the development of subscapular skinfold through out the mean age range while for the development of median triceps skinfold for both boys and girls was high through out the age span compared to the Ellisras pre school sample [fig_ref] Figure 1: Development of median triceps skinfold of Ellisras rural children and NHANES III [/fig_ref]. The development of central patterns of body fat in the Ellisras children started in both childhood and adolescent spurt and was consistent with findings from other studies [bib_ref] Critical period in childhood for the development of obesity, Dietz [/bib_ref] [bib_ref] Adiposity in Czech children followed from 1 moth of age to adulthood:..., Prokopec [/bib_ref] [bib_ref] Sex differences in body fat of rural South African school children: Implications..., Mantsena [/bib_ref] [bib_ref] Body composition in children and adolescent born after Vitro fertilization or spontaneous..., Manon [/bib_ref] [bib_ref] Development and tracking of central pattern of subcutaneous fat in adolescence and..., Van Lenthe [/bib_ref] [bib_ref] Longitudinal analysis adolescent growth in height, fatness and fat pattering in rural..., Cameron [/bib_ref] [bib_ref] Principles of growth standards, Tanner [/bib_ref] [bib_ref] Tracking body fat distribution during growth: using measurements at two occasion vs..., Mueller [/bib_ref] [bib_ref] Factors affecting tracking of coronary heart disease risk factors in children. The..., Mahoney [/bib_ref]. Furthermore, similar to the current study the skinfold ratio (suscapular/subscapular + triceps skinfold) was consistently higher for boys compared to girls through out the childhood and adolescent spurt [bib_ref] Longitudinal analysis adolescent growth in height, fatness and fat pattering in rural..., Cameron [/bib_ref] [bib_ref] Tracking of fat pattern indices in childhood: Melbourne Growth Study, Baumgartner [/bib_ref] [bib_ref] Siervogel RM: Body mass index during childhood, adolescence and young adulthood in..., Guo [/bib_ref]. In contrast, Cronk et al [bib_ref] Changes in triceps and subscapular skinfold thickness during adolescent, Cronk [/bib_ref] , Malina and Bouchard [bib_ref] Subcutaneous fat distribution during growth, Malina [/bib_ref] and Koziel and Malina [bib_ref] Variation in relative fat distribution associated with maturational timing: The Wroclaw Growth..., Koziel [/bib_ref] reported a small increase in subscapular thickness while triceps skinfold did not show an increase during the adolescent spurt.
There are a few longitudinal studies in which the stability coefficient for subcutaneous fat variables was assessed over more or less the same length where the current study was carried out. In the Amsterdam Longitudinal Growth and Health Study (r = 0.60 to 0.81 between the ages 13 and 16 years) [bib_ref] Development and tracking of central pattern of subcutaneous fat in adolescence and..., Van Lenthe [/bib_ref] and the Muscatine Study [bib_ref] Factors affecting tracking of coronary heart disease risk factors in children. The..., Mahoney [/bib_ref] the stability coefficients of subcutaneous fat variables were higher compared to our study. Tracking of the skinfold ratios reported from the Paris Growth Study [bib_ref] Influence of body fat distribution during childhood on body fat distribution in..., Roland-Cachera [/bib_ref] (r = 0.40 to 0.50 between the ages 16 and 21 years) and Project Development of median triceps skinfold of Ellisras rural chil-dren and NHANES IIIreference population HeartBeat! [bib_ref] Tracking body fat distribution during growth: using measurements at two occasion vs..., Mueller [/bib_ref] (r = 0.4 to 0.81 between the ages 8, 11, 14 examine over 1 to 3 years) was similar to the current study. Similar to the current analysis, we found a significant tracking of body mass index for these children (preschool children (B = 0.6 (95% CI 0.6-0.7) and for primary school children (B = 0.6 (95%CI 0.5-0.6) [bib_ref] Development and tracking of body mass index from preschool aged into adolescence..., Monyeki [/bib_ref].
In previous report, the current sample was reported to exhibit high prevalence of stunting and wasting particularly at an older age while thinness (preschool children ranged from 39.4 to 42.6% and primary school children ranged from 23.7 to 30.0%) was a major public health problem compared to overweight (pre school children ranged from 0 to 3.9% and primary school children ranged from 0 to 15.5%) [bib_ref] Obesity: does it occur in African children in a rural community in..., Monyeki [/bib_ref] [bib_ref] Growth and nutritional status of rural South African children 3-10 years old:..., Monyeki [/bib_ref] [bib_ref] Development and tracking of body mass index from preschool aged into adolescence..., Monyeki [/bib_ref]. Cameron and Demerath [bib_ref] Critical period in human growth and their relationship to diseases of aging, Cameron [/bib_ref] reported that growth retardation or malnutrition in early foetal development alters hypothalamic development such that appetite control or energy maintenance functions are permanently reset to high energy efficiency to promote the rapid gain of weight.
The influence of fat distribution in the Ellisras population may also be mediated by the sex steroid hormones which could be described as android patterns for males and gynoid patterns for females. The android pattern is central or visceral and the gynoid pattern is peripheral and mostly notable at the gluteal-femoral region. Visceral fat is more mobile and gluteo-femoral is described as sluggish in ST = subscapular/triceps ratio, SST = subscapular/subscapular +triceps ratio, SBT = subscapular + suprailiac/biceps + triceps + subscapular + suprailiac, * = not significant, ** = * Longitudinal tracking coefficient derived from GEE analysis, CI confidence interval and partial correlation coefficient controlled for age and maturation for fat pattern variables of Ellisras rural children.
## Table 3: regression coefficient, 95% confidence interval and p-value controlled for age and maturation in the association of the initial fat pattern variable measurements with subsequent measurements
Mean age in years 4.9 5.5 5.9 6.5 6.9 7.5 7.9 8.5 8.9 9.9 10.9 11.5 terms of lipid mobilization [bib_ref] Critical period in human growth and their relationship to diseases of aging, Cameron [/bib_ref]. Bjorntorpdescribe the gluteo-femoral deposition of fat as related to high lipo-protein lipase and slow lipid mobilization. The increase in size of gluteo-femoral deposits in girls around the time of menarchy may thus be because of perimenarcheal increase in progesterone level [bib_ref] Critical period in human growth and their relationship to diseases of aging, Cameron [/bib_ref]. Furthermore, serum testosterone is associated with an increase in subcutaneous trunk fat in pubertal males while higher concentration of oestrogen in early pubertal girls is associated with a gynoid distribution of body fat [bib_ref] Tracking of physical activity and physical fitness across the lifespan, Malina [/bib_ref]. However, the relationships among steroid hormones adiposity and relative fat distribution are complex and may be mediated by sex steroid stimulated growth hormone release [bib_ref] Hormonal changes during puberty and their relationship to fat distribution, Roemmich [/bib_ref]. Cameron and Demerath [bib_ref] Critical period in human growth and their relationship to diseases of aging, Cameron [/bib_ref] reported high subcutaneous level at the subscapular and the triceps sites to be positively related to both insulin concentration and insulin resistance.
[formula] ǩ Ǩ ǩ Ǩ ǩ Ǩ ǩ Ǩ ǩ Ǩ ǩ Ǩ ǩ Ǩ ǩ Ǩ ǩ Ǩ ǩ Ǩ ǩ Ǩ ǩ Ǩ B B B β β β B β β B B β β B B β β β β β B β β β CIǩ Ǩ ǩ Ǩ ǩ Ǩ ǩ Ǩ ǩ Ǩ ǩ Ǩ ǩ Ǩ ǩ Ǩ ǩ Ǩ ǩ Ǩ ǩ Ǩ ǩ Ǩ B B B β β β B β β B B β β B B β β β β β B β β β [/formula]
Tracking of the central patterns of body fat in this study may also be affected by the developmental features of children. There was more heterogeneity among Ellisras girls than boys for the tracking correlation coefficient either through GEE or partial correlation coefficient or linear regression. The onset of puberty affects different anatomic well defined body sites of fat differently [bib_ref] Principles of growth standards, Tanner [/bib_ref]. This could be supported by slightly higher tracking coefficients in Ellisras girls compared to boys over time. The eight years duration of the study with measurements carried out twice yearly not only provide accurate tracking measurements of the central patterns of body fat but also account for the slightly lower tracking coefficient in the last three measurements for both preschool and primary school children as it was the case in other studies [bib_ref] Development of fat tissue and body mass index from infancy to adulthood, Gasser [/bib_ref] [bib_ref] Development and tracking of central pattern of subcutaneous fat in adolescence and..., Van Lenthe [/bib_ref] [bib_ref] Reference data for obesity: 85 th and 95 th percentile of body..., Must [/bib_ref].
Selection of skinfold and skinfold ratios as indicators for the central patterns of body fat in children is of real concern given the challenges in measuring due to slightly larger inter and intra tester reproducibility of the skinfold measurements as it was also the case in the present study [bib_ref] Waist circumference and cardiovascular risk factors in pre-pubertal children, Maffeis [/bib_ref] [bib_ref] Relationship between anthropometric variables and lipids levels among school children: the Taipei..., Chu [/bib_ref] [bib_ref] Body fatness: Longitudinal relationship of body mass index and the sum of..., Twisk [/bib_ref]. However, currently skinfolds are the most suitable indicators until such time when indicators can be found after the patho-physiological mechanism relating to central pattern of body fat to cardiovascular disease morbidity and mortality is clarified [bib_ref] Development and tracking of central pattern of subcutaneous fat in adolescence and..., Van Lenthe [/bib_ref]. Furthermore, the assessment of breast development was also problematic. Although we were able to obtain a visual assessment of breast development rather than relying on the self reports from girls, fat tissue can be mistaken for breast tissue in cases where the breast is not palpated. However, a key advantage of this method is that it is widely used by researchers and clinicians thereby increasing its applicability. Physical activity and fitness of these children were not controlled in the analysis. Finally, in our study girls were never asked if they had once given birth as some subjects missed measurements sessions more than one occa-sion and rejoined the study thereafter. Many adolescent girls, particularly in rural areas of South Africa today, have multiple pregnancies as a results of poverty and other social factors. It is common that initially women become overweight after their first birth child [bib_ref] Nutrition transition in Morocco, Benjelloun [/bib_ref]. Very few, if none do engage in hard physical activity, sports or work after pregnancy hence they do not lose weight.
# Conclusion
The accumulation of central patterns of body fat for Ellisras children starts in childhood and adolescent stage. Both Ellisras pre school and primary school girls showed a high subcutaneous fat compared to boys from the childhood stage and beyond. Preschool and primary school boys showed a consistent high skinfold ratios compared to girls over time. Tracking coefficient for skinfold thickness was significantly high for both preschool and primary school children, while it was slightly lower for the skinfold ratios. Investigation of nutritional intake and physical activity patterns over time will shed light on how healthy these children are and their lifestyle is. Community awareness on healthy life style may have a key role in the prevention of obesity later in life.
[fig] Figure 1: Development of median triceps skinfold of Ellisras rural children and NHANES III (Frisancho, 1990) reference population. of median subscapular skinfold of Ellisras rural children and the NHANES III (Frisancho, 1990) reference population Figure 2 Development of median subscapular skinfold of Ellisras rural children and the NHANES III (Frisancho, 1990) reference population. [/fig]
[table] Table 3: shows regression coefficient, 95% confidence interval and p-value in the association of the initial fat pattern variables measurements and the subsequent measurements adjusted for age and maturation. Pre- [/table]
[table] Table 1: Differences in the descriptive statistics for overlapping mean ages# [/table]
[table] Table 2: Specific tracking coefficient* between the first and the subsequent measurements [/table]
|
A boy with developmental delay and mosaic supernumerary inv dup(5)(p15.33p15.1) leading to distal 5p tetrasomy – case report and review of the literature
Background: With only 11 patients reported, 5p tetrasomy belongs to rare postnatal findings. Most cases are due to small supernumerary marker chromosomes (sSMCs) or isochromosomes. The patients share common but unspecific symptoms such as developmental delay, seizures, ventriculomegaly, hypotonia, and fifth finger clinodactyly. Simple interstitial duplications leading to trisomies of parts of 5p are much more frequent and better described. Duplications encompassing 5p13.2 cause a defined syndrome with macrocephaly, distinct facial phenotype, heart defects, talipes equinovarus, feeding difficulties, respiratory distress and anomalies of the central nervous system, developmental delay and hypotonia.Case presentation: We present a boy with dysmorphic features, developmental delay, intellectual disability and congenital anomalies, and a mosaic sSMC inv dup(5)(p15.33p15.1). He is the fourth and the oldest reported patient with distal 5p tetrasomy. His level of mosaicism was significantly different in lymphocytes (13.2%) and buccal cells (64.7%). The amplification in our patient is smaller than that in the three previously published patients but the only phenotype difference is the absence of seizures in our patient. Conclusions: Our observations indicate that for the assessment of prognosis, especially with respect to intellectual functioning, the level of mosaicism could be more important than the extent of amplification and the number of extra copies. Evaluation of the phenotypical effect of rare chromosomal aberrations is challenging and each additional case is valuable for refinement of the genotype-phenotype correlation. Moreover, our patient demonstrates that if the phenotype is severe and if the level of sSMC mosaicism is low in lymphocytes, other tissues should be tested.
# Background
Tetrasomies of a part or the whole 5p belong to rare postnatal findings and can be the result of a small supernumerary marker chromosome (sSMC) with inverted duplication (inv dup) or an isochromosome. Clinical outcomes of sSMCs vary from an unaffected to a severely affected status with major anomalies and intellectual disability (ID). sSMCs are often present in mosaics and can even be absent in some tissues. The clinical manifestation is frequently influenced by the level of mosaicism in specific tissues [bib_ref] Clinical impact of somatic mosaicism in cases with small supernumerary marker chromosomes, Liehr [/bib_ref].
Only 11 postnatal patients of 5p tetrasomy have been reported, seven with supernumerary i(5)(p10) [bib_ref] Mosaic 5p tetrasomy, Stanley [/bib_ref] [bib_ref] Tetrasomy 5p mosaicism due to an extra i(5p) in a severely affected..., Lorda-Sánchez [/bib_ref] [bib_ref] Pigmentary mosaicism with mosaic chromosome 5p tetrasomy, Hansen [/bib_ref] [bib_ref] Tetrasomy 5p mosaicism due to an additional isochromosome 5p in a man..., Venci [/bib_ref] [bib_ref] Mosaic tetrasomy 5p resulting from an isochromosome 5p marker chromosome: case report..., Brock [/bib_ref] , three with amplification of the distal part of 5pone interstitial triplication [bib_ref] Partial tetrasomy with triplication of chromosome (5) (p14-p15.33) in a patient with..., Harrison [/bib_ref] and two supernumerary inv dup [bib_ref] A supernumerary marker chromosome with a neocentromere derived from 5p14->pter, Fritz [/bib_ref] [bib_ref] Unusual isochromosome 5p marker chromosome, Roulet-Coudrier [/bib_ref] , and one patient had proximal 5p tetrasomy but his phenotype description and sSMC characterization were rather incomplete [bib_ref] Multicolor FISH used for the characterization of small supernumerary marker chromosomes (sSMC)..., Brecevic [/bib_ref]. The features of 5p tetrasomy are developmental delay, seizures, ventriculomegaly, hypotonia, short stature or growth delay, and fifth finger clinodactyly [bib_ref] Unusual isochromosome 5p marker chromosome, Roulet-Coudrier [/bib_ref]. A single transverse palmar crease, recurrent infections, abnormalities of the diaphragm, abnormalities of the pinna, microretrognathia, abnormalities of the philtrum and thin upper lip vermilion are also frequently observed. Interestingly, the phenotype seems to be rather similar irrespective of whether the whole 5p or just its distal part is amplified.
In contrast to 5p tetrasomies, partial or complete 5p trisomies are much more frequent. Proximal 5p trisomies are associated with a specific phenotype with macrocephaly, facial anomalies (upslanted palpebral fissures, hypertelorism, epicanthus, depressed nasal bridge, midface retrusion, micrognathia, and abnormality of the pinna), short neck, hypoplasia of the abdominal wall musculature, congenital heart defects, talipes equinovarus, feeding difficulties, respiratory distress, recurrent respiratory infections, and anomalies of the central nervous system (CNS), especially ventriculomegaly. Generalized hypotonia, seizures and ID are also common. The critical region for this phenotype was primarily placed to 5p11-p13 by Chia et al. [bib_ref] A case report of a de novo tandem duplication (5p) (p14--pter), Chia [/bib_ref] and later refined to 5p13.1-p13.3 by Loscalzo et al. [bib_ref] A patient with an interstitial duplication of chromosome 5p11-p13.3 further confirming a..., Loscalzo [/bib_ref]. The 5p13.2 duplication syndrome was finally defined by Yan et al. [bib_ref] Genomic duplication resulting in increased copy number of genes encoding the sister..., Yan [/bib_ref] , and NIPBL was suggested as its critical gene [bib_ref] 5p13 microduplication syndrome: a new case and better clinical definition of the..., Novara [/bib_ref].
We present a 5-year-old boy with dysmorphic features, feeding difficulties, hypotonia, developmental delay, ID, autistic features, and mosaic sSMC inv dup(5)(p15. 33p15.1) resulting in tetrasomy of distal 5p not involving the 5p13.2 region. We compared the genotype and phenotype of our patient with previously described patients of 5p tetrasomy and reviewed the literature for clinical effects of tetrasomy and trisomy of distinct parts of 5p. We found just slight and non-specific differences among the phenotypes of the patients compared. Our observations indicate that for the assessment of prognosis, especially with respect to intellectual functioning, the level of mosaicism could be more important than the extent of the region amplified and the number of extra copies.
## Case presentation
## Clinical report
The boy was born via spontaneous delivery from the second uneventful pregnancy to healthy unrelated Caucasian parents. The age of the mother and the father was 31 and 32 years, respectively. Birth weight was 3650 g (82nd centile) and birth length was 53 cm (87th centile). No congenital anomalies were observed in the newborn, and the neonatal period was unremarkable. During the infant period the boy suffered from feeding difficulties and recurrent upper airways infections. His psychomotor development was delayed. Pyelectasia, foramen ovale apertum and hiatal hernia were detected at the age of 14 months, and brain MRI revealed a cyst in posterior cranial fossa.
He was examined again at the age of 3 years and 5 months due to global developmental delay. He was not able to walk but he started to stand with support. The speech was absent and he used only vocalizations. He had problems with chewing and he could eat only mashed food. His parents described frequent aggressive behavior with biting. Hypotonia (central hypotonic syndrome), hypermobility of joints and pedes planovalgi were also present. His height was 96 cm (11th centile), weight was 13.5 kg (7th centile) and head circumference was 52.5 cm (96th centile). Dysmorphic features included dolichocephaly, frontal bossing, low set ears, abnormality of the pinna, hypotelorism, downslanted palpebral fissures, epicanthus, depressed nasal bridge, low hanging columella, long philtrum, thin upper lip vermilion, short chin, slight midface retrusion, single transverse palmar crease on the right hand, pectus excavatum, and kyphosis .
At the last examination at the age of 5 years and 1 month his height was 110 cm (18th centile; however, according to growth prediction from the mid parent height, the height of the boy would be expected to be around 97th centile, and thus the actual observation might rather indicate a more significant growth delay), weight was 15.5 kg (2th centile) and head circumference was 54 cm (95th centile). Speech was still absent and feeding and chewing difficulties persisted, but the boy was able to walk independently. He showed symptoms of an autism spectrum disorder such as stereotypical hand movements, frequent bruxism and a very sporadic eye contact. Macrocephaly, dolichocephaly, central hypotonic syndrome and hypermobility of the joints persisted. Pedes planovalgi were accompanied by valgus deformity of the knees. In addition, a small umbilical hernia was present. Midface retrusion and short chin observed previously were not present, but the forehead was even more prominent .
# Laboratory analysis
The research was prospectively reviewed and approved by a duly constituted ethics committee. Informed consent was obtained from the parents of the patient. Examination of G-banded chromosomes from peripheral blood lymphocytes was performed using standard protocols. For fluorescence in situ hybridization (FISH) analysis of blood lymphocytes and buccal cells, Cytocell Aquarius Cri-duchat and SOTOS Probe Combination (LPU 013) (Oxford Gene Technology, UK) was used. High-resolution array comparative genomic hybridization (aCGH) and single nucleotide polymorphism (SNP) analysis of the lymphocyte genomic DNA isolated using AutoGen Flex STAR (Autogen, USA) employed the SurePrint G3 ISCA V2 CGH 8x60K Microarray and the SurePrint G3 ISCA CGH + SNP 4x180K Microarray, respectively, according to the protocol of the manufacturer (Agilent Genomics, USA).
# Results
The patient had a mosaic sSMC and karyotype mos 47,XY,+mar [bib_ref] A supernumerary marker chromosome with a neocentromere derived from 5p14->pter, Fritz [/bib_ref] /46,XY [bib_ref] Identification of mammalian mediator subunits with similarities to yeast mediator subunits Srb5,..., Sato [/bib_ref]. Parental karyotypes were normal and the sSMC occurred de novo. aCGH analysis of the patient revealed gain of material in 5p15.33p15. 1, in the region chr5:22149_15009591 (hg19) . FISH analysis confirmed the origin of the sSMC from 5p15. 33p15.1 and revealed its inv dup structure , with the final karyotype mos 47,XY,+dup(5)(pter→p15.1::p15. 1 → pter)dn [bib_ref] A supernumerary marker chromosome with a neocentromere derived from 5p14->pter, Fritz [/bib_ref] /46,XY [bib_ref] Identification of mammalian mediator subunits with similarities to yeast mediator subunits Srb5,..., Sato [/bib_ref]. The region of tetrasomy was 15 Mb long and encompassed 137 genes including 54 protein-coding genes, of which 14 are known diseasecausing genes (Additional file 1). Considering the tetrasomic nature of the aberration, the mean log 2 ratio of 0.176 observed in aCGH indicated a mosaic of 13.0% according to the formula by Valli et al. [bib_ref] Evaluating chromosomal mosaicism by array comparative genomic hybridization in hematological malignancies: the..., Valli [/bib_ref]. FISH analysis of 409 cell nuclei of peripheral lymphocytes showed the sSMC in 13.2% of cells. In contrast, interphase FISH analysis of 312 nuclei of buccal mucosa cells showed the sSMC in 64.7% of cells. SNP array CGH analysis identified 71 at least partly informative SNPs in the interval of the 5p tetrasomy allowing to deduce the parental origin of the sSMC. The analysis was complicated by the low level of mosaicism of the sSMC in lymphocytes, and the almost 1% error rate of the array as indicated by genome-wide discrepancies with Mendelian inheritance. Nevertheless, 18 informative SNPs clearly indicated the paternal origin of the sSMC, and 51 additional SNPs were not at odds with this scenario. Of the SNPs which could discriminate between the paternal alleles, 10 clearly indicated that both SNP alleles on the sSMC were identical to the allele present on the normal chromosome 5 inherited by the patient from the father, and 8 additional SNPs were not at odds with this scenario (showing that the sSMC alleles could originate either from the paternal chromosome 5 inherited by the patient, or from the other paternal chromosome 5). Of the 51 SNPs which were not at odds with the paternal origin, 20 showed alleles identical to the sSMC and the normal paternal chromosome 5, and 29 could not discriminate between the two scenarios.
# Discussion and conclusions
According to literature reports, isochromosomes of the whole 5p (i(5)(p10)) [bib_ref] Mosaic 5p tetrasomy, Stanley [/bib_ref] [bib_ref] Tetrasomy 5p mosaicism due to an extra i(5p) in a severely affected..., Lorda-Sánchez [/bib_ref] [bib_ref] Pigmentary mosaicism with mosaic chromosome 5p tetrasomy, Hansen [/bib_ref] [bib_ref] Tetrasomy 5p mosaicism due to an additional isochromosome 5p in a man..., Venci [/bib_ref] [bib_ref] Mosaic tetrasomy 5p resulting from an isochromosome 5p marker chromosome: case report..., Brock [/bib_ref] are more frequent than inv dups of distal 5p [bib_ref] Partial tetrasomy with triplication of chromosome (5) (p14-p15.33) in a patient with..., Harrison [/bib_ref] [bib_ref] A supernumerary marker chromosome with a neocentromere derived from 5p14->pter, Fritz [/bib_ref] [bib_ref] Unusual isochromosome 5p marker chromosome, Roulet-Coudrier [/bib_ref]. Our patient is the fourth and the oldest case with tetrasomy of distal 5p. He is the only patient old enough to allow proper evaluation of his intellectual functioning, not only developmental delay, but also of all other features. The patient at the age of 3 years and 5 months (a, b) and 5 years and 1 month (c, d). Apparent features include low set ears, hypotelorism, downslanted palpebral fissures, epicanthus, depressed nasal bridge, low hanging columella, long philtrum, and thin upper lip vermilion. While short chin is present only at younger age, frontal bossing is more remarkable at older age Phenotypes of the tetrasomic patients are similar and only few unspecific differences can be observed (Additional file 2). Recurrent respiratory infections were observed in our patient and in two other patients with distal 5p tetrasomy [bib_ref] A supernumerary marker chromosome with a neocentromere derived from 5p14->pter, Fritz [/bib_ref] [bib_ref] Unusual isochromosome 5p marker chromosome, Roulet-Coudrier [/bib_ref] , and another patient died at the age of 35 days because of bronchopneumonia [bib_ref] Partial tetrasomy with triplication of chromosome (5) (p14-p15.33) in a patient with..., Harrison [/bib_ref]. Respiratory infections were also described in a patient with mos i(5)(p10) [bib_ref] Tetrasomy 5p mosaicism due to an extra i(5p) in a severely affected..., Lorda-Sánchez [/bib_ref]. A possible explanation for this feature was proposed in a patient with trisomy of the whole 5p by Grosso et al. [bib_ref] De novo complete trisomy 5p: clinical and neuroradiological findings, Grosso [/bib_ref] , who found low level of secretory immunoglobulins A (IgA). Nevertheless, detailed immunological examination was not performed and the finding could be coincidental. Patients with 5p tetrasomy published in the literature were not tested for IgA and in our patient it was within the normal range. More detailed immunological examination of patients with 5p amplifications is needed to elucidate the cause of recurrent infections.
Compared to the other patients with distal 5p tetrasomy [bib_ref] Partial tetrasomy with triplication of chromosome (5) (p14-p15.33) in a patient with..., Harrison [/bib_ref] [bib_ref] A supernumerary marker chromosome with a neocentromere derived from 5p14->pter, Fritz [/bib_ref] [bib_ref] Unusual isochromosome 5p marker chromosome, Roulet-Coudrier [/bib_ref] , the tetrasomy in our patient does not encompass six of their amplified genes. The previous patients manifested seizures in infancy but our patient did not, and thus one of these genes could cause the seizures. Inactivating RETREG1 mutations cause autosomal recessive hereditary sensory and autonomic neuropathy IIB (OMIM #613115), but our patient does not show this phenotype. No diseases are associated with the other genes. While the FBXL7 [bib_ref] Prediction of the coding sequences of unidentified human genes. XII. The complete..., Nagase [/bib_ref] , MARCH11 [bib_ref] MARCH-XI, a novel transmembrane ubiquitin ligase implicated in ubiquitindependent protein sorting in..., Morokuma [/bib_ref] , ZNF622 [bib_ref] Phosphorylation of a novel zincfinger-like protein, ZPR9, by murine protein serine/threonine kinase..., Seong [/bib_ref] , and BASP1 [bib_ref] Characterization of bovine and human cDNAs encoding NAP-22 (22 kDa neuronal tissueenriched..., Park [/bib_ref] genes are expressed in brain and could possibly contribute to seizures, no data exist on MYO10 expression in CNS.
The comparison of phenotypes associated with 5p tetrasomy and 5p trisomy is summarized in Additional file 3. The common features of all 5p amplifications are anomalies of CNS, hypotonia, seizures, and ID. Although the critical region of the 5p13.2 duplication syndrome is involved neither in the tetrasomic region of our patient nor in the distal 5p tetrasomy regions of the previous patients [bib_ref] Partial tetrasomy with triplication of chromosome (5) (p14-p15.33) in a patient with..., Harrison [/bib_ref] [bib_ref] A supernumerary marker chromosome with a neocentromere derived from 5p14->pter, Fritz [/bib_ref] [bib_ref] Unusual isochromosome 5p marker chromosome, Roulet-Coudrier [/bib_ref] , many of their traits are similar to those of patients with the 5p13.2 duplication syndrome, including congenital heart defects and seizures (in two and three of the four distal 5p tetrasomy patients, respectively). The presence of these symptoms in patients with tetrasomies not involving NIPBL indicates that other gene(s) for these conditions may exist in 5p14 and 5p15. Phenotypic overlap among carriers of amplifications of non-overlapping regions of 5p was noted by Cervera et al. [bib_ref] Trisomy of the short arm of chromosome 5 due to a de..., Cervera [/bib_ref]. Their patient with dup(5)(p13.3p15.3) had severe a Partial G-banded karyotype of normal chromosomes 5 and the sSMC. b Scheme showing the banding pattern of 5p, the result of array CGH in the present patient, FISH probes for the cri-du-chat region employed in our study (5p15.31 (FLJ25076) in green, 5p15.2 (CTNND2) in red), disease-causing genes discussed (orange), the extent of 5p amplification in the present patient (thick dark blue bar) and previously published tetrasomies (inv dups, isochromosomes, an interstitial triplication; thin dark blue bars) and trisomies (terminal and interstitial duplications, unbalanced translocations; thin light blue bars). Uncertain ranges in published patients are hatched. A megabase scale bar is also shown. c FISH examination of normal chromosomes 5 and the sSMC. See (b) for 5p probe colors. The 5q35 (NSD1) probe is in green. The pattern of signals on the sSMC indicates its structure of inv dup of distal 5p ID, seizures, macrocephaly, upslanted palpebral fissures, epicanthal folds, depressed nasal bridge, and abnormal pinna, which are also characteristic for the 5p13.2 duplication syndrome. Also de Carvalho et al. [bib_ref] Partial 5p monosomy or trisomy in 11 patients from a family with..., De Carvalho [/bib_ref] reported t(5;15)(p13.3;p12) leading to distal 5p trisomy in five individuals with features resembling the 5p13.2 duplication syndrome. This reopens the question if the 5p trisomy phenotype is caused solely by trisomy of 5p13.2 with NIPBL as the possible candidate gene [bib_ref] Mb tandem microduplication in chromosome 5p13.1-p13.2 associated with developmental delay, macrocephaly, obesity,..., Oexle [/bib_ref]. Conversely, also a normal individual with a terminal 5p trisomy resulting from der(15)t(5;15)(p15.1;p13) was reported [bib_ref] Partial distal 5p trisomy resulting from paternal translocation (5;15)(p15.1;p13) in a boy..., Baialardo [/bib_ref].
The tetrasomy in our patient affected a total of 137 genes. Extensive data collected on their disease involvement, expression and sensitivity to variation are in Additional file 1. The majority of the protein-coding genes are expressed in all tissues including the brain, and some of them show lack of tolerance to loss-offunction variation. Evidence for sensitivity to increased copy number is generally sparser, and the ClinGen Dosage Sensitivity Map (https://www.ncbi.nlm.nih.gov/projects/dbvar/clingen/) indicates triplosensitivity in none of these genes. Evidence exists for haploinsufficiency of TERT and CTNND2, but neither the phenotype caused by the loss of these genes nor a "mirror" phenotype could be observed in our patient. The TRIO gene has the strongest evidence for haploinsufficiency and causes autosomal dominant mental retardation 44 (MRD44, OMIM #617061). Our patient and also other patients with 5p tetrasomy showed some features of this disorder such as downslanting palpebral fissures, short nose, micrognathia, facial asymmetry, large ears, clinodactyly, feeding difficulties, developmental delay, ID, poor speech, autistic features and recurrent infections, while microcephaly and seizures are present only in some of them [bib_ref] Partial tetrasomy with triplication of chromosome (5) (p14-p15.33) in a patient with..., Harrison [/bib_ref] [bib_ref] A supernumerary marker chromosome with a neocentromere derived from 5p14->pter, Fritz [/bib_ref] [bib_ref] Unusual isochromosome 5p marker chromosome, Roulet-Coudrier [/bib_ref].
Moreover, according to the STRING database (https:// string-db.org/), the TRIO protein activates RAC1, and the RAC1 gene causes autosomal dominant mental retardation 48 (MRD48, OMIM #617751) with a very variable phenotype. Other possibly interesting protein interactions may exist. For example, the NDUFS6 gene amplified in our patient encodes the NADH:ubiquinone oxidoreductase subunit S6 of the mitochondrial complex I (MCI). Biallelic NDUFS6 mutations cause MCI deficiency (OMIM #252010) but no evidence exists for the sensitivity of this gene to amplification. A duplication of NDUFS4, an interaction partner of NDUFS6, in case 331431 from DECIPHER was considered to be possibly pathogenic. The phenotype and severity of MCI deficiency is highly variable and it is difficult to decide if the phenotype of our patient could fit at least partly this condition because of his rather unspecific symptoms; his phenotype is definitely not discordant with mild MCI deficiency. Similarly, the product of MED10 is a component of the Mediator complex, a coactivator for DNAbinding factors activating transcription by RNA polymerase II [bib_ref] Identification of mammalian mediator subunits with similarities to yeast mediator subunits Srb5,..., Sato [/bib_ref]. Mutations in a possible interacting partner of MED10, MED12, cause X-linked ID (OMIM #309520, #300895, #305450). Another highly interconnected protein encoded by the sSMC is CCT5, a subunit of the chaperonin containing TCP1 complex involved in folding of cytoskeletal proteins. Multiple subunits of this complex have been associated with various neurodevelopmental disorders [bib_ref] Low load for disruptive mutations in autism genes and their biased transmission, Iossifov [/bib_ref] [bib_ref] Genome-wide prediction and functional characterization of the genetic basis of autism spectrum..., Krishnan [/bib_ref]. It must be noted that it is unclear from in silico analysis how much the levels of proteins encoded by genes in the amplified interval are changed, if these changes cause any disturbances of stechiometric ratios in the protein complexes, and if these disturbances impair the functions of these complexes.
As the sSMC does not contain the normal chromosome 5 centromere, it is likely to carry a neocentromere (reviewed in [bib_ref] Neocentromeres: a place for everything and everything in its place, Scott [/bib_ref]. A variable postzygotic mitotic stability of the sSMC in individual tissues (possibly due to the differences in the time of neocentromere activation [bib_ref] Inverted duplications on acentric markers: mechanism of formation, Murmann [/bib_ref] or a different proliferation disadvantage of different cell types with the sSMC could cause the different level of mosaicism observed in lymphocytes and buccal cells of our patient. The SNP array CGH analysis revealed the paternal origin of the sSMC and showed that the two copies of distal 5p present on the sSMC are identical with the corresponding part of the normal paternal chromosome 5 present in the patient. This could point to the paternal meiotic origin of the sSMC from an acentric fragment via hairpin formation, reduplication and neocentromere acquisition described by Murmann et al. [bib_ref] Inverted duplications on acentric markers: mechanism of formation, Murmann [/bib_ref]. In contrast to usually non-mosaic 5p trisomies, the phenotype of 5p tetrasomy due to a sSMC can be influenced by the level of mosaicism. Three patients with mosaic i(5)(p10) in fibroblasts but a normal karyotype in lymphocytes were reported. A girl had a 62% mosaic of i(5)(p10) in fibroblasts and developmental delay, seizures, mild ID and hypotonia [bib_ref] Mosaic 5p tetrasomy, Stanley [/bib_ref]. Another girl with a milder phenotype (borderline intelligence and psychomotor skills, but with complex partial seizures) had only 10% mosaic in fibroblasts [bib_ref] Pigmentary mosaicism with mosaic chromosome 5p tetrasomy, Hansen [/bib_ref]. Finally, a boy having in addition ventriculomegaly, short stature, and macrocephaly, had a 14% mosaic in fibroblasts [bib_ref] Mosaic tetrasomy 5p resulting from an isochromosome 5p marker chromosome: case report..., Brock [/bib_ref]. All three children exhibited mosaic hyperpigmentation of skin, and therefore the biopsy should target the hypo-or hyperpigmented lesions generally following the lines of Blaschko [bib_ref] Association of pigmentary anomalies with chromosomal and genetic mosaicism and chimerism, Thomas [/bib_ref]. Skin biopsy can be replaced by buccal smear but if the examination is negative and the suspicion for mosaicism persists, invasive sampling is necessary to reach the diagnosis. Because aCGH cannot distinguish trisomy from 50% mosaic of tetrasomy, FISH should be used to elucidate the type of aberration. However, in contrast to i(5)(p10) cases, skin signs of mosaicism have not been observed in patients with mosaic distal 5p tetrasomy. The pitfalls of mosaicism are illustrated also by an infertile man with no dysmorphic features, ID nor congenital defects who carried mos i(5) (p10) in 16% of lymphocytes, but no such sSMC in skin fibroblasts and urothelial cells [bib_ref] Tetrasomy 5p mosaicism due to an additional isochromosome 5p in a man..., Venci [/bib_ref]. The similarly low level of mosaicism in lymphocytes of our patient initially also shed doubts on its causality, until the buccal analysis was performed which showed a much higher percentage of cells carrying the sSMC. Similarly to Pallister-Killian syndrome with supernumerary isochromosome 12p in skin fibroblasts (and other tissues), but with no or insignificant representation in lymphocytes [bib_ref] Pallister-Killian syndrome: cytogenetic and molecular studies, Peltomaki [/bib_ref] , our case showed that the examination of another tissue is essential, and the easily accessible buccal cells could serve this aim. The decision if the sSMC is causal or if it is just a coincidental finding not explaining the phenotype is very important as it is crucial for planning of possible additional examinations.
## Additional files
Additional file 1:
[table] Table S1: Detailed description of encompassed genes. (XLS 328 kb) Additional file 2: Table S2. Comparison of clinical features of patients with 5p tetrasomies. (XLS 41 kb) Additional file 3: Table S3. Comparison of phenotypes caused by distal 5p tetrasomy, i(5)(p10) and dup5p13.2. (XLS 42 kb) Abbreviations aCGH: Array comparative genomic hybridization; CNS: Central nervous system; FISH: Fluorescence in situ hybridization; inv dup: Inverted duplication; SNP: Single nucleotide polymorphism; sSMC: Small supernumerary marker chromosome [/table]
|
A functional SNP regulates E-cadherin expression by dynamically remodeling the 3D structure of a promoter-associated non-coding RNA transcript
Transcription of E-cadherin, a tumor suppressor which plays critical roles in cell adhesion and the epithelial-mesenchymal transition, is regulated by a promoter-associated non-coding transcript.This RNA includes a functional C/A single nucleotide polymorphism (SNP rs16260). The A-allele is linked to decreased transcriptional activity and increased prostate cancer risk. This single nucleotide change affects recruitment of an iso-miRNA and epigenetic enzymes to regulate the promoter, yet it is distant from the isomiR-binding site in both primary sequence and secondary structure, raising the question of how regulation occurs. Here we report the 3D NMR structure of the domain of 90 nucleotides within the paRNA which includes the SNP and the isomiR-binding site. We show that the A->C mutation alters the locally dynamic structure of the paRNA, revealing that the mutation regulates the E-cadherin promoter through its effect on RNA structure.
# Introduction
Classical cadherins, such as E-cadherin 1 , are transmembrane glycoprotein components of adherens junctions which promote intercellular communication [bib_ref] The cell-cell adhesion molecule E-cadherin, Van Roy [/bib_ref] [bib_ref] Evolution: structural and functional diversity of cadherin at the adherens junction, Oda [/bib_ref]. The E-cadherin gene 4 (CDH1) generates a 120 kDa protein 5 whose cytoplasmic domain links various catenins to the actin cytoskeleton and facilitates downstream signaling through multiple pathways, including Wnt and TGF-β [bib_ref] The cytoplasmic domain of the cell adhesion molecule uvomorulin associates with three..., Ozawa [/bib_ref] [bib_ref] E-cadherin mediates contact inhibition of proliferation through Hippo signaling-pathway components, Kim [/bib_ref]. Misfunction of E-cadherin is linked to invasiveness and advanced tumor stage in many epithelial cancers [bib_ref] Global cancer statistics 2018: GLOBOCAN estimates of incidence and mortality worldwide for..., Bray [/bib_ref] [bib_ref] China Ka-doorie Biobank Collaborative Group, Cancer incidence and mortality: a cohort study..., Pan [/bib_ref]. In fact, reduced E-cadherin expression is a hallmark of the epithelial-mesenchymal transition (EMT) [bib_ref] Epithelial-mesenchymal transition: at the crossroads of development and tumor metastasis, Yang [/bib_ref] , and inhibition of E-cadherin function provokes separation and invasion of cancer cells [bib_ref] Molecular mechanisms of epithelial-mesenchymal transition, Lamouille [/bib_ref]. Thus, E-cadherin is a tumor suppressor whose deregulation promotes carcinogenesis [bib_ref] The transcription factor snail controls epithelial-mesenchymal transitions by repressing Ecadherin expression, Cano [/bib_ref].
Hypermethylation of the CDH1 promoter has been observed in human breast, prostate, and hepatocellular tumors carrying a wildtype CDH1 gene, leading to reduced expression of Ecadherin [bib_ref] E-cadherin expression is silenced by DNA hypermethylation in human breast and prostate..., Graff [/bib_ref] [bib_ref] Silencing of the Ecadherin invasion-suppressor gene by CpG methylation in human cancers, Yoshiura [/bib_ref]. In addition, a C->A (SNP rs16260) polymorphism at position -160 from the transcriptional start site decreases the activity of the CDH1 promoter by about 70% and is linked to increased risk for prostate cancer [bib_ref] A Single Nucleotide Polymorphism in the E-cadherin Gene Promoter Alters Transcriptional Activities, Li [/bib_ref]. The C allele more robustly recruits transcription factors compared to the A allele [bib_ref] A Single Nucleotide Polymorphism in the E-cadherin Gene Promoter Alters Transcriptional Activities, Li [/bib_ref]. Mechanistically, it was demonstrated that silencing of the CDH1 promoter requires formation of a microRNA (miRNA)-guided Argonaute 1 (AGO1) complex on an independently transcribed sense promoter-associated transcript, which encompasses the SNP and recruits the SUV39H1 methyltransferase to affect chromatin modifications [bib_ref] A promoter-proximal transcript targeted by genetic polymorphism controls E-cadherin silencing in human..., Pisignano [/bib_ref] [fig_ref] Figure 1: A [/fig_ref]. It was also demonstrated that the SNP-160 (C/A) influences the ability of isomiR-4534 and AGO1 to interact with the paRNA [bib_ref] A promoter-proximal transcript targeted by genetic polymorphism controls E-cadherin silencing in human..., Pisignano [/bib_ref]. However, it remains unclear how the polymorphism affects transcriptional regulation, since it does not overlap with the isomiR binding site in sequence or secondary structure. The A and C variants differed substantially in the pattern of SHAPE reactivity at and near SNP rs16260, revealing different secondary structures for the two polymorphic transcripts [bib_ref] A promoter-proximal transcript targeted by genetic polymorphism controls E-cadherin silencing in human..., Pisignano [/bib_ref] , but how the secondary structure changes is unclear too, because the A is single stranded in the proposed secondary structure model.
We report the three-dimensional structure of the 90 nucleotides domain within the promoter associated RNA containing the A-SNP, determined using NMR spectroscopy. We show that the paRNA folds into a well-defined three-dimensional structure, and that the A->C mutation unfolds a partially dynamic three-way junction where the SNP is located. Propagation of this rearrangement to a neighboring dynamic internal loop extends the structural reorganization to the helix that links the three-way junction to the isomiR binding site. Thus, we suggest a new mechanism for regulation of ncRNA function which emphasizes the functional role of RNA structure and its modulation by genetic variation. Since changing the structure of ncRNAs at polymorphic sites and through somatic mutations can affect the epigenetic landscape of human disease [bib_ref] Silencing of the Ecadherin invasion-suppressor gene by CpG methylation in human cancers, Yoshiura [/bib_ref] [bib_ref] A Single Nucleotide Polymorphism in the E-cadherin Gene Promoter Alters Transcriptional Activities, Li [/bib_ref] [bib_ref] A promoter-proximal transcript targeted by genetic polymorphism controls E-cadherin silencing in human..., Pisignano [/bib_ref] , including cancer, the potential for therapeutic targeting should be considered.
# Results
Secondary structure of the E-cadherin transcript with the -160A allele (A-paRNA). We recorded NMR spectra for the A-paRNA, corresponding to the domain of the sense transcript with the -160A allele. The 1D spectrum contain many NH peaks in the region where base pairs can be directly monitored , revealing a well-folded structure [fig_ref] Figure 1: A [/fig_ref]. However, the relatively poor quality of the spectra, as reflected as well in broad and overlapped cross-peaks in 2D nonexchangeable protons [fig_ref] Figure 1: A [/fig_ref] , would prevent structure determination. In fact, even for exchangeable peaks, below 15 °C , the spectra become broadened to the point of being nearly useless. We encountered this problem with several other RNAs studied in the laboratory, including the CssA thermometer 17 and the stem loop within the c-JUN 5′ UTR recognized by eIF3 during specialized translation initiation [bib_ref] Structure of the RNA Specialized Translation Initiation Element that Recruits eIF3 to..., Walker [/bib_ref]. As was done in those projects, we introduced tetraloops in the A-paRNA sequence, in place of three existing loops, to generate A-paRNA-TL, as shown in [fig_ref] Figure 1: A [/fig_ref] , based on the SHAPE-generated secondary structure [bib_ref] A promoter-proximal transcript targeted by genetic polymorphism controls E-cadherin silencing in human..., Pisignano [/bib_ref]. This is not unlike what is often done in RNA x-ray crystallography, where protein binding sites and tertiary contacts are engineered to stabilize crystal contacts and facilitate crystallization. We reasoned that introducing tetraloops would stabilize the secondary structure and reduce aggregation; indeed, the spectral quality improved considerably [fig_ref] Figure 1: A [/fig_ref].
To examine whether introduction of the tetraloops altered the RNA structure, we mapped the NH assignments for A-paRNA-TL (with tetraloops), obtained as described below, onto the A-paRNA (wild type loops) spectra. An overlay of the imino region spectra for the two RNAs show many overlapping peaks, indicative of very similar secondary structures for both constructs [fig_ref] Figure 1: A [/fig_ref]. Thus, introducing tetraloops in the A-paRNA sequence does not change its secondary structure but stabilizes it and reduces aggregation, making high resolution NMR studies possible.
The sharp and well-resolved H2 protons allowed us to identify U-A base pairs, while G-C base pairs were identified from the strong cross peaks between GH1 and the pair of Cytidine amino protons. Wobble base pairs were identified from the characteristically strong cross peaks between GNH1 and UNH3 within the non-canonical base-pairing range . Based on this attribution, and the canonical pattern of sequential NOEs involving imino protons in helical regions, we assigned all slowly exchanging NH chemical shifts for the A-paRNA-TL and for a shorter variant lacking the terminal stem-loop (called A-paRNA-TL-tr, , designed to isolate the 3-way junction where the SNP is located, the two stem-loops emanating from it (H4 and H5), as well as helix H3, but missing the larger 3-way junction near the boundary of the domain formed by helices H1, H2 and H3. All assignments were consistent with the predicted secondary structure, and all base pairs predicted from the SHAPE analysis were observed, except for nucleotides that are paired but at the end of helices; imino protons for unpaired nucleotides were instead not observed; overall, the results are fully consistent with the secondary structure generated from the SHAPE analysis.
In order to investigate whether the C/A polymorphism would alter the secondary structure, we prepared and purified the C-paRNA construct with the -160C allele in place of A . As before, we introduced tetraloop, consistent with the secondary structure established from SHAPE, to improve spectral quality and obtain complete assignments. An overlay of the 1D imino region for the A-paRNA-TL (-160A allele) and C-paRNA-TL (-160C) show that the NMR spectra of A and C variants differ substantially, reflecting changes in the secondary structure due to the SNP. To confirm this conclusion, we assigned the NHs of the C-paRNA-TL (-160C) as well, and established its secondary structure based on the NH assignments . Based on this analysis, we conclude that the region of the two RNAs that comprise the isomiR binding site (helix H2 in [fig_ref] Figure 1: A [/fig_ref] adopts the same secondary structure regardless of the A-or C-allele at position -160.
Furthermore, the longer stem-loop emanating from the three-way junction where the SNP is located (helix H5 in [fig_ref] Figure 1: A [/fig_ref] also retains the same secondary structure, as does helix H1 that defines the boundaries of the domain. However, the single nucleotide A-C change at position -160 within the three-way junction region alters the structure of the junction and results in a new secondary structure for the junction itself, of helix H4 as well as helix H3 that links the SNP to the domain that recruits the isomiR. As a consequence of the rearrangement of helix H3, the other three-way junction is also remodeled, so that helix H2 now emanates from a larger multi-helix junction. It is conceivable that this new structure would change the thermodynamic or kinetic stability of helix H2, allowing differential access by isomiR-4534 which initiates assembly of SUV39H1.
In summary, the analysis of the NMR spectra of the paRNA domains containing either A-and C-alleles demonstrates that both RNAs are well-folded, consistent with the SHAPE analysis [bib_ref] A promoter-proximal transcript targeted by genetic polymorphism controls E-cadherin silencing in human..., Pisignano [/bib_ref] , and that the secondary structures of the transcripts of the two polymorphic sequences differ in the region linking the SNP to the isomiR binding site. To understand which structural features of the RNA are responsible for the conformational change, we set up to determine the three-dimensional structure of the A-paRNA-TL construct.
'Divide-and-conquer' allows nearly complete spectral assignments and the collection of a large NOE dataset. The quality of the NMR spectra of the A-paRNA-TL construct is very high considering the size of the RNA, but it remains very challenging to obtain complete spectral assignments and collect a large number of experimental NOEs, due to fast relaxation leading to peak broadening, as well as very crowded NMR spectra. 3D and higher dimensionality experiments based on 13 C editing are of limited use due to fast relaxation, as is segmental labeling that would reduce spectral overlap but not the broad lines. Limited experimental NMR data (the few NOEs from slowly relaxing protons like AH2 obtained by extensive per-deuteration and selective carbon labeling) have been combined with RDC's and SAXS to generate 3D structures, but these approaches rely extensively on modeling assumptions to overcome the paucity of direct experimental observations [bib_ref] Structure of the Intact Stem and Bulge of HIV-1 ψ-RNA Stem-Loop SL1, Lawrence [/bib_ref] [bib_ref] NMR Studies of the Structure and Function of the HIV-1 5'-Leader, Keane [/bib_ref] [bib_ref] Structural basis of the non-coding RNA RsmZ acting as a protein sponge, Duss [/bib_ref] [bib_ref] Isotope labeling and segmental labeling of larger RNAs for NMR structural studies, Duss [/bib_ref]. Unsurprisingly, the quality of NMR structures of RNA improves significantly when the density of experimental NOEs is high [bib_ref] Structure and mechanism of a molecular rheostat an RNA thermometer that modulates..., Barnwal [/bib_ref] [bib_ref] Structure of the RNA Specialized Translation Initiation Element that Recruits eIF3 to..., Walker [/bib_ref] [bib_ref] NMR structure of Dengue West Nile viruses stem-loop B: A key cis-acting..., Sharma [/bib_ref].
To address these limitations, we used the proven 'divide and conquer' approach to complete chemical shift assignments and obtain a large number of NMR constraints for structure determination, starting with smaller independently folded sub-domains. This approach, illustrated in and , is grounded in the modular nature of RNA structure; in analogy to multidomain proteins, RNA secondary structure domains fold independently and tertiary interactions stabilize but only seldom rearrange the secondary structure of the RNA [bib_ref] RNA folding causes secondary structure rearrangement, Wu [/bib_ref] [bib_ref] Visualizing the higher order folding of a catalytic RNA molecule, Celander [/bib_ref]. Thus, several smaller constructs of varying length were generated (18-58 nts; . When common and overlapping base pairs are superposed, these constructs combine to generate the complete A-paRNA-TL (Figs. S3 and S4). The larger fragments (called A-paRNA-TL-tr and A-paRNA-TLtr-1) comprise the 3-way junction formed by helices H3, H4 and H5 [fig_ref] Figure 1: A [/fig_ref] where the A/C SNP is located, and isolate it within smaller RNAs to reduce spectral overlap and allow closer investigation of its folding; the stem-loop derived from it (A-paRNA-TL-tr-2) isolates the longer stem-loop emerging from the 3-way junction, helix H5; finally, A-paRNA-2 isolates the isomiR-4534 binding site, helix H2.
When we examined the isolated sub-domains, we observed all the imino peaks in the 1D relevant illustrative spectra are shown in Figs. S6-S10. Overlaying the 1D and 2D-NOESY NMR spectra (with and without per-deuteration) of the smaller subdomains on the full A-paRNA-TL spectra revealed a highly transferable pattern of chemical shifts and NOESY cross-peaks for both the exchangeable and non-exchangeable protons (Figs. S4-S5 and S11), demonstrating that the structure found in the complete A-paRNA is retained in each fragment. Once the secondary structure of each fragment was verified, we recorded 2D and 3D-NOESY spectra at different mixing times for both exchangeable and non-exchangeable protons to obtain nearly complete and unambiguous chemical shift assignments. To achieve further simplification of the spectra, we also prepared per-deuterated RNA samples for the three larger fragments, A-paRNA-tr-1, A-paRNAtr-2 and A-paRNA-tr (representative sections of the spectra are shown in Figs. S12-S14) and compared the resulting NMR spectra with the spectra of the perdeuterated A-paRNA-TL (Figs. S11 and S15). Briefly, imino protons for A-paRNA-TL establish its secondary structure and were initiated with the U28-G32, G70-U75 and U71-G74 wobble base pairs identified by strong cross peaks between GH1 and UH3 within the non-canonical base-pairing range (12-10 ppm) [fig_ref] Figure 1: A [/fig_ref].
A series of D 2 O 2D NOESY experiments with mixing times between 100 and 300 ms were used to assign the non-exchangeable protons. The overlap in the NMR ribose spectra was resolved through deuteration of the H3′, H4′ and H5′/ 5′′ positions in the sugar and the H5 positions of C/U, which simplified the spectra and sharpened the linewidth by reducing dipolar relaxation [fig_ref] Figure 1: A [/fig_ref].
Nearly complete spectral assignments were obtained in this manner. Altogether, with the aid of deuteration and 3D 13 C-edited NOESY spectra, we assigned 91.5% of resonances for A-paRNA-TL G36, A55, C58 and U82, which we were unable to assign). Based on these assignments, we then generated a comprehensive list of distance restraints for each of the individual fragments, which were merged to generate a restrain table for the A-paRNA-TL . These restraints were used in the structure calculation of A-paRNA-TL.
Quality of the structure. The 'divide and conquer' approach allowed us to obtain a much larger number of constraints for structure determination than would have been possible if we just studied the complete RNA, whose large and asymmetric structure leads to rapid relaxation and broadening of the NMR signal. These physical limitations cannot be overcome by selective isotopic labeling.
We generated the restraint set by combining NOE-derived distance and torsion angle restraints from each fragment and the structure was then calculated using a well-tested simulated annealing protocol (as discussed in methods section) within NIH-XPLOR [bib_ref] The Xplor-NIH NMR molecular structure determination package, Schwieters [/bib_ref] This dynamic is reflected in higher local uncertainty in the structure and higher RMSD, as discussed in the previous paragraph. As discussed next, the conformational flexibility of the threeway junction where the SNP is located as well as internal loop in helix H3 is likely to be functionally significant.
3D structure of the CDH1 paRNA (-160A SNP). The first feature of the structure, starting from its 5′-and 3′-ends that come together to form helix H1 that defines the boundaries of the domain, is the purine-rich three-way junction abutted by helices H1, H2 and H3 (Figs. 2C and 3B). No evidence of stable cross-strand base pairing was found based on the absence of NOEs between guanine NH2 or AH2 to the cross strand H1′, or the observation of slow exchanging NHs.
However, the stretch of purines G83-G89 bases all stack on top to each other , as reflected in the characteristic sequential pattern of NOEs from H2′/ H1′ to H6/H8 [fig_ref] Figure 1: A [/fig_ref] observed in the highly deuterated sample; only residue U82 is bulged out. U8 is inserted between A7 and A9 but not bulged out, allowing for stacking interactions between A7 and A9 . As a result, two of the helices stack coaxially [fig_ref] Figure 1: A [/fig_ref] , while the third helix (H2) is oriented more or less orthogonally to the coaxial stack . This topological arrangement matches the 'family A' topology of three-way junctions [bib_ref] Topology of three-way junctions in folded RNAs, Lescoute [/bib_ref] , a very common tertiary motif in many three-way RNA junctions [bib_ref] Topology of three-way junctions in folded RNAs, Lescoute [/bib_ref] [bib_ref] Analysis of four-way junctions in RNA structures, Laing [/bib_ref].
The most important structural and functional feature of the paRNA is the three-way junction formed by helices H3, H4 and H5 , where the SNP is located. Within this junction, residues A43 and A44 stack on top of each other while A55 is bulged out ; residues U17 and A18 are sandwiched between the G56-C16 and G19-C42 base pairs .
In the NMR structure, residue A44, which coincides with the SNP-160A, is unpaired and stacked on A43 but does not form any cross strands interactions. Pairs of unpaired A's are often found in 3-ways junctions and internal loops, for example in ribosomal RNAs [bib_ref] A story: unpaired adenosine bases in ribosomal RNAs, Gutell [/bib_ref] [bib_ref] Comparative anatomy of 16S-like ribosomal RNA, Gutell [/bib_ref]. These residues often orient and stabilize coaxially stacked helices at three-way junctions and are important for the formation of tertiary motifs in these RNAs [bib_ref] A story: unpaired adenosine bases in ribosomal RNAs, Gutell [/bib_ref] [bib_ref] Comparative anatomy of 16S-like ribosomal RNA, Gutell [/bib_ref].
This 'SNP-three-way' junction shows similar topology and arrangements of helices as the first three-way junction, with two helices coaxially stacked (H3 and H5) and the third helix (H4) orthogonal to them . However, this second three-way junction is less conformationally rigid than the first , with a local RMSD=1.98 Å. This conformational flexibility could facilitate the structural rearrangement that occurs when A-160 is changed to C. In the C-paRNA (-160C), the AA motif is changed to AC. When this mutation is introduced, the entire three-way junction unfolds in conjunction with helix H3 linking it to the larger three-way junction near the isomiR binding site. As a result, a much larger four-way junction forms . This reorganization is most likely facilitated by the small internal loop within H3 just two base pairs away from the SNP junction, which introduces a structural distortion that destabilizes helix H3, which is conformationally dynamic, as revealed by the mixed sugar conformation observed for several residues.
In the internal loop within helix H3 involving nucleotides 13-15 and 57-60, A14 is inserted between the G60-C13 and G57-C15 base pairs and the only deviation from A-form helical pattern is found in the C15 backbone angles, as required to accommodate A14. The two bases opposite, C58 and C59, however, are displaced from helical stacking and experience conformational exchange between 2′-endo and 3′-endo pucker, like U17 in the 3-way junction.
The three unpaired bases A14, and especially C58 and C59, provide conformational flexibility to the internal loop and destabilize helix H3, which could facilitate the rearrangement of the entire helix when the A-C mutation occurs, which is necessary to remodel the secondary structure.
To test the hypothesis that the stability of the internal loop would affect the conformation of the paRNA domain, we synthesized two additional RNAs with wild type loops corresponding to either A and C-containing paRNA, truncated to helix H3 [fig_ref] Figure 1: A [/fig_ref]. The CC motif (nts 58 and 59) is replaced with U in each of the two RNAs to create a perfectly base paired helix, containing either A-and C-allele at the neighboring 3-way junction. Comparison of the 1D imino 1 H NMR spectra [fig_ref] Figure 1: A [/fig_ref] of the three RNAs shown in [fig_ref] Figure 1: A [/fig_ref] are very similar, except for the appearance of a strong AU NH signal. most likely corresponding to the newly created base pair by the mutation in helix H3. Thus, once helix H3 is stabilized by replacing unpaired nucleotides within the internal loop to create a perfect helix, the A-C substitution at position -160 has no effect on the structure.
The fourth significant feature in the 3D structure is a small internal loop in helix H5 (five nts) involving nucleotides 24-25 and 34-37 . This is the most flexible region of the RNA. Nucleotides A35 and G36 are transiently solvent exposed, leading to relatively poor convergence for the neighboring base pairs as well . However, the entire helix retains the same structure with both allelic variants, suggesting it is not likely to be important for differential regulation of the promoter activity.
# Discussion
E-cadherin acts as a tumor suppressor in many carcinomas [bib_ref] Global cancer statistics 2018: GLOBOCAN estimates of incidence and mortality worldwide for..., Bray [/bib_ref] [bib_ref] China Ka-doorie Biobank Collaborative Group, Cancer incidence and mortality: a cohort study..., Pan [/bib_ref] , and its reduced expression is a hallmark of the epithelial-to-mesenchymal transition [bib_ref] Epithelial-mesenchymal transition: at the crossroads of development and tumor metastasis, Yang [/bib_ref] [bib_ref] Molecular mechanisms of epithelial-mesenchymal transition, Lamouille [/bib_ref]. One mechanism of regulation of Ecadherin expression involves the recruitment of histone-modifying enzymes to the CDH1 promoter, which is regulated by a sense promoter-associated ncRNA transcribed independently of the main promoter [bib_ref] A promoter-proximal transcript targeted by genetic polymorphism controls E-cadherin silencing in human..., Pisignano [/bib_ref]. As illustrated schematically in [fig_ref] Figure 1: A [/fig_ref] , the A-allele is linked to decreased transcriptional activity and silencing of the E-cadherin gene requires formation of a microRNA (miRNA)-guided Argonaute 1 complex on the sense paRNA, which then recruits the SUV39H1 methyltransferase to affect chromatin modifications, thereby reducing E-cadherin expression and increasing cancer risk. Within this non-coding RNA, an A/C SNP at position -160 from the CDH1 transcription start site is associated with significantly increased prostate cancer risk as a result of increased recruitment of epigenetic enzymes and reduced activity of the CDH1 promoter [bib_ref] A promoter-proximal transcript targeted by genetic polymorphism controls E-cadherin silencing in human..., Pisignano [/bib_ref].
However, the SNP is distant from the isomiR-binding site in primary sequence and secondary structure, raising the question of how this single nucleotide change affects transcription. A possibility is that the effect of the SNP occurs at the three-dimensional structure level, yet the role of RNA structure in noncoding RNA function has been suggested much more often than it has been demonstrated, with some notable exceptions [bib_ref] Challenges and perspectives for structural biology of lncRNAs-the example of the Xist..., Jones [/bib_ref] [bib_ref] Structural insights into the stabilization of MALAT1 noncoding RNA by a bipartite..., Brown [/bib_ref].
In order to address the role of RNA structure in the regulation of E-cadherin, we determined the three-dimensional NMR structure of the 90 nucleotides domain within the paRNA responsible for recruitment of the isomiR and epigenetic enzymes, by using a divide and conquer approach.
We show here that the paRNA folds into a well-defined three-dimensional structure, and that the structures of A and C variant SNPs differ significantly. This conformational change extends from the three-way junction where the SNP is located to reorganize helix H3, propagating the effect of the single nucleotide change to the larger three-way junction from which the isomer binding site emanates [bib_ref] A promoter-proximal transcript targeted by genetic polymorphism controls E-cadherin silencing in human..., Pisignano [/bib_ref]. The conformational rearrangement is initiated by the disruption of the structure of the SNP three-way junction as a result of the AA to AC change [fig_ref] Figure 4: Schematic diagram showing the impact of SNP rs16260 on the structure of... [/fig_ref] ; however, the presence of a second distortion nearby, the asymmetric three nucleotides internal loop in helix H3 which is only two base pairs away from the junction, facilitates reorganization of the entire helix. This conformational change is facilitated by local conformational flexibility observed for both the threeway junction and the internal loop [fig_ref] Figure 4: Schematic diagram showing the impact of SNP rs16260 on the structure of... [/fig_ref]. In fact, when the helix is stabilized by mutating the internal loop into a perfectly paired helix, the NMR spectra of A-and C-alleles are very similar, implying very similar structures [fig_ref] Figure 1: A [/fig_ref]. Thus, internal loop facilitates reorganization of the 3way junction and helix H3.
Through this structural mechanism, although the SNP is located approximately 50 nucleotides away from the isomiR-4534 binding site, the conformational rearrangement of helix H3
communicates the single nucleotide change to the entire paRNA domain, generating a plausible way to affect recruitment of factors that modulate E-cadherin expression. Our NMR structural studies provide a mechanistic basis for the effect of a single nucleotide change located away from the functional site and demonstrate that even minimal changes in sequence can significantly affect the function of ncRNAs by modulating RNA structure [fig_ref] Figure 4: Schematic diagram showing the impact of SNP rs16260 on the structure of... [/fig_ref]. These results highlight a new mechanism by which polymorphic sites and somatic mutations in noncoding regions can affect the epigenetic landscape, inviting consideration of therapeutic targeting with small molecules. Several efforts have been made to discover compounds that impact cellular E-cadherin levels, but the molecular mechanism of action remains unclear [bib_ref] Targeting E-cadherin expression with small molecules for digestive cancer treatment, Song [/bib_ref] [bib_ref] Restoring E-cadherin Expression by Natural Compounds for Anticancer Therapies in Genital and..., Song [/bib_ref]. Our NMR structural studies represent a first step towards the identification of small molecules that could restore E-cadherin expression in epithelial cancers by regulating protein expression at the RNA level. NMR spectral assignments. In order to determine the structure of the A-paRNA domain including the A-allele, we used a divide-and-conquer approach and divided the RNA into four segments , which overlap to generate the complete structure. We observed many NH peaks in the 1D NMR spectrum for the for the full paRNA domain and the shorter fragments revealing a well-folded structure . Overlaying the 1D and 2D NOESY NMR spectra of the smaller segments on the spectra of the complete paRNA domain, revealed a highly transferable pattern of chemical shifts and NOESY cross-peaks for both the exchangeable and non-exchangeable protons (Figs S4-S5 and S11). In order to further simplify the spectra, we also prepared per-deuterated RNA samples (D-H5, H3′, H4′, H5′ and H5′′) for the full paRNA domain and the smaller fragments shown in , and recorded 2D-NOESY spectra at different mixing times as well. Overlaying the highly simplified, deuterated 2D NOESY spectra for the smaller domains on those of the full paRNA, we also observed considerable similarities in chemical shifts of the ribose sugar and aromatic region, to demonstrate the presence of highly transferable chemical shifts [fig_ref] Figure 1: A [/fig_ref] Spectral assignments of all fragments were facilitated by predicted RNA chemical shifts for helical regions [bib_ref] NMR investigation of RNA structure, Varani [/bib_ref] [bib_ref] RNA structure and NMR spectroscopy, Varani [/bib_ref] and confirmed using established NOE helical "walk" patterns 37,38 (Figs. S12-S15). Altogether, the analysis of multiple fragments enabled the bulk of the chemical shift assignments and NMR restraints to be obtained for the structural calculation of the complete A-paRNA domain.
# Methods
NMR structural constraints and structure calculations. In order to determine the 3D structure of the A-paRNA, we used the same divide and conquer approach used to assign chemical shifts because it allowed us to obtain a much larger number of distance restraints for structure calculation than would be possible if we just examined the complete RNA 17 . Its large molecular weight and asymmetric structure leads to rapid peak broadening due to fast relaxation of the NMR signal [fig_ref] Figure 1: A [/fig_ref]. This approach is possible because RNA structure is highly modular, and individual stem-loops fold independently of each other;
tertiary interactions, if present, are often weak and involve weaker long-range interactions between preformed RNA secondary structure motifs that only seldom disrupt the secondary structure [bib_ref] RNA folding causes secondary structure rearrangement, Wu [/bib_ref] [bib_ref] Visualizing the higher order folding of a catalytic RNA molecule, Celander [/bib_ref].
After dividing the A-paRNA domain into four partially overlapping smaller segments which, together, generate the complete A-paRNA structure, we obtained the bulk of distance restraints from NOE intensities used for the structure calculation by examining the smaller domains and by comparing them with the A-paRNA. This is possible because overlaying the 1D 1 H and 2D A summary of the NMR restraints used in our structure calculation is provided in .
Convergence was established when we observed no NOE violation greater than 0.5 Å or dihedral angle violations greater than 5- for the majority of structures within the final ensemble. The 3D structures were visualized with PyMol 40 or Chimera 41 and structural quality was analyzed using Molprobity 42 .
# Funding
The study was supported by NIH grant R35GM126942.
# Author contribution
Conceptualization, methodology, investigation, writing, reviewing and editing; graphics: SS Experimental design, writing, reviewing and editing: GV
## Declaration of competing interest
The authors declare conflict of interest. GV is a co-founder of Ithax Pharmaceuticals and Ranar Therapeutics. caused poor spectral quality. C) Comparison of the 1D imino 1 H NMR spectra of the two RNAs, with assignments for the A-paRNA-TL, obtained as presented in the text, mapped onto the spectrum of the sequence with wild type loops. The spectra are very similar, indicative of conserved structures; (*) identify resonances originating from the UUCG, GAAA and UACG tetraloops, within the A-paRNA-TL construct.
## Figure legends
## Fig. 2.
A) Secondary structure of A-paRNA-TL-tr, which was designed to isolate the three-way junction where the SNP is located, as established by NMR; B) 2D-15 N-HSQC spectra of the same RNA construct, with NMR assignments. C) Secondary structure of A-paRNA-TL, as established by NMR; D) 2D-15 N-HSQC spectra of the same RNA construct, with NMR assignments. . A) 10 lowest-energy structures superposed to generate the structural ensemble of A-paRNA-TL. B) Close up view of the 3-way junction from which the isomiR binding site emanates, formed by helices H1, H2 and H3. The unpaired bases G83-G89 stack on top to each other, while U82 is bulged out and U8 is inserted between residues A7 and A9 but not bulged out; as a result, helices H1 and H2 are coaxially stacked while helix H3 is in an orthogonal orientation; C) Close up view of the 13-15; 57-60 internal loop within helix H3; the unpaired A14 nucleotide is inserted between the G60-C13 and G57-C15 base pairs; the two bases on the opposite strand, C58 and C59, are displaced from the helix to accommodate imperfect stacking of the G60-C13 and G57-C15 base pairs; D) In the internal loop within helix H5, A35 and G36 are transiently bulged out to accommodate the U34-A25 and U24-A37 base pairs; this region of the structure displays relatively high flexibility, resulting in less precisely defined local structure; E) close up view of the three-way junction formed by helices H3, H4 and H5, where the SNP is located; A43 and A44 stack on top of each other and A55 is bulged out, while U17 and A18 are sandwiched between the G56-C16 and G19-C42 base pairs; as a result, helices H3 and H5 coaxially stack while the third helix H4 is in an orthogonal orientation.
[fig] 600: RNA synthesis, purification and NMR sample preparation. All RNA constructs were prepared by in vitro transcription using in house purified T7 RNA polymerase and synthetic DNA templates (purchased from Integrated DNA Technologies), generally as reported37,38 . 2′-O-methyl groups were incorporated with the last two residues at the 5′ end of the templates to reduce the addition of untemplated nucleotides at the 5′-end. ′Ultramer′ oligonucleotides were used for RNAs longer than 75 nucleotides. Deuterated RNA samples were prepared in the same way using selective deuterated rNTPs (D-H5, H3′, H4′, H5′ and H5′′) and 13 C/ 15 N-labeled samples were prepared using labeled rNTPs (Cambridge Isotope Laboratories). The resulting RNAs were purified for NMR studies by gel electrophoresis, electroelution and extensive dialysis, according to standard methods37,38 . RNA concentrations used for NMR experiments were 0.5-1 mM.NMR experiments.All experiments were recorded on Bruker Avance 800 or/and resonance cryogenic probes. Before each experiment, samples were freshly annealed by quick cooling after heating to 90 °C . The 1D [/fig]
[fig] Figure 1: A) Schematic diagram of the impact of SNP rs16260 on transcriptional regulation of E-cadherin. B) Secondary structure of the domain of the CDH1 paRNA that regulates E-cadherin transcription, named A-paRNA, and of its variant A-paRNA-TL, where tetraloops were introduced to stabilize the RNA structure and reduce aggregation which [/fig]
[fig] Figure 4: Schematic diagram showing the impact of SNP rs16260 on the structure of the CDH1 paRNA. A) and B) are cartoon representation of the secondary structures derived by NMR and SHAPE of the C-and A-paRNAs; C) and D) are close up view of the of the local secondary structure rearrangement induced by the C to A mutation at position -160; E) close up view of the 3D structure of A-paRNA-TL corresponding to the secondary structure of panel D. [/fig]
[table] Table legend Table 1: NMR and Structural Statistics for Structure Determination of A-paRNA-TL. [/table]
[bib_ref] Cadherin cell adhesion receptors as a morphogenic regulator, Takeichi [/bib_ref] [bib_ref] Water suppression that works. Excitation sculpting using arbitrary waveforms and pulsed field..., Hwang [/bib_ref] [bib_ref] E-cadherin expression is silenced by DNA hypermethylation in human breast and prostate..., Graff [/bib_ref] [bib_ref] Cadherin cell adhesion receptors as a morphogenic regulator, Takeichi [/bib_ref] |
Atypical femoral fracture as the cause of greater trochanteric pain syndrome – a case report
Greater trochanteric pain syndromeLateral hip pain Magnetic resonance imaging Looser zone a b s t r a c t Greater trochanteric pain syndrome may be caused by atypical femoral fractures, and this should be taken into consideration in the diagnostic workout. A 63-year-old woman was referred to our orthopedic outpatient hip clinic with a history of greater trochanteric pain syndrome without known trauma for 1 year. Initially X-ray of the hip and magnetic resonance imaging were found without pathology, and she was given a diagnosis of gluteus medius tendinopathy. As physiotherapy and steroid injections did not resolve her pain, a second look on the magnetic resonance imaging and X-ray revealed a discrete atypical femoral fracture in the lateral cortex with the presence of an isolated Looser zone, which were attributed to her pain syndrome. Two years after onset of symptoms, and with no pain relief on medical treatment, she was treated with an intramedullary nail. One-year postoperative the patient was pain free. This case emphasizes the important utility of magnetic resonance imaging in refractory greater trochanteric pain syndrome.
# Introduction
Greater trochanteric pain syndrome (GTPS) challenges the clinician in the diagnostic setup [bib_ref] Greater trochanteric pain syndrome: defining the clinical syndrome, Fearon [/bib_ref]. Many structures are acknowledged as potential lateral hip pain generators in GTPS.
✩ Competing Interests: The authors have declared that no competing interests exist. * Corresponding author.
E-mail address: [email protected] (J. Lange).
Trochanteric bursitis has historically been identified as the main pain generator, but also deep gluteal structures, spinal nerve entrapment, hip arthrosis, and hip abductors are included in the spectrum. This highlights the importance of appropriate medical imaging [bib_ref] MRI and US of gluteal tendinopathy in greater trochanteric pain syndrome, Kong [/bib_ref] [bib_ref] Does ultrasound correlate with surgical or histologic findings in greater trochanteric pain..., Fearon [/bib_ref] [bib_ref] Sonography of greater trochanteric pain syndrome and the rarity of primary bursitis, Long [/bib_ref] , especially in refractory GTPS cases [bib_ref] Tendinosis and tears of gluteus medius and minimus muscles as a cause..., Kingzett-Taylor [/bib_ref]. Atypical femoral fractures need to been recognized as a potential lateral hip pain generator in patients with GTPS.
## Case
A 63-year-old woman was referred to the senior author by her general practitioner with a history of GTPS for 1 year. The pain had started during the summer of 2016. She stated that the lateral hip pain was not caused by any trauma, but had increased gradually during a period of high activity. Her pain was accentuated when standing, and walking on slopes. But also, classically, no pain was reported, when riding a bicycle. The pain was unilateral, and she had no other musculoskeletal complains to the affected limb. She had a medical history of myxedema treated with levothyroxine since 2010, nonerosive, sero-negative rheumatoid arthritis treated with methotrexate since 2006, and osteoporosis treated with a 150 mg tablet of Bonviva (Atnahs Pharma) every fourth week, since 2008. The patient reported no history of alcohol or tobacco use.
At the initial physical examination in July 2017 by the senior author, severe pain at palpation were detected at the attachment of the gluteus medius tendon at the greater trochanter, and this was identified like the pain she normally had. There was normal 2-legged squat, negative 1-legged Trendelenburg stance, and negative Obers test. Standard X-ray examination of the hip was described as normal by a dedicated musculoskeletal radiologist.
A diagnosis of GTPS interpreted as caused by gluteus medius tendinopathy was made, and she was treated with a steroid injection in the greater trochanteric bursae and subsequently physiotherapy.
As the treatment had no effect on the lateral hip pain, a magnetic resonance imaging (MRI) scan of the hip was performed in September 2017. The MRI was initially interpreted without any abnormalities in relation to the suspected area of pathology -the hip abductors and the greater trochanteric bursae -by a dedicated musculoskeletal radiologist, and the patient was continued on physiotherapy.
But as the pain persisted, she was seen in December 2017, where scrutiny of the medical imaging did indeed show a minute, very localized inflammation in the lateral femoral cortex on MRI, which had not been identified initially, and the X-ray did reveal an atypical femoral fracture with the presence of a Looser zone in the lateral cortex at the same location of the inflammation [fig_ref] Figure 1 -: X-ray showing a Looser zone in the lateral cortex 5 [/fig_ref].
The findings were interpreted as a Looser-Milkman insufficiency fracture, and were attributed as the cause of her GTPS. Watchful waiting combined with limited weight bearing was initiated. At this point the patient had normal vitamin D and parathyroid hormone status, and normal serum calcium and alkaline phosphatase. Thyroid stimulating hormone level were consistent with thyrotoxicosis. A dual-energy X-ray absorptiometry in April 2018 showed osteoporosis with decreased T-scores (lumbar score -1.4 and hip score -2.5). The patient was seen by an endocrinologist and thyrotoxicosis was a suspected contributor to the progression in her osteoporosis Over the next 7 months the pain worsened. A renewed X-ray and MRI showed no signs of regression, nor progression, of her Looser zone. She was finally treated in September 2018 with a long intramedullary nail for pain relief [fig_ref] Figure 3 -: X-ray after insertion of a long intramedullary nail [/fig_ref] , and to avoid secondary fracture. At 1-year follow-up, there were radiological signs of healing , and the patient was free of any lateral hip pain, and reported no sequelae to the surgery.
# Discussion
We describe a case of GTPS, which were attributed to an atypical femoral fracture presumably caused by prolonged bisphosphonate use, with the imaging feature of an isolated Looser zone. The presentation of GTPS in our case was identical to what we normally see in our outpatient hip clinic with gluteal pathologies leading to lateral hip pain: a female over the age of 50 with increasing lateral hip pain, no trauma, and pain over the greater trochanter area on clinical exam.
As in our case, radiological findings of Looser zones are easily overseen, and difficult to diagnose, and MRI can play an important role in early diagnosis [bib_ref] Stress fractures in the lower extremity. The importance of increasing awareness amongst..., Berger [/bib_ref]. In 1920, Looser described these zones as transverse or oblique fissures or band-like radiolucency in cortical bone . They are often bilateral, symmetric, and lie perpendicular to the cortical margins of bones. They are most commonly found at the femoral neck, on the medial part of the femoral shaft under the lesser trochanter, and in the pubic and ischial rami. The margins of the radiolucent bands are well defined, mildly sclerotic, and there is no evidence of formation of callus. They are associated with little or no trauma, and rarely heal without treatment [bib_ref] Clinician approach to diagnosis of stress fractures including bisphosphonate-associated fractures, Mckenna [/bib_ref]. Looser explained the lesions on the basis of excessive strain, both muscular and postural, on already weakened bone [bib_ref] Uber pathologische Formen von Infraktionen und Callusbildungen bei Rachitis und osteomalakie und..., Looser [/bib_ref]. Presence of a Looser zone is now recognized as a stress fracture. Stress fractures are typically classified into 2 types depending on whether the bone is normal (fatigue fracture) or abnormal (insufficiency fracture). Risk factors include rheumatoid arthritis, metabolic bone disease, corticosteroid therapy, among others [bib_ref] Stress fractures: pathophysiology, clinical presentation, imaging features, and treatment options, Matcuk [/bib_ref]. In our case, this has the anatomic location, radiologic appearance, and clinical features of a bisphosphonate-related insufficiency fracture. Thorough medical examination including repeated medical imaging is recommended, and internal fixation could be offered if an insufficiency fracture is not healing to avoid secondary fracture. In our case, the patient showed no signs of healing or callus on repeat X-ray and MRI, and as no medical or conservative treatment helped the pain, internal fixation was performed with good pain relief.
MRI is the gold standard in early recognition of stress fractures [bib_ref] Stress fractures in the lower extremity. The importance of increasing awareness amongst..., Berger [/bib_ref]. We have identified a similar case [bib_ref] Subtrochanteric osteoid osteoma: A misdiagnosed case complicated by a hip fracture, Sferopoulos [/bib_ref] in which GTPS was believed to be caused by pain generated by a femoral osteoid osteoma, but were only identified after a hip fracture had occurred. This case could have benefitted from a first-line MRI exam.
Our case highlights the application of MRI as an important diagnostic modality, with a low threshold for performing, in refractory GTPS. Atypical femoral fractures must be kept in mind in refractory GTPS.
## Patient consent
The patient was informed that data concerning the case would be submitted for publication, and the patient has agreed to this.
[fig] Figure 1 -: X-ray showing a Looser zone in the lateral cortex 5.4 cm beneath the greater trochanter. [/fig]
[fig] Figure 2 -: Magnetic Resonance Imaging of the Looser zone. Pathological abnormalities (Looser zone and edema of the bone) were seen in STIR and T1 sequences. [/fig]
[fig] Figure 3 -: X-ray after insertion of a long intramedullary nail. Fig. 4 -One-year postoperative X-ray show radiological signs of healing. [/fig]
|
Epidemiology of diabetic retinopathy and maculopathy in Africa: a systematic review
Aim To summarize findings from studies reporting the prevalence and incidence of diabetic retinopathy and diabetic maculopathy in African countries in light of the rising prevalence of diabetes mellitus.Methods Using a predefined search strategy, we systematically searched MEDLINE, EMBASE, Science Citation index and Conference Proceedings Citation index, African Index Medicus and the grey literature database 'OpenSIGLE' for studies published between January 1990 and February 2011. Included studies reported prevalence or incidence of diabetic retinopathy or diabetic maculopathy of subjects with diabetes resident in African countries.Results Sixty-two studies from 21 countries were included: three population-based surveys; two cohort studies; five case-control studies; 32 diabetes clinic-based, nine eye clinic-based and 11 other hospital-based surveys. Included studies varied considerably in terms of patient selection, method of assessing the eye and retinopathy classification. In population-based studies, the reported prevalence range in patients with diabetes for diabetic retinopathy was 30.2 to 31.6%, proliferative diabetic retinopathy 0.9 to 1.3%, and any maculopathy 1.2 to 4.5%. In diabetes clinic-based surveys, the reported prevalence range for diabetic retinopathy was 7.0 to 62.4%, proliferative diabetic retinopathy 0 to 6.9%, and any maculopathy 1.2 to 31.1%. No obvious association between prevalence and income level of the country was detected.Conclusions Large, community-based cross-sectional and cohort studies are needed to investigate rates and determinants of prevalence of diabetic retinopathy, incidence and progression in Africa. Consensus is needed on the most appropriate methods of identification and classification of retinopathy for research and clinical practice. Estimates of prevalence of diabetic retinopathy, proliferative diabetic retinopathy and maculopathy are comparable with recent European and American studies.
# Introduction
The International Diabetes Federation (IDF) has estimated that the number of adults with diabetes in Africa will expand by 98%, from 12.1 million in 2010 to 23AE9 million in 2030 [1]-a consequence of urbanization, sedentary lifestyles, obesity, and population growth and ageing (in part as a result of successes in combating communicable diseases) [bib_ref] Global estimates of the prevalence of diabetes for 2010 and 2030, Shaw [/bib_ref]. Thirty-one of the 48 least economically developed countries, as defined by the Uni-ted Nations, are situated in Africa. The epidemic rise in diabetes therefore poses significant public health and socioeconomic challenges for the continent.
Diabetes causes visual impairment through cataract and diabetic retinopathy, a progressive disease of the retinal microvasculature. Diabetic retinopathy can be broadly divided into two clinical categories: non-proliferative and proliferative diabetic retinopathy. The pathophysiology of non-proliferative diabetic retinopathy is characterized by abnormal permeability of retinal capillaries leading to retinal oedema, and closure of capillaries leading to retinal non-perfusion and ischaemia. Diabetic maculopathy occurs when these processes affect the macula and are therefore a threat to visual functioning. Clinically significant macular oedema (CSMO) is a term from the Early Treatment of Diabetic Retinopathy Study (ETDRS)and is an evidencebased threshold for laser photocoagulation treatment.
Proliferative diabetic retinopathy occurs when retinal ischaemia is sufficiently severe to lead to the formation of new vessels. Visual loss occurs in proliferative diabetic retinopathy when these vessels bleed, or tractional retinal detachment ensues from fibrovascular proliferation. Without treatment, 50% of patients with proliferative diabetic retinopathy will become blind within 5 years. Diabetic retinopathy can be graded on the basis of the clinical features. The grades of retinopathy correlate with likelihood of development of proliferative diabetic retinopathy and can be standardized by standard retinal photographs, as used in the Early Treatment of Diabetic Retinopathy Study. The aim of this systematic review was to summarize findings from reliable research studies of estimates of the prevalence and incidence of diabetic retinopathy and maculopathy in African countries.
# Methods
## Data sources and search strategy
A systematic narrative review of published literature was performed according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement. Relevant studies published between 1948 and February 2011 were identified by searching, using a predefined strategy, the following electronic databases: MEDLINE (PubMed), EMBASE (OVID) and EMBASE Classic, Science Citation index and Conference Proceedings Citation index (ISI Web of Science). The following were also searched: the African regional database 'African Index Medicus', the grey literature database 'Open-SIGLE', the World Health Organization (WHO) International Clinical Trials Registry and the meta-Register of Controlled Trials (mRCT). Customized searches were developed by one of the authors (PIB) in conjunction with a Cochrane Collaborationtrained trials coordinator. Search histories are reproduced in the Supporting Information (Appendix S1). No language, publication status, time limits or language restrictions were applied to electronic searches. Search results were merged using reference management software (Endnote, Thomson Reuters) and duplicate records removed. The reference lists of articles identified, including existing reviews, were hand-searched.
## Selection criteria
The following were included: studies reporting prevalence or incidence or progression of diabetic retinopathy or diabetic maculopathy; cross-sectional or cohort study design; studies of subjects with diabetes mellitus resident in African countries. Exclusion criteria were: studies with fewer than 50 subjects; studies of populations of African origin residing outside the continent; reports not published in English; case series and conference abstracts. To improve the current relevance of the review those reports published before 1990 were excluded.
The method used to apply selection criteria was as follows. Titles and abstracts were examined by one investigator (PIB) and obviously irrelevant reports removed. Full text copies of the potentially relevant reports were retrieved. Multiple reports of the same study were linked together. Full-text reports were examined independently by two investigators (PIB and IJCM) for compliance with eligibility criteria. Disagreements were resolved by discussion.
## Data extraction and assessment of risk of bias
Major outcome variables were extracted independently by two investigators (PIB and IJCM) into a spreadsheet (Excel, Microsoft) with a standardized approach. Any disagreement was resolved by discussion. The main outcome variables extracted were the prevalence of diabetic retinopathy, proliferative diabetic retinopathy and diabetic maculopathy and the incidence of diabetic retinopathy, proliferative diabetic retinopathy and diabetic maculopathy. Prevalence of grades of retinopathy were recorded by patient according to the worse eye and, unless stated, are presented as such below. Studies were stratified by the source of the population sample (with community studies more likely to give a more accurate population-based assessment of prevalence); and risk of bias was assessed by seeking evidence of incomplete outcome data (missing data, patients excluded from report, patients lost to follow-up in cohort studies).
# Results
The literature search yielded 380 citations, of which 142 were reviewed in full text; 71 met the inclusion criteria and reported on a total of 62 studies , and see also Supporting Information (Appendix S2 [82-98]). Literature search report reproduced in the Supporting Information [fig_ref] Table 1: Community-based cross-sectional studies reporting prevalence of diabetic retinopathy in Africa Stratified random... [/fig_ref].
## Characteristics of included studies
Characteristics of the included studies are summarized in the Supporting Information [fig_ref] Table 2: Cohort studies reporting prevalence and incidence of diabetic retinopathy in Africa [/fig_ref].
## Design
Only three community-based studies were identified [bib_ref] Prevalence and risk factors for diabetic retinopathy in the multiethnic population of..., Dowse [/bib_ref] [bib_ref] Comparison of fasting and 2-hour glucose and HbA 1c levels for diagnosing..., Engelgau [/bib_ref] [bib_ref] Nigeria National Blindness and Visual Impairment Study Group. Causes of blindness and..., Abdull [/bib_ref]. In Mauritius, in 1998, researchers followed up the populationbased study performed in 1992 [bib_ref] Prevalence and risk factors for diabetic retinopathy in the multiethnic population of..., Dowse [/bib_ref] with a survey of the same cohort 6 years later [bib_ref] Six-year incidence and progression of diabetic retinopathy: results from the Mauritius diabetes..., Tapp [/bib_ref]. An additional cohort study followed a group of patients with Type 1 diabetes identified from a hospital clinic [bib_ref] Intensive health screening of young black diabetics, Gill [/bib_ref] [bib_ref] Mortality and outcome of insulindependent diabetes in Soweto, Gill [/bib_ref] [bib_ref] Long-term (20 years) outcome and mortality of Type 1 diabetic patients in..., Gill [/bib_ref]. All other studies were clinic-based surveys or casecontrol studies; the majority were undertaken in diabetes clinics (hospital or primary care) or hospital ophthalmology clinics.
## Distribution
The 62 studies were performed in 21 countries. Geographical distribution of studies was uneven and, within geographical regions, certain countries were over-represented. All of 19 studies undertaken in Western Africa took place in Nigeria, except one that covered Nigeria and Ghana [bib_ref] Prevalence and determinants of diabetic retinopathy and cataracts in West African type..., Rotimi [/bib_ref] and one from Mali [bib_ref] Diabetic hand infections in hospital practice in Bamako, Sidibe [/bib_ref]. Within East Africa, two studies were conducted in the Seychelles [bib_ref] Using insulin U 100 conversion to estimate the prevalence and problems of..., Sullivan [/bib_ref] [bib_ref] Diabetic eye disease: a natural history study, Taylor [/bib_ref] and two in Mauritius [bib_ref] Prevalence and risk factors for diabetic retinopathy in the multiethnic population of..., Dowse [/bib_ref] [bib_ref] Six-year incidence and progression of diabetic retinopathy: results from the Mauritius diabetes..., Tapp [/bib_ref] : relatively wealthy, ethnically diverse, small island nations. There was no clear correlation between the average standard of living in a country, as measured by per capita gross domestic product (GDP) and reported prevalence of diabetic retinopathy [fig_ref] FIGURE 2: Prevalence of diabetic retinopathy in patients with diabetes according to national per... [/fig_ref] or proliferative diabetic retinopathy (see also Supporting Information, . Only five studies specifically reported data from rural populations [bib_ref] Prevalence and risk factors for diabetic retinopathy in the multiethnic population of..., Dowse [/bib_ref] [bib_ref] Comparison of fasting and 2-hour glucose and HbA 1c levels for diagnosing..., Engelgau [/bib_ref] [bib_ref] Nigeria National Blindness and Visual Impairment Study Group. Causes of blindness and..., Abdull [/bib_ref] [bib_ref] Diabetic complications and glycaemic control in remote North Africa, Gill [/bib_ref] [bib_ref] Diabetes in rural South Africa-an assessment of care and complications, Rotchford [/bib_ref].
## Patient selection
Clinic-based studies were highly heterogeneous in patient selection in relation to age range, gender, ethnicity, duration and type of diabetes and co-morbidity. Of those studies conducted in diabetes clinics, 18 included all patients with diabetes attending the clinic, while 14 confined their study to a subgroup; for example, subjects with Type 2 diabetes [bib_ref] Features of non-insulindependent diabetes mellitus (NIDDM) in the Sudan, Elmahdi [/bib_ref] , children 5-18 years [bib_ref] Survey on acute and chronic complications in children and adolescents with type..., Majaliwa [/bib_ref] or patients with duration of diabe-tes > 5 years [bib_ref] Care-related risk factors for chronic diabetic complications in developing countries: a case..., El-Shazly [/bib_ref]. Of the nine studies conducted in ophthalmology clinics, four studied patients with a particular diagnosis (neovascular glaucoma [bib_ref] Neovascular glaucoma in a Nigerian African population, Ashaye [/bib_ref] , retinal disease [bib_ref] Retinal diseases in Ibadan, Oluleye [/bib_ref] [bib_ref] Pattern of retinal diseases at a teaching eye department, Teshome [/bib_ref] , blindness [bib_ref] Causes of blindness in Northern Tanzania: a hospital and rural health centre..., Poole [/bib_ref] , one studied patients attending specific diabetes eye clinics [bib_ref] Diabetic eye disease: a natural history study, Taylor [/bib_ref] and four studied a cross section of all eye patients [bib_ref] Ophthalmology in Luanda (Angola): a hospital based report, Carreras [/bib_ref] [bib_ref] The burden and spectrum of vitreoretinal diseases among ophthalmic outpatients in a..., Eze [/bib_ref] [bib_ref] Prevalence and pattern of retinal diseases at the Guinness Eye Hospital, Nwosu [/bib_ref] [bib_ref] Retinal diseases in a tertiary hospital: the need for establishment of a..., Onakpoya [/bib_ref]. In those studies that differentiated Type 1 and Type 2 diabetes (30 studies), most used studyspecific definitions, making inter-study comparisons problematic.
## Assessment of retinopathy
Methods of assessment and classification of retinopathy varied widely. Only nine studies used retinal photography [bib_ref] Prevalence and risk factors for diabetic retinopathy in the multiethnic population of..., Dowse [/bib_ref] [bib_ref] Six-year incidence and progression of diabetic retinopathy: results from the Mauritius diabetes..., Tapp [/bib_ref] [bib_ref] Effective and accurate screening for diabetic retinopathy using a 60 degree mydriatic..., Carmichael [/bib_ref] [bib_ref] Striving for the impossible dream: a community-based multi-practice collaborative model of diabetes..., Distiller [/bib_ref] [bib_ref] Carotid artery intima-media complex thickening in patients with relatively long-surviving type 1..., Distiller [/bib_ref] [bib_ref] Diabetes mellitus in Egypt: glycaemic control and microvascular and neuropathic complications, Herman [/bib_ref] [bib_ref] Ethnic differences in the clinical and laboratory associations with retinopathy in adult..., Kalk [/bib_ref] [bib_ref] Screening for diabetic retinopathy in primary care with a mobile fundal camera-evaluation..., Mash [/bib_ref] [bib_ref] Ophthalmoscopy versus nonmydriatic fundus photography in the detection of diabetic retinopathy in..., Mollentze [/bib_ref] and six of these were conducted in South Africa [bib_ref] Effective and accurate screening for diabetic retinopathy using a 60 degree mydriatic..., Carmichael [/bib_ref] [bib_ref] Striving for the impossible dream: a community-based multi-practice collaborative model of diabetes..., Distiller [/bib_ref] [bib_ref] Carotid artery intima-media complex thickening in patients with relatively long-surviving type 1..., Distiller [/bib_ref] [bib_ref] Ethnic differences in the clinical and laboratory associations with retinopathy in adult..., Kalk [/bib_ref] [bib_ref] Screening for diabetic retinopathy in primary care with a mobile fundal camera-evaluation..., Mash [/bib_ref] [bib_ref] Ophthalmoscopy versus nonmydriatic fundus photography in the detection of diabetic retinopathy in..., Mollentze [/bib_ref]. Thirty studies classified retinopathy simply as present or absent; 32 used a recognized grading system. Most used an adaptation of the Early Treatment of Diabetic Retinopathy Study grading system. However, the application and its reporting varied widely. In no study was an external validation of the practitioner's grading reported.
## Evidence of bias
There was evidence of incomplete outcome data in a number of studies. In the majority of clinic-based studies, the number of patients approached to participate was not reported, making selection bias difficult to assess. Many studies reported prevalence of a number of diabetic complications. In some studies, a low proportion of patients were examined for retinopathy. For example, in Harzallah et al. [bib_ref] Quality of care of patients with type 2 diabetes in a Tunisian..., Harzallah [/bib_ref] only 19% of 593 patients underwent retinal examination. Many studies excluded patients with significant cornea or media opacities [bib_ref] Epidemiology of Diabetic Retinopathy in Egypt: a hospital-based study, Macky [/bib_ref] [bib_ref] Refractive errors in type 2 diabetic patients, Mwale [/bib_ref] or with ungradeable photographs [bib_ref] Effective and accurate screening for diabetic retinopathy using a 60 degree mydriatic..., Carmichael [/bib_ref].
## Community-based studies
We identified three community-based studies [fig_ref] Table 1: Community-based cross-sectional studies reporting prevalence of diabetic retinopathy in Africa Stratified random... [/fig_ref]. In Egypt, in 1991-1994, researchers examined the prevalence of diabetes and the relationship between HbA 1c and retinopathy [bib_ref] Comparison of fasting and 2-hour glucose and HbA 1c levels for diagnosing..., Engelgau [/bib_ref] [bib_ref] Diabetes mellitus in Egypt: glycaemic control and microvascular and neuropathic complications, Herman [/bib_ref] [bib_ref] Screening for diabetic retinopathy: the utility of non-mydriatic retinal photography in Egyptian..., Penman [/bib_ref] [bib_ref] The onset of NIDDM and its relationship to clinical diagnosis in Egyptian..., Thompson [/bib_ref]. Articles by Herman et al. [bib_ref] Diabetes mellitus in Egypt: glycaemic control and microvascular and neuropathic complications, Herman [/bib_ref] and Penman et al. [bib_ref] Screening for diabetic retinopathy: the utility of non-mydriatic retinal photography in Egyptian..., Penman [/bib_ref] report different prevalence of diabetic retinopathy as graded from retinal photography: 35.4% in 376 subjects in the former and 31.6% in 335 subjects in the latter. No explanation for this difference between the two reports is offered, suggesting missing data in one or both analyses. Herman et al. [bib_ref] Diabetes mellitus in Egypt: glycaemic control and microvascular and neuropathic complications, Herman [/bib_ref] demonstrated in multivariate analysis that diabetic retinopathy was associated with longer duration of diabetes (per 10 years) (odds ratio = 1.37, 95% CI 1.09-1.73) and higher HbA 1c (per unit) (odds ratio = 1.15, 95% CI 1.03-1.27).
In 1992, researchers in Mauritius [bib_ref] Prevalence and risk factors for diabetic retinopathy in the multiethnic population of..., Dowse [/bib_ref] investigated prevalence of and risk factors for diabetic retinopathy in Asian Indian, Chinese and Creole Mauritians. This high-quality study demonstrated a high prevalence of diabetic retinopathy in all major ethnic groups in Mauritius. The prevalence of diabetic retinopathy and proliferative diabetic retinopathy were particularly high in known diabetes: 44.3 and 2.3%, respectively. Muslim Indians had the lowest prevalence of retinopathy (10.8 and 34.0% for new and known diabetes, respectively); significantly lower than Creoles (18.8 and 53.8%, respectively). The following were independently associated with retinopathy: duration of diabetes, fasting plasma glucose, systolic blood pressure, albuminuria and decreasing BMI. [fig_ref] Table 2: Cohort studies reporting prevalence and incidence of diabetic retinopathy in Africa [/fig_ref] summarizes cohort studies of diabetic retinopathy conducted in Africa. In Mauritius, in 1998, researchers followed up the population-based study performed in 1992 [bib_ref] Prevalence and risk factors for diabetic retinopathy in the multiethnic population of..., Dowse [/bib_ref] with a survey of diabetes complications [bib_ref] Six-year incidence and progression of diabetic retinopathy: results from the Mauritius diabetes..., Tapp [/bib_ref]. Of subjects with diabetes in the initial survey, 40.5% were re-examined. The 6-year incidence of diabetic retinopathy and proliferative diabetic retinopathy in subjects with diabetes but no diabetic retinopathy in the first survey was 23.8 and 0.4%, respectively. The incidence of proliferative diabetic retinopathy was much higher in subjects with mild non-proliferative diabetic retinopathy (5.2%) and moderate non-proliferative diabetic retinopathy (29.4%) in the first survey. Duration of diabetes and fasting blood glucose were independently associated with incidence of retinopathy. In South Africa, in 1982-2002, Gill and co-workers identified a cohort of patients with diabetes requiring insulin therapy diagnosed before age 30 years [bib_ref] Intensive health screening of young black diabetics, Gill [/bib_ref]. In those subjects seen at 10 years, prevalence of diabetic retinopathy had increased from 6 to 52% and proliferative diabetic retinopathy from 0 to 3% [bib_ref] Mortality and outcome of insulindependent diabetes in Soweto, Gill [/bib_ref]. In subjects seen at 20 years, prevalence of diabetic retinopathy had increased from 12 to 59% [bib_ref] Long-term (20 years) outcome and mortality of Type 1 diabetic patients in..., Gill [/bib_ref].
## Cohort studies
No other prospective cohort studies were identified. However, studies reflecting cumulative incidence of diabetic retinopathy are available. In South Africa, Distiller et al. [bib_ref] Striving for the impossible dream: a community-based multi-practice collaborative model of diabetes..., Distiller [/bib_ref] reported on 1520 patients with Type 1 diabetes and 8026 patients with Type 2 who had maintained membership for ‡ 5 years of a community-based, privately funded diabetes management programme. In subjects with Type 1 diabetes, prevalence of any retinopathy at baseline and at 5 years was 22.3 and 28%, respectively, and in subjects with Type 2 diabetes was 20.5 and 26.6%, respectively. In retrospective studies of patients with diabetes of long duration, Lester [bib_ref] Clinical status of Ethiopian diabetic patients after 20 years of diabetes, Lester [/bib_ref] showed a prevalence of diabetic retinopathy of 45.5% in 121 Ethiopian patients with duration of diabetes > 20 years, while Distiller et al. [bib_ref] Carotid artery intima-media complex thickening in patients with relatively long-surviving type 1..., Distiller [/bib_ref] reported presence of diabetic retinopathy in 14.8% of 148 South African Caucasian patients with Type 1 diabetes of > 18 years duration.
## Hospital-based and primary care-based surveys
Tables 3, 4 and 5 summarize hospital-based and primary carebased surveys reporting prevalence of diabetic retinopathy using a recognized grading system. The most recent large study from Northern Africa was conducted in Cairo during 2007-2008 in endocrinology clinics in two major teaching hospitals [bib_ref] Epidemiology of Diabetic Retinopathy in Egypt: a hospital-based study, Macky [/bib_ref]. Prevalence of proliferative diabetic retinopathy (2.3%) and clinically significant macular oedema (11.5%) reported in this study was high. Of four studies from Western Africa [bib_ref] Morbidity in relation to stage of diabetic nephropathy in type 2 diabetic..., Co [/bib_ref] [bib_ref] Diabetic retinopathy in a Nigerian community, Omolase [/bib_ref] [bib_ref] Determinants of previous dilated eye examination among type II diabetics in Southwestern..., Onakpoya [/bib_ref] [bib_ref] What does the presence of hypertension portend in the Nigerian with non-insulin-dependent..., Ikem [/bib_ref] , none reported the prevalence of maculopathy [fig_ref] Table 3: Hospital-based surveys of patients with diabetes reporting prevalence of diabetic retinopathy using... [/fig_ref]. Population-based study of prevalence of diabetes and diabetic retinopathy: methodology outlined in [fig_ref] Table 1: Community-based cross-sectional studies reporting prevalence of diabetic retinopathy in Africa Stratified random... [/fig_ref] All
## Diabeticmedicine
Diabetic retinopathy in Africa - P. I. Burgess et al. Not reported *Three reports [bib_ref] Effective and accurate screening for diabetic retinopathy using a 60 degree mydriatic..., Carmichael [/bib_ref] [bib_ref] Ethnic differences in the clinical and laboratory associations with retinopathy in adult..., Kalk [/bib_ref] [bib_ref] Screening for diabetic retinopathy in South Africa with 60 degrees retinal colour..., Joannou [/bib_ref] described grades of diabetic retinopathy in overlapping populations. Figure for any diabetic retinopathy taken from the largest report [bib_ref] Effective and accurate screening for diabetic retinopathy using a 60 degree mydriatic..., Carmichael [/bib_ref] (n = 1517); associations of diabetic retinopathy taken from smaller report (n = 507) [bib_ref] Ethnic differences in the clinical and laboratory associations with retinopathy in adult..., Kalk [/bib_ref]. Any maculopathy. àPercentage of eyes (not patients) with specified grade of diabetic retinopathy.
Only three were identified from Middle Africa [bib_ref] Care-related risk factors for chronic diabetic complications in developing countries: a case..., El-Shazly [/bib_ref] [bib_ref] Higher pulse pressure, systolic arterial hypertension, duration of diabetes and family history..., Longo-Mbenza [/bib_ref] [bib_ref] Microalbuminuria and retinopathy in a diabetic population of Cameroon, Sobngwi [/bib_ref]. Longo-Mbenza et al. [bib_ref] Higher pulse pressure, systolic arterial hypertension, duration of diabetes and family history..., Longo-Mbenza [/bib_ref] studied 3010 patients with diabetes attending diabetes primary care facilities using retinal photography; prevalence of diabetic retinopathy was 31.6%. Hospital-based surveys from Eastern Africa cover nine countries showing a general trend of increasing prevalence of diabetic retinopathy from earlier to more recent studies (Table 4). Diabetes clinic-based surveys from Southern Africa in general report higher prevalence of diabetic retinopathy and proliferative diabetic retinopathy than comparable clinics in other regions of Africa [fig_ref] Table 5: Hospital-based and primary care-based surveys of patients with diabetes reporting prevalence of... [/fig_ref]. Proliferative diabetic retinopathy prevalence > 4% was recorded in three studies of unselected diabetes clinic attendees from South Africa [bib_ref] Diabetes in rural South Africa-an assessment of care and complications, Rotchford [/bib_ref] [bib_ref] Screening for diabetic retinopathy in primary care with a mobile fundal camera-evaluation..., Mash [/bib_ref] [bib_ref] Audit of public sector primary diabetes care in Cape Town, South Africa:..., Levitt [/bib_ref]. Data on prevalence of diabetic maculopathy were limited from all regions. However, eight studies suggest high prevalence [bib_ref] Diabetic eye disease: a natural history study, Taylor [/bib_ref] [bib_ref] Diabetes in rural South Africa-an assessment of care and complications, Rotchford [/bib_ref] [bib_ref] Pattern of retinal diseases at a teaching eye department, Teshome [/bib_ref] [bib_ref] Screening for diabetic retinopathy in primary care with a mobile fundal camera-evaluation..., Mash [/bib_ref] [bib_ref] Epidemiology of Diabetic Retinopathy in Egypt: a hospital-based study, Macky [/bib_ref] [bib_ref] Audit of public sector primary diabetes care in Cape Town, South Africa:..., Levitt [/bib_ref] [bib_ref] Retinopathy in diabetic patients evaluated at a primary care clinic in Cape..., Read [/bib_ref] [bib_ref] Prevalence of diabetic retinopathy, cataract and visual impairment in diabetic patients in..., Glover [/bib_ref]. Of note, three South African, primary care-based studies were identified. Levitt et al. [bib_ref] Audit of public sector primary diabetes care in Cape Town, South Africa:..., Levitt [/bib_ref] , Mash et al. [bib_ref] Screening for diabetic retinopathy in primary care with a mobile fundal camera-evaluation..., Mash [/bib_ref] and Read and Cook [bib_ref] Retinopathy in diabetic patients evaluated at a primary care clinic in Cape..., Read [/bib_ref] reported high prevalence of proliferative diabetic retinopathy and maculopathy [fig_ref] Table 5: Hospital-based and primary care-based surveys of patients with diabetes reporting prevalence of... [/fig_ref] , comparable with hospital-based surveys in the same country and higher than hospital-based surveys elsewhere in Africa.
Two studies from South Africa compared prevalence of diabetic retinopathy in different ethnic groups [bib_ref] Ethnic differences in the clinical and laboratory associations with retinopathy in adult..., Kalk [/bib_ref] [bib_ref] Retinopathy in diabetic patients evaluated at a primary care clinic in Cape..., Read [/bib_ref]. The authors acknowledge the effect of environmental factors on different racial communities, even in the post-apartheid era. Kalk et al. [bib_ref] Ethnic differences in the clinical and laboratory associations with retinopathy in adult..., Kalk [/bib_ref] studied 507 'poor or indigent' patients attending a free hospital diabetes clinic. Prevalence of diabetic retinopathy was similar in patients of African (37%), European (41%) or Indian (37%) heritage. However, 'severe diabetic retinopathy' (study-specific classification) was significantly more frequent in Africans (52%) and Indians (41%) compared with Europeans (26%). Read and Cook [bib_ref] Retinopathy in diabetic patients evaluated at a primary care clinic in Cape..., Read [/bib_ref] found no relationship between ethnicity and diabetic retinopathy prevalence.
## Studies reporting visual acuity
Nineteen studies reported visual acuity in subjects with diabetes; parameters reported varied widely between studies. Only the Nigerian national blindness and visual impairment survey [bib_ref] Nigeria National Blindness and Visual Impairment Study Group. Causes of blindness and..., Abdull [/bib_ref] tested logarithm of the minimum angle of resolution (log-MAR) acuity. The population-based Mauritius diabetes complication study [bib_ref] Prevalence and risk factors for diabetic retinopathy in the multiethnic population of..., Dowse [/bib_ref] reported best-corrected visual acuity < 6 ⁄ 12 in 7.1% of subjects with diabetes at baseline. There was no difference in this figure for subjects with and without retinopathy. The Diabetes in Egypt project [bib_ref] Screening for diabetic retinopathy: the utility of non-mydriatic retinal photography in Egyptian..., Penman [/bib_ref] reported visual acuity in 427 subjects with diabetes. Of these, 31 (7.3%) were blind (defined as best-corrected visual acuity in the better eye less than 6 ⁄ 60); 239 (56%) had a best-corrected visual acuity between 6 ⁄ 12 and 6 ⁄ 60. It is likely that media opacities accounted for a proportion of this visual impairment: 123 eyes had cataract; 11 had corneal opacity; 17 had both.
The Nigerian national blindness and visual impairment survey was conducted between 2005 and 2007 [bib_ref] Nigeria National Blindness and Visual Impairment Study Group. Causes of blindness and..., Abdull [/bib_ref]. Diabetic retinopathy was identified as the primary cause of visual impairment in 0.29% of 3129 subjects with uncorrected visual acuity worse than 6 ⁄ 12 and in 0.5% of those with acuity less than 3 ⁄ 60. This study is likely to underestimate the visual impact of diabetic retinopathy as examiners were instructed to preferentially record treatable, then preventable causes of visual impairment; i.e. cataract would be recorded in preference to diabetic retinopathy if both were affecting visual acuity to similar degrees.
# Discussion
This systematic narrative review describes 62 studies reporting the prevalence and incidence of diabetic retinopathy and maculopathy in Africa. The methodological approach used standard inclusion, appraisal and data extraction approaches. Few high-quality, population-based studies were identified: the majority of studies were surveys of hospital clinic attendees. Identified studies were highly heterogeneous in terms of patient selection and method of assessment and classification of retinopathy. Despite these inconsistencies between studies, the review identified rates of prevalence of diabetic retinopathy in many areas of Africa comparable with high-income countries. Prevalence of proliferative diabetic retinopathy and maculopathy was high in recent studies, particularly those from Southern and Eastern Africa. Common themes were identified in the associations of diabetic retinopathy and impact on vision.
## Methodology of included studies
The review identified three high-quality, population-based, cross-sectional studies of diabetic retinopathy epidemiology [bib_ref] Prevalence and risk factors for diabetic retinopathy in the multiethnic population of..., Dowse [/bib_ref] [bib_ref] Comparison of fasting and 2-hour glucose and HbA 1c levels for diagnosing..., Engelgau [/bib_ref] [bib_ref] Nigeria National Blindness and Visual Impairment Study Group. Causes of blindness and..., Abdull [/bib_ref]. Only two cohort studies were identified. Large epidemiological studies are expensive; the population-based studies were conducted in states with relatively greater resources: Nigeria, Mauritius and Egypt. The lack of studies from Middle Africa is likely to reflect lack of resources, poor health infrastructure and deficiency of trained medical professionals. The relatively small number of studies identified from Northern Africa is partially explained by the tendency of francophone countries to publish in French.
The literature is dominated by studies of urban populations reflecting the distribution of major health facilities. Urbanization is seen as an important factor driving the diabetes epidemic [bib_ref] Diabetes in Africa: epidemiology, management and health care challenges, Levitt [/bib_ref] ; studies of urban populations may overestimate diabetic retinopathy prevalence. A caveat is that, in resource-poor settings, patients travel long distances to health facilities and rural patients may therefore be included. The majority of studies identified were hospital clinic-based surveys; selection bias is a major issue and the findings should be generalized to other settings with caution. Another bias is that clinics are seen by many as a point to collect medication; patients with diet-controlled diabetes may be under-represented. The classification of diabetes in Africa is problematic, particularly where investigations are limited. Disease characteristics differ from Caucasian populations. For example, peak age of onset of Type 1 diabetes
## Diabeticmedicine
Diabetic retinopathy in Africa - P. I. Burgess et al.
is later in African communities, typically 22-29 years [bib_ref] Diabetes in sub-Saharan Africa, Mbanya [/bib_ref].
Other phenotypes of diabetes are recognized in patients of African origin, including 'atypical African diabetes' and 'malnutrition-related diabetes' [bib_ref] A sub-Saharan African perspective of diabetes, Gill [/bib_ref]. Adaptations of the Early Treatment of Diabetic Retinopathy Study grading system have become the accepted reference standard for classifying retinopathy in research settings. Despite this, its use in everyday clinical practice is difficult because of a large number of levels requiring correlations with standard photographs and grading rules that must be remembered. General ophthalmologists and physicians in resource-poor settings may not be able to use this system to a reproducible level. Stereoscopic photography with validated grading is rapidly becoming the reference standard for assessing retinopathy. Digital photography allows transfer of images to distant reading centres, as was used in the Diabetes in Egypt project [bib_ref] Comparison of fasting and 2-hour glucose and HbA 1c levels for diagnosing..., Engelgau [/bib_ref] [bib_ref] Diabetes mellitus in Egypt: glycaemic control and microvascular and neuropathic complications, Herman [/bib_ref] [bib_ref] Screening for diabetic retinopathy: the utility of non-mydriatic retinal photography in Egyptian..., Penman [/bib_ref] [bib_ref] The onset of NIDDM and its relationship to clinical diagnosis in Egyptian..., Thompson [/bib_ref]. While expensive, this may be the direction of future research.
## Prevalence and incidence of diabetic retinopathy and diabetic maculopathy
Community-based studies identified in this review reported prevalence rates of diabetic retinopathy and proliferative diabetic retinopathy comparable with American and European populations with diabetes. The Diabetes in Egypt project [bib_ref] Comparison of fasting and 2-hour glucose and HbA 1c levels for diagnosing..., Engelgau [/bib_ref] [bib_ref] Diabetes mellitus in Egypt: glycaemic control and microvascular and neuropathic complications, Herman [/bib_ref] [bib_ref] Screening for diabetic retinopathy: the utility of non-mydriatic retinal photography in Egyptian..., Penman [/bib_ref] [bib_ref] The onset of NIDDM and its relationship to clinical diagnosis in Egyptian..., Thompson [/bib_ref] reported a prevalence of diabetic retinopathy and proliferative diabetic retinopathy in subjects with diabetes of 31.6 and 0.9%, respectively. The Mauritius diabetes complication study [bib_ref] Prevalence and risk factors for diabetic retinopathy in the multiethnic population of..., Dowse [/bib_ref] reported 30.2% diabetic retinopathy and 1.3% proliferative diabetic retinopathy; the prevalence of proliferative diabetic retinopathy in subjects with known diabetes was 2.3%. In comparison, a 2005-2008 cross-sectional sample of US adults with diabetes aged 40 years and older estimated prevalence of diabetic retinopathy and proliferative diabetic retinopathy as 28.5% and 1.5%, respectively. Recent population-based studies in Europe have reported similar rates [bib_ref] Prevalence of diabetic eye disease in patients entering a systematic primary care-based..., Younis [/bib_ref] [bib_ref] Prevalence of diabetic retinopathy in relation to age at onset of the..., Henricsson [/bib_ref] [bib_ref] Screening for eye disease in type 2 diabetes mellitus, Kristinsson [/bib_ref] [bib_ref] UK Prospective Diabetes Study, 30: diabetic retinopathy at diagnosis of non-insulin-dependent diabetes..., Kohner [/bib_ref] [bib_ref] The implementation of prompted retinal screening for diabetic eye disease by accredited..., Burnett [/bib_ref]. Younis et al. [bib_ref] Prevalence of diabetic eye disease in patients entering a systematic primary care-based..., Younis [/bib_ref] studied 8062 patients with diabetes entering an English primary care-based screening programme. The prevalences of any retinopathy and proliferative diabetic retinopathy in Type 1 diabetes were 45.7 and 3.7%, respectively, and in Type 2 diabetes were 25.3 and 0.5%, respectively.
The lack of community-based studies from Sub-Saharan Africa is important. Very high prevalence of diabetic retinopathy, proliferative diabetic retinopathy and maculopathy has been reported in notable high-quality, clinic-based surveys in the last decade: in Eastern Africa by Glover et al. [bib_ref] Prevalence of diabetic retinopathy, cataract and visual impairment in diabetic patients in..., Glover [/bib_ref] (32.0% diabetic retinopathy, 5.7% proliferative diabetic retinopathy, 15% sight-threatening maculopathy), and in South Africa by Mash et al. [bib_ref] Screening for diabetic retinopathy in primary care with a mobile fundal camera-evaluation..., Mash [/bib_ref] (62.4% diabetic retinopathy, 6.1% proliferative diabetic retinopathy, 15.2% any maculopathy) and Rotchford and Rotchford [bib_ref] Diabetes in rural South Africa-an assessment of care and complications, Rotchford [/bib_ref] (40.3% diabetic retinopathy, 5.6% proliferative diabetic retinopathy, 10.3% clinically significant macular oedema). These figures are likely to reflect factors including ethnicity, poor access to medical services, late diagnosis, and co-pathology including infection (importantly HIV and malaria), hypertension, malnutrition, and anaemia. We found no clear relationship between per capita gross domestic product and prevalence of diabetic retinopathy or proliferative diabetic retinopathy. However, the increased infrastructure to detect disease in states with greater resources is an important confounding factor.
The influence of ethnicity on diabetic retinopathy prevalence in populations of African origin has yet to be determined. In the USA, Zhang et al.reported prevalence of both diabetic retinopathy and vision-threatening retinopathy (defined as Early Treatment of Diabetic Retinopathy Study severe nonproliferative diabetic retinopathy, proliferative diabetic retinopathy, or clinically significant macular oedema) to be higher in non-Hispanic black subjects (38.8 and 9.3%, respectively) compared with non-Hispanic white subjects (26.4 and 3.2%, respectively). Previous studies have shown similar results [bib_ref] Diabetic retinopathy in a multiethnic cohort in the United states, Wong [/bib_ref] [bib_ref] Is the risk of diabetic retinopathy greater in non-Hispanic blacks and Mexican..., Harris [/bib_ref]. However, differences were attributable to risk factors for retinopathy [bib_ref] Is the risk of diabetic retinopathy greater in non-Hispanic blacks and Mexican..., Harris [/bib_ref]. Therefore, while associations between polymorphisms of specific genes and diabetic retinopathy have been described in African populations [bib_ref] Polymorphism of the endothelial nitric oxide synthase gene is associated with diabetic..., Chen [/bib_ref] , no ethnic propensity to retinopathy has been identified.
Neither of the two cohort studies identified by this review reported two-or three-step progression on the Early Treatment of Diabetic Retinopathy Study scale, as used in recent European studies [bib_ref] Effects of medical therapies on retinopathy progression in type 2 diabetes, Accord Study Group [/bib_ref]. The Mauritius diabetes complication study [bib_ref] Six-year incidence and progression of diabetic retinopathy: results from the Mauritius diabetes..., Tapp [/bib_ref] reported 6-year incidence of diabetic retinopathy (23.8%).
Six-year progression to proliferative diabetic retinopathy was reported from no diabetic retinopathy (0.4%), mild non-proliferative diabetic retinopathy (5.2%) and moderate non-proliferative diabetic retinopathy (29.4%). The UK Prospective Diabetes Study (UKPDS) reported similar 6-year incidence of diabetic retinopathy: 22% [bib_ref] UKPDS 50: Risk factors for incidence and progression of retinopathy in Type..., Stratton [/bib_ref]. However, the UKPDS population were studied from a later time point: clinical diagnosis of diabetes. In the Wisconsin Epidemiological Study of Diabetic Retinopathy (WESDR) 4-year progression of diabetic retinopathy and progression to proliferative diabetic retinopathy was observed in 41.2 and 10.5% of subjects with Type I diabetes, 34 and 7.4% of insulin-treated patients with Type 2 diabetes and 24.9 and 2.3% of non-insulin treated patients, respectively [bib_ref] The Wisconsin Epidemiological Study of Diabetic Retinopathy: a review, Klein [/bib_ref].
## Impact of diabetic retinopathy on vision
Estimates of the proportion of African patients with diabetes who are visually impaired are high even compared with older European and American studies. Of subjects in the Diabetes in Egypt project [bib_ref] Screening for diabetic retinopathy: the utility of non-mydriatic retinal photography in Egyptian..., Penman [/bib_ref] , 7.3% had best-corrected visual acuity in the better eye < 6 ⁄ 60. In contrast, of the population in the Wisconsin Epidemiological Study of Diabetic Retinopathy, 3.6% of patients aged < 30 years at diagnosis, and 1.6% of patients aged ‡ 30 years at diagnosis were legally blind according to US standards [bib_ref] The Wisconsin Epidemiologic Study of Diabetic Retinopathy. II. Prevalence and risk of..., Klein [/bib_ref] [bib_ref] The Wisconsin Epidemiologic Study of Diabetic Retinopathy. III. Prevalence and risk of..., Klein [/bib_ref]. The World Health Organization estimates that, in the USA and Canada, 17% of blindness is attributable to diabetic retinopathy [bib_ref] Global data on visual impairment in the year 2002, Resnikoff [/bib_ref]. While data are sparse, the proportion of visual impairment and blindness as a result of diabetic retinopathy in Africa appears to be considerably less. However, the prevalence of visual impairment and blindness is significantly higher in Africa [bib_ref] Global data on visual impairment in the year 2002, Resnikoff [/bib_ref] , reflecting high prevalence of pathologies including uncorrected refractive error, cataract, corneal opacities and glaucoma.
The findings of this review have important implications for both research and clinical practice. Large, community-based cross-sectional and cohort studies are needed to investigate rates and determinants of diabetic retinopathy prevalence, incidence and progression across Africa. Consensus is needed on standardized data sets and the most appropriate methods of identification and classification of diabetic retinopathy in Africa. In Europe and America, there is strong evidence for the role of poor glycaemic control and co-pathology, including hypertension in development and progression of diabetic retinopathy. Similarly, a strong evidence base exists for the treatment of sight-threatening diabetic retinopathy with laser photocoagulation and intravitreal agents. This evidence has yet to be accrued in African settings. Management of systemic disease and screening and treatment of retinopathy requires substantial infrastructure, which is currently lacking in many African states. The public health and health economic challenges for policymakers across Africa are significant.
# Funding sources
None.
## Competing interests
Nothing to declare.
## Supporting information
Additional Supporting Information may be found in the online version of this article: . Prevalence of proliferative diabetic retinopathy in patients with diabetes according to national per capita gross domestic product. [fig_ref] Table 1: Community-based cross-sectional studies reporting prevalence of diabetic retinopathy in Africa Stratified random... [/fig_ref]. Literature search report for articles reporting prevalence, incidence or progression of diabetic retinopathy or diabetic maculopathy in African countries. [fig_ref] Table 2: Cohort studies reporting prevalence and incidence of diabetic retinopathy in Africa [/fig_ref]. Characteristics of 62 studies reporting prevalence of diabetic retinopathy and maculopathy in Africa.
Appendix S1. Search histories.
Appendix S2. Supplementary references.
## Diabeticmedicine
Diabetic retinopathy in Africa - P. I. Burgess et al.
[fig] FIGURE 2: Prevalence of diabetic retinopathy in patients with diabetes according to national per capita gross domestic product. Red markers: population-based studies. Blue markers: cohort and clinic-based studies. For cohort studies prevalence in baseline survey shown. Gross domestic product per capita figures: International Monetary Fund (2011)[79].DIABETICMedicineDiabetic retinopathy in Africa • P. I. Burgess et al. [/fig]
[table] Table 1: Community-based cross-sectional studies reporting prevalence of diabetic retinopathy in Africa Stratified random sampling of persons ‡ 20 years in urban and rural areas near Cairo. 4620 adults underwent random glucose testing. Those at high risk of diabetes and a sample of those at low risk (total 1451) had a fasting glucose test. Diabetes diagnosed by World Health Organization criteria (see also Supporting Information, Appendix S1 [101]). Retinal photography graded according to Airlie House Classification and binocular indirect ophthalmoscopy examination by ophthalmologist. Those ungradeable on photography persons in 14 geographically defined clusters underwent glucose tolerance test. In 11 clusters, all adults aged 25-74 years were invited to attend; in three clusters, age-stratified sampling of adults aged 35-64 years performed. Those with diabetes and 25% of those with impaired glucose tolerance [World Health Organization criteria (115)] had 3-field, 45°stereoscopic retinal photography of the right eye. Grading by certified assessor according to modified Airlie House criteria. The prevalence of diabetic retinopathy in Egypt, 1991-1994, was reported in four publications: *denotes data from Penman et al. (1998) [41], denotes data from Herman et al. (1998) [34]. àMaculopathy in Penman et al. (1998) [41] defined as any exudates present in the macular region. The Authors. Diabetic Medicine published by John Wiley & Sons Ltd on behalf of Diabetes UK. [/table]
[table] Table 2: Cohort studies reporting prevalence and incidence of diabetic retinopathy in Africa [/table]
[table] Table 3: Hospital-based surveys of patients with diabetes reporting prevalence of diabetic retinopathy using a recognized grading system in Northern, Western and Middle Africa The Authors. Diabetic Medicine published by John Wiley & Sons Ltd on behalf of Diabetes UK. [/table]
[table] Table 4: Hospital-based surveys of patients with diabetes reporting prevalence of diabetic retinopathy using a recognized grading system in Eastern Africa with overlapping populations have emanated from the diabetes clinic at Yekatit 12 Hospital, Addis Ababa, Ethiopia. For the purposes of this review, these papers are viewed as one study. Data presented fromLester 1992 [56] andLester 1993 [57]. [/table]
[table] Table 5: Hospital-based and primary care-based surveys of patients with diabetes reporting prevalence of diabetic retinopathy using a recognized grading system in Southern Africa [/table]
|
Case Report: A case of radioactive iodine-refractory thyroid cancer accompanying cervical lymph node metastasis treated with US-guided RFA combined with 125I seed implantation
Background: Local control of metastases is critical to improving the life quality of patients with radioactive iodine-refractory (RAIR) thyroid cancer accompanying regional lymph node metastasis.Case report: The reported patient suffered from RAIR thyroid cancer accompanying poorly controlled cervical lymph node metastasis. The patient's lesions were controlled through 125 I seed implantation combined with ultrasound-guided radio-frequency ablation (US-guided RFA). Such a combination therapy has not been reported to date.Conclusion: This study found US-guided RFA combined with 125 I seed implantation to be safe and effective for the control of cervical local metastases in patients with RAIR thyroid cancer. KEYWORDS ultrasound-(US) guided, radiofrequency ablation (RFA), 125 I seed implantation, radioactive iodine-refractory (RAIR) thyroid cancer, cervical lymph node metastasis
© 2022 Zhai, Shao and Li. 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.
Patients with RAIR-DTC accompanying regional lymph node metastasis struggle with poor prognosis because they cannot benefit from 131 I diagnosis and treatment. Although conventional diagnosis and treatment methods employed in clinical practices, such as targeted medical therapy and surgery, can improve the prognosis of patients with RAIR-DTC to a certain extent, the overall response is not satisfactory. Some patients suffer from skin rupture caused by poorly controlled metastases, which seriously affects their quality of life. Studies have shown that 125 I seed implantation in the treatment of RAIR-DTC accompanying regional lymph node metastasis can safely deliver a good local control effect, making it of clinical value [bib_ref] Clinical study on clinical study on 125I seeds implantation for colon cancer..., Ye [/bib_ref]. However, the effectiveness of this treatment is usually limited to metastases of a small size. When lymph node metastases are large, with skin rupture and internal liquefactive necrosis, 125 I seed implantation alone has limitations. For example, in metastases of this kind, it is not easy to control the location of the implanted seeds since they are likely to fall into the liquid components, thereby reducing the therapeutic effect. New and effective methods are needed for controlling metastasis locally. As a minimally invasive treatment method, RFA has been used to treat benign thyroid nodules, papillary thyroid cancer and recurrent thyroid cancer [bib_ref] Efficacy and safety of ultrasound-guided radiofrequency ablation for treating low-risk papillary thyroid..., Zhang [/bib_ref] [bib_ref] Thyroid radiofrequency ablation guideline: Korean society of thyroid radiology, Kim [/bib_ref] [bib_ref] Successful radiofrequency ablation strategies for benign thyroid nodules, Lee [/bib_ref]. It is safe and feasible, especially when patients refuse surgery, have high requirements for aesthetics or have surgical contraindications, RFA as an alternative treatment can reduce patients' anxiety and improve their quality of life [bib_ref] Long-term prognosis of unilateral and multifocal papillary thyroid microcarcinoma after unilateral lobectomy..., Jeon [/bib_ref]. In our case,during the operation the patients was under monitoring and questioning, the operation was sucessfull and safe.
This report presents the results achieved using ultrasound-(US) guided aspiration of metastatic liquid components, radiofrequency ablation (RFA), and 125 I seed implantation in the local control of metastases.
## Case report
A 70-year-old woman was admitted to our hospital with "lung metastasis after papillary thyroid carcinoma surgery and new cervical masses detected over 2 years." Two years earlier, in May 2004, she had received a thyroid mass resection under local anesthesia in a local hospital to remove the cervical mass; subsequently, the postoperative pathology revealed lymph node metastasis of papillary thyroid carcinoma. With a recurrent cervical mass, the patient visited our hospital for the first time in August 2006 and underwent a radical resection of a left thyroid carcinoma under general anesthesia. The postoperative pathology indicated papillary thyroid carcinoma. The patient continued to suffer regional lymph node metastasis and received lymph node dissection twice in combination with 131 I therapy.
## Operation history
In October 2010, the patient received a left neck dissection under general anesthesia in our hospital. Postoperative pathology indicated that one in five level-III and -IV lymph nodes had metastases of papillary thyroid carcinoma, with level VI showing the formation of a cancerous node. On March 26, 2015, the patient received a radical resection of recurrent thyroid cancer and a right recurrent laryngeal nerve anatomic exploration. The postoperative pathology indicated the infiltration of papillary thyroid carcinoma at levels IV and VI of the left neck. At levels IV and VI of the right neck, all the four lymph nodes detected were found with metastases of papillary thyroid carcinoma, accompanied by the formation of a cancerous node. According to the patient's latest physical examination, she had multiple palpable enlarged lymph nodes in the neck, with tough texture and poor mobility. In 2014, new masses were detected in the right neck, which eventually ruptured in 2019. The patient received a lymph node puncture in the right neck on November 30, 2019, in the Nuclear Medicine Department of the Shanghai Sixth People's Hospital (affiliated with Shanghai Jiao Tong University). An analysis showed papillary thyroid carcinoma due to the BRAF V600E mutation. The ruptured lesion in the right neck was not under control, and a new mass was found in the left neck, protruding through the skin. This mass had a diameter of around 30 mm and was accompanied by bruising of the local skin. The patient's quality of life was significantly affected. The patient was not provided with a therapeutic regimen after visiting upper-level hospitals several times, so she returned to our hospital. After admission, the possibility of lymph node metastasis was considered based on a contrast-enhanced computed tomography (CT) scan of the neck showing multiple nodules and tumor shadows in the thyroid region, bilateral parotid region, bilateral neck, bilateral subclavian and supraclavicular fossae, and the mediastinum. Thyroglobulin in the puncture fluid from the right cervical ruptured lesion was measured at 413 ng/ml, and the lesion was considered to be a thyroid lymph node metastasis. In recent years, the patient had undergone several surgeries, 131 radioiodine therapy, systemic anti-tumor therapies, and treatment with sorafenib, vemurafenib, and pembrolizumab. Unfortunately, the effects were not satisfactory, and the metastases in the cervical lymph nodes remained uncontrolled. At that point, the patient's condition was severe, and there was no effective anti-tumor therapy.
## History of 131 iodine therapy
After admission, the patient was given targeted systemic anti-tumor treatment comprising anlotinib (1 tablet/day) and Euthyrox (112.5 µg/day) as inhibition and replacement therapy. As the left cervical mass was relatively large, local cytoreductive surgery was considered for lowering the risk of skin ulceration. Given the large volume of the left cervical mass, with a diameter of approximately 30 mm, a total of 33 125 I seeds were implanted one by one at an interval of 5 mm under CT scan guidance to achieve local cytoreduction for the metastases and reduce the risk of skin rupture.
Three days after 125 I seed implantation in the left cervical metastases, the patient reported feeling slightly better. However, the mass in her left neck remained large; it was accompanied by liquefactive necrosis, and seeds were entering the liquid components. To prevent further enlargement and development of the metastases, as well as to avoid radiation damage to adjacent tissues and organs, we proposed during multidisciplinary consultations that RFA be used to reduce the metastases in combination with other therapies for symptomatic treatment. Under the guidance of US, the size of the left cervical mass was measured to be 33 × 19 mm, which protruded through the surface of the skin. Accordingly, we aspirated the liquid components in the lesion, considering the continuous new production of fluid, seed implantation was not carried out, but RFA inactivation was performed on the solid portion at the same time as fluid extraction [fig_ref] FIGURE 1 US: -guided RFA for left cervical mass [/fig_ref]. Postoperative ultrasonography showed no obvious contrast enhancement in the metastatic lymph nodes of the left neck [fig_ref] FIGURE 2: No contrast enhancement in left cervical mass according to postoperative ultrasonography [/fig_ref]. The left lesion recover over the month following the RFA. After 23 days, the metastases no longer protruded through the skin, and all that could be seen on the cervical skin was a slight crust. Thirty days after the RFA, a contrast-enhanced CT scan of the neck revealed that the cervical anterior subcutaneous nodule could no longer be seen, which was a postoperative change. The patient's left neck completely recovered 32-35 days after the RFA.
The findings of the contrast-enhanced CT scan and the recovery of the patient's left neck provide evidence that USguided RFA combined with 125 I seed implantation can deliver a complete local response for large metastases in a safe, effective manner. US-guided RFA is particularly effective in treating lymph node lesions with fluid.
As the therapeutic effect of US-guided RFA for the patient's left cervical metastases was quite considerable, we continued with RFA of a fluid-containing lymph node metastasis in the patient's right neck at a safe location. The US-guided measurements showed one of the lymph nodes in the right neck was around 21 × 11 mm, another was around 24 × 17 mm , and there were heterogeneous internal echoes.
We aspirated the liquid components and performed RFA. According to postoperative ultrasonography, no obvious contrast enhancement was found in the cervical metastatic lymph nodes, and there was no residual ablation. The day following the right cervical lymph node RFA, there was a special location of one ulcerated lymph node on the right neck, and there was no safe entry point, so seeds were implanted in the patient's right cervical ruptured lymph node. The ruptured mass shrank following the implantation of nine 125 I one by one at an interval of 5 mm. Two months later, the mass had gone. The ablated cervical mass disappeared over 40 days. Following this success, the masses of the patients in the subsequent treatment were all solid masses, and the location was deep adjacent to the important cervical nerve, so the patient had three more occasions when 125 I seeds were implanted into the cervical lymph nodes: 9, 16, and 9 seeds were implanted, respectively. Now, the patient is recovering well, with the local metastases in the cervical lymph nodes well under control. She was discharged from hospital and is receiving systemic targeted drug therapy.
# Discussion
Most patients with DTC achieve a good prognosis after receiving surgery, 131 I radiotherapy, thyrotropin inhibition, and other therapies, with a 10-year survival rate of up to 90%. However, about 15% of these patients develop local recurrence Metastatic lymph nodes accompanying internal liquefactive necrosis in right neck as shown in ultrasonography. Frontiers in Oncology frontiersin.org and/or distant metastases and have a 10-year survival rate below 10%. As some cases are not sensitive to radioactive iodine therapy, they may develop RAIR-DTC, which is characterized by fast progression and high mortality. It is difficult for these patients to achieve satisfactory results with the traditional methods referred to above. Currently, the diagnosis and treatment of RAIR-DTC are controversial. According to the usual definition of RAIR-DTC, in the absence of exogenous iodine load interference when the level of thyroid-stimulating hormone is greater than 30 m IU/L, the lesion loses its iodine uptake capacity. Consequently, favorable results are not expected from 131 I treatment. In the case of the present study, the patient's condition was severe. Her cervical metastases could not be surgically removed, and a high dose of radioactive iodine, if allowed to accumulate around her neck, would likely have caused radiation damage to the larynx and tracheal mucosa, resulting in radiation tracheitis or radiation pneumonia. In this case, the patient was treated as RAIR-DTC because radioactive iodine therapy was not feasible for her. The treatment for RAIR-DTC is controversial, and cases accompanying cervical regional lymph node metastasis remain a problem to be solved [bib_ref] Global cancer statistics 2018: GLOBOCAN estimates of incidence and mortality worldwide for..., Bray [/bib_ref]. Surgery, chemotherapy, targeted therapy, and other related methods are now applied in clinical practices. In this case, the patient's condition was severe, and her cervical metastases could not be surgically removed. As studies in molecular biology on thyroid cancer have advanced, a large number of targeted therapeutic drugs have been developed in recent years [bib_ref] Sorafenib in radioactive iodine-refractory, locally advanced or metastatic differentiated thyroid cancer: A..., Brose [/bib_ref] [bib_ref] Exploratory analysis of biomarkers associated with clinical outcomes from the study of..., Tahara [/bib_ref] [bib_ref] Vemurafenib in patients with BRAFV600E-positive metastatic or unresectable papillary thyroid cancer refractory..., Brose [/bib_ref] , making targeted drug therapy a major option for patients with RAIR-DTC. However, this patient did not respond well to sorafenib, vemurafenib, pembrolizumab, and other targeted therapy drugs. Chemotherapy is mainly used for terminally ill patients with obvious invasive symptoms which are uncontrollable via radioactive iodine therapy or surgery. However, as most patients with RAIR-DTC are not sensitive to chemotherapy, traditional chemotherapy drugs are unsuitable [bib_ref] A phase II trial of doxorubicin and interferon alpha 2b in advanced,..., Argiris [/bib_ref].
For 15 years [fig_ref] FIGURE 1 US: -guided RFA for left cervical mass [/fig_ref] , the patient suffered from recurrent cervical lymph node metastases, which progressively increased in size and number and were not effectively controlled. Patients need an effective local therapy that includes surgery and external irradiation, with a focus on cytoreduction, to improve the quality of life (15). However, in this case, the patient's condition was severe, and the cervical metastases could not be surgically removed. The patient was not provided with any effective therapeutic regimen despite repeatedly visiting several hospitals. After the patient came to our hospital, we initially treated the cervical metastases with 125 I seed implantation and US-guided RFA.
The lymph node metastases in her left neck, which were large with liquefactive necrosis, showed a positive therapeutic effect in the early stage of 125 I seed implantation. Due to a short radiation distance, the 125 I seeds can effectively kill tumor cells in lesions with little damage to surrounding tissues, thus achieving highly conformal brachytherapy [bib_ref] Clinical study on clinical study on 125I seeds implantation for colon cancer..., Ye [/bib_ref]. Nonetheless, its effect on killing cancer cells depends on precise positioning and reasonable arrangement. Studies have shown that the larger the lesion, the lower the therapeutic effect. This has been attributed to seed displacement due to liquefactive necrosis in the large lesions, which makes it impossible to precisely position and reasonably arrange the seeds. This leads to a lower conformal index of dose distribution in the lesion, thereby weakening the therapeutic effect. Furthermore, dose deposit by Right cervical lymph node metastasis. seed displacement poses a greater risk of radiation damage to healthy tissues, making inflammatory reactions more likely, forming adhesions between lesions and vital organs (such as peripheral vessels), and thus making further treatment, including replantation, more difficult. In this case, the patient had large lymph node metastases accompanied by liquefactive necrosis, and the US showed that implanted seeds had migrated into the liquid components. Considering the unsatisfactory nature of iodine seed implantation, it is important to explore new therapies for controlling local lesions. During multidisciplinary consultations, our department suggested that the liquid components of metastasis be treated with US-guided aspiration and metastasis be treated with RFA. The RAF can be used for patients with recurrent or metastatic lymph nodes after standard surgical resection and neck lyphnode dissection (17). The patient had undergone at least one operation before, the anatomical structure of the operation area has changed and adhesions would be severe. High risk and difficulty hindered the second operation. besides, the patient is not easy to accept the second operation. In addition, some patients may be ineffective to the iodine therapy, these patients requires further local surgery or RFA. RFA has the unique advantages of safety and simple operation, which provides a more reliable treatment option [bib_ref] Long-term prognosis of unilateral and multifocal papillary thyroid microcarcinoma after unilateral lobectomy..., Jeon [/bib_ref].
The RFA technique has been increasingly applied in clinical practice since it is minimally invasive, has a beneficial therapeutic effect, and is well tolerated. The RFA needle is inserted into lesions under the guidance of US, inducing a thermal effect via a highfrequency alternating current in the targeted tissues, thereby inactivating lesion cells.
In this case, the left cervical lymph node metastasis was large (33mm×19mm) and accompanied with liquefaction necrosis. 125I was not effective at first for the particles fell into the liquid components. Under the guidance of ultrasound, we aspirated the liquid components of the left cervical lesion. Considering that the liquid would be generated continuely, we did not implant the particles again, but chose to inactivate the solid parts while aspirating the liquid. RFA was then performed on metastatic tumors at safety site of the right cervical. The solid mass was deep adjacent to important cervical nerves, so 125I was selected for the following treatment.
After RFA, the metastases in the patient's left and right neck shrank. The metastases that ruptured in the upper part also shrank after the implantation of 125 I seeds. In summary, the patient obtained the beneficial effect of local control of the metastases through US-guided RFA combined with 125 I seed implantation. This is a safe method with high clinical application value.
In this case, RFA plays an important role in the safe location of fluid lymph node lesions, and seed implantation plays an important role in the subsequent treatment of lymph node lesions located deep adjacent to important cervical nerves in the right neck. The implantation of 125I seeds has good local control effect, high safety and high clinical application value for the patients with RAIR-DTC with regional lymph node metastasis. Especially, when the tumor showed liquid necrosis, the implantation position of 125I particles would be uncontrolled. If the particles fall off into the liquid components, the effectiveness will be greatly reduced. RFA has been used to treat benign thyroid nodules, papillary thyroid cancer, recurrent thyroid cancer and lymph node metastasis safely. The coagulative necrotic tissue gradually becomes smaller after the RFA operation, and the clinical symptoms caused by nodules are also significantly improved, which is safe and feasible. Besides, the solid part can be inactivated while the fluid is pumped for mixed lesions.
# Conclusions
In conclusion, patients with RAIR-DTC with accompanying cervical lymph node metastases face limited therapy options and a poor prognosis. In addition to conventional therapies, such as surgery, chemotherapy, and targeted drug treatment, new local therapies are needed, particularly for lesions with large local mass accompanied by liquefactive necrosis. For such lesions, US-guided RFA is a valuable contribution to these therapies. Our patient was treated with targeted drugs and thyrotropin inhibition, combined with US-guided RFA and 125 I seed implantation for local control, to achieve a complete clinical response for RAIR-DTC with accompanying cervical lymph node metastases.
# Data availability statement
The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.
# Ethics statement
This study was conducted with approval from the Ethics Committee of the Affiliated Cancer Hospital of Zhengzhou University & Henan Cancer Hospital. The patients/participants provided their written informed consent to participate in this study.
# Author contributions
Conception and design of the research: YZ, YS. Acquisition of data: QL. Analysis and interpretation of the data: YZ. Statistical analysis: YZ, YS. Obtaining financing: YZ. Writing of the manuscript: YZ, Critical revision of the manuscript for intellectual content: QL. All authors read and approved the final draft.
# Funding
This study was funded by the Scientific and technological project in Henan Province (No.212102310637).
[fig] FIGURE 1 US: -guided RFA for left cervical mass. [/fig]
[fig] FIGURE 2: No contrast enhancement in left cervical mass according to postoperative ultrasonography. [/fig]
|
A rapidly growing mature mediastinal teratoma with a testicular epidermoid cyst and familial Mediterranean fever
A B S T R A C TAnterior mediastinal teratomas are common and are generally characterized as slow growing tumors. Very few reports documenting rapidly growing tumors exist. Here, we describe a case of a mature teratoma showing rapid growth in 1 year treated with complete surgical resection.A 25-year-old man who had a history of familial Mediterranean fever was referred to our hospital for further evaluation and treatment of an anterior mediastinal tumor with the largest diameter of 12 cm. A followup chest computed tomography of familial Mediterranean fever performed in the previous year showed no abnormal findings, therefore, he was suspected of having a malignant mediastinal tumor. During a systemic examination, a left testicular cyst was identified incidentally. We performed a complete resection of the mediastinal tumor and the left testicular cyst simultaneously. Following this, a benign mature teratoma in the anterior mediastinum and an epidermoid cyst in left testicle were pathologically diagnosed. The postoperative course was uneventful, and no evidence of recurrence was indicated 1 year after the surgery. We should be aware of the rapid growth potential in any mature teratoma and follow up accordingly.
A 25-year-old man who had a history of familial Mediterranean fever was referred to our hospital for further evaluation and treatment of an anterior mediastinal tumor with the largest diameter of 12 cm. A followup chest computed tomography of familial Mediterranean fever performed in the previous year showed no abnormal findings, therefore, he was suspected of having a malignant mediastinal tumor. During a systemic examination, a left testicular cyst was identified incidentally. We performed a complete resection of the mediastinal tumor and the left testicular cyst simultaneously. Following this, a benign mature teratoma in the anterior mediastinum and an epidermoid cyst in left testicle were pathologically diagnosed. The postoperative course was uneventful, and no evidence of recurrence was indicated 1 year after the surgery. We should be aware of the rapid growth potential in any mature teratoma and follow up accordingly.
# Introduction
Teratomas in the anterior mediastinum are common tumors [1] that generally grow slowly; they are often asymptomatic and identified incidentally on chest X-rays [bib_ref] Benign teratomas of the mediastinum, Lewis [/bib_ref]. However, there are some reported cases of benign mature teratomas in the anterior mediastinum that demonstrated rapid growth. The detailed mechanism of rapid growth is unclear in most of the previous cases [bib_ref] Lifethreatening and rapidly growing teratoma in the anterior mediastinum, Omachi [/bib_ref] [bib_ref] Mediastinal teratoma presenting as a rapidly enlarging paracardial mass, Hussain [/bib_ref] [bib_ref] Rapidly growing mature teratoma of the mediastinum: do sex hormones affect growth..., Uyama [/bib_ref] [bib_ref] Giant mature teratoma in the mediastinum presenting with rapid growth, Fujita [/bib_ref]. Here, we describe a case of a mature teratoma showing rapid growth in one year treated effectively with complete surgical resection.
## Case report
A 25-year-old man who had a history of familial Mediterranean fever (FMF) complained of sudden chest pain. His chest X-ray showed a large mass shadow in the left hilum, and subsequent chest computed tomography (CT) revealed an anterior mediastinal tumor, with the largest diameter of 12 cm, containing multiple cysts [fig_ref] Figure 1: Radiological examinations [/fig_ref]. However, there were no abnormal findings on the follow-up chest CT of FMF performed in the previous year [fig_ref] Figure 1: Radiological examinations [/fig_ref]. He was initially suspected of having a malignant mediastinal tumor and referred to our hospital for further examination and treatment. Our differential diagnoses included malignant mediastinal tumors, such as germ cell tumor, thymic epithelial tumor, or malignant lymphoma. However, serum tumor markers including β-human chorionic gonadotropin (HCG), alpha-fetoprotein (AFP), squamous cell carcinoma antigen (SCC) and soluble-interleukin 2 receptor (IL2R) were all within normal limits. Moreover, 18F-fluorodeoxyglucose-positron emission tomography/CT (PET/CT) showed low uptake of fluorodeoxyglucose (SUV max ¼ 3.6) in the mass, indicating a benign or low-grade malignant feature [fig_ref] Figure 1: Radiological examinations [/fig_ref]. Magnetic resonance imaging (MRI) demonstrated a huge mass with multiple cysts and solid components which suggested a mature teratoma [fig_ref] Figure 1: Radiological examinations [/fig_ref]. We then considered the possibility of a rare rapidly growing mature teratoma. We thought a percutaneous needle biopsy should be avoided considering the risk of tumor rupture or dissemination; therefore, surgery was planned for diagnosis and treatment. During a preoperative systemic examination, a left testicular cyst was identified incidentally [fig_ref] Figure 2: T2-weighed MRI showing a homogenous high intensity mass [/fig_ref]. Eventually, we performed a complete resection of the left testicular cyst and the mediastinal tumor, included a part of the left lung that was adhered to the mediastinal tumor .
Microscopically, the tumor consisted of several cysts, which were lined by mature epithelium, containing sebaceous glands, nests of respiratory epithelium, cartilage, and parathyroid glands and C). There were no immature components in the specimen and therefore these findings confirmed the diagnosis of a mature teratoma. There were no characteristic findings to explain the rapid tumor growth. In addition, pathological findings of the testicular cyst, which was lined internally with squamous epithelium, revealed a testicular epidermoid cyst. The postoperative course was uneventful, and he was discharged without any complications. No disease recurrence was seen in the follow-up radiological findings 1.5 years after surgery.
# Discussion
We encountered a very rare case of a mature mediastinal teratoma showing rapid growth in 1 year accompanied by testicular epidermoid cyst, and FMF. A teratoma is a tumor that arises from germ cells and is usually found in the gonads; however, there are some extragonadal sites where teratomas can develop, including the anterior mediastinum which is the most common extragonadal site [bib_ref] Mediastinal germ cell tumors. Clinical features and biologic correlates, Nichols [/bib_ref]. Mediastinal teratomas are generally asymptomatic and commonly found incidentally on chest radiography [bib_ref] Benign teratomas of the mediastinum, Lewis [/bib_ref]. Most symptoms, including chest pain, dyspnea and cough, are related to compression of neighboring structures. When the tumor ruptures, these symptoms worsen, and some patients need to undergo an emergent operation [bib_ref] Lifethreatening and rapidly growing teratoma in the anterior mediastinum, Omachi [/bib_ref]. In previous reports, most of the tumors had a slow-growing nature, while there were five reported cases associated with rapid growing tumors [bib_ref] Lifethreatening and rapidly growing teratoma in the anterior mediastinum, Omachi [/bib_ref] [bib_ref] Mediastinal teratoma presenting as a rapidly enlarging paracardial mass, Hussain [/bib_ref] [bib_ref] Rapidly growing mature teratoma of the mediastinum: do sex hormones affect growth..., Uyama [/bib_ref] [bib_ref] Giant mature teratoma in the mediastinum presenting with rapid growth, Fujita [/bib_ref] [bib_ref] Rapid growing huge teratoma: complete surgical resection, Kim [/bib_ref]. The cause of rapid growth among the tumors was unclear; however, bleeding, rupture, female hormones and pancreatic enzymes were thought to be associated factors [bib_ref] Lifethreatening and rapidly growing teratoma in the anterior mediastinum, Omachi [/bib_ref] [bib_ref] Mediastinal teratoma presenting as a rapidly enlarging paracardial mass, Hussain [/bib_ref] [bib_ref] Rapidly growing mature teratoma of the mediastinum: do sex hormones affect growth..., Uyama [/bib_ref] [bib_ref] Giant mature teratoma in the mediastinum presenting with rapid growth, Fujita [/bib_ref]. In this case, we could not identify pathological findings of bleeding, rupture, and pancreatic tissue in the resected specimen. Also, as this patient is a male, female hormones were less likely to be the cause of rapid growth.
Testicular epidermoid cysts can be classified into a prepubertal type of teratoma according to the 2016 WHO classification [bib_ref] Testicular epidermoid cysts: a reevaluation, Anheuser [/bib_ref]. From a clinical perspective, it seems that there is a relationship between mediastinal teratomas and testicular cysts. However, the histopathological findings of a testicular cyst show limited evidence of teratoma differentiation because mesodermal and endodermal components are absent [bib_ref] Epidermoid cyst and teratoma of the testis: sonographic and histologic similarities, Maizlin [/bib_ref]. Therefore, each disease is considered to be independent. Moreover, our patient had a history of FMF. Some authors have previously reported that FMF is associated with the pathogenesis of benign and malignant lesions such as multiple pelvic cysts [bib_ref] Multiple pelvic cysts in a patient with familial Mediterranean fever: benign cystic..., Eryilmaz [/bib_ref] and pericardial cysts [bib_ref] Case images: a pericardial cyst due to familial Mediterranean fever, Çelik [/bib_ref]. Therefore, the chronic inflammation associated with FMF may have some effect on the pathogenesis or rapid growth of a teratoma. In addition, tumor rupture, often accompanied by slight fever and chest pain, would be difficult to identify since these symptoms are also caused by FMF. Hence, a mediastinal teratoma may have been missed without the radiological findings.
This case is the first report of co-existence of a mature mediastinal teratoma, testicular epidermoid cyst, and FMF. Although each disease is considered to be independent, we suspect that each condition may have some relevance at the genetic level.
# Conclusion
We encountered a very rare case of a mature mediastinal teratoma showing rapid growth. We should be aware of the rapid growth potential in any mature teratoma and follow up accordingly.
## Declaration of competing interestcoi
The authors declare that they have no conflict of interest.
# Funding acknowledgements
None to declare.
# Informed consent statement
The patient provided written informed consent.
[fig] Figure 1: Radiological examinations. A: Contrast-enhanced CT showing a large heterogenous mass shadow containing multiple cysts in the anterior mediastinum. B: There were no abnormal findings on the chest CT taken last year. C: PET-CT showing a low uptake only in the upper area of the mass. D: T2-weighed MRI showing high intensity areas predominantly consistent with multiple cysts. [/fig]
[fig] Figure 2: T2-weighed MRI showing a homogenous high intensity mass (white arrow) in the patient's left testicle. [/fig]
[fig] Figure 3, Figure 4: Macroscopic findings. The mediastinal tumor, a solid and white colored mass, was resected completely with a part of the left lung (white arrow) and thymus (yellow arrow). (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.) Pathological findings. Histopathological analysis showing multiple areas of mature epidermis with respiratory epithelium (A), sebaceous glands (B), and parathyroid glands (C) in the mediastinal mass. [/fig]
|
Evaluation of 3‐ and 4‐Phenoxybenzamides as Selective Inhibitors of the Mono‐ADP‐Ribosyltransferase PARP10
## General procedures
General Procedure A (GP A) -for the syntheses of compounds C, G, and K: In a sealed Schlenk tube, the 4-bromo-substituted arene (10 mmol) and the respective phenole derivative (10 mmol) were mixed with palladium-(II)-acetate (44 mg, 0.2 mmol, 2 mol %) and mmol, 3 mol %) under an atmosphere of argon. Then, dry tripotassium phosphate (20 mmol, generated from 4.6 g of tripotassium phosphate monohydrate by heating at 160 °C under vacuum for 16 h) followed by dry toluene (30 mL) were added. The mixture was heated to 140 °C for the described time under an atmosphere of argon. Cooling to room temperature, adding water (25 mL) and extracting with ethyl acetate (3 x 30 mL) led to a mixture, which was washed with water and dried over magnesium sulfate. After filtration, evaporation of the organic components gave a crude product mixture, and the product was purified by column chromatography using the described mixture of cyclohexane and ethyl acetate.
## General procedure b (gp b) -for the syntheses of compounds d, h, l, 15, and 31:
The ester (3 mmol) resulting from GP A and lithium hydroxide monohydrate (157.3 mg, 3.75 mmol) were kept in 36 mL of a mixture of methanol/water (v/v = 3:1) at 70 °C for 16 h. After evaporation, the residue was treated with water (5 mL), and the mixture was set to pH 3 with hydrochloric acid . Suction, washing with water, and drying at 40 °C for 16 h gave a colorless solid of the corresponding benzoic acid.
General Procedure C (GP C) -for the syntheses of compounds 1, 4, 8, 20, 21, 29, 30, 32, and 33: The benzonitrile (1.1 mmol) was stirred with sodium perborate tetrahydrate (520 mg, 3.4 mmol) in 10 mL of a mixture of acetone and water (v/v=1:1) in a sealed tube at 70 °C for 3 days. Then, the reaction mixture was cooled, evaporated in vacuum, and the residue was boiled up in water (10 mL). The suspension was filtered through a G3 glass frit, and the filter cake was washed with boiling water (3 x 5 mL). Drying at 40 °C for 16 h yielded colorless phenyoxy benzamides.
## General procedure d (gp d) -for the syntheses of compounds 5 and 22:
The respective benzoic acid (1 mmol) was suspended in a mixture of 2 mL of methanol and 5 mL of dry toluene and cooled to 0 °C. After dropwise addition of 2.17 mL of a 0.6 M solution of trimethylsilyl diazomethane in hexane the reaction was allowed to come to room temperature during 2 hours. Evaporation and column chromatography with ethyl acetate/cyclohexane (6:1) gave the colorless solids of the corresponding methyl esters.
## S10
General Procedure E (GP E) -for the syntheses of compounds 6, 12, 13, 17, and 18: The respective benzamide (0.6 mmol) was heated with Lawsson's reagent (1.32 mmol, 2.2 equiv) in hexamethyl phosphoric acid triamide (2 mL) at 80 °C (except for 17 where the temperature was 100 °C) for the specified period of time. (Note: In the preparation of 18, only 1 equiv of Lawsson's reagent was used.) Each product was isolated by direct column chromatography of the resulted reaction solution with the described mixtures of eluents to yield the corresponding thiobenzamide.
## General procedure f (gp f) -for the synthesis of compound 7:
Benzoic acid D (1.67 mmol) was suspended in dry toluene (15 mL). After dropwise addition of thionyl chloride (0.4 mL, 5.5 mmol) the mixture was heated at 120 °C for 16 h.
Evaporation of the reaction solution and drying in vacuum gave the crude acid chloride that was reacted further directly. A solution of the acid chloride in dry THF (5 mL) was cooled to 0 °C, and hydrazine hydrate (2 mL) was carefully dropped in with stirring while the reaction mixture was allowed to come to room temperature during 2 h. After extraction with ethyl acetate (3 x 20 mL), washing with water and drying over magnesium sulfate, the organic extract was filtered and evaporated in vacuum to yield benzoic acid hydrazide 7.
General Procedure G (GP G) -for the synthesis of compound 9: A mixture of benzoic acid D (0.5 mmol), magnesium chloride hexahydrate (0.5 mmol), and sodium hydrosulfide hydrate (0.5 mmol) in dimethyl formamide (5 mL) was stirred at room temperature for 3 days. After diluting with water (10 mL), the reaction mixture was set to pH 3 with 2M hydrochloric acid. Extraction with ethyl acetate (3 x 15 mL) was followed by drying over magnesium sulfate, filtration and evaporation. The product was then purified by column chromatography (ethyl acetate/cyclohexane = 2:1) to yield 9 as a bright yellowish solid.
## General procedure h (gp h) -for the syntheses of compounds 10, 19, and 25:
The respective benzoic acid (1.67 mmol) resulting from GP B was suspended in dry toluene (15 mL). After dropwise addition of thionyl chloride (0.4 mL, 5.5 mmol) the mixture was heated at 120 °C for 16 h. Evaporation and drying in vacuum gave a crude acid chloride that was dissolved in dry THF (5 mL). While cooling in an ice bath, ammonia (3 mL, 25% water solution) was carefully dropped into the solution followed by stirring for up to 2 h at room temperature. After extraction with ethyl acetate (3 x 20 mL), washing with water, and drying over magnesium sulfate the organic extract was filtered and evaporated in vacuum to yield the corresponding amide as a colorless solid.
## S11
General Procedure I (GP I) -for the syntheses of compounds 11, 23, and 26: A mixture of the respective aryl nitrile (0.5 mmol), sodium azide (0.75 mmol) and a catalytic amount of copper sulfate pentahydrate (5 mg) in dimethyl sulfoxide (3 mL) was heated at 140 °C for 2 days. After cooling, the reaction mixture was acidified with 1 mL of 2 M hydrochloric acid, extracted with ethyl acetate (3 x 10 mL) and washed with water to neutrality. The product was purified by column chromatography (ethyl acetate/ethanol = 3:1)
to yield the colorless solid tetrazoles.
General Procedure J (GP J) -for the synthesis of compound 14: Aryl nitrile C (0.5 mmol), sodium azide (0.75 mmol) and a catalytic amount of iodine (10 mg, 8 mol %) in dry dimethyl formamide (5 mL) were heated at 120 °C for 16 h. Then, the reaction mixture was cooled to room temperature, acidified with 1 mL of 2 M hydrochloric acid and extracted with ethyl acetate (3 x 10 mL). Decolorization of the organic phase was achieved by washing with a diluted solution of sodium thiosulfate. Drying over magnesium sulfate, filtration and evaporation led to a crude product mixture, and the product was purified by column chromatography (ethyl acetate/ethanol = 3:1) to yield 14 as a colorless solid.
General Procedure K (GP K) -for the synthesis of compound 16: A mixture of 4-cyano bromobenzene (A, 15 mmol), 4-cyanophenol (E, 15 mmol), potassium carbonate (15.5 mmol), copper(I) iodide (57.3 mg, 2 mol %), and ferrum-III-acetonyl acetonate (212.1 mg, 4 mol %) in dry dimethyl formamide (10 mL) was heated to 140 °C under an atmosphere of argon for 2 days. After cooling, the reaction mixture was treated with 10 mL of a 2.5 M sodium hydroxide solution and extracted with dichloromethane (3 x 30 mL). Drying over magnesium sulfate, filtration and evaporation gave a crude product mixture that was further evaporated in high vacuum at 60 °C to get rid of most of the dimethyl formamide. For crystallization the crude product was dissolved in ethyl acetate (15 mL) under reflux and cyclohexane was dropped in (about 24 mL) until a bright cloudiness occurred. Then, the mixture was slowly cooled to room temperature and put into the fridge at 2 °C for 16 h. The crystallized product was filtered off and dried at 40 °C for 16 h to yield 16 as a colorless solid.
General Procedure L (GP L) -for the syntheses of compounds 24 and 27: A mixture of the respective aryl nitrile (0.5 mmol), diethyl ammonium hydrochloride (1.75 mmol), and sodium hydrogensulfide hydrate (1.75 mmol) was stirred in dimethyl formamide S12 (5 mL) at room temperature for 3 days. Dilution with water (10 mL), acidification with 2 M hydrochloric acid to pH 3 and extraction with ethyl acetate (3 x 10 mL) led to an extract that was dried over magnesium sulfate, filtered and evaporated to give a crude product.
Purification by column chromatography (ethyl acetate/cyclohexane = 3:1) yielded the products as yellowish solids.
General Procedure M (GP M) -for the syntheses of compounds P. R, S, and 28: A mixture of the respective halo benzene (2 mmol), the substituted phenol (2.2 mmol), copper(I) iodide (0.2 mmol), cesium carbonate (4 mmol), and N,N-dimethyl glycine hydrochloride (0.6 mmol) in 1,4-dioxane (5 mL) was heated to 100 °C for 7 days. After cooling to room temperature, the reaction mixture was treated with water (10 mL), extracted with ethyl acetate (3 x 10 mL), dried over magnesium sulfate, filtered and finally evaporated to give a crude product. Purification by column chromatography (cyclohexane/ethyl acetate = 3:1) yielded the products as colorless solids.
## Analytical data
The presentation follows the sequence of depicted products shown in Schemes S1-S19
(under section 1.2.).
## 4-methyl-(4-cyanophenoxy)benzoate (c)
Following GP A, 4-bromobenzonitrile (A, 1.82 g, 10 mmol) and methyl-4-hydroxy benzoate (B, 1.52 g, 10 mmol) were reacted with heating for 3 days, and then the product was purified by column chromatography (cyclohexane/ethyl acetate = 6:1) to yield 2.32 g (92%) of C as a colorless solid. The analytical data are in accord with those from literature. Mp. 104-105 °C. IR (KBr): u = 3408,
## 4-(4-cyanophenoxy)benzoic acid (d) [s15-s18]
Following GP B, 4-methyl-(4-cyanophenoxy)benzoate (C, 760 mg, 3 mmol) yielded 667 mg (93%) of benzoic acid D as a colorless solid. .5.
## 4-(4-carbamoylphenoxy)benzoic acid (4) [s14]
Following GP C, the yield was 238 mg (93%) as a colorless solid.
## Methyl 4-(4-carbamoylphenoxy)benzoate (5)
Following GP D, the yield was 258 mg (95%) as a colorless solid.
## Methyl 4-(4-carbamothioylphenoxy)benzoate (6)
Following GP E, the reaction was performed at
## 4-(4-cyanophenoxy)benzohydrazide (7)
Following GP F, the yield of 7 was 317 mg (70%) as a colorless solid.
## 4-[4-(hydrazinecarbonyl)phenoxy]benzamide (8)
Following
## 4-(4-carbamothioylphenoxy)benzoic acid (9)
Following GP G, the yield of 9 was 120 mg (95%) as a bright yellowish solid. 1 H NMR (
## 4-(4-cyanophenoxy)benzamide (10)
Following GP H, 4-(4-cyanophenoxy)benzoic acid (D, 400 mg, 1.67 mmol) yielded 384.9 mg (96%) of benzamide 10 as a colorless solid.
## 4-[4-(1h-tetrazol-5-yl)phenoxy]benzamide (11)
Following GP I, the yield of 11 .0, 119.9.
## 4-[4-(1h-tetrazol-5-yl)phenoxy)]benzothioamide (12)
Following GP E, the reaction was performed at .6.
## 4-(4-cyanophenoxy)benzothioamide (13)
Following GP E, the yield of 13 was 97.7 mg (57%) as a bright yellowish solid. .1.
## Methyl 4-[4-(1h-tetrazol-5-yl)phenoxy]benzoate (14)
Following GP J, the yield of 14 .6.
## 4-[4-(1h-tetrazol-5-yl)phenoxy]benzoic acid (15)
Following .
## 4,4'-oxybenzonitrile (16) [s19]
Following GP K, the yield of 16 was 1.
## 4,4'-oxybenzothioamide (17)
Following GP E, the reaction was performed at 100 °C for 2 d. .8.
## 4-(4-carbamothioylphenoxy)benzamide (18)
Following GP E, the reaction was performed with 1.0 equiv of Lawsson's reagent at
## Methyl 3-(4-cyanophenoxy)benzoate (g)
Following GP A, 4-bromobenzonitrile (A, 1.82 g, 10 mmol) and methyl-3-hydroxy benzoate (F, 1.52 g, 10 mmol) were reacted by heating for 3 days. Then, the product was purified by column chromatography (cyclohexane/ethyl acetate = 7:1) to yield 2.26 g (89%) of product G as a colorless solid. The analytical data were in agreement to those reported in the literature . .
## 3-(4-cyanophenoxy)benzoic acid (h) [s24]
Following GP B, 3-methyl-(4-cyanophenoxy)benzoate (G, 760 mg, 3 mmol) yielded 683 mg (95%) of benzoic acid H as a colorless solid. 6.
## 3-(4-cyanophenoxy)benzamide (19)
Following GP H, .
## 3-(4-carbamothioylphenoxy)benzamide (24)
Following GP L, the yield of 24 was 133.4 mg (98%) as a bright yellowish solid.
## Methyl 4-(3-cyanophenoxy)benzoate (k) [s25]
Following GP A, the reaction time was 3 d, and the product was purified by column chromatography (cyclohexane/ ethyl acetate = 7:1) to yield 1.36 g (54%) of K as a colorless
## 4-(3-cyanophenoxy)benzoic acid (l) [s14]
Following GP B, the yield of L was 697 mg (97%) as a colorless solid. . Titrations of the compounds 10 and 20 on PARP2 and PARP10 as indicated (related to . Example curves are shown with data as means ± standard deviations of four replicates. Lowest and highest compound concentrations are controls without an inhibitor and without enzyme, respectively. . Solvent accessible surface areas of compound 1 (OUL35) in complex with PARP14 and PARP10 (related to .
## Biochemistry and biology
## Selectivity
## In silico modeling and structural analyses
A. and B. The 1-PARP14 and 1-PARP10 complexes, as indicated, were analyzed by molecular dynamics and the solvent accessible surface areas (SASA) of compound 1 are shown as grey volumes. The part of the protein that contains the catalytic domain and more specifically interacts with the buried part of the ligand are shown as yellow volumes.
PARP14 and PARP10 are in green and pink cartoon, while the ligand is in stick following the same color code.
C. SASA as a function of time is shown obtained from molecular dynamics simulations. PARP14 PARP10 Glide Score (kcal/mol) . Data collection and refinement statistics of the PARP15-20 complex (related to .
## Data processing refinement
Resolution ( |
Mechanical determinants of chromatin topology and gene expression
The compaction of linear DNA into micrometer-sized nuclear boundaries involves the establishment of specific three-dimensional (3D) DNA structures complexed with histone proteins that form chromatin. The resulting structures modulate essential nuclear processes such as transcription, replication, and repair to facilitate or impede their multi-step progression and these contribute to dynamic modification of the 3D-genome organization. It is generally accepted that protein-protein and protein-DNA interactions form the basis of 3D-genome organization. However, the constant generation of mechanical forces, torques, and other stresses produced by various proteins translocating along DNA could be playing a larger role in genome organization than currently appreciated. Clearly, a thorough understanding of the mechanical determinants imposed by DNA transactions on the 3D organization of the genome is required. We provide here an overview of our current knowledge and highlight the importance of DNA and chromatin mechanics in gene expression.ARTICLE HISTORY
## Overview
## Basics of three-dimensional genome organization
The mechanical contributions of genome folding into gene regulation are multi-layered involving: DNA base pairing, nucleosome formation, nucleosome organization, chromatin loops, and the formation of Topologically Associating Domains (TAD) [bib_ref] understanding and engineering chromatin as a dynamical system across length and timescales, Johnstone [/bib_ref].
(1) DNA base-pairing: DNA sequences determine DNA base pairing. The disruption of DNA base pairing occurs during processes where the genetic information is copied into an RNA (transcription) or a DNA molecule (replication). Special DNA sequences are also prone to form non-B DNA structures that are functionally important in a variety of physiological and pathological conditions [bib_ref] Supercoil-driven DNA structures regulate genetic transactions, Kouzine [/bib_ref]. A prerequisite for the formation of these structures is duplex destabilization. Finally, recognition and binding of regulatory factors depends on local DNA structure. To implement these regulatory events, the stability and stiffness of the DNA double helix might be overcome by forces and torques applied to DNA and chromatin.
(2) Nucleosome formation: The basic unit of chromatin is a nucleosome which consists of ~147 bp of core DNA wrapped nearly two times around a conserved histone-octamer composed of H2A, H2B, H3, and H4 [bib_ref] Crystal structure of the nucleosome core particle at 2.8 A resolution, Luger [/bib_ref]. Nucleosome formation is an important regulatory event in genomic processes as it interferes with protein factors binding to DNA. The affinity of the histones to DNA is assumed to provide the energy required to wrap DNA around the core histones [bib_ref] Physics behind the mechanical nucleosome positioning code, Zuiddam [/bib_ref]. Various biochemical assays and computational modeling have shown that nucleosomes are dynamic in nature and exhibit robust response to mechanical stimuli. Thus, the contribution from forces applied to the DNA might supplement or oppose the energy of histone-DNA interactions.
(3) Nucleosome organization: Individual nucleosomes are separated by a short linker DNA, giving the classical 10 nm 'beads-on -a-string' appearance. Early in vitro studies under high salt concentrations showed a transition of 10 nm to 30-nm condensed structures. However, recent experiments argue against the existence of these 30 nm structures in vivo. Super-resolution and live-cell imaging revealed an irregular transition from nucleosome free regions to 10-nm structures and to clusters of nucleosomes with different degrees of compaction [bib_ref] Dynamic chromatin organization without the 30-nm fiber, Maeshima [/bib_ref]. The interactions between distantly spaced genomic segments are influenced by the nucleosome organization.
(4) Chromatin loop: The next level of chromatin folding is loop-formation, considered as a ubiquitous feature in the regulation of genomic transactions. The base of loops display increased chromosomal contact between separated genomic regions, perhaps established by protein(s) bound on one side of the loop bridging with protein(s) bound to the other [bib_ref] The self-organizing genome: principles of genome architecture and function, Misteli [/bib_ref]. Loops ranging in size of tens kbs often bring enhancers to the targeted promoters [bib_ref] Developmental enhancers and chromosome topology, Furlong [/bib_ref]. The regulatory specificity of enhancers is thought to be controlled by their sequence and by the binding of transcription factors. In the current models, loop formation is not a spontaneous process but demands energy to bring interacting partners together.
(5) Topologically Associating Domains (TADs) formation: Development of chromosome conformation capture-based techniques (Hi-C) has shown that at a sub-megabase scale, chromatin is folded into TADs [bib_ref] A 3D map of the human genome at kilobase resolution reveals principles..., Rao [/bib_ref]. On a population level, each TAD is characterized by a relatively high number of genomic contacts. TADs preferentially encompass chromatin loops established by interaction between enhancers and gene promoters, and so confine the zone of enhancer action. Currently, the most popular model of TADs formation is loop extrusion driven by cohesin proteins and stabilized by CCCTC-binding factor (CTCF) [bib_ref] Formation of chromosomal domains by loop extrusion, Fudenberg [/bib_ref]. The importance of this mechanism has been questioned by recent super-resolution imaging experiments, which found that domain boundaries are highly stochastic at the single-cell level [bib_ref] Extensive heterogeneity and intrinsic variation in spatial genome organization, Finn [/bib_ref] [bib_ref] Live-cell imaging reveals enhancer-dependent Sox2 transcription in the absence of enhancer proximity, Alexander [/bib_ref].
Changes in genome organization at each of these levels have been linked to functional changes suggesting an intimate role of 3D-chromatin structure in genome function [bib_ref] understanding and engineering chromatin as a dynamical system across length and timescales, Johnstone [/bib_ref] [bib_ref] The self-organizing genome: principles of genome architecture and function, Misteli [/bib_ref]. Here, we aim to emphasize that reverse is also true: those genomic transactions strongly influence genome folding and organization.
## Basics of dna mechanics
DNA can be regarded as a physical object that is subject to various forces in a chromatin environment. All genomes are constantly remodeled by proteins acting on DNA that generate torsion and apply tension on the double helix [bib_ref] unpack, bend, twist, pull, push: the physical side of gene expression, Pack [/bib_ref]. The best studied mechanism for generating torsion is the progressive motion of protein complexes translocating on DNA. In particular, elongating RNA polymerase tracks along the right-handed DNA double helix causing axial rotation of the transcribed DNA relative to the polymerase. The 'twin-supercoiled domain model' proposes that hindering the free rotation of DNA ends causes the double helix to become over-twisted in front of the polymerase, and under-twisted behind the polymerase, leading to torsional stress [bib_ref] Supercoiling of the DNA template during transcription, Liu [/bib_ref]. Geometrical parameters of the double helix (overtwisting/undertwisting and associated coiling of the axis of the double helix) are referred as DNA supercoiling [fig_ref] Figure 1: Basic of DNA Supercoiling [/fig_ref]. The original model suggested that accumulation of torsional stress might occur under conditions that prevented the free rotation of DNA, or in the absence of DNA topoisomerases. DNA topoisomerases are enzymes that remove torsional stress by transiently breaking and resealing DNA strands [bib_ref] Roles of eukaryotic topoisomerases in transcription, replication and genomic stability, Pommier [/bib_ref]. About a decade later, reevaluation of this model suggested that transcription can induce significant torsional stress even in linear unanchored DNA and in the presence of active DNA topoisomerases [fig_ref] Figure 2: DNA Supercoiling in vivo [/fig_ref] , A) [bib_ref] Transport of torsional stress in DNA, Nelson [/bib_ref] [bib_ref] The functional response of upstream DNA to dynamic supercoiling in vivo, Kouzine [/bib_ref] [bib_ref] The dynamic response of upstream DNA to transcription-generated torsional stress, Kouzine [/bib_ref]. Finally, a variety of biochemical studies, single-molecule experiments, and in vivo assays allowed careful characterization of transcriptionally generated torsion [fig_ref] Figure 2: DNA Supercoiling in vivo [/fig_ref] , B) and suggested its importance in genomic transactions [bib_ref] Controlling gene expression by DNA mechanics: emerging insights and challenges, Levens [/bib_ref] [bib_ref] DNA torsion as a feedback mediator of transcription and chromatin dynamics, Teves [/bib_ref] [bib_ref] Divergent RNA transcription: a role in promoter unwinding?, Naughton [/bib_ref] [bib_ref] Interplay between DNA supercoiling and transcription elongation, Ma [/bib_ref]. The supercoiled domain model applies with minor modification to all activities that force DNA to revolve around its axis such as movement of replisome, helicases, and type I restriction enzyme activity [bib_ref] Protein tracking-induced supercoiling of DNA: a tool to regulate DNA transactions in..., Droge [/bib_ref]. DNA loop extruding enzymes might generate torsional stress inside the loop that is topologically isolated from the rest of the DNA. Selective DNA relaxation in the loop by topoisomerases would then yield a net supercoiling across the genome after loop dismantlement .
DNA is supercoiled as a result of coiling its axis around core histones on the nucleosome in a lefthanded direction [bib_ref] Crystal structure of the nucleosome core particle at 2.8 A resolution, Luger [/bib_ref]. However, this supercoiling is constrained by DNA-protein interactions and cannot be resolved by action of DNA [bib_ref] unpack, bend, twist, pull, push: the physical side of gene expression, Pack [/bib_ref] , usually ascribed to circular DNA, such as plasmids. Ligation of ends of a linear DNA results in a unconstrained planar circle of duplex DNA made of two strands. Two strands of circular DNA are interlinked and the number of interlinks is called the linking number (LK). LK can only change if one or both DNA strands are transiently cut. The linking number of 'relaxed' DNA (Lk 0 ) reflects the geometry of the double helix: each 10.5 bp of the helical repeat produces one interlink. A relaxed circular DNA with 21 helical turns has Lk = 21. Lk is a function of the twist (Tw) and writhe (Wr): Lk = Tw+Wr. In the first approximation, twist is a measure of the winding of DNA strands around each other. Therefore, for relaxed DNA shown in this figure Lk 0 = Tw 0 = 21 (A, left). Because double helix resists bending and twisting, changing the Tw is compensated by coiling of the double helix axis which is measured by Wr. If the Twist number is altered before ligation, the DNA molecule adopts a supercoiled conformation (A, right). Topologically, immobilizing the end of DNA fragment fixes the number of links between the two DNA strands, mirroring the ligation of DNA fragment to form a circle. Thus, supercoiling can also be imposed on topologically constrained noncircular DNA molecules (b). Negative supercoils (-Sc) is generated by un-twisting (B, right), while positive supercoils (+Sc) is due to over-twisting of double helix (B, left). Supercoiled DNA molecule is under torsional stress. Accordingly, transient propagation of torsional stress along the DNA axis away from its mechanical source results in dynamic supercoiling (c). Although the Lk in topologically constrained DNA cannot be changed without breaking DNA strands, several processes alter distribution of torsional stress along the molecule. Writing part of supercoiling can be manifested as plectoneme, or as toroid when constrained in a nucleosome (d). Constrained supercoiling does not impose torsional stress on adjacent regions until liberated. Torsional stress in negatively supercoiled DNA promotes strand-separation and can be released by formation of melted DNA bubble (e) or by formation of other non-B DNA structures (f). These structures form on tracts of low complexity sequences that are abundant in genomes and occurs at specific genomic locations, supporting a functional role of non-B DNA structures in genomic transactions [bib_ref] Supercoil-driven DNA structures regulate genetic transactions, Kouzine [/bib_ref]. topoisomerase until released from the nucleosome [fig_ref] Figure 1: Basic of DNA Supercoiling [/fig_ref]. Chromatin remodeler complexes modify, slide, or remove nucleosomes [bib_ref] Collaboration through chromatin: motors of transcription and chromatin structure, Gamarra [/bib_ref]. Such reorganization of eukaryotic chromatin releases negative DNA supercoiling [bib_ref] Formation of nucleosomes on positively supercoiled DNA, Clark [/bib_ref]. In addition, in vitro experiments show that in the process of nucleosome destabilization, the chromatin remodelers might generally and directly introduce negative torsional stress into the DNA [bib_ref] Generation of superhelical torsion by ATP-dependent chromatin remodeling activities, Havas [/bib_ref]. It should be noted that transcription continuously generates DNA supercoiling at a much higher rate in comparison with one-step removal of nucleosomes. Consequently, in vivo confirmation of the functional role of DNA supercoiling generated during nucleosome reorganization is scarce and largely qualitative [bib_ref] Regulation of CSF1 promoter by the SWI/SNF-like BAF complex, Liu [/bib_ref] [bib_ref] Cooperative activity of BRG1 and Z-DNA formation in chromatin remodeling, Liu [/bib_ref]. The intimate relationship between nucleosome structure and supercoiling indicates that DNA torsional stress has a strong impact on nucleosome structure and stability [bib_ref] DNA torsion as a feedback mediator of transcription and chromatin dynamics, Teves [/bib_ref].
## Step by step transcription
The transcriptional machinery moves to productive RNA elongation through multiple steps: nucleosome remodeling at the promoters; transcription factor binding; RNA Polymerase II (Pol II) recruitment; melting of DNA during open promoter complex formation; promoter clearance; promoter-proximal pausing of Pol II; pause release and RNA elongation [bib_ref] Understanding transcription across scales: from base pairs to chromosomes, Vos [/bib_ref]. Gene expression is the outcome of regulation at each of these steps as well as promoter-enhancer interactions and TAD formation [bib_ref] The self-organizing genome: principles of genome architecture and function, Misteli [/bib_ref] [bib_ref] Understanding transcription across scales: from base pairs to chromosomes, Vos [/bib_ref]. Key to understand gene expression is an in-depth knowledge of how chromatin structure can be dynamically reorganized [bib_ref] understanding and engineering chromatin as a dynamical system across length and timescales, Johnstone [/bib_ref]. Transcriptional control in prokaryotes and eukaryotes is markedly different because the eukaryotic genome is packed into chromatin [bib_ref] Fundamentally different logic of gene regulation in eukaryotes and prokaryotes, Struhl [/bib_ref] [bib_ref] Old cogs, new tricks: the evolution of gene expression in a chromatin..., Talbert [/bib_ref]. This difference is essential for the diverse pattern of eukaryotic gene expression. While prokaryotic cells possess many analogous mechanisms that translates mechanical stimuli on DNA to gene regulation [bib_ref] The regulatory role of DNA supercoiling in nucleoprotein complex assembly and genetic..., Muskhelishvili [/bib_ref] [bib_ref] A common topology for bacterial and eukaryotic transcription initiation?, Travers [/bib_ref] , we put aside the bacteria kingdom in our review. Here, our goal is to demonstrate that DNA mechanical constraints introduced into the chromatin by transcription work as a feedback mechanism to regulate gene expression across multiple levels of 3D genome organization.
## Transcription factors binding and pol ii recruitment
Early plasmid-based experiments showed that negatively supercoiled DNA yields higher levels of gene expression than relaxed, suggesting that this topological state creates a favorable environment for general transcription factors (TFs) and RNA polymerases [bib_ref] Topological effects of the TATA box binding protein on minicircle DNA and..., Kahn [/bib_ref] [bib_ref] Negative supercoiling of DNA facilitates an interaction between transcription factor IID and..., Mizutani [/bib_ref] [bib_ref] DNA topology and a minimal set of basal factors for transcription by..., Parvin [/bib_ref]. In vitro transcription performed with a minimal set of factors showed that DNA supercoiling facilitates the binding of transcription factors TFIID and TATA binding protein (TBP) to the promoter [bib_ref] Topological effects of the TATA box binding protein on minicircle DNA and..., Kahn [/bib_ref] [bib_ref] Negative supercoiling of DNA facilitates an interaction between transcription factor IID and..., Mizutani [/bib_ref]. Importantly, TBP is a key component of the transcription initiation machinery, considered as a major interaction hub within the preinitiation complex (PIC) [bib_ref] Structure and mechanism of the RNA polymerase II transcription machinery, Schier [/bib_ref]. The dissection of the different steps in transcription reveals that DNA supercoiling promotes DNA melting and consequent formation of an open complex for transcription initiation [bib_ref] DNA topology and a minimal set of basal factors for transcription by..., Parvin [/bib_ref]. The highly specific binding of TFs to their corresponding DNA targets is established by the direct readout of the target sequence as well as by the geometry of the double helix. The shape of the DNA constitutes, in fact, a second constraint recognized through shape readout mechanisms [bib_ref] P22 c2 repressor-operator complex: mechanisms of direct and indirect readout, Watkins [/bib_ref] [bib_ref] Specifically bound BZIP transcription factors modulate DNA supercoiling transitions, Horberg [/bib_ref]. Evidently, DNA supercoiling changes the geometry of the double helix modulating the affinity of certain TFs to DNA. Conversely, TF binding by itself has the capacity to modify DNA response to supercoiling and modulate the affinity of other factors for their targets on the promoter [bib_ref] Specifically bound BZIP transcription factors modulate DNA supercoiling transitions, Horberg [/bib_ref] [bib_ref] Sequencespecific dynamics of DNA response elements and their flanking sites regulate the..., Horberg [/bib_ref]. Although it is well established that TF-DNA interactions is key to transcriptional control in eukaryotic cells, our understanding of the mechanistic and dynamic aspects of these interactions is still somewhat rudimentary.
Promising advances in technology have enhanced our understanding of the role of DNA supercoiling in the specific targeting of TFs to the chromatin. A recent genome-wide study has compared the occupancy of various chromatin-binding proteins (by ChIP-seq) and DNA supercoiling map obtained by psoralen intercalation analysis (TMP-seq) [bib_ref] Satb1 integrates DNA binding site geometry and torsional stress to differentially target..., Ghosh [/bib_ref]. From a cohort of 10 proteins with different functions, all 4 transcriptional factors Fos, Jun, JunB1 and Satb1 have been shown to have sharp preference for localization on undertwisted, negatively supercoiled DNA. This result [bib_ref] Transport of torsional stress in DNA, Nelson [/bib_ref]. Although this supercoiling regulates the variety of DNA transactions, excessive DNA supercoils will halt the further progression of translocating complex if not properly resolved. (b) Methods for detection of DNA supercoiling in vivo. Top panel: DNA supercoiling have been most frequently probed with psoralen. Psoralen freely crosses cellular membranes, intercalates between DNA bases and forms crosslinks between the two strands when exposed to UV light [bib_ref] A method for genome-wide analysis of DNA helical tension by means of..., Bermudez [/bib_ref]. It has a different preference for relaxed, positively supercoiled, and negatively supercoiled DNA (blue curve). Taking advantage of this psoralen property, supercoiled DNA have been mapped in bacteria, yeast, Drosophila, and human cells [bib_ref] Transcriptiondependent dynamic supercoiling is a short-range genomic force, Kouzine [/bib_ref] [bib_ref] Transcription-generated torsional stress destabilizes nucleosomes, Teves [/bib_ref] [bib_ref] Transcription forms and remodels supercoiling domains unfolding large-scale chromatin structures, Naughton [/bib_ref] [bib_ref] Genome scale patterns of supercoiling in a bacterial chromosome, Lal [/bib_ref]. Recently developed GapR-seq assay is based on the ability of the bacterial protein GapR to preferentially recognize overtwisted DNA (green curve). Chromatin immunoprecipitation of GapR combined with high-throughput sequencing was used to generate maps of positive supercoiling in bacteria and yeast [bib_ref] Highresolution, genome-wide mapping of positive supercoiling in chromosomes, Guo [/bib_ref]. Detection of topoisomerase activity sites (Middle panel) and non-B DNA structures (Bottom panel) are also powerful methods to predict DNA supercoiling in vivo [bib_ref] RNA Polymerase II regulates topoisomerase 1 activity to favor efficient transcription, Baranello [/bib_ref] [bib_ref] Topoisomerase 1 activity during mitotic transcription favors the transition from mitosis to..., Wiegard [/bib_ref] [bib_ref] Permanganate/S1 nuclease footprinting reveals non-B DNA structures with regulatory potential across a..., Kouzine [/bib_ref]. There has been considerable concordance between the studies supporting the main prophecies of the twin-supercoiled domain model: negative torsional stress accumulated at the upstream promoter region of the active genes, while positive torsional stress accrues in a transcription-dependent manner in gene bodies and downstream to the 3' ends of genes. -Sc (negatively supercoiled DNA); R (relaxed DNA); +Sc (positively supercoiled DNA). Blue triangle (Non-B DNA). might represent only the tip of the iceberg and the dependence of DNA recognition by regulatory factors on mechano-sensors might be much broader than we imagine. Of note is also the discovery of very high conformational diversity of individual negatively supercoiled DNA minicircles by high resolution atomic force microscopy and molecular dynamics simulations [bib_ref] Base-pair resolution analysis of the effect of supercoiling on DNA flexibility and..., Pyne [/bib_ref]. This diversity is proposed to allow many structural perturbations which could better accommodate the binding of molecular partner. In addition, a cooperative effect between global DNA conformation and molecular recognition of the short sequences has been observed [bib_ref] Base-pair resolution analysis of the effect of supercoiling on DNA flexibility and..., Pyne [/bib_ref] , which clearly indicates how information from the binding of a TF can be transferred through DNA to a distal TF by the imposition of torsional stress through double helix.
Genome-wide studies show that TFs bind to sites that are largely cleared of nucleosomes [bib_ref] Bidirectional transcription arises from two distinct hubs of transcription factor binding and..., Scruggs [/bib_ref]. Nucleosomes are dynamic structures that must be modified for the precise control of gene expression [bib_ref] Nucleosome-mediated cooperativity between transcription factors, Mirny [/bib_ref]. Current evidence favors active nucleosome eviction or depletion by factors like SOX2 and OCT4 and/or by multiple nucleosome remodeling factors [bib_ref] Mammalian ISWI and SWI/SNF selectively mediate binding of distinct transcription factors, Barisic [/bib_ref] [bib_ref] Mechanisms of action and regulation of ATP-dependent chromatin-remodelling complexes, Clapier [/bib_ref]. Common thinking is that nucleosome remodeling provides a mean of regulating genomic accessibility [bib_ref] Understanding transcription across scales: from base pairs to chromosomes, Vos [/bib_ref] , allowing DNA binding proteins to gain access to their otherwise nucleosome-protected target sites. As discussed above, DNA accessibility alone is not the only determinant of DNA-protein interaction. Another key factor is the affinity between DNA and regulatory factors that may be modified by mechanical forces acting on DNA. What forces act on DNA prior to TFs binding? Core histone rearrangement by all Snf2p-related nucleosome remodelers and/or acetylation by factors such as p300/CBP generate torsional tension in DNA by un-restraining negative supercoils held by nucleosomes [fig_ref] Figure 1: Basic of DNA Supercoiling [/fig_ref] [bib_ref] Generation of superhelical torsion by ATP-dependent chromatin remodeling activities, Havas [/bib_ref] [bib_ref] Protein lysine acetylation by p300/CBP, Dancy [/bib_ref] [bib_ref] Role of histone N-terminal tails and their acetylation in nucleosome dynamics, Morales [/bib_ref] [bib_ref] Histone acetylation reduces nucleosome core particle linking number change, Norton [/bib_ref]. Torsional stress generated by ATP-driven remodelers is high enough to drive transition from B-DNA into unusual DNA conformations, so-called non-B DNAs [bib_ref] Supercoil-driven DNA structures regulate genetic transactions, Kouzine [/bib_ref]. It has been shown that the silent CSF1 promoter is activated by the chromatin remodeling enzyme BRG1, which removes the nucleosome from the promoter and drives Z-DNA formation [bib_ref] Regulation of CSF1 promoter by the SWI/SNF-like BAF complex, Liu [/bib_ref] [bib_ref] Cooperative activity of BRG1 and Z-DNA formation in chromatin remodeling, Liu [/bib_ref]. Z-DNA is then involved in locally preserving the remodeled state of the chromatin. All these considerations lead to a model where local changes in chromatin structure during transcription initiation, introduce torsional stress into the DNA which favors loading of TFs, RNA polymerase II recruitment and finally PIC formation [fig_ref] Figure 3: Chromatin mechanics and gene expression [/fig_ref] , A). If this model is true, then the first punch of DNA torsional stress required for transcription activation could be created not only from nucleosome eviction but also from upstream transcriptional activity. Indeed, a classical study demonstrated that gene expression can be switched ON by the negative supercoiling diffusing from a nearby divergent promoter [bib_ref] Local domains of supercoiling activate a eukaryotic promoter in vivo, Dunaway [/bib_ref]. Later, this observation and additional studies led to the elegant hypothesis that widespread divergent transcription initiation in the mammalian genome is necessary to keep promoters under torsional stress to facilitate transcription factor binding and ultimately mRNA production [bib_ref] The functional response of upstream DNA to dynamic supercoiling in vivo, Kouzine [/bib_ref] [bib_ref] Divergent RNA transcription: a role in promoter unwinding?, Naughton [/bib_ref] [bib_ref] Divergent transcription: a new feature of active promoters, Seila [/bib_ref]. In accord with this prediction, identification of promoters supporting divergent transcription in the mouse genome revealed a high level of TFs binding [bib_ref] Bidirectional transcription arises from two distinct hubs of transcription factor binding and..., Scruggs [/bib_ref].
## Open complex formation
Once the PIC is formed, the promoter DNA is melted locally, and the template DNA strand is stabilized within the Pol II active site forming a transcriptional bubble 12-15 base pairs long [bib_ref] Double-stranded DNA translocase activity of transcription factor TFIIH and the mechanism of..., Fishburn [/bib_ref] [bib_ref] Near-atomic resolution visualization of human transcription promoter opening, He [/bib_ref] [bib_ref] Promoter distortion and opening in the RNA polymerase II cleft, Dienemann [/bib_ref]. DNA melting is favored by the protein complex TFIIH. The XPB subunit of TFIIH translocates on double helix away from the PIC. Because translocation is constrained due to interaction of TFIIH with other PIC components, the promoter DNA is rotated, leading to DNA untwisting [bib_ref] Double-stranded DNA translocase activity of transcription factor TFIIH and the mechanism of..., Fishburn [/bib_ref] [bib_ref] Near-atomic resolution visualization of human transcription promoter opening, He [/bib_ref]. Torsional stress results in disruption of base pairing, creating a melted DNA bubble that is fed into the Pol II active site. Thus, the formation of an open complex is based on the DNA mechanical constraints. However, this mechanism is not universal [bib_ref] Transcription without XPB establishes a unified helicase-independent mechanism of promoter opening in..., Alekseev [/bib_ref] , and the degree of dependence on TFIIH for DNA opening varies for different promoters. Promoters that are prone to melt easily in response to negative supercoiling can initiate transcription even without TFIIH [bib_ref] Promoter distortion and opening in the RNA polymerase II cleft, Dienemann [/bib_ref]. It has been predicted and recently shown that localized disruption of base pairing is widespread in supercoiled DNA [bib_ref] Supercoiling and looping promote DNA base accessibility and coordination among distant sites, Fogg [/bib_ref] [bib_ref] B-DNA under stress: over-and untwisting of DNA during molecular dynamics simulations, Kannan [/bib_ref] , suggesting that this might be exploited as a regulatory mechanism in open complex formation. At the unstable promoters, DNA strand separation is spontaneously nucleated in under-twisted regions [bib_ref] Promoter distortion and opening in the RNA polymerase II cleft, Dienemann [/bib_ref] explaining why the translocase activity of TFIIH is not required at certain promoters and may be circumvented if promoter DNA is supercoiled [bib_ref] DNA topology and a minimal set of basal factors for transcription by..., Parvin [/bib_ref] [bib_ref] The requirement for the basal transcription factor IIE is determined by the..., Holstege [/bib_ref]. Taken together, these considerations highlight The pre-initiation complex formation often involves the recruitment of chromatin remodeling complexes and histone acetyltransferases on the promoter. Core histone rearrangement and/or acetylation release negative supercoils previously constrained by the nucleosomes. Negative supercoiling increases affinity of TFs to promoter DNA, helps recruitment of transcription machinery and assist promoter DNA melting. (b) Nucleosome destabilization in the gene body is a mechanism to achieve high elongation efficiency. Positive supercoiling in front of transcribing Pol II propagates faster than the rate of elongation. The resulting torsional stress weakens the contacts between DNA and core histone by promoting H2A/H2B dimer eviction from the nucleosomes. Chromatin responds to DNA supercoiling by confinement of gene domain. This confined state of chromatin enhances the frequency of interaction among distal transcription regulators and Pol II.
that the mechanical features of the DNA are important for open complex formation [fig_ref] Figure 3: Chromatin mechanics and gene expression [/fig_ref] , A), a regulatory step in gene transcription [bib_ref] Promoter distortion and opening in the RNA polymerase II cleft, Dienemann [/bib_ref] [bib_ref] Global regulation of promoter melting in naive lymphocytes, Kouzine [/bib_ref].
## Early elongation and pol ii promoter-proximal pausing
Once the open complex is formed, the Pol II catalyzes the formation of nascent RNA, a step known as early elongation. During this step, the transcriptional machinery still has tight contacts with promoter DNA and TFs [bib_ref] Near-atomic resolution visualization of human transcription promoter opening, He [/bib_ref]. Instead of polymerase moving forward along the DNA, downstream DNA is pulled into the early elongating complex, causing extension of transcription bubble [bib_ref] The role of the transcription bubble and TFIIB in promoter clearance by..., Pal [/bib_ref] [bib_ref] Real-time observation of the initiation of RNA polymerase II transcription, Fazal [/bib_ref]. Further progression of RNA polymerase requires promoter clearance. This necessitates the collapse of the upstream part of the extended transcriptional bubble, resulting in the abrupt reannealing of the two DNA strands [bib_ref] The role of the transcription bubble and TFIIB in promoter clearance by..., Pal [/bib_ref] [bib_ref] Real-time observation of the initiation of RNA polymerase II transcription, Fazal [/bib_ref]. Only about 15 base pairs remain melted in the Pol II active site. If DNA is negatively supercoiled, DNA can undergo spontaneous strand separation, exposing the two strands in a single-stranded bubble [bib_ref] Supercoil-driven DNA structures regulate genetic transactions, Kouzine [/bib_ref] [bib_ref] B-DNA under stress: over-and untwisting of DNA during molecular dynamics simulations, Kannan [/bib_ref]. When the two strands of the bubble reanneal, supercoiling is released back into the surrounding DNA domain [fig_ref] Figure 1: Basic of DNA Supercoiling [/fig_ref]. Thus, the collapse of extended transcriptional bubble imposes high torsional stress on promoter. While the exact molecular mechanism driving promoter clearance is currently unknown [bib_ref] Promoterproximal pausing of RNA polymerase II: a nexus of gene regulation, Core [/bib_ref] , it is reasonable to propose that the high level of DNA supercoiling released on the promoter DNA drives extensive reorganization of both promoter and elongating complexes enabling promoter escape. Keeping the promoter under torsional stress might also be a prerequisite for the next round of transcription initiation [bib_ref] Divergent RNA transcription: a role in promoter unwinding?, Naughton [/bib_ref] [bib_ref] Divergent transcription: a new feature of active promoters, Seila [/bib_ref].
Pol II escaped from the promoter produces a short, nascent RNA before it usually pauses 50 to 100 bp downstream from the transcription start site (TSS) [bib_ref] Promoterproximal pausing of RNA polymerase II: a nexus of gene regulation, Core [/bib_ref]. This promoter-proximal pausing is considered as a rate-limiting step in the regulation of transcription [bib_ref] Getting up to speed with transcription elongation by RNA polymerase II, Jonkers [/bib_ref]. A variety of factors converge to establish paused Pol II and mediate its release. In current models, the sequence of RNA-DNA hybrid, DNA elements around paused site, stability of nucleosomes and different positive and negative elongation factors such as NELF, DSIF, and P-TEFb can each influence the efficiency of nucleotide incorporation and pausing [bib_ref] Promoterproximal pausing of RNA polymerase II: a nexus of gene regulation, Core [/bib_ref]. Early in vitro studies have shown the importance of Top1 in PIC assembly and transcription initiation [bib_ref] Topoisomerase I enhances TFIID-TFIIA complex assembly during activation of transcription, Shykind [/bib_ref] [bib_ref] DNA topoisomerase I is involved in both repression and activation of transcription, Merino [/bib_ref]. As the catalytic activity of Top1 is not required for its role in initiation, it has been suggested that Top1 plays the role of an architectural factor, stabilizing bent or irregular DNA structure within the PIC [bib_ref] Topoisomerase I enhances TFIID-TFIIA complex assembly during activation of transcription, Shykind [/bib_ref]. Psoralen-based mapping of DNA supercoiling near promoters of a human cell line revealed that paused genes have much higher level of negative supercoiling at their TSS compared with elongating genes [bib_ref] Transcriptiondependent dynamic supercoiling is a short-range genomic force, Kouzine [/bib_ref]. This might suggest that even if Top1 is localized in the vicinity of the paused Pol II, it cannot exert its ability to relax torsional tension. Torsional stress generated during early elongation can inhibit Pol II translocation by increasing Pol II stalling frequency and/or duration or by supporting inhibitory architecture within the early elongation complex. Singlemolecule assays revealed that bacterial RNA polymerase may be stalled by torsional stress that accumulates both downstream (overtwisted) or upstream (undertwisted) DNA regions [bib_ref] Transcription under torsion, Ma [/bib_ref]. It has also been shown that the pausing of RNA polymerase triggered by the torsional stress could be relieved upon release of the opposing force [bib_ref] Transcription under torsion, Ma [/bib_ref]. Although the effect of supercoiling on stalling eukaryotic Pol II is still unknown, in vivo experiments suggest that excessive torsional stress does inhibit Pol II translocation [bib_ref] Transcriptional inhibition by DNA torsional stress, Roca [/bib_ref] [bib_ref] Transcription-generated torsional stress destabilizes nucleosomes, Teves [/bib_ref]. It has been proposed that the pause and release into productive elongation are established by the specific DNA supercoiling balance at promoters [bib_ref] DNA topology and transcription, Kouzine [/bib_ref].
A recent study using human HCT116 cells has shown that Top1 was catalytically inactive when associated with the early elongation complex before pause release, but active when binding with elongating complexes, demonstrating a role for Top1 in regulating Pol II promoter proximal pausing [bib_ref] RNA Polymerase II regulates topoisomerase 1 activity to favor efficient transcription, Baranello [/bib_ref]. During this step, a positive feedback loop is established between Pol II and Top1 based on their physical interaction. Upon phosphorylation of the Pol II carboxy-terminal domain, the DNA relaxing activity of Top1 is enhanced and this, in turn, promotes pause release. This suggests that the low activity of Top1 at paused Pol II is not sufficient to remove the torsional stress that inhibits elongation, thus contributing to efficient pausing. After pause release, phosphorylated Pol II stimulates Top1 activity to remove mechanical impediments on Pol II translocation and promote productive elongation [bib_ref] RNA Polymerase II regulates topoisomerase 1 activity to favor efficient transcription, Baranello [/bib_ref].
## Elongation
Pol II released from pausing into productive elongation is a highly processive enzyme. At full speed, it can translocate up to a remarkable 5 kb per minutes [bib_ref] Getting up to speed with transcription elongation by RNA polymerase II, Jonkers [/bib_ref]. The main obstacles for efficient elongation are thought to be nucleosomes [bib_ref] Factor-Stimulated rna polymerase-ii transcribes at physiological elongation rates on naked DNA but..., Izban [/bib_ref] [bib_ref] Structure of transcribed chromatin is a sensor of DNA damage, Pestov [/bib_ref]. While traveling through a nucleosome array, the transcriptional machinery must pass a nucleosome every few seconds. However, early in vitro studies revealed that even a single nucleosome can impose a strong barrier for a translocating polymerase, highlighting the importance of mechanism/s allowing effective elongation in vivo [bib_ref] Factor-Stimulated rna polymerase-ii transcribes at physiological elongation rates on naked DNA but..., Izban [/bib_ref]. Single-molecule studies have provided unprecedented clarity in examining the structural dynamics of nucleosomes [bib_ref] Structural dynamics of nucleosomes at single-molecule resolution, Choy [/bib_ref]. It has become clear that nucleosomes take spontaneous excursions between wrapped and unwrapped DNA states displaying dynamic 'breathing' [bib_ref] Rapid spontaneous accessibility of nucleosomal DNA, Li [/bib_ref]. Pol II does not actively break contacts between histone core and DNA but waits until short stretches of DNA transiently unwrap from the core histones and then advances until the nucleosome is finally upstream of the translocating polymerase [bib_ref] Nucleosomal fluctuations govern the transcription dynamics of RNA polymerase II, Hodges [/bib_ref] [bib_ref] Nucleosomal elements that control the topography of the barrier to transcription, Bintu [/bib_ref]. This ratchet-like moving through DNA of a nucleosome can be promoted by a variety of seemingly synergistic factors such as histone marks that loosen DNA-histone contacts, elongation factors, chromatin remodelers and underlying DNA sequences.
Experimental advances illuminated the role that DNA mechanical constraints could have in regulating nucleosome stability [bib_ref] Structure of transcribed chromatin is a sensor of DNA damage, Pestov [/bib_ref] [bib_ref] Nucleosome assembly depends on the torsion in the DNA molecule: a magnetic..., Gupta [/bib_ref] [bib_ref] Torque modulates nucleosome stability and facilitates H2A/H2B dimer loss, Sheinin [/bib_ref]. Torsional stress might change the conformation of the DNA in such a way that it becomes refractory to tight interaction with histones. It has been shown that single-stranded DNA breaks present in the non-template strand strongly affect the rate of transcription through the nucleosome but not the rate of transcription of naked DNA [bib_ref] Structure of transcribed chromatin is a sensor of DNA damage, Pestov [/bib_ref]. Nicked DNA cannot be supercoiled; thus, preventing the buildup of local torsional stress which is required for nucleosome destabilization and Pol II passage through the nucleosome [bib_ref] Structure of transcribed chromatin is a sensor of DNA damage, Pestov [/bib_ref]. Single-molecule study of nucleosome assembly on topologically constrained DNA has shown that nucleosome formation is inhibited by DNA positive supercoiling [bib_ref] Nucleosome assembly depends on the torsion in the DNA molecule: a magnetic..., Gupta [/bib_ref]. The nucleosome consists of an octameric protein core, formed by central H3/H4 tetramer flanked by two H2A/H2B dimers [bib_ref] Crystal structure of the nucleosome core particle at 2.8 A resolution, Luger [/bib_ref]. In studies that examine the effects of torsion on preassembled nucleosomes, even moderate positive torsion led to almost complete loss of H2A-H2B dimers from the nucleosome core, suggesting a possible mechanism for loosening the nucleosome barrier [bib_ref] Torque modulates nucleosome stability and facilitates H2A/H2B dimer loss, Sheinin [/bib_ref]. Indeed, the eviction of one H2A/H2B dimer results in unwrapping about 40 base-pairs of DNA which decreases nucleosome stability [bib_ref] Structural analysis of the hexasome, lacking one histone H2A/ H2B dimer from..., Arimura [/bib_ref]. At physiological ionic strengths a nucleosome can block Pol II elongation, however with increasing ionic strength Pol II can bypass this block. With these conditions, in vitro transcription causes the loss of a single H2A-H2B dimer [bib_ref] Nucleosome remodeling induced by RNA polymerase II: loss of the H2A/H2B dimer..., Kireeva [/bib_ref]. Therefore, it seems that positive torsional stress induces nucleosome destabilization by dimer eviction. In vivo, the factors that increase the torsional stress experienced by DNA favor nucleosome destabilization, whereas factors that decrease topological stress favor nucleosome stabilization [bib_ref] Transcription and remodeling produce asymmetrically unwrapped nucleosomal intermediates, Ramachandran [/bib_ref] [bib_ref] Heavy transcription of yeast genes correlates with differential loss of histone H2B..., Cole [/bib_ref]. Mapping of unwrapped nucleosomes genome-wide revealed that inhibiting topoisomerase activity in vivo increased unwrapping, whereas reducing Pol II elongation decreased the unwrapping of nucleosomes, specifically within promoter-proximal regions [bib_ref] Transcription and remodeling produce asymmetrically unwrapped nucleosomal intermediates, Ramachandran [/bib_ref]. This argues that the positive torsional stress, generated in front of a transcribing Pol II, induces the loss of core histone-DNA contact to facilitate transcriptional elongation [fig_ref] Figure 3: Chromatin mechanics and gene expression [/fig_ref].
In Drosophila cells, transcription-generated positive torsional stress preferentially drives loss of contacts between DNA and promoter-distal H2A/H2B [bib_ref] Transcription and remodeling produce asymmetrically unwrapped nucleosomal intermediates, Ramachandran [/bib_ref] , mirroring early in vitro experiments [bib_ref] Mechanism of chromatin remodeling and recovery during passage of RNA polymerase II, Kulaeva [/bib_ref]. What is the reason for the asymmetrical nucleosome destabilization? The intrinsic bendability of short DNA fragments from the genome of Saccharomyces cerevisiae were analyzed by 'loop-seq' assay [bib_ref] Measuring DNA mechanics on the genome scale, Basu [/bib_ref]. As the nucleosome assembly requires extensive DNA bending [bib_ref] New DNA sequence rules for high affinity binding to histone octamer and..., Lowary [/bib_ref] , differential bendability across a nucleosome could potentially dictate which region of the nucleosome is destabilized first under positive torsional stress.
The authors found that DNA at +1 and +2 nucleosomes has higher intrinsic bendability on the promoter-proximal face than on the distal face of the nucleosome [bib_ref] Measuring DNA mechanics on the genome scale, Basu [/bib_ref]. This observation suggests that Pol II overcomes a nucleosome-imposed barrier better when promoter-distal face is destabilized by torsional stress. Interestingly, if the analyzed sequences were altered to use alternative codons for the same amino acids, the characteristic bendability pattern was lost. Thus, DNA mechanical properties dictate the asymmetry in nucleosome destabilization required for efficient elongation in vivo. These properties are under selective pressure to preserve nucleosome response to the torsional stress [bib_ref] Old cogs, new tricks: the evolution of gene expression in a chromatin..., Talbert [/bib_ref].
In principle, the diffusion of torsional stress through the chromatin should dynamically affect the organization of chromatin. Simplified modeling suggested that the wavefront of the altered chromatin progresses ahead of an elongating Pol II ~10 times faster than the rate of elongation [bib_ref] Transcription within condensed chromatin: steric hindrance facilitates elongation, Becavin [/bib_ref]. This model is based on the ability of a nucleosome to adopt different entry and exit linker DNA configurations under a small positive torsional stress [bib_ref] Synergistic coordination of chromatin torsional mechanics and topoisomerase activity, Le [/bib_ref] [bib_ref] Nucleosome chiral transition under positive torsional stress in single chromatin fibers, Bancaud [/bib_ref] [bib_ref] Nucleosome assembly dynamics involve spontaneous fluctuations in the handedness of tetrasomes, Vlijm [/bib_ref]. Nucleosome conformational changes help smooth the elongation process by buffering torsional stress experienced by DNA [bib_ref] Chromatin fibers stabilize nucleosomes under torsional stress, Kaczmarczyk [/bib_ref]. This wavefront is expected to stop at barriers which prevent torsional stress diffusion, such as insulators and boundary elements. A pioneering study of chromatin architecture at the Hsp70 gene of D. melanogaster provides evidence for rapid nucleosome disruption across the entire gene within 30 after activation, much faster than the rate of Pol II transcription [bib_ref] Rapid, transcription-independent loss of nucleosomes over a large chromatin domain at Hsp70..., Petesch [/bib_ref]. The nucleosome disruption extends beyond Hsp70 and stops at insulating boundary elements. Subsequently, direct measurement of DNA torsional stress across genome of D. melanogaster revealed that accumulation of torsional stress results in increased nucleosome turnover, providing direct evidence for an in vivo influence of DNA torsion on nucleosome dynamics [bib_ref] Transcription-generated torsional stress destabilizes nucleosomes, Teves [/bib_ref].
## Enhancer-promoter communication
Large-scale chromatin movements have been shown to depend on transcription and topoisomerase activity, implicating DNA supercoiling in 3D dynamics of the genome [bib_ref] Micron-scale coherence in interphase chromatin dynamics, Zidovska [/bib_ref]. Recent progress in super-resolution imaging combined with singlenucleosome tracking promises to improve our understanding of this phenomenon [bib_ref] Dynamic interplay between enhancer-promoter topology and gene activity, Chen [/bib_ref] [bib_ref] Real-Time imaging of a single gene reveals transcription-initiated local confinement, Germier [/bib_ref] [bib_ref] Single nucleosome imaging reveals loose genome chromatin networks via active RNA polymerase..., Nagashima [/bib_ref]. Contrary to the common view that transcribed regions are more open, in human cells, transcription activation resulted in threefold more confinement of an mRNA-producing gene domain within minutes [bib_ref] Real-Time imaging of a single gene reveals transcription-initiated local confinement, Germier [/bib_ref]. This confinement is consistent with the response of chromatin to the axial rotation of DNA observed in single-molecule experiments [fig_ref] Figure 3: Chromatin mechanics and gene expression [/fig_ref]. Rotations that produce positive supercoiling dramatically compact the chromatin fiber [bib_ref] Nucleosome chiral transition under positive torsional stress in single chromatin fibers, Bancaud [/bib_ref]. Further, constrained mobility was an immediate response to transcription initiation and was not dependent on ongoing polymerase elongation. Single-nucleosome imaging in human cells revealed that active Pol II constrains chromatin movement globally. Rapid inhibition or depletion of Pol II was able to release the chromatin constraints [bib_ref] Single nucleosome imaging reveals loose genome chromatin networks via active RNA polymerase..., Nagashima [/bib_ref]. Similar spatial compaction and temporal stabilization of chromatin in response to immediate transcription has been reported in Drosophila embryos [bib_ref] Dynamic interplay between enhancer-promoter topology and gene activity, Chen [/bib_ref]. Cross-correlation between population averaged genome-wide studies and global imaging of chromatin dynamics in single cells suggests that local Pol II activity is sufficient to alter the chromatin environment even at a distance. This process likely involves dynamic positive supercoiling acting as a means for signal propagation between TSS proximal regions and regions far downstream in gene bodies. Discovery of GapR, a bacterial protein that preferentially recognizes overtwisted DNA can be used as a tool for the imaging of positive DNA supercoiling in the gene body, and might provide direct evidence to this suggestion [bib_ref] Highresolution, genome-wide mapping of positive supercoiling in chromosomes, Guo [/bib_ref]. In fact, the GapRsequencing in yeast already has revealed that positive DNA supercoiling accumulates near the 3' ends of transcribed genes and correlates with the transcriptional activity of the gene, thus, confirming the prediction of the twin-supercoiled domain model [bib_ref] Highresolution, genome-wide mapping of positive supercoiling in chromosomes, Guo [/bib_ref]. Lastly, GapR sequencing results are in line with psoralen intercalation studies in D. melanogaster cells [bib_ref] Transcription-generated torsional stress destabilizes nucleosomes, Teves [/bib_ref].
Confinement of the transcribed locus due to positive supercoiling is expected to increase the frequency of direct interaction between distal transcription regulators and Pol II, bringing them into proximity. The question is whether cells evolved to use these phenomena for regulatory reasons. The rapid development of Hi-C methods during the last decade has enabled the detection of loci in physical proximity on a genomic scale [bib_ref] A 3D map of the human genome at kilobase resolution reveals principles..., Rao [/bib_ref]. Enhancer-promoter (E-P) interactions were detected as local loops in Hi-C maps. Although sometimes the E-P interaction is established over 100 kb distance, on an average, enhancers map ~10 kbs to their targets in the mammalian genomes [bib_ref] Developmental enhancers and chromosome topology, Furlong [/bib_ref]. Specific transcription factors bind enhancer regions and recruit Mediator, a multisubunit protein complex generally required for transcription [bib_ref] Transcription regulation by the mediator complex, Soutourina [/bib_ref]. The mediator then contacts the PIC assembled at a promoter to activate specific gene expression programs [bib_ref] Transcription regulation by the mediator complex, Soutourina [/bib_ref]. Due to the long separation between enhancer and promoter, the chromatin fiber can display random movements which can become an energetic barrier impeding the establishment of a productive communication. The most popular model of E-P communication is the looping of the intervening DNA to juxtapose the enhancer and the target promoter [fig_ref] Figure 4: Enhancer-Promoter [/fig_ref]. Different models of functional loop establishment have been proposed: linking promoter and enhancer through protein interaction; Pol II translocation; or loop extrusion driven by cohesin complexes [bib_ref] Developmental enhancers and chromosome topology, Furlong [/bib_ref]. However, the direct visualization of enhancer-promoter interaction at the single-cell level has shown high variability between cells suggesting that stable loop formation might not be required for gene activation [bib_ref] The self-organizing genome: principles of genome architecture and function, Misteli [/bib_ref] [bib_ref] Extensive heterogeneity and intrinsic variation in spatial genome organization, Finn [/bib_ref] [bib_ref] Live-cell imaging reveals enhancer-dependent Sox2 transcription in the absence of enhancer proximity, Alexander [/bib_ref]. This evidence had led to the models of E-P communication where the requirement of physical proximity of enhancers with their target promoters rather than their physical association is most important for a productive communication [fig_ref] Figure 4: Enhancer-Promoter [/fig_ref] [bib_ref] Developmental enhancers and chromosome topology, Furlong [/bib_ref] [bib_ref] The transcription factor activity gradient (TAG) model: contemplating a contact-independent mechanism for..., Karr [/bib_ref] [bib_ref] Enhancer-promoter communication: hubs or loops?, Lim [/bib_ref]. The mechanism of proximity establishment is currently unknown and subject of speculation [bib_ref] Tracking and interpreting long-range chromatin interactions with super-resolution live-cell imaging, Brandao [/bib_ref].
Active enhancers are often transcribed [bib_ref] Widespread transcription at neuronal activity-regulated enhancers, Kim [/bib_ref] [bib_ref] Transcribed enhancers lead waves of coordinated transcription in transitioning mammalian cells, Arner [/bib_ref] , but the role of enhancer transcription remains a matter of debate [bib_ref] Diversity and emerging roles of enhancer rna in regulation of gene expression..., Arnold [/bib_ref]. Functional requirement for transcription at enhancers has been recently shown in activated mouse B cells [bib_ref] Spt5-mediated enhancer transcription directly couples enhancer activation with physical promoter interaction, Fitz [/bib_ref]. The authors found that 50% of paired enhancers-promoters are coordinately expressed, and that E-P communication cannot be established if enhancer transcription is switched off, regardless of other features of the active enhancer. The reactivation of the enhancer transcription instantly activated the transcription from the target promoter. More importantly, enhancer transcription was dispensable if E-P interaction was established and the target promoter was already switched on, at least short term [bib_ref] Spt5-mediated enhancer transcription directly couples enhancer activation with physical promoter interaction, Fitz [/bib_ref]. The direct effect of gene activation on the E-P communication came from the direct visualization of the gene regulatory element and transcription at the single-cell level in Drosophila embryos [bib_ref] Dynamic interplay between enhancer-promoter topology and gene activity, Chen [/bib_ref]. Transcription from the target promoter endorses temporal stability of the proximity between the promoter and enhancer as well as spatial compaction, demonstrating how the act of transcription can dramatically affect the 3D topology of chromatin.
Thus, we propose here that a two-step pathway might occur when E-P communication is dependent on enhancer transcription [fig_ref] Figure 4: Enhancer-Promoter [/fig_ref]. First, translocation activity of Pol II at the enhancer introduces positive torsional stress and results in spatial compaction of genomic region between enhancer and promoter. This increased compaction favors enhancer-promoter proximity for sustained transcription. The second step is activation of the targeted-promoter transcription. Consequently, the function of spatial compaction is transferred to Pol II elongating along the gene. Once the second step is reached, enhancer transcription may no longer be required. Indeed, analysis of a wide range of cell types has shown that a rapid burst of enhancer transcriptional activity is frequently followed by its fast return to baseline [bib_ref] Transcribed enhancers lead waves of coordinated transcription in transitioning mammalian cells, Arner [/bib_ref].
## Transcription on a megabase scale
In the current paradigm, miswiring of enhancers to non-target promoters is prevented by the formation of large-scale, up to Mbs in length, chromatin loops and TADs [bib_ref] The self-organizing genome: principles of genome architecture and function, Misteli [/bib_ref] [bib_ref] A 3D map of the human genome at kilobase resolution reveals principles..., Rao [/bib_ref] [bib_ref] Breaking TADs: how alterations of chromatin domains result in disease, Lupianez [/bib_ref]. Methods of chromosome conformation capture have shown that these structures impose physical proximity on gene clusters and gene regulatory elements which is thought to represent an important feature in controlling gene expression [fig_ref] Figure 5: Model of chromatin topology Chromatin fibers are partitioned into topologically associating domains [/fig_ref] [bib_ref] A 3D map of the human genome at kilobase resolution reveals principles..., Rao [/bib_ref]. TADs and loops boundaries are enriched with cohesin complexes residing inside the domain, topoisomerase 2 enzyme residing outside the domain and the architectural factor CTCF at base of the domain [bib_ref] Genome organization drives chromosome fragility, Canela [/bib_ref].
The most accepted mechanism of domain formation is loop extrusion upon entrapment of the chromatin fiber by cohesin [fig_ref] Figure 5: Model of chromatin topology Chromatin fibers are partitioned into topologically associating domains [/fig_ref]. Extrusion proceeds until cohesin becomes stalled at the CTCF-bound sites. However, the mechanism that drives cohesin translocation on the extruded loop is unknown [bib_ref] Loop extrusion: theory meets single-molecule experiments, Banigan [/bib_ref]. Singlemolecule experiments have demonstrated that cohesin and its mitosis-specific analog condensin can extrude DNA loops that are kilobases in length. Importantly, loop-extrusion requires adenosine triphosphate (ATP) hydrolysis both in vitro and in vivo [bib_ref] Human cohesin compacts DNA by loop extrusion, Kim [/bib_ref] [bib_ref] The energetics and physiological impact of cohesin extrusion, Vian [/bib_ref] and is sensitive to mechanical forces. Extrusion is inhibited at <1 pN stall force [bib_ref] Human cohesin compacts DNA by loop extrusion, Kim [/bib_ref] , which is an order of magnitude TADs in the genome are detected by Hi-C method. First, cells are fixed by formaldehyde treatment, which crosslinks chromatin segments that are in proximity. After digestion with restriction enzyme(s), DNA fragments are re-ligated. Proximity between chromatin segments results in the higher incidence that fragments are ligated together (A, left). Deep sequencing of the ligated fragments and mapping sequencing reads on the genome enables genome-wide identification of contact frequencies among different genomic loci (A, right). TADs preferentially self-associate to create discrete structural blocks which appear as triangles in the Hi-C map. Within TADs, sub-domains with higher contact frequencies are formed, often representing the confinement of region between enhancer and promoter. Chromatin loop extrusion is mediated by cohesin and CTCF proteins (b). Cohesin loads on DNA and begins translocating along DNA, resulting in loop extrusion. This process halts when cohesin encounters CTCF molecules, forming a chromatin loop with cohesin and CTCF present at its base. smaller than stalling forces for RNA polymerases [bib_ref] Transcription against an applied force, Yin [/bib_ref] [bib_ref] Force and velocity measured for single molecules of RNA polymerase, Wang [/bib_ref] [bib_ref] Backtracking determines the force sensitivity of RNAP II in a factor-dependent manner, Galburt [/bib_ref]. Cohesin is also able to compact DNA in the absence of ATP, however it has a strong preference for compacting positively supercoiled DNA [bib_ref] The SMC1-SMC3 cohesin heterodimer structures DNA through supercoiling-dependent loop formation, Sun [/bib_ref]. Indeed, it was shown that cohesin recruitment in vivo is enhanced in regions where positive supercoiling is generated: ahead of transcribed Pol II [bib_ref] Highresolution, genome-wide mapping of positive supercoiling in chromosomes, Guo [/bib_ref] or in front of progressing DNA replication forks [bib_ref] Cohesin causes replicative DNA damage by trapping DNA topological stress, Minchell [/bib_ref]. Conceivably, the mechanical forces acting on transcribed DNA might be a key modulator of loop-extrusion, favoring or inhibiting domain formation [bib_ref] DNA supercoiling, topoisomerases, and cohesin: partners in regulating chromatin architecture?, Bjorkegren [/bib_ref]. While the contribution of transcription to chromatin folding is still debated, increasing evidence points to a connection between Pol II binding to chromatin and TADs/loops formation [bib_ref] Single nucleosome imaging reveals loose genome chromatin networks via active RNA polymerase..., Nagashima [/bib_ref] [bib_ref] Active chromatin and transcription play a key role in chromosome partitioning into..., Ulianov [/bib_ref] [bib_ref] Differential contribution of steady-state RNA and active transcription in chromatin organization, Barutcu [/bib_ref]. Transcription and topoisomerase activities are involved in maintaining a steady state profile of torsional stress within the chromatin regions [bib_ref] Transcriptiondependent dynamic supercoiling is a short-range genomic force, Kouzine [/bib_ref] [bib_ref] Transcription-generated torsional stress destabilizes nucleosomes, Teves [/bib_ref] [bib_ref] Transcription forms and remodels supercoiling domains unfolding large-scale chromatin structures, Naughton [/bib_ref]. The discovery that TOP2 interacts with CTCF and cohesin at the borders of chromosomal domains led to the hypothesis that the DNA supercoiling generated by transcription along with TOP2 help to form chromosomal domains [bib_ref] Topoisomerase II beta interacts with cohesin and CTCF at topological domain borders, Uuskula-Reimand [/bib_ref] [bib_ref] Transcriptioninduced supercoiling as the driving force of chromatin loop extrusion during formation..., Racko [/bib_ref].
Two recent studies shine new light on how transcription modulates proper domain formation in the chromatin. In the first study, Pol II was acutely depleted in human DLD-1 cell line to assess its contribution to genome folding [bib_ref] RNA polymerase II is required for spatial chromatin reorganization following exit from..., Zhang [/bib_ref]. Hi-C analysis of chromatin folding on G1-sorted cells after extended depletion (14 hours) revealed only a mild disruption of TADs. Hi-C data from cells depleted of Pol II for 2 hours did not reveal any changes in 3D genome organization. However, when cells were synchronized in G2, and Pol II depleted, then released via mitosis into G1, a strong and widespread destruction of domain structures was observed. The model predicting a role of DNA supercoiling in TADs formation envisages that in the absence of transcription, less cohesin will be translocated toward the CTCFbound sites [bib_ref] Transcriptioninduced supercoiling as the driving force of chromatin loop extrusion during formation..., Racko [/bib_ref]. Indeed, in Pol II depleted cells entering G1 phase, the cohesin signal at CTCF sites was significantly reduced. These results indicate that transcription is involved in reestablishing chromatin folding during mitotic exit and hence suggests that transcription-generated DNA torsional stress is partnering with cohesin and CTCF in the loop extrusion process [fig_ref] Figure 6: TADs formation Schematic showing active transcription supports TAD formation in DNA supercoiling... [/fig_ref].
In another study, chromatin loops decorated with cohesin and Pol II were visualized in human HeLa cells [bib_ref] Transcription-mediated supercoiling regulates genome folding and loop formation, Neguembor [/bib_ref]. Depletion of the cohesinreleasing factor WAPL (a human ortholog of the Drosophila wings-apart like protein (Wapl)) caused chromatin condensation in interphase cells through the process of enhanced loop extrusion [bib_ref] Wapl is an essential regulator of chromatin structure and chromosome segregation, Tedeschi [/bib_ref]. Upon WAPL knockout, transcription inhibition disrupts loop formation and alters the cohesin distribution. Remarkably, a similar outcome was observed when topoisomerases were inhibited. Because both scenarios lead to a net imbalance in the level of supercoiling, the findings indicate that fine-tuning of supercoiling in chromatin favors loops extrusion [fig_ref] Figure 6: TADs formation Schematic showing active transcription supports TAD formation in DNA supercoiling... [/fig_ref]. To validate this hypothesis, the authors probed the levels of DNA supercoiling by combining psoralen assay with imaging approaches. While high levels of psoralen intercalation were detected in the WAPL deficient cells compared to normal cells, upon inhibition of transcription and topoisomerase activity, the incorporation of psoralen was strongly decreased. Following assumptions derived from dynamic simulation studies [bib_ref] Transcriptioninduced supercoiling as the driving force of chromatin loop extrusion during formation..., Racko [/bib_ref] , the authors concluded that cohesin loop extrusion is promoted by the accumulation of transcriptiongenerated negative supercoiling. Although possible, these experiments could not clearly indicate whether positive or negative DNA supercoiling controls cohesin activity. Unfortunately, very high concentrations of biotinylated psoralen were used in these experiments. Under these conditions, psoralen binding to DNA mostly reflects chromatin structure at the nucleosome level [bib_ref] Intercalation of small molecules into DNA in chromatin is primarily controlled by..., Bosire [/bib_ref] [bib_ref] Psoralen photocrosslinking, a tool to study the chromatin structure of RNA polymerase..., Toussaint [/bib_ref]. Two orders of magnitude less psoralen would be required to assess DNA torsional stress [bib_ref] A method for genome-wide analysis of DNA helical tension by means of..., Bermudez [/bib_ref] [bib_ref] Single-Molecule techniques to study chromatin, Chanou [/bib_ref]. Nevertheless, the high affinity of psoralen for DNA in cells with increased loop extrusion, suggests global nucleosome destabilization, as expected for a positively supercoiled genome [bib_ref] Transcription-generated torsional stress destabilizes nucleosomes, Teves [/bib_ref] [bib_ref] Torque modulates nucleosome stability and facilitates H2A/H2B dimer loss, Sheinin [/bib_ref].
## Mechanical epigenetics
DNA supercoiling may be considered as an epigenetic mark, playing an important role in establishing cell identity [bib_ref] unpack, bend, twist, pull, push: the physical side of gene expression, Pack [/bib_ref]. However, the mechanism of its inheritance, how the topological information from parental cells is transmitted to daughter cells, has been debated. Two main hypotheses for the supercoiling-based mechanism of transmitting cellular memory through mitosis have surfaced: conservation of the distribution of nucleosomes constraining supercoiling in chromatin domains [bib_ref] The physics of epigenetics, Cortini [/bib_ref] , and the preservation of non-B DNA structures near the TSS of genes that are scheduled for activation in G1 [bib_ref] Marking of active genes on mitotic chromosomes, Michelotti [/bib_ref].
A highly significant study has shown now that cell fate or memory is tightly linked to positive supercoiling acquired during transcription in the previous cell cycle. If the genomic distribution of positive supercoiling is disturbed, the next cell cycle is impaired. The study shows that during mitotic transcription [bib_ref] Mitotic transcription and waves of gene reactivation during mitotic exit, Palozola [/bib_ref] , Pol II and Top1 coordinate their activities [bib_ref] Topoisomerase 1 activity during mitotic transcription favors the transition from mitosis to..., Wiegard [/bib_ref]. Deregulated Pol II-Top1 coordination or acute degradation of Top1 cause mitotic defects, cell cycle delays, and impaired transcription in the following cell cycle. Importantly, impaired Top1 activity results in a strong increase of negative supercoiling near genes active in mitosis, indicating that DNA supercoiling in mitotic chromatin is a critical dominant of cellular fate. Previous reports pointed to the importance of supercoiling in the mitotic function of condensin complexes [bib_ref] Condensins and the evolution of torsion-mediated genome organization, Hirano [/bib_ref]. Condensin is a close analog of cohesin and is also able to extrude chromatin loops, which is thought to be important for mitotic chromosome formation. Condensin-mediated loop extrusion favors positively supercoiled over negatively supercoiled and relaxed DNA [bib_ref] Human cohesin compacts DNA by loop extrusion, Kim [/bib_ref]. This preference indicates that positive supercoiling acquired in TADs and loops during the cell cycle is important for appropriate condensin function. If the required level of positive supercoiling is lost because of excessive annihilating negative supercoiling generated by mitotic transcription, the transition to the new cell cycle is delayed and the daughter cells experience a 'temporal amnesia' [bib_ref] Topoisomerase 1 activity during mitotic transcription favors the transition from mitosis to..., Wiegard [/bib_ref].
## Supercoiling -belief versus reality
Taken together, our interpretation of the available data is that: transcription and 3D genome organization are influenced by negative DNA supercoiling across short scales and by positive supercoiling at the long scales. Negative supercoiling operates at 1-10 kbs-scale [bib_ref] Transcriptiondependent dynamic supercoiling is a short-range genomic force, Kouzine [/bib_ref] [bib_ref] Transcription-generated torsional stress destabilizes nucleosomes, Teves [/bib_ref] [bib_ref] Transcription forms and remodels supercoiling domains unfolding large-scale chromatin structures, Naughton [/bib_ref] mostly near the promoter regions regulating initiation (reviewed here) and fine-tuning of transcription (reviewed elsewhere [bib_ref] Supercoil-driven DNA structures regulate genetic transactions, Kouzine [/bib_ref] [bib_ref] Divergent RNA transcription: a role in promoter unwinding?, Naughton [/bib_ref] [bib_ref] DNA topology and transcription, Kouzine [/bib_ref]. Positive supercoiling operates at longer 10-100 kbs-scale regulating (1) nucleosome conformational changes that help to smooth the elongation process by buffering mechanical stress and facilitate H2A/H2B dimer loss for nucleosome destabilization; (2) physical proximity between enhancers and their target promoters; (3) the construction of TADs by loop extrusion; and (4) propagation of transcriptional memory from parental to daughter cells.
The long-range operation of positive supercoiling is explained by single-molecule experiments which show that the chromatin fiber could accommodate DNA overtwisting without a global buildup of torque [bib_ref] Synergistic coordination of chromatin torsional mechanics and topoisomerase activity, Le [/bib_ref] [bib_ref] Nucleosome chiral transition under positive torsional stress in single chromatin fibers, Bancaud [/bib_ref] [bib_ref] Chromatin fibers stabilize nucleosomes under torsional stress, Kaczmarczyk [/bib_ref]. Without the accumulation torque, positive supercoiling escapes the relaxation activity of the most abundant cellular topoisomerase Top1, a stress-sensitive enzyme that quickly responds to changes in DNA twisting [bib_ref] Friction and torque govern the relaxation of DNA supercoils by eukaryotic topoisomerase..., Koster [/bib_ref]. Thus, the plasticity of chromatin to positive supercoiling allows for much further diffusion of supercoiling through the chromatin fiber. The topological plasticity of chromatin seems highly relevant to TADs/loop formation based on the single-molecule experiments [bib_ref] Chromatin fibers stabilize nucleosomes under torsional stress, Kaczmarczyk [/bib_ref] [bib_ref] The texture of chromatin, Kouzine [/bib_ref]. Curiously, an in vivo study using yeast mini-chromosomes shows that positive supercoiling promotes the approximation intramolecular DNA segments [bib_ref] Transcriptional supercoiling boosts topoisomerase II-mediated knotting of intracellular DNA, Valdes [/bib_ref] , a feature in accord with TADs/loops on Hi-C maps [bib_ref] A 3D map of the human genome at kilobase resolution reveals principles..., Rao [/bib_ref]. Unlike positive supercoiling, unwinding of DNA in chromatin is directly converted to under-twisting of linker DNA [bib_ref] Nucleosome chiral transition under positive torsional stress in single chromatin fibers, Bancaud [/bib_ref] [bib_ref] Chromatin fibers stabilize nucleosomes under torsional stress, Kaczmarczyk [/bib_ref] which is efficiently relaxed by Top1, limiting diffusion of negative supercoiling through the chromatin fiber [bib_ref] Transcriptiondependent dynamic supercoiling is a short-range genomic force, Kouzine [/bib_ref]. However, in the current literature, the tendency is to consider more negative supercoiling as an active player in regulation of transcription and 3D chromatin structure, while the existence of positive DNA supercoiling is considered equivocal. The recent studies proposing supercoiling as the driving force for 3D genome re-organization are often based on the loose assumption that only negative supercoiling is persistent in genomes, while positive supercoiling is rapidly and/or preferentially removed by DNA topoisomerases [bib_ref] Transcriptioninduced supercoiling as the driving force of chromatin loop extrusion during formation..., Racko [/bib_ref] [bib_ref] Transcription-mediated supercoiling regulates genome folding and loop formation, Neguembor [/bib_ref] [bib_ref] Are TADs supercoiled?, Racko [/bib_ref] [bib_ref] Chromatin-based mechanisms to coordinate convergent overlapping transcription, Inagaki [/bib_ref] [bib_ref] Entropic competition between supercoiled and torsionally relaxed chromatin fibers drives loop extrusion..., Ruskova [/bib_ref]. The basis for this assumption arises mainly from the overinterpretation of studies performed in Levens' and Roca' laboratories [bib_ref] Transcriptiondependent dynamic supercoiling is a short-range genomic force, Kouzine [/bib_ref] [bib_ref] RNA Polymerase II regulates topoisomerase 1 activity to favor efficient transcription, Baranello [/bib_ref] [bib_ref] Chromatin regulates DNA torsional energy via topoisomerase II-mediated relaxation of positive supercoils, Fernandez [/bib_ref] as well as oversimplification of psoralen-based experiments.
(1) We would like to take this opportunity to clarify that none of our work shows that Top1 localizes in front of Pol II and preferentially relaxes positive supercoiling as too often is suggested. Catalytic activation of Top1 during pause release does indicate that torsional stress contributes to efficient pausing by creating mechanical impediments [bib_ref] RNA Polymerase II regulates topoisomerase 1 activity to favor efficient transcription, Baranello [/bib_ref]. However, this stress might equally derive from either positive or negative supercoiling. Given the high topological plasticity of chromatin to the positive supercoiling, our current thinking is that positive supercoiling has very little impact on the processivity of the Pol II and is even required to establish an elongation 'friendly' environment [bib_ref] DNA torsion as a feedback mediator of transcription and chromatin dynamics, Teves [/bib_ref] [bib_ref] Synergistic coordination of chromatin torsional mechanics and topoisomerase activity, Le [/bib_ref] [bib_ref] Nucleosome chiral transition under positive torsional stress in single chromatin fibers, Bancaud [/bib_ref] [bib_ref] Chromatin fibers stabilize nucleosomes under torsional stress, Kaczmarczyk [/bib_ref]. Only transcriptions through long genes can generate enough positive supercoiling to necessitate special mechanisms to remove it [bib_ref] Topoisomerases facilitate transcription of long genes linked to autism, King [/bib_ref]. At the same time, high negative supercoiling detected near the sites of pausing suggest that this supercoiling reinforces Pol II pausing and so needs to be adjusted by activated Top1 at to commence productive elongation [bib_ref] Transcriptiondependent dynamic supercoiling is a short-range genomic force, Kouzine [/bib_ref]. In line with this rationale, it has been shown that impairment of the bromodomain chromatin factor BRD4, which serves as mediator for Top1 activation [bib_ref] RNA Polymerase II regulates topoisomerase 1 activity to favor efficient transcription, Baranello [/bib_ref] , leads to an accumulation of RNA:DNA hybrids (R-loops) [bib_ref] BRD4 prevents r-loop formation and transcription-replication conflicts by ensuring efficient transcription elongation, Edwards [/bib_ref]. R-loops formation could be dangerous to the genome, perhaps causing lethal DNA damage [bib_ref] Genome-wide map of r-loop-induced damage reveals how a subset of r-loops contributes..., Costantino [/bib_ref]. They are also known to be directly sponsored by aberrant buildup of negative supercoiling [bib_ref] Emerging roles for R-loop structures in the management of topological stress, Chedin [/bib_ref] , stressing the importance of removing excess negative supercoiling.
(2) Roca's group elaborated on the observation that transcriptionally active circular minichromosomes in yeast acquire high levels of negative supercoiling in the absence of Top1. Under variety of conditions, they concluded that Top2 removes positive supercoiling faster than negative, whereas Top1 relaxes supercoiling indiscriminately [bib_ref] Chromatin regulates DNA torsional energy via topoisomerase II-mediated relaxation of positive supercoils, Fernandez [/bib_ref]. Although these results are solid, care must be taken to not over-generalize them to the linear genome organized by different architectural and mechanical constraints. Indeed, mapping positive supercoiling by GapR-seq assay in yeast cells has indicated that positive supercoiling is a persistent feature of the yeast genome, confirming all the predictions of the 'twin-domain' model [bib_ref] Highresolution, genome-wide mapping of positive supercoiling in chromosomes, Guo [/bib_ref].
(3) DNA supercoiling has been detected genome-wide primarily through binding of psoralen to DNA inside living cells [fig_ref] Figure 2: DNA Supercoiling in vivo [/fig_ref]. Since the difference between binding affinity of psoralen to different topological forms of DNA is small, statistically significant detection of positive versus negative supercoiling requires the application of complex mathematical analysis or deep sequencing to reduce noise in genome-wide studies [bib_ref] A method for genome-wide analysis of DNA helical tension by means of..., Bermudez [/bib_ref]. The topological plasticity of chromatin further increases the difficulty of ascribing the psoralenbinding pattern to positively supercoiled DNA [bib_ref] Nucleosome chiral transition under positive torsional stress in single chromatin fibers, Bancaud [/bib_ref]. Despite the introduction of high levels of supercoiling in front of the elongating polymerase, only a fraction of it distributes into overtwisting of the double helix. Because psoralen affinity is lessened with DNA overtwisting, detecting positive supercoiling is challenging. Consequently, the conclusion that positive supercoiling is preferentially relaxed by topoisomerases in vivo has not been rigorously established [bib_ref] Transcriptioninduced supercoiling as the driving force of chromatin loop extrusion during formation..., Racko [/bib_ref] [bib_ref] Transcription-mediated supercoiling regulates genome folding and loop formation, Neguembor [/bib_ref] [bib_ref] Are TADs supercoiled?, Racko [/bib_ref]. Perhaps psoralen-based deep sequencing approaches may definitively reveal and map positive supercoiling genome wide [bib_ref] Transcription-generated torsional stress destabilizes nucleosomes, Teves [/bib_ref] ; this will allow comparison with newly developed GapR-seq assay [bib_ref] Highresolution, genome-wide mapping of positive supercoiling in chromosomes, Guo [/bib_ref] and potential decoding of DNA conformationfunction in genome biology.
# Conclusion
Our understanding of the critical role played by DNA and chromatin mechanics in genomic transactions has only recently received its due attention. The role of mechanical effects such as DNA supercoiling in eukaryotes remained an unfamiliar theme for many years, with studies mostly restricted to bacteria and plasmid systems. However, recent evidence demonstrates that virtually all DNA-based processes both control and are controlled by dynamic forces and torques acting on DNA in the chromatin. DNA supercoiling is a mediator that connects genome functions to structural reorganization of the genome, which in turn feedback to regulate the functions themselves. Incredible progress in genome-wide detection of DNA torsional constraints, theoretical studies, and single-molecular experiments have led to the conclusion that DNA supercoiling is an epigenetic mark possessing many layers of feedback regulations. Yet, although many scientists working in the fields of DNA and chromatin topology recognize the importance of DNA mechanics to most fundamental nuclear processes such as transcription, 3D genome organization, and replication, overall acceptance by the broader community remains to be achieved. Many peerreviewed papers attempt to explain genomic processes mainly through the static architecture of chromatin. These studies largely ignore the functional dynamics of chromatin and the role of DNA mechanical constraints. We argue that these studies will prove insufficient to explain the biology of chromatin until the physical characteristics and the dynamics of the material from which it is built are considered. Therefore, combined efforts of multidisciplinary studies are imperative to understand how DNA mechanics affect genome function and structure. We imagine that new tools will ultimately allow the visualization of DNA supercoiling in single cells with high spatial and temporal resolution. We hope that this review will help stimulate the curiosity and intellectual excitement of bright minds in different disciplines to join in the investigations of this exciting field.
[fig] Figure 1: Basic of DNA Supercoiling. DNA supercoiling is a physical property of the DNA double helix [/fig]
[fig] Figure 2: DNA Supercoiling in vivo (a) Twin-supercoiled domain model explaining the generation of dynamic positive supercoils ahead and negative supercoils behind of the protein complex that translocates along the DNA double helix [/fig]
[fig] Figure 3: Chromatin mechanics and gene expression (a) [/fig]
[fig] Figure 4: Enhancer-Promoter (e-p) communication There are at least three models by which an enhancer-promoter communication is established (a). The classical model (i) where transcription factors bind within their target enhancer and promoter form a stable complex between enhancer and promoter to stabilize the chromatin loop. In 'kiss-and-run' model (II), only transient physical contact between enhancer and promoter is required to regulate promoter activity. In proximity model (III), the enhancer communicates with the target promoter in a distance-dependent manner through the high local concentrations of transcription factors established by 'hub' or 'condensate' formation. Enhancer (red rectangle) is located far from the promoter (green rectangle) and may not communicate in a linear scale (B, Top panel). Bidirectional transcription at the enhancer region induces positive torsional stress resulting in confinement of region between enhancer and promoter (B, Middle panel). Enhanced spatial exploration of chromatin fiber promotes establishing functional E-P communication. Upon activation of the targeted promoter, the enhancer transcription is no longer required (B, Bottom panel). For clarity, transcription factors and Pol II complex have been omitted. [/fig]
[fig] Figure 5: Model of chromatin topology Chromatin fibers are partitioned into topologically associating domains (TADs) (A, left). [/fig]
[fig] Figure 6: TADs formation Schematic showing active transcription supports TAD formation in DNA supercoiling dependent fashion. [/fig]
|
The ethnic distribution of sickle cell disease in Sudan
Sickle cell disease (SCD) is one of the most common inherited disorders of haemoglobin in Africa and it is expected that sickle cell trait varies in frequency in different areas in Sudan. An extensive literature search was carried out accessing the US National Library of Medicine, the WHO Eastern Mediterranean Region resources, the Catalogue for Transmission Genetics in Arabs and papers and documents published in Sudan that included data on the prevalence of sickle cell anaemia and trait. Rates of SCA and trait varied in different areas in Sudan with the highest rates reported from Western and Eastern Sudan where one in every 123 children born in Messeryia tribe in Western Sudan is at risk of having SCD. High consanguinity rates and malaria endemicity are strong related factors with sickle cell gene in Sudan. This review will present what is known about the rates of sickle cell gene in different ethnic groups in Sudan.
# Introduction
Sudan includes variable ethnic groups that range from Arabs to African and Afro-Arabs tribes. These ethnic groups include groups with Negroid genetic characteristics with an established history in the area such as Nuba and Nilotes. Other groups include Arab, Hausa and Copt who migrated to the area in different times in history, as well as the Arab-negroid admixture tribes [bib_ref] Genetic variation and population structure of Sudanese populations as indicated by 15..., Babiker [/bib_ref] [bib_ref] Genetic heterogeneity among the Negroid and Arab Tribes of the Sudan, Tay [/bib_ref]. A study was carried out to analyze genetically Fur, Beja, Gaalin, Hawazma, and Messeryia tribes, which belong to different ethnic and linguistic groups in Sudan. Fur tribe has been found to have intermediate genetic characteristics between the Arabs and Negroids. Beja and Gaalin tribes have more Arab genetic characteristics when compared to Hawazma, and Messeryia tribes which have more Negroid admixture [bib_ref] Genetic heterogeneity among the Negroid and Arab Tribes of the Sudan, Tay [/bib_ref] [fig_ref] FiguresFigure 1: Map of Sudan showing different areas and the distribution of ethnic groups [/fig_ref]. With the diversity in ethnic groups, Sudan has a total of 133 different local languages that belong to three major African linguistic families, the Niger-Congo, Nilo-Saharan and Afro-Asiatic language families.The ethnic diversity, rapid increase in the population, high fertility rate, large family size, high consanguinity rate and the historical, cultural, traditional and religious background of these ethnic groups, highlight the interest of genetic studies in Sudan. First cousin marriage rates in Sudan are amongst the highest worldwide reaching 40-45% of all marriages [bib_ref] Inbreeding effects on reproductive outcome in a Sudanese population, Saha [/bib_ref] [bib_ref] Inbreeding levels in Khartoum, Saha [/bib_ref]. This review presents the rates of sickle cell gene in different ethnic groups in Sudan. The haemoglobinopathies are inherited as autosomal recessive (AR) disorders, where carrier parents could transmit the abnormal genes to their offspring. If both, the father and mother are heterozygotes for HbS, there is chance of 25 % of having a homozygous HbSS (Sickle cell anemia, SCA) child. If one parent is a carrier for HbS and the other is carrier for one of the abnormal hemoglobins, it results in a double heterozygote state. Heterozygotes are generally asymptomatic carriers (traits), while the SCD is expressed in the homozygotes and the double heterozygotes for two abnormal haemoglobin genes or HbS and the thalassaemias.
# Methods
An extensive literature search was carried out accessing the US National Library of Medicine [7], and the WHO resources including the Index Medicus for Eastern Mediterranean Region [8]. Keywords used for the search included "Sudan" combined with each of the following search terms: Sickle cell anaemia, sickle cell trait, haemoglobinopathy. The Catalogue for Transmission Genetics in Arabs, a database on genetic disorders in Arab populations maintained by the Centre for Arab Genomic Studies (CAGS) was also searched for relevant data. Papers and documents published in Sudan that included data on the prevalence of sickle cell anaemia and trait were searched and cited. There are no exclusion criteria for citing published data because of the general dearth of studies on sickle cell anaemia in Sudan.
# Results
## Sickle cell gene in sudan
The Origin of Sickle cell gene in Sudan : Based on analysis of Y-chromosome haplogroups, the sickle cell gene may have been preferentially introduced through males of migrating West African tribes, particularly Hausa-Fulani, and Bagara in the large migrations that began in the eighteenth century and escalated during the nineteenth and early twentieth century [bib_ref] Co-introgression of Y-chromosome haplogroups and the sickle cell gene across Africa's Sahel, Bereir [/bib_ref]. The haplotypes associated with the S gene in Sudan are most likely to be the Cameroon, Benin, Bantu and Senegal haplotypes rather than Saudi-Asian haplotype. Among 40 clinically and electrophoretically confirmed SCA cases, the Cameroon and Benin haplotypes accounted for 25% each of the samples [bib_ref] Relationship of the sickle cell gene to the ethnic and geographic groups..., Attalla [/bib_ref]. The most frequent haplotype among 143 chromosomes with S gene was the Cameroon (35.0%), followed by the Benin (29.4%), the Senegal (18.2%) and the Bantu (2.8%). The Indian-Arab haplotype was not observed. Three atypical haplotypes were identified in 17 patients, occurring at a combined frequency of 14.6%. One of these, found at the high frequency of 11.8%, possibly represented a new Sudan haplotype [bib_ref] Molecular analysis of the β-globin gene cluster haplotypes in a Sudanese population..., Elderdery [/bib_ref].
## Occurrence and distribution of sickle cell anemia among the sudanese
In 1950, the first case of HbS gene was reported in Sudan [bib_ref] Sickle cell disease in Middle East Arab countries, El-Hazmi [/bib_ref]. Later studies showed that sickle cell gene frequencies vary from region to another in Sudan as well as within the same region.
Central Sudan: Sickle cell gene is known to be prevalent in the Khartoum area, which is the capital of the country and situated in central Sudan. In the 1980s when drought and famine struck western Sudan, a huge number of migrations took place and many tribes settled around Khartoum. This unique situation made Khartoum a multiethnic area, with a blend of almost all the Sudanese tribes. Among 632 patients attending various clinics at the Khartoum Teaching Hospital, there were 5.1% with Hb AS and 0.8% with Hb SS [bib_ref] Tribal distribution of haemoglobinopathies in a Sudanese patient population, Elderdery [/bib_ref]. Sickle cell disease is the major haemoglobinopathy seen in the Khartoum, the capital of Sudan. This may be attributed to the migration of tribes from western Sudan as a result of drought and desertification in the 1970s and 1980s, and the conflicts in Darfur in 2005. The rate is highest in Western Sudanese ethnic groups particularly in Messeryia tribes in Darfur and Kordofan regions. In the Blue Nile area, where groups of indigenous population live, the prevalence ranges from 0-5% in addition to a rate of 16% among some immigrant tribes from western Sudan and West Africa in the area [bib_ref] Tribal distribution of haemoglobinopathies in a Sudanese patient population, Elderdery [/bib_ref]. The SCA presentation is usually severe and accompanied with major complications, and could be fatal in early childhood [bib_ref] Sickle cell disease in Middle East Arab countries, El-Hazmi [/bib_ref].
## Northern sudan:
Although the data about sickle cells gene in the north of Sudan is incomplete, it seems that this area shows a low frequency of SCA. A study conducted in the north of Sudan in Shagia and Manaseer tribes confirmed that the sickle cell gene is lower in the north of Sudan than in other areas [bib_ref] Sickle Cell Anemia-Associated Beta-Globin Mutation in Shagia and Manasir Tribes from Sudan, Podhorodecka [/bib_ref]. Shagia are partly nomadic, isolated, and an agricultural population. Therefore, it is difficult to determine significantly whether they are Arab or African. Manaseer tribe is of Arab origin. Both of them inhabit the 4th cataract region. [bib_ref] Sickle Cell Anemia-Associated Beta-Globin Mutation in Shagia and Manasir Tribes from Sudan, Podhorodecka [/bib_ref] Eastern Sudan: Eastern region of Sudan is composed of three states, Gedarif, Kasala, and Red Sea states. Most studies on sickle cell gene done in this region were conducted in Gedarif state, since the majority of the population migrated to and settled in this state in different decades in the past century. Blood samples tested for SCA among 100 individuals from different tribes in Gedarif state showed that 20 samples had HbSS, 55 samples had HbAS and 25 samples had HbAA. The results of this study cannot be generalized for the population of the area due to the low sample number. Among the population of the same area , a higher number of individuals were studied (261 from Hausa and 285 from Massaleet tribes), showing that sickle cells trait Hb AS was found in approximately 35% of study subjects in Hausa and 24% in Massaleet, whereas Hb SS was reported as 6% and 5% in Hausa and Massaleet respectively [bib_ref] Loss of balancing selection in the betaS globin locus, Salih [/bib_ref].
## Western sudan:
The presence of HbS is already well documented among Kordofan and Darfur region inhabitants, especially Albaggara, an Afro-Arab constellation of tribes with a predominantly African descent [bib_ref] A study of some genetic characteristics of the Fur and Baggara tribes..., Bayoumi [/bib_ref]. Some findings of a study conducted in Elobied hospital in north Kordofan state, showed that sickle cell trait in relatives of patients suffering from sickle cell disease (SCD) who were referred to this Hospital, was 54% of target samples, which concentrated mainly in two tribes, Bederia and Fulani. Sickle cell disease in Messeryia of Darfur and Messeryia Hummer of Kordofan showed a prevalence of 30.4% and 18% respectively. It is estimated that one in every 123 children born in Messeryia tribe is at risk of having SCD [bib_ref] Tribal distribution of haemoglobinopathies in a Sudanese patient population, Elderdery [/bib_ref]. Many indigenous tribes that inhabit Darfur region and belong to the Negroid ethnic group and are a part of Nilo-Saharan language family such as the Berge, Fur and Masaleet had the highest frequencies of the S gene among them [bib_ref] Relationship of the sickle cell gene to the ethnic and geographic groups..., Attalla [/bib_ref].
# Discussion
This review of published data on the prevalence of SC trait and disease in Sudan showed the dearth of studies addressing this issue. The available data revealed the wide range of SC disease frequencies in different areas of Sudan ranging from 0.8% in central Sudan to 30.4% in Western Sudan. The Messeryia tribe (a branch of the Baggara tribes) in Kordofan and Darfur showed the highest rate of sickle cell disease where it is estimated that one in every 123 children born is at risk of having SCD. Gedarif state in Eastern Sudan also showed high rate of sickle cell gene among the population that migrated from Western Africa and Sudan. The high prevalence of sickle cell gene in Sudan could be attributed to the following factors: Malaria is endemic in Sudan. The advantage of sickle cell trait (Hb AS) and its strong association with protection against all forms of clinical P. falciparum malaria is well established. Heterozygotes for the sickle gene (AS) are relatively protected against the danger of dying of malaria through the plausibile mechanism that in AS heterozygotes P falciparum-infected red cells sickle preferentially and are then removed by macrophages. The clinically relevant consequence of this process is to keep parasitemia relatively low in AS heterozygotes, [bib_ref] Sickle cell anaemia and malaria, Luzzatto [/bib_ref] ; The high consanguinity rates in Sudan and the rate of first cousin marriages is the highest when compared with the other Arab countries (which exceeds 40 % ), which could increase the prevalence of autosomal recessive diseases such as SCD. Moreover, the traditional tribal society is still existent in Sudan [bib_ref] Genetic disorders in the Arab world, Al-Gazali [/bib_ref] ; The lack of public health measures and services for the prevention of genetic disorders in general; The selective termination of pregnancy of an affected fetus is illegal in Sudan.
# Conclusion
There is a need to initiate systematic epidemiological studies to assess the prevalence rates of SCD and SCT in different areas in Sudan. Recently, many Arab countries have ongoing premarital screening programs for haemoglobinopathies, with success stories of reducing the birth rate of affected with SCD, such as in Bahrain [bib_ref] Campaign to control genetic blood diseases in Bahrain, Arrayed [/bib_ref]. It is recommended that such programs for premarital screening of haemoglobinopathies are initiated in Sudan, especially in known areas with high prevalence rates of SCA such as Darfur and Kordofan regions, to allow couples to take an informative decision when they are both carriers of the gene.
[fig] FiguresFigure 1: Map of Sudan showing different areas and the distribution of ethnic groups (From http://nealrauhauser.wordpress.com/2012/12/25/sudan-africasyugoslavia/sudanethnicgroups/) [/fig]
|
Impact of a school-based intervention on nutritional education and physical activity in primary public schools in Chile (KIND) programme study protocol: cluster randomised controlled trial
Background: Chile has suffered a fast increase in childhood obesity in the last 10 years. As a result, several school programmes have been implemented, however the effectiveness of these needs to be evaluated to identify and prioritize strategies to curve this trend. Methods: Cluster randomized controlled trial. Twelve primary public schools chosen at random over three regions of the country will take part in this study. The sample size consisted of a total of 1,655 children. For each region one school will be selected for each of the three nutritional intervention modes and one school will be selected as the control group. The intervention modes consist of the following:Healthy Kiosk and nutritional education (KSEAN); Optimized physical activity (AFSO); Healthy Kiosk and nutritional education (KSEAN) + optimized physical activity (AFSO); Control group.The effectiveness of each intervention will be evaluated by determining the nutritional condition of each child by measuring percentage of body fat, BMI and the z-score of the BMI. This study will also identify the eating behaviours, nutritional knowledge and fitness of each child, along with the effective time of moderate activity during physical education classes.
(Continued from previous page) Discussion: A protocol to evaluate the effectiveness of a school based intervention to control and/or reduce the rates of childhood obesity for children between 6 and 10 years of age was developed. The protocol was developed in line with the Declaration of Helsinski, the Nüremberg Code and the University of Chile Guidelines for ethical committees, and was approved by the INTA, Universidad de Chile ethical committee on Wednesday 12 March 2014. There is consensus among researchers and health and education personnel that schools are a favourable environment for actions to prevent and/or control childhood obesity. However a lack of evidence on the effectiveness of interventions to date has led some to question the wisdom of allocating resources to programmes. This is the first study of this kind in Chile and could be an important first step to provide guidance to political authorities in relation to which food and nutrition strategies to prioritize to curve this alarming trend.
Trial registration: ISRCTN32136790, registered retrospectively on 05 September 2014.
Keywords: Childhood obesity, Food education, Healthy kiosk, Physical activity, Schools
# Background
Chile has experienced an accelerated process of epidemiologic and nutritional transition. During this process, very rapid changes have taken place, going from a pre-transition condition where infectious maternal and child diseases predominated public health issues during the decade of the 60s, to a post-transitional condition where chronic, noninfectious diseases dominate (ENT) [bib_ref] Nutrition transition in Chile revisited: mid-term evaluation of obesity goals for the..., Vio [/bib_ref]. Studies have identified that this has resulted due to a broader dietary offering, changes in eating patterns and a considerable increase in sedentary behaviours/lifestyle [bib_ref] Estrategia de promoción de la salud en escolares de educación básica municipalizada..., Kain [/bib_ref] [bib_ref] Smart food policies for obesity prevention, Hawkes [/bib_ref].
The Junta de Auxilio Escolar y Becas (JUNAEB), a Chilean government organization, measures the height and weight of all 6-year-old children attending year one of primary school and found that during 2013 the obesity prevalence for this group of children was 25.3% [bib_ref] Informe Mapa Nutricional, Junaeb [/bib_ref].
The World Health Organization (WHO) has indicated that there is convincing evidence that a sedentary lifestyle along with a diet consisting of a high consumption of foods high in calories and deficient in fruits, vegetables, legumes and fat-free dairy products increases the risk of obesity. However, a suitable home and school environment that promotes the choice and consumption of healthy foods could reduce this risk in children [bib_ref] nutrition and the prevention of chronic diseases, Who [/bib_ref].
Childhood obesity is one of the main issues affecting public health, not only due to its increased incidence, but also because the obesity is maintained throughout adolescence and adulthood [bib_ref] Childhood obesity and adult morbidities, Biro [/bib_ref] [bib_ref] Long-term impact of overweight and obesity in childhood and adolescence on morbidity..., Reilly [/bib_ref]. It is associated with a higher risk of cardiovascular diseases [bib_ref] Obesity in childhood and vascular changes in adulthood: insights into the Cardiovascular..., Raitakari [/bib_ref] , diabetes [bib_ref] Relationship between obesity and diabetes in a US adult population: findings from..., Nguyen [/bib_ref] , some types of cancer [bib_ref] Overweight, obesity, and mortality from cancer in a prospectively studied cohort of..., Calle [/bib_ref] [bib_ref] Obesity and cancer risk: evidence, mechanisms, and recommendations, Vucenik [/bib_ref] , depression [bib_ref] Depression and obesity: a meta-analysis of community-based studies, De Wit [/bib_ref] , discrimination [bib_ref] Social marginalization of overweight children, Strauss [/bib_ref] and weight-related problems, as well as other illnesses, where the short-and long-term risks of childhood obesity translate to a decrease in quality of life [bib_ref] Calidad de vida relacionada con la salud y obesidad en un centro..., Barajas Gutiérrez [/bib_ref] , requiring earlier interventions that result in healthy behaviours within an obesogenic environment [bib_ref] Strategies for the prevention and control of obesity in the school setting:..., Katz [/bib_ref].
Schools have been identified as a key setting in the reduction or prevention of the prevalence of overweight and obesity [bib_ref] Preventing Childhood Obesity: Health in the Balance, Institute Of Medicine [/bib_ref] , as they provide continuous and intensive contact with children during their formative years. The school infrastructure and the physical environment, policies, programmes and staff have great potential to provide a positive influence on the health of children [bib_ref] Strategies for the prevention and control of obesity in the school setting:..., Katz [/bib_ref]. However, most research has been based on diverse strategies, covering one or more components (nutritional education, reducing time spent in front of the television, providing reading material, modifying the school menu, reducing sedentary behaviours) [bib_ref] Effectiveness of a randomized controlled lifestyle intervention to prevent obesity among Chinese..., Xu [/bib_ref] [bib_ref] Effect of school based physical activity programme (KISS) on fitness and adiposity..., Kriemler [/bib_ref] , and has not been able to show convincing results regarding the effectiveness of school-based programmes in reducing overweight and obesity [bib_ref] A systematic review of controlled trials of interventions to prevent childhood obesity..., Connelly [/bib_ref].
Even though school-based interventions have been clearly shown to be more effective than interventions in other settings, it has not been demonstrated that interventions consisting of multiple elements are more advantageous than ones consisting of a single element [bib_ref] Effect of school based physical activity programme (KISS) on fitness and adiposity..., Kriemler [/bib_ref] [bib_ref] The costs and cost-effectiveness of a school-based comprehensive intervention study on childhood..., Meng [/bib_ref] [bib_ref] Pathways: a school-based, randomized controlled trial for the prevention of obesity in..., Caballero [/bib_ref]. This has resulted in the growing need to generate highquality evidence to guide public policymaking in this area.
Since the year 2000, a series of structural and individual initiatives, linked with the promotion of healthy lifestyles to prevent obesity in the population, have been promoted in Chile. These initiatives have implemented programmes to promote healthy lifestyles in schools, such as VIDA CHILE [bib_ref] resultados y desafíos de la política de promoción de la salud en..., Salinas [/bib_ref] and the Global Strategy against Obesity (EGO-CHILE). Additionally, in 2012, the National Education Council approved changes to the school curriculum that increased the number of hours dedicated to physical activity in schools to 3 or 4 hours per week, and promulgated a new law, the "Ley de Composición Nutricional de los Alimentos y su Publicidad", which prohibits the marketing/advertising and the sale inside schools of foods high in calories, saturated fats, sugars and sodium to children under the age of 14 years.
The KIND study looks at addressing the deficiencies of previous studies and as such should enable evaluation of the effectiveness of an integral school-based intervention in diet and physical activity. The study is aimed at controlling the increase in obesity in children aged between 6 and 10 years, from a medium-low and low socio-economic status, that attend public schools. This study will be conducted in three regions of the country and will provide valuable information that should enable the development of an integral view of the dietary and nutritional status of children attending public schools in Chile.
# Methods
## Design
The proposed study is a cluster randomized controlled trial to evaluate the effect of a school-based intervention in nutritional education and physical activity over two school years in children aged between 6 and 10 years, attending primary public schools. Three modes of intervention will be implemented and the results will be measured alongside those of a control group. The intervention modes consist of the following:
Intervention 1: Healthy Kiosk and nutritional education (KSEAN); Intervention 2: Optimized physical activity (AFSO), where the physical education classes will be taken by a specialized physical education teacher or a primary teacher with a specialization in physical education. The effective class time will be a minimum of 70 min, during which half of the time should involve undertaking activities of moderate to vigorous intensity; Intervention 3: Healthy kiosk and nutritional education (KSEAN) + Optimized physical activity (AFSO); Control group: Physical education classes will follow the curriculum as indicated for the subject of physical education and health (AFS).
The study will be conducted in 12 clusters (schools) in three regions of Chile (VIII, VI and Metropolitan), distributed in four schools per region, randomly assigned to KSEAN, AFSO, KSEAN+ AFSO or control.
## Study hypotheses
Integrating school-based interventions over a period of two school years that cover nutritional education, along with the implementation of a healthy kiosk and 4 h per week of physical education classes, where 50% of the time involves physical activity of moderate to vigorous intensity, is more effective in controlling obesity in children between 6 and 10 years of age than implementing each of these interventions separately.
Incorporating nutritional education into the curriculum, where the dietary message provided in the "Chilean Dietary Guidelines"published in 2012 is conveyed, along with the implementation of healthy kiosks at schools, improves the dietary knowledge and behaviours of children between 6 and 10 years of age and guides them to choose foods low in calories inside schools, thereby helping to manage the increase in childhood obesity.
Optimized physical education and health classes, i.e. those undertaken by a specialized physical education teacher, with a total duration of 4 hours per week in separate blocks, where 50% of the effective class time involves conducting activities of moderate to vigorous intensity designed to improve the fitness levels of children aged between 6 and 10 years, helps in managing the increase in childhood obesity.
## Cluster inclusion criteria
The inclusion criteria included:
Primary public schools located within the Metropolitan Region, Region VI and Region VIII of Chile. Schools that are classed with a high school Vulnerability Index by JUNAEB (IVE ≥ 60 JUNAEB). Schools that have not been included in previous interventions with programmes that promote healthy lifestyles. Full-time co-educational schools with a minimum of 600 students. Schools that have specialized physical education teachers.
## Participant inclusion criteria
Students attending public primary schools regularly between the ages of six and ten.
## Participant exclusion criteria
Children excluded from physical education classes due to medical reasons.
## Recruitment of school (clusters)
For the purpose of this effectiveness study, all of the information regarding public schools was made available to the municipal education departments. This did not require any additional effort or resources for the purpose of recruitment.
From a total of 71 full-time public primary schools within the three regions previously indicated (17 in the Metropolitan Region, 29 in Region VI and 25 in Region VIII), schools with an IVE ≤ 60 were excluded, leaving a total of 58 schools available. Then schools that had previously undergone intervention as part of healthy lifestyle programmes were also excluded, leaving 45 schools, of which only 28 were co-educational and had more than 600 students between 6 and 10 years of age.
Teaching staff qualifications at each school were reviewed and only 18 schools had specialized physical education teachers (six in the Metropolitan Region, eight in Region VI and six in Region VIII). From these 18 schools, four schools per region were chosen at random.
## Recruitment of participants
Selection of potential participants for each intervention mode and for the control group is to be conducted after randomly selecting the four schools per region.
Meetings are to be held with representatives from the education board and the principals from the participating schools to explain in detail the stages of this research study and the activities that will be undertaken. Afterwards, each school is to invite the parents of the potential participants to meet with the research team, who will explain in detail the aim of the research study. The parents will also be provided with an information sheet detailing the nature and importance of this study and indicating explicitly that it is necessary to have the parents'/guardians' consent before a child can participate in this study.
A phone number will also be provided to the parents to use in case of any doubts or concerns related to this study. The children will be invited to participate in the project at this time. All parents that indicate that they do not want to participate will be thanked for their time and cooperation.
Children older than 8 years of age that are motivated to participate will be invited to sign an agreement, which will later be used to inform the parents and invite them to participate by signing the consent form. Both the agreement and the consent form have been previously approved by the ethics committee of the Nutrition and Food Technology Institute (INTA) from the Universidad de Chile.
Once the consent forms have been signed, a third meeting is to be organized, where the activities related to this study to be conducted during the school year will be described in detail.
Parents will be informed that this study will determine the nutritional status of each child by measuring weight, height, skinfolds and waistline. These measurements only require that the children remove their shoes and any other heavy element that they might be carrying. It will also be indicated that they will need to answer a survey aimed at determining their current dietary behaviours and that the children's fitness levels, along with the intensity of the activities conducted in the physical education classes, will be determined by conducting the following tests:
1. Grip strength: consists of squeezing a dynamometer with both right and left hands. This instrument measures the grip strength of the child [bib_ref] ALPHA-Fitness: test de campo para la evaluación de la condición física relacionada..., Ruiz [/bib_ref]. 2. Measuring the distance of a jump done with both feet together and without a run-up [bib_ref] ALPHA-Fitness: test de campo para la evaluación de la condición física relacionada..., Ruiz [/bib_ref].
3. Measuring the intensity of the physical activity by using a pedometer [bib_ref] Validity of the New lifestyles NL-1000 accelerometer for measuring time spent in..., Mcminn [/bib_ref] , which is a portable device used to count steps that will be placed on the waist of children and secured with a belt while they participate in the physical education classes.
It will be indicated to the parents that all of these tests will be conducted under the supervision of a qualified physical education teacher with experience in conducting all of the tests along with a teacher from the school. It will also be explained that all of the tests should take approximately 15 min to complete and that these will be conducted at the beginning and end of the school year during class time. The only measurement that will be taken throughout the whole year is the intensity of the physical education class. The child's nutritional status will be informed by means of a letter addressed to the parents.
It will also be indicated that the participation in this study is voluntary and that a child can choose to stop participating at any moment. Additionally, it will be made clear that the participation in this study will not incur any cost for the families of the children nor the school, and that all of the information collected will be kept confidential and will only be used for this study.
The diagram shown in [fig_ref] Figure 1: Participant recruitment process [/fig_ref] below, shows the flow of the suggested recruitment process for this study:
## Each component of the intervention modes is described below:
a) Nutritional Education Intervention: This intervention is based on the "Social Cognitive Learning Theory", which incorporates the interdependency relationship between personal characteristics, behavioural factors and environmental influences. Teachers will provide knowledge and skills in relation to choosing healthy foods at school and at home by providing learning material. Year one to year four students participating in the programme will work with learning material based on the dietary guidelines for the Chilean population that provides nutritional concepts that are reinforced with theory and practical activity, along with the creation of healthy messages. To enhance the effectiveness, the nutritional education activities will be scheduled over 16 sessions of 90 min each. As this is not part of the curriculum, the teacher in charge of each teaching unit (UTP), at each individual school, will define the time when this content will be presented to the students. The implementation of the learning material, including teacher interviews and display of the work conducted by the children in wall displays, will be supervised by a nutritionist, who is part of the research team. b) Healthy Kiosk Implementation: The model described by Bustos and colleagues in the Manual de Implementación de un Espacio y Punto de Venta Saludable en Escuelas Básicas de Chilewill be replicated in this study. This considers the following:
The construction of a Healthy Space, located within the school. This space is defined as a place that will encourage healthy lifestyles by promoting recreational activities, the sale of healthy foods and the implementation of various educational strategies aimed at modifying and strengthening healthy behaviour amongst schoolchildren. The Healthy Space will consist of a healthy kiosk, surrounded by tables, chairs and a number of games of an attractive design and painted in bright colours. The kiosk will be built in accordance with what is established in the "Reglamento Sanitario de los Alimentos"and its design will promote the visibility of the food that will be displayed in refrigerated display cases and shelves inside the kiosk. The food offering will consist of avocado and fresh cheese or tomato sandwiches, fat-free yoghurt and milk, milk-based desserts, jelly, sugar-free biscuits, cereals, fresh and dried fruits, sugar-free juices and drinks, mineral water and other snacks low in calories. Processed foods will be determined to be low in calories if the nutritional information on the packaging shows the portion sizing and the nutritional information per portion and if this does not exceed 130 Kcal, 3 g of total fats, 20 g of carbohydrates and 140 mg of sodium. For fruits and vegetables a standard portion of 150 g is considered acceptable, as is 30 g for nuts. Also, it will be established that dairy products must be fat-free or low in fat and that other drinks must be sugar-free. Strategies for promoting the sale of healthy foods and processed foods low in calories will be established based on behavioural economics, which places the food recommended by the Chilean Dietary Guidelines, such as fruits, vegetables and dairy products, amongst others, in the most visible place within the kiosk. Alongside this, positive reinforcement statements promoting the benefits of consuming healthy foods will be displayed to promote their sale. Also, the kiosks will be fitted with an information board displaying various themes related to the promotion of a healthy lifestyle and a price list of foods available for purchase.
Training will be provided for the kiosk staff in topics such as healthy eating, how to read and understand the nutritional information displayed on food packaging, hygiene, food preparation and handling, and communicational and marketing strategies for promoting the sale of healthy foods. c) Physical Activity Intervention: The optimized physical education classes will consider what is already established in the curriculum for the subject physical education and health, with four teaching hours per week, distributed in blocks of 90 min each on different days of the week. Each school selected for this mode of intervention will have a specialized physical education teacher or a teacher with specific training in physical education. Teachers will receive further training to ensure: 1) that the effective class time has a duration of 70 min; 2) that 50% of the effective time consists of activities that demand a moderate to vigorous intensity. Monitoring will be carried out of all of the physical education classes conducted for years 1 to 4, by external physical education teachers that have been previously trained and standardized. They will measure the effective class time and will select children at random to wear the pedometers during the class.
## Possible adverse effects of the physical activity intervention
A register will be designed and implemented to identify any potential adverse effect on the children taking part in the intervention that could occur as a result of the increase in physical activity from moderate to vigorous. An injury management protocol will also be designed and provided. With regard to nutritional education, a form will be designed to identify the causes for not complying with the planned activities.
## Control group
The participants in this group will participate in physical education classes that follow the specifications of the physical education and health curriculum as indicated by the Education Ministry. They will not participate in nutritional education classes and a healthy kiosk will not be implemented at their school. The control group participants will be provided with the same level of support and information, and will be subject to the same evaluations as the participants of the intervention groups in this study.
At the end of the study all of the participants will be invited to take part in nutritional education talks and all participating schools will have the infrastructure of their kiosks repaired. In parallel, teachers from all participating schools will be provided with training in nutritional education and optimized physical education classes.
## Outcome measures healthy kiosk with nutritional education and optimized physical activity intervention (ksean + afso)
Primary result variable Nutritional status of the participants after a 2-year intervention.
Operational definition of nutritional status Obese >2 Z of BMI, Overweight >1 Z of BMI, Normal ≥ − 1 y ≤ 1 Z of BMI.
## Secondary result variable percentage of body fat.
Optimized physical activity (AFSO) primary dependent variable Fitness as measured in the lower body by the distance, in centimetres (cm), of a jump done with both feet together and without a run-up, and in the upper body by a dynamometer, which measures the isometric force in kilograms (kg) of the upper body by participants squeezing the instrument with the right and left hand.
## Healthy kiosk with nutritional education intervention (ksean)
Primary dependent variable Intake of foods measured as consumed and not consumed.
Control variable Physical activity intensity measured using a pedometer during the optimized physical activity classes.
Determination of sample size The sample size was determined as that required to obtain a standard size of the effect of 0.2 on the BMI z-score, based on a previous nutritional and physical education intervention study conducted in primary schools to prevent childhood obesity by Kain et al. in order to identify this effect in four clusters per intervention mode, with a power of 0.8, a significance of 5% and an inter-cluster coefficient of variance of 0.00012 [bib_ref] Two-year controlled effectiveness trial of a school-based intervention to prevent obesity in..., Kain [/bib_ref] , 140 children per cluster are required along with a minimum of 11 clusters, giving a total of 1540 children. Given that this study was conducted in three regions and that each intervention mode must be conducted in each region, a total of 12 clusters and 1680 subjects are required. Assuming a participation loss of 10% in 2 years of intervention, the total sample size would be 1848 children.
It is anticipated that the grip strength dependent variable will be improved by 20% compared to the study conducted by Rojas C et al. [bib_ref] Dinamometria de manos en estudiantes de Merida, Rojas [/bib_ref]. With the sample size calculated in this manner, in order to have a power of 0.8 at a 95% confidence level, 130 subjects per cluster are required, thereby reaching a total of 1560 children between year 1 and year 4.
## Study schedule
The schedule of enrolment, interventions, and assessments is illustrated in . In this figure time point t 1 corresponds to the time at which the baseline measurements of all variables will be taken and each intervention mode will commence, this time will be aligned with the beginning of the school year. Time point t 2 corresponds to the first evaluation after commencing each intervention mode, at this time all variables will be measured on every participating child and will be done 8 months later in line with the end of the school year. Time point t 3 corresponds to the second evaluation, once again all variables will be measured and this time will correspond with the children returning to school from their summer holidays that is 12 months after time point t 1 . Time point t 4 corresponds to the final evaluation, where once again all variables will be measured. Time point t 4 corresponds to the end of the second school year of the intervention that is 20 months after time point t 1 . Final result of the study will be delivered to each school and FOSIS 24 months after commencing the interventions at time point t 5 .
## Recruitment status
To obtain a sample size of 1656 children, a total of 1900 needed to be asked to participate in this study. This is based on an estimate of the proportion of children that would potentially satisfy the inclusion criteria and that would be motivated to participate in this study.
Following a review of the sample strategy and the recruitment of the field personnel, the selection of the potential participants commenced on 8 th March 2014. By 25 th March 2014 a total of 1923 children had been recruited, exceeding the sample size required.
## Data collection baseline data collection
During the baseline evaluation, participating children will be interviewed by the field personnel, who have been previously trained for this task. The evaluation is to take place at each participating school, following prior authorization by the relevant municipal authorities and school principals.
The following information will be registered. from summer holidays) and after 20 months of intervention (at the end of the second school year).
## Data handling data acquisition, management and transfer
All of the participants in this study will be identified by their unique national identification number (RUT). The data collected in the field by the field coordinators will be maintained under confidentiality between the field coordinators and the INTA data coordination centre. A unique identification number will be assigned to each participant once they have been registered in the study and their details have been recorded. This unique identification number will be used from that point forward on all relevant documentation. All participant files will be stored and locked in a secured location. Access to these files will be controlled by the field coordinator. All of the field data collected will be taken to the INTA on a weekly basis, where they will be validated to identify any inconsistencies and rectify errors. Once validated, the data will be entered into the study database.
The INTA data coordination centre will be responsible for registering all of the data files as they are delivered and notifying the field coordinator if any data are missing. The INTA data coordination centre will also be responsible for the safe keeping of the anonymous and encrypted data.
# Data analysis
The data will be validated by the minimum and maximum values. All the statistical analysis will consider the cluster design, in an intention of treating analysis. Additionally, each intervention protocol will be analysed separately to determine the effectiveness of each intervention.
The normality of the data will be assessed by determining the goodness of fit using the Shapiro-Wilk test, while the homogeneity of the variance will be assessed by using the Bartlett and Levene tests. The results from the Shapiro-Wilk test will be used to describe each variable depending on their normality, variables that are normally distributed will be described by the mean ± SD and range of the 95% CI, while variables that do not follow a normal distribution will be described by their percentile distribution. The results will be shown with their respective 95% confidence intervals.
The difference between each group (intervention mode) will be assessed by conducting parametric tests for normally distributed continuous variables (student t-test and ANOVA), and non-parametric tests (Wilcoxon, Kruskal-Wallis and Friedman tests) for non-normally distributed continuous variables. The Chi-square test will be conducted for categorical variables (non-continuous). Logistic regression models will also be developed to analyse the probability of change in the result variables for each intervention mode, controlling by age and gender of the participants and considering the impact of the secondary variables. To evaluate the difference between modes of intervention for continuous dependent variables, a multilevel regression analysis with mixed effects will be conducted to make adjustments by co-variables, as these models do consider the experimental design. StataCorp) and SAS 9.1 (Copyright (c) 2002-2003 by SAS Institute Inc., Cary, NC, USA) will be used to conduct all of the statistical tests and analysis.
# Conclusion
In relation to the increase in childhood obesity in Chile, there is consensus among researchers, educators and health and education personnel that schools are a favourable environment for actions to prevent and/or control childhood obesity.
Despite the apparent advantages of dealing with obesity in schools, a lack of evidence on the effectiveness of interventions to date has led some to question the wisdom of allocating resources to programmes.
Clearly further studies are required to provide more information on these aspects, and to achieve this, political authorities require specific information on what food and nutrition strategies to prioritize, which, along with physical activity, will provide encouraging results for the control and/or reduction of childhood obesity.
[fig] Figure 1: Participant recruitment process [/fig]
|
Cardiac Imaging and Management of Prosthetic Valve Candida Parapsilosis Endocarditis
# Introduction
Fungal infective endocarditis (IE) is a rare and serious form of IE with mortality rates up to 50% [bib_ref] Fungal endocarditis, Yuan [/bib_ref]. Typical risk factors include prosthetic valve implantation, cardiac implantation devices, and intravenous drug use [bib_ref] Fungal endocarditis, Yuan [/bib_ref]. In general, the most common etiologic causes of fungal IE are the Candida and Aspergillus species [bib_ref] Fungal endocarditis, Yuan [/bib_ref]. A six-year case review examining 12 separate cases of Candida IE found the most common organisms were Candida parapsilosis (n = 8, 67%), Candida glabrata (n = 3, 25%), and Candida albicans (n = 1, 8%) [bib_ref] Candida infective endocarditis during the infectious diseases and substance use disorder syndemic:..., Sankar [/bib_ref]. Current guidelines by the Infectious Diseases Society of America Candidiasis and American Heart Association Endocarditis recommend treatment of Candida IE with either Amphotericin B with or without Flucytosine or high-dose echinocandin therapy, followed by life-long maintenance therapy with an oral azole [bib_ref] Candida infective endocarditis during the infectious diseases and substance use disorder syndemic:..., Sankar [/bib_ref]. However, treatment failure with medical management alone is common because Candida species have adapted survival strategies including the formation of biofilms on native and prosthetic heart valves, which can lead to poor antifungal activity [bib_ref] Fungal endocarditis, Yuan [/bib_ref]. Therefore, it is imperative that the treatment of native valve endocarditis includes a potential surgical option, especially in patients with prosthetic valves since early intervention has been proven beneficial [bib_ref] Fungal endocarditis, Yuan [/bib_ref]. Despite combined medical and surgical treatment, fungal IE has a poor overall prognosis, and there is an imperative need for aggressive risk factor management.
We present a case of a 45-year-old male with Candida parapsilosis IE in the setting of a bioprosthetic mitral valve and highlight the computed tomography (CT) abdomen and pelvis, transthoracic echocardiography (TTE), and transesophageal echocardiography (TEE) image findings associated with this devastating disease process.
## Case presentation
A 45-year-old male with a history of intravenous drug use, liver cirrhosis secondary to hepatitis C, recent mitral valve endocarditis with a bioprosthetic mitral valve in 2019, gastroesophageal reflux disease (GERD), and pancytopenia presented with four months of progressive abdominal distention. The patient lost his insurance during the coronavirus-19 pandemic and was not following up with routine healthcare. His abdominal distention was causing decreased appetite and early satiety and led to a 40-pound unintentional weight loss. The patient denied associated symptoms of nausea, vomiting, or diarrhea and felt that his symptoms of GERD were well-controlled with daily pantoprazole. In addition, the patient was having intermittent fevers for two weeks and shortness of breath for three days prior to admission.
On admission, the patient was febrile with a temperature of 101 degrees Fahrenheit, normotensive with blood pressure 116/77 millimeters of mercury (mmHg), tachycardia with a heart rate of 114 beats per minute, and saturating 98% on room air. Initial labs are listed in [fig_ref] TABLE 1: Patient's labs on presentation [/fig_ref].
## Name of lab
## Patient value normal range
High-sensitivity C-reactive protein [bib_ref] Candida parapsilosis prosthetic valve endocarditis, Silva-Pinto [/bib_ref] CT abdomen and pelvis were remarkable for splenomegaly with new acute to subacute splenic infarcts, the presence of cirrhosis with moderate volume ascites, and portal enteropathy [fig_ref] FIGURE 1: Computed tomography of the abdomen, transthoracic echocardiography, and transesophageal echocardiography findings D [/fig_ref]. Chest x-ray showed worsening bilateral airspace disease with perihilar distribution, and CT angiography of the pulmonary arteries was negative for acute central pulmonary embolus. Blood cultures were ordered, and the patient was started on empiric Cefepime. Thoracentesis was performed to relieve his respiratory symptoms. Approximately 1200 milliliters of cloudy amber fluid was drained out, which had moderate white blood cells without the growth of any organisms. TTE showed a preserved left ventricular ejection fraction of 55%-60% and a 31-millimeter (mm) Edwards bioprosthetic valve with severely thickened leaflets and restricted motion. TTE was unequivocal in completely excluding any vegetation or mass on the bioprosthetic valve. The mean mitral valve gradient was elevated at 28 mm Hg (normal < 13 mm Hg), and the pulmonary artery pressure was 74 mm Hg (normal 18-25 mm Hg) [fig_ref] FIGURE 1: Computed tomography of the abdomen, transthoracic echocardiography, and transesophageal echocardiography findings D [/fig_ref]. Initial blood cultures grew in Candida parapsilosis, and the patient was started on intravenous (IV) Micafungin 100 mg daily. TEE study performed subsequently during the same anesthesia showed a 1.23 cm x 0.55 cm lesion and 1.02 cm x 0.55 cm lesion on the Edwards 31 mm Magna bioprosthetic mitral valve consistent with vegetation versus thrombus . The patient was deemed not a surgical candidate due to concomitant liver cirrhosis, esophageal varices, ascites, and elevated model for end-stage liver disease (MELD) score of 13-15. He was temporarily treated with IV Amphotericin B 200 mg daily for seven days and was ultimately discharged with six weeks of IV Fluconazole 400 mg daily followed by lifelong oral Fluconazole 200 mg daily.
## Figure 2: three-dimensional transesophageal echocardiogram of surgeon's view during diastole
Three-dimensional planimeter valve area: 0.77-0.87 cm 2 (Severe < 1.0 cm 2 ).
# Discussion
Fungal prosthetic valve IE is a rare presentation of IE affecting people worldwide. In a prior literature review, a total of 152 cases were identified from 1995 to 2000, with intravenous injection drug use as an identified risk factor in only 4.1% of cases. Fungal IE is associated with high incidences of severe morbidity and mortality, ranging from overall morbidity of 67% and six-month mortality risk of 37% [bib_ref] Prosthetic valve Candida spp. endocarditis: new insights into longterm prognosis-the ESCAPE study, Rivoisy [/bib_ref] [bib_ref] Fungal prosthetic valve endocarditis: Mayo Clinic experience with a clinicopathological analysis, Jennifer [/bib_ref]. The rarity of published reports and the medical guidelines for the appropriate choice and duration of antifungal therapy has been limited. Of those case reports published, medical management with liposomal Amphotericin B (LAmB) 5 mg/kg/day (300 mg/day) and Flucytosine 150 mg/kg/day (9 g/day) as part of an initial treatment after positive blood cultures for Candida is the suggested intervention [bib_ref] Candida parapsilosis prosthetic valve endocarditis, Silva-Pinto [/bib_ref]. The latest Infectious Diseases Society of America (IDSA) from 2016 provides guidelines regarding the treatment of Candida infections with relationship to infected pacemakers, implantable cardiac defibrillators, ventricular assist devices, along with the treatment of native valve and prosthetic valve IE [bib_ref] Clinical practice guideline for the management of Candidiasis: 2016 update by the..., Pappas [/bib_ref]. These guidelines recommend LAmB of 3-5 milligrams (mg) per kilogram (kg), with or without Flucytosine 25 mg/kg four times daily, or high-dose echinocandin therapy (Caspofungin 150 mg daily, Micafungin 150 mg daily, or Anidulafungin 200 mg daily) [bib_ref] Clinical practice guideline for the management of Candidiasis: 2016 update by the..., Pappas [/bib_ref]. In addition, for patients unable to undergo valve replacement, there is a strong recommendation for chronic long-term suppression with Fluconazole 400-800 mg daily [bib_ref] Clinical practice guideline for the management of Candidiasis: 2016 update by the..., Pappas [/bib_ref]. Our patient continues to follow up with cardiology as an outpatient. [bib_ref] Prosthetic valve Candida spp. endocarditis: new insights into longterm prognosis-the ESCAPE study, Rivoisy [/bib_ref]. In addition, 21 patients received long-term maintenance therapy with Fluconazole, with an average of 13 months, and demonstrated minimal adverse effects [bib_ref] Prosthetic valve Candida spp. endocarditis: new insights into longterm prognosis-the ESCAPE study, Rivoisy [/bib_ref]. Also, the 19 patients who underwent a cardiac surgical procedure did not have improved survival outcomes over a sixmonth period as compared to the group treated with medical management [bib_ref] Prosthetic valve Candida spp. endocarditis: new insights into longterm prognosis-the ESCAPE study, Rivoisy [/bib_ref]. These results offer some contradiction to the Infectious Disease of America and European Society of Clinical Microbiology and Infectious Diseases guidelines, which recommend early surgical intervention for all patients with prosthetic IE [bib_ref] Fungal prosthetic valve endocarditis: Mayo Clinic experience with a clinicopathological analysis, Jennifer [/bib_ref] [bib_ref] A meta-analysis of medical versus surgical therapy for Candida endocarditis, Steinbach [/bib_ref]. Despite the poor statistical power with a small sample size, this study provides an insight into medical management as an alternative to surgical therapy.
# Conclusions
Candida IE is a rare pathological process with a high propensity to involve and compromise implanted prosthetic valves and implanted cardiac devices. Individuals with a history of drug use are among the greatest risk in the general population. In our case, the patient had a previous history of bioprosthetic mitral valve replacement and was found to have recurrent IE on his bioprosthetic mitral valve. Ultimately, the patient was not a surgical candidate and was treated with Amphotericin B and long-term Fluconazole. This case focuses on the presentation of Candida parapsilosis IE with subsequent imaging findings of this rare disease. In addition, this case highlights the importance of reviewing current clinical guidelines and providing appropriate medical management.
# Additional information disclosures
Human subjects: Consent was obtained or waived by all participants in this study. 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 1: Computed tomography of the abdomen, transthoracic echocardiography, and transesophageal echocardiography findings D: Doppler findings; mitral mean velocity: 2.7 m/sec; mitral mean gradient: 17 mm Hg (severe > 10 mm Hg) [/fig]
[table] TABLE 1: Patient's labs on presentation [/table]
|
Contrecoup Extradural Hematoma Without Fracture: A Case Report and Review of Literature
Extradural hemorrhages are commonly seen in coup head injuries, rarely seen in contrecoup head injuries. Acute extradural hemorrhage in the coup head injuries associated with a fracture is common, but the incidence of acute contrecoup extradural hemorrhage not associated with the fracture is extremely rare. Only 21 cases have been reported previously. A 28-year-old male patient presented to the emergency department with complaints of sustaining injuries in a road traffic accident by fall from a two-wheeler. No history of loss of consciousness, vomiting, seizures, and ear/nose bleed. On examination, the patient was conscious and coherent with a Glasgow Coma Scale score of 15/15 and a laceration on the right frontotemporal region which was sutured. Contrast tomography of the brain revealed an extradural hemorrhage of 10 mm thickness in the left parieto-occipital region with soft tissue swelling in the right temporoparietal region, without any evident fractures in the calvarium. The patient was managed conservatively. Contrecoup extradural hematoma is a rare entity, and those without fracture are extremely rare. Early diagnosis, careful observation, and management lead to a good outcome.
# Introduction
An extradural hemorrhage is a collection of blood in the space between the inner table of calvarium and dura mater. It can occur in coup and countercoup injuries. Extradural hematomas (EDHs) due to coup type of head injuries are quite common and most often associated with overlying calvarium fractures, whereas contrecoup EDHs are very rare entities and are not associated with any overlying fractures. There are only a few reports of contrecoup EDHs published in the literature.
## Case report
A 28-year-old male alleged to have sustained head injuries in a road traffic accident by fall from a two-wheeler under the influence of alcohol presented to the emergency department with no complaints of loss of consciousness, vomiting, seizures, and ear/nose bleed. A laceration was noted on the right temporoparietal region which was immediately sutured. On examination, the patient was conscious and coherent with a Glasgow Coma Scale Score of 15/15 without any focal neurological deficits; all vitals were within normal limits. Computed tomography of the brain was done which revealed a 10 mm thick extradural hemorrhage in the left parieto-occipital region and soft tissue thickening in the right temporoparietal region [ [fig_ref] Figure 1: Computed tomography scan of the brain shows a 10 mm thick hyperdense... [/fig_ref] ]. No calvarial fractures were noted. A computed tomography scan of the brain was repeated after 24 h which revealed no increase in the thickness of extradural hemorrhage. The patient was managed conservatively, and the follow-up was uneventful.
# Discussion
At our center, we came across one such case where the patient had a contrecoup EDH without an overlying fracture of the calvarium, and a thorough review of literature reiterates the fact that this presentation is a rare variant and hence this case report. EDHs are one of the most common presentations in head injury. It accounts for 1%-3% of all the head injuries. [bib_ref] Acute epidural hematoma caused by contrecoup head injury -Case report, Mitsuyama [/bib_ref] They are usually located at the site of impact and occur due to the rupture of a middle meningeal artery or calvarial fracture resulting in accumulation of blood in between the inner calvarium and dura mater due to the separation of dura mater from the calvarium. These types of EDHs are referred to as coup EDHs and are more commonly noted. [bib_ref] Surgical management of acute epidural hematomas, Bullock [/bib_ref] They develop below the impact point and are accompanied by the linear fracture in most of the cases. [bib_ref] Acute epidural hematoma caused by contrecoup head injury -Case report, Mitsuyama [/bib_ref] [bib_ref] Frontal acute extradural hematoma due to contrecoup injury: A case report, Miyazaki [/bib_ref] [bib_ref] Acute epidural hematoma caused by contrecoup injury, Motohashi [/bib_ref] Contrecoup EDHs, however, are a rare variant those not associated with a fracture are extremely rare, and only 21 cases have been previously reported in the literature. [bib_ref] Acute contrecoup epidural hematoma that developed without skull fracture in two adults:..., Andoh [/bib_ref] They are not associated with any overlying fractures and are usually found at a location opposite to the site of impact. Contrecoup EDHs are more commonly seen in the females in the fifth to sixth decade. [bib_ref] Administration of recombinant tissue plasminogen activator to a case of cerebral infarction..., Takeuchi [/bib_ref] They are more commonly seen in the frontal regions, which can be explained by the fact that the dura mater can easily be detached from the inner table of the lateral frontal region. [bib_ref] Acute epidural hematoma caused by contrecoup head injury -Case report, Mitsuyama [/bib_ref] Contrecoup acute epidural hemorrhage without fracture in the occipital region is considered due to the skull development. The occipital bone develops from two types of tissues membranous and cartilaginous tissues, and the transverse sinus is present in the boundary between these tissues. Thus, this region is easily deformed and reported to be a region with reduced resistance against external forces. [bib_ref] Acute epidural hematoma of the posterior fossa caused by Fronto-temporal impact. Case..., Shigemori [/bib_ref] Cavitation effect and inertia loading at the countercoup site may be the reason for countercoup hematomas. [bib_ref] Revisiting the contrecoup extradural hematoma, Sahoo [/bib_ref] The negative pressure created at the opposite site may cause the cavitation and the tensile strain in the angular movement of the head may cause vascular injury. [bib_ref] Biomechanical basis of traumatic brain injury, Meaney [/bib_ref] Both of these mechanisms explain the genesis of contrecoup intraparenchymal hematomas. However, mechanism of contrecoup EDH remains unclear.
Few have proposed that it might be generated by the buckling effect of the skull exactly opposite to the site of impact. [bib_ref] A case of coup and contrecoup extradural hematoma, Balasubramaniam [/bib_ref] Others have proposed that compression wave from the site of impact produces a relative movement in between the dura mater and skull bone resulting in the stripping of dura mater and collection of EDH at the diagonally opposite side.
Here, we report a case of the left parieto-occipital EDH in a 28-year-old male who was managed conservatively with a good outcome. Few other cases reported by others have been noted below.
Okamoto reported a case of the left frontal contrecoup hematoma in a 51-year-old female due to fall which was surgically treated. [bib_ref] Acute epidural hematoma caused by contrecoup injury, Okamoto [/bib_ref] Balasubramaniam and Ramesh reported a case of a 21-year-old male patient with a contrecoup EDH at the left frontal region which was managed surgically. Miyazaki et al. reported a case of right frontal contrecoup EDH in a 52-year-old female who was managed surgically. [bib_ref] Frontal acute extradural hematoma due to contrecoup injury: A case report, Miyazaki [/bib_ref] Motohashi reported a case of a 59-year-old female patient with a left frontal contrecoup EDH who was managed conservatively. [bib_ref] Acute epidural hematoma caused by contrecoup injury, Motohashi [/bib_ref] Mitsuyama et al. reported a case of a 50-year-old female patient with right frontal contrecoup EDH who was managed conservatively.
# Conclusion
EDH is a neurosurgical emergency. Contrecoup variants of EDH are a rare entity, and those not associated with fracture are extremely rare. Hence, careful observation and timely management are required. Here, we report a case of contrecoup variant of EDH which is not associated with fracture.
## Declaration of patient consent
The authors certify that they have obtained all appropriate patient consent forms. In the form the patient(s) has/have given his/her/their consent for his/her/their images and other clinical information to be reported in the journal. The patients understand that their names and initials will not be published and due efforts will be made to conceal their identity, but anonymity cannot be guaranteed.
## Financial support and sponsorship
Nil.
## Conflicts of interest
There are no conflicts of interest.
[fig] Figure 1: Computed tomography scan of the brain shows a 10 mm thick hyperdense collection in the left parieto-occipital region with a subcutaneous swelling in the right temporoparietal region. No calvarial fracture noted [/fig]
|
Rouse model with transient intramolecular contacts on a timescale of seconds recapitulates folding and fluctuation of yeast chromosomes
DNA folding and dynamics along with major nuclear functions are determined by chromosome structural properties, which remain, thus far, elusive in vivo. Here, we combine polymer modeling and single particle tracking experiments to determine the physicochemical parameters of chromatin in vitro and in living yeast. We find that the motion of reconstituted chromatin fibers can be recapitulated by the Rouse model using mechanical parameters of nucleosome arrays deduced from structural simulations. Conversely, we report that the Rouse model shows some inconsistencies to analyze the motion and structural properties inferred from yeast chromosomes determined with chromosome conformation capture techniques (specifically, Hi-C). We hence introduce the Rouse model with Transient Internal Contacts (RouseTIC), in which random association and dissociation occurs along the chromosome contour. The parametrization of this model by fitting motion and Hi-C data allows us to measure the kinetic parameters of the contact formation reaction. Chromosome contacts appear to be transient; associated to a lifetime of seconds and characterized by an attractive energy of -0.3 to -0.5 k B T. We suggest attribut-ing this energy to the occurrence of histone tail-DNA contacts and notice that its amplitude sets chromosomes in 'theta' conditions, in which they are poised for compartmentalization and phase separation.
# Introduction
The genome is organized into higher-order functional domains and territories (1) that contribute to the regulation of gene expression and to cell differentiation [bib_ref] Dynamic organization of chromatin domains revealed by super-resolution live-cell imaging, Nozaki [/bib_ref]. In interphase of budding yeast cells cycle, chromosomes adopt a Rabl-like conformation [bib_ref] Principles of chromatin organization in yeast: relevance of polymer models to describe..., Wang [/bib_ref] , which is characterized by the positioning of centromeres and the nucleolus at opposite ends in the nucleoplasm. This large-scale organization has adequately been reproduced in silico by modeling chromosomes as homogeneous polymers (homopolymers) with some structural constraints associated to peripheral tethering of telomeres and centromeres [bib_ref] Physical tethering and volume exclusion determine higher-order genome organization in budding yeast, Tjong [/bib_ref] [bib_ref] A predictive computational model of the dynamic 3D interphase nucleus, Wong [/bib_ref] [bib_ref] Chromosome positioning and the clustering of functionally related loci in yeast is..., Gehlen [/bib_ref]. A detailed analysis of chromosome fluctuations with the Rouse model, which is a generic polymer model relevant to crowded environments, further corroborated the relevance of this homopolymer model in yeast [bib_ref] Dynamical modeling of three-dimensional genome organization in interphase budding yeast, Tokuda [/bib_ref]. Indeed, the dynamics of chromosome loci, as measured by the temporal variation of the mean square displacement (MSD), appeared to follow an anomalous response characterized by:
[formula] MSD (τ ) = τ α(1) [/formula]
with α the diffusion exponent, the amplitude, and τ the time interval. The exponent α was in the range of 0.4-0.6 and the amplitude appeared to be homogeneous in the genome except for the proximity of telomeres and centromeres [bib_ref] SAGA interacting factors confine sub-diffusion of transcribed genes to the nuclear envelope, Cabal [/bib_ref] [bib_ref] Correlations of three-dimensional motion of chromosomal loci in yeast revealed by the..., Backlund [/bib_ref] [bib_ref] Analysis of single locus trajectories for extracting in vivo chromatin tethering interactions, Amitai [/bib_ref] [bib_ref] Systematic characterization of the conformation and dynamics of budding yeast chromosome XII, Albert [/bib_ref] [bib_ref] Visualization of chromatin decompaction and break site extrusion as predicted by statistical..., Amitai [/bib_ref] [bib_ref] Inferring the physical properties of yeast chromatin through Bayesian analysis of whole..., Arbona [/bib_ref] [bib_ref] Chromatin stiffening underlies enhanced locus mobility after DNA damage in budding yeast, Herbert [/bib_ref]. Accordingly, the Rouse model predicts that the MSD should increase over time with an exponent of 0.5 and that the amplitude of fluctuations is determined by the local properties of chromosomes, specifically the bending flexibility and solvent friction resisting monomer displacements. Using the Kuhn length b, which is twice the persistence length of the polymer, as a proxy of the bending stiffness, and the monomer friction coefficient ζ , the Rouse model predicts the MSD in 2D:
[formula] MSD (τ ) = 16b 2 k B T 3πζ τ 0.5(2) [/formula]
with k B T the Boltzmann thermal energy. Note that this expression is obtained with the assumption that the chain contains a large number of monomers and behaves as a 'phantom polymer', i.e. without effects of volume exclusion between monomers. Taking the standard approximation that the friction coefficient ζ is proportional to the Kuhn length, the amplitude of MSD fluctuations is expected to provide a means to infer the bending flexibility of chromosomes in vivo. However, direct adjustment with the Rouse theory showed that the in vivo value of ≈ 0.01μm 2 .s −0.5 corresponded to a Kuhn length as small as 1-5 nm [bib_ref] High throughput chromatin motion tracking in living yeast reveals the flexibility of..., Hajjoul [/bib_ref]. This estimate appeared to be incompatible with structural and mechanical models of nucleosome fibers [bib_ref] The physics of chromatin, Schiessel [/bib_ref] , as well as the recurrent detection of irregular 10-nm fibers by electron microscopy (EM) of thin nuclear sections [bib_ref] Liquid-like behavior of chromatin, Maeshima [/bib_ref] [bib_ref] Chromatin structure: does the 30-nm fibre exist in vivo?, Maeshima [/bib_ref] [bib_ref] ChromEMT: Visualizing 3D chromatin structure and compaction in interphase and mitotic cells, Ou [/bib_ref]. This inconsistency shows that the precise mechanisms underlying chromosome motion remain largely unclear. The study of chromosome higher-order structure based on chromosome conformation capture techniques (specifically, Hi-C) has recently shed light on the contribution of two essential mechanisms, which are not included in the simple Rouse model: active processes involving loop extrusion or transcription, and segregation through the random association of monomers along the chain [bib_ref] Evolutionarily conserved principles predict 3D Chromatin Organization, Rowley [/bib_ref] [bib_ref] Chromatin organization by an interplay of loop extrusion and compartmental segregation, Nuebler [/bib_ref] [bib_ref] Modeling epigenome folding: formation and dynamics of topologically associated chromatin domains, Jost [/bib_ref]. Theoretical studies have suggested that both mechanisms could modulate the amplitude of chromatin fluctuations [bib_ref] Linear viscoelastic properties of transient networks formed by associating polymers with multiple..., Indei [/bib_ref] [bib_ref] Dynamics of active Rouse chains, Osmanović [/bib_ref]. Consequently, we were motivated to clarify if and whether additional mechanisms had to be integrated into the Rouse model in order to recapitulate chromosome dynamics in yeast. In order to test polymer models in controlled settings, we first design a biomimetic system to validate the relevance of the Rouse model to analyze DNA and chromatin fluctuation in bulk. This study shows that fluctuation data are quantitatively fitted with the Rouse model parametrized with mechanical parameters of DNA and chromatin. Conversely, in living yeast, the inconsistency of the Rouse model to fit fluctuation and Hi-C data forced us to introduce an equilibrium polymer model, the RouseTIC model, in which random Transient Internal Contacts along the chromosome contour were considered. After a rigorous fitting of the kinetic parameters of the chromosome contact formation reaction, we show that this model recapitulates Hi-C and MSD measurements. Chromosome contacts appear to be transient, lasting a few seconds, and characterized by an in-teraction energy in the range of -0.3 to -0.5 k B T. We finally speculate that these labile interactions along the chromosome contour may originate from unspecific histone tail interactions and constitute a key mechanism for the formation of dynamic compartments in eukaryotes.
# Materials and methods
## Dna and chromatin preparation and characterization
Unless stated, chemical and biochemical reagents were purchased from Sigma-Aldrich or New England Biolabs, respectively. Chromosome fragments were purified from U2OS cells, which were synchronized at the beginning of S phase, then released in the presence of dUTP-Cy3 and scraped from glass surfaces in order to allow fluorescent nucleotide entry into the cells and incorporation into the genome during DNA replication (see [bib_ref] Analysis of DNA replication by optical mapping in nanochannels, Lacroix [/bib_ref] for details of the protocol). Cells were eventually embedded in agarose plugs, and treated with 5% SDS and 100 g/ml proteinase-K during two days in order to purify genomic DNA. The resulting chromosome fragments were recovered from agarose plugs using -Agarase treatment during 2 h. Nucleosome arrays were assembled with a reconstitution kit (Active Motif) using ∼10 ng of purified chromosome fragments mixed with 1 g of unlabeled -DNA.
Structural characterization of chromatin fibers was carried out by electron microscopy using the protocol of (28). Briefly, 5 l of reaction product diluted to ∼10 −12 M in 10 mM Tris-HCl pH 7.5, 5 mM MgCl 2 , 50 mM, NaCl was deposited on a 600-mesh copper grid covered with a thin carbon film, activated by glow-discharge in the presence of pentylamine. Grids were washed with aqueous 2% (w/v) uranyl acetate, dried and observed in the annular dark-field mode, using a Zeiss 902 transmission electron microscope. Images were captured at a magnification of 85 000 or 140 000 with a Veletta CCD camera and analyzed with iTEM software (Olympus Soft Imaging Solution).
For imaging experiments, DNA or chromatin fibers were diluted 1000-fold in a low salt buffer (Tris-HCl 89 mM, 89 mM boric acid, 2 mM EDTA; pH 8.3) supplemented with 360 or 40 kDa PVP (Sigma-Aldrich). For DNA motion analysis, the viscosity was 5.4 or 2.3 mPa.s with the same mass:volume fraction of 2% of 360 or 40 kDa PVP, respectively. The 360 kDa PVP concentration was set to 3.2% and the viscosity to 15 mPa.s for chromatin loci tracking experiments. The overlapping concentration of these PVP solutions has been estimated in ref. [bib_ref] Poly(ethyleneoxide) (PEO) and poly(vinyl pyrolidone) (PVP) induce different changes in the colloid..., Mcfarlane [/bib_ref]. Note that very low concentrations of 100 nm carboxylated polystyrene fluorescent beads (Invitrogen) were added to the buffer in order to measure the viscosity by particle tracking.
## Yeast culture
We used a fluorescent yeast strain containing one fluorescent locus with 224 tetO sequences to maximize signal-tonoise ratio located at genomic position 530 kb on chromosome XII [bib_ref] Systematic characterization of the conformation and dynamics of budding yeast chromosome XII, Albert [/bib_ref]. Cells were grown overnight at 30 - C in YP media containing 2% glucose. They were then diluted at 10 6 cells/ml and harvested when OD 600 reached 4 × 10 6 cells/ml. Cells were spread on slides coated with corresponding patch containing 2% agar and 2% glucose. Cover Nucleic Acids slides were sealed with 'VaLaP' (1/3 vaseline, 1/3 lanoline, 1/3 paraffin). Microscopy was performed during the first 10-20 min after the encapsulation of the cells in the chamber. Characterization of the signal-to-noise ratio of the fluorescent locus is reported in Supplementary [fig_ref] Figure 1: Analysis of DNA fluctuations in vitro [/fig_ref] with one example of a time lapse of a fluorescent focus.
## Microscopy and trajectory analysis
DNA/chromatin or chromosome loci were tracked with a Zeiss AxioObserver or a Nikon TI-E/B inverted microscope, respectively. The Nikon status was equipped with an EM-CCD IxonULTRA DU897 camera (Andor), a 488 nm laser illumination (Sapphire 488, Coherent), and 100× oil immersion objective (numerical aperture = 1.4) with a 1.5× magnification placed ahead of the camera. Pixel size was 107 nm. The Zeiss status was equipped with an Zyla 4.2 sC-MOS camera (Andor), a Sola light engine (Lumencor), a 40X oil immersion objective (NA = 1.4) placed ahead of the camera. Pixel size was 162.5 nm.
Trajectories with more than ∼100 consecutive positions were subsequently extracted with the TrackMate Plugin [bib_ref] TrackMate: an open and extensible platform for single-particle tracking, Tinevez [/bib_ref]. We processed the MSD and the kurtosis K, as defined by (31):
[formula] MSD (τ ) = (x (t) − x (t + τ )) 2 (3) K (τ ) = (x (t) − x (t + τ )) 4 − MSD(τ ) 2 MSD(τ ) 2(4) [/formula]
with τ the time interval. We also extracted the step distribution function (SDF) for a given time interval τ . For each trajectory, these MSD and kurtosis were computed for smaller than 30% of the total duration of the track in order to obtain statistically significant values in single cell datasets [bib_ref] Lateral diffusion in an archipelago. Single-particle diffusion, Saxton [/bib_ref].
## Monte carlo simulations of nucleosome arrays
The chromatin fiber was modelled as a string of coarsegrained DNA linkers and nucleosome core particles following the methodology described in [bib_ref] In silico single-molecule manipulation of DNA with rigid body dynamics, Carrivain [/bib_ref]. The nucleosomal DNA consists of 14 segments of 10.5 bp, anchored to the histone core by 14 minor groove locations, usually referred as Super Helix Locations [bib_ref] Crystal structure of the nucleosome core particle at 2.8 A resolution, Luger [/bib_ref]. The linker DNA is made of an integer number of identical segments, each one containing a number of base pairs as close as possible to 10.5 bp. For example, we modeled a 22 bp linker with two segments of 11 bp or a 18 bp linker with two segments of 9 bp. Note that the length l of each linker segment is proportional to the number n of base pairs with l = n × 0.32 nm and the helical angle between consecutive base pairs is 34 - . The articulation between linker DNA segments is modelled as a ball-andsocket joint with an energy penalty corresponding to the bending and twisting restoring torques of the DNA double helix. The bending energy is
[formula] E b = g b (1 − cos) [/formula]
where θ is the bending angle between two consecutive segments and g b is calculated from L ( g b k B T ) = (b − l)/(b + l) with b = 100 nm and L the Langevin function. The twisting energy is given by E t = g t φ 2 /2 in the harmonic approximation where φ is the twist angle of two consecutive segments. The twisting energy constant g t between two connected segments of length l is given by g t /k B T = l t /l with l t = 95 nm the twist persistence length of naked DNA. At each step of the simulation, we chose one segment at random and rotated all the next segments around a random axis with a random angle (Pivot move). The resulting conformation was accepted following the Metropolis criterion.
## Extraction of chromatin kuhn length from nucleosome array monte carlo simulations
We used the end-to-end distance distribution as readout of Monte-Carlo simulations to extract the Kuhn length of every nucleosome array configuration (see details in Supplementary Material 1). Briefly, we assumed that nucleosome arrays behaved according to the Worm-Like Chain model, which is defined by the persistence length l p as half the Kuhn length and the contour length L of the chain. We developed a specific approach to estimate these two parameters independently from the analysis of the end-to-end distance distribution based on [bib_ref] End-to-end distribution function of stiff polymers for all persistence lengths, Hamprecht [/bib_ref]. The second and fourth moments of the distribution can indeed be expressed as a function of L and l p , and are the most relevant whenever the contour length L is comparable to 4l p . Therefore, we simulated sufficiently, but not overly, long fibers with up to 100 nucleosomes and then, using an inversion method, extracted both L and l p from the second and fourth moments. In order to limit end effects, we performed this extraction for central sub-chains containing 5 to 95 nucleosomes (Supplementary [fig_ref] Figure 2: Nucleosome array fluctuations in vitro [/fig_ref]. The stationary value of the persistence length detected for a contour length L ∼ 4l p was used to estimate l p .
## Kinetic monte carlo simulations of coarse-grained homopolymers with internal contacts
We used a coarse-grained self-avoiding homopolymer model for chromosome structure and fluctuations, as described in [bib_ref] The folding landscape of the epigenome, Olarte-Plata [/bib_ref]. Briefly, we modeled a typical chromosome containing 10 6 bp by a polymer chain of N monomers, each of size b and corresponding to n base pairs (N*n = 10 6 ). The dynamics of the chain were simulated on a face-centeredcubic lattice with periodic boundary conditions using a simple kinetic Monte-Carlo scheme with local moves: at each trial move, a monomer is randomly picked and randomly displaced to a nearest-neighbor (NN) site on the lattice. The move is accepted if the trial position is empty and maintains the chain connectivity. The density of monomer in the simulation box was set to 15% to effectively account for the presence of other chromosomes.
For the simulations of the Rouse model with transient internal contacts, we assumed that two monomers that occupied NN sites have the possibility to form a rigid pair at binding rate k b and that this complex can be disassembled with a rate of unbinding k u . We assumed that a monomer can be involved at most in only one pair. A paired monomer could then move only if it remained connected to its partner, i.e. only if its new location remained a NN site of its partner.
For each investigated situation, we initiated the system with random configurations and then let it reach equilibrium before measuring the average contact probabilities, the mean squared distances, mean squared displacements and kurtosis by averaging over 1000 trajectories. The time step of the simulation was 0.55 ms.
In order to evaluate the effective energy E 0 of contact between monomers, we used the detailed balance of the Monte Carlo simulation between a micro-state i where two monomers were NN on the lattice and a micro-state j where they were not in contact. According to Boltzmann statistics,
[formula] we have P(i → j )/P( j → i ) = exp(E 0 /k B T). [/formula]
Considering reaction rates k b and k u , the detailed balance yields during a time step dt:
[formula] P(i → j ) P( j → i ) = 1 dt k b k u + k b k u dt + k u k u + k b (1 − k b dt) = k u k u + k b (5) leading to E 0 = k B T log k u k u +k b . [/formula]
# Results
## Dna flexibility can be determined from the analysis of its fluctuations
We first wished to validate our assumption that the bending stiffness of a polymer, equivalently its Kuhn length, can be inferred from the analysis of its spatial fluctuations with the Rouse model. Because DNA Kuhn length of b ∼100 to 110 nm in low salt conditions is well-documented in the literature (37), we used naked DNA as a model biopolymer. In order to generate long deproteinized DNA molecules containing randomly-distributed short fluorescent stretches, we purified chromosome fragments from human osteosarcoma cells treated with fluorescent nucleotides during DNA replication (see Materials and Methods). The size of the stretches of ∼50 kb and the contour length of the molecules of 500 kb or more have been characterized by DNA combing and optical mapping in nanochannels, respectively [bib_ref] Analysis of DNA replication by optical mapping in nanochannels, Lacroix [/bib_ref]. Labeled DNA molecules were then dispersed in a crowded solution containing a high concentration of the neutral polymer poly-vinylpyrrolidone (PVP), equal to three times the overlapping concentration, i.e. the threshold concentration for which PVP chains become intermingled [fig_ref] Figure 1: Analysis of DNA fluctuations in vitro [/fig_ref] , (29)). DNA loci movements were finally recorded during >60 images using wide field fluorescence microscopy at an inter-frame interval of 39 ms. The MSD in 2D was finally extracted for each trajectory (gray datasets in [fig_ref] Figure 1: Analysis of DNA fluctuations in vitro [/fig_ref]. The average MSD response shown with the red dataset in [fig_ref] Figure 1: Analysis of DNA fluctuations in vitro [/fig_ref] was consistent with the Rouse model using anomalous diffusion parameters ( ,α) of (0.26 ± 0.02 m 2 /s 0.53 , 0.53 ± 0.03), as plotted with the corresponding black curve, using 95% confidence intervals for the uncertainty. Taking b∼100 nm for DNA [bib_ref] Dependence of DNA persistence length on ionic strength of solutions with monovalent..., Brunet [/bib_ref] and using the friction coefficient of a cylindrical monomer of diameter d ∼2 nm for DNA ζ = 3πηb/ln(b/d) with η the solvent viscosity of 5.4 mPa.s, we deduce from equation (2) that = 0.23 m 2 /s 0.5 , in good agreement with experimental data. More quantitatively, we could estimate the Kuhn length by measuring the sum of the squared residuals between the fit and the MSD curve [fig_ref] Figure 1: Analysis of DNA fluctuations in vitro [/fig_ref] , which reached a marked minimum for b = 119 nm. In addition, the same model enabled fitting the step distribution function (SDF) for three time intervals of 0.12, 0.23 and 0.35 s (Supplementary [fig_ref] Figure 3: Inferring the mechanical properties of nucleosome arrays [/fig_ref].
We next performed a series of experiments using a PVP concentration three times lower than the overlapping concentration (see Methods), i.e. in conditions where crowding becomes negligible because the PVP chains are no longer intermingled [bib_ref] Molecular crowding: analysis of effects of high concentrations of inert cosolutes on..., Minton [/bib_ref]. For this, we used the same PVP concentration but the chains were characterized by a molecular weight 10 times smaller. The resulting viscosity of η = 2.3 mPa.s was expectedly lower. In this regime where hydrodynamic interactions between DNA monomers contribute to the dynamics of the chain, the Zimm model is expected to be valid with the corresponding temporal variation of the MSD (39):
[formula] MSD (τ ) = 2 × (1/3) π 2 × √ 3π k B T η τ 2/3(6) [/formula]
with (1/3) = 2.679. Note that equation (6) is dependent on the viscosity η and unaffected by DNA persistence length. The Zimm model was in agreement with our measurements (green line in [fig_ref] Figure 1: Analysis of DNA fluctuations in vitro [/fig_ref] , whereas the Rouse model yielded an inconsistent prediction using a Kuhn length of 100 nm (black curve in [fig_ref] Figure 1: Analysis of DNA fluctuations in vitro [/fig_ref]. In addition, the Zimm model was not consistent with the dynamics of DNA in crowded conditions (green curve in [fig_ref] Figure 1: Analysis of DNA fluctuations in vitro [/fig_ref]. We note that the consistency of the Zimm model to analyze fluctuation data in 'uncrowded' conditions was already reported in published fluorescence correlation spectroscopy studies [bib_ref] Precise characterization of the conformation fluctuations of freely diffusing DNA: beyond rouse..., Mchale [/bib_ref] [bib_ref] Dynamics of large semiflexible chains probed by fluorescence correlation spectroscopy, Lumma [/bib_ref]. We conclude that the motion of DNA loci can be accurately analyzed with polymer models, and that DNA bending stiffness can be inferred from realtime microscopy studies using the Rouse model in the case of crowded environments.
## Nucleosome array fluctuations are consistent with the rouse model
We then focused on the analysis of nucleosome array fluctuations in vitro. Because the Kuhn length of chromatin remains debated in the literature (18), we separated the presentation of experimental results from the analysis of the amplitude of spatial fluctuations with the Rouse model. We assembled nucleosome arrays on the same long DNA fragments with fluorescent stretches using human core histones and chromatin assembly factors to produce nucleosome arrays with a repeat length of ∼168 bp, as characterized by micrococcal nuclease digestion (Supplementary [fig_ref] Figure 4: Analysis of MSD and Hi-C data in yeast with the Rouse model [/fig_ref]. This nucleosome repeat length matches that of yeast chromatin, which is known to have an average nucleosome repeat of 167 bp that is equivalent to 20-22 bp of linker DNA [bib_ref] Nucleosome spacing and chromatin higher-order folding, Grigoryev [/bib_ref]. Electron microscopy revealed the characteristic pattern of 10 nm fibers, with gaps and random clustering of assembled chromatin [fig_ref] Figure 2: Nucleosome array fluctuations in vitro [/fig_ref] , see Methods). The resulting chromatin fibers were diluted in a crowded buffer, and the motion of chromatin loci was tracked by microscopy to measure the MSD and the SDF [fig_ref] Figure 2: Nucleosome array fluctuations in vitro [/fig_ref]. The aver- age MSD plotted in red in [fig_ref] Figure 2: Nucleosome array fluctuations in vitro [/fig_ref] was consistent with the Rouse model, as shown from the fit with an anomalous diffusion response characterized by parameters ( ,α) of (0.043 ± 0.005 m 2 /s 0.54 , 0.54 ± 0.03). Furthermore, despite the broad variability of the MSD response from fiber to fiber (gray datasets in [fig_ref] Figure 2: Nucleosome array fluctuations in vitro [/fig_ref] , possibly caused by the heterogeneous arrangements of nucleosome in single molecules and/or the residual random clustering observed in the electron micrographs [fig_ref] Figure 2: Nucleosome array fluctuations in vitro [/fig_ref] , the SDF appeared to adopt a Gaussian shape (solid lines in [fig_ref] Figure 2: Nucleosome array fluctuations in vitro [/fig_ref]. This behavior is predicted by the Rouse model [bib_ref] Quantitative analysis of single particle trajectories: mean maximal excursion method, Tejedor [/bib_ref]. More quantitatively, the width of the SDF is expected to be determined by the amplitude of the MSD:
[formula] SDF (x.τ ) = 2 π × MSD (τ ) e −x 2 /MSD(τ )(7) [/formula]
By plugging the measured MSD in [fig_ref] Figure 2: Nucleosome array fluctuations in vitro [/fig_ref] , we obtained a good fit between the experimental SDF and that predicted by the Rouse model (curves and corresponding fits in [fig_ref] Figure 2: Nucleosome array fluctuations in vitro [/fig_ref]. Finally, exploiting the spontaneous destabilization of nucleosome arrays in diluted conditions [bib_ref] Binding of ethidium bromide causes dissociation of the nucleosome core particle, Mcmurray [/bib_ref] , we compared the fluctuations of nucleosome arrays after assembly, and two days after their reconstitution, and detected a motion characterized by an amplitude 3.5-fold enhanced (data not shown). This response matched that of purified DNA molecules in these experimental conditions,
showing that the amplitude of the fluctuations of nucleosome arrays are lower than that of DNA.
## Inferring the mechanical properties of nucleosome arrays to test the rouse model
In order to check the relevance of the Rouse model for chromatin, we had to determine the monomer friction coefficient ζ and the Kuhn length b from structural models of nucleosome arrays and integrate these parameters in equation [bib_ref] Dynamic organization of chromatin domains revealed by super-resolution live-cell imaging, Nozaki [/bib_ref]. In order to sample the conformational space of nucleosome arrays thoroughly, we performed simulations with the nucleosome conformation of the crystal structure, i.e. with two turns of DNA wrapped around the histone core, hereafter referred to as the 'closed negative' structure (see Materials and Methods, [fig_ref] Figure 3: Inferring the mechanical properties of nucleosome arrays [/fig_ref] ,. We also considered an 'open' nucleosome state, in which the most distal histone-DNA binding sites at super-helix location ±6.5 are disrupted [bib_ref] Crystal structure of the nucleosome core particle at 2.8 A resolution, Luger [/bib_ref] , as detected in molecular biology assays (44) as well as single molecule techniques ((45) [fig_ref] Figure 3: Inferring the mechanical properties of nucleosome arrays [/fig_ref]. For both nucleosome configurations, nucleosome arrays were simulated with variable lengths of DNA linkers in the range 18 to 22 bp, i.e. for nucleosome repeat lengths of ∼167 ± 2 bp. Finally, in order to account for the structural defects of nucleosome arrays observed by electron microscopy [fig_ref] Figure 2: Nucleosome array fluctuations in vitro [/fig_ref] , we introduced some randomness in the positioning of nucleosomes associated to a deviation [fig_ref] Figure 3: Inferring the mechanical properties of nucleosome arrays [/fig_ref]. For each condition, we simulated 1000 independent chromatin fibers structures containing 100 nucleosomes. Note that nucleosome-nucleosome interactions were neglected in the simulations because our experiments have been carried out in low salt conditions with minimal oligomerization or intramolecular contacts, as inferred from analytical sedimentation [bib_ref] Conformational Dynamics of the chromatin fiber: determinants, mechanisms, and functions, Hansen [/bib_ref].
We then extracted the Kuhn length of each array using the end-to-end distance distribution as readout of the simulations [fig_ref] Figure 3: Inferring the mechanical properties of nucleosome arrays [/fig_ref] , see Materials and Methods). We found that the Kuhn length was larger for closed negative vs. open nucleosomes, altogether spanning 85-140 nm (third column of [fig_ref] Table 1: Mechanical, structural and hydrodynamic modeling of nucleosome arrays by Monte Carlo simulations... [/fig_ref]. Specifically, for regular arrays of negative nucleosomes with a repeat length of 165-169 bp, the Kuhn length was roughly constant in the range 120-140 nm, whereas fibers with open nucleosomes appeared to be slightly more flexible with a Kuhn length spanning 85-110 nm (Supplementary [fig_ref] Figure 2: Nucleosome array fluctuations in vitro [/fig_ref] and [fig_ref] Table 1: Mechanical, structural and hydrodynamic modeling of nucleosome arrays by Monte Carlo simulations... [/fig_ref]. These values of the Kuhn length matched the predictions of other simulations for closed negative nucleosomes in the range of 102-154 nm for nucleosome repeat length of 169 and 168 bp, respectively [bib_ref] Nucleosome positioning and composition modulate in silico chromatin flexibility, Clauvelin [/bib_ref] [bib_ref] Nucleosome geometry and internucleosomal interactions control the chromatin fiber conformation, Kepper [/bib_ref]. Measurements based on Hi-C in yeast also indicated that the Kuhn length was in a comparable size range of 110-230 nm [bib_ref] Mapping in vivo Chromatin interactions in yeast suggests an extended chromatin fiber..., Dekker [/bib_ref]. Last but not least, analytical models support the stiffness of fibers with short linkers of ∼20 bp with a Kuhn length of 100 nm or more [bib_ref] Chromatin: a tunable spring at work inside chromosomes, Ben-Haïm [/bib_ref].
These simulations also provided an estimate of the density of nucleosomes per 11 nm knowing the contour length and the number of nucleosomes for each fiber configuration (fourth column of [fig_ref] Table 1: Mechanical, structural and hydrodynamic modeling of nucleosome arrays by Monte Carlo simulations... [/fig_ref]. Nucleosome arrays with closed negative nucleosomes were more compact than fibers with open nucleosomes with a typical density in the range of 2.5-3.0 versus <2.0 nucleosomes/11 nm. Finally, we noticed that the addition of randomness in the positioning of nucleosomes tended to reduce the Kuhn length and to increase the density of nucleosomes [fig_ref] Table 1: Mechanical, structural and hydrodynamic modeling of nucleosome arrays by Monte Carlo simulations... [/fig_ref].
We subsequently inferred the monomer friction coefficient ζ by performing simulations based on stochastic rotation dynamics (see details in Supplementary Material 2). These simulations are based on the explicit description of solvent particles, which transfer momentum to floating particles [bib_ref] Stochastic rotation dynamics simulation of electro-osmosis, Ceratti [/bib_ref]. They were performed at the level of one Kuhn segment, which was modeled as an ensemble of 'spherical' nucleosomes of 10 nm in diameter arranged in space according to the structural models shown in e.g. [fig_ref] Figure 3: Inferring the mechanical properties of nucleosome arrays [/fig_ref] -C. We took three regularly repeated chromatin fibers with nucleosome densities of 1.2, 2.1, and 3.2 nucleosomes/11 nm. These simulations showed that the monomer friction coefficient ζ was proportional to the number of nucleosomes in one Kuhn segment for a low density of nucleosomes , whereas it was equal to half the number of nucleosomes for a high density of 3.2 nucleosomes/11 nm. Therefore, the friction coefficient appeared to decrease with the density of nucleosomes, confirming a trend documented for concentrated colloidal suspensions [bib_ref] Long-time dynamics of charged colloidal suspensions: hydrodynamic interaction effects, Nägele [/bib_ref]. Our results were finally extended to every simulation using linear interpolation of ζ as a function of the nucleosome density. For clarity, we report the hydrodynamic diameter of the Kuhn segment as defined by ζ /3πη in the fifth column of [fig_ref] Table 1: Mechanical, structural and hydrodynamic modeling of nucleosome arrays by Monte Carlo simulations... [/fig_ref].
Using equation (2), we computed the MSD amplitude for every fiber structure (sixth column of [fig_ref] Table 1: Mechanical, structural and hydrodynamic modeling of nucleosome arrays by Monte Carlo simulations... [/fig_ref]. Given that Γ = 0.043 ± 0.005 m 2 /s 0.5 in the experiments, the predictions of the simulations were in agreement with the fluctuation data if some degree of variability in the positioning of closed negative nucleosomes was allowed. The defects in nucleosome arrangements observed in [fig_ref] Figure 2: Nucleosome array fluctuations in vitro [/fig_ref] may therefore account for the variability of MSD responses but also constitute an essential ingredient to recapitulate the amplitude of in vitro fluctuation data. Altogether, our results confirm that the mechanical parameters inferred from nucleosome array models integrated into the Rouse model recapitulate motion data in vitro.
## Inconsistency of the rouse model for dynamic and hi-c data in living cells
We next investigated if and whether the Rouse model could be applied to study chromosomes in living budding yeast cells. We studied chromatin dynamics by recording 110 trajectories (49 800 displacements) for a locus on chromosome XII at an inter-frame interval of 0.2 s. The average MSD shown in the red dataset of [fig_ref] Figure 4: Analysis of MSD and Hi-C data in yeast with the Rouse model [/fig_ref] followed a temporal response consistent with the Rouse model associated to anomalous diffusion parameters ( ,␣) of (0.010 ± 0.001 m 2 /s 0.54 , 0.54) (black fit in [fig_ref] Figure 4: Analysis of MSD and Hi-C data in yeast with the Rouse model [/fig_ref] , in agreement with our previous measurements [bib_ref] High throughput chromatin motion tracking in living yeast reveals the flexibility of..., Hajjoul [/bib_ref]. Interpreting this data with equation (2), as derived in ref. [bib_ref] High throughput chromatin motion tracking in living yeast reveals the flexibility of..., Hajjoul [/bib_ref] , leads to a Kuhn length of less than 5 nm. We tested the consistency of this estimate with respect to published Hi-C data in the form of probability of contact normalized at 20 kbp vs. genomic distance (black, blue, red and burgundy curves in [fig_ref] Figure 4: Analysis of MSD and Hi-C data in yeast with the Rouse model [/fig_ref]. For this, we ran a coarse-grained lattice-based polymer simulation (see Methods) using n = 200 bp per monomer, a Kuhn length b = 5 nm, and the concentration of DNA set to 3 × 10 −3 bp/nm 3 as input parameters [bib_ref] BioNumbers--the database of key numbers in molecular and cell biology, Milo [/bib_ref]. We observed that the occurrence of distant contacts in the simulation was underestimated by more than one order of magnitude (cyan curve in [fig_ref] Figure 4: Analysis of MSD and Hi-C data in yeast with the Rouse model [/fig_ref]. Furthermore, despite the contradiction with MSD data, we could run the simulation for a Kuhn length of b ∼ 50 nm for n = 2 kb per monomer, because this value is more consistent with in vitro data [bib_ref] Mapping in vivo Chromatin interactions in yeast suggests an extended chromatin fiber..., Dekker [/bib_ref]. Yet, this parametrization only marginally improved the fitting of Hi-C data (green curve in [fig_ref] Figure 4: Analysis of MSD and Hi-C data in yeast with the Rouse model [/fig_ref]. Last, using our large ensemble of displacement measurements on chromosome XII, we compiled the SDF and detected a deviation to the Gaussian behavior expected for the Rouse model with a peaked distribution for time intervals of 0.4 and 0.8 s [fig_ref] Figure 4: Analysis of MSD and Hi-C data in yeast with the Rouse model [/fig_ref]. Overall, the Rouse model appeared to show some inconsistencies with structural and motion data in living yeast, suggesting that additional molecular parameters should be integrated in the model.
## Transient chromatin contacts account for chromosome dynamics and folding in yeast
Recent EM studies of thin nuclear sections of budding yeast have hinted to the existence of intra-chromosomal contacts [bib_ref] Budding yeast chromatin is dispersed in a crowded nucleoplasm in vivo, Chen [/bib_ref]. Although intramolecular contacts have not been thoroughly discussed in yeast, their contribution to higher-order chromosome folding in mammalian cells or drosophila has been proposed many times [bib_ref] Chromatin organization by an interplay of loop extrusion and compartmental segregation, Nuebler [/bib_ref] [bib_ref] TADs are 3D structural units of higher-order chromosome organization in Drosophila, Szabo [/bib_ref] [bib_ref] Predictive polymer modeling reveals coupled fluctuations in chromosome conformation and transcription, Giorgetti [/bib_ref]. We thus set out to explore if a Rouse model with transient internal contacts (RouseTIC) could fit chromosome structure and motion in vivo. We used the same simulations scheme, but introduced non-specific self-interaction in between the monomers (36). We defined k b and k u the binding and unbinding rates between monomers, respectively. In addition to these two kinetic parameters, the RouseTIC model relies on the defini-tion of a concentration of DNA set to 3 × 10 −3 bp/nm 3 , the Kuhn length and the characteristic time for the fluctuations. In order to estimate k b and k u , we first performed simulations of the contact probability normalized at 20 kb versus genomic distance and compared them with Hi-C measurements. These simulations were performed setting the number of base-pair by Kuhn length to 2 kb and 50 nm, i.e. with a density of 2.6 nucleosomes/11 nm that roughly match the results of our simulations in [fig_ref] Table 1: Mechanical, structural and hydrodynamic modeling of nucleosome arrays by Monte Carlo simulations... [/fig_ref] and those of experiments of ∼2.0 nucleosomes/11 nm based on Hi-C [bib_ref] Mapping in vivo Chromatin interactions in yeast suggests an extended chromatin fiber..., Dekker [/bib_ref]. Notably, Hi-C data map the steady-state conformation of a polymer, hence the occurrence of contacts can be recapitulated with one equilibrium kinetic parameter, namely the reaction constant k b /k u . These simulations were in good agreement with the experiments for k b /k u in the range 0.3-0.70 [fig_ref] Figure 5: Analysis of MSD and Hi-C data in yeast with the RouseTIC model [/fig_ref]. As an argument of self-consistency, we checked the validity of the RouseTIC model by comparing its predictions to distance measurements in between chromosome loci in fixed cells derived from ref. [bib_ref] The genome folding mechanism in yeast, Kimura [/bib_ref] [fig_ref] Figure 5: Analysis of MSD and Hi-C data in yeast with the RouseTIC model [/fig_ref]. For k b /k u spanning 0.3-0.7, we adjusted the Kuhn length and confirmed that its value of b = 53 ± 5 nm was compatible with the data.
We then wished to analyze motion data. In order to adjust the time step of the simulation, we first ran simulations with no interaction and compared their predictions to our in vitro MSD measurements [fig_ref] Figure 5: Analysis of MSD and Hi-C data in yeast with the RouseTIC model [/fig_ref]. The temporal parameter of the simulation was 0.55 ms. We finally focused on the motion of chromosome loci in living yeast with the RouseTIC model and k b /k u in the range 0.3-0.70 [fig_ref] Figure 5: Analysis of MSD and Hi-C data in yeast with the RouseTIC model [/fig_ref]. Because the amplitude of fluctuations appeared to primarily depend on the unbinding rate k u , we could determine both reaction kinetic parameters. We proceeded by setting k b /k u in the range 0.3-0.7 and then varied k u to fit MSD data [fig_ref] Figure 5: Analysis of MSD and Hi-C data in yeast with the RouseTIC model [/fig_ref] and Supplementary [fig_ref] Figure 5: Analysis of MSD and Hi-C data in yeast with the RouseTIC model [/fig_ref]. In this parameter space, the typical unbinding time 1/k u varied from tens of seconds to seconds (inset of [fig_ref] Figure 5: Analysis of MSD and Hi-C data in yeast with the RouseTIC model [/fig_ref]. Notably, the fit of MSD data was significantly better for k b /k u of 0.7 than 0.3 (see Supplementary [fig_ref] Figure 5: Analysis of MSD and Hi-C data in yeast with the RouseTIC model [/fig_ref].
In order to reinforce the relevance of the RouseTIC model, we focused on higher-order moments of the SDF [fig_ref] Figure 4: Analysis of MSD and Hi-C data in yeast with the Rouse model [/fig_ref]. The anomaly of the SDF with respect to the Gaussian response can be measured from the kurtosis, which is computed from the fourth-order moment of the SDF (see Equationin Materials and Methods). In the Rouse model, the kurtosis is equal to 2 ((31), dashed black line in [fig_ref] Figure 6: Analysis of chromatin motion based on the kurtosis [/fig_ref] , whereas an excess of kurtosis is detected for step distribution functions more peaked than a Gaussian function, as noticed for time intervals of 0.4 and 0.8 s in [fig_ref] Figure 4: Analysis of MSD and Hi-C data in yeast with the Rouse model [/fig_ref]. Expectedly, the trajectory-averaged kurtosis appeared to be equal to 2.4 in the small time limit for chromosomes in vivo (blue solid line in [fig_ref] Figure 6: Analysis of chromatin motion based on the kurtosis [/fig_ref]. In addition, we noticed a clear temporal signature in the kurtosis, which decreased after 1 to 3 s. Conversely, the kurtosis measured for purified chromatin fibers was lower and equal to ∼1.75, and roughly constant over a narrower temporal window (black solid line in [fig_ref] Figure 6: Analysis of chromatin motion based on the kurtosis [/fig_ref].
We finally computed the kurtosis of the simulated trajectories with the RouseTIC model. We used the fitted parameters of k b / k u = 0.4 and 0.7 and k u = 0.2 and 1 s −1 , respectively, because these values fit MSD and Hi-C data. Simulations showed an excess of kurtosis equal to ∼2.7 in the small time limit (dashed lines in [fig_ref] Figure 6: Analysis of chromatin motion based on the kurtosis [/fig_ref] , and a decrease after a typical time defined by the unbinding reaction time scale 1/k u . These predictions were in qualitative agreement with our data, as could be clearly evidenced by the phenomenological normalization of in vitro and in vivo kurtosis data by the same factor of 0.9, and the resulting superposition of experiments and simulations [fig_ref] Figure 6: Analysis of chromatin motion based on the kurtosis [/fig_ref]. Altogether, the analysis of the kurtosis tends to support the consistency of the RouseTIC model to reproduce structural and dynamic data of yeast chromosome in vivo with a steady-state polymer model that includes two kinetic parameters for the contact formation reaction.
# Discussion
## Origin of molecular interactions in rousetic
In this study, we monitored chromatin motion in vitro and in vivo by fluorescence microscopy, and analyzed experimental data with chromatin structure models and kinetic Monte Carlo simulations of self-interacting polymers. We propose that the Rouse model with transient internal contacts accounts for the in vivo dynamics of chromosomes [fig_ref] Figure 7: Molecular scenario for the RouseTIC model [/fig_ref] and determines the kinetic parameters of the association reaction for the first time in living yeast. Internal contacts appear to be detectable from the occurrence of frequent contacts along the chromosome contour [fig_ref] Figure 4: Analysis of MSD and Hi-C data in yeast with the Rouse model [/fig_ref] and the slow motion of chromatin loci with a deviation of the step distribution from the Gaussian response in the short time limit [fig_ref] Figure 4: Analysis of MSD and Hi-C data in yeast with the Rouse model [/fig_ref]. The latter consequence is qualitatively explained by the formation of transient contacts that tend to decrease the fraction of time a monomer freely diffuses, and restrain the displacement of freely diffusing monomers inside boundaries along the chromosome contour [bib_ref] Linear viscoelastic properties of transient networks formed by associating polymers with multiple..., Indei [/bib_ref]. Interestingly, the existence of these transient contacts is reminiscent of the mechanism of internal friction [bib_ref] Quantifying internal friction in unfolded and intrinsically disordered proteins with single-molecule spectroscopy, Soranno [/bib_ref] , which manifests itself if intramolecular conformational barriers lead to the dissipation of energy into internal degrees of freedom. Already proposed to be relevant for chromosomes [bib_ref] Effect of internal friction on biofilament dynamics, Poirier [/bib_ref] , internal friction is known to lead to slow fluctuations, as observed in our datasets.
Nevertheless, the molecular origin of the contacts in the RouseTIC model remains to be clarified. By simulating the structure of a fiber with 1000 nucleosomes at a volume density of 15% with the same structural models of chromatin as for deriving the Kuhn length, we suggest that numerous contacts are likely to pre-exist within chromosomes [fig_ref] Figure 7: Molecular scenario for the RouseTIC model [/fig_ref] , which could be stabilized by nucleosome interactions. By scaling the effective energy of the binding events, the RouseTIC model can help narrow down possible molecular origin of these interactions. In addition to the transient nature of binding events with a lifetime on the order of seconds, the effective energy E 0 of contacts between monomers, as determined from the ratio of binding and unbinding rates k b /k u (see Equation (5)), is weak in the range of E 0 ∼-0.3 to -0.5 k B T. Such strength for E 0 is far smaller than the energy associated with ATP hydrolysis of ∼20 k B T, but also than the attractive stacking energy between nucleosomes of -2.7 k B T (60). Thus, this energy cannot be directly attributed to stacking between two nucleosomes through specific contacts, such as those described between the histone H4 N-terminal tail of one nucleosome with the 'acidic patch' of an adjacent one [bib_ref] Higher-Order structures of chromatin: the elusive 30 nm fiber, Tremethick [/bib_ref]. Rather, we hypothesize that contacts predominantly correspond to transient, weak and labile interactions. Interestingly, an interaction energy E 0 of -0.3 to -0.5 k B T compares to the conditions of equilibrium between histone-tail DNA attraction and DNA-DNA electrostatic repulsion in physiological salt conditions, as inferred from X-ray scattering of nucleosome core particles [bib_ref] Salt-Induced conformation and interaction changes of nucleosome core particles, Mangenot [/bib_ref]. Thus, contacts may originate from unspecific and dynamic histone tail-DNA contacts within the chromatin fiber as observed in recent cryo-EM studies on nucleosome core particles (63), potentially mediated or perturbed by chromatin-binding proteins. The higher chromatin density in the nucleus of living cells in comparison to in vitro experiments is expected to favor the formation of contacts. Yet the contribution of these contacts on chromatin fluctuations in vitro may be investigated by modulating the ionic strength of the solution. Using analytical centrifugation, it has indeed been shown that addition of 2 mM divalent salt enhances chromatin compaction and oligomerization mediated by histone tails interactions [bib_ref] Conformational Dynamics of the chromatin fiber: determinants, mechanisms, and functions, Hansen [/bib_ref] [bib_ref] Statics and dynamics of the freely jointed polymer chain with Lennard-Jones interaction, Baumgärtner [/bib_ref].
## Relevance of contact reaction for the analysis of hi-c data in metazoan
RouseTIC may also be relevant to chromatin compartment segregation, in which the occurrence of chromosome intramolecular contacts mediated by nucleosomes has been suggested [bib_ref] ChromEMT: Visualizing 3D chromatin structure and compaction in interphase and mitotic cells, Ou [/bib_ref]. Compartments inferred from Hi-C contact maps could be modeled assuming limited attraction between the monomers [bib_ref] Chromatin organization by an interplay of loop extrusion and compartmental segregation, Nuebler [/bib_ref] [bib_ref] Modeling epigenome folding: formation and dynamics of topologically associated chromatin domains, Jost [/bib_ref] , although the amplitude and origin of this interaction remains mostly elusive. The estimate of monomer-monomer interaction energy in the range of E 0 ∼-0.3 to -0.5 k B T is notably close to the energy between monomers of -0.3 k B T for a polymer chain in theta conditions [bib_ref] Liquid droplet formation by HP1␣ suggests a role for phase separation in..., Larson [/bib_ref] , i.e. with balanced monomer-monomer and monomer-solvent interactions. Because polymers in theta conditions are poised to fold into globules or coils whether monomer-monomer interaction increases or decreases, respectively, we suggest that the residual interactions between monomers can be modulated to trigger compartment segregation. This mechanism is likely necessary but not sufficient to model metazoan datasets due to the essential contribution of active processes involving loop extrusion or transcription [bib_ref] Evolutionarily conserved principles predict 3D Chromatin Organization, Rowley [/bib_ref] [bib_ref] Chromatin organization by an interplay of loop extrusion and compartmental segregation, Nuebler [/bib_ref] [bib_ref] Modeling epigenome folding: formation and dynamics of topologically associated chromatin domains, Jost [/bib_ref]. Nevertheless, our recent analysis of chromosome structure in drosophila appears to corroborate the globule to coil transition (data not shown), suggesting that nucleosome-nucleosome interactions stabilization by e.g. heterochromatin-specific proteins, such as HP1, could promote local collapse into globular state and phase separation [bib_ref] HP1 and the dynamics of heterochromatin maintenance, Maison [/bib_ref] [bib_ref] Formation of chromatin subcompartments by phase separation, Erdel [/bib_ref]. Despite a strong impact on chromosome large-scale organization, the number of proteins implicated in this structural transition could be limited. This hypothesis deserves further investigation using heteropolymer models [bib_ref] Modeling epigenome folding: formation and dynamics of topologically associated chromatin domains, Jost [/bib_ref] and combining them to chromatin structure models in order to reach a molecular description of segregation.
[fig] Figure 1: Analysis of DNA fluctuations in vitro. (A) Schematics of the experiment: long DNA molecules with fluorescent stretches, which are shown with black and red segments, respectively, are dispersed in a crowded solution composed of a high concentration of PVP chains (gray lines). (B) The graph presents the temporal evolution of the MSD as a function of time for fluorescent DNA loci (gray dataset). The average response (red squares) is fitted with the Rouse model (black line), but not with the Zimm model (green line). (C) The graph shows the residuals of the fit with the Rouse model as a function of the Kuhn length. (D) The red dataset represents the average MSD over time for DNA loci dispersed in a solution without crowding. The green and black solid lines show the predictions of the Zimm and Rouse models, respectively, obtained by setting the Kuhn length to 100 nm and the viscosity to 2.3 mPa.s. [/fig]
[fig] Figure 2: Nucleosome array fluctuations in vitro. (A) Electron micrographs of chromatin fibers reconstituted on -DNA molecules. (B) The graph shows the temporal evolution of the MSD of chromatin fibers dispersed in a crowded solution with individual trajectory reported in gray and the average response in red with the standard error. The viscosity of the solution was 15 mPa.s. The black curve is the fit with a power-law response characterized by an anomalous exponent of 0.54. (C) The plot shows the step distribution function extracted from the same set of data as in (B) for three consecutive time lags indicated in the caption. The corresponding Gaussian fits are directly expressed from equation(7). [/fig]
[fig] Figure 3: Inferring the mechanical properties of nucleosome arrays. (A) Structure of a nucleosome array with closed negative nucleosomes and a linker length of 20 bp. Red spheres correspond to histone octamers. Gray segments represent DNA and the blue segments specifically show entry exit DNAs. (B) Structure of a nucleosome array with open nucleosomes and a linker length of 20 bp. (C) Structure of a nucleosome array with open nucleosomes and a variable linker length of 20 ± 5 bp. (D) The graph shows the end-to-end distance distribution as extracted from the simulation for a nucleosome array containing 20, 30, and 60 closed negative nucleosomes and a linker length of 20 bp. (E) The plot shows the contour length vs. Kuhn length deduced from the analysis of the moments of the end-to-end distance distribution. The colored dots correspond to the fit obtained with chromatin fibers containing 5-95 nucleosomes (plotted with dark blue to red colors, respectively). The three arrows show the results inferred from the analysis of the histograms shown in (D). The red line corresponds to the conditions for which the contour length is equal to two times the Kuhn length. range of 3 or 5 bp, i.e. linkers of 17 to 23 bp or 15 to 25 bp, respectively ( [/fig]
[fig] Figure 4: Analysis of MSD and Hi-C data in yeast with the Rouse model. (A) The graph shows 110 MSD responses in gray, the average response in red with the standard error to the mean, and the corresponding fit with the Rouse model in black. (B) The plot presents normalized contact frequency vs. genomic distance inferred from Hi-C measurements by Mercy et al. in red (68), Duan et al. in black (69), Marie-Nelly et al. in blue (70) and Belton et al. in burgundy (71). The orange sector serves to define the limits of experimental responses. The green and cyan datasets are the results of simulations for one homopolymer freely diffusing in space and characterized by a Kuhn length of 50 or 5 nm, respectively. (C) The plot presents the SDF at four consecutive time lags, as indicated in the legend. The trajectories are the same as in A. (D) Same data as in (C) in log-lin. [/fig]
[fig] Figure 5: Analysis of MSD and Hi-C data in yeast with the RouseTIC model. (A) The graph shows the comparison of experimental and simulated Hi-C contact frequency versus genomic distance. Experimental data are within the two orange limits defined in Figure 4B. The output of simulations fit experimental data for k b /k u in the range 0.3 to 0.7, as indicated in the legend at the top. (B) The plot shows the Kuhn length as a function of k b /k u that best fits 3D mean square distance measurements with the RouseTIC model. The inset shows an example of fit for k b / k u = 0.7. Note the error bars for the determination of the Kuhn length corresponds to acceptable fits within the gray limits in the inset. (C) The orange dataset shows the experimental MSD versus time for nucleosome arrays in vitro, as reported in Figure 2B, and the corresponding black curve is obtained with the RouseTIC model for a simulation time step of 0.55 ms. (D) The plot presents the MSD as a function of time for chromosomes in living yeast. The dashed orange line is the experimental measurement of 0.01 × t 0.54 and the upper and lower gray lines correspond to limits characterized by 0.013 × t 0.54 and 0.008 × t 0.54 . The blue, black and green curves are obtained from the RouseTIC model with k b / k u = 0.7 and different values of k u , as indicated in the legend. [/fig]
[fig] Figure 6: Analysis of chromatin motion based on the kurtosis. (A) The kurtosis is plotted as a function of time for the locus on chromosome XII or for reconstituted nucleosome arrays (blue and black datasets, respectively). The dashed black line corresponds to the response for a Gaussian process, as in e.g. Brownian motion or the Rouse model (31). The dashed gray and blue curves are deduced from simulations of the RouseTIC model using kinetic parameters that fit Hi-C and MSD data, which are reported in the caption. (B) Same graph as in (A) with the two experimental datasets divided by 0.9. [/fig]
[fig] Figure 7: Molecular scenario for the RouseTIC model. (A) The sketch represents the Rouse TIC model with the definition of the reaction rates k u and k b . The fiber displayed in the inset is obtained for k u /k b = 0.5. (B) Simulation of a chromatin fiber with a linker size of 20 bp and closed negative nucleosomes with a concentration of nucleosomes set to 3 × 10 −3 bp/nm 3 . 127 nucleosomes are highlighted in green because they potentially form intramolecular contacts, as defined by an inter-nucleosome distance in trans of <15 nm, i.e. comparable to the typical length of histone tails, and a genomic distance in cis of more than 1650 bp (10 nucleosomes). (C) The same simulation is carried out for a fiber with linker size of 22 bp and open nucleosomes. 456 nucleosomes are marked in green. [/fig]
[table] Table 1: Mechanical, structural and hydrodynamic modeling of nucleosome arrays by Monte Carlo simulations using different nucleosome conformations and repeat lengths. The last column corresponds to the amplitude of chromatin loci fluctuations according to the Rouse model [/table]
|
Concurrent Acute Monoblastic Leukemia and Multiple Myeloma in a 66-Year-Old Chemotherapy-Naive Woman
Concurrent acute myeloid leukemia (AML) and multiple myeloma (MM) is rare, more so in chemotherapy-naive patients. Concurrent occurrence of these two malignancies portends poor prognosis. Although anthracycline-based AML regimen, allogeneic hematopoietic stem cell transplantation, tipifarnib and bortezomib have shown promising results in small number of patients, there is a lack of established therapy. We describe a case of concurrent AML and MM in a 66-year-old woman and review previously published literature.
# Introduction
Concurrent acute myeloid leukemia (AML) and multiple myeloma (MM) is rare and presents unique therapeutic challenges. Here, we describe a case of concurrent AML and MM and review previously published literature.
## Case report
A 66-year-old woman, complaining of fatigue, loss of appetite and 25-pound weight loss for 8 weeks, was seen in the ambulatory clinic at a different institution after detection of abnormalities in blood work. Her medical history was significant for 80 pack-year smoking, chronic obstructive pulmonary disease, hypercholesterolemia, hypertension, coronary artery disease, coronary stent placement, tonsillectomy and appendectomy. The patient had a good performance status. Family history was significant for breast cancer in her mother and esophageal cancer in her father. Her medications included isosorbide mononitrate, aspirin, simvastatin, metoprolol, inhalational ipratropium, albuterol and fluticasone.
Physical examination revealed temperature of 96 °F, blood pressure of 107/73 mmHg, pulse of 98/min and respiratory rate of 16/min. The patient was calm and oriented. She had pallor but no icterus, petechiae, or any other evidence of bleeding. Lymph nodes were not palpable. There was no bony tenderness. Abdominal examination did not reveal hepatosplenomegaly or any other masses. Rest of the physical examination was also unremarkable.
Laboratory tests revealed white count of 64,500/µL with 17% granulocyte, 11.9% monocyte and 6.2% lymphocyte, hemoglobin of 7.5 g/dL, platelet of 100,000/µL and serum lactate dehydrogenase level of 306 IU/L. Peripheral blood smear showed 3% blasts, 73% monocytes with some immature forms, 10% neutrophils, 2% myelocytes, metamyelocytes and promyelocytes, and 4% lymphocytes. Glucose, electrolytes, coagulation profile, renal and liver function tests were within normal limits.
A bone marrow biopsy done at the other institution revealed hypercellular bone marrow with decreased bone marrow precursor cells and infiltration with plasma cells, which were kappa restricted. Flow cytometry of the bone marrow aspirate was significant for two abnormal populations: 1) myeloid cells (27% of non-erythroid cells) with immunophenotype suggestive of monocytic differentiation; 2) plasma cells (33% of total cells) positive for monoclonal IgG, kappa-restricted, positive for CD38, CD117 and CD56 consistent with MM. The patient was referred to our institution for further management.
Peripheral blood smear at our institution showed 12% blasts with Auer rods. A repeat bone marrow biopsy showed hypercellular bone marrow (70-80%) with 50-60% blasts, decreased megakaryocytes, maturing myeloid and erythroid The patient was started on allopurinol and hydroxyurea, and the latter was subsequently stopped once chemotherapy was initiated. She was induced as an in-patient with 7-day low-dose cytarabine (100 mg/m 2 /day) and 3-day idarubicin (12 mg/m 2 /day).
Her hospitalization was complicated by Clostridium difficile diarrhea, neutropenic sepsis, and hospital acquired pneumonia secondary to Stenotrophomonas maltophila and vancomycin resistant Enteroccocus faecium. She was aggressively treated in intensive care unit but she developed multiorgan failure and died on hospital day 23.
# Discussion
AML in a patient with MM is very rare and often occurs after chemotherapy for MM [bib_ref] Multiple myeloma and acute myelomonocytic leukemia, Kyle [/bib_ref] [bib_ref] The chemotherapy on plasma-cell myeloma and the incidence of acute leukemia, Bergsagel [/bib_ref]. Concurrent AML and MM in chemotherapy-naive patients is extremely rare with only few cases reported in the literature [fig_ref] Table 1: Manuscript accepted for publicationDecember 18, 2013 Review of previous reported cases of... [/fig_ref] [bib_ref] Simultaneous acute myeloid leukemia and multiple myeloma successfully treated with allogeneic stem..., Kim [/bib_ref] [bib_ref] Simultaneous appearance of dual malignancies of hematopoietic systemmultiple myeloma and acute myeloid..., Shukla [/bib_ref] [bib_ref] Simultaneous occurrence of multiple myeloma and acute myeloid leukaemia, Attili [/bib_ref] [bib_ref] Simultaneous presentation of multiple myeloma and acute monocytic leukemia, Luca [/bib_ref] [bib_ref] Coexistence of myelomonocytic leukemia and monoclonal gammopathy or myeloma. Simultaneous presentation in..., Raz [/bib_ref] [bib_ref] Simultaneous presentation of acute myelomonocytic leukemia and multiple myeloma, Cleary [/bib_ref] [bib_ref] Simultaneous occurrence of acute myeloblastic leukaemia and multiple myeloma without previous chemotherapy, Tursz [/bib_ref] [bib_ref] Report of 12 cases and review of the literature, Rosner [/bib_ref] [bib_ref] Concomitant myelomonocytic leukemia and multiple myeloma, Taddeini [/bib_ref] [bib_ref] A case of acute myelomonocytic leukaemia associated with myelomatosis, Parker [/bib_ref] [bib_ref] Multiple myeloma and acute leukemia, Kastanas [/bib_ref] [bib_ref] Multiple myeloma and acute myelomonocytic leukemia: simultaneous occurrence without previous chemotherapy, Annino [/bib_ref] [bib_ref] Simultaneous presentation of acute myelomonocytic leukaemia and multiple myeloma, Parapia [/bib_ref] [bib_ref] Simultaneous occurrence of multiple myeloma and acute myeloblastic leukemia: fact or myth?, Rosner [/bib_ref]. Acute myeloblastic or myelo-monocytic leukemia are the most common AML sub-types encountered in this setting. Postulated mechanisms to explain the concurrent occurrence of AML and MM include disorder of multi-potent stem cell [bib_ref] Simultaneous presentation of acute myelomonocytic leukemia and multiple myeloma, Cleary [/bib_ref] [bib_ref] Simultaneous presentation of plasma cell and monocytic leukemia with a subacute clinical..., Naparstek [/bib_ref] , exposure to common risk factors [bib_ref] Simultaneous presentation of multiple myeloma and acute monocytic leukemia, Luca [/bib_ref] , leukemic or myeloma cells stimulating proliferation of bone marrow cells with subsequent development of a second hematologic malignancy [bib_ref] Coexistence of myelomonocytic leukemia and monoclonal gammopathy or myeloma. Simultaneous presentation in..., Raz [/bib_ref] or AML occurring co-incidentally while monoclonal gammopathy of undetermined significance is progressing to MM [bib_ref] Simultaneous acute myeloid leukemia and multiple myeloma successfully treated with allogeneic stem..., Kim [/bib_ref].
Reactive plasmacytosis is common in AML patients with an incidence of up to 6% and can be associated with monoclonal paraproteins in the absence of other components of the diagnostic criteria of MM [bib_ref] Reactive plasmacytosis and lymphocytosis in acute myeloid leukemia, Rosenthal [/bib_ref]. Therefore, it is important to rule out reactive plasmacytosis before making a diagnosis of concurrent AML and MM.
Given the rarity of the disease, there has been no established treatment and hence, prognosis remains extremely poor. Patients are often treated with therapy for AML since AML is more aggressive and anthracyclines are effective against MM as well. There are two reported cases of concurrent MM and AML in chemotherapy-naive patients, who had better outcomes compared to other patients. Raz et al describe a 68-year-old man who was diagnosed in 1978 with MM and AML without any "bizarre chromosomal changes" on cytogenetic studies. The patient was initially treated with cyclophosphamide with disappearance of monoblasts. The disease, however, recurred in 2 months when he was treated with melphalan without success and thereafter with combination chemotherapy consisting of vincristine, melphalan, cyclophosphamide and prednisone. This resulted in the disappearance of monoblasts as well as significant decline in the serum level of paraprotein. Six months later, unfortunately he developed end-organ damage related to MM and became unresponsive to chemotherapy. He died of septic shock and severe bleeding tendency 2.5 years from the time of initial diagnosis [bib_ref] Coexistence of myelomonocytic leukemia and monoclonal gammopathy or myeloma. Simultaneous presentation in..., Raz [/bib_ref]. Kim et al describe a 51-year-old previously healthy man who was diagnosed with concurrent MM and AML with complex cytogenetics and immunoglobulin heavy chain rearrangement. The patient was initially treated with one-cycle of bortezomib and then with cytarabine (ara-C) and idarubicin, along with bortezomib. This was followed by re-induction with mitoxantrone and high-dose ara-C and mitoxantrone (HAM regimen) for induction failure. Because of incomplete response, the patient subsequently went on receiving myeloablative allogeneic hematopoietic stem cell transplantation (allo-SCT) after conditioning with busulfan and cyclophosphamide. The patient remains to be diseasefree and well at 421 days post-SCT [bib_ref] Simultaneous acute myeloid leukemia and multiple myeloma successfully treated with allogeneic stem..., Kim [/bib_ref]. There has been some promising result in a preclinical study in which tipifarnib and bortezomib have been shown to be synergistic in MM and AML cell lines [bib_ref] Tipifarnib and bortezomib are synergistic and overcome cell adhesionmediated drug resistance in..., Yanamandra [/bib_ref]. Although the usefulness of these therapies needs further evaluation, because of the rarity of the disease, it will require an international registry to be able to do so.
# Conclusion
Although very rare, AML and MM can present simultaneously even in chemotherapy-naive patients. Concurrent occurrence of these two malignancies portends poor prognosis. Anthracycline-based AML regimens are often used because AML is more aggressive and anthracyclines are effective against MM as well. Allogeneic stem cell transplantation as well as tipifarnib and bortezomib have shown promising re-sult.
## Financial support
None.
[table] Table 1: Manuscript accepted for publicationDecember 18, 2013 Review of previous reported cases of concurrent AML and MM *Patient had pancytopenia for 3 years prior to the diagnosis. †Patient had pancytopenia for 5 years prior to the diagnosis. **Cases described in Rosner and Grunwald[17]. K: kappa light chain; λ: lambda light chain; F: female; M: male; NA: not available. [/table]
|
Toxin yet not toxic: Botulinum toxin in dentistry
Paracelsus contrasted poisons from nonpoisons, stating that ''All things are poisons, and there is nothing that is harmless; the dose alone decides that something is a poison". Living organisms, such as plants, animals, and microorganisms, constitute a huge source of pharmaceutically useful medicines and toxins. Depending on their source, toxins can be categorized as phytotoxins, mycotoxins, or zootoxins, which include venoms and bacterial toxins. Any toxin can be harmful or beneficial. Within the last 100 years, the perception of botulinum neurotoxin (BTX) has evolved from that of a poison to a versatile clinical agent with various uses. BTX plays a key role in the management of many orofacial and dental disorders. Its indications are rapidly expanding, with ongoing trials for further applications. However, despite its clinical use, what BTX specifically does in each condition is still not clear. The main aim of this review is to describe some of the unclear aspects of this potentially useful agent, with a focus on the current research in dentistry. Ó 2015 The Author. Production and hosting by Elsevier B.V. on behalf of King Saud University. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).Abbreviations: BTX, botulinum neurotoxin; SNARE, soluble N-ethylmaleimide-sensitive factor attachment protein receptor; SNAP-25, synaptosomal-associated protein; MPDS, myofacial pain dysfunction syndrome; EMG, electromyography; TGF-b1, transforming growth factor b-1
# Introduction
Botulinum neurotoxin (BTX) is a neurotoxic protein produced by the Gram-positive, rod-shaped, spore-forming, and strictly anaerobic bacterium Clostridium botulinum and, rarely, by Clostridium butyricum and Clostridium baratii, commonly found on plants and in soil, water, and animal intestinal tracts. Although once considered lethal, BTX is now used as a therapeutic drug. BTX exhibits transient, nondestructive, dosedependent, and localized actions, with minimal systemic side effects [bib_ref] Postsurgical role of botulinum toxin-A injection in patients with head and neck..., Marchese [/bib_ref] , underlying its wide use in various orofacial and dental disorders. The exact mechanism of action, dosage, and delivery procedure of BTX are very important. In addition to conditions for which BTX is currently used as a therapeutic agent, evidence supports the expansion of its indications in dentistry. The purpose of this review is to provide insights into the current indications of BTX, highlight its expanding use, and review recent advances in the use of BTX in dentistry.
## Sites and modes of action of btx
Modes of action of BTX are summarized in [fig_ref] Table 1: Action of botulinum toxin [/fig_ref] [bib_ref] Botulinum toxins: pharmacology and its current therapeutic evidence for use, Muthane [/bib_ref]. BTX induces muscle weakness by inhibiting transmission of alpha motorneurons at the neuromuscular junction. Release of acetylcholine (ACh) is mediated by the assembly of synaptic fusion complexes-a set of soluble N-ethylmaleimide-sensitive factor attachment protein receptor (SNARE) proteins, including synaptobrevin, synaptosomal-associated protein (SNAP), and syntaxin. Seven BTX serotypes have been identified. BTX types B, D, F, and G cleave synaptobrevin; types A, C, and E cleave SNAP-25; and type C cleaves syntaxin [bib_ref] Therapeutic and cosmetic uses of botulinum toxin, Kant [/bib_ref] [bib_ref] Botulinum toxin. From poison to medicine, Davis [/bib_ref].
## Forms of btx
Commercially available various forms of BTX are summarized in [fig_ref] Table 2: Forms of botulinum toxin [/fig_ref] [bib_ref] Application of Botulinum toxin type A: an arsenal in dentistry, Rao [/bib_ref]. Forms of BTX range in weight [bib_ref] Botulinum toxins: pharmacology and its current therapeutic evidence for use, Muthane [/bib_ref]. Crude toxin is purified and diluted with human serum albumin. Each vial of 100 units (U) is reconstituted with preservative-free normal saline immediately before use. BTX begins to act within 24-48 h of administration, peaking at 2-3 weeks and maintaining its efficacy for about 3-4 months [bib_ref] Botulinum toxins: pharmacology and its current therapeutic evidence for use, Muthane [/bib_ref]. Secondary nonresponsiveness could be encountered due to the production of neutralizing antibodies. To overcome this effect, newer toxins with higher activity levels are being engineered. Antibody production could be prevented by avoiding repeated injections and keeping the dose as low as possible [bib_ref] Botulinum toxin, Nigam [/bib_ref].
## Injection procedure
The BTX dose should be tailored to the severity of the condition. Toxin is injected with a 1-to 1.5-inch, 25-to 30-gauge needle, with electromyography (EMG) monitoring. Subsequent injections can be given according to the response after 3 months.
## Adverse effects
Side effects of BTX use [fig_ref] Table 3: Side effects of botulinum toxin use [/fig_ref] are generally transient, but could last up to several months after administration.4.1. Pain disorders BTX type A inhibits the calcium-dependent release of substance P in the embryonic dorsal root ganglia, producing an analgesic effect through peripheral inhibition of C and A delta fibers [bib_ref] Capsaicinstimulated release of substance P from cultured dorsal root ganglion neurons: involvement..., Purkiss [/bib_ref]. Peripheral and central analgesic effects of BTX arise from the direct and indirect (retrograde transport) effects of the toxin on peripheral nociceptive neurons.
## Applications in dentistry
## Trigeminal and postherpetic neuralgia
When used to treat intractable and idiopathic neuralgia, BTX type A acts by inhibiting the exocytosis of ACh and other neurotransmitters. This action could be analgesic if it prevents the release of neuropeptides from nociceptive nerve endings. BTX type A inhibits the release of norepinephrine and ATP from postganglionic sympathetic nerve endings, providing an analgesic effect and reducing central and peripheral sensitization. The appropriate dose of BTX for treating trigeminal neuralgia is 20-50 U, injected at the trigger zone or into the masseter muscle [bib_ref] Beneficial effects of botulinum toxin type A in trigeminal neuralgia, Zuniga [/bib_ref].
Recently, BTX type A has been used as an alternative treatment modality for refractory cases of postherpetic neuralgia [bib_ref] The efficacy of intradermal injection of botulinum toxin in patients with post-herpetic..., Emad [/bib_ref]. When provided intradermally at 15 U, BTX inhibits the release of formalin-induced glutamate, substance P, and calcitonin gene-related polypeptide (CGRP), with direct effects on sensory neurons and indirect effects on the central nervous system (CNS). Pain decreases to mild and tolerable levels within the first week of administration.
## Headache and migraine
Headache may be due to abnormal excitation of the peripheral nociceptive afferent fibers, leading to central sensitization and an increase in pericranial muscle hardness and tenderness. Headache may also be due to enhanced responsiveness of the trigeminal nucleus caudalis neurons, leading to the generation of pain signals that decrease pain modulation involving serotonin and norepinephrine. Migraine is a neurovascular pain syndrome associated with sterile inflammation and vasodilatation, which activate the trigeminal afferents on the vessel wall and cause pain. In cluster headache, the ophthalmic branch of the trigeminal nerve relays a pain signal, leading to the release of substance P and CGRP, vasodilatation of the dural blood vessels, and neurogenic inflammation. 1. Oro-facial pain conditions like trigeminal neuralgia, postherpetic neuralgia, migraine, headache and myofacial pain dysfunction 2. Salivary gland disorders like sialorrhea, sialocele, Frey's syndrome etc. 3. Hypertrophy of masseter muscle 4. TMJ disorders like dislocation, bruxism, and oromandibular dystonia and arthritis 5. Trismus 6. Gummy smile 7. Disorders of the facial nerve i.e. Facial nerve palsy/paresis 8. Cancer therapy 9. Botulinum toxin as a carrier for oral vaccines 10. Preparation of oral cavity for microsurgical reconstruction 11. During dental implant, jaw and periodontal surgeries. 12. In wound healing 13. Treatment of hypertrophic scars Four modes of action for BTX in migraine and headache have been proposed. (1) BTX may decrease muscle contractility by preventing ACh release from the presynaptic terminal. However, this theory does not explain why pain relief with BTX often occurs before muscle relaxation [bib_ref] Botulinum toxin A in the treatment of headache syndromes and pericranial pain..., Gobel [/bib_ref]. (2) BTX may inhibit the extrafusal muscle fibers and normalize excessive levels of muscle spindle activity [bib_ref] Extrafusal and intrafusal muscle effects in experimental toxin A injection, Rosales [/bib_ref] [bib_ref] Botulinum toxin A in the treatment of headache syndromes and pericranial pain..., Gobel [/bib_ref]. Excessive release of ACh at the neuromuscular junction leads to abnormally high levels of end-plate activity, resulting in extrafusal muscle contraction. BTX inhibits transmission of neurotransmitters at the gamma motorneurons in the muscle spindle and decreases muscle activity. (3) BTX may enter the CNS, modulating pain by inhibiting the release of substance P from the trigeminal nerve endings and activating the expression of substance P in the raphe nuclei. The BTX-mediated inhibition of SNAP-25 blocks neurotransmitter exocytosis, leading to a decreased pain through the colocalization of vasoactive intestinal peptide and neuropeptide Y with ACh in the parasympathetic neurons. Further contributing to the analgesic effect is the decompression of afferent nociceptive neurons, which leads to a decrease in excitatory metabolite levels secondary to muscle relaxation [bib_ref] The action of Botulinum A neurotoxin on the inhibition by antidromic stimulation..., Weigand [/bib_ref]. (4) BTX may reduce parasympathetic outflow, leading to analgesia. For this reason, BTX use is indicated in cases of chronic headache [bib_ref] Human in vivo evidence for trigeminovascular activation in cluster headache: neuropeptide changes..., Goadsby [/bib_ref].
Methods of BTX administration currently used include a fixed-site approach, following the pain, direct injection into the tender muscles, and a combination of these methods. Glabellar injections can provide complete relief. Other injection sites include the suboccipital region, where higher or lower doses are necessary for the posterior or anterior (e.g., frontal and temporal) regions, respectively. Doses of BTX used for headache and migraine range from 10 to 150 U, with clinical improvement within the first 2 weeks of injection, maximum benefit at about 6 weeks, and efficaciousness for 3 months. For cluster headache, commons doses are 24-150 U of BTX type A or the equivalent BTX type B (1200 U) [bib_ref] Botulinum toxins in the treatment of migraine and tension-type headaches, Winner [/bib_ref].
## Myofacial pain dysfunction syndrome (mpds)
A complex pain syndrome with an unclear etiology, MPDS is associated with pain and tenderness of the muscles, especially those involved in mastication, and with trigger points/bands. Injection of 50 U of BTX type A is a simple and effective means to reduce the muscle hyperactivity of MPDS, by blocking ACh release from the neuromuscular junction [bib_ref] Randomized controlled trial of botulinum toxin A for chronic myogenous facial pain, Nixford [/bib_ref]. Although this mechanism of BTX has been studied extensively, the results have been inconclusive owing to the unclear etiology. For example, EMG studies of MPDS patients have not consistently shown muscle hyperactivity. Other studies reported myositis as the underlying cause of pain in MPDS; however, myositis cannot be treated by BTX as the toxin does not have anti-inflammatory activity. Furthermore, the presence of trigger points in MPDS has not been confirmed. Overall, insufficient prospective randomized controlled studies have been performed to prove the effectiveness of BTX use in MPDS. Until recently, BTX has only been used as a temporary therapy to alleviate pain and dysfunction in the disorder [bib_ref] Uses of botulinum toxin a for treatment of myofacial pain and dysfunction, Fallah [/bib_ref] [bib_ref] Botulinum toxin A in the treatment of myofacial pain and dysfunction: the..., Laskin [/bib_ref].
## Disorders of the salivary glands
BTX has been used in the treatment of disorders of the salivary glands, such as sialorrhea, sialocele, and Frey's syndrome. When used in the treatment of hypersalivation (sialorrhea), BTX acts on the cholinergic nerve endings, causing proteolysis of SNAP-25, chemical denervation, and loss of neuronal activity. BTX is injected intraglandularly, mainly to the parotid gland, at a dose that depends on the condition. Common dose ranges are 10-100 U of Botox or 20-300 U of Dysport [bib_ref] Salivary gland application of botulinum toxin for the treatment of sialorrhea, Fuster Torres [/bib_ref]. Sialocele is an accumulation of saliva surrounded by a tissue reaction, which develops as a postoperative complication after parotidectomy. When used for this condition, BTX type A acts by blocking ACh release from the secretomotor parasympathetic autonomic nerve. Doses of 50-70 U are given percutaneously in the parotid region [bib_ref] Use of botulinum toxin type A in a case of persistent parotid..., Chow [/bib_ref]. Another common complication of parotidectomy, Frey's syndrome is characterized by facial hyperhidrosis in the preauricular region initiated by a gustatory stimulus. The mode of action of BTX type A in Frey's syndrome involves inhibition of ACh release at the nerve endings and in the muscles, and blockade of the motor end plates. At the autonomic level, sweat secretion is blocked in glands that depend on ACh release for their activation. BTX is typically administered at 30 U, and its efficacy lasts 6-15 months. Repeated administration decreases the symptom severity, reduces the extent of the affected area, and increases the time to relapse .
## Masseter muscle hypertrophy
Injection of BTX type A is a minimally invasive treatment modality for massetric hypertrophy, defined as the asymptomatic enlargement of one or both masseter muscles. The masseter muscle is injected with 100 U of BTX type A in 2 ml of sterile saline. The only limitation of this therapy is recurrence after 6 months, when the procedure must be repeated [bib_ref] Treatment of masseteric hypertrophy with botulinum toxin: a report of two cases, Bas [/bib_ref].
## Temporomandibular disorders (tmds)
TMDs are a group of nonodontogenic facial pain disorders associated with the temporomandibular joint (TMJ) or associated muscles.
## Tmj dislocation
Dislocation of the TMJ is caused by excessive forward movement of the condyle beyond the articular eminence, with complete separation of the articular surfaces and positional fixation. By inhibiting ACh release at the neuromuscular junction and weakening the lateral pterygoid muscles via chemodenervation, BTX type A causes an imbalance between the muscles used for opening and closing the jaw. These effects of BTX limit mouth opening and help to prevent dislocation [bib_ref] Longterm efficacy of botulinum toxin type A for the treatment of habitual..., Fu [/bib_ref] , with effects lasting 2-4 months. The BTX dose for this purpose is 25-50 U, injected percutaneously into each lateral pterygoid muscle.
## Bruxism
Observed in individuals while awake or asleep, bruxism is an involuntary disorder manifested by jaw clenching, tooth gnashing, and grinding. For bruxism treatment, BTX type A is injected into the masseter muscle at 60 U per side. A dose range of 25-100 U elicits a good response for up to 3-4 months [bib_ref] Treating severe bruxism with botulinum toxin, Tan [/bib_ref].
## Oromandibular dystonia
These involuntary spasms of the masticatory, lingual, and pharyngeal muscles result in distortions of the oral position and function. Modes of action of BTX type A in oromandibular dystonia include local chemodenervation of the motor end plates and central intracortical inhibition, which normalizes the distorted primary motor cortex projection. Which muscle is injected depends on the form of dystonia, with the bilateral masseter muscles being injected for jaw-closure dystonia, lateral pterygoids with the anterior belly of the omohyoid muscle for jaw-opening dystonia, and tongue muscles for lingual dystonia. The recommended dose of BTX is 30 U per side.
## Arthritis.
This inflammatory joint disease manifests as joint pain and dysfunction, with joint contractures and muscle atrophy in advanced stages. The pain in chronic arthritis is amplified by neuropeptide release in the periphery. BTX type B inhibits neuropeptide release, thereby altering the nociceptor function and reducing pain and neurogenic inflammation. BTX also causes chemodenervation of the articular pain fibers [bib_ref] Analgesic effects of intra-articular botulinum toxin Type B in a murine model..., Anderson [/bib_ref].
## Trismus
Trismus is defined as a motor disturbance of the trigeminal nerve, especially spasm of the masticatory muscles, with difficulty in opening the mouth. In this disorder, BTX type A acts at the synaptic terminal of the cholinergic lower motorneuron and causes flaccid paralysis due to blockade of neuroexocytosis at the lower motorneuron terminal presynaptically [bib_ref] Botulinum toxin A for trismus in cephalic tetanus, Andrade [/bib_ref]. The recommended dose of BTX for trismus is 25 U injected into each masseter muscle and 10 U into the temporalis muscle.
## Gummy smile
Gummy smile is defined as the display of excessive gingival tissue in the maxilla upon smiling, caused by hyperfunctional muscles of the upper lip. Treatment with BTX type A provides effective, minimally invasive, and temporary improvement (for 3-6 months) of gummy smile [bib_ref] Botulinum toxin type A in the treatment of excessive gingival display, Polo [/bib_ref]. BTX acts by cleaving SNAP-25, which blocks ACh release from motorneurons and enables repolarization of the postsynaptic terminal, resulting in partial chemodenervation and blockade of muscular contraction. Muscles are injected close to the nasalis or orbicularis oculi, with some muscle fibers intermeshing the levator labii superioris, levator labii superioris alaeque nasi, levator anguli oris, and zygomaticus major and minor. The ideal dose of BTX is about 2.5 U per side at the levator labii superiori and zygomaticus sites, and 1.25 U per side at the orbicularis oculi sites.
## Facial nerve palsy/paresis
Facial palsy refers to both incomplete loss (paresis) and complete loss (paralysis) of facial nerve function. Unilateral palsy affects the balance between the right and left sides of the face, causing asymmetry. Injection of BTX type A into the contralateral lower facial muscles weakens these muscles and restores facial symmetry. BTX acts at the neuromuscular junction by inhibiting the release of ACh and preventing muscle contraction. A dose of 10-80 U of BTX in saline is given intramuscularly, with the precise dose being tailored to each patient and monitored by EMG. On average, the effect begins within 6 days and lasts 7-24 weeks. BTX can be re-administered, depending on the response. A minor and self-limiting side effect of the treatment is drooling of saliva [bib_ref] Botulinum toxin to improve lower facial symmetry in facial nerve palsy, Sadiq [/bib_ref].
## Cancer therapy
Adjunctive treatment with BTX type A can be used to potentiate the tumor response to chemo-or radiotherapy, by opening the vascular bed. Local administration of the toxin promotes tumor perfusion and oxygenation, and modulates the vasoreactivity of vessels. BTX potentiates with the release of noradrenalin, a vasomodulator that maintains sympathetic vascular tone through activation of the vascular smooth muscle adrenoceptors, in arterioles co-opted by the surrounding tumor. Greatest effects are seen when BTX is injected 3 days before beginning anticancer treatment [bib_ref] Botulinum toxin potentiates cancer radiotherapy and chemotherapy, Ansiaux [/bib_ref].
## Carrier for oral vaccines
Recently, molecular biological techniques have been used to generate an expression product of BTX type A. While losing its neurotoxic effect, the expression product retains the abilities to escape the gut, reach the general circulation, and evoke an immune response. After the toxin is ingested, it traverses a portion of the gastrointestinal system and is transcytosed from the gut lumen to the general circulation. Circulating toxin binds to peripheral cholinergic nerve endings, is endocytosed, and acts as a metalloendoprotease to cleave essential polypeptides for exocytosis. The most important mechanism for the toxin to penetrate the gut cells is its specific binding to receptors on the mucosal side of polarized gut cells. Bound toxin is actively transported across cells and delivered intact and unmodified on the serosal side of the monolayer [bib_ref] Botulinum toxin as a carrier for oral vaccines, Simpson [/bib_ref].
## Oral cavity reconstruction
In spite of good postoperative care, most oral cancer patients who are treated by tumor excision, neck dissection, and reconstruction will encounter complications, including infections, wound dehiscence, and fistula formation. These complications are caused by saliva stagnation due to reduced saliva clearance, limited capacity to swallow, and increased saliva production. Infiltration of BTX type A into the major salivary gland 4 days before surgery can help overcome these complications. Before injection, sialometry and sialography are performed, and gland markings are made. Then, 3-4 injections are given to each gland, with a total dose of 80-100 U of BTX. The peak effect is seen on days 5-8 after injection. Treatment results in a 50-70% reduction of salivary secretion [bib_ref] Botulinum toxin A for oral cavity cancer patients: in microsurgical patients BTX..., Corradino [/bib_ref].
## Dental implants, and jaw and oral surgery procedures
Failure of implant placement is mainly due to the lack of osseous integration, which could be due to excessive functional forces in patients with parafunctional habits. Treatment of maxillofacial (e.g., zygomatic and condylar) fractures requires multiple fixation sites and hardware to overcome the forces of masticatory musculature that prevent callus formation. Prophylactic injection of 100 U of BTX type A into the masseter muscle bilaterally 12-48 h before surgery could be beneficial in reducing these forces. In periodontal surgeries, BTX injection can reduce periodontal trauma due to excessive muscular function [bib_ref] Application of Botulinum toxin type A: an arsenal in dentistry, Rao [/bib_ref].
## Wound healing
Healing of traumatic, surgical, or other wounds (e.g., fissures and ulcers) involves multiple processes (e.g., hemostasis, inflammation, tissue proliferation, and remodeling), disruption of which can lead to a chronic wound. Increases in metabolic activity and inflammation during the healing process induce muscle contraction around the wound edges. Recently, experimental treatment with BTX type A has been attempted for wound healing, based on the ability of BTX to eliminate dynamic tension on and around healing tissues. This chemoimmobilization can potentially improve healing and minimize scarring for optimal esthetics [bib_ref] Kinetic and reaction pathway analysis in the application of botulinum toxin A..., Lebeda [/bib_ref].
Cleft lip and palate repair is generally associated with distorted facial growth and retarded development of the midfacial region. Causes of these effects have been attributed to a tense cheiloplasty and excessive lifting of the soft tissue, which causes tension on the healing wound. Intraoperative injection of BTX type A into the medial and lateral portions of the orbicularis oris muscle has been shown to decrease muscle activity, thereby decreasing tension at the surgical wound and resulting in better healing. There has been an encouraging trend in support of using BTX in wound-healing paradigms, although additional studies are necessary to determine a standardized approach [bib_ref] Use of botulinum toxin in cheiloplasty: a new method to decrease tension, Gala´rraga [/bib_ref].
## Hypertrophic scars
Increased deposition of collagen fibers and extracellular matrix can lead to hypertrophic scars. BTX type A influences the fibroblast cell cycle, causing decreased proliferation and increased apoptosis. These effects, in turn, lead to decreased expression of TGF-b1 protein. Injection with BTX type A decreases tension at the healing site to prevent scar formation. Studies of BTX use for hypertrophic scars have only been performed in vitro, but the results are encouraging for further in vivo studies to elucidate the mechanism and standardized procedure for BTX use in this context [bib_ref] Effects of botulinum toxin type A on collagen deposition in hypertrophic scars, Xiao [/bib_ref].
## Contraindications for btx use
Contraindications and precautions for BTX use in dentistry are summarized in [fig_ref] Table 5: Contraindications/precautions for use of botulinum toxin [/fig_ref] [bib_ref] Botulinum toxins: pharmacology and its current therapeutic evidence for use, Muthane [/bib_ref].
# Conclusions
As a group, the BTXs are the most potent of known neurotoxins. BTXs are clinically useful in the management of various dental and orofacial disorders involving the muscles and glands. BTX can be used as a helpful and minimally invasive treatment option to improve the quality of life of patients. As a versatile treatment option with a rapidly expanding list of uses, this toxin offers a reversible alternative to numerous aggressive procedures.
[table] Table 1: Action of botulinum toxin. Absorption via the GI tract or through tissue Reaches the lymphatic channels and the blood stream Circulates in the blood until it reaches cholinergic synapses Binds with the help of binding domain Cholinergic neuronal cell membrane at nerve terminal Enters neuron by endocytosis light chain of the toxin Crosses the membrane of the endocytic vesicle Enters cytoplasm of the pre synaptic terminal cleaves to sites on SNARE proteins Prevents assembly of synaptic fusion complex Blocks docking, fusion and acetylcholine release from 300 to 900 kD, with the pure toxin, including both light and heavy chains, typically weighing about 150 kD [/table]
[table] Table 2: Forms of botulinum toxin. [/table]
[table] Table 3: Side effects of botulinum toxin use. [/table]
[table] Table 5: Contraindications/precautions for use of botulinum toxin. [/table]
|
Effect of light-curing units in shear bond strength of metallic brackets: an in vitro study
bjectives: To determine the influence of the light curing units on the shear bond strength of orthodontic brackets. Material and Methods: Seventy-two premolars were divided into six groups (n=12): Group I: brackets bonded with Transbond and polymerization with halogen light; Group II: Transbond and LED; Group III: Fuji Ortho and halogen light; Group IV: Fuji Ortho and LED; Group V: Fuji Ortho, without acid and halogen light; Group VI: Fuji Ortho, without acid and LED. The groups were tested to shear strength in a universal testing machine at a crosshead speed of 0.5 mm/min. Data were analyzed statistically by ANOVA and Tukey's test. Results: The composite resin presented higher shear bond strength than the resin-modified glass ionomer cement (p<0.05). The halogen light and LED sources produced similar shear bond strength (p>0.05). Conclusion: The shear bond strength was influenced by the material but not by the light-curing unit. The use of LED reduced the experimental time by approximately 60%, with the same curing efficiency.
# Introduction
Dentistry has experienced a remarkable progress, starting from the technique of enamel acid etching introduced by Buonocore 6 (1955).
In the same way, the direct bonding of brackets to the teeth revolutionized Orthodontics.
Most orthodontic bonding materials use as the activation mechanism the luminous energy, like quartz-tungsten-halogen (QTH) visible light, xenon light and light-emitting diode (LED) [bib_ref] Light-emitting diodes in composite resin photopolymerization, Duke [/bib_ref] [bib_ref] Polymerization efficiency of chemically cured and visible light-cured orthodontic adhesives: degree of..., Eliades [/bib_ref]. In GII, after bonding as described in GI,
[formula] Transbond [/formula]
# Results
There was no statistically significant difference .
The ARI scores were distributed as shown in
[fig] Figure 1 -: Buccal surface positioned against a glass plate, fastened with wax 7, PVC ring positioned and the acrylic resin flowed [/fig]
[fig] Figure 2 -: Groups according to the bonding material and light-curing unit used [/fig]
[fig] Figure 4: Most specimens of Groups V and VI failed at the enamel/adhesive interface, which means that the whole adhesive layer remained on the bracket. In the specimens of the other 4 groups, great part of the adhesive remained on the enamel, with the impression of the bracket base on the remainder. When the ARI is analyzed comparing the materials, failure at the adhesive/ bracket interface (score 3) was more common in the specimens of the Groups I and II, while in the specimens of Groups III and IV there was an even distribution among scores 2 and 3, though [/fig]
[fig] Figure 3 -: Specimens (A and B) stressed in a universal testing machine at a crosshead speed of 0.5 mm/min [/fig]
[fig] CONCLUSIONS: The following conclusions may be drawn from the obtained results: 1. The light-curing units(halogen or LED) did not influence the shear bond strength of orthodontic brackets to enamel, but the orthodontic material influence bracket adhesion; 2. No acid conditioning of enamel influenced the bond strength of brackets bonded with the RMGIC (Fuji Ortho LC), resulting in values that are not acceptable for clinical conditions; 3. The use of LED reduced the experimental time by approximately 60%, with the same curing efficiency. ACKNOWLEDGMENTS To the Professors of the Discipline of Dental Materials of the Pontifical University Catholic of Rio Grande do Sul, Brazil, who made the J Appl Oral Sci. 2010;18(1):68-74 [/fig]
|
Rpn (YhgA-Like) Proteins of Escherichia coli K-12 and Their Contribution to RecA-Independent Horizontal Transfer
Bacteria use a variety of DNA-mobilizing enzymes to facilitate environmental niche adaptation via horizontal gene transfer. This has led to real-world problems, like the spread of antibiotic resistance, yet many mobilization proteins remain undefined. In the study described here, we investigated the uncharacterized family of YhgA-like transposase_31 (Pfam PF04754) proteins. Our primary focus was the genetic and biochemical properties of the five Escherichia coli K-12 members of this family, which we designate RpnA to RpnE, where Rpn represents recombinationpromoting nuclease. We employed a conjugal system developed by our lab that demanded RecA-independent recombination following transfer of chromosomal DNA. Overexpression of RpnA (YhgA), RpnB (YfcI), RpnC (YadD), and RpnD (YjiP) increased RecA-independent recombination, reduced cell viability, and induced the expression of reporter of DNA damage. For the exemplar of the family, RpnA, mutational changes in proposed catalytic residues reduced or abolished all three phenotypes in concert. In vitro, RpnA displayed magnesium-dependent, calciumstimulated DNA endonuclease activity with little, if any, sequence specificity and a preference for double-strand cleavage. We propose that Rpn/YhgA-like family nucleases can participate in gene acquisition processes. IMPORTANCE Bacteria adapt to new environments by obtaining new genes from other bacteria. Here, we characterize a set of genes that can promote the acquisition process by a novel mechanism. Genome comparisons had suggested the horizontal spread of the genes for the YhgA-like family of proteins through bacteria. Although annotated as transposase_31, no member of the family has previously been characterized experimentally. We show that four Escherichia coli K-12 paralogs contribute to a novel RecA-independent recombination mechanism in vivo. For RpnA, we demonstrate in vitro action as a magnesium-dependent, calcium-stimulated nonspecific DNA endonuclease. The cleavage products are capable of providing priming sites for DNA polymerase, which can enable DNA joining by primer-template switching.
I n prokaryotes, horizontal gene transfer (HGT; also called lateral gene transfer) is a massive force of evolutionary change and adaptation. It promotes the acquisition of new genes that allow bacteria to adapt to ecological niches and survive under stressful conditions when traditional gene regulation is not sufficient [bib_ref] Lateral genetic transfer: open issues, Ragan [/bib_ref]. To illustrate the magnitude of the issue, consider Escherichia coli. Approximately 40% of a typical E. coli genome and 90% of the E. coli species-wide pangenome consist of foreign gene islands [bib_ref] Organised genome dynamics in the Escherichia coli species results in highly diverse..., Touchon [/bib_ref] [bib_ref] Core and panmetabolism in Escherichia coli, Vieira [/bib_ref] , in the sense that they are not shared by all E. coli isolates. However, many aspects of HGT are still poorly understood, and its overall effect on genomic evolution is the subject of active research.
Both homologous and nonhomologous recombination processes contribute to pangenome assembly [bib_ref] Impact of homologous and non-homologous recombination in the genomic evolution of Escherichia..., Didelot [/bib_ref]. Homologous recombination is a universally conserved process mediated by strand transfer proteins, such as the RecA/RadA family of proteins. It acts efficiently to disseminate advantageous genetic material [bib_ref] The population genetics of commensal Escherichia coli, Tenaillon [/bib_ref] , but its dependence on sequence identity (Ͼ96 to 97%) limits the exchange to close relatives [bib_ref] Genetic barriers among bacteria, Matic [/bib_ref]. In contrast, nonhomologous recombination can come in many forms and is often less efficient but can operate across large phylogenetic distances because it does not depend on extensive DNA sequence similarity. The two processes complement each other: a nonhomologous recombination event can add a novel capability to one member of a population, and homologous exchange can then spread that capability within the population more efficiently than the original mechanism that introduced it (10).
Most nonhomologous gene addition mechanisms involve a DNA-mobilizing protein or complex that places its own gene(s) into a new location. Such action may also move cargo genes-nonmobile genetic material-along with the gene for the mobile element [bib_ref] Genomic islands: tools of bacterial horizontal gene transfer and evolution, Juhas [/bib_ref]. Common examples include transposases [bib_ref] Heteromeric transposase elements: generators of genomic islands across diverse bacteria, Peters [/bib_ref] , site-specific recombinases [bib_ref] Mechanisms of site-specific recombination, Grindley [/bib_ref] , and integrases [bib_ref] Specificity in DNA recognition by phage integrases, Campbell [/bib_ref].
The YhgA-like family (Pfam PF04754) [bib_ref] Sequence, structure and functional diversity of PD-(D/E)XK phosphodiesterase superfamily, Steczkiewicz [/bib_ref] has been proposed to represent a class of DNA-mobilizing enzymes on the basis of bioinformatic analysis: the genes involved are sporadically distributed among a wide variety of bacteria, often with multiple paralogs in each genome [bib_ref] TIGRFAMs: a protein family resource for the functional identification of proteins, Haft [/bib_ref]. PF04754 is designated a putative transposase family (trans-posase_31) by the TigrFam [bib_ref] TIGRFAMs: a protein family resource for the functional identification of proteins, Haft [/bib_ref] and Pfam [bib_ref] Pfam: the protein families database, Finn [/bib_ref] databases. Separately, members of this protein family were predicted to encode a PD-(D/E)XK phosphodiesterase domain [bib_ref] Identification of new homologs of PD-(D/E)XK nucleases by support vector machines trained..., Laganeckas [/bib_ref] [bib_ref] Realm of PD-(D/E)XK nuclease superfamily revisited: detection of novel families with modified..., Knizewski [/bib_ref]. This domain is prevalent in nucleases [bib_ref] The PD-(D/E)XK superfamily revisited: identification of new members among proteins involved in..., Kosinski [/bib_ref] but has also been found in enzymes connected to HGT [bib_ref] The restriction fold turns to the dark side: a bacterial homing endonuclease..., Zhao [/bib_ref] [bib_ref] Unexpected structural diversity in DNA recombination: the restriction endonuclease connection, Hickman [/bib_ref].
Though some YhgA-like proteins have been analyzed in silico [bib_ref] Identification of new homologs of PD-(D/E)XK nucleases by support vector machines trained..., Laganeckas [/bib_ref] [bib_ref] Realm of PD-(D/E)XK nuclease superfamily revisited: detection of novel families with modified..., Knizewski [/bib_ref] , we are the first to investigate these proteins experimentally. Our investigation centered on the five E. coli K-12 family members. These are renamed here to reflect their functional characterization presented below: RpnA (YhgA), RpnB (YfcI), RpnC, (YadD), RpnD (YjiP), and RpnE (YfaD). We show that overexpression of RpnA to RpnD increases recombination efficiency in our conjugal system, reduces cell viability in a recA-deficient background, and induces a reporter of DNA damage in a recA wild-type background, while RpnE is inactive in these assays. We then focus on the exemplar RpnA to show that the predicted PD-(D/E)XK domain is responsible for these phenotypes in vivo and that purified RpnA exhibits calcium-stimulated, magnesium-dependent DNA endonuclease activity in vitro. We also provide suggestive evidence that family member rpnC has been acquired twice at a syntenic location in enteric bacteria.
# Results
Background: the conjugal system and YhgA-like paralogs. (i) The conjugal system. The Escherichia coli mating system consists of a tetracycline-resistant donor and a streptomycin-resistant recipient, both of which are RecA deficient. In the donor, an integrated F plasmid (Hfr) lacking vegetative replication functions promotes DNA transfer (see in the supplemental material). Recombinants are selected with both drugs following the conjugal transfer of donor chromosomal DNA to the recipient. Recombinants result when a donor marker (mrr::tetRA) is added to or replaces a segment of the recipient chromosome. Under the basal condition with this system, most recombinants are found to be replacements [bib_ref] Novel recA-independent horizontal gene transfer in Escherichia coli K-12, Kingston [/bib_ref]. These recombination events occur at a low frequency (ϳ10 Ϫ10 /recipient/h), and the exchange can range from just a portion of the genome island originally interrogated (Ͻ16 kb) to over half of the genome (Ͼ2.4 Mb); a majority of events have replaced over 400 kb [bib_ref] Novel recA-independent horizontal gene transfer in Escherichia coli K-12, Kingston [/bib_ref].
(ii) YhgA-like proteins. The first experimental evidence linking YhgA-like proteins (PF04754) to HGT was discovered while testing this conjugal system. The PF04754 member rpnD (formerly yjiP) was investigated due to its close proximity to the genome island under study [bib_ref] Cassette-like variation of restriction enzyme genes in Escherichia coli C and relatives, Sibley [/bib_ref]. Though rpnD did not appear to be a part of the island, its expression promoted the RecA-independent recombination encompassing it [bib_ref] Novel recA-independent horizontal gene transfer in Escherichia coli K-12, Kingston [/bib_ref]. The E. coli K-12 genome encodes 4 additional YhgA-like proteins (RpnA, RpnB, RpnC, and RpnE) with a conserved 5= domain .
## Yhga-like protein expression in vivo promotes reca-independent recombination and damages the genome. (i) increased reca-independent recombination.
RecA-independent recombinants were recovered at higher rates upon overexpression of RpnA to RpnD but not RpnE. We fused each of the E. coli K-12 YhgA-like proteinencoding genes to the rhamnose-inducible promoter rhaBp and integrated these constructs into the Tn7 attachment site of ΔrecA recipient strains [bib_ref] Novel recA-independent horizontal gene transfer in Escherichia coli K-12, Kingston [/bib_ref] [bib_ref] A versatile element for gene addition in bacterial chromosomes, Sibley [/bib_ref]. We mated these recipients to the ΔrecA donor in media with and without 0.2% rhamnose and recorded the recombination efficiency as the proportion of recipients that acquired tetracycline resistance from the donor. Rhamnose had no significant effect on recombination efficiency in a control mating or when it induced the expression of rpnE. However, rhamnose-induced expression of rpnC, rpnD, rpnB, and rpnA significantly increased the recombination efficiency by 2.9-, 4.7-, 19-, and 49-fold, respectively ( (ii) Toxicity in ⌬recA cells. RpnA to RpnD were also toxic to the ΔrecA recipients. After each mating, the number of CFU of the recipient per milliliter was counted to assess cell viability. The number of CFU per milliliter of the rhaBp-rpnE recipient resembled that of the control, but rhamnose reduced the number of CFU per milliliter of the rhaBp-rpnC, rhaBp-rpnD, rhaBp-rpnB, and rhaBp-rpnA recipients by 59%, 94%, 98.7%, and 98.0%, respectively. This cell toxicity could be related to the effect of YhgA-like protein expression on recombination efficiency, but the two factors were not completely correlated. For example, RpnA was the strongest activator of RecAindependent recombination but had the second weakest cell toxicity.
(iii) Induced DNA damage (SOS) response. YhgA-like proteins could be toxic to ΔrecA cells due to DNA damage: the PD-(D/E)XK domain found in YhgA-like proteins is best known as the active site in restriction endonucleases [bib_ref] The PD-(D/E)XK superfamily revisited: identification of new members among proteins involved in..., Kosinski [/bib_ref] , and RecA-deficient cells are hypersensitive to DNA damage.
Consistent with this hypothesis, all the active E. coli K-12 YhgA-like proteins induced a reporter of the SOS response when expressed in a Rec ϩ host. Plasmids with rhamnose-inducible constructs were introduced into an SOS indicator strain: lacZ was fused to the DNA damage-inducible dinD locus on the E. coli genome [dinDp-lacZ(Ts)] [bib_ref] Isolation of temperature-sensitive McrA and McrB mutations and complementation analysis of the..., Piekarowicz [/bib_ref]. All these strains except for the strain carrying rhaBp-rpnE gave a blue color on X-Gal (5-bromo-4-chloro-3-indolyl--D-galactopyranoside) plates when they were induced with rhamnose (data not shown).
Quantitative -galactosidase assays showed that the intensity of the SOS response induced by each YhgA-like protein (in Rec-positive [Rec ϩ ] strains) was proportionate to the cell-killing activity of the protein (in Rec-negative [Rec Ϫ ] strains). To understand the induction kinetics and the dynamic range of this indirect assay, a rhamnose-inducible lacZ gene in the Rec ϩ strain was analyzed in the same way. The most lethal paralogs, RpnB and RpnD, were the first to activate the SOS-responsive dinDp-lacZ(Ts) reporter, at a time when the rhaBp promoter had just begun to activate lacZ . RpnA and RpnC were substantially weaker in this regard, in that they activated the dinDp-lacZ(Ts) reporter only after the lacZ activity driven directly from the rhamnose promoter reached a maximum. Expression of rpnE for 24 h did not affect -galactosidase activity, consistent with its lack of activity in other assays.
(iv) Interpreting initial results. Taken together, these findings establish that YhgA-like proteins participate in a RecA-independent recombination mechanism. The most direct evidence of such a mechanism is that YhgA-like protein expression increased productive recombination, while the cell toxicity and SOS induction phenotypes suggest that these proteins act on the recipient genome.
Comparison of the YhgA-like family to known DNA damage processes using the conjugal system. (i) Comparison to the chain-terminating nucleotide AZT. Azidothymidine (AZT) is a thymidine analogue that acts as a chain terminator [bib_ref] Antibacterial activity and mechanism of action of 3=-azido-3=-deoxythymidine (BW A509U), Elwell [/bib_ref]. It is known to promote template switching during replication [bib_ref] Azidothymidine and other chain terminators are mutagenic for template-switch-generated genetic mutations, Seier [/bib_ref]. We earlier suggested the involvement of template switching to explain recombination in our system [bib_ref] Novel recA-independent horizontal gene transfer in Escherichia coli K-12, Kingston [/bib_ref]. Addition of AZT to a wild-type (WT) mating (ER3435 ϫ ER3473) significantly increased the recombination efficiency: at a concentration high enough to reduce cell viability by 4.8-fold, the recombination frequency increased by 31-fold (P ϭ 0.01) [fig_ref] FIG 4: Site-specific endonucleases and AZT have disparate effects in this system [/fig_ref].
(ii) Comparison to nucleases. Nuclease action alone is not sufficient to increase recombination efficiency in this system. We overexpressed the native endonuclease McrA, the foreign endonuclease AsiSI, and the foreign nickase BsrDIB in ΔrecA recipients. McrA served as a negative control since the donor and recipient genomes are resistant to this endonuclease [bib_ref] Transposon-mediated linker insertion scanning mutagenesis of the Escherichia coli McrA endonuclease, Anton [/bib_ref]. As expected, it had no effect on recombination efficiency or cell toxicity when expressed [fig_ref] FIG 4: Site-specific endonucleases and AZT have disparate effects in this system [/fig_ref]. Expression of AsiSI or BsrDIB reduced cell viability (by 97% and 88%, respectively), but recombination was not stimulated in either case.
Differentiating basal recombination from RpnA-promoted recombination. Our next goal was to determine whether YhgA-like proteins were increasing basal recombination events (those that already occur in the conjugal system) or were promoting a different pathway. Prior work had shown only that basal recombination events are infrequent and that recombinants carried replacements of large segments of genomic DNA [bib_ref] Novel recA-independent horizontal gene transfer in Escherichia coli K-12, Kingston [/bib_ref].
Here, we found a larger proportion of shorter (but still large) replacement segments when RpnA participates than when it does not. To estimate the size of replacement segments, we monitored markers located at various distances from the selected donor marker (mrr::tetRA) [fig_ref] FIG 5: Distribution of genomic exchanges in recombinants [/fig_ref]. The npt and cat resistance cassettes were within 16 kb proximal to the mrr locus in the recipient genome, and the fhuA::IS2 and lacZ-positive (lacZ ϩ ) markers were 223 kb and 418 kb distal to the mrr::tetRA locus in the donor genome, respectively. We analyzed 96 and 170 recombinants produced under basal (WT) conditions and during RpnA (RpnA-promoted) overexpression, respectively, representing at least 9 independent matings in each case. The frequency of additions (those recombinants acquiring tetRA and also retaining mrr in a PCR screen) was under 4% for both conditions [fig_ref] FIG 5: Distribution of genomic exchanges in recombinants [/fig_ref] and C). Seventy percent of the WT recombinants received all the donor markers and therefore exchanged at least 434 kb of genomic DNA; only 11% received just the tetRA cassette (maximum exchange of 236 kb, less than 13 kb of which could be proximal to tetRA). In contrast, 35% of the RpnA-promoted recombinants received all the donor markers and 53% received just the tetRA cassette. The recombinant distribution with AZT treatment resembled that found with RpnA overproduction but was more pronounced; 74% of the recombinants acquired only the tetRA cassette.
Consistent with the changed recombinant distribution, deletion of the genes encoding the four active E. coli K-12 YhgA-like proteins in the recipient strain (producing ΔrpnA, ΔrpnB, ΔrpnC, and ΔrpnD strains) did not affect the frequency of recombinant formation. These paralogs thus do not contribute to the basal recombination frequency or distribution. We also tested the potential role of the Rac prophage-encoded RecET recombinase. This is normally silent [bib_ref] Homologous pairing proteins encoded by the Escherichia coli recE and recT genes, Kolodner [/bib_ref] [bib_ref] Physical characterisation of the "Rac prophage" in E. coli K12, Kaiser [/bib_ref] , but rare unscheduled expression can be imagined. Deletion of the entire sequence for the prophage did not affect the frequency recombinant formation.
Three segments of YhgA-like proteins determine in vivo activity. YhgA-like proteins can be roughly divided into three segments: a predicted PD-(D/E)XK structural core [bib_ref] Identification of new homologs of PD-(D/E)XK nucleases by support vector machines trained..., Laganeckas [/bib_ref] [bib_ref] Realm of PD-(D/E)XK nuclease superfamily revisited: detection of novel families with modified..., Knizewski [/bib_ref] which comprises the N-terminal portion of the conserved longer trans-posase_31 domain; the remaining portion of the transposase_31 domain, which is also highly conserved in the family; and a variable C-terminal tail . We investigated the function of each segment by expressing a series of YhgA-like protein variants in the mating system.
(i) The PD-(D/E)XK structural core. The role PD-(D/E)XK structural core was probed by the use of mutations in predicted active-site residues. We mutated each predicted signature residue with alanine substitutions; in one case, we also changed a glutamine residue characteristic of this family but unusual in other PD-(D/E)XK enzymes [bib_ref] Realm of PD-(D/E)XK nuclease superfamily revisited: detection of novel families with modified..., Knizewski [/bib_ref] to lysine. We expected this mutant series [fig_ref] FIG 6: Phenotypes of mutated RpnA proteins in vivo [/fig_ref] to exhibit low or diminished activity. The mutants were placed in the same expression environment as the parent and tested for the three phenotypes: increased recombination efficiency, cell toxicity, and SOS induction. The isolates carrying RpnA with the D11A mutation (RpnA-D11A), RpnA-D63A, RpnA-E82A, and RpnA-Q84K lost all three phenotypes [fig_ref] FIG 6: Phenotypes of mutated RpnA proteins in vivo [/fig_ref] and C). RpnA-Q84A and -R94A did not promote recombination but still reduced cell viability (by 98.1% and 94.6%, respectively, which made them 8.1-and 3.0-fold more toxic than WT RpnA, respectively). Both mutants elicited an SOS response on the plates as well (not shown). but are still highly conserved among YhgA-like proteins [bib_ref] Pfam: the protein families database, Finn [/bib_ref]. In particular, a PDDEI motif within this segment is identical in all E. coli K-12 YhgA-like proteins [bib_ref] Pfam: the protein families database, Finn [/bib_ref]. Mutating this acidic cluster to create RpnA-D165A unexpectedly yielded a hyperactive RpnA variant: it was 2.2-fold more effective than the WT in promoting recombination and 42-fold more toxic than the WT to cells [fig_ref] FIG 6: Phenotypes of mutated RpnA proteins in vivo [/fig_ref]. This high-activity mutation confirms the functional relevance of the non-PD-(D/E)XK portion of the transposase_31 domain.
(iii) The variable C-terminal tail. An RpnD variant lacking the last 45 residues of its variable C-terminal tail was substantially less active than WT RpnD but not completely dead. Expression of rhaBp-rpnD with a deletion of nucleotides 786 to 921 increased recombination (2.3-fold) and reduced cell viability (1.6-fold) compared with the results for the control strain . These effects were statistically significant but lower than the effects of WT RpnD, which increased recombination 2.0-fold more (4.5-fold total) and was 7.8 times more toxic to cell viability.
Migration history of YhgA-like proteins suggests independent rpnC (yadD) insertion events at the same place. We undertook a limited reinvestigation of the distribution of rpnC (formerly yadD) to find support for its self-mobility. Our work flow is fully described in Text S1, and key findings are summarized here.
The distribution of rpnC orthologs is consistent with two separate introductions into the intergenic region between panC and panD (Text S1). We identified by BLAST analysis panCD DNA segments in Enterobacteriaceae with similarity to the sequence of the K-12 genome at the 5= end of panC and the 3= end of panD. A collection of 32 sequences, 18 of which had similarity to rpnC, were compared using nucleotide sequence alignment and phylogenetic tree construction [fig_ref] FIG 4: Site-specific endonucleases and AZT have disparate effects in this system [/fig_ref] ; . Two clusters were more divergent from each other than the flanking panC and panD homologs were. They could be visually distinguished by aligning the panC-panD intergenic regions and creating an evolutionary tree. Comparison of the phylogeny of these intergenic regions [fig_ref] FIG 4: Site-specific endonucleases and AZT have disparate effects in this system [/fig_ref] with that of the flanking core genes [fig_ref] FIG 4: Site-specific endonucleases and AZT have disparate effects in this system [/fig_ref] strongly suggested that two distinct versions of rpnC (yadD) were introduced separately into ancestral pan operons. Alternative explanations are described in Text S1.
RpnA exhibits nonspecific DNA endonuclease activity in vitro. (i) Purification of RpnA. RpnA and two RpnA variants were purified to characterize the activity of Recombinants were screened for the cat, npt, mrr, fhuA, and lacZ markers. Recombinants containing both tetRA and mrr were classified as having genome additions, and the results for these recombinants are not shown. Horizontal bars indicate the extent of donor DNA (orange) that we inferred replaced the recipient genome (blue) during the recombination event. NA, not available. (C) Proportion of recombinants in each class from a basal mating (WT; ER3435 ϫ ER3473) with or without AZT treatment or from a mating in which rpnA was overexpressed (RpnA; ER3435 ϫ ER3514). For the WT, most recombinants were created by large replacements of over 400 kb of genomic DNA. In the RpnA and AZT matings, large replacements were less frequent, and over half of the genomic replacements were within the 236-kb segment between the npt and fhuA markers flanking the selected marker, tetRA.
YhgA-like proteins in vitro. Purification by affinity and anion-exchange chromatography yielded chromatographically pure RpnA, RpnA-D63A (which is inactive in vivo), and RpnA-D165A (which is hyperactive in vivo).
(ii) RpnA has low DNA endonuclease activity. Purified RpnA exhibited low but detectable DNA endonuclease activity in vitro. Supercoiled pUC19 was initially used as the DNA substrate because a single nick relaxes the plasmid, a single cleavage linearizes it, and all three species (supercoiled, relaxed, and linear) can be easily distinguished on an agarose gel [fig_ref] FIG 7: In vitro analysis of RpnA endonuclease activity [/fig_ref]. RpnA initially digests pUC19 to the linear species with a small increase in the amount of nicked plasmid. Time course assays showed that less than 20% of the supercoiled plasmid is nicked at any one time [fig_ref] FIG 7: In vitro analysis of RpnA endonuclease activity [/fig_ref] , and the entire plasmid is eventually digested to a smear of DNA. In contrast, the well-characterized and highly active nonspecific nicking enzyme DNase I (32) converted 40% of the total DNA to the nicked product before substantial linearization occurred and digested DNA 2.8 ϫ 10 7 -fold faster than RpnA [fig_ref] FIG 6: Phenotypes of mutated RpnA proteins in vivo [/fig_ref].
(iii) Mutated RpnA proteins display activity consistent with their in vivo phenotypes. With pUC19 and the linear bacteriophage DNA substrate, RpnA-D63A had [bib_ref] Realm of PD-(D/E)XK nuclease superfamily revisited: detection of novel families with modified..., Knizewski [/bib_ref]. RpnA variants with mutations in these active-site residues were created; an additional mutation in which a conserved PDDE motif was converted to PDAE (D165A) was made. (B) Recombination efficiency of matings between the ΔrecA donor (ER3435) and either the control ΔrecA recipient (ER3473) or ΔrecA recipients with rhamnose-inducible rpnA or its mutants (WT, ER3514; D11A, ER3552; D63A, ER3553; E82A, ER3554; Q84A, ER3556; Q84K, ER3555; R94A, ER3557; and D165A, ER3558) with and without induction. (C) Toxicity of induction during these matings, reflected in viability decline. (Continued on next page) Rpn Proteins of E. coli K-12
Journal of Bacteriology negligible activity (Ͻ0.006-fold the WT activity) and RpnA-D165A digested the substrates ϳ2-fold faster than the WT enzyme [fig_ref] FIG 6: Phenotypes of mutated RpnA proteins in vivo [/fig_ref] to D). Since the activities of these variants mirror the effects of their overexpression in vivo, we conclude that the purification strategy successfully removed potentially contaminating nucleases and that these assays reflect the RpnA endonuclease activity.
(iv) Digestion patterns do not suggest sequence specificity. No evidence of sequence specificity was detected with the substrates used, in that no banding pattern was detected [fig_ref] FIG 6: Phenotypes of mutated RpnA proteins in vivo [/fig_ref]. In long digests, RpnA degraded DNA substrates to small fragments: bacteriophage DNA was reduced to a smear of between 100 and 500 bp on an agarose gel [fig_ref] FIG 7: In vitro analysis of RpnA endonuclease activity [/fig_ref] , which disappeared when four times as much RpnA was used. We did not observe a band of recalcitrant DNA that would constitute a sequence resistant to RpnA digestion in these extended digests. Similar results were obtained when RpnA was incubated with E. coli K-12 genomic DNA or a 2-log DNA ladder [fig_ref] FIG 7: In vitro analysis of RpnA endonuclease activity [/fig_ref]. The disappearance of the upper bands of the DNA ladder first [fig_ref] FIG 7: In vitro analysis of RpnA endonuclease activity [/fig_ref] is consistent with primary endonuclease action; exonuclease action would affect small bands first.
(v) DNA conformation does not affect RpnA activity; dsRNA is not cleaved. Substrates sensitive to RpnA included covalently closed circular pUC19, linear molecular weight ladders, full-length bacteriophage DNA, and single-stranded DNA (ssDNA) of the M13 virion. When ssDNA and double-stranded DNA (dsDNA) substrates were digested together or separately, the two substrates were digested at comparable rates [fig_ref] FIG 7: In vitro analysis of RpnA endonuclease activity [/fig_ref]. RpnA did not exhibit a preference for nicked pUC19 or for the cruciform extrusion in pUC(AT) (data not shown). No digestion was observed with a double-stranded RNA (dsRNA) ladder after 18 h [fig_ref] FIG 7: In vitro analysis of RpnA endonuclease activity [/fig_ref]. In our hands, single-stranded RNA was too unstable to be assessed in this way (not shown).
(vi) RpnA is a broadly active Mg 2؉ -dependent enzyme with unusual stimulation by Ca 2؉ . Buffer optimization yielded the RpnA buffer described in Materials and Methods. The full process by which we developed this buffer is described in , and key findings derived from this process are highlighted below.
As is true with most nucleases (33), Mg 2ϩ was required: removal of MgCl 2 by omission or by adding EDTA inactivated the enzyme [fig_ref] FIG 7: In vitro analysis of RpnA endonuclease activity [/fig_ref]. Mn 2ϩ supported some action of RpnA in the absence of Mg 2ϩ but was inhibitory in its presence, as was Zn 2ϩ . RpnA activity was slightly higher in the presence of dithiothreitol (DTT) and -mercaptoethanol, as might be predicted from the cysteine content (5 residues; [fig_ref] FIG 7: In vitro analysis of RpnA endonuclease activity [/fig_ref].
Surprisingly, Ca 2ϩ stimulated activity 6-fold [fig_ref] FIG 7: In vitro analysis of RpnA endonuclease activity [/fig_ref]. This was unusual because Ca 2ϩ usually inhibits PD-(D/E)XK nucleases by competing with Mg 2ϩ for the active site [bib_ref] Metal activation of enzymes in nucleic acid biochemistry, Cowan [/bib_ref]. With bacteriophage DNA as the substrate, RpnA activity was the highest when the Ca 2ϩ concentration] was 1 to 2 times the Mg 2ϩ concentration, but at higher ratios calcium was inhibitory .
RpnA tolerates a wide range of pH and salt. Using bacteriophage DNA as a substrate, we found activity between pH 7.5 and pH 10.5, with maximal activity occurring at a pH of 9.0 . NaCl concentrations ranging from 0 to 200 mM were acceptable, with a broad plateau taking place between 10 mM and 75 mM .
(vii) RpnA cleavage results in 3= hydroxyl DNA ends. Fluorescent end-labeling experiments revealed that RpnA cleavage produces fragments extendable by DNA polymerase. pUC19 was digested to a smear of DNA ranging from 100 bp to 1 kb with either RpnA, DNase I, or micrococcal nuclease. Cleavage products were then incubated with the Klenow fragment and fluorescein-labeled deoxynucleoside triphosphates (dNTPs). The positive-control DNase I-digested smear had a strong fluorescent signal, confirming that the polymerase could effectively label 3= hydroxyl ends, while the 3= phosphate ends of the negative-control micrococcal nuclease-digested DNA [bib_ref] The purification and properties of micrococcal nuclease, Alexander [/bib_ref] were not labeled at all [fig_ref] FIG 7: In vitro analysis of RpnA endonuclease activity [/fig_ref]. The RpnA-digested DNA exhibited a fluorescent signal, confirming the presence of 3= hydroxyl ends.
# Discussion
Overproduction phenotypes implicate YhgA-like proteins in DNA transactions. When they were artificially overproduced, four of the five endogenous E. coli K-12 paralogs in the YhgA-like transposase_31 family promoted RecA-independent recombinationand produced DNA damage in vivo in RecA ϩ cells , likely accounting for their toxicity in RecA Ϫ cells. These proteins were not responsible for the basal level of RecA-independent recombination, since the multipledeletion strain lacking the four active paralogs showed the same level of recombinant formation as the wild type (seein the supplemental material). The prophageencoded recET recombination system was also not required for basal recombinant formation.
Recombinants produced under both basal and RpnA-stimulated conditions carried large segmental replacements of recipient genes with donor genes, rather than additions of donor DNA to the recipient genome. Those formed with overproduced RpnA were smaller, on average, than those formed in its absence [fig_ref] FIG 5: Distribution of genomic exchanges in recombinants [/fig_ref]. This distribution agrees with that observed with the overproduced paralog RpnD [bib_ref] Novel recA-independent horizontal gene transfer in Escherichia coli K-12, Kingston [/bib_ref]. The change in distribution strongly suggests that these proteins play a role in generating both proximal and distal crossover events, when present.
In vitro analysis of RpnA reveals a novel nuclease activity. The transposase annotation of YhgA-like proteins implies a transesterase activity able to reconnect the phosphodiester bonds in donor and recipient strands, sometimes via a protein-DNA covalent intermediate [bib_ref] Heteromeric transposase elements: generators of genomic islands across diverse bacteria, Peters [/bib_ref] [bib_ref] Mechanisms of site-specific recombination, Grindley [/bib_ref] [bib_ref] Specificity in DNA recognition by phage integrases, Campbell [/bib_ref]. However, nuclease activity was also expected from bioinformatic assignment of this family to the PD-(D/E)XK nuclease clan . This clan includes structure-specific nucleases, exonucleases, and auxiliary transposon components as well as large numbers of restriction endonucleases. One homing endonuclease belongs to this clan, although most such mobile, highly specific nucleases belong to other nuclease clans [bib_ref] Specificity in DNA recognition by phage integrases, Campbell [/bib_ref] [bib_ref] Structural, functional and evolutionary relationships between homing endonucleases and proteins from their..., Taylor [/bib_ref].
RpnA exhibited an in vitro nuclease activity plausibly related to the biological properties described above [fig_ref] FIG 7: In vitro analysis of RpnA endonuclease activity [/fig_ref]. A mutation abrogating the recombination, toxicity, and DNA damage phenotypes [fig_ref] FIG 6: Phenotypes of mutated RpnA proteins in vivo [/fig_ref] also destroyed DNA degradation in vitro, while a different mutation that enhances those phenotypes also enhanced DNA degradation [fig_ref] FIG 6: Phenotypes of mutated RpnA proteins in vivo [/fig_ref] to D). RpnA had low nuclease activity compared to that of the digestive enzyme DNase I and did not show evidence of sequence specificity in the degradation process. Both dsDNA and ssDNA were substrates [fig_ref] FIG 7: In vitro analysis of RpnA endonuclease activity [/fig_ref]. The primary action could be either nicking or double-strand cleavage: some members of the PD-(D/E)XK clan nick two strands sequentially [bib_ref] DNA nicking by HinP1I endonuclease: bending, base flipping and minor groove expansion, Horton [/bib_ref] [bib_ref] One recognition sequence, seven restriction enzymes, five reaction mechanisms, Gowers [/bib_ref] , others cut two strands in concert [bib_ref] One recognition sequence, seven restriction enzymes, five reaction mechanisms, Gowers [/bib_ref] , and others nick in collaboration with partner proteins [bib_ref] Unexpected structural diversity in DNA recombination: the restriction endonuclease connection, Hickman [/bib_ref] [bib_ref] Cleavage of individual DNA strands by the different subunits of the heterodimeric..., Bellamy [/bib_ref] or act at structural features [bib_ref] A conserved nuclease domain in the archaeal Holliday junction resolving enzyme Hjc, Kvaratskhelia [/bib_ref] [bib_ref] Survey and summary: Holliday junction resolvases and related nucleases: identification of new..., Aravind [/bib_ref].
The fact that RpnA action provided a 3= hydroxyl [fig_ref] FIG 7: In vitro analysis of RpnA endonuclease activity [/fig_ref] was expected from the general properties of the PD-(D/E)XK family of enzymes [bib_ref] Type II restriction endonucleases-a historical perspective and more, Pingoud [/bib_ref] [bib_ref] Grouping together highly diverged PD-(D/E)XK nucleases and identification of novel superfamily members..., Bujnicki [/bib_ref] [bib_ref] Topology of type II REases revisited; structural classes and the common conserved..., Niv [/bib_ref] and provides further support for the assignment of PF04754 to this clan. It is also consistent with recombination models that rely on polymerase template switching at the cleavage site, since the 3= end provides a polymerase-priming capacity without further processing.
RpnA is novel in its stimulation by Ca 2ϩ [fig_ref] FIG 7: In vitro analysis of RpnA endonuclease activity [/fig_ref]. We could find no report of a PD-(D/E)XK nuclease stimulated by Ca 2ϩ or any other intracellular Ca 2ϩ -stimulated bacterial nuclease at all. This result strongly implies two metal-binding sites in the protein, which has plenty of precedent in general, but reports are in conflict for PD-(D/E)XK nucleases [bib_ref] Structural, functional and evolutionary relationships between homing endonucleases and proteins from their..., Taylor [/bib_ref] [bib_ref] Type II restriction endonucleases-a historical perspective and more, Pingoud [/bib_ref].
Though several types of Ca 2ϩ -stimulated/dependent nucleases have been discovered in eukaryotes [bib_ref] Ca 2ϩ -dependent activity of human DNase I and its hyperactive variants, Pan [/bib_ref] [bib_ref] Ca 2ϩ /Mg 2ϩ -dependent endonuclease from human spleen: purification, properties, and..., Ribeiro [/bib_ref] , prokaryotic examples are rare. PD-(D/E)XK enzymes are usually inhibited by Ca 2ϩ because it competes with an essential Mg 2ϩ ion at the active site [bib_ref] Metal activation of enzymes in nucleic acid biochemistry, Cowan [/bib_ref] , as it indeed did for RpnA when it was present in sufficient excess . The only Ca 2ϩ -dependent bacterial nuclease family that we found in the literature is the secreted micrococcal nuclease family (also known as Staphylococcus nuclease [SNase; PF00565]) [bib_ref] Pfam: the protein families database, Finn [/bib_ref] or related enzymes with an SNase-like domain (IPR016071) [bib_ref] The InterPro protein families database: the classification resource after 15 years, Mitchell [/bib_ref]. These are quite distinct from RpnA: they require Ca 2ϩ for activity, while RpnA is only stimulated by it; the SNase fold responsible for nuclease activity is structurally distinct from the PD-(D/E)XK domain; and most importantly, the products of SNase cleavage carry 3= phosphate ends [bib_ref] The purification and properties of micrococcal nuclease, Alexander [/bib_ref] , while RpnA cleavage yielded 3= hydroxyl ends that allowed DNA polymerase action [fig_ref] FIG 7: In vitro analysis of RpnA endonuclease activity [/fig_ref].
A demonstration of calcium regulation in vivo is lacking at present, yet the effect of calcium on RpnA could be relevant to its role in recombination. Most studies estimate the intracellular calcium concentration in prokaryotes to be from 200 to 300 nM [bib_ref] Direct measurement of free Ca(2ϩ) shows different regulation of Ca(2ϩ) between the..., Jones [/bib_ref] [bib_ref] An assessment of the role of intracellular free Ca 2ϩ in E...., Holland [/bib_ref] , which is much lower than both the millimolar concentrations of Ca 2ϩ required to activate RpnA in vitro and the concentration of cytosolic Mg 2ϩ [bib_ref] Metal activation of enzymes in nucleic acid biochemistry, Cowan [/bib_ref]. However, studies have shown that E. coli can raise total Ca 2ϩ levels [bib_ref] Cytosolic Ca 2ϩ regulates protein expression in E. coli through release from..., Naseem [/bib_ref] and may be able to direct Ca 2ϩ to specific regions of the cell [bib_ref] Direct measurement of free Ca(2ϩ) shows different regulation of Ca(2ϩ) between the..., Jones [/bib_ref]. It is therefore possible that E. coli directs Ca 2ϩ to RpnA to promote DNA transactions.
Interpreting recombinant formation in the conjugal system. The recombination events observed here may require collaboration with other endogenous DNA transaction proteins. The recombination transaction is not set in train simply by introducing DNA cleavage: damaging the recipient genome with a restriction endonuclease or a nicking enzyme did not increase recombinant formation [fig_ref] FIG 4: Site-specific endonucleases and AZT have disparate effects in this system [/fig_ref]. Similarly, the degree of damage reflected in toxicity did not directly correlate with effective recombination, whether comparing paralogsor comparing mutated enzymes (RpnA-Q84A and -R94A) [fig_ref] FIG 6: Phenotypes of mutated RpnA proteins in vivo [/fig_ref].
YhgA-like proteins could nevertheless be causing DNA damage that provokes further processing. This damage could simply be the continued association of the nuclease with its site of action, resulting in a polymerase roadblock. Candidates for repair processing include primase (priA), involved in replisome assembly at stalled forks [bib_ref] Recruitment to stalled replication forks of the PriA DNA helicase and replisome-loading..., Gabbai [/bib_ref] ; DnaK, required for RecA-independent replication fork repair (50); or YoaA, a helicase thought to assist with the removal of blocked termini by displacement of the primer terminus [bib_ref] Connecting replication and repair: YoaA, a helicase-related protein, promotes azidothymidine tolerance through..., Brown [/bib_ref]. Single-strand annealing processes to promote primer-template switching (as observed in Salmonella [bib_ref] Recombination and annealing pathways compete for substrates in making rrn duplications in..., Reams [/bib_ref] and inducible RecA-independent repair mechanisms (such as the RpoS-mediated response [bib_ref] A DNA damage response in Escherichia coli involving the alternative sigma factor, Merrikh [/bib_ref] could also participate in recombinant formation.
AZT-stimulated recombination: support for a template-switch model of RpnAstimulated recombination. One model for recombination involves polymerase template switching, in which a 3= end is freed from one homolog, anneals to the other homolog, and is extended by DNA polymerase. At the proximal crossover, such an event would connect the donor's antibiotic resistance cassette to the recipient replication origin. A similar event distal to the donor antibiotic resistance marker would complete the substitution. RpnA could be contributing to these switching events through genomic disruptions that lead to polymerase dissociation or by the creation of free 3= ends that dissociated polymerase can act on.
The effect of azidothymidine (AZT) on the conjugal system is compatible with this model. AZT is a chain-terminating thymidine analog that causes DNA polymerase to dissociate from the genome and reveals ssDNA gaps to which the polymerase primer can anneal [bib_ref] Toxicity and tolerance mechanisms for azidothymidine, a replication gap-promoting agent, in Escherichia..., Cooper [/bib_ref]. This activity increases the frequency of template-switch-generated mutations in E. coli, which are conceptually similar to the template-switching events proposed by our model [bib_ref] Azidothymidine and other chain terminators are mutagenic for template-switch-generated genetic mutations, Seier [/bib_ref]. Switching provoked by AZT would therefore be expected to increase the frequency of RecA-independent recombination and, if it is frequent enough, could reduce the overall size of genomic exchanges. Both AZT treatment and RpnA expression increased the recombination frequencyand reduced the size of genomic exchange [fig_ref] FIG 5: Distribution of genomic exchanges in recombinants [/fig_ref]. In this respect, the model is supported.
Searching for a true target site: site specificity of gene location for an rpnA paralog. Although RpnA degrades DNA with low activity and little to no specificity with the limited sequence universe tested in vitro, YhgA-like proteins might target specific DNA sites in vivo, as do homing endonucleases [bib_ref] Structural, functional and evolutionary relationships between homing endonucleases and proteins from their..., Taylor [/bib_ref]. Unlike insertion sequence (IS) elements, the genes for members of the RpnA family do not often move within laboratory lineages, making it problematic to identify sites of action. However, genome mining can suggest possibilities. Bioinformatic analysis of the rpnC (yadD) gene distribution suggests two independent insertions into the same genomic locus among the Enterobacteriaceae [fig_ref] FIG 4: Site-specific endonucleases and AZT have disparate effects in this system [/fig_ref]. Acquisition could involve either the autonomous action of the enzyme to move its own gene or an action to import a copy from a distant relative by stimulating localized recombination or mutagenesis. The variable C-terminal tail of YhgA-like proteins might determine different DNA sequence preferences, resulting in different insertion positions for the different paralogs. Variable domains often function in sequence recognition [bib_ref] Movement of DNA sequence recognition domains between non-orthologous proteins, Furuta [/bib_ref] [bib_ref] Segmentally variable genes: a new perspective on adaptation, Zheng [/bib_ref]. Our own experimentation has shown for RpnD that truncation of the tail reduces but does not eliminate the recombination efficiency and cell toxicity in vivo .
Assignment of function to the uncharacterized PF04754 protein family. The results of our genetic and biochemical studies support the assignment of a nuclease function to this family, broadly validating the predicted motif. The low activity and lack of sequence specificity among the tested paralogs are consistent with the possibility that these enzymes are eroded versions of ancient mobile elements. Limited reinvestigation of the phylogeny of rpnC (yadD) in its gene neighborhood is compatible with independent insertion at the same position on two occasions, leaving open the possible sequence specificity of the primary action, followed by the loss of activity due to the absence of selection.
The properties reported here could then represent secondary phenotypes, similar to the off-target effects of a homing endonuclease [bib_ref] Structural, functional and evolutionary relationships between homing endonucleases and proteins from their..., Taylor [/bib_ref] [bib_ref] Frequent endonuclease cleavage at off-target locations in vivo, Petek [/bib_ref] or transposase (e.g., see reference 58). Given the ubiquity of the family, such off-target effects may be relevant to genome island assembly, generating substrates for microhomology-mediated sequence assembly.
Conclusion: the YhgA-like family represents a novel class of DNA-active proteins. The experimental data presented in this paper confirm the biological relevance of the YhgA-like family to gene mobility. The fact that YhgA-like protein expression increases RecA-independent recombination is the central piece of evidence supporting this argument because it directly shows these proteins contributing to HGT events. The results of the cell toxicity, SOS induction, and in vitro nuclease activity experiments further confirm that YhgA-like proteins interact with DNA and suggest an HGT mechanism that involves DNA cleavage to create polymerase-extendable ends. A distinguishing property is that RpnA-promoted recombinants tend to exhibit exchanges of DNA shorter than those that occur with naturally occurring recA-independent recombination in the system.
# Materials and methods
Strains, plasmids, and growth conditions. All strains, plasmids, and oligonucleotides used in this study are listed in in the supplemental material. Bacteria were routinely grown in liquid Luria broth (LB; 10 g/liter tryptone, 5 g/liter yeast extract, 10 g/liter NaCl, 1 g/liter dextrose, 1 g/liter MgCl 2 ·6H 2 O) or rich broth (RB; 10 g/liter tryptone, 5 g/liter yeast extract, 5 g/liter NaCl, pH 7.2) medium at 37°C with vigorous shaking or on solid LB or RB medium containing 1.5% agar with appropriate selection. Ampicillin (Ap; 100 g/ml), streptomycin (Sm; 100 g/ml), kanamycin (Kn; 40 g/ml), chloramphenicol (Cm; 30 g/ml), and tetracycline (Tc; 20 g/ml) were used for selections and screens. Where appropriate, 40 g/l X-Gal (5-bromo-4-chloro-3-indolyl--D-galactopyranoside) was added to score the lac phenotype. Plasmids were prepared from E. coli Turbo cells (catalog number C2984; New England BioLabs [NEB]). For temperature-sensitive plasmids, incubation at 30°C was used to permit plasmid replication and incubation at 42°C was used to remove the plasmid. Integrated pDEL-R6K vectors [bib_ref] Leveraging modern DNA assembly techniques for rapid, markerless genome modification, Tikh [/bib_ref] were excised by plating on LB supplemented with 50 g/liter sucrose, 0.5 mM IPTG (isopropyl--Dthiogalactopyranoside), and 0.2% rhamnose.
Genetic and molecular techniques. Linear and plasmid DNA transformations were performed as described previously (60), as were P1vir transductions [bib_ref] Bacteriophage-mediated generalized transduction in Escherichia coli and Salmonella typhimurium, Sternberg [/bib_ref]. DNA constructs were created using an NEBuilder HiFi DNA assembly kit (catalog number E5520; NEB). Chromosomal gene deletions were generated using a bacteriophage Red recombinase system [bib_ref] One-step inactivation of chromosomal genes in Escherichia coli K-12 using PCR products, Datsenko [/bib_ref] or a Fast genome engineering system [bib_ref] Leveraging modern DNA assembly techniques for rapid, markerless genome modification, Tikh [/bib_ref]. All PCR products used in the constructions were generated using E. coli MG1655 chromosomal DNA as the template, and the sequences of all strains were verified by PCR and/or sequence analysis (NEB DNA Sequencing Facility). PCRs used to generate sequencing templates or genetic constructions were performed with Q5 high-fidelity DNA polymerase (catalog number M0491; NEB), while diagnostic PCRs used the Hot Start Taq 2ϫ master mix (catalog number M0496; NEB).
Matings. Donor and recipient cultures were grown at 37°C to an optical density at 600 nm (OD 600 ) of ϳ1.0 in RB with shaking. Mating was initiated by mixing the cultures in a 1:1 ratio and lasted for 18 h at 37°C. For rhamnose induction, 0.2% L-rhamnose monohydrate (catalog number R3875; Sigma-Aldrich) was added to the mating mixture at the start of mating. For AZT treatment, 2.5 ng/ml 3=-azido-3=-deoxythymidine (catalog number A2169; Sigma-Aldrich) was added to the mating mixture at the start of mating.
Mating mixtures for donors (Tc), for recipients (Sm), and for recombinants (Tc and Sm) were plated at dilutions that would yield at least 10 to 100 colonies per plate. In matings with a low recombination efficiency, up to 6 ml of the mating culture was centrifuged, resuspended in residual broth, and spread on multiple recombinant selective plates. Plates were incubated for 48 h at 37°C, and the colonies were counted to calculate the recombination efficiency and cell survival. Unmated donor and recipient cultures were subjected to identical procedures to determine the number of CFU per milliliter in unmated cells.
Characterization of recombinants. Recombinant colonies were purified once on the same selection medium and tested for markers that distinguish the donor and recipient genomes.
lacZ screen for large replacements. To score the frequency of large replacement recombinants among the total, we tested for the presence of lacZ, located 418 kb distal to the selected tetRA cassette in the donor genome but absent in the recipient . RB X-Gal plates were incubated at 37°C for 2 days; blue colonies were classified as lacZ ϩ .
Antibiotic screens. All recombinants were screened for unselected drug markers chloramphenicol (cat) and kanamycin (npt).
PCR screens. The mrr locus was screened with primers pER91 and pER92; the fhuA locus was screened with primers oTK148 and oTK149.
Assessing the SOS response. We tested the effect of YhgA-like protein overexpression on the SOS response using the dinD-lacZ DNA damage response reporter [bib_ref] DNA-damaging agents stimulate gene expression at specific loci in Escherichia coli, Kenyon [/bib_ref] [bib_ref] SOS induction as an in vivo assay of enzyme-DNA interactions, Heitman [/bib_ref]. The dinD2::MudI1734 [Kan r lacZ(Ts)] version of this reporter isolated by Piekarowicz et al. [bib_ref] A new method for the rapid identification of genes encoding restriction and..., Piekarowicz [/bib_ref] was used because the indication is less ambiguous than that obtained with the leaky wild-type dinD-lacZ reporter [bib_ref] Direct selection of binding proficient/ catalytic deficient variants of BamHI endonuclease, Dorner [/bib_ref]. RB X-Gal plates with or without 0.2% rhamnose were used for initial assessment. Quantitative assays used cultures grown to an OD 600 of 0.2 in RB at 30°C. Rhamnose (0.2%) was added, and growth was continued at 30°C with continuous shaking. Samples collected at various times after rhamnose addition were lysed by sonication, and -galactosidase assays were performed.
Characterizing RpnA in vitro. (i) Purifying RpnA and its variants. The expression and purification of RpnA employed the Impact system (catalog number E6901; NEB) [bib_ref] Single-column purification of free recombinant proteins using a selfcleavable affinity tag derived..., Chong [/bib_ref]. Briefly, rpnA was inserted into the pTXB1 vector (pTK038) and transformed into the E. coli T7 expression strain (strain ER2566; NEB). RpnA was purified from the resulting strain (ER3573) as described in the kit's manual and then concentrated with a Vivaspin20 centrifugal ultrafiltration device (molecular weight cutoff, 10,000; GE Healthcare Life Sciences), dialyzed overnight in diluent A (catalog number B8001S; NEB), and kept at Ϫ20°C.
RpnA purified with the Impact system was further purified by anion-exchange chromatography. Concentrated protein was diluted to 50 ml in buffer (20 mM Tris-HCl, pH 8.0, 100 mM NaCl) and loaded onto a HiTrap Q HP 5-ml column using an Äkta fast-performance liquid chromatography system (model P-920; GE Healthcare Life Sciences). The protein was eluted with a 150-ml linear gradient of 100 to 600 mM NaCl over 38 fractions. The majority of the sample eluted as a single broad peak, which was pooled, concentrated, and dialyzed in diluent A to a final concentration of ϳ10 mg/ml. The final purified product was confirmed to be RpnA via liquid chromatography (LC)-mass spectrometry (MS). The single band of the protein sample isolated by SDS-PAGE was excised, digested with trypsin, and injected onto a 25-cm 3-m C 18 analytical column with a 1/4-cm Poros R1 plug. LC was accomplished with a Proxeon Easy-nLC II liquid chromatograph (Thermo Fisher), and MS data were collected from an LTQ Orbitrap XL mass spectrometer (Thermo Fisher) using a 60-min collision-induced dissociation/electron transfer dissociation (CID/ETD) data-dependent method. Data were analyzed with Proteome Discover (version 2.0) software (Thermo Fisher) and searched against the E. coli proteome to identify RpnA as the major protein in the sample.
Purification of the RpnA variants followed a similar protocol, in which expression strains ER3609 (RpnA-D63A) and ER3610 (RpnA-D165A) were used. The purities of the final products were confirmed by SDS-PAGE.
(ii) Assessing RpnA activity. RpnA endonuclease activity was determined by measuring DNA digestion with Tris-borate-EDTA (TBE)-agarose (1.5% or 0.75%) visualization by UV with ethidium bromide (EtBr) staining. Unless otherwise noted, the reaction conditions were as follows: purified protein and DNA substrate were combined in standard RpnA buffer (10 mM Tris-HCl, pH 9.0, 50 mM NaCl, 10 mM MgCl 2 , 15 mM CaCl 2 , 1 mM DTT) and incubated at 37°C. Substrates (all from NEB) were pUC19 (catalog number N3041), bacteriophage DNA (catalog number N3011), M13mp18 ssDNA (catalog number N4040), a dsRNA ladder (catalog number N0363), pUC19 nicked with Nb.BtsI (catalog number R0707), or pUC(AT), a pUC derivative carrying an extruded cruciform [bib_ref] Changing the enzymatic activity of T7 endonuclease by mutations at the -bridge..., Guan [/bib_ref]. Reactions were stopped with heat (70°C 10 min) or EDTA (30 mM). The 1.5% gels were run at 150 V for pUC19; the 0.75% gels were run at 100
[fig] Figure 2: (P ϭ 0.002, 0.018, 0.037, and 0.029, respectively). [/fig]
[fig] FIG 2: YhgA-like proteins increase recombination efficiency and are toxic. (A) Frequency of recombination during matings between the standard ΔrecA donor (strain ER3435) and either the ΔrecA recipient (control, ER3473) or a ΔrecA recipient with a rhamnose-inducible overexpression construct (rpnD, ER3481; rpnC, ER3512; rpnE, ER3513; rpnA, ER3514; and rpnB, ER3511) that was uninduced (no rhamnose) or induced (0.2% rhamnose). (B) Toxicity of expression was measured as the percent reduction in the number of recipient CFU per milliliter during rhamnose treatment relative to that for the control during the matings indicated in panel A. [/fig]
[fig] FIG 4: Site-specific endonucleases and AZT have disparate effects in this system. (A) Recombination efficiency: mating of the ΔrecA donor (ER3435) with and without rhamnose induction of the ΔrecA recipient (control, ER3473) or inducible rpnA (ER3514), mcrA (ER3533), asiSI (ER3535), or bsrDIB (ER3541) and with azidothymidine during mating (AZT; 2.5 ng/ml) (ER3435 ϫ ER3473). (B) Toxicity of the expressed proteins/treatments during these matings. [/fig]
[fig] FIG 5: Distribution of genomic exchanges in recombinants. (A) Diagram of markers that distinguish the donor and recipient genomes and distances from the selected tetRA cassette. The chromosome segregation site dif is also shown. ICR, restriction enzyme gene cluster known as the immigration control region. (B) [/fig]
[fig] FIG 6: Phenotypes of mutated RpnA proteins in vivo. (A) The RpnA domain structure, PD-(D/E)XK core motifs, and active-site residues predicted by Knizewski et al. [/fig]
[fig] FIG 7: In vitro analysis of RpnA endonuclease activity. (A) WT RpnA cleaves pUC19, RpnA-D63A does not cleave pUC19, and RpnA-D165A is more active on pUC19. The pUC19 DNA (29 nM, 50 g/ml) is initially supercoiled but can be relaxed by nicks, linearized by double-strand cleavage, or cleaved further. The supercoiled (control), relaxed (Nb.BtsI), and linear (HindIII) forms are indicated. pUC19 was treated with RpnA-inactive RpnA-D63A or hyperactive RpnA-D165A (15 M, 45 min). (B) Time course of an RpnA (7.5 M)-pUC19 (29 nM) digest. Band intensity was compared to determine the relative amounts of supercoiled, nicked, and linear pUC19 at each time point. Over 90% of the supercoiled pUC19 was digested within 180 min. (C) RpnA endonuclease activity depends on divalent cation and is stimulated by Ca 2ϩ . The reaction buffer was 50 mM NaCl and 10 mM Tris, pH 8.0; the indicated additives were at 10 mM each. RpnA at 3.8 M was added for 18 h. (D) RpnA cleavage products provide a DNA polymerase primer. [/fig]
|
Organizing pneumonia as the first manifestation of anti-synthetase syndrome
Background: Anti-synthetase syndrome associated interstitial lung disease can occur either simultaneously, before, or after the development of polymyositis/dermatomyositis. Histology of interstitial lung disease can be nonspecific interstitial pneumonia, usual interstitial pneumonia, diffuse alveolar damage, organizing pneumonia. Organizing pneumonia associated anti-synthetase syndrome is a rare finding especially as the first manifestation.Case presentation:We report a 41 year old male patient who presented with organizing pneumonia and 2 years following the onset, developed polymyositis with anti-JO-1 antibody positivity.Conclusion:It is important to screen patients with organizing pneumonia for anti-synthetase syndrome which can be manifested later.
# Background
Polymyositis (PM) and dermatomyositis (DM) are systemic inflammatory disorders predominantly affecting skeletal muscles and skin respectively. They also affect the oesophagus, lungs and the heart. Pulmonary involvement is common, and may be a major cause of morbidity. It commonly manifests as an interstitial lung disease (ILD) which may progress rapidly and be fatal [bib_ref] Interstitial lung disease in polymyositis and dermatomyositis, Marie [/bib_ref]. Antisynthetase syndrome is characterized by serum antibodies to aminoacyl-tRNA synthetase and constellation of manifestations, including fever, PM-DM, ILD, arthritis, "mechanics hands". Organizing pneumonia (previously known as bronchiolitis oblitarence with organizing pneumonia/BOOP) is rarely reported in these patients, as the first presenting symptom [bib_ref] Bronchiolitis obliterans organizing pneumonia as the first manifestation of polymyositis, Fata [/bib_ref] [bib_ref] Polymyositis and Sjögren's syndrome associated with bronchiolitis obliterans organizing pneumonia, Imasaki [/bib_ref] [bib_ref] Bronchiolitis obliterans with organizing pneumonia (BOOP) heralding anti-Jo-1-positive polymyositis, Xing [/bib_ref]. We report a case of anti-synthetase syndrome initially presented to us with organizing pneumonia and 2 years later presented as polymyositis.
## Case presentation
A 41-year-old male with uncomplicated type 2 diabetes mellitus presented with fever and progressive exertional dyspnoea for 1 week. His respiratory system examination revealed bilateral fine end-inspiratory crackles. High resolution computed tomography (HRCT) of the chest showed sub pleural patchy air space consolidation and ground glass opacification with air bronchogram in bilateral basal regions with sparing of upper and middle zones [fig_ref] Figure 1: HRCT of the chest showing sub pleural patchy airspace consolidation and ground... [/fig_ref]. Trans-bronchial lung biopsy showed alveoli which are filled with sheets of foamy macrophages with few scattered neutrophils and fibrotic plugs. Alveolar septae appeared thickened. A diagnosis of OP was made based on these findings. During this presentation he did not complain any muscle pain, weakness, joint pains or skin thickening in fingers. He was successfully managed with steroids, which were tailed off after 10 months. Subsequently he was lost to follow up as he failed to showup at scheduled clinic visits. However 2 years later, he presented with a 2 week history of progressively worsening proximal muscle pain mainly involving the neck and upper limbs. He also had low-grade fever with no skin rashes, arthralgia or respiratory symptoms. There was no history of exposure to dust or medication except
## Open access
## Bmc research notes
*Correspondence: [email protected] National hospital of Sri Lanka, Colombo, Sri Lanka metformin. On examination both upper and lower limbs were neurologically normal except proximal muscle tenderness. His respiratory system examination was unremarkable.
Laboratory investigations revealed creatine phosphokinase (CPK) 14820 U/L, C reactive protein (CRP) 110 mg/ dL, erythrocyte sedimentation rate (ESR) 70 mm/1st h, thyroid stimulating test (TSH) 1.84 mIU/L and free T4 0.94 ng/dL. The electro-myography (EMG) did not show evidence of myositis or myopathy. Anti-Jo-1 antibody and the anti-nuclear-antibody were positive. Deltoid muscle biopsy and magnetic resonance imaging (MRI) were compatible with polymyositis. Repeat HRCT and CXR were compatible with a relapse of an ILD.
Based on the clinical presentation and investigations, diagnosis of anti-synthetase syndrome was made. He was treated with high dose prednisolone and azathioprine. After 1 month of treatment the CPK and inflammatory markers normalized.
# Discussion
Anti-synthetase syndrome is a rare systemic autoimmune syndrome, characterized by the presence of anti-aminoacyl-tRNA antibodies (anti-Jo-1) accompanied by a constellation of clinical findings including PM-DM, ILD, "mechanics hand" appearance, fever and Reynaud's phenomenon [bib_ref] Antisynthetase syndrome presenting as cryptogenic organizing pneumonia, Haydour [/bib_ref] [bib_ref] Antisynthetase syndrome, Tzioufas [/bib_ref] [bib_ref] Antisynthetase syndrome: not just an inflammatory myopathy, Chatterjee [/bib_ref]. This occurs mainly in adults and more common in females and etiology is not known [bib_ref] Antisynthetase syndrome, Tzioufas [/bib_ref]. The anti-aminoacyl-tRNA antibodies directed toward the attachment of particular amino acid to its transfer RNA (tRNA). There are several anti-synthetase antibodies and anti-Jo-1 is the commonest and occurs in 80 % of patients with anti-synthetase syndrome [bib_ref] Fellow's forum case report: diagnosing antisynthetase syndrome, Aslam [/bib_ref] , others are PL-7, PL-12, OJ, EJ [bib_ref] Antisynthetase syndrome, Tzioufas [/bib_ref] [bib_ref] Antisynthetase syndrome: not just an inflammatory myopathy, Chatterjee [/bib_ref] [bib_ref] Anti-synthetase syndrome positive for anti-isoleucyl-tRNA synthetase antibodies: an unusual case overlapping with..., Hervier [/bib_ref].
Myositis, ILD and joint involvement are the classic tried in anti-synthetase syndrome. Myositis occurs in more than 90 % of patients and ILD found in more than 60 % of patients. ILD in anti-synthase syndrome is a major cause of morbidity and it can occur in the absence of myositis (amyopathic ILD) [bib_ref] Antisynthetase syndrome: not just an inflammatory myopathy, Chatterjee [/bib_ref] [bib_ref] Fellow's forum case report: diagnosing antisynthetase syndrome, Aslam [/bib_ref]. Because of this antisynthetase antibodies, particularly anti-Jo-1 should be performed in all patients with ILD without an obvious etiology [bib_ref] Fellow's forum case report: diagnosing antisynthetase syndrome, Aslam [/bib_ref]. Identification of anti-synthetase syndrome in patients with amyopathic ILD would be important as there are therapeutic implications [bib_ref] Antisynthetase syndrome: not just an inflammatory myopathy, Chatterjee [/bib_ref]. Studies have demonstrated the efficacy of immunosuppressive agents in ILD associated with anti-synthetase syndrome whereas lung transplantation has so far been the only treatment option in idiopathic pulmonary fibrosis [bib_ref] Antisynthetase syndrome: not just an inflammatory myopathy, Chatterjee [/bib_ref].
Histology may show different patterns including nonspecific interstitial pneumonia (NSIP), diffuse alveolar damage (DAD), usual interstitial pneumonia (UIP) or organizing pneumonia (OP). The prevalence of these histological features varies between reports and NSIP is the commonest pattern [bib_ref] Antisynthetase syndrome presenting as cryptogenic organizing pneumonia, Haydour [/bib_ref] [bib_ref] Antisynthetase syndrome: not just an inflammatory myopathy, Chatterjee [/bib_ref].
Although OP is commonly seen with rheumatoid arthritis (RA), manifesting after the onset of arthritis, it is rare with PM-DM and manifests before the onset of myositis as in our patient [bib_ref] Polymyositis and Sjögren's syndrome associated with bronchiolitis obliterans organizing pneumonia, Imasaki [/bib_ref]. However organizing pneumonia complicating polymyositis carries a better prognosis than UIP or DAD [bib_ref] Interstitial lung disease in polymyositis and dermatomyositis, Marie [/bib_ref].
Joint involvement occurs in more than 50 % of patients with anti-synthetase syndrome and it can range from simple arthralgia to arthritis which can be erosive [bib_ref] Antisynthetase syndrome, Tzioufas [/bib_ref]. Submit your next manuscript to BioMed Central and we will help you at every step:
"Mechanics hands" occurs in 30 % of patients and Raynaud phenomenon occurs in 40 % [bib_ref] Antisynthetase syndrome: not just an inflammatory myopathy, Chatterjee [/bib_ref].
Although over 90 % patients with polymyositis typically present with proximal muscle weakness, mild myalgias and muscle tenderness occur in 25-50 % of cases and 11 % of patients show normal EMG as in our patient [bib_ref] A computer-assisted analysis of 153 patients with polymyositis and dermatomyositis, Bohan [/bib_ref]. Positive anti-Jo-1 antibody, muscle histology and muscle MRI confirmed the diagnosis of polymyositis in this patient.
As our patient had positive anti JO-1 antibody together with fever, PM and ILD the diagnosis of anti-synthetase syndrome was made. At the time of diagnosis Raynaud phenomenon, joint involvement, "mechanics hands" were absent in our patient but these can be manifested later [bib_ref] Antisynthetase syndrome: not just an inflammatory myopathy, Chatterjee [/bib_ref].
The presence of Anti-Jo-1 is known to be associated with poor survival, lesser response to steroids and a higher incidence of flare-ups when steroids are tapered off. This patient however responded to steroids and is currently stable on a tail off regime of steroids and azathioprine.
# Conclusion
Organizing pneumonia associated anti-synthetase syndrome is a rare finding especially as the first manifestation. This case signifies the importance of screening patients with OP and other ILD without obvious etiology for anti-synthetase syndrome and arranging long term follow-ups [bib_ref] Fellow's forum case report: diagnosing antisynthetase syndrome, Aslam [/bib_ref].
Abbreviations PM: polymyositis; DM: dermatomyositis; ILD: interstitial lung disease; BOOP: bronchiolitis obliterance organizing pneumonia; OP: organizing pneumonia; NSIP: nonspecific interstitial pneumonia; DAD: diffuse alveolar damage; UIP: usual interstitial pneumonia.
[fig] Figure 1: HRCT of the chest showing sub pleural patchy airspace consolidation and ground glass opacification with airbronchogram in bilateral basel regions with sparing of upper and middle zones, which was compatible with OP Fig. 2 Histology of muscle biopsy showing varying sized muscle fibers with focal degenerative and regenerative changes and a chronic inflammatory infiltrate destroying myocytes (H&E) Priyangika et al. BMC Res Notes (2016) 9:290 • We accept pre-submission inquiries • Our selector tool helps you to find the most relevant journal • We provide round the clock customer support • Convenient online submission • Thorough peer review • Inclusion in PubMed and all major indexing services • Maximum visibility for your research Submit your manuscript at www.biomedcentral.com/submit [/fig]
|
The role of microRNAs in respiratory viral infection: friend or foe?
S U M M A RYMicroRNAs (miRNAs) have emerged as a class of regulatory RNAs in host-pathogen interactions. Aberrant miRNA expression seems to play a central role in the pathology of several respiratory viruses, promoting development and progression of infection. miRNAs may thus serve as therapeutic and prognostic factors for respiratory viral infectious disease caused by a variety of agents. We present a comprehensive review of recent findings related to the role of miRNAs in different respiratory viral infections and discuss possible therapeutic opportunities aiming to attenuate the burden of viral infections. Our review supports the emerging concept that cellular and viral-encoded miRNAs might be broadly implicated in human respiratory viral infections, with either positive or negative effects on virus life cycle. Figure 1. Following viral infection, host cells alter their microRNAs (miRNAs) expression as a defense against infection, while viruses can circumvent host defense and promote their own propagation by affecting host cellular miRNAs expression or by expressing their own miRNAs 391 MicroRNAs and respiratory viral infection
# Introduction
MicroRNAs (miRNAs) are small endogenous, noncoding RNAs, approximately 20-25 nt long. They are RNA-sequence-specific post-transcriptional regulators of gene expression [bib_ref] MicroRNAs: genomics, biogenesis, mechanism, and function, Bartel [/bib_ref]. miRNAs are expressed in a wide variety of organisms and originate in the nucleus as primary miRNA transcripts (~1000 nt), which are processed by the dsRNAspecific endonuclease Drosha into precursor miRNAs (pre-miRNA). The pre-miRNA (~70 nt) are transported to the cytoplasm, and further processed by Dicer into mature miRNAs. A single strand of mature miRNA is incorporated into the RNA-induced silencing complex (RISC), which binds to the three prime untranslated region (3 -UTR) of target mRNA, and exerts direct effects by blocking the translational process or inducing mRNA degradation, and indirect effects by influencing methylation or targeting of transcriptional factors [bib_ref] The widespread regulation of microRNA biogenesis, function and decay, Krol [/bib_ref] [bib_ref] MicroRNA-29 family reverts aberrant methylation in lung cancer by targeting DNA methyltransferases..., Fabbri [/bib_ref]. Over 2000 human miRNAs are currently recognized in the comprehensive miRNA database miRBase [bib_ref] miRBase: microRNA sequences, targets and gene nomenclature, Griffiths-Jones [/bib_ref] , and the function of many of these miRNAs in various biological processes including differentiation, proliferation, metabolism, and apoptosis is well established [bib_ref] MicroRNAs: genomics, biogenesis, mechanism, and function, Bartel [/bib_ref]. It is estimated that about 60% of human genes may be subjected to miRNA regulation. miRNA systems constitute complex combinatorial networks, where one miRNA may regulate many mRNA, and conversely, one mRNA may be regulated by several miRNAs.
Given the breadth of miRNA-mediated regulation of various biological process and immunity in mammals, the role of miRNAs has recently been highlighted in host-pathogen interactions [bib_ref] Physiological and pathological roles for microRNAs in the immune system, O'connell [/bib_ref] [bib_ref] MicroRNAs and the immune response to respiratory virus infections, Globinska [/bib_ref] [bib_ref] MicroRNAs in development and disease, Sayed [/bib_ref].
Host-pathogen interaction is the most important dynamic system in nature, and epigenetic modifications and post-transcriptional regulation through miRNA systems may provide an accessory source of fast-acting and readily available phenotypic variation that can be directly carried out by both host and pathogen selection pressures. Over the past decade, our knowledge of miRNA processes in various biological systems and hostpathogen interactions has rapidly advanced, but the precise role of miRNAs in the host-pathogen interactions is still unclear [bib_ref] miR-126 is downregulated in cystic fibrosis airway epithelial cells and regulates TOM1..., Oglesby [/bib_ref]. A number of studies in recent years report differential expression and biological function of miRNAs in airway cells [bib_ref] Therapeutic modulation of miRNA for the treatment of proinflammatory lung diseases, Hassan [/bib_ref] [bib_ref] The emerging role of microRNAs in regulating immune and inflammatory responses in..., Foster [/bib_ref]. miRNAs play an important role in physiological and pathological aspects of airway cells including pulmonary development, immune function, fibrosis, and cancer [bib_ref] Role of MicroRNAs in lung disease, Angulo [/bib_ref] [bib_ref] MicroRNAs in the lung, Sittka [/bib_ref]. In airway epithelial cells, miRNAs have been shown to affect numerous processes pertaining to respiratory pathogens, such as modulation of innate and adaptive immune responses, cell cycle progression, and apoptosis induction [bib_ref] MicroRNAs and the immune response to respiratory virus infections, Globinska [/bib_ref] [bib_ref] Respiratory syncytial virus infection of airway cells: role of microRNAs, Rossi [/bib_ref].
In this comprehensive review, we discuss recent findings that indicate an important role for miRNAs in various respiratory viral infections. We also discuss the putative significance of these effects on respiratory viral replication, viral cytopathogenicity, and the immune response. Identifying the role of miRNAs in respiratory viral infections may enhance our understanding of the mechanisms of infection and also indicate a potential future for miRNA-based therapies.
## Microrna and virus interaction
Viruses are obligate intracellular infectious agents that use the host cellular machinery to ensure their own fitness and survival. The success of viruses principally depends on their capability to efficiently use the host machinery to take advantage of basic biological processes [bib_ref] Viruses, masters at downsizing, Dimaio [/bib_ref] [bib_ref] Virus meets host microRNA: the destroyer, the booster, the hijacker, Guo [/bib_ref] [bib_ref] Hostvirus interaction: a new role for microRNAs, Scaria [/bib_ref]. miRNA systems have several features that make them ideal tools for virus propagation. They are potent post-transcriptional gene expression regulators. They are both small and non-antigenic and can modulate expression of several critical cellular pathways [bib_ref] RNAi and cellular miRNAs in infections by mammalian viruses, Haasnoot [/bib_ref] [bib_ref] Five questions about viruses and microRNAs, Cullen [/bib_ref] [bib_ref] The role of RNAi and microRNAs in animal virus replication and antiviral..., Umbach [/bib_ref]. miRNA systems modulate viral replication and pathogenesis in several ways: (i) Host cell miRNAs can positively or negatively affect viral replication and pathogenesis as a result of their biological functions. (ii) Some viruses also encode miRNAs. Current evidence indicates that viral-encoded miRNAs target several cellular genes involved in cell proliferation and survival, stress responses, and anti-viral response. (iii) Virus-encoded miRNAs may also regulate viral gene expression [bib_ref] The role of RNAi and microRNAs in animal virus replication and antiviral..., Umbach [/bib_ref] [bib_ref] Viruses and microRNAs, Cullen [/bib_ref] [bib_ref] MicroRNAs in viral replication and pathogenesis, Dykxhoorn [/bib_ref] [bib_ref] MicroRNAs in human virus genomes: helping hands for viral Infection, Liu [/bib_ref] [bib_ref] Virus-encoded microRNAs, Grundhoff [/bib_ref] [bib_ref] Viruses, microRNAs, and host interactions, Skalsky [/bib_ref] [bib_ref] Virus-encoded microRNAs: an overview and a look to the future, Kincaid [/bib_ref]. Thus, host-encoded miRNAs, virus-encoded miRNAs, and miRNA targets together form a novel regulatory system (miRNA system) between the host and the virus, which contributes to the outcome of infection [bib_ref] RNAi and cellular miRNAs in infections by mammalian viruses, Haasnoot [/bib_ref] [bib_ref] Hostvirus genome interactions: macro roles for microRNAs, Scaria [/bib_ref].
Host cellular miRNA expression is profoundly altered following viral infection , which affects the viral life cycle including viral replication, immune responses, and infection outcome [bib_ref] Viruses and microRNAs, Cullen [/bib_ref]. On the one hand, these changes could represent a host defense against infection and might therefore act to inhibit virus replication. On the other hand, changes in cellular miRNA may be induced by viruses to prepare a suitable environment for productive viral infection and/or latency [bib_ref] The role of RNAi and microRNAs in animal virus replication and antiviral..., Umbach [/bib_ref]. Host cellular miRNAs may have direct effects on viral replication, through positive or negative interactions with viral genomes or other viral factors [bib_ref] Roles of microRNAs in the life cycles of mammalian viruses, Gottwein [/bib_ref]. An miRNA system may also contribute to anti-viral host defense . Evidence suggests that miRNA can positively or negatively regulate innate and adaptive immune responses [bib_ref] MicroRNAs and the immune response to respiratory virus infections, Globinska [/bib_ref]. Furthermore, host cell miRNAs have potential to regulate virus tissue tropisms [bib_ref] Viruses and microRNAs, Cullen [/bib_ref]. Thus, miRNAs are utilized by viruses to invade host cells, replicate in host cell, evade host immune response, and establish and maintain virus latency [bib_ref] RNAi and cellular miRNAs in infections by mammalian viruses, Haasnoot [/bib_ref] [bib_ref] MicroRNAs in human virus genomes: helping hands for viral Infection, Liu [/bib_ref].
There are several reports demonstrating that some viruses take advantage of cellular miRNAs by enhancing the expression of specific cellular miRNAs. This can enhance virus replication, apparently by down-regulating specific cellular mRNA targets with anti-viral potential . Alternatively, viruses may use the miRNA system to limit their own replication in infected cells, allowing evasion of the host immune response, survival of infected cells, establishment of viral latency, and increased spread to other individuals in a population . Furthermore, other studies indicate that viral gene products inhibit cellular miRNA expression . Thus, viruses can induce certain cellular miRNAs that affect the virus life cycle positively and inhibit those that affect the virus life cycle negatively [bib_ref] Roles of microRNAs in the life cycles of mammalian viruses, Gottwein [/bib_ref]. Interestingly, some viruses can propagate despite the presence of host cell-encoded inhibitory miRNA. Viruses may avoid inhibition by cellular miRNAs by several methods including: (i) blocking cellular miRNA biogenesis; (ii) inhibiting cellular miRNA function; or (iii) evolving 3 -UTR sequences that miRNAs are unable to bind to because they are not complementary to miRNA, are too short, or have complex secondary structures that could restrict binding by RISCs . However, some reports demonstrated that specific cellular miRNAs can negatively inhibit virus replication .
There are several ways in which the association of cellular and viral-encoded miRNAs with pathology and their targets can be identified . Computational analyses (in silico) for predicting miRNAs and their targets are applied by most studies as the first step of a survey. However, computer-based predictions of miRNA-target interactions may or may not exist in reality and should be verified by in vitro and/or in vivo investigations, often involving addition and removal of miRNAs from a system. In recent years, rapid advances in next generation sequencing have been successfully incorporated to analyze miRNAs and their targets . In addition, deep sequencing of small RNAs isolated from virusinfected cells may provide valuable information .
## Micrornas and respiratory viruses
Respiratory viruses are the most common global health problem with morbidity and mortality worldwide . Respiratory viral infections are responsible for an enormous economic burden, precipitating considerable absence from school and work, large numbers of visits to clinicians, and also represent a major cause of exacerbations of chronic respiratory disease such as asthma and chronic obstructive pulmonary disease . Viruses most commonly associated with respiratory infections are orthomyxoviruses, adenoviruses, paramyxoviruses, coronaviruses, picornaviruses, human bocavirus, and human herpesviruses . The availability of effective vaccines against respiratory viral infections is limited, and other than the anti-influenza medications oseltamivir and zanamivir, no clinical anti-viral treatments for common respiratory viruses are available . Novel anti-viral therapeutic approaches to prevent and treat respiratory viral infection are needed according to the WHO initiative Battle against Respiratory Viruses .
In recent years, significant progress has been made in understanding the molecular mechanisms underlying respiratory virus infection and host interaction. Identification and characterization of the miRNA expression profile following respiratory viral infection and its implication in viral infection is an important tool for understanding host-virus interaction, mechanisms of infection, and also therapy strategy development. Here, we summarize the literature data on such hostrespiratory virus implications in humans and discuss how these implications can be used as research tools or targets in the development of novel antiviral therapeutics [fig_ref] Table 1: miRNAs effects in respiratory viral infection [/fig_ref].
## Rna viruses
Unlike DNA viruses, RNA viruses usually do not encode their own miRNA, and the reasons behind this discrepancy are debated theoretically [bib_ref] Viruses and microRNAs: RISCy interactions with serious consequences, Cullen [/bib_ref]. The majority of RNA viruses replicate in the cytoplasm where they cannot access the nuclear enzyme Drosha, which is required for miRNA processing. Those RNA viruses, which do have access to the nucleus (e.g. influenza and HIV-1), may avoid encoding their own miRNAs because excision of a primary miRNA from RNA virus genome would induce cleavage and destruction of viral genome [bib_ref] Five questions about viruses and microRNAs, Cullen [/bib_ref] [bib_ref] Virus-encoded microRNAs: an overview and a look to the future, Kincaid [/bib_ref]. In addition, viruses that undergo short lytic replication cycles are less likely to encode miRNAs [bib_ref] Viruses and microRNAs, Cullen [/bib_ref].
## Influenza virus
Influenza virus is a common respiratory pathogen that primarily infects airway epithelial cells and leads to clinical outcomes ranging from mild upper respiratory infection to severe pneumonia [bib_ref] Pathogenesis of influenza-induced acute respiratory distress syndrome, Short [/bib_ref]. The host cellular response, specifically miRNA dysregulation, is likely to play a critical role in influenza infection outcome . Recent studies show distinct miRNA expression profiles in ill patients with influenza A (H1N1), that is, down-regulation of miR-29a, miR-29c, let-7g, miR-146b-5p, miR-150, miR-342-3p, miR-769-5p, miR-30b, miR-31, miR-361-3p, miR-362-3p, miR-342, miR-155, miR-210, and miR-192. These miRNAs are involved in the regulation of important biological pathways during virus infection, such as mitogen-activated protein kinase, epidermal growth factor receptor, and toll-like receptor signaling pathways [bib_ref] Microarray analysis of microRNA expression in peripheral blood mononuclear cells of critically..., Song [/bib_ref]. Additional studies detected high expression of miR-299-5p and miR-335 in influenza patients [bib_ref] Microarray analysis of microRNA expression in peripheral blood mononuclear cells of critically..., Song [/bib_ref] [bib_ref] Let-7 microRNA-mediated regulation of IL-13 and allergic airway inflammation, Kumar [/bib_ref]. In contrast, miR-765, miR-34b, miR-519e, miR-18a, miR-628-3p, miR-185, miR-576-3p, miR-519d, miR-28-5p, miR-26a, miR-1285, miR-665, and miR-30a were down-regulated in H1N1 patients, and interestingly, miR-576-3p could affect viral entry into cells by regulating AP1G1 expression [60]. Furthermore, miR-17, miR-20a, miR-106a, and miR-376c were significantly elevated in H7N9 patients [bib_ref] Comprehensive characterization of serum microRNA profile in response to the emerging avian..., Zhu [/bib_ref].
The virulence of influenza virus may be mediated in part by host cellular miRNAs via dysregulation of pathways critical for anti-viral immune responses . Influenza infection up-regulates miR-29 expression, which is involved in regulation of both innate and adaptive immune responses through protection of A20 mRNA . miR-29 acts as an RNA decoy to prevent HuR (human antigen R) from binding to the A20 3 -UTR and recruiting the RISC [bib_ref] miR-29 acts as a decoy in sarcomas to protect the tumor suppressor..., Balkhi [/bib_ref]. A20 is a deubiquitinating enzyme known to play an important role in terminating the anti-viral immune response by inhibiting nuclear factor kappa B (NF-kB) and interferon regulatory factor pathways. Influenza infection of A549 cells induces expression of miR-146a, also a negative regulator of NF-kB [51]. Interestingly, one study showed that the zoonotic respiratory hendra virus induces miR-146a, which promotes viral replication by targeting ring finger protein 11 . In a study by Huang, et al., up-regulation of several miRNAs including miR-15b-3p, miR-24-2-5p, miR-331-3p, miR-124-3p, and miR-337-5p was demonstrated following H1N1 infection. These miRNAs participate in toll-like receptor and RIG-I-like receptor signaling pathways, and also regulate IL-1β and TNF receptor-associated factor 3 [52]. Furthermore, miR-7, miR-132, miR-146a, miR-187, miR-200c, and miR-1275 accumulate in human lung cell lines in response to infection with two influenza A virus strains, A/Udorn/72 and A/WSN/33, causing down-regulation of anti-viral proteins such as interleukin-1 receptor-associated kinase (IRAK1) and mitogen-activated protein kinase 3 [53].
Virulence of highly pathogenic influenza viruses may be mediated in part by host cellular miRNAs. For example, the highly pathogenic H5N1 virus induces miR-141 shortly after infection, which suppresses the expression of transforming growth factor-β in lung epithelial cells . Without sufficient transforming growth factor-β, the proinflammatory response might not be tightly controlled in cases of highly pathogenic H5N1 infection [bib_ref] TGFβ directs gene expression of activated microglia to an anti-inflammatory phenotype strongly..., Paglinawan [/bib_ref]. The 1918 pandemic influenza virus induces a distinct miRNA expression profile in mice compared with non-lethal influenza A/Texas/36/91, including down-regulation of miR-200a and up-regulation of miR-223; miR-223 is a negative modulator of neutrophil activation, and miR-200a has a role in the type I IFN response [59]. miVARNAs expression Down-regulates the HDGF expression [bib_ref] Adenovirus-encoding virus-associated RNAs suppress HDGF gene expression to support efficient viral replication, Kondo [/bib_ref] miR-214 expression Inhibits virus replication [bib_ref] E1A expression might be controlled by miR-214 in cells with low adenovirus..., Yanagawa-Matsuda [/bib_ref] Apoptosis is characteristic of influenza virus infection, and the mechanisms underlying this have advanced understanding of influenza virus replication . According to Othumpangat, et al., in the first hours after influenza infection, down-regulation of miR-4276 increases cytochrome c oxidase VIC expression, inhibiting viral replication by inducing the apoptotic protein caspase-9. However, after 6 to 9 h, this effect is completely reversed, thereby prolonging cell survival. This may suggest that influenza virus is able to induce miR-4276 and inhibit cytochrome c oxidase VIC and caspase-9 expression, thus promoting viral replication . Recent studies have revealed that miR-29 family members are up-regulated during influenza infection, especially miR-29c, which targets the anti-apoptotic factor B-cell lymphoma 2like 2 contributing to virus-mediated apoptosis . Furthermore, down-regulation of miR-30 family members and up-regulation of miR-223 during influenza infection lead to increased apoptosis .
The expression of host genes required for influenza virus replication can be regulated by multiple cellular miRNAs, for example, miR-106b, miR-124, and miR-1254 regulate human protease genes (ADAMTS7, CPE, DPP3, MST1, and PRSS12) that are essential for influenza replication . In addition, down-regulation of miR-24 with a concomitant up-regulation of furin mRNA has been demonstrated during the influenza H5N1 infection in A549 cells. miR-24 regulates furin-mediated activation of influenza hemagglutinin precursor and subsequent production of fusion-competent virions in the host secretory pathway . Some miRNAs play important roles in priming airway cells for repair and regeneration following influenza infection [bib_ref] Micro-RNAs in regenerating lungs: an integrative systems biology analysis of murine influenza..., Tan [/bib_ref]. Elevated expression of miR-21 throughout repair and regeneration corresponds with increased cell proliferation in repairing lungs, because miR-21 targets proliferationsuppressing factors . The miR-30 family was significantly down-regulated during repair consistent with increased expression of its main target p53, which promotes proliferation in recovering lung tissues [57].
While influenza virus can clearly take advantage of cellular miRNAs via their promotion or inhibition, it is also revealed that certain cellular miRNAs can inhibit replication of influenza viruses in infected cells . miR-323, miR-491, and miR-654 inhibit replication of the H1N1 influenza A virus in MDCK cells by targeting the same conserved region in the influenza PB1 gene .
Furthermore, let-7c inhibits M1 protein expression of the H1N1 influenza A virus in A549 cells .
Overall, these results suggest that influenza respiratory infection induces or inhibits expression of certain miRNAs in airway cells that favor viral replication, pathogenesis, and also suppress anti-viral responses [fig_ref] Figure 2: Influenza infection of airway epithelial cells induces or inhibits certain cellular microRNAs [/fig_ref]. Thus, cellular miRNAs associated with immune response, apoptosis, and protease genes could be the best candidates for development of miRNA-based therapies for influenza disease. However, caution must be taken due to the immunopathogenic character of influenza infection.
## Respiratory syncytial virus
RSV is a leading cause of viral respiratory tract disease among infants and young children [bib_ref] Infectious burden of respiratory syncytial virus in relation to time of birth..., Birkhaug [/bib_ref] [bib_ref] Local IL-17A potentiates early neutrophil recruitment to the respiratory tract during severe..., Stoppelenburg [/bib_ref]. Worldwide, 33.8 million episodes of RSV-associated acute lower respiratory tract infections are estimated to occur in children <5 years of age [bib_ref] Global burden of acute lower respiratory infections due to respiratory syncytial virus..., Nair [/bib_ref]. In developed countries, the RSV-associated mortality rates are reported to be approximately three deaths per 100 000 in children younger than 1 year [bib_ref] Respiratory syncytial virus-associated mortality in hospitalized infants and young children, Byington [/bib_ref] [bib_ref] Mortality caused by influenza and respiratory syncytial virus by age group in..., Hardelid [/bib_ref] [bib_ref] Impact of respiratory syncytial virus in the United States, Robinson [/bib_ref] [bib_ref] Mortality associated with seasonal and pandemic influenza and respiratory syncytial virus among..., Tempia [/bib_ref]. RSV infection is a common example of viruses that modulate host miRNA expression to influence the outcome of the anti-viral host response and viral replication [bib_ref] MicroRNAs and the immune response to respiratory virus infections, Globinska [/bib_ref] [bib_ref] Respiratory syncytial virus infection of airway cells: role of microRNAs, Rossi [/bib_ref] [bib_ref] Reduced Dicer expression in the cord blood of infants admitted with severe..., Inchley [/bib_ref]. Distinct immuneassociated miRNA expression profiles have been detected in the nasal epithelium of RSV-positive infants; down-regulation of miR-34b, miR-34c, miR-125b, miR-29c, miR125a, miR-429, and miR-27b, and up-regulation of miR-155, miR-31, miR-203a, miR-16, and let-7d were detected in these patients. In addition, miR-125a and miR-429 were downregulated in mild disease, but not in severe disease, and the lack of down-regulation in severe disease may rationalize the observed differences in disease manifestations following RSV infection [bib_ref] Nasal mucosal microRNA expression in children with respiratory syncytial virus infection, Inchley [/bib_ref]. miR-125a regulates the expression of NF-kB by suppressing the inhibitor protein A20, and chemokine (C-C motif) ligand (CCL5), an important cytokine in both innate and adaptive immune systems [bib_ref] Identifying functional microRNAs in macrophages with polarized phenotypes, Graff [/bib_ref].
RSV infection of A549 cells induced let-7f, miR-24, miR-337-3p, miR-26b, and miR-520a-5p and repressed miR-198 and miR-59 expression. Let-7f expression was RSV G protein dependent, and its expression likely contributes to delayed viral clearance by targeting CCL7 and suppressor of cytokine signaling 3, which are involved in anti-viral response . In another study, let-7b, let-7c, let-7i, and miR-30b were up-regulated on RSV infection of monocyte derived dendritic cells and human bronchial epithelial cells, and associated with IFN-β and/or NF-kB activation. Interestingly, RSV nonstructural proteins NS1 and NS2 antagonized the up-regulation of let-7i and miR-30b, a process that may favor viral replication [bib_ref] Respiratory syncytial virus regulates human microRNAs by using mechanisms involving beta interferon..., Thornburg [/bib_ref]. The miRNAs described in these studies have a number of experimentally confirmed targets that are associated with RSV replication and pathology. For example, an experimentally confirmed target of the let-7 family is IL-13, which appears to enhance the severity of disease [bib_ref] Proinflammatory role for let-7 microRNAS in experimental asthma, Polikepahad [/bib_ref].
RSV infection modifies the expression of critical neurotrophic factors and receptors such as nerve growth factor (NGF), and its cognate high-affinity receptor tropomyosin-related kinase A (TrkA), which prevents apoptosis by increasing expression of the anti-apoptotic Bcl-2 family members [bib_ref] NGF is an essential survival factor for bronchial epithelial cells during respiratory..., Othumpangat [/bib_ref]. In human bronchial epithelial cells, high levels of intracellular miR-221 reduced NGF and TrkA expression, which favor the apoptotic death of infected cells, and attenuate virus infection. RSV infection reduces miR-221 expression, thus interfering with the apoptotic death of infected cells by increasing NGF and TrkA expression and ultimately promoting viral replication [bib_ref] MicroRNA-221 modulates RSV replication in human bronchial epithelium by targeting NGF expression, Othumpangat [/bib_ref].
Overall, these findings suggest that following RSV respiratory infection, an altered expression profile of distinct immune-associated miRNAs occurs in the airway cells that inhibit viral replication and preserve the airway epithelial barrier. However, the virus concurrently induces or inhibits the expression of other miRNAs that favor viral replication. These conflicting miRNA effects during RSV infection may provide treatment options in susceptible individuals. However, attempts to modulate RSV pathology in clinical practice should be made with caution as RSV immunopathogenesis is complicated and an early RSV vaccine candidate caused serious adverse events during natural RSV infection [bib_ref] Respiratory syncytial virus disease in infants despite prior administration of antigenic inactivated..., Kim [/bib_ref].
## Coronaviruses
Coronaviruses can cause a wide spectrum of respiratory infections ranging from mild, upper respiratory tract infections to severe and life-threatening lower respiratory tract infections. There are no in vivo studies regarding the role of miRNAs in coronaviruses infection, but the OC43 virus has been investigated in vitro, and severe acute respiratory syndrome-coronavirus (SARS-CoV) and Middle East respiratory syndrome-coronavirus (MERS-CoV) were analyzed by in silico methods.
The coronavirus OC43 contributes to the common cold worldwide [bib_ref] Genetic drift of human coronavirus OC43 spike gene during adaptive evolution, Ren [/bib_ref]. Coronavirus N protein is essential for replication and binds to genomic RNA to form a helical capsid. OC43 N protein potentiates NF-kB activation by binding to its negative regulator miR-9. It is not clear whether NF-kB activation is directly beneficial to viral replication, or whether this is an incidental effect that limits viral virulence. Compared with more pathogenic coronaviruses, reduced OC43 virulence with limited symptoms may promote contact between infected and non-infected individuals and thus promote spread of the virus within a population [bib_ref] Human Coronavirus OC43 Nucleocapsid Protein Binds MicroRNA 9 and Potentiates NF-kB Activation, Lai [/bib_ref]. This novel mechanism of miRNA-binding to promote gene activity may provide insight into the mechanisms by which successful RNA viruses avoid the host immune system or cause pathology.
Severe acute respiratory syndrome-coronavirus is a novel coronavirus that threatened to cause a global pandemic of the severe acute respiratory syndrome in 2002-2003 [bib_ref] A novel coronavirus associated with severe acute respiratory syndrome, Ksiazek [/bib_ref]. An in silico analysis of miRNA interactions with SARS-CoV mRNA suggested that the virus might suppress its own replication early during infection by up-regulation of miR-17, miR-574-5p, and miR-214. These host miRNAs target all four virulent viral proteins, spike (S), nucleocapsid (N), matrix (M), and envelope (E) [bib_ref] MicroRNome analysis unravels the molecular basis of SARS infection in bronchoalveolar stem..., Mallick [/bib_ref]. Suppression of viral replication may aid evasion of immune surveillance until successful infection of other cells. These results demonstrate how SARS-CoV might alter host miRNA expression profile to its own advantage.
Middle East respiratory syndrome, caused by a novel human coronavirus MERS-CoV, has emerged recently. An in silico analysis identified miR-628-5p, miR-6804-3p, miR-4289, miR-208a-3p, miR-510-3p, miR-18a-3p, miR-329-3p, miR-548ax, miR-3934-5p, miR-4474-5p, miR-7974, miR-6865-5p, and miR-342-3p as having significant sequence similarity to hairpin structures in the MERS-CoV genome, and they may thus down-regulate viral gene expression to inhibit viral replication [bib_ref] A computational approach for predicting role of human microRNAs in MERS-CoV genome, Hasan [/bib_ref].
This knowledge may help us to better understand host-virus interactions with the intention to develop new anti-viral therapies against MERS-CoV, a highly lethal respiratory disease.
## Rhinoviruses
Rhinoviruses are members of the Picornaviridae family. Rhinoviruses cause respiratory infection in humans with severity ranging from the common cold to viral bronchiolitis, and exacerbations of asthma and chronic obstructive pulmonary disease [bib_ref] Infantile respiratory syncytial virus and human rhinovirus infections: respective role in inception..., Rossi [/bib_ref]. Bondanese, et al. showed that cellular miRNAs miR-128 and miR-155 with putative sites in the rhinovirus-1B coding region can inhibit virus replication. miR-128 inhibition seemed to increase viral replication by inducing apoptosis. The detection of miR-155-mediated anti-viral activity in bronchial epithelial cells is very relevant because this miRNA has a central role in innate and adaptive immunity. As an example, miR-155 has been shown to target suppressor of cytokine signaling 1, an inhibitor of type I IFN signaling [38]. . Following RSV respiratory infection, an altered expression profile of certain cellular miRNAs, specifically immune-associated miRNAs, occurs in order to inhibit viral replication and preserve the airway epithelial barrier; meanwhile, the virus induces or inhibits the expression of other miRNAs that favor viral replication. The RSV G protein enhances let-7f, the RSV NS1/NS2 proteins decrease miR-30b and enhance let-7i, and RSV infection decrease miR-221, which is an advantage for the virus A minor group of rhinoviruses including subtypes 1A, 1B, 2, 23, 25, 29, 30, 31, 44, 47, 49, and 62 commonly utilize the very low-density lipoprotein receptor (VLDLR) for entry into host cells, and cause disease more often than the major group. Recent evidence published by Ouda, et al., showed that down-regulation of VLDLR by miR-23b is of significance for host defense against the minor group of rhinoviruses. miR-23b was induced by RIG-I-like receptor signaling resulting in suppression of respiratory infections caused by minor group viruses, specifically rhinovirus-1B through down-regulation of its receptor VLDLR [bib_ref] Retinoic acid-inducible gene I-inducible miR-23b inhibits infections by minor group rhinoviruses through..., Ouda [/bib_ref]. In conclusion, these results suggest that miRNAs play an important role in human anti-viral responses against rhinovirus infection [fig_ref] Figure 5: In rhinoviruses infection, cellular microRNAs play anti-viral responses against viruses [38], human... [/fig_ref].
## Human metapneumovirus
Human metapneumovirus causes acute respiratory disease in infants, the elderly, and immunocompromised individuals ranging from mild upper respiratory illness to more serious lower respiratory illness [bib_ref] A newly discovered human pneumovirus isolated from young children with respiratory tract..., Van Den Hoogen [/bib_ref]. Limited literature is available regarding the role of miRNAs in human metapneumovirus (HMPV) infection. Deng, et al. reported that host airway epithelial cells alter their miRNA expression profile upon HMPV infection as a defense mechanism against the virus. The HMPV M2-2 protein acted as a key viral protein that regulated host cell miRNA expression, specifically antagonizing miR-30a and miR-16 [fig_ref] Figure 5: In rhinoviruses infection, cellular microRNAs play anti-viral responses against viruses [38], human... [/fig_ref]. Interestingly, M2-2mediated miR-16 suppression was interferon dependent, whereas suppression of miR-30a was interferon independent [bib_ref] Human metapneumovirus infection induces significant changes in small noncoding RNA expression in..., Deng [/bib_ref].
The aforementioned data suggest a new way in which HMPV regulates the host cell response to infection. There are currently no licensed therapeutics or vaccines against HMPV. These and future studies may help the development of effective miRNA-based therapies.
## Dna viruses
Many DNA viruses encode their own miRNAs, because they generally replicate in the nucleus and have access to the canonical miRNA pathway (except poxviruses) [bib_ref] Virus-encoded microRNAs: an overview and a look to the future, Kincaid [/bib_ref]. Most DNA viruses establish long-term latent or persistent infections and take advantage of virus-encoded and host cell miRNAs . Coronaviruses interact with the host cell at the onset of infection and induces several changes in host cellular microRNAs (miRNAs) expression profile to their own advantage; severe acute respiratory syndrome-coronavirus (SARS-CoV) uses cellular miRNAs machinery to evade immune elimination [bib_ref] MicroRNome analysis unravels the molecular basis of SARS infection in bronchoalveolar stem..., Mallick [/bib_ref] ; in Middle East respiratory syndrome-coronavirus (MERS-CoV), host cells miRNAs would be an anti-viral therapeutic agent [bib_ref] A computational approach for predicting role of human microRNAs in MERS-CoV genome, Hasan [/bib_ref] , and the N protein of OC43-coronavirus (OC43-CoV) causes potentiation of nuclear factor kappa B (NF-kB) activation via binding to its negative regulator miR-9 [bib_ref] Human Coronavirus OC43 Nucleocapsid Protein Binds MicroRNA 9 and Potentiates NF-kB Activation, Lai [/bib_ref] [22]. Because of the fact that viral miRNAs, unlike viral proteins, are non-immunogenic, viruses have developed their own miRNAs in order to escape and suppress both host innate and adaptive immune responses [bib_ref] Temporal landscape of microRNA-mediated host-virus crosstalk during productive human cytomegalovirus infection, Kim [/bib_ref].
## Adenoviruses
Adenoviruses cause mild to serious respiratory tract infections in many age groups [bib_ref] Adenovirus, Lynch [/bib_ref]. Adenovirus infection has a great impact on cellular miRNA expression profiles [bib_ref] High-throughput sequencing of microRNAs in adenovirus type 3 infected human laryngeal epithelial..., Qi [/bib_ref] [bib_ref] Fluctuating expression of microRNAs in adenovirus infected cells, Zhao [/bib_ref]. A total of 44 miRNAs demonstrated high expression and 36 miRNAs low expression following adenovirus type 3 infection in human laryngeal epithelial cells [bib_ref] High-throughput sequencing of microRNAs in adenovirus type 3 infected human laryngeal epithelial..., Qi [/bib_ref]. A temporal study demonstrated dramatic changes in cellular miRNA expression patterns during the course of adenovirus type 2 infection in lung fibroblast cells; up-regulation of miR-22, miR-320, let-7, miR-181b, miR-155, miR-125, miR-27, and miR-191 and downregulation of miR-21, miR-31, let-7 family, miR-30 family, and miR-23/27 cluster was detected. These miRNAs have been associated with host immune evasion and inflammatory responses, as well as in virus entry, replication, and propagation [bib_ref] Fluctuating expression of microRNAs in adenovirus infected cells, Zhao [/bib_ref].
Adenoviruses encode a set of highly abundant miRNAs that are generated by Dicer-mediated cleavage of the larger non-coding virus-associated RNAs (VARNAs) I and II. VARNAs are dsRNA molecules similar in structure to cellular pre-miRNAs. They are transported by exportin 5 into the cytoplasm, and processed to functional viral miRNAs (miVARNAs) [bib_ref] Adenovirus virus-associated RNA is processed to functional interfering RNAs involved in virus..., Aparicio [/bib_ref]. miVARNAs actively target the expression of cellular genes involved in cell proliferation, DNA repair, or RNA regulation. VARNAs are expressed at very high levels in adenovirus-infected cells and potently inhibit human pre-miRNA via inhibition of nuclear export of pre-miRNA, competition for exportin 5 to facilitate their transportation, and inhibition of Dicer activity by direct binding to Dicer [bib_ref] Identification of RISC-associated adenoviral microRNAs, a subset of their direct targets, and..., Bellutti [/bib_ref] [bib_ref] Adenovirus VA1 noncoding RNA can inhibit small interfering RNA and MicroRNA biogenesis, Lu [/bib_ref]. Adenovirus miVARNAs target cellular and viral genes that are important for the virus cell cycle. Hepatoma-derived growth factor inhibits adenovirus growth. However, the expression level of hepatomaderived growth factor significantly decreased in response to miVARNAs under replication-deficient conditions, and this suppression was also observed during the early phase of viral infection under replication-competent conditions [bib_ref] Adenovirus-encoding virus-associated RNAs suppress HDGF gene expression to support efficient viral replication, Kondo [/bib_ref]. Adenovirus miVARNAs also target cellular genes involved in cell growth, gene expression and DNA repair. The TIA-1 (cytotoxic granule-associated RNA binding protein) is down-regulated at mRNA and protein levels in infected cells expressing functional miVARNAs and in transfected cells [bib_ref] Adenovirus VA RNA-derived miRNAs target cellular genes involved in cell growth, gene..., Aparicio [/bib_ref].
Conversely, cellular miRNAs may play a role in anti-adenovirus replication by regulating virus gene expression. It was shown that cellular miR-214 inhibits adenovirus replication by regulating the translation of viral E1A protein, which is key [bib_ref] Human metapneumovirus infection induces significant changes in small noncoding RNA expression in..., Deng [/bib_ref] , and HHV-6A miR-U86 targets the HHV-6A IE gene U86, thereby regulating virus lytic replication [bib_ref] A human herpesvirus 6A-encoded microRNA: role in viral lytic replication, Nukui [/bib_ref].
to the activation of other adenovirus genes, while inhibition of miR-214 increases the productive efficiency of the virus [bib_ref] E1A expression might be controlled by miR-214 in cells with low adenovirus..., Yanagawa-Matsuda [/bib_ref]. showed that cellular miR-466 can effectively down-regulate human Coxsackie virus and adenovirus receptor protein expression [bib_ref] miR-466 is putative negative regulator of Coxsackie virus and adenovirus receptor, Lam [/bib_ref]. Furthermore, a subset of cellular miRNAs including miR-1, miR-34, miR-22, miR-365, miR-29, miR-145, and let-7 was shown to coordinately target retinoblastoma-dependent cell cycle and DNA replication mRNAs to restrict proliferation [bib_ref] Differentiation-associated microRNAs antagonize the Rb-E2F pathway to restrict proliferation, Marzi [/bib_ref].
Taken together, these results suggest that miVARNA-mediated silencing can represent a novel mechanism used by adenoviruses to control cellular or viral gene expression, and are potential therapeutic targets. The actions of cellular miRNAs may also be exploited to combat adenovirus infection .
## Human cytomegalovirus
Human cytomegalovirus (HCMV), a DNA virus, infects a broad range of human cell types and disrupts cellular processes through a variety of mechanisms. For example, HCMV uses several of its own encoded proteins to disrupt the MHC class I pathway [bib_ref] Human cytomegalovirus protein pp 71 disrupts major histocompatibility complex class I cell..., Trgovcich [/bib_ref] and the fine balance between a beneficial and a destructive immune response , and uses several of its own encoded miRNAs to disrupt a variety of cellular pathways such as TLR2/IRAK1/NF-kB signaling. This virus therefore induces a complex and diverse pathogenesis, and is an opportunistic pathogen causing lung infection in immunocompromised individuals [bib_ref] Asymptomatic carriage of Pneumocystis jirovecii and cytomegalovirus in lungs of immunocompetent patients, Shteinberg [/bib_ref]. Host cell miRNA expression levels may determine the cellular site of HCMV infection. As an example, host miR-200 family members target the HCMV protein UL112 resulting in repression of this viral protein, and cells permissive for lytic HCMV replication demonstrate low levels of these miRNAs [bib_ref] Host microRNA regulation of human cytomegalovirus immediate early protein translation promotes viral..., O'connor [/bib_ref]. However, HCMV can also selectively alter the expression of some cellular miRNAs to help its own replication [bib_ref] Human cytomegalovirus infection alters the expression of cellular microRNA species that affect..., Wang [/bib_ref]. For example, significant up-regulation of miR-96, miR-182, and miR-183 have been observed following infection. A study by Fu, et al., indicated that expression of host miRNAs may be affected by latent HCMV; at least 49 miRNAs were differentially expressed; . Adenovirus encodes viral miRNAs (miVARNAs) that potently inhibit human pre-microRNA (miRNA) via inhibition the nuclear export of pre-miRNA, competition for the exportin 5, and inhibition of Dicer activity by direct binding of Dicer. The miVARNAs are able to target cellular and viral genes that are important for virus cell cycle. Adenovirus miVARNAs target cellular genes involved in cell proliferation, DNA repairing, and RNA regulation. However, cellular miRNAs may play a role in anti-adenovirus replication by regulating virus gene expression. 39 were up-regulated and 10 were down-regulated accordingly [bib_ref] Human cytomegalovirus latent infection alters the expression of cellular and viral microRNA, Fu [/bib_ref]. In addition, HCMV encodes its own miRNAs that target both viral and cellular genes in order to regulate viral replication, viral latency, cell survival, and anti-viral immunity [79].
Human cytomegalovirus has miRNAs that help escape and suppress both host innate and adaptive immune responses [bib_ref] Temporal landscape of microRNA-mediated host-virus crosstalk during productive human cytomegalovirus infection, Kim [/bib_ref]. HCMV-encoded miR-UL148D modulates host immune response by directly targeting the mRNA of human chemokine CCL5 . HCMV miR-UL112 attenuates NK cell activity by inhibition of type I IFN secretion [bib_ref] Hcmv-miR-UL112 attenuates NK cell activity by inhibition type I interferon secretion, Huang [/bib_ref] , down-regulation of IL-32 expression [bib_ref] Human cytomegalovirus clinical strain-specific microRNA miR-UL148D targets the human chemokine RANTES during..., Kim [/bib_ref] , and TLR2 targeting, causing significant modulation of the downstream signaling pathway (TLR2/ IRAK1/NF-kB). HCMV may evade CD8 + T-cells by altering MHC class 1 antigen expression; HCMV miR-US4-1 targets the endoplasmic reticulum aminopeptidase 1, a key step in the MHC class I antigen-processing pathway . Furthermore, HCMV expresses miR-US25-2 and, in addition, increases cellular miR-17p expression, both of which target tissue inhibitor of metalloproteinase 3. Reduced tissue inhibitor of metalloproteinase 3 expression following HCMV infection reduces signaling via the MHC class I-like ligand MICA .
Some HCMV-encoded miRNAs suppress virus replication and lytic infection, which could help the virus to establish or maintain latent infection. It has been reported that HCMV miR-US25-1-5p was highly expressed during lytic and latent infections, and inhibited viral replication [83]; HCMV miR-US25-2 reduces viral replication by targeting the RNA helicase eIF4A1, which is a requisite for translation of viral mRNA [bib_ref] Over-expression of human cytomegalovirus miR-US25-2-3p downregulates eIF4A1 and inhibits HCMV replication, Qi [/bib_ref] , and HCMV miR-US33 negatively influences virus replication, possibly by suppression of the HCMV gene US29 [84] or cellular syntaxin 3 expression [131]. Premature expression of HCMV miR-UL112-1 during infection resulted in a significant decrease in genomic viral DNA levels, suggesting a functional role for miR-UL112-1 in regulating the expression of genes involved in viral replication . Finally, HCMV encodes latency-associated CMV-IL-10, a homologue for cellular IL-10 associated with latent infection. Latency-associated CMV-IL-10 has been shown to suppress miR-92a, resulting in up-regulation of its target CCL8. The mechanisms for both miR-92a suppression and how CCL8 up-regulation might . Human cytomegalovirus, a DNA virus, encodes its own microRNAs, and human cytomegalovirus microRNAs target both viral and cellular genes in order to; first regulation of viral replication, second regulation of viral latency infection, and third regulation of cellular anti-viral immunity.
promote latent infection are unclear, but seem to be associated with increased immune-regulatory cellular IL-10. These results provide insight into how HCMV can alter host gene expression .
Human cytomegalovirus-encoded miRNAs do not only suppress virus lytic replication but can also enhance virus replication. For example, HCMV-restricting cellular BclAF1 is downregulated late in infection by HCMV-encoded miR-UL112-1 to promote virus production [87]. Furthermore, multiple HCMV-encoded miRNAs coordinately regulate reorganization of the secretory pathways responsible for controlling cytokine secretion and facilitate formation of the viral assembly compartment for efficient infectious virus production. In this aspect, HCMV-encoded miRNAs such as miR-UL112-1, miR-US5-1, and miR-US5-2 target multiple components of the host secretory pathways, including VAMP3, RAB5C, RAB11A, SNAP23, and CDC42 [bib_ref] Cytomegalovirus miRNAs target secretory pathway genes to facilitate formation of the virion..., Hook [/bib_ref]. Additionally, HCMV employs its miRNA repertoire to counter cellular apoptosis and autophagy, particularly the mitochondrial-dependent intrinsic pathway of apoptosis. The pro-apoptotic genes MOAP1, PHAP, and ERN1 are identified as potential targets for miR-UL70-3p and miR-UL148D, respectively [bib_ref] Role of HCMV miR-UL70-3p and miR-UL148D in overcoming the cellular apoptosis, Babu [/bib_ref]. Finally, a viral intergenetic non-coding RNA element, composed of highly conserved sequences throughout HCMV clinical strains, selectively degrades the cellular miR-17 family members of the miR-17-92 cluster and accelerates virus production [bib_ref] Selective degradation of host MicroRNAs by an intergenic HCMV noncoding RNA accelerates..., Lee [/bib_ref].
Overall, these results suggest that identification and characterization of the HCMV-encoded miRNAs that are expressed during lytic and latent infection are crucial to understanding their roles in HCMV persistence, pathogenesis, and disease. Knowledge of host and viral miRNAs expressed during HCMV infection can thus provide a precise insight into viral pathogenesis and may help researchers to develop new therapeutic approaches.
## Human herpesvirus 6
HHV-6, a DNA virus in the betaherpesvirus subfamily, is associated with several human diseases. Complications of acute respiratory tract infection such as pneumonia and sinusitis in young children are associated with HHV-6 as is limbic encephalitis following hematopoietic stem cell transplantation. In addition, HHV-6 salivary gland replication and subsequent secretion in saliva is the epidemiologically proven source of transmission [bib_ref] Coinfection with human herpesvirus 6 variants A and B in lung tissue, Cone [/bib_ref]. As discussed previously, herpesvirus-derived miRNAs play considerable roles in modulating both cellular and viral gene expression, thereby facilitating a suitable environment for productive viral infection and/or latency. Like other human herpesviruses, HHV-6 encodes its own miRNAs, promoting efficient viral infection [bib_ref] A human herpesvirus 6A-encoded microRNA: role in viral lytic replication, Nukui [/bib_ref] [bib_ref] Small RNA deep sequencing identifies microRNAs and other small noncoding RNAs from..., Tuddenham [/bib_ref]. An miRNA encoded by HHV-6A (miR-U86) targets the HHV-6A IE gene U86, thereby regulating lytic replication, as revealed by growth analyses of mutant viruses [fig_ref] Figure 5: In rhinoviruses infection, cellular microRNAs play anti-viral responses against viruses [38], human... [/fig_ref] [bib_ref] A human herpesvirus 6A-encoded microRNA: role in viral lytic replication, Nukui [/bib_ref]. However, HHV-6B encodes at least four pre-miRNAs at two positions within the genome in an antisense orientation related to predicted HHV-6B-specific genes [bib_ref] Small RNA deep sequencing identifies microRNAs and other small noncoding RNAs from..., Tuddenham [/bib_ref]. These data suggest that HHV-6, like other herpesviruses, encodes its own miRNAs, but the precise function of these miRNAs in HHV6B requires further investigation.
# Conclusions
This comprehensive review attempts to highlight the role of miRNAs in replication as well as pathogenesis of respiratory viral infections. miRNAs modulate a variety of cellular processes by regulating multiple targets, promoting or inhibiting the development of viral infection [bib_ref] Roles of microRNAs in the life cycles of mammalian viruses, Gottwein [/bib_ref]. Increasing evidence regarding disrupted miRNA expression and function following viral infection makes them promising targets for therapeutic interventions. The development of miRNA-based therapy for respiratory infection is less advanced compared with other viral infections, such as hepatitis C. Improved knowledge on the cross-talk between host cells and viruses should increase our understanding of the molecular basis for viral pathogenesis and may enable us to develop better therapeutic strategies [bib_ref] MicroRNAs in viral replication and pathogenesis, Dykxhoorn [/bib_ref].
Therapeutic modulation of miRNAs can be achieved through miRNA inhibitors to disrupt miRNA function or miRNA mimics to increase miRNA function. The application of miRNA-based therapies is in its beginning, and important difficulties remain. A significant barrier to miRNA-based therapy is the development of essential pharmaceutical strategies for targeted delivery to specific sites with minimum toxicity [bib_ref] MicroRNAs in disease and potential therapeutic applications, Soifer [/bib_ref] [bib_ref] Inhaling medicines: delivering drugs to the body through the lungs, Patton [/bib_ref]. In support of this, novel nanotechnologies and delivery methods are under development for efficient and effective delivery [bib_ref] Inhaling medicines: delivering drugs to the body through the lungs, Patton [/bib_ref] [bib_ref] Chitosan nanoparticles as a potential nonviral gene delivery for HPV-16 E7 into..., Tahamtan [/bib_ref].
Alongside the critical role of miRNAs in the regulation of viral respiratory infection and their potential to be targeted by new therapeutics, caution must be taken because excessive inhibition or overexpression of miRNAs might predispose patients to cellular abnormalities, impaired immunity, or even cancer. The relevance of miRNAs in viral infection has been proven broadly; however, the exact role of each miRNA on viral pathogenesis remains to be determined, and future studies are warranted. Enhancing the knowledge on miRNAs may open opportunities to use them in clinical practice in order to develop more accurate and powerful diagnostic and therapeutic strategies.
[fig] Figure 2: Influenza infection of airway epithelial cells induces or inhibits certain cellular microRNAs (miRNAs) expression in favor of viral replication, pathogenesis, and also suppress anti-viral responses. However, certain cellular miRNAs can inhibit replication of influenza in infected cells, and certain miRNAs play important roles in priming airway cells for repair and regeneration following influenza infection [/fig]
[fig] Figure 3: Following RSV respiratory infection, an altered expression profile of certain cellular miRNAs, specifically immune-associated miRNAs, occurs in order to inhibit viral replication and preserve the airway epithelial barrier; meanwhile, the virus induces or inhibits the expression of other miRNAs that favor viral replication. The RSV G protein enhances let-7f, the RSV NS1/NS2 proteins decrease miR-30b and enhance let-7i, and RSV infection decrease miR-221, which is an advantage for the virus [/fig]
[fig] Figure 4: Coronaviruses interact with the host cell at the onset of infection and induces several changes in host cellular microRNAs (miRNAs) expression profile to their own advantage; severe acute respiratory syndrome-coronavirus (SARS-CoV) uses cellular miRNAs machinery to evade immune elimination [64]; in Middle East respiratory syndrome-coronavirus (MERS-CoV), host cells miRNAs would be an anti-viral therapeutic agent [111], and the N protein of OC43-coronavirus (OC43-CoV) causes potentiation of nuclear factor kappa B (NF-kB) activation via binding to its negative regulator miR-9 [65] [/fig]
[fig] Figure 5: In rhinoviruses infection, cellular microRNAs play anti-viral responses against viruses [38], human metapneumovirus (HMPV) M2-2 regulates the host cell microRNAs response to infection [/fig]
[fig] 52: Huang L, Ma J, Sun Y, et al. Altered splenic miRNA expression profile in H1N1 swine influenza. Archives of Virology 2015; 160: 979-85. DOI:10.1007/s00705-015-2351-0. 53. Buggele WA, Johnson KE, Horvath CM. Influenza A virus infection of human respiratory cells induces primary microRNA expression. Journal of Biological Chemistry 2012; 287: 31027-40. DOI:10.1074/jbc. M112.387670. 54. Meliopoulos VA, Andersen LE, Brooks P, et al. MicroRNA regulation of human protease genes essential for influenza virus replication. PLoS One 2012; 7e37169: . DOI:10.1371/journal.pone.0037169. 55. Loveday E-K, Diederich S, Pasick J, et al. Zhong Z, Dong Z, Yang L, et al. miR-21 induces cell cycle at S phase and modulates cell proliferation by down-regulating hMSH2 in lung cancer. Journal of Cancer Research and Clinical Oncology 2012; 138: 1781-8. DOI:10.1007/s00432-012-1287-y. 57. Li J, Donath S, Li Y, et al. miR-30 regulates mitochondrial fission through targeting p53 and the dynamin-related protein-1 pathway. PLoS Genetics 2010; 6e1000795: . DOI:10.1371/journal.pgen.1000795. 58. Lam W-Y, Yeung AC-M, Ngai KL-K, et al. Effect of avian influenza A H5N1 infection on the expression of microRNA-141 in human respiratory epithelial cells. BMC Microbiology 2013; 13: 104. DOI:10.1186/ 1471-2180-13-104. 59. Li Y, Chan EY, Li J, et al. MicroRNA expression and virulence in pandemic influenza virus-infected mice. Journal of Virology 2010; 84: 3023-32. DOI:10.1128/ viral and host small RNAs during human cytomegalovirus infection. Journal of Virology 2012; 86: 226-35. DOI:10.1128/ JVI.05903-11. 78. Kim S, Lee S, Shin J, et al. Human cytomegalovirus microRNA miR-US4-1 inhibits CD8+ T cell responses by targeting the aminopeptidase ERAP1. Nature Immunology 2011; 12: 984-91. DOI:10.1038/ni.2097. 79. Hook L, Hancock M, Landais I, et al. Cytomegalovirus microRNAs. Current Opinion Virology 2014; 7: 40-6. DOI:10.1016/j. coviro.2014.03.015. 80. Huang Y, Qi Y, Ma Y, et al. Immunology 1997; 92: 138-45. 83. Jiang S, Qi Y, He R, et al. Human cytomegalovirus microRNA miR-US25-1-5p inhibits viral replication by targeting multiple cellular genes during infection. Gene 2015; 570: 108-14. DOI:10.1016/j. gene.2015.06.009. 84. Shen Z-Z, Pan X, Miao L-F, et al. Comprehensive analysis of human cytomegalovirus microRNA expression during lytic and quiescent infection. PLoS One 2014; 9e88531: . DOI:10.1371/journal. pone.0088531. 85. Grey F, Meyers H, White EA, et al. A human cytomegalovirus-encoded microRNA regulates expression of multiple viral genes involved in replication. PLoS Pathogens 2007; 3e163: . DOI:10.1371/journal. ppat.0030163. 86. Poole E, Avdic S, Hodkinson J, et al. Latency-associated viral interleukin-10 (IL-10) encoded by human cytomegalovirus modulates cellular IL-10 and CCL8 secretion during latent infection through changes in the cellular microRNA hsa-miR-92a. Journal of Virology 2014; 88: 13947-55. DOI:10.1128/JVI.02424-14. 87. Lee SH, Kalejta RF, Kerry J, et al. BclAF1 restriction factor is neutralized by proteasomal degradation and microRNA repression during human cytomegalovirus infection. Proceedings of the National Academy of Sciences USA 2012; 109: 9575-80. DOI:10.1073/pnas.1207496109. 88. Gambarino S, Mantovani S, Astegiano S, et al. Lower respiratory tract viral [/fig]
[table] Table 1: miRNAs effects in respiratory viral infection [/table]
[table] 30: Cullen BR. MicroRNAs as mediators of viral evasion of the immune system. Nature DOI:10.1038/ni.2537. 31. Bakre A, Mitchell P, Coleman JK, et al. Re-32. Zhang X, Dong C, Sun X, et al. Induction of the cellular miR-29c by influenza virus inhibits the innate immune response through protection of A20 mRNA. Biochemical and Biophysical Research Communications 2014; 450: 755-61. DOI:10.1016/j. bbrc.2014.06.059. 33. Stewart CR, Marsh GA, Jenkins KA, et al. Promotion of Hendra virus replication by microRNA 146a. Journal of Virology 2013; 87: 3782-91. DOI:10.1128/JVI.01342-12. 34. Müller S, Imler J-L. Dicing with viruses: microRNAs as antiviral factors. Immunity DOI:10.1016/j. immuni.2007.07.003. 35. Sullivan CS, Ganem D. MicroRNAs and viral infection. Molecular Cell 2005; 20: 3-7. DOI:10.1016/j.molcel.2005.09.012. 36. Cullen BR. How do viruses avoid inhibition by endogenous cellular microRNAs? PLoS Pathogens 2013; 9e1003694: . DOI:10.1371/journal.ppat.1003694. 37. Song L, Liu H, Gao S, et al. Cellular microRNAs inhibit replication of the H1N1 influenza A virus in infected cells. Journal of Virology 2010; 84: 8849-60. DOI:10.1128/JVI.00456-10. 38. Bondanese VP, Francisco-Garcia A, Bedke N, et al. Identification of host miRNAs that may limit human rhinovirus replication. World Journal of Biological Chemistry 2014; 5: 437. DOI:10.4331/ wjbc.v5.i4.437. 39. Ma YJ, Yang J, Fan XL, et al. Cellular microRNA let-7c inhibits M1 protein expression of the H1N1 influenza A virus in infected human lung epithelial cells. Journal of Cellular and Molecular Medicine 2012; 16: 2539-46. DOI:10.1128/JVI.00456-10. 40. Cristiano F, Veltri P. Methods and Techniques for miRNA data analysis. Methods in Molecular Biology 2016; 1375: 11-23. DOI:10.1007/7651_2015_238. 41. Svoboda P. A toolbox for miRNA analysis. FEBS Letters 2015; 589: 1694-701. DOI:10.1016/j.febslet.2015.04.054. 42. Salimi V, Tavakoli-Yaraki M, Yavarian J, et al. Prevalence of human respiratory syncytial virus circulating in Iran. Journal of Infection and Public Health 2015; S1876-0341: 00102-1. DOI: 10.1016/j.jiph.2015.05.005. 43. Kurai D, Saraya T, Ishii H, et al. Virus-induced exacerbations in asthma and COPD. Frontiers in Microbiology 2013; 4: 293. DOI:10.3389/fmicb.2013.00293. 44. del Valle MJ, Cornejo-Tapia A, Weilg P, et al. Incidence of respiratory viruses in Peruvian children with acute respiratory infections. Journal of Medical Virology 2015; 87: 917-24. DOI:10.1002/jmv.24159. 45. Tahamtan A, Tavakoli-Yaraki M, Rygiel TP, et al. Effects of cannabinoids and their receptors on viral infections. Journal of DOI:10.1002/jmv.24292. 46. Salimi V, Hennus MP, Mokhtari-Azad T, et al. Opioid receptors control viral replication in the airways. Critical Care Medicine CCM.0b013e31826767a8. 47. Legand A, Briand S, Shindo N, et al. Addressing the public health burden of respiratory viruses: the Battle against Respiratory Viruses (BRaVe) Initiative. Future Virology 2013; 8: 953-68. DOI:10.2217/fvl.13.85. 48. Li Y, Li J, Belisle S, et al. Differential microRNA expression and virulence of avian, 1918 reassortant, and reconstructed 1918 influenza A viruses. Virology 2011; 49. Guan Z, Shi N, Song Y, et al. Induction of the cellular microRNA-29c by influenza virus contributes to virus-mediated apoptosis through repression of antiapoptotic factors BCL2L2. Biochemical and Biophysical Research Communications 2012; 425: 662-7. DOI:10.1016/j.bbrc.2012.07.114. 50. Othumpangat S, Noti JD, Beezhold DH. Lung epithelial cells resist influenza A infection by inducing the expression of cytochrome c oxidase VIc which is modulated by miRNA 4276. Virology 2014; 468: 256-64. DOI:10.1016/j.virol.2014.08.007. 51. Terrier O, Textoris J, Carron C, et al. Host microRNA molecular signatures associated with human H1N1 and H3N2 influenza A viruses reveal an unanticipated antiviral activity for miR-146a. The Journal of General Virology 2013; 94: 985-95. DOI:10.1099/vir.0.049528-0. [/table]
|
Efficacy and safety of fixed-dose combination therapy with olmesartan medoxomil and rosuvastatin in Korean patients with mild to moderate hypertension and dyslipidemia: an 8-week, multicenter, randomized, double-blind, factorial-design study (OLSTA-D RCT: OLmesartan rosuvaSTAtin from Daewoong)
# Introduction
The coexistence of hypertension and dyslipidemia, which are central to the pathogenesis of coronary heart disease, has been reported to be prevalent. [bib_ref] Risk stratification in hypertension: new insights from the Framingham study, Kannel [/bib_ref] [bib_ref] Should angiotensin II receptor blockers and statins be combined?, Nickenig [/bib_ref] [bib_ref] Ethnic and sex differences in the prevalence, treatment, and control of dyslipidemia..., O'meara [/bib_ref] [bib_ref] Prevalence and prognostic significance of hypercholesterolemia in men with hypertension. Prospective data..., Stamler [/bib_ref] The risk of coronary submit your manuscript | www.dovepress.com
## Dovepress
## Dovepress
## 2600
Park et al heart disease with the coexistence of hypertension and dyslipidemia has been reported to be higher than the sum of the risks of coronary heart disease with each of the component factors. [bib_ref] Prevalence and prognostic significance of hypercholesterolemia in men with hypertension. Prospective data..., Stamler [/bib_ref] [bib_ref] A population at risk. Prevalence of high cholesterol levels in hypertensive patients..., Castelli [/bib_ref] [bib_ref] Evaluating the impact of population and high-risk strategies for the primary prevention..., Emberson [/bib_ref] As cardiovascular risk factors interact with each other, comprehensive control of both blood pressure (BP) and blood cholesterol level is effective for reducing the risk of future cardiovascular events. [bib_ref] Evaluating the impact of population and high-risk strategies for the primary prevention..., Emberson [/bib_ref] [bib_ref] Treatment with drugs to lower blood pressure and blood cholesterol based on..., Jackson [/bib_ref] In clinical practice, the pill burden in patients with both hypertension and dyslipidemia can result in poor adherence and persistence with the prescribed drugs. [bib_ref] Predictors of adherence with antihypertensive and lipid-lowering therapy, Chapman [/bib_ref] A fixed-dose combination (FDC) of a BP-lowering agent and statin could improve adherence and persistence in patients with multiple risk factors, resulting in a reduction of the risks of future cardiovascular events.
In our previous study, the coadministration of ol mesartan medoxomil (40 mg) and rosuvastatin (20 mg) did not significantly influence each other's pharmacokinetics without adverse events (AEs). [bib_ref] Pharmacokinetic interaction between rosuvastatin and olmesartan: a randomized, openlabel, 3-period, multiple-dose crossover..., Roh [/bib_ref] In healthy volunteers, FDC therapy with olmesartan medoxomil (40 mg) and rosuvastatin (20 mg) had a similar pharmacokinetic profile to that of coadministration of each drug as individual tablets. [bib_ref] Pharmacokinetics of rosuvastatin/olmesartan fixed-dose combination: a single-dose, randomized, open-label, 2-period crossover study..., Son [/bib_ref] The present study aimed to evaluate the efficacy and safety of FDC therapy with olmesartan medoxomil (40 mg) and rosuvastatin in Korean patients with mild to moderate hypertension and dyslipidemia.
## Materials and methods study design
This was a randomized, double-blind, factorial-design study performed at 25 locations in Korea between September 2012 and May 2013 . This study was designed to adhere to the Korean Good Clinical Practice guidelines, related regulations in Korea, and the Declaration of Helsinki, and it was approved by the Ministry of Food and Drug Safety, and the institutional review boards of each of the participating institutions (Table S1) (ClinicalTrials.gov Identifier: NCT01764295).
Screening was performed after patients signed a written informed consent form for participation in this study. After assessing the screening results of the patients, those who satisfied the inclusion criteria underwent therapeutic lifestyle change for a period of .4 weeks. After the therapeutic lifestyle change period, central laboratory tests and BP measurements for final decisions were performed at the baseline visit. After a qualification period of ,1 week, the selected patients were randomly allocated to the following four groups: the FDC therapy group , Crestor ® , AstraZeneca plc, London, UK); and placebo group. Each placebo tablet had an appearance and an odor identical to that of the active tablets. The pills were completely indistinguishable. All randomly assigned subjects took three tablets of investigational drugs orally once a day for 8 weeks at the same time each day. For randomization, this study used a stratified block randomization method stratified according to the low-density lipoprotein cholesterol (LDL-C) (100 mg/dL # LDL-C ,130 mg/dL, 130 mg/dL # LDL-C ,160 mg/dL, LDL-C $160 mg/dL) level and diastolic blood pressure (DBP) (90 mmHg # DBP ,100 mmHg, DBP $100 mmHg, in case of subjects with diabetes or chronic renal disease, 80 mmHg # DBP ,90 mmHg, DBP $90 mmHg). The randomization code was generated with the proc plan procedure using SAS version 9.2 (SAS Institute Inc., Cary, NC, USA) by an independent statistician of the contract research organization. The independent statistician made an extra "randomization list with investigational drug number". The investigators and the pharmacists used this list for prescription of investigational drugs. All the investigators, participants, and study staffs remained blinded to treatment group until study completion.
Patients stopped taking any antihypertensive drugs at least 2 weeks before randomization and stopped taking lipidlowering drugs during the entire therapeutic lifestyle change period. In addition, antihypertensive and lipid-lowering drugs that could interact with the study drugs were discontinued during the treatment period. During the study period, the patients visited the participating institutions five times as follows: screening visit, baseline visit, randomization visit, and visits at weeks 4 and 8 after starting treatment. The following procedures were carried out at each visit: physical examination, vital signs (DBP/systolic blood pressure [SBP], temperature, and pulse), laboratory tests (hematology, chemistry, and urinalysis), assessment of compliance, and AEs.
When the subject showed signs or symptoms of hypotension with SBP ,90 mmHg or DBP ,60 mmHg, hypertension with SBP $180 mmHg or DBP $110 mmHg, and abnormal results values of liver function (aspartate aminotransferase and alanine aminotransferase three times greater than upper limit of normal level), the subject had to discontinue this study for his or her safety.
## Inclusion and exclusion criteria
The study recruited patients aged $20 years with mild to moderate essential hypertension and dyslipidemia, as defined by the Seventh Report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure .
Patients were excluded if they had secondary hypertension (medical history of secondary hypertension or suspected secondary hypertension by physician) or dyslipidemia; hypersensitivity to olmesartan medoxomil or rosuvastatin; uncontrolled diabetes mellitus (hemoglobin A1c $9% or fasting plasma glucose level $160 mg/dL); myocardial infarction, transient ischemic attack, percutaneous transluminal coronary angioplasty, or unstable angina within the previous 6 months; severe heart failure (New York Heart Association class 3 and 4); thyroid stimulating hormone levels $1.5 times the upper normal limit; creatinine level $1.5 times the upper normal limit; creatinine kinase, aspartate aminotransferase, and alanine aminotransferase levels $2 times the upper normal limits; triglyceride levels $400 mg/dL; or any disease that could influence the study results.
## Objectives and outcome measures
The primary objectives were to determine the superiority of FDC therapy over olmesartan medoxomil (40 mg) for the percentage change from baseline in the LDL-C level and the superiority of FDC therapy over rosuvastatin (20 mg) for the change from baseline in DBP at week 8.
The secondary objectives were to compare the FDC therapy to olmesartan medoxomil (40 mg) for the change from baseline in DBP and the FDC therapy to rosuvastatin (20 mg) for the percentage change from baseline in the LDL-C level at week 8. The additional secondary objectives were to compare the FDC therapy to olmesartan medoxomil (40 mg) and rosuvastatin (20 mg) for the percentage change from baseline in the total cholesterol, triglyceride, and highdensity lipoprotein cholesterol levels at weeks 4 and 8; the change from baseline in SBP at weeks 4 and 8; and the percentage of patients who achieved the treatment goals (LDL-C ,160 mg/dL, 130 mg/dL, 100 mg/dL each category according to the risk factors and a 10-year risk assessment, SBP/DBP ,140/90 mmHg, in case of subjects with diabetes or chronic renal disease, 130/80 mmHg) defined by the NCEP ATP III and JNC VII at week 8.
For reliability evaluations, the percentage change from baseline in the LDL-C level and the change from baseline in DBP were compared between FDC therapy and placebo at week 8.
For safety evaluations, the dates of onset of AEs and termination, severity of AEs, actions taken for the AEs, and relationships of the AEs with the study products were assessed at each visit. In addition, abnormal vital signs, laboratory test results (including hematology, biochemistry, and urinalysis), physical examination results, and echocardiography results were recorded.
# Statistical analysis
The hypotheses being tested were that the FDC therapy was superior to olmesartan medoxomil (40 mg) in reducing the LDL-C level and superior to rosuvastatin (20 mg) in reducing DBP. The expected difference of the mean percentage change from baseline in the LDL-C level between FDC therapy and olmesartan medoxomil (40 mg) was −53.8% (standard deviation: 20%) and the expected difference of the mean change from baseline in DBP between FDC therapy and rosuvastatin (20 mg) was −6 mmHg (standard deviation: 8.7 mmHg). Finally, the sample sizes were calculated using the percent change of LDL-C level and the change in DBP. The sample size identified for assessing the change in DBP, which was larger, was selected. For collecting more safety data of FDC therapy, the randomization ratio was set into 2:1:1:1. According to the randomization ratio of 2:1:1:1 and a 20% drop-out rate, a sample size of 150 patients was calculated (60 patients in the FDC therapy group and 30 patients each in the olmesartan medoxomil, rosuvastatin, and placebo groups). As both the hypotheses required significant findings Inclusion criteria according to risk factors and a 10-year risk assessment Category
## Ldl-c level (mg/dl) dbp (mmhg)
1. Person without risk factors a other than hypertension and dyslipidemia 160-250 90-109 (patients with DM, cKD: 80-99) 2. Person with more than one risk factor a other than hypertension and dyslipidemia and with a ,10% risk in the 10-year risk assessment
## 160-250
3. Person with more than one risk factor a other than hypertension and dyslipidemia and with a 10%-20% risk in the 10-year risk assessment
## 130-250
4. Person with coronary heart disease or equivalent b or with a .20% risk in the 10-year risk assessment
## 100-250
Notes: a Risk factors: a) Cigarette smoking; b) HDL-C level ,40 mg/dL; c) family history of premature coronary heart disease (male first degree relative ,55 years; female first degree relative ,65 years); and d) age (males $45 years; females $55 years). HDL-C level $60 mg/dL counts as a "negative" risk factor; its presence excludes one risk factor from the total count. b Patients with carotid artery disease, peripheral vascular disease, abdominal aortic aneurysm, and type 2 DM. Abbreviations: CKD, chronic kidney disease; DBP, diastolic blood pressure; DM, diabetes mellitus; HDL-C, high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol. for acceptance, each individual significance level was set at 5% for the entire hypothesis and the statistical power for each hypothesis was set at 80%. Continuous data were summarized using descriptive statistics, and the treatment groups were compared using analysis of covariance, with baseline values, stratification factors (risk factors and BP), and drug interaction variables as covariates. Categorical data were analyzed using logistic regression models, with stratification factors and drug interaction variables as covariates. All analyses were two-sided, and the significance level was α=0.05. The analyses were performed using SAS version 9.2 (SAS Institute Inc.).
# Results
## Patients' characteristics
A total of 423 patients underwent the screening examination, and of these patients, 183 who were found to be suitable for this study were randomized. Of these 183 patients, 181 patients were administered the investigational products. However, 19 patients were excluded based on the inclusion/ exclusion criteria. Therefore, 162 patients completed the treatment and were included in the full analysis set [fig_ref] Figure 1: Study flowchart [/fig_ref].
The demographics of the full analysis set according to the treatment group are presented in [fig_ref] Table 2: Demographics and baseline characteristics prior to randomization in the full analysis set [/fig_ref]. The mean (standard deviation) age of the patients was 61.4 (7.8) years, and the mean body mass index was 25.4 (2.7) kg/m 2 . The mean SBP was 150.5 (13.5) mmHg, and the mean DBP was 92.6 (6.6) mmHg. The mean LDL-C level was 154.5 (31.7) mg/dL, and the mean high-density lipoprotein cholesterol, triglyceride, and total cholesterol levels were 50.0 (11.4) mg/dL, 147.7 (67.2) mg/dL, and 230.2 (36.3) mg/dL, respectively. There were no significant differences in demographic characteristics, except for family history of premature coronary heart disease, among the treatment groups (P=0.0118). The least square mean percentage changes (standard error) from baseline in the LDL-C level 8 weeks after treatment were −52.3% (2.8%) in the FDC therapy group, −0.6% (3.5%) in the olmesartan medoxomil group, and −46.9% (3.5%) in the rosuvastatin group. The difference between the FDC and olmesartan medoxomil groups was −51.7% (4.1%) (95% confidence interval [CI]: −59.8% to −43.6%), and the percentage change was significantly higher in the FDC therapy group than in the olmesartan medoxomil group (P,0.0001). The difference between the FDC therapy and rosuvastatin groups was −5.4% (4.1%) (95% CI: −13.5% to 2.7%), and the percentage change was not significantly different between the FDC and rosuvastatin groups (P=0.1864).
The percentage changes in LDL-C levels at weeks 4 and 8 are presented in [fig_ref] Table 3: Changes in the low-density lipoprotein cholesterol level at weeks 4 and 8... [/fig_ref] , and the percentage changes in other lipid parameters at weeks 4 and 8 are presented in [fig_ref] Table 4: Least square mean percentage change from baseline in the total cholesterol, triglyceride... [/fig_ref]. The treatment goal was achieved at 8 weeks in 90.2% (55/61) of patients from the FDC therapy group, 16.7% (6/36) of patients from the olmesartan medoxomil group, 86.1% (31/36) of patients from the rosuvastatin group, and 17.2% (5/29) of patients from the placebo group. The percentage of patients who achieved the treatment goal was significantly higher in the FDC therapy group than in the olmesartan medoxomil and placebo groups (both P,0.0001). There was no significant difference in the percentage of patients between the FDC therapy and rosuvastatin groups (P=0.5111, [fig_ref] Figure 2: Percentage of patients who achieved the treatment goals of low-density lipoprotein cholesterol... [/fig_ref].
## Blood pressure
The least square mean changes (standard error) from baseline in DBP at 8 weeks after treatment were −10.4 (1.2) mmHg in the FDC therapy group, 0.1 (1.6) mmHg in the rosuvastatin group, and −8.1 (1.5) mmHg in the olmesartan medoxomil group. The difference between the FDC therapy and rosuvastatin groups was −10.5 (1.8) mmHg (95% CI: −14.1 to −6.9 mmHg), and the change was significantly higher in the FDC therapy group than in the The changes in DBP at 4 weeks and the changes in SBP at 4 and 8 weeks were similar to the changes in DBP at 8 weeks [fig_ref] Table 5: Changes in blood pressure at weeks 4 and 8 in the full... [/fig_ref]. The treatment goal was achieved at 8 weeks in 57.4% (35/61) of patients from the FDC therapy group, 11.1% (4/36) of patients from the rosuvastatin group, 41.7% (15/36) of patients from the olmesartan medoxomil group, and 20.7% (6/29) of patients from the placebo group. The percentage of patients who achieved the treatment goal was significantly higher in the FDC therapy group than in the rosuvastatin and placebo groups (P,0.0001, P=0.0018, respectively). There was no significant difference in the percentage of patients between the FDC and olmesartan medoxomil groups (P=0.1360, [fig_ref] Figure 2: Percentage of patients who achieved the treatment goals of low-density lipoprotein cholesterol... [/fig_ref]. There were no significant differences in the incidences of AEs and adverse drug reactions (ADRs) among the treatment groups (P=0.9202, and P=0.5990, respectively). Most of the AEs were mild, and severe AEs were not reported in any of the treatment groups.
Serious AEs occurred in two patients (1.1%, two cases): one patient had myocardial infarction and was from the olmesartan medoxomil group, while the other patient who had subarachinoid hemorrhage was from the rosuvastatin group. The investigational products were immediately discontinued in these patients. However, all serious AEs were not likely related to investigational drugs and they were resolved without sequelae.
A total of five patients experienced eight ADRs during the study period. In the FDC therapy group, two patients reported ADRs. Of these two patients, one had increases in aspartate aminotransferase and alanine aminotransferase levels and the other had an increase in blood creatinine levels and a decrease in creatinine clearance. All the ADRs were expected side effects of the approved drugs.
# Discussion
The present study demonstrated that FDC therapy with olmesartan medoxomil (40 mg) and rosuvastatin (20 mg) was highly effective for achieving the therapeutic goals of the LDL-C level and BP. In the reduction of the LDL-C level, the effectiveness of FDC therapy was not different from that of rosuvastatin (20 mg), and in the reduction of BP, the effectiveness of FDC therapy was not different from that of olmesartan medoxomil (40 mg). Additionally, we found that FDC therapy was generally safe and well tolerated.
Olmesartan medoxomil is a selective angiotensin II type 1 receptor blocker (ARB) with proven BP-lowering efficacy. [bib_ref] Antihypertensive activity of angiotensin II AT1 receptor antagonists: a systematic review of..., Fabia [/bib_ref] The antihypertensive efficacy of ARBs has been shown to be at least equivalent to the efficacies of other major classes of antihypertensive agents but with a better tolerability profile. [bib_ref] Angiotensin II type 1 receptor blockers, Burnier [/bib_ref] Several studies have demonstrated that ARBs have positive effects on left ventricular hypertrophy, endothelial dysfunction, and atherosclerosis, suggesting that ARBs offer cardiovascular protective benefits in addition to their favorable effects on BP. [bib_ref] Acute and chronic angiotensin-1 receptor antagonism reverses endothelial dysfunction in atherosclerosis, Prasad [/bib_ref] [bib_ref] Effect of AT1 receptor blockade on endothelial function in essential hypertension, Klingbeil [/bib_ref] Olmesartan medoxomil has a more rapid onset of action than that of other ARBs, with significant improvements in efficacy. [bib_ref] Antihypertensive activity of angiotensin II AT1 receptor antagonists: a systematic review of..., Fabia [/bib_ref] The ability of olmesartan medoxomil to effectively reduce BP suggests that it is a good therapeutic option for intensive treatment in patients with mild to moderate hypertension.Statins are usually used to treat dyslipidemia and manage patients with ischemic heart disease. However, with the completion of many large clinical trials on statins over the past 10 years,
## 2607
Efficacy and safety of fixed-dose combination their use has been extended to preventive treatment for a variety of cardiovascular diseases. [bib_ref] Should angiotensin II receptor blockers and statins be combined?, Nickenig [/bib_ref] Rosuvastatin is more effective than other statins for achieving LDL-C goals and producing favorable changes in the atherogenic lipid profile. [bib_ref] Comparison of the efficacy and safety of rosuvastatin versus atorvastatin, simvastatin, and..., Jones [/bib_ref] [bib_ref] Efficacy and safety of rosuvastatin and atorvastatin in patients with hypercholesterolemia and..., Schwartz [/bib_ref] Previous studies have shown that statins have direct effects on plaque stability, nitric oxide metabolism, inflammation, endothelial function, and oxidative stress. [bib_ref] Pleiotropic effects of HMG-CoA reductase inhibitors, Werner [/bib_ref] [bib_ref] Interrelationship of free oxygen radicals and endothelial dysfunction -modulation by statins, Wassmann [/bib_ref] Additionally, statins have been shown to significantly reduce cardiovascular mortality and morbidity in patients at risk for cardiovascular diseases. [bib_ref] Pleiotropic effects of HMG-CoA reductase inhibitors, Werner [/bib_ref] [bib_ref] Effects of statins on the vasculature: Implications for aggressive lipid management in..., Sowers [/bib_ref] The interplay between hypertension and dyslipidemia acts through the renin-angiotensin system to increase cardiovascular risk. Hypertension and dyslipidemia result in the release of angiotensin II, which acts on angiotensin 1 receptors. Activation of angiotensin 1 receptors stimulates NADH oxidase production in endothelial cells, resulting in the generation of reactive oxygen species in vascular cells and eventually endothelial dysfunction and decreased nitric oxide production. [bib_ref] The AT(1)-type angiotensin receptor in oxidative stress and atherogenesis: part I: oxidative..., Nickenig [/bib_ref] [bib_ref] The AT(1)-type angiotensin receptor in oxidative stress and atherogenesis: Part II: AT(1)..., Nickenig [/bib_ref] Combinations of ARBs and statins could be atheroprotective and effective in improving endothelial function through their synergistic mode of action on angiotensin 1 receptors, resulting in the reduction of cardiovascular morbidity and mortality. 2 FDC therapy with olmesartan medoxomil (40 mg) and rosuvastatin (20 mg) could be highly effective for the prevention of cardiovascular events through cardiovascular protective benefits beyond comprehensive control of both BP and blood cholesterol and could potentially increase treatment adherence in patients prescribed long-term polymedication.
Olmesartan medoxomil is not metabolized by the cytochrome P450 system and has no effect on P450 enzymes. Rosuvastatin clearance is not dependent on metabolism by cytochrome P450 3A4 to a clinically significant extent. Thus, interactions with drugs that inhibit or induce those enzymes, or are metabolized by these enzymes are not expected. A previous study showed that FDC therapy with olmesartan medoxomil (40 mg) and rosuvastatin (20 mg) has a similar pharmacokinetic profile to that of coadministration of each drug as individual tablets, without serious AEs. [bib_ref] Pharmacokinetics of rosuvastatin/olmesartan fixed-dose combination: a single-dose, randomized, open-label, 2-period crossover study..., Son [/bib_ref] These results suggested that FDC therapy could be used interchangeably with the conventional formulation of the coadministration of each drug separately. In the present study, we demonstrated that the efficacy and safety of FDC therapy with olmesartan medoxomil (40 mg) and rosuvastatin (20 mg) were similar to those of each drug in the combination in patients with both hypertension and dyslipidemia.
The relatively small sample size would be a limitation of the present study. Further investigation of a larger patient population over a longer period will be needed to confirm the clinical benefit.
# Conclusion
For patients who have hypertension and dyslipidemia concomitantly, FDC therapy with olmesartan medoxomil (40 mg) and rosuvastatin (20 mg) is a good therapeutic option with appropriate efficacy and safety. Such a combo-pill may help enhance the compliance of the patients with large pill burden due to comorbidities.
[fig] Figure 1: Study flowchart. Abbreviations: BP, blood pressure; FAS, full analysis set; FDC, fixed dose combination; IP, investigational product; SAEs, serious adverse events. Drug Design, Development and Therapy 2016:10 submit your manuscript | www.dovepress. [/fig]
[fig] Figure 2: Percentage of patients who achieved the treatment goals of low-density lipoprotein cholesterol levels and blood pressure at week 8 in the full analysis set. Notes: (A) Percentage of patient who achieved the treatment goal of low-density lipoprotein cholesterol levels defined by the National Cholesterol Education Program Adult Treatment Panel III. Goal was defined as low-density lipoprotein cholesterol level ,100, 130, or 160 mg/dL depending on the risk profile. (B) Percentage of patients who achieved the treatment goal of blood pressure defined by the Seventh Report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure. Goal was defined as DBP ,90 mmHg (or ,80 mmHg in patients with diabetes mellitus or chronic kidney disease). Abbreviations: DBP, diastolic blood pressure; FDC, fixed-dose combination. [/fig]
[table] Table 2: Demographics and baseline characteristics prior to randomization in the full analysis set [/table]
[table] Table 3: Changes in the low-density lipoprotein cholesterol level at weeks 4 and 8 in the full analysis set [/table]
[table] Table 4: Least square mean percentage change from baseline in the total cholesterol, triglyceride and high-density lipoprotein cholesterol levels at weeks 4 and 8 in the full analysis set [/table]
[table] Table 5: Changes in blood pressure at weeks 4 and 8 in the full analysis set [/table]
[table] Table 6: summary of aes and adverse drug reactions in the safety set Notes: *Data are presented as number of patients (%) [number of events] and the denominator for % is the number of patients in the column. **Chi-square or Fisher's exact test. Abbreviations: AE, adverse event; ALT, alanine aminotransferase; AST, aspartate aminotransferase; Ccr, creatinine clearance; FDC, fixed-dose combination; SAEs, serious adverse events. Drug Design, Development and Therapy 2016:10 submit your manuscript | www.dovepress.com [/table]
|
COVID-19–Associated Bone Marrow Necrosis—A Case Report
# Introduction
The novel coronavirus disease 2019 (COVID-19 pneumonia) patients typically present with fever, shortness of breath, and cough. However, musculoskeletal complaints are seldom encountered as red flags. Here, we report the first presumptive case of COVID-19-associated vertebral (and also pelvic and femoral) bone marrow necrosis, a rare disorder that has been associated with malignancies, sickle cell anemia, and other infections but is yet to be illustrated through the role of imaging in COVID-19-recovered patients.
It is imperative to begin with a few anatomical pearls pertinent to our case for a comprehensive understanding of Keywords ► avascular necrosis ► bone marrow necrosis ► chemical shift imaging ► computed tomography ► COVID-19 ► disseminated intravascular coagulation ► opposed phase ► magnetic resonance imaging ► opposed phase ► positron emission tomography ► signal intensity index
# Abstract
We report, herein, a rare case of vertebral bone marrow necrosis in a patient at 1-month post-novel coronavirus disease 2019 (COVID-19) pneumonia complicated with disseminated intravascular coagulation (DIC). The commonly observed radiological features on the imaging modalities like computed tomography (CT), magnetic resonance imaging (MRI), and 18-F fluorodeoxyglucose positron emission tomography (FDG PET) have been discussed here followed by a brief discussion on the role of inphase and opposed-phase imaging in differentiating the disease from malignant infiltrative pathologies. Histopathological findings on bone marrow smear that confirm the diagnosis have also been illustrated.
## Case presentation
A 72-year-old male diabetic and hypertensive patient was admitted at an outside hospital, in mid-February of 2021, with the complaints of fever, cough, anosmia, and ageusia after his oral and nasopharyngeal swab reverse-transcription polymerase chain reaction (RT-PCR) assay was found to be positive. As per protocol, he received high-dose steroids, two doses of remdesivir which was discontinued on account of rising liver function tests. After spending 5 days there, he was shifted to our hospital where he received two doses of tocilizumab and was continued on high-dose intravenous dexamethasone. His blood sugar was controlled with human active insulin. His complete hemogram showed a decreasing platelet count which was tackled with platelet transfusions. Upon further investigation, it was detected that he had developed disseminated intravascular coagulopathy (DIC) due to Klebsiella pneumoniae sepsis for which filgrastim (for falling white blood cells and meropenem (for Klebsiella) were initiated.
Ten days after his discharge from the COVID-19 unit, he presented with severe, radiating pain in both gluteal regions which aggravated both on sitting and walking. He was admitted in late March 2021 for this complaint and a whole spine screening magnetic resonance imaging (MRI) was performed which revealed abnormal marrow signal intensity in all vertebral bodies. Short-tau inversion imaging (STIR) and T1-hypointense areas with lack of signal drop on opposed-phase gradient images (signal intensity index [SII] of 1.0 or more; SII ¼ SI on opposed-phase image/SI on in-phase image) were noted along the posterior third of all vertebral bodies. STIR hyperintensity was noted with an appropriate signal drop (SII < 1.0) on opposed-phase gradient images consistent with viable fatty marrow in the remaining anterior two thirds of the vertebrae (►Figs. 2-5).
Suh et al showed that in-and opposed-phase chemical shift imaging (CSI) has excellent diagnostic performance for differentiating benign and malignant vertebral bone marrow lesions with pooled sensitivity of 0.92 and specificity of 0.89. They found that the proposed signal intensity ratio cut-off values (opposed-phase-in-phase) were similar (0.8-1) in seven studies. They concluded that a signal intensity ratio of <0.8 indicated significant signal drop on opposed-phase images and the presence of bone marrow fat, which favors a benign condition as fatty replacement, is associated with acute benign conditions. However, a signal intensity ratio >1 did not always mean that a malignancy is present but stem from increased amounts of blood and cell water. [bib_ref] Diagnostic performance of in-phase and opposed-phase chemical-shift imaging for differentiating benign and..., Suh [/bib_ref] To further evaluate the condition, the patient underwent 18-F fluorodeoxyglucose positron emission tomographycomputed tomography (FDG PET-CT) which demonstrated abnormal heterogeneous areas of increased tracer uptake in all the vertebral bodies (►Fig. 6) and showed mild sclerotic density with loss of normal trabecular pattern on CT corresponding to the abnormal T1 hypointense areas seen on the whole spine MRI (►Fig. 7). This was suggestive of an infiltrative marrow disorder or metastasis. The differential of avascular bone marrow necrosis was included to cover the imaging spectrum of morphological appearances. FDG uptake has been reported in a case of avascular marrow necrosis by Grigolon and Delbeke. It is well reported that FDG tracer is taken up in inflammatory or infectious processes akin to the inflammatory exudates occurring in the pathogenesis of osteonecrosis. [bib_ref] F-18 FDG uptake in a bone infarct: a case report, Grigolon [/bib_ref] Sharply defined serpiginous areas of STIR hyperintensity were also noted in both halves of his pelvis and both femora, with presence of the well-described "double line sign," suggestive of avascular bone necrosis (►Fig. 8).
This diagnostic conundrum, therefore, necessitated bone marrow trephine aspiration and biopsy which demonstrated areas of necrosis with neutrophilic debris, lymphohistiocytic aggregates, and proliferating fibroblasts in an oedematous background (►Fig. 9). Scattered reactive plasma cells showing immunopositivity for CD138 (cluster of differentiation), kappa and lambda immunostains were noted. An ill-formed granuloma comprising of lymphocytes, plasma cells, and few histiocytic cells was also seen.
COVID-19-associated bone marrow necrosis has commonly been reported with use of glucocorticoids in the conditions such as acute respiratory distress syndrome (ARDS). Glucocorticoids are widely used to hinder the progression of acute lung injury and ARDS in patients with severe acute respiratory syndrome (SARS) and COVID-19. [bib_ref] Dexamethasone in hospitalized patients with Covid-19-preliminary report, Horby [/bib_ref] A host of studies scrutinizing the use of steroids in viral respiratory diseases showed expected complications of avascular necrosis and diabetes, increased mortality, and secondary infections in influenza and reduced clearance of viral particles in SARS and Middle East respiratory syndrome (MERS) coronavirus outbreaks. [bib_ref] Early changes in immune cell subsets with corticosteroids in patients with solid..., Marté [/bib_ref] Patients plagued with the severe form of COVID-19 disease often find themselves faced with such challenges as coagulopathies and DIC. Infection-induced endothelial dysfunction causes an excess of procoagulant thrombin with concomitant shutting down of the fibrinolytic cascade resulting in a hypercoagulable state which can spur thrombosis by increasing blood viscosity and coupled with hypoxia can activate the hypoxia-induced transcription factor-dependent signaling. To add to that, prolonged immobilization in these COVID-19 afflicted patients is associated with a much higher risk of developing venous thromboembolism. 6
# Conclusion
Based on the above clinical, radiological, PET, and bone marrow biopsy findings, we presume that the MRI features are highly suggestive of diffuse bone marrow necrosis associated with COVID-19 infection following treatment with high-dose pulse steroids. Our case report will help radiologists to consider bone marrow necrosis as a strong differential in the current pandemic as more of these kinds of rare observations of COVID-19 marrow involvement get reported.
## Note
This article has not been presented or published anywhere else.
# Funding
No support or funding has been requested or granted for this article.
## Conflict of interest
There is no conflict of interest. Coronal STIR screening sequence of pelvis with both hips showing well demarcated serpiginous STIR hyperintensity in both femora with the well-described "double line sign" (yellow arrow). STIR, short-tau inversion imaging. Bone marrow trephine biopsy sections (A and B; Â40 magnification) with hematoxylin & eosin stain show lymphohistiocytic aggregates (black arrow) and areas of necrosis (white arrow) with neutrophilic debris (yellow arrow) and proliferating fibroblasts in an oedematous background.
[fig] Figure 1: A) Axial schematic diagram of blood supply to bony vertebra showing anterior (AC) and posterior central (PC) arteries; black area represents watershed area between AC and PC arteries. (B) Midsagittal schematic diagram of two adjacent vertebral bodies with upper vertebra showing vascular territories of AC and PC arteries and lower vertebra showing watershed area (WSA; gray circles) in deep medullary portion and end arterial territory (EAT; white circles) near end plate. Black circles denote areas located within both EAT and WSA and may represent areas more vulnerable to ischemia. [/fig]
[fig] Figure 2: A) STIR mid sagittal image showing hyperintensity in the anterior two-third and hypointensity in the posterior one-third of the vertebral bodies. (B) T1 mid sagittal image showing hypointensity in the posterior one-third of the vertebral bodies. [/fig]
[fig] Figure 3: Axial T1 (A) and T2 (B) images of L2 lumbar vertebra showing well-defined hypointensity on both T1 and T2 in the posterior onethird of the L2 vertebral body. [/fig]
[fig] Figure 6: 18-F FDG PET-CT fusion images (A and B) showing heterogeneously increased tracer uptake in the posterior parts of vertebral bodies. CT, computed tomography; FDG PET, fluorodeoxyglucose positron emission tomography. [/fig]
[fig] Figure 4: In-phase (A) and opposed-phase (B) images of cervicodorsal spine showing signal drop on opposed-phase images in the anterior two-thirds of the vertebral bodies signifying viable fatty marrow. [/fig]
[fig] Figure 5: In-phase (A) and opposed-phase (B) images of dorsolumbar spine showing signal drop on opposed-phase images in the anterior two-thirds of the vertebral bodies signifying viable fatty marrow. [/fig]
[fig] Figure 7: Reformatted mid-sagittal (A, B) and axial (C) CT images showing mild sclerosis with loss of normal trabecular pattern in the posterior parts of the vertebral bodies corresponding to the abnormal T1hypointense areas seen in ►Fig. 2. CT, computed tomography. Indian Journal of Radiology and Imaging Vol. 31 No. 3/2021 © 2021. Indian Radiological Association. All rights reserved. [/fig]
|
Remarkable hexafunctional anion receptor with operational urea-based inner cleft and thiourea-based outer cleft: Novel design with high-efficiency for sulfate binding
The recognition of anions by designed receptors has attracted much attention in recent days. In particular, the selective binding of sulfate with artificial receptors is important because of its relevance to many biological and environmental applications. However, the development of organized molecular receptors with high-efficiency for sulfate binding still remains a significant challenge. We report a novel para-phenylene-bridged hexafunctional tripodal receptor that contains a urea-based inner cleft and a thiourea-based outer cleft, providing perfect sites for step-wise binding of two anions within a single cavity. The new receptor was synthesized in a three-step process, and was investigated for its anion binding properties by 1 H NMR titrations, 2D NOESY experiments and computational studies. As indicated by solution binding studies, the receptor selectively binds sulfate over other oxoanions, forming a 1:2 stoichiometric complex that is stabilized via strong H-bonding interactions. High-level DFT calculations reveal that the receptor, owing to the enhanced H-bonding ability of thiourea groups, initially encapsulates one sulfate in its thiourea-based outer cleft, followed by a second encapsulation in its urea-based inner cleft. Such a functionalized receptor with the unique combination of urea-based cleft and thiourea-based cleft in a single receptor has not been reported previously.Anion binding and sensing with designed abiotic receptors has become an active area of current research 1-6 . Molecular clefts with organized functional groups in defined geometries have recently gained significant attention in anion coordination chemistry 7-10 . Among the various anions, sulfate is particularly important because of its relevance to biological and environmental implications with respect to nuclear waste management 11 , biosynthesis 12 , and protein binding 13 . Thus, there is a growing interest in developing suitable molecular receptors that can strongly and selectively bind sulfate anions 14 . The use of urea groups appended to the tripodal framework in anion binding is well documented, showing complementary binding with sulfate 15-29 . For instance, tren-based urea receptors substituted with m-cyanophenyl 15 , p-cyanophenyl 16 , m-nitrophenyl 17 , and 3-pyridyl 18 were shown to bind sulfate by hydrogen bonding interactions. Although receptors incorporating urea functional groups have been shown as useful binding motifs for anions through NH···O interactions, it has been reported that the incorporation of thiourea groups to synthetic receptors leads to an enhanced acidity of NH 25 , thereby providing strong affinity for anions[30][31][32][33][34][35][36][37][38]. Furthermore, recent reports have established the link between the anion binding and certain catalytic reactions, especially for thiourea-based anion receptors 39-41 . For example, Jacobsen et al. have reported a thiourea-based compound that, upon the binding of a fluoride ion, can catalyze the acylation of silyl ketene acetals with acyl fluorides39.It has been shown that the free energies of association for host-guest interactions are dependent on the number of rotatable bonds formed by the hosts and guests 42 ; therefore, the increased number of binding sites within a single host designed from structural manipulations could lead to the enhancement of its binding ability and
selectivity for a specific guest. Recently, Wu et al. have reported tripodal-based hexaurea receptors containing ortho-phenylene-bridged bisurea moieties that encapsulated sulfate through H-bonds, forming 1:1 complexes [bib_ref] Highly efficient extraction of sulfate ions with a tripodal hexaurea receptor, Jia [/bib_ref] [bib_ref] Stepwise Encapsulation of Sulfate Ions by Ferrocenyl-Functionalized Tripodal Hexaurea Receptors, Huang [/bib_ref]. In the pursuit of achieving the higher order of binding sites, we synthesized a pentaflouro-substituted hexaurea receptor that formed an encapsulated complex with a carbonate ion [bib_ref] Absorption of atmospheric CO 2 as carbonate inside the molecular cavity of..., Pramanik [/bib_ref].
Herein, we report a novel para-phenylene-bridged hexafunctional mixed urea/thiourea receptor L that contains one inner cleft with three urea groups and one outer cleft with three thiourea groups. We hypothesized that such an organization with two operational clefts linking through rotatable spacers in a single molecule could provide perfect sites for hosting two anions, each within one cleft. Because of the enhanced acidity of the thiourea groups, the receptor would possibly show preference to bind the first anion at the outer cleft (instead of inner cleft). This assumption was further supported by the electrostatic potential surfaces of L calculated at the M06-2X/6-31G(d,p) level of theory, showing more positive potential on the outer cleft than the inner cleft. Such an assembled, exceptional anion receptor with an elite blend of urea-based cleft and thiourea-based cleft in a single tripodal receptor has not been reported previously.
# Results and discussion
As demonstrated by the electrostatic potential map as well as by the optimized structures (discussed later), the receptor adopts into a C 3 symmetric cone shape, owing to the presence of three identical (para-phenylene-bridged) arms linked to the tertiary amine. A strong electrostatic positive potential is created within both the inner and outer clefts due to the urea and thiourea moieties, potentially making the molecule a ditopic receptor for anions. Through the analysis of 1 H NMR binding isotherms and NMR NOESY experiments, we have shown that the receptor binds sulfate selectively over other oxoanions, forming a 1:2 stoichiometric complex. High level DFT calculations further support that the receptor efficiently encapsulates two sulfate ions in its inner cleft and outer cleft.
## Synthesis.
The new hexafunctional mixed urea/thiourea receptor L was prepared by three-step synthetic strategy, with about 40% overall yield. Tris(2-aminoethyl) amine (tren) (1) and p-nitrophenyl isocyanate were reacted to give the nitro-functionalized tris-urea 2 which was reduced with hydrazine hydrate and Pd/C (10%) to produce the amino-functionalized tris-urea 3. The final coupling was achieved by reacting 3 with p-cyanophenyl isothiocyanate to form the p-phenylene bridged hexafunctional mixed urea/thiourea L. The receptor is stable under normal conditions and soluble in DMSO, but hardly soluble in water and other common organic solvents. Attempts to isolate X-ray quality crystals of L with anions were unsuccessful. NMR studies. The binding properties of L towards various oxoanions (SO 4 2− , HSO 4 − , H 2 PO 4 − , ClO 4 − and NO 3 − ), which were added as their tetrabutylammonium (TBA) salts, were investigated in DMSO-d 6 by using proton NMR titration techniques at room temperature. The free receptor shows four NH resonances at 6.16 (NH1), 8.58 (NH2), 9.99 (NH3) and 10.02 (NH4) ppm in its NMR spectrum: two for ureas (NH1 and NH2), and the other two for thioureas (NH3 and NH4), indicating the C 3 conformation of L (The NH signals were assigned by NOESY NMR spectroscopy, see below. Seefor the numbering of the NH protons).shows the 1 H NMR spectra of the free L and its mixture with 5 equivalents of different anions. After the addition of SO [bib_ref] Artificial receptors for the recognition of phosphorylated molecules, Hargrove [/bib_ref] 2− to L, the NH resonances significantly shifted downfield showing ΔNH2 = 0.89, ΔNH3 = 0.27 and ΔNH4 = 0.35, while H1 signal overlapped with one aromatic proton at 7.13 ppm (ΔNH1 = 0.97 ppm), suggesting the interactions of L with sulfate anions. In addition to the shifting of NH signals, the aromatic signals also shifted. In particular the upfield shift of peripheral signals (Hd and Hb) on p-cyanophenyl and p-phenylene groups were observed, which may be due to a shielding effect induced by the encapsulated sulfate inside the outer cleft [bib_ref] Synthesis and anion binding studies of tris(3-aminopropyl)amine-based tripodal urea and thiourea receptors:..., Khansari [/bib_ref] [bib_ref] Highly efficient extraction of sulfate ions with a tripodal hexaurea receptor, Jia [/bib_ref]. Notably, the shift difference of NH resonances of urea groups is much larger than that of thiourea groups, suggesting a possible cavity strain due to the encapsulation of sulfate anion into the inner cavity. The addition of HSO 4 − to L induced small but considerable changes in the chemical shifts of both urea and thiourea groups, as it contains two charges. The addition of H 2 PO 4 − to L resulted in downfield shifts of NH1 and NH2 signals, while both NH3 and NH4 signals were broadened as observed previously for related ligands [bib_ref] Highly efficient extraction of sulfate ions with a tripodal hexaurea receptor, Jia [/bib_ref] [bib_ref] Stepwise Encapsulation of Sulfate Ions by Ferrocenyl-Functionalized Tripodal Hexaurea Receptors, Huang [/bib_ref]. In contrast, the addition of ClO 4 − or NO 3 − to L did not show any noticeable change in the shifts of NH or aromatic protons (see the Supporting Information, Figs S12 and S13), thus indicating weaker interactions between the perchlorate or nitrate anions and the receptor.shows the stacking of 1 H NMR spectra of L with varying amount of SO 4 2− (0 to 10 eq.) in DMSO-d 6 , exhibiting gradual downfield shifts of NH signals. The shift changes for NH resonances [fig_ref] Figure 3 1: H NMR titration curves of L [/fig_ref] , however, were not consistent with a purely 1:1 binding model as commonly observed for related molecules [bib_ref] Synthesis and anion binding studies of tris(3-aminopropyl)amine-based tripodal urea and thiourea receptors:..., Khansari [/bib_ref]. Therefore, they were analyzed with a 1:2 (L:sulfate) binding model using the EQNMR program, displaying the binding constants (in log K) of 3.06(2) and 2.56(4) for L + SO [bib_ref] Artificial receptors for the recognition of phosphorylated molecules, Hargrove [/bib_ref] 2 These results clearly indicate the stepwise binding of two sulfates, one with the inner cleft and other with the outer cleft. Gunnlaugsson et al. reported ortho-, meta-, and para-phenylene bridged acyclic urea-amide based receptors, demonstrating that an anion recognition at the first binding moiety may lead to a "positive allosteric effect" for the second functionality toward anions, thus promoting the formation of a 1:2 complex 47 . A similar effect was recently described by Wu et al. for a ferrocenyl-functionalized hexaurea receptor that contains two urea groups separated by a meta-phenylene group, showing both 1:1 and 1:2 complexes with sulfate anions, as supported by [bib_ref] Anion receptor chemistry: highlights from 2010, Wenzel [/bib_ref] H NMR and theoretical calculations [bib_ref] Stepwise Encapsulation of Sulfate Ions by Ferrocenyl-Functionalized Tripodal Hexaurea Receptors, Huang [/bib_ref]. In an earlier report, we also observed that a para-xylene bridged hexaprotonated azamacrocycle was capable of hosting two chlorides at its two binding moieties via trigonal recognition of two clefts [bib_ref] Charge-assisted encapsulation of two chlorides by a hexaprotonated azamacrocycle, Hossain [/bib_ref]. In the present work, the receptor L featuring two clefts with different functionalities (urea and thiourea) can readily host two tetrahedral sulfate anions in its two clefts. Owing to an enhanced binding ability as well as the structural complementarity of thiourea functionalities, it is suggested that the first binding occurs at the outer cleft followed by the second binding at the inner cleft. This is further supported by optimizing the geometries and calculating the respective binding energies using high-level density functional theory (discussed later). As shown in, the larger change in chemical shifts are observed within 0 to 1 equivalents of SO 4 2− , implying a 1:1 complex, while the formation of the 1:2 species is dominant after one equivalent of the added anion. The stepwise binding constants of L for SO [bib_ref] Artificial receptors for the recognition of phosphorylated molecules, Hargrove [/bib_ref] 2− have been shown as log K 1 = 3.07(3) and log K 2 = 2.56(4) for the first and second sulfate, respectively. Since the first binding constant is higher than the second binding constant, this binding process can be considered as "non-cooperative" [bib_ref] What is cooperativity?, Hunter [/bib_ref]. The titrations of L with HSO 4 − or H 2 PO 4 − also suggest the stepwise formation of both 1:1 and 1:2 complexes, and the calculated binding constants are provided in . The higher binding constant for the first step as compared to that for the second step for each complex implies that the outer cavity is the preferential binding site for sulfate, presumably the enhanced acidity of thioureas. Further, the overall binding trend in the order of SO [bib_ref] Artificial receptors for the recognition of phosphorylated molecules, Hargrove [/bib_ref] 2
[formula] − = [L(SO 4 )] 2− and [L(SO 4 )] 2− + SO 4 2− = [L(SO 4 ) 2 ] 4− , respectively.− > HSO 4 − > H 2 PO 4 − > ClO 4 − or NO 3 − [/formula]
, suggests that the receptor can selectively bind sulfate over other anions studied.
The solution binding mode of L for sulfate anion was further evaluated by 2D NOESY NMR experiments [fig_ref] Figure 4: 2D NOESY NMR experiment of [/fig_ref] , as reported before by us [bib_ref] Seven-coordinate anion complex with a tren-based urea: Binding discrepancy of hydrogen sulfate..., Pramanik [/bib_ref] and others [bib_ref] Oxyanion-encapsulated caged supramolecular frameworks of a tris (urea) receptor: evidence of hydroxide-and..., Dey [/bib_ref] [bib_ref] Complexation of anions including nucleotide anions by open-chain host compounds with amide,..., Werner [/bib_ref]. As shown in [fig_ref] Figure 4: 2D NOESY NMR experiment of [/fig_ref] , the free receptor of L shows two strong cross peaks for NH1NH2 and NH3NH4 of urea and thiourea moieties, respectively; indicating that these protons are close in space [bib_ref] Seven-coordinate anion complex with a tren-based urea: Binding discrepancy of hydrogen sulfate..., Pramanik [/bib_ref]. In addition, two strong couplings for NH2CHa and NH3CHb with aromatic protons were observed. However, the addition of two equivalents of sulfate anions to L resulted in the complete loss of NH1NH2 contacts, implying a possible rotation of the two sites of a thiourea unit in order to bind a sulfate anion. On the other hand, the NH3NH4 contacts from urea groups were retained [fig_ref] Figure 4: 2D NOESY NMR experiment of [/fig_ref] , suggesting that these protons remain in a close distance after the encapsulation of sulfate. Indeed, as shown in the optimized structure of the sulfate complexes [fig_ref] Figure 5: Optimized structures of [/fig_ref] , the NH sites of a single urea are twisted to bind two oxygen atoms of sulfate inside the inner cleft, while this is not the case for thiourea groups, showing the respective sites bonded to a single oxygen atom. . a The binding constants were determined using a 1:2 binding model for the following reaction: Our previous work has shown that the M06-2X functional accurately predicts the binding energy trends for non-covalent interactions between anions and organic receptors [bib_ref] Synthesis and anion binding studies of tris(3-aminopropyl)amine-based tripodal urea and thiourea receptors:..., Khansari [/bib_ref]. To this end, the initial equilibrium geometry for the free receptor L was first optimized at the M06-2X/6-31G(d,p) level of theory [bib_ref] Ab initio study of solvated molecules: a new implementation of the polarizable..., Cossi [/bib_ref]. From this equilibrium geometry, the sulfate anion with different orientations was placed in a single (inner or outer) cleft or both clefts, and molecular geometries of the various sulfate-bound complexes were fully optimized at the M06-2X/6-31G(d,p) level of theory and corrected for zero-point energies (ZPE) in gas phase as well as in a solvent phase to approximate a DMSO environment (dielectric constant = 46.8) using a polarizable continuum model (PCM). With the optimized geometry, the binding energies of L for SO 4 2− were calculated using the equation:
[formula] + L L L A [ A] [ A ] K A K 2 1 2 .ΔE = E(complex) − [E( receptor) + E(anion)]. [/formula]
As shown in [fig_ref] Figure 5: Optimized structures of [/fig_ref] , the optimized structure of the receptor adopts a perfect C 3 symmetric cone shape, due to the presence of three identical arms linked to the tertiary amine. The inner cleft of the receptor is decorated with six intra-molecular H-bonds, where each oxygen atom of one urea group is H-bonded with both NH of the adjacent urea unit. Further, all three NH groups of the outer cleft are pointed inside the cavity, making it a preferred binding site for a C 3 symmetric sulfate anion. With this optimized geometry, we first attempted to organize all NH groups of L around a tetrahedral sulfate; however, due to the lack of complementarity, the receptor could not be optimized with a single anion bonded to both clefts simultaneously. Therefore, we proceeded to optimize with one sulfate added separately at each cleft or two sulfates at both clefts of L. The optimized structure of the thiourea-bound sulfate complex as shown in [fig_ref] Figure 5: Optimized structures of [/fig_ref] , reveals that the sulfate binds to the outer cleft through a total of six NH···O bonds (NH···O = 2.78-2.93 Å). The calculated binding energy of this complex was found to be −203 kcal/mol in gas phase, while it was much lower (−96 kcal/mol) in solvent phase, due to the polarity effect of DMSO solvent included in the calculations [bib_ref] Synthesis and anion binding studies of tris(3-aminopropyl)amine-based tripodal urea and thiourea receptors:..., Khansari [/bib_ref].
In contrast to the thiourea-bound sulfate complex, the receptor significantly deformed in the urea-bound sulfate complex [fig_ref] Figure 5: Optimized structures of [/fig_ref] to encapsulate a sulfate anion within its inner cleft, yielding the binding energies as −151 and −77 kcal/mol in gas and solvent phase, respectively. The calculated binding energies for thiourea-bound complex (ΔE = −203 kcal/mol) and for urea-bound complex (ΔE = −151 kcal/mol) are comparable to our previous report on sulfate binding with a tris-thiourea (ΔE = −200 kcal/mol) and a tris-urea (ΔE = −173 kcal/mol) in gas phase [bib_ref] Synthesis and anion binding studies of tris(3-aminopropyl)amine-based tripodal urea and thiourea receptors:..., Khansari [/bib_ref]. The higher binding energy for the thiourea-bound complex [fig_ref] Figure 5: Optimized structures of [/fig_ref] than that for the urea-bound complex [fig_ref] Figure 5: Optimized structures of [/fig_ref] demonstrates that the outer cleft is the preferential binding site for the first sulfate, which is in agreement with the experimental results. As mentioned previously, these results further support our assumption that the binding of the first sulfate at the outer cleft [fig_ref] Figure 5: Optimized structures of [/fig_ref] may allow the second sulfate to bind at the inner cleft.
Considering that the first binding occurs at the outer cleft (thiourea groups), followed by the second binding at the inner cleft (urea groups), as proposed by NMR titration studies, we proceeded to re-optimize the receptor with two sulfate ions by incorporating both clefts, each with a single sulfate. The calculated binding energies were found to be −161 and −87 kcal/mol in gas and solvent phase, respectively. The optimized structure, as displayed in [fig_ref] Figure 6: Optimized structures of 1 [/fig_ref] , reveals that both the inner cleft and the outer cleft are occupied by sulfate anions that are bound through strong H-bonding interactions (NH···O < 2.94 Å), thereby overcoming the expected electrostatic repulsion due to the encapsulation of two anions in a single molecule. It is noteworthy that the receptor, in a 1:1 complex (urea-or thiourea-bound sulfate), readjusted its geometry to implement maximum interactions for sulfate that is bonded through six NH···O bonds (see bond distances in . While the thiourea-bound complex adopted a perfect C 3 symmetry, leaving the urea-cleft open for a second sulfate (see [fig_ref] Figure 5: Optimized structures of [/fig_ref] ; the urea-bound complex deviated from its C 3 conformation, adopting a folded umbrella that could not allow to bind another sulfate due to the nonexistence of the outer cavity [fig_ref] Figure 5: Optimized structures of [/fig_ref]. On the other hand, the receptor is stabilized with two sulfates, each with six NH···O bonds from six NH binding sites from a single cleft (inner or outer), creating a perfect C 3 symmetric 1:2 complex.
# Conclusion
We have designed and synthesized a novel para-phenylene-bridged hexafunctional tripodal receptor consisting of two different functionalized clefts (urea-based inner cleft and thiourea-based outer cleft). As demonstrated by experimental studies and theoretical calculations, the receptor can effectively bind sulfate anions in a two-step binding process, leading to a well-defined 1:2 stoichiometric complex that is stabilized through complementary H-bonding interactions. Our results suggest that the unique combination of two different functionalities makes the receptor ideal to bind the first sulfate at the thiourea-based outer cleft and the second sulfate at the urea-based inner cleft. The preferred binding at the outer cleft is due to the enhanced H-bonding ability as well as of the structural complementarity of thiourea functionalities, leading to stronger interactions with the anion than those with its urea analogue. This binding propagation was further supported by DFT calculations, illustrating that the thiourea-bound complex is energetically more favorable than the urea-bound complex. Therefore, we conclude that the binding of one sulfate at the outer cleft assists the receptor to bind the second sulfate at the inner cleft. To the best of our knowledge, such an assembled multifunctional anion receptor with the unique combination of a urea-based cleft and a thiourea-based cleft has not been reported previously. Understanding and being able to accurately predict the interactions between synthetic receptors and guests is a key step towards elucidating the complex mechanisms in living systems. Taken together, the results from our study may be useful in developing highly organized molecular receptors for extraction, catalysis and drug design for environmental and biomedical applications.
# Methods
General. All reagents and solvents were purchased as reagent grade and were used without further purification. Nuclear magnetic resonance (NMR) spectra were recorded on a Varian Unity INOVA 500 FT-NMR. Chemical shifts for samples were measured in DMSO-d 6 and calibrated against sodium salt of 3-(trimethylsilyl) propionic-2,2,3,3-d 4 acid (TSP) as an external reference in a sealed capillary tube. NMR data were processed and analyzed with MestReNova Version 6. [bib_ref] Artificial receptors for the recognition of phosphorylated molecules, Hargrove [/bib_ref] 2-) at neutral pH. Initial concentrations were [ligand] 0 = 2 mM, and [anion] 0 = 20 mM. Each titration was performed by 13 measurements at room temperature. The association constant K was calculated by fitting of several independent NMR signals using a 1:2 (L:anion) binding model using the EQNMR program. Error limit in K was less than 15%.
## Tris-(4-aminophenyl)-urea (3).
Computational studies. Interaction energies and geometry optimization of sulfate complexes were performed with density functional theory (DFT) calculations 51 . All calculations were carried out using Gaussian 09 package of programs 52 . Data Availability. All data generated or analysed during this study are included in this published article and its Supplementary Information files.
[fig] Figure 1: (a) Synthetic scheme for L ((i) p-nitrophenyl isocyanate, (ii) Hydrazine hydrate and Pd/C (10%), and (iii) p-cyanophenyl isothiocyanate), and (b) electrostatic potential map for L, calculated at the M06-2X/6-31G(d,p) level of theory (red = negative potential, and blue = positive potential). Scientific REPoRts | 7: 6032 | DOI:10.1038/s41598-017-05831-x showing ΔNH1 = 0.37, ΔNH2 = 0.28, ΔNH3 = 0.06 and ΔNH4 = 0.09 ppm. The larger shift change (ΔNH) in the respective NH signal due to the addition of SO 4 2− than that of HSO 4 − indicates stronger interactions of [/fig]
[fig] Figure 2: (a) Partial 1 H NMR spectra of L (2 mM) in the presence of 5 equivalents of different anions in DMSO-d 6 ; (b) partial 1 HNMR titration of L showing changes in the NH chemical shifts of L (2 mM) with an increasing amount of SO 4 2− (20 mM) in DMSO-d 6 . (H1 = CH 2 NHCO, H2 = CONHAr, H3 = ArNHCS, H4 = CSNHAr); and (c) proposed binding mechanism of L with [/fig]
[fig] Figure 3 1: H NMR titration curves of L (2 mM) with an increasing amount of various oxoanions (R = [anion] 0 / [L] 0 ) in DMSO-d 6 . [/fig]
[fig] Figure 4: 2D NOESY NMR experiment of (a) Free L, and (b) L in the presence of one equivalent of sulfate anion in DMSO-d 6 at room temperature. Scientific REPoRts | 7: 6032 | DOI:10.1038/s41598-017-05831-x Computational studies. In an effort to understand the interactions and structural aspects of the new receptor with sulfate, theoretical calculations based on density functional theory (DFT) were performed with hybrid meta-exchange correlation functional M06-2X 51 , using the Gaussian 09 package of programs 52 . [/fig]
[fig] Figure 5: Optimized structures of (a) L, (b) thiourea-bound 1:1 complex [L(SO 4 )] 2− , and (c) urea-bound 1:1 complex [L(SO 4 )] 2− , calculated at the M06-2X/6-31G(d,p) level of theory. [/fig]
[fig] Figure 6: Optimized structures of 1:2 complex [L(SO 4 ) 2 ] 4− complex showing (a) perspective view, and (b) space filling model, calculated at the M06-2X/6-31G(d,p) level of theory. [/fig]
|
Major histocompatibility complex (MHC) class II alleles, haplotypes and epitopes which confer susceptibility or protection in systemic sclerosis: analyses in 1300 Caucasian, African-American and Hispanic cases and 1000 controls
Objective To determine human leucocyte antigenclass II (HLA-class II) (DRB1, DQB1, DQA1 and DPB1) alleles, haplotypes and shared epitopes associated with scleroderma (systemic sclerosis (SSc)) and its subphenotypes in a large multi-ethnic US cohort by a case-control association study. Patients and methods 1300 SSc cases (961 white, 178 black and 161 Hispanic subjects) characterised for clinical skin forms (limited vs diffuse), SScspecifi c autoantibodies (anticentromere (ACA), anti-topoisomerase I (ATA), anti-RNA polymerase III (ARA), anti-U3 ribonucleoprotein (fi brillarin)) and others were studied using molecular genotyping. Statistical analyses in SSc itself by ethnicity, gender, skin type and autoantibodies were performed using exact logistic regression modelling for dominant, additive and recessive effects from HLA. Results The strongest positive class II associations with SSc in white and Hispanic subjects were the DRB1*1104, DQA1*0501, DQB1*0301 haplotype and DQB1 alleles encoding a non-leucine residue at position 26 (DQB1 26 epi), while the DRB1*0701, DQA1*0201, DQB1*0202 haplotype and DRB1*1501 haplotype were negatively correlated and possibly protective in dominant and recessive models, respectively. These associations did not discriminate between limited and diffuse SSc. SSc in black subjects was associated with DRB1*0804, DQA1*0501, DQB1*0301 alleles. DPB1*1301 showed the highest odds ratio for ATA (OR = 14). Moreover, it showed no linkage disequilibrium or gene interaction with
# Introduction
Scleroderma (systemic sclerosis, SSc) is a chronic complex autoimmune disease characterised by (a) organ fi brosis involving the skin, lungs, gastrointestinal tract and/or heart; (b) a proliferative vasculopathy primarily affecting small blood vessels and capillaries; (c) immune activation with production of disease-specifi c autoantibodies. [bib_ref] Systemic sclerosis: hypothesis-driven treatment strategies, Charles [/bib_ref] [bib_ref] Following the molecular pathways toward an understanding of the pathogenesis of systemic..., Jimenez [/bib_ref] [bib_ref] Scleroderma (systemic sclerosis): classifi cation, subsets and pathogenesis, Leroy [/bib_ref] [bib_ref] Scleroderma epidemiology, Mayes [/bib_ref] The disease is further classifi ed into limited and diffuse forms based on extent and distribution of cutaneous thickening.
The most common SSc-specifi c autoantibodies are directed against centromeric proteins anticentromere (ACA) (CENP B and A), anti-topoisomerase I (ATA) (also termed Scl-70) and anti-RNA polymerase III (ARA); however, a variety of less common specifi cities, typically antinucleolar, can be found, which include anti-U3 ribonucleoprotein (fi brillarin) AFA), anti-Th/To, anti-Pm-Scl, anti-RNA polymerase I and anti-U1-ribonucleoprotein (RNP). [bib_ref] HLA and autoimmunity in scleroderma (systemic sclerosis), Arnett [/bib_ref] Importantly, each patient with SSc typically produces only one of these autoantibodies and each one currently serves as a biomarker for different patterns of skin and visceral involvement, as well as prognosis. In addition, certain SScspecifi c antibodies occur in different frequencies among different ethnic groups.
Scleroderma is thought to be a complex genetic disease, infl uenced by multiple genes, with a substantial environmental or non-germline component based on twin studies. [bib_ref] Analysis of systemic sclerosis in twins reveals low concordance for disease and..., Feghali-Bostwick [/bib_ref] African-American and Hispanic patients with scleroderma tend to have more severe disease than their Caucasian counterparts and disease in African-American subjects begins at an earlier age. [bib_ref] Systemic sclerosis in 3 US ethnic groups: a comparison of clinical, sociodemographic,..., Reveille [/bib_ref] The Choctaw Indians of southeastern Oklahoma have a nearly 10-fold prevalence of the disease compared with other ethnicities. [bib_ref] Increased prevalence of systemic sclerosis in a Native American tribe in Oklahoma...., Arnett [/bib_ref] It appears likely that there are different combinations of genes, the interacting effects of which infl uence disease susceptibility and severity. Recently, we have found that the same polymorphism in the PTPN22 gene associated with rheumatoid arthritis, systemic lupus erythematosus and other autoimmune diseases is also associated with SSc, especially in those patients having ATA or ACA antibodies. [bib_ref] Association of the PTPN22 R620W polymorphism with anti-topoisomerase I-and anticentromere antibody-positive systemic..., Gourh [/bib_ref] [bib_ref] Association of a functional singlenucleotide polymorphism of PTPN22, encoding lymphoid protein phosphatase,..., Orozco [/bib_ref] [bib_ref] Replication of putative candidate-gene associations with rheumatoid arthritis in >4,000 samples from..., Plenge [/bib_ref] Additional 'autoimmunity' genes now reported to be associated with SSc include allograft infl ammatory factor, IL1A, IRF5, STAT4 and FAS. [bib_ref] An allograft infl ammatory factor 1 (AIF1) single nucleotide polymorphism (SNP) is..., Alkassab [/bib_ref] [bib_ref] Association between the IRF5 rs2004640 functional polymorphism and systemic sclerosis: a new..., Dieudé [/bib_ref] [bib_ref] Association of IL1A gene polymorphisms with susceptibility to and severity of systemic..., Kawaguchi [/bib_ref] [bib_ref] The STAT4 gene infl uences the genetic predisposition to systemic sclerosis phenotype, Rueda [/bib_ref] [bib_ref] The -670G>A polymorphism in the FAS gene promoter region infl uences the..., Liakouli [/bib_ref] Major histocompatibility complex (MHC) or human leucocyte antigen-class II (HLA-class II) allelic associations with SSc have been reported in European and North American Caucasian subjects (DRB1*0301, DRB1*11, DRB1*07) and Japanese and Koreans (DRB1*1502) over the past two decades but have been relatively weak. Much stronger correlations, however, have been demonstrated between certain HLA-class II alleles and each of the SSc-specifi c autoantibodies. Different HLA-DRB1, DQB1, DQA1 and/ or DPB1 alleles, or combinations thereof, are associated with expression of ACA (DQB1*0501 and other DQB1 alleles encoding non-leucine residues at position 26 in the peptide binding groove), ATA (DRB1*11, especially the DRB1*1104, DQB1*0301 haplotype) in Caucasian subjects and DRB1*1502, DQB1*0601 in Japanese and Korean subjects, 21-23 AFA antibody (the DRB1*1302, DQB1*0604 haplotype), [bib_ref] Autoantibodies to fi brillarin in systemic sclerosis (scleroderma). An immunogenetic, serologic, and..., Arnett [/bib_ref] and anti-PM-Scl (DRB1*0301). [bib_ref] The clinical and immunogenetic features of patients with autoantibodies to the nucleolar..., Marguerie [/bib_ref] No consistent HLA correlations heretofore have been made with ARA. [bib_ref] Studies of HLA-DR and DQ alleles in systemic sclerosis patients with autoantibodies..., Falkner [/bib_ref] Only a few studies have examined HLA-DPB1 alleles in SSc; however, one such allele (DPB1*1301) has been associated with ATA. [bib_ref] HLA-DPB1 alleles and autoantibody subsets in systemic lupus erythematosus, Sjögren's syndrome and..., Reveille [/bib_ref] [bib_ref] Class II HLA associations with autoantibodies in scleroderma: a highly signifi cant..., Gilchrist [/bib_ref] [bib_ref] The HLA-DP locus in systemic sclerosis--no primary association, Stephens [/bib_ref] It has been unclear whether this allele is an independent disease correlate or the result of linkage disequilibrium (LD) with HLA-DRB1 and DQB1 haplotypes. Although genetic infl uences are thought to have an important role in susceptibility to SSc, genetic studies of scleroderma have included relatively small numbers of patients, especially black and Hispanic subjects. HLA associations with specifi c autoantibodies may be clinically relevant because each of the autoantibody subsets of scleroderma is associated with certain disease features and different prognostic implications.
Thus, the overall aims of this study were to determine specifi c HLA-class II alleles, haplotypes and epitopes infl uencing susceptibility to, and/or expression of, SSc itself, its limited or diffuse forms, or its various autoantibody subsets across ethnic lines in the largest cohort yet of American patients.
# Methods
## Selection of ssc cases and normal controls
A case-control association study was performed (total SSc cases 1300; total controls 1000) along with subanalyses by ethnicity, gender, clinical subsets and specifi c autoantibody profi les. Patients were included from the NIH/NIAMS Scleroderma Family Registry and DNA Repository, [bib_ref] The establishment and utility of a population-based registry to understand the epidemiology..., Mayes [/bib_ref] the Genetics versus Environment in Scleroderma Outcomes Study (GENISOS) cohort followed up in the NIH/NIAMS Center of Research Translation in Scleroderma, [bib_ref] Analysis of systemic sclerosis in twins reveals low concordance for disease and..., Feghali-Bostwick [/bib_ref] and the UT-Houston Rheumatology Division. Patients with SSc included 961 Caucasian, 178 African-American and 161 Hispanic subjects; normal local control totals were 539 Caucasian, 263 African-American and 198 Hispanic subjects. Only Hispanic patients with SSc and controls of Mexican or Central American ancestry were included. HLA-DPB1 alleles were determined in 705 Caucasian patients with SSc and 287 Caucasian controls in the Scleroderma Registry. Only 82 African-American and Hispanic cases were studied for HLA-DPB1 alleles but were added to the Caucasian subjects for an overall comparison. All patients fulfi lled the preliminary American College of Rheumatology criteria for the diagnosis of SSc 34 or had three of the fi ve clinical features of the CREST syndrome (calcinosis, Raynaud's phenomenon, esophageal dysfunction, sclerodactyly or telangiectasia). Patients having overlapping Sjögren's syndrome, myositis or features of rheumatoid arthritis or systemic lupus were not excluded. Results of HLA allele frequencies in some of these patients have been reported previously and are included here. Because of the crosssectional nature of two of these three cohorts, reliable clinical information on specifi c organ involvement, such as pulmonary fi brosis, pulmonary hypertension or renal crisis, was not available to assess HLA associations. HLA registry controls were primarily spouse or friend controls, and GENISOS/Division controls were healthy medical centre personnel or blood bank donors from the local Houston area. All controls were screened for a history of any autoimmune diseases and excluded if positive. All study subjects provided written informed consent and the study was approved by the UTH Committee for the Protection of Human Subjects.
## Autoantibody identifi cation
Antinuclear antibodies and ACA were determined in all patients with SSc by indirect immunofl uorescence on HEp-2 cells (Antibodies, Davis, California, USA). Immunodiffusion against calf thymus extract was used to determine the presence of ATA (Scl-70), anti-Ro/SS-A, anti-La/SS-B, anti-Smith (Sm) and anti-RNP autoantibodies (Inova Diagnostics, San Diego, California, USA). ARA were determined using a commercially available enzyme-linked immunoassay (EIA) kit (MBL, Nagoya, Japan). AFA were determined only in those Division and GENISOS patients with SSc with a nucleolar antinuclear antibody pattern using immunoprecipitation.
## Hla-class ii genotyping
HLA-DQA1, -DQB1 and -DPB1 alleles were oligotyped and DRB1 alleles directly sequenced from extracted genomic DNA as previously described.
## Statistical analyses
We used statistical analysis software SAS 9.1.3 and the SAS Genetic Package (SAS Institute, Cary, North Carolina, USA). χ 2 Tests or Fisher's exact tests were used to compare HLA-class II allele frequencies between the normal control and SSc groups. Mantel-Haenszel tests were performed for the analysis of HLA-class II allelic frequencies to control for the confounding effects of ethnicity. In the online supplementary tables, we only reported the odds ratio (OR) and the corresponding 95% CI for p values <0.017, which accounts for the Bonferroni corrections by the number of comparisons for each HLA genotype (0.017 = 0.05/3). LD coeffi cients and the corresponding p values were computed to examine the allelic associations that occur between alleles at different loci using SAS Genetics software.
These χ 2 data are presented in the supplementary tables. We then performed exact logistic regression for each of the HLA variants using dominant (D), additive (A) and recessive (R) modelling. To account for the multiple comparisons, we used a false discovery rate approach. [bib_ref] Controlling the false discovery rate: a practical and powerful approach to multiple..., Benjamini [/bib_ref] We used an experimental threshold of α = 0.05 and accounted for 1000 potential multiple comparisons that we performed. Using these values for α and number of tests performed, a point signifi cance of 0.004 is considered statistically signifi cant. Nonetheless, p values of <0.05 also are selectively shown for alleles of interest and those which have been associated with SSc or its subgroups in other reports. Use of the 0.004 as the threshold to declare signifi cance leads to a rough false discovery rate of 0.025. For LD analyses we used the procedure haplotype in the SAS Genetic Package. In addition, we examined frequencies of the presence of non-leucine residues at position 26 in the DQ β chain based on our previous studies implicating this 'epitope' (DQB1*26 epi) in susceptibility to the ACA in SSc.
# Results
## Hla-class ii alleles in ssc
HLA-class II (DRB1, DQA1 and DQB1) genotyping was completed in 1300 SSc cases and 1000 normal controls and HLA-DPB1 in 705 white SSc cases and 287 white controls (tables 1-3 and supplementary tables 1 and 2).
In African-Americans, the HLA-DRB1*0804/DQA1*0501/ DQB1*0301 haplotype was signifi cantly increased, while DRB1*1104 was not associated with SSc [fig_ref] Table 2: HLA-class II allelic frequencies signifi cantly associated with African-American patients with systemic... [/fig_ref].
Signifi cant associations with SSc in Hispanic subjects were found for DRB1*1104 (OR = 4.99), DQA1*0501 (OR = 4.09) and DQB1*0301 (OR = 2.13), along with the DQB1*26 epi (OR = 1.81). Again, DRB1*0701 and DQA1*0201 were decreased in frequency.
HLA-DQA1*0501 was signifi cantly more frequent in male (101/165 or 61%) than in female (583/1135 or 51%) patients with SSc (p = 0.02, OR = 1.49, 95% CI 1.06 to 2.12), thus confi rming one earlier report.
## Hla alleles in limited versus diffuse ssc
HLA associations with both limited and diffuse forms of SSc largely mirrored the fi ndings in the disease group as a whole, except in black patients with limited disease whose numbers were small (n = 54) (supplementary tables 3a and 4a by logistic regression and supplementary tables 3, 4 and 5 by χ 2 ).
## Hla allele frequencies in autoantibody subsets of ssc
Frequencies and ethnic differences in SSc autoantibodies are shown in online supplementary table 5. ACA were found to be associated with several class II haplotypes in white subjects, including HLA-DRB1*0101, DQA1*0101 and DQB1*0501, DRB1*04, especially DRB1*0404, but weakly DRB1*0401 which carried DQB1*0301 alleles and weakly with DRB1*08 (*0801) and DQA1*0401 [fig_ref] Table 4: HLA-class II allelic frequencies signifi cantly associated with anticentromere antibody positive Caucasian... [/fig_ref] No HLA alleles were associated with ACA in black subjects and only HLA-DRB1*0407 in Hispanic subjects.
ATA in white subjects were strongly associated with the HLA-DRB1*1104, DQA1*0501, DQB1*0301 haplotype (OR = 6.93) and even more so with HLA-DPB1*1301 (OR = 14.02), both showing dominant models [fig_ref] Table 5: HLA-class II allelic frequencies signifi cantly associated with antitopoisomerase I antibody positive... [/fig_ref].
The DRB1*1104 association also was present in the black subjects and Hispanic subjects, as was the DQB1 26 epitope [fig_ref] Table 5: HLA-class II allelic frequencies signifi cantly associated with antitopoisomerase I antibody positive... [/fig_ref]. In addition, HLA-DRB1*08 alleles, DRB1*0804 in black subjects (OR = 3.42) and DRB1*0802 in Hispanic subjects (OR = 1.91) also were increased in these groups showing both additive and dominant effects. There appeared to be no negative or 'protective' effect from the DRB1*0701 haplotype. HLA-DPB1*1301 was highly signifi cantly associated with For all white patients with SSc versus ethnically matched controls, a strong association was found with HLA-DRB1*11, but specifi cally with DRB1*1104 and not with DRB1*1101, DQA1*0501 and DQB1*0301 also were signifi cantly increased. These three alleles constitute a haplotype with most DRB*11 alleles including both DRB1*1101 and DRB1*1104.Given the highest odds ratio (OR = 2.48) being conferred by DRB1*1104 rather than DQA1*0501 (OR = 2.29) or DQB1*0301 (OR = 1.50), the primary associated allele appeared likely to be DRB1*1104 in a dominant model. Also, the shared DQB1*26 epitope (absence of leucine in position 26) was increased (OR = 1.59). HLA-DRB1*0404 also was signifi cantly increased (OR = 2.33); however, the strongest HLA-class II association with SSc was with HLA-DPB1*1301 (OR = 3.18). Testing for LD between HLA-DPB1*1301 and DRB1*1104 showed an r 2 value of 0.0001 and for DQB1*0301 0.009, thus demonstrating no signifi cant LD of this DR/DQ haplotype with DPB1*1301.
The HLA-DRB1*0701, DQA1*0201, DQB1*0202 haplotype was negatively associated with SSc in an additive model (HLA-DRB1*0701, DQA1*0201) and in dominant model (HLA-DQB1*0202) suggesting that it conferred a 'protective' effect [fig_ref] Table 1: HLA-class II allelic frequencies signifi cantly associated with Caucasian patients with systemic... [/fig_ref]. Similarly, the HLA-DRB1*1501, DQA1*0102, DQB1*0602 haplotypes were signifi cantly decreased but in a recessive model. ATA when all ethnic groups were combined (p<0.0001, OR = 9.96) (supplementary table 7). Using logistic regression, there was no evidence of gene-gene interaction between DPB1*1301 and DRB1*1104 (p = 0.9863) or between DPB1*1301 and DQB1*0301 (p = 0.9999) in the ATA positive group or in the total SSc group.
ARA were found to be most strongly associated in white subjects with HLA-DRB1*0404 (OR = 5.13) and DRB1*11 alleles (OR = 1.55 for one copy, 6.78 for two copies), especially DRB1*1104 in additive and dominant models, respectively [fig_ref] Table 6: HLA-class II allelic frequencies signifi cantly associated with anti-RNA polymerase III positive... [/fig_ref]. In addition, a positive recessive effect was seen from DQB1*03 alleles (OR = 2.38), especially DQB1*0301 (OR = 1.50 for one copy, 3.77 for two copies). In black patients, the strongest associations were with DRB1*08 alleles (OR = 3.92), primarily DRB1*0804 (OR = 2.98), along with DQA1*0501 (OR = 3.10 for one copy, 6.03 for two copies) in an additive model and DQB1*0301 (OR = 3.60) showing a dominant effect. Among Hispanic subjects, DRB1*11, DQA1*0501 and the DQB1 26 epitope showed possible weak positive correlations, but the strongest association was with DQB1*0301 (OR = 4.07).
## Other autoantibodies and hla
The HLA-DRB1*1302, DQB1*0604 haplotype was most strongly associated with AFA in white subjects (p=0.0003, OR=6.87), along with the DQB1 26 epitope (p=0.0009, OR=1.60), while in black subjects DRB1*08 alleles (p=0.0003, OR=5.76), especially DQB1*0804 (p=0.002, OR=5.70), showed the strongest associations (supplementary tables 9 and 9a).
Anti-RNP, anti-Smith (Sm), anti-Ro/SSA and anti-La/SSB antibodies showed no signifi cant HLA associations.
# Discussion
new to this study was the fi nding that the African HLA-DRB1*0804 allele, along with DQA1*0501 and DQB1*0301, showed the most signifi cant associations in black subjects and recurred in those with diffuse disease, ATA, ARA and AFA.
Perhaps equally important was the susceptibility effect of other DQB1 alleles (besides *0301 and *0501) encoding polar amino acids at position 26 of the DQ β chain (DQB1*26 epi). Previously, we reported this shared epitope in the antigen binding cleft to be most important in the ACA autoimmune response [bib_ref] Association of polar amino acids at position 26 of the HLA-DQB1 fi..., Reveille [/bib_ref] ; however, this current larger study suggests that it occurs in the majority of patients with SSc regardless of ethnicity or autoantibody.
A second unreported and potentially 'protective' haplotype in white subjects, HLA-DRB1*1501, DQA1*0102, DQB1*0602, was signifi cantly decreased in a recessive model in Caucasian subjects with SSc overall and in both limited and diffuse subgroups, as well as in those patients with ACA.
An HLA-DPB1 allele, DPB1*1301 in a dominant model, showed the highest odds ratios in SSc (OR = 3.18) and in both the limited (OR = 4.20) and diffuse (OR = 4.38) forms in Caucasian subjects; however, these associations were completely explained by its strong correlation with ATA (OR = 14.02). Testing for LD of this DPB1 allele with the DRB1*1104 and DQB1*0301 alleles showed that these were independently associated class II effects. Interestingly, the ATA response is strongly associated with pulmonary fi brosis in SSc and certain HLA-DP alleles have been clearly shown to promote susceptibility to occupationally acquired berylliosis, another fi brosing lung disease. [bib_ref] Immunogenetic basis of environmental lung disease: lessons from the berylliosis model, Saltini [/bib_ref] Finally, ARA characteristically are markers of rapidly progressive skin thickening and renal crises and this study is the fi rst to demonstrate that they occur in similar prevalences in each of these ethnic groups and are associated with certain class II MHC alleles.
In previous studies of HLA and ARA, there have been contradictory fi ndings, perhaps because small samples have been studied. [bib_ref] Studies of HLA-DR and DQ alleles in systemic sclerosis patients with autoantibodies..., Falkner [/bib_ref] Funding NIH.
[table] Table 2: HLA-class II allelic frequencies signifi cantly associated with African-American patients with systemic sclerosis (SSc) compared with normal African-American controls using logistic regression [/table]
[table] Table 1: HLA-class II allelic frequencies signifi cantly associated with Caucasian patients with systemic sclerosis (SSc) compared with Caucasian normal controls using logistic regression [/table]
[table] Table 4: HLA-class II allelic frequencies signifi cantly associated with anticentromere antibody positive Caucasian and Hispanic patients with systemic sclerosis (SSc) compared with ethnically matched normal controls using logistic regression [/table]
[table] Table 5: HLA-class II allelic frequencies signifi cantly associated with antitopoisomerase I antibody positive Caucasian, African-American and Hispanic patients with systemic sclerosis (SSc) compared with ethnically matched normal controls using logistic regression 59 1.1 to 6.1 †SSc cases (n =154) versus normal controls (n = 539) in Caucasian subjects; ‡SSc cases (n = 40) versus normal controls (n = 263) in African-American subjects; §SSc cases (n = 37) versus normal controls (n = 198) in Hispanic subjects. [/table]
[table] Table 6: HLA-class II allelic frequencies signifi cantly associated with anti-RNA polymerase III positive antibody positive Caucasian, African-American and Hispanic patients with systemic sclerosis (SSc) compared with ethnically matched normal controls using logistic regression.Model (D/R/A) p Value OR95% CI 39 1.1 to 5.4 †SSc cases (n = 182) versus normal controls (n = 539) in Caucasian subjects; ‡SSc cases (n = 33) versus normal controls (n = 263) in African-American subjects; §SSc cases (n = 40) versus normal controls (n = 198) in Hispanic subjects; NS=not signifi cant for any of the three models [/table]
|
Biorefinery of Biomass of Agro-Industrial Banana Waste to Obtain High-Value Biopolymers
On a worldwide scale, food demand is increasing as a consequence of global population growth. This makes companies push their food supply chains' limits with a consequent increase in generation of large amounts of untreated waste that are considered of no value to them. Biorefinery technologies offer a suitable alternative for obtaining high-value products by using unconventional raw materials, such as agro-industrial waste. Currently, most biorefineries aim to take advantage of specific residues (by either chemical, biotechnological, or physical treatments) provided by agro-industry in order to develop high-value products for either in-house use or for sale purposes. This article reviews the currently explored possibilities to apply biorefinery-known processes to banana agro-industrial waste in order to generate high-value products out of this residual biomass source. Firstly, the Central and Latin American context regarding biomass and banana residues is presented, followed by advantages of using banana residues as raw materials for the production of distinct biofuels, nanocellulose fibers, different bioplastics, and other high-value products Lastly, additional uses of banana biomass residues are presented, including energy generation and water treatment.
# Introduction
The rising development of industries all over the world has brought a consequential increase in their residue generation, especially in the field of agro-industry. This waste can be denominated as "food supply chain waste" (FSCW) and can be defined as "organic material produced for human consumption lost or degraded primarily at the manufacturing and retail stages". This concept has emerged in the context of the current vast inefficiency of the food supply chain business. For instance, the Food and Agriculture Organization (FAO) revealed in 2011 that up to a third of the food aimed at human consumption is wasted every year globally. The environmental and economic impacts of this worrying situation have driven the development of technologies pursuing not only conventional waste management and disposal, but also the extraction of as much value as possible out of any given agro-industrial waste.
## Agro-industry residues as biomass sources
Agro-industries have slowly come to realize that valorization of biomass residues (either by using them as raw materials for the development of high-value products, or investing in recirculating processes that make use of these residues to obtain income in the long run) is not only beneficial from an environmental perspective, but can also help to minimize economic losses or even raise the net value of companies.
Inadequate treatment of these biomass residues has a negative impact on the environment, mainly generating greenhouse gases, contaminating water sources, and causing ecological problems. Biorefinery technologies rise as suitable alternatives to mitigate these impacts as they can assist on reducing waste volume, but also on producing high-value goods out of revalorized biomass waste for circular economy purposes.
Waste valorization converts polymeric substrates into useful products such as chemicals, materials, and fuels, often by extraction, chemical conversion, or degradation. Historically, the utilization of complex biomasses led the way in pulp and paper production, or biotechnological production of furfural, ethanol, or short chain organic acids since the 19th century. However, it was not until the 1990s when the term "biorefinery" became widespread in the industry, once biomass started to be used as a source of higher-value products.
## The concept of biorefinery
Out of the many published definitions for the term "biorefinery", perhaps the one from the American National Renewable Energy Laboratory (NREL) seems to fit best the approach of this review: "A biorefinery is a facility that integrates biomass conversion processes and equipment to produce fuels, power, and chemicals from biomass". Definitions such as this one comprise the conversion of biomasses not only into biofuels, biopolymers, high-value products, and fine chemicals, but also include the generation of power (heat and electricity) analogous to today's petroleum-based refineries.
In addition to lignocellulosic biomass-based industries, more sectors have shown interest in applying biorefinery approaches to organic waste, for instance: food waste, nonedible oilssewage sludge, and municipal solid wastejust to name a few. Biorefinery technologies enable more efficient use of agricultural resources and sustainable food production. Taking advantage of biorefinery technologies represents a valuable strategy for agro-industries and governments; hence, they can easily navigate the challenges of the green economy era.
## Applications of agro-industrial residues
Agro-industrial waste is mainly composed of lignocellulose biomass. Lignocellulose waste has been gaining increasing attention due to its mechanical and thermal properties, renewability, wide availability, non-toxicity, low cost, and biodegradability. The vast range of lignans and celluloses comprised in agro-industrial residues grants them tremendous potential as feedstocks for chemical and biotechnological conversion processes. For instance, enzymatic breakdown of cellulose and hemicelluloses into glucose and xylose allows further fermentation of these monosaccharides into ethanol by fermentative microorganisms. Furthermore, pyrolysis and anaerobic digestion of lignocellulose biomasses can yield combustion gases such as H 2 and CH 4.
Currently, organic and agro-industrial residues take up a large portion of overall global waste, which is one of the reasons to make good use of it. The abundance of biomass feedstocks gives a positive prospect for their future utilization in biorefinery technologies. Estimates on biomass crop residue flows in Latin America show that most of lignocellulose containing biomasses are mostly made of maize, soybean, and sugarcane residues; banana residues are not found within the main agro-industrial residues of developing countries to be used as biorefinery biomass sources, though in many locations banana waste treatment remains a problem that needs to be addressed, as we discuss in the following section.
## Generation of banana residues
A banana plant is a tall and sturdy herbaceous plant, with a succulent and very juicy tubular stem, composed of leaf-petiole sheaths consisting of long and strongly overlapping fibers called pseudostem. Each pseudostem bears fruit only once, before dying and being replaced by a new one; this pseudostem consists of concentric layers of a leaf sheath and a crown of large leaves.
Banana biomass mainly consists of four elements, namely: pseudostems, leaves, rachis, and skins. Additionally, a significant number of rejected bananas provide starchy feedstock to feed biorefineries. Feedstock derived from rejected bananas can reach up to 30 wt% of the total production (remaining unsold overripe fruits also fall in this category). All these biomass residues are normally dumped in rivers, oceans, landfills, and unregulated dumping grounds, creating huge decaying deposits that can lead to the spread of diseases, contamination of water sources, generation of foul odors, and attraction of rodents, insects, and scavengers. Some of the possible ways that enable the utilization of banana waste include compost production and food wrapping. However, these solutions do not always prevent the material from reaching the wasteland after serving its purpose. Recently, a craft type paper of good strength has been made from crushed, washed, and dried banana pseudostems.
## Banana residues in central and latin america
Banana is one of the most cultivated fruit crops worldwide (~106.7 million tons of production in 2013). Many industries take advantage of banana pulp, but discard lignocellulosic biomass, including pseudostems, stalks, leaves (normally found in the field), and rachis of the fruit bunches (gathered usually in the packing plants). Leaves, the pseudostem, stalk, and peel generate a huge amount of waste; for instance, banana peels account for more than 41.3 million tons per year, therefore serving as a potential biomass feedstock.
The estimated amount of agricultural residue available in Central America in 2011 was about 192 Petajoules (PJ); the countries with the highest energy potentials are Guatemala with 79 PJ and Honduras with 29 PJ, followed by Costa Rica with 22 PJ. Banana residues represent an important fraction of these wastes, as the Central American region provides excellent environmental conditions for optimal development for the banana plant. In this fashion, banana crops rank in the top six agriculture residues in countries such as Belize, Costa Rica, Guatemala, Honduras, and Panama.
Nevertheless, Central American producers are far more focused on commercializing the crop itself than valorizing the corresponding waste. In 2011, about 2.9 million tons of banana residue (wet basis) were produced in Central America. However, it is worth mentioning that these residues cannot be fully recovered, as part of them must be left in situ to avoid soil degradation (i.e., reduction of carbon stock in the soil), while other residues have found uses as fertilizers, fodder, and domestic fuel. Nonetheless, there is still a large proportion that can find applications as biorefinery feedstocks.
Other Latin American nations face similar realities when it comes to banana production. For instance, Brazil produces around 82.8 million tons of bananas annually; each produced ton leaves behind 100 kg of rejected fruit and some 4 tons of lignocellulosic waste (3 tons pseudostems, 480 kg leaves, 160 kg rachis, and 440 kg skins).
Nations such as Ecuador have come up with a series of initiatives regarding bioethanol production using lignocellulosic biomass from banana crops. For instance, Guerrero and collaborators developed a process of production of bioethanol from banana rachis with a positive energy balance.presents schematics of the production and use of second-generation ethanol from banana waste and its further blending with regular gasoline, this study employed a Well-to-Wheel (WtW) perspective and concludes that this strategy has great potential to reduce greenhouse gas emissions and fossil depletion, as a consequence of an overall positive energy balance for the process. Bioethanol production from banana wastes is further discussed in Section 3.1 of this review.
## Potential biorefinery use of different banana residues
Before addressing the possible ways to convert banana residues into high-value products through biorefinery, it is important to describe their physicochemical properties, as this information allows for their maximum exploitation as raw materials and would help in developing more eco-friendly approaches too. Banana peel and rachis waste are composed mainly of biopolymers such as lignin, pectin, cellulose, hemicellulose, fiber, proteins, and some low-molecular-weight compounds such as chlorophylls, phenolic compounds, water-soluble sugars, and minerals.shows the composition of banana peel and rachis on a dry matter basis. It is worth mentioning that the high moisture content of banana residues promotes their biodegradability before processing, thus affecting their handling, transportation, storage, and further uses in biorefinery technologies. The composition of fruit peel residues varies according to species, seasonal variations, geographic location, variety, and stage of maturation. Lignin contents are greater in banana rachis, and banana leaves are rich in holocellulose, hemicellulose, and lignin, all promising compounds for biorefinery processing. Lignin is particularly valuable, accounting for 25 wt% of banana leaves, which is higher than other important agro-industry residues such as cotton or straw. Banana rachis and pseudostem residues can be used as biomass feedstocks, both biomasses have a high content of carbohydrates such as hemicellulose, starch, and lignin. Banana stem residues also contain lignins, glucans, and most abundantly xylans and ashes.
Though mechanicaland hydrothermal pretreatments of these residues are still energy and time consuming, they have proved necessary for further steps in biorefinery operations. For instance, steam explosion pretreatment increases cellulose content compared to raw materials (from 20.1 to 54.4 wt% going from raw to pretreated pseudostem, and 26.1 to 57.1 wt% going from raw to pretreated rachis); pretreatment also increases free glucose content for further biotransformations.
## High-value products obtained from banana residues
As shown above, a large number of parts of banana residues can be used as biomass sources; chemical composition of the raw material in question will determine its further suitability as a biomass source. In this section we describe a series of high-value products obtained from banana residues via biorefinery technologies.
## Biofuels
Chemical composition of banana stems provides an indicator of their feasibility for production of fermentable sugars as a function of moisture content, as well as cellulose and hemicellulose contents. In fact, saccharification and further fermentation of banana lignocellulosic content for ethanol production has been extensively investigated. Research by Duque and collaborators shows that the potential of ethanol production is 0.259 kg of ethanol per kg of banana stem raw material. Research by Guerrero and co-workers has found high solid loading, low enzyme dosage and a short period conversion process as optimal conditions for bioethanol production from banana pseudostem and rachis, yielding ethanol solutions of 4.0 v/v % (87% yield) and 4.8 v/v % (74% yield), respectively.
Another study by Ingale and co-workers also employed banana stem waste and found that alkali treated banana pseudostems followed by enzymatic saccharification yield higher contents of reducing sugars than those alkali treated only, the corresponding increase in ethanol production was observed. Not only has lignocellulosic waste from banana production been transformed into ethanol, but fermentation of banana pulp and fruit can yield comparable biofuel efficiencies as corn.
Even though further technological improvements are still required, bioethanol yield from banana waste presents a promising alternative for the production of this biofuel. Further research on improved enzymatic cocktail formulations, more robust microorganism strains, as well as optimization of industrial conditions, such as reaction time, water content, and ethanol separation technologies, are still required.
## Fibers for mechanical reinforcement
Banana fiber has traditionally found a place in a number of manufactured products such as paper, ropes, table mats, and handbags. Though these materials are inexpensive, biodegradable, and produced from renewable sources, biorefinery approaches offer the possibility to generate higher value outcomes from their raw material. In this fashion, lignocellulosic micro/nanofibers (LCMNF) can be produced from banana leaf residues; Tarrés and co-workers' results show that banana leaves can yield up to 82.44% LCMNF, a significantly high value compared to other agricultural wastes that typically yield around 15%.
LCMNFs from banana leaf residue have been used to restore mechanical properties of recycled fluting paper. A study found that incorporation of only 1.5 wt% of LCMNF can restore the original properties of fluting paper, with a low impact in pulp drainability, while increasing the life span of the resulting recycled products.
## Nanocellulose fibers
Banana residues are a great source of cellulosic materials (see; cellulose provides stiffness and strength to the plants' structure, and approximately a third of the plant's anatomy is composed of this polysaccharide. Since cellulose is greatly present in banana peel and rachis, those represent biomass sources suitable to obtain nanocellulose fibers (NCFs).
NCFs exhibit many attractive physicochemical properties such as a high bending strength (~10 GPa), a Young's modulus of approximately 150 GPa, a high aspect ratio, and a high specific surface area. Therefore, NCFs have been used as reinforcements for polymer matrixes, and as additives for papermaking. Suspensions of NCFs improve the mechanical strength and density of paper while reducing its porosity. The surface of NCFs is decorated with polar hydroxyl groups, which confer high moisture adsorption capacity and surface reactivity. NCFs present a strong potential as oil-water suspension stabilizers in the food industry; as mechanical reinforcementin drug delivery, enzyme supports, biosensors, and scaffolds for tissue engineering applications.
Acid hydrolysis of cellulose is the most common process for obtaining NCFs, as fractions containing amorphous material can be hydrolyzed with HCl and sulfuric acid, while those containing crystalline cellulose are typically recovered by centrifugation. NCFs have been isolated from banana peel using different processes, involving alkaline treatment and bleaching, followed by acid hydrolysis with sulfuric acid and high-pressure homogenization. Transmission electron microscopy (TEM) and atomic force microscopy (AFM) investigations have revealed the clearance of large amounts of amorphous materials to afford highly crystalline NCFs, these NCFs showed good cytocompatibility with human epithelial colorectal adenocarcinoma cells (Caco-2 cell line) in concentrations below 1000 mg/mL, these NCFs exhibit promising features as reinforcement material in composites too.
## Bioplastics
Poly-hydroxybutyrates (PHBs) are value-added biocompatible, biodegradable, thermoplastic biopolymers that can be synthesized by microorganisms from diverse carbon sources. Polysaccharides in banana peel can be either chemically or microbiologically transformed into PHBs, these biopolymers are hydrophobic, and bear similar mechanical properties to polypropylene or polyethylene. Getachew and collaborators have reported on a series of strains of Bacillus sp. able to yield up to 27 w/w % of PHB content after the fermentation of hydrolyzed banana peel residues. PHB production from banana waste is still not affordable on an industrial scale, though efforts to couple this production as part of a multiproduct biorefinery are moving the field in this direction; for instance, Naranjo and coworkers reported on how such a kind of integration might save energetic costs and water waste via the fermentation of banana peel hydrolysates using Burkholderia sacchari IPT101.
Poly-(l-lactic acid) (PLA) is a biodegradable and renewable polyester with many industrial and biomedical applications, including drug delivery systems, bioabsorbable fixation devices, bone regeneration, and tissue engineering scaffolds. PLA has been obtained through fermentation of banana (and also pineapple) waste hydrolisates using Lactobacillus casei (subspecies rhamnosus), and further microwave-assisted polymerization, as reported by Jiménez-Bonilla and collaborators. This direct melt polycondensation method afforded PLA oligomers with low oxidation losses, better stereopurity and lower energetic cost than conventional heating methods.
## Enzymes and food additives
Banana stalk residues have been valorized in the bio assisted production of enzymes like laccase, different oxidases, and endoglucanases too. For instance, Reddy and co-workers investigated the use of Pleurotus ostreatus and P. sajor-caju, to produce lignolytic and cellulolytic enzymes such as laccase, lignin peroxidase, xylanase, endo-1,4-β-D-glucanase (CMCase), and exo-1,4-β-D-glucanase using banana wastes as solid substrate fermentation. Both microorganisms originated comparable levels of enzyme activities and patterns of production.
Leaf biomass was found to be an appropriate substrate (compared to pseudostems) for enzyme production. Banana peel extracts have been studied as antioxidants in fresh orange juices, finding that free radical scavenging capacity increased by adding banana peel extracts to juice formulations. In addition, remarkable increases in antioxidant capacity using 2,2 -azino-bis-(3-ethylbenzothiazoline)-6-sulfonic acid (ABTS) radicals were observed when equal or greater than 5 mg of banana peel extract per ml of freshly squeezed juice was added, though no clear effects were observed in its ability to reduce the extent of lipid peroxidation.
## Additional uses of banana residues
## Energy generation
There seems to be a worldwide agreement on shifting towards green energy production and lessening the current dependence on fossil fuels. Agro-industrial biomass sources are a natural choice when it comes to exploring environmentally friendly ways to produce energy, and banana wastes are no exception to this. Banana residues have made it into the energy production sector only in recent times and their potential as energetic biomass is becoming evident; for instance, energy generation from dry banana peel can yield up to 18.89 MJ/kg. The reader is invited to consult referencefor an extensive assessment of banana biomass as an energy source in the Central American region.
Currently, there are two approaches for the conversion of banana biomass into energy: thermal and biological conversion. The former involves direct combustion and gasification, while the latter involves anaerobic digestion as shown in. Compared to other types of waste substrates (such as human sewage, piggery, or feedlot waste), banana residues produce a very clean form of biogas (mostly made of methane and carbon dioxide, with little noxious odors). Pisutpaisal and collaborators have demonstrated that size reduction of banana peel raw material and its fungal pretreatment might improve methane yield. Theoretical estimates for potential power generation based on both banana waste and banana peels in Malaysia (by Tock and coworkers for the period 2003-2008) suggest that banana biomass is a suitable renewable energy source in this one and other similar tropical nations. This study calculated potential power assuming that 1 PJ can be converted into 46 MW of electrical energy with 21% electrical conversion efficiency. This study estimates that using whole banana residues might generate some 80-95 MW yearly, while banana peels only would generate 12-25 MW per year (15% and 25% of energy production using the whole residue) as detailed in.
## Water treatment
Banana peel has been reported to be used as a bio adsorbent for the removal of contaminants such as heavy metals, dyes, and organic pollutants from wastewaters. A study by Pathak and co-workers reports on the adsorptive removal of benzoic acid (BA) and salicylic acid (SA) using banana peel; the authors report on removal efficiencies between 60-80% for the removal efficiency of these contaminants, with the advantage of possible reuse of the banana peel adsorbent in gasification for power generation(though adsorbing contaminants that may be detrimental for biogas quality must be avoided). Banana pith has been used to produce activated carbons to be employed in divalent heavy metal cations and dye removal from aqueous solutions with satisfactory results.
# Conclusions
Banana plants exhibit high growth rates and carbon neutrality; therefore, their agro-industrial residues are a promising alternative to be used as feedstock for biorefinery technologies, though challenges regarding composition variations in the wastes must be addressed, i.e., geographic location, plant variety, ripening stage, among others that difficult standardization for biorefinery processing.
Current technologies require further developments in order to extract as much value as possible from banana waste, for instance, by achieving positive energy balances by full integration of different biorefinery processes. Efforts involving the production of bioethanol and biogas are currently moving in this direction. Although efforts are being made on the production of enzymes and food additives derived from different components of banana waste, the most promising potential for these residues rely on high-value biopolymers, i.e., micro and nanofibrillar mechanical reinforcements such as nanocellulose fibers, as well as the biotechnologically and chemically assisted production of bioplastics such as PHBs and PLA. More research on biorefinery approaches will be required if banana residues' biomass potential is to be increased. As a result of this, small communities from developing countries, as well as agro-industrial, chemical, and pharmaceutical industries, are most likely to benefit.
## Conflicts of interest:
The authors declare no conflict of interest. |
Spermine and Citrate as Metabolic Biomarkers for Assessing Prostate Cancer Aggressiveness
Separating indolent from aggressive prostate cancer is an important clinical challenge for identifying patients eligible for active surveillance, thereby reducing the risk of overtreatment. The purpose of this study was to assess prostate cancer aggressiveness by metabolic profiling of prostatectomy tissue and to identify specific metabolites as biomarkers for aggressiveness. Prostate tissue samples (n = 158, 48 patients) with a high cancer content (mean: 61.8%) were obtained using a new harvesting method, and metabolic profiles of samples representing different Gleason scores (GS) were acquired by high resolution magic angle spinning magnetic resonance spectroscopy (HR-MAS). Multivariate analysis (PLS, PLS-DA) and absolute quantification (LCModel) were used to examine the ability to predict cancer aggressiveness by comparing low grade (GS = 6, n = 30) and high grade (GS$7, n = 81) cancer with normal adjacent tissue (n = 47). High grade cancer tissue was distinguished from low grade cancer tissue by decreased concentrations of spermine (p = 0.0044) and citrate (p = 7.73?10 24 ), and an increase in the clinically applied (total choline+creatine+polyamines)/citrate (CCP/C) ratio (p = 2.17?10 24 ). The metabolic profiles were significantly correlated to the GS obtained from each tissue sample (r = 0.71), and cancer tissue could be distinguished from normal tissue with sensitivity 86.9% and specificity 85.2%. Overall, our findings show that metabolic profiling can separate aggressive from indolent prostate cancer. This holds promise for the benefit of applying in vivo magnetic resonance spectroscopy (MRS) within clinical MR imaging investigations, and HR-MAS analysis of transrectal ultrasound-guided biopsies has a potential as an additional diagnostic tool.
# Introduction
Currently there are no objective clinical tools that can accurately discriminate aggressive from indolent prostate cancer. The Gleason scoring systemis the most important prognostic tool in treatment planning, but it is dependent on subjective factors in the evaluation of aggressiveness and is limited by underestimation due to under-sampling of biopsies. New diagnostic and prognostic tools for evaluating prostate cancer aggressiveness are therefore urgently needed. Metabolic alteration is an emerging hallmark of cancer [bib_ref] Hallmarks of Cancer: The Next Generation, Hanahan [/bib_ref] , and metabolic profiling of prostate tissue using magnetic resonance spectroscopy (MRS) can provide additional information about tumor behaviour [bib_ref] Magnetic Resonance Metabolomics of Intact Tissue: A Biotechnological Tool in Cancer Diagnostics..., Bathen [/bib_ref] , especially with the possibility to translate findings from ex vivo tissue samples to in vivo measurements in patients using MRS imaging (MRSI).
Metabolic differences between prostate cancer and normal tissue are documented both in vivo by MRSI [bib_ref] Quantitative 1H MR spectroscopic imaging of the prostate gland using LCModel and..., García-Martín [/bib_ref] [bib_ref] Standardized Threshold Approach Using Three-Dimensional Proton Magnetic Resonance Spectroscopic Imaging in Prostate..., Fütterer [/bib_ref] [bib_ref] Multiparametric magnetic resonance imaging in prostate cancer: present and future, Kurhanewicz [/bib_ref] [bib_ref] Three-dimensional proton MR spectroscopy of human prostate at 3 T without endorectal..., Scheenen [/bib_ref] and ex vivo using high resolution magic angle spinning MRS (HR-MAS) [bib_ref] Proton HR-MAS spectroscopy and quantitative pathologic analysis of MRI/3D-MRSI-targeted postsurgical prostate tissues, Swanson [/bib_ref] [bib_ref] Quantitative analysis of prostate metabolites using 1H HR-MAS spectroscopy, Swanson [/bib_ref] [bib_ref] Evaluation of lactate and alanine as metabolic biomarkers of prostate cancer using..., Tessem [/bib_ref]. In some hospitals, MRSI has already been implemented into clinical practice, making use of the (total choline+creatine+polyamines)/citrate (CCP/C) ratio or the (total choline+creatine)/ citrate (CC/C) ratio which is increased in malignant prostate tissue [bib_ref] Standardized Threshold Approach Using Three-Dimensional Proton Magnetic Resonance Spectroscopic Imaging in Prostate..., Fütterer [/bib_ref] [bib_ref] Proton HR-MAS spectroscopy and quantitative pathologic analysis of MRI/3D-MRSI-targeted postsurgical prostate tissues, Swanson [/bib_ref] [bib_ref] Prostate MRI and 3D MR Spectroscopy: How We Do It, Verma [/bib_ref] [bib_ref] In Vivo Assessment of Prostate Cancer Aggressiveness Using Magnetic Resonance Spectroscopic Imaging..., Kobus [/bib_ref]. The total choline signal measured in vivo can be separated by HR-MAS into the choline-containing metabolites [free choline (Cho), phosphocholine (PCho) and glycerophosphocholine (GPC)] [bib_ref] Proton HR-MAS spectroscopy and quantitative pathologic analysis of MRI/3D-MRSI-targeted postsurgical prostate tissues, Swanson [/bib_ref] [bib_ref] Quantitative analysis of prostate metabolites using 1H HR-MAS spectroscopy, Swanson [/bib_ref] [bib_ref] Correlation of phospholipid metabolites with prostate cancer pathologic grade, proliferative status and..., Keshari [/bib_ref]. Lactate and alanine are also reported to be increased in cancer compared to normal tissues [bib_ref] Evaluation of lactate and alanine as metabolic biomarkers of prostate cancer using..., Tessem [/bib_ref] , while the prostate-specific metabolites citrate and the polyamines (spermine, spermidine, and putrescine) are found in lower concentrations in cancer tissue [bib_ref] Quantitative analysis of prostate metabolites using 1H HR-MAS spectroscopy, Swanson [/bib_ref] [bib_ref] Non-destructive quantitation of spermine in human prostate tissue samples using HRMAS 1H..., Cheng [/bib_ref].
HR-MAS is a well-established technique for analyzing biological tissue, leaving the samples unprocessed for subsequent histopathological evaluation or other molecular methods such as gene expression profiling [bib_ref] Changes in Gene Transcription Underlying the Aberrant Citrate and Choline Metabolism in..., Bertilsson [/bib_ref] [bib_ref] Metabolic, pathologic, and genetic analysis of prostate tissues: quantitative evaluation of histopathologic..., Santos [/bib_ref]. We have previously confirmed that there is a significant correlation between results from ex vivo HR-MAS analyses and in vivo MRSI from spatially matched regions, proving that the translation from ex vivo to in vivo is valid [bib_ref] Spatially matched in vivo and ex vivo MR metabolic profiles of prostate..., Selnaes [/bib_ref]. The overall aim of this study was to investigate the possibility of assessing prostate cancer aggressiveness by HR-MAS analysis of human prostate tissue, and to identify specific metabolites as biomarkers for cancer aggressiveness. The study was performed using fresh frozen tissue samples extracted from radical prostatectomy specimens using a novel method allowing samples with a high cancer content to be included [bib_ref] A new method to provide a fresh frozen prostate slice suitable for..., Bertilsson [/bib_ref]. Both metabolic profiles and individual metabolite concentrations were used to discriminate between the histologically determined Gleason score (GS) which was evaluated from a cryosection of each tissue sample. The value of HR-MAS as an additional tool to complement histopathological scoring, and the improvement the results add to in vivo MRSI examinations, will be discussed.
# Materials and methods
## Patient and tumor characteristics
Since 2007, all prostate cancer patients at St. Olavs Hospital, Trondheim University Hospital, Norway, scheduled for radical prostatectomy have been invited to sign an informed consent form to donate tissue for research. From each patient a 2 mm transversal prostate tissue slice has been collected for storage in the Regional Research Biobank of Central Norway. The study has been approved by the Regional Committees for Medical and Health Research Ethics (REC) Central, Norway, and the Data Inspectorate of Norway. The current study includes 48 patients with no previous prostate cancer treatment and with a tumor volume .5% of the gland, estimated by histopathology. Patient characteristics are described in [fig_ref] Table 1: Characteristics of patients and prostate tissue samples [/fig_ref].
## Harvesting method and sselection of hr-mas samples
On average 15 minutes after surgical removal of the prostate gland, a tissue slice (2 mm) was obtained by transection through its middle, perpendicularly to the urethra [bib_ref] A new method to provide a fresh frozen prostate slice suitable for..., Bertilsson [/bib_ref]. The slice was snap frozen by clamping between two metal plates precooled in liquid nitrogen and stored at 280uC. The two remaining halves were stitched to a cork board, in order to avoid disturbances in the histopathological evaluation of the surgical margin. After fixation in formalin, both halves were further sliced (4 mm thick slices) and paraffin embedded. Microscopic sections were made and stained with hematoxylin, erythrosine and saffron (HES) for diagnostic purposes. The HR-MAS samples were excised from the frozen prostate slice using a novel harvesting method described by Bertilsson et al. [bib_ref] A new method to provide a fresh frozen prostate slice suitable for..., Bertilsson [/bib_ref]. By using this method, summarized in [fig_ref] Figure 1: The prostate sample harvesting method after radical prostatectomy [/fig_ref] , tissue samples of predetermined histopathological GS are obtained from the slice. During sample extraction, the frozen tissue slice was placed on an aluminium plate in direct contact with liquid nitrogen, preventing the tissue from thawing and thus reducing molecular degradation. Several samples from each slice (range: 1-7 samples per slice (median: 3) depending on tumor size) were selected from malignant areas of different GS and from normal adjacent areas, using the HES stained slides from neighboring tissue blocks as a guide. Thus, a total of 162 HR-MAS samples was obtained. Normal adjacent samples are defined as samples not showing signs of cancer, thus containing only benign glandular and/or stromal tissue, and these samples were excised as far away from the cancer as possible. To assess the GS of each HR-MAS sample (2 mm thick), and to determine the amount of cancer tissue, stroma, and glandular tissue, a 4 mm cryosection was cut from one side of the extracted sample and HES stained, and the tissue composition was evaluated by an experienced pathologist specialized in uropathology before the HR-MAS procedure. The samples were not thawed before the
## Hr-mas mrs experiments
A PBS solution (3 ml) containing trimethylsilyl 3-propionic acid sodium salt (TSP, 5 mM) and formate (25 mM) was added to disposable Kel-F HR-MAS inserts (30 ml, Bruker Biospin, Germany). Each prostate tissue sample (mean weight: 12.7 mg, range: 3.0-21.9 mg) was transferred to a HR-MAS insert using a sterile biopsy punch (2 mm, Miltex Gmbh, Germany), and the insert was placed into the zirconium rotor (4 mm). HR-MAS was performed on a Bruker Avance DRX600 (14.1 T) spectrometer (Bruker BioSpin, Germany) equipped with a 1 H/ 13 C MAS probe. Proton spectra were acquired at 4uC with a spin rate of 5 kHz. Pulse-acquired spectra were obtained with a presaturation delay of 3.0s and acquisition time of 3.27s. A Carr-Purcell-Meiboom-Gill (CPMG) spin echo sequence [90u-(t-180u-t) n -acquisition] was used to suppress signals from lipids and macromolecules with an effective echo time of 60 ms. One hundred and twenty-eight scans over a spectral region of 10 kHz were collected into 64k points for both sequences. The spectra were Fourier transformed with a line broadening of 0.30 Hz. Chemical shifts were referenced to the lactate peak (left peak of the doublet) at 1.336 ppm and a linear baseline correction was applied (Topspin 3.1, Bruker Biospin, Germany). Peak assignments were set according to the human metabolomics database and previous published papers using HR MAS on prostate tissue [bib_ref] Quantitative analysis of prostate metabolites using 1H HR-MAS spectroscopy, Swanson [/bib_ref] [bib_ref] Evaluation of lactate and alanine as metabolic biomarkers of prostate cancer using..., Tessem [/bib_ref] [bib_ref] Single-voxel oversampled J-resolved spectroscopy of in vivo human prostate tissue. Magnetic resonance..., Swanson [/bib_ref].
# Multivariate analysis
The spectral data between 1.46 and 4.66 ppm from the CPMG spectra were used for multivariate analysis. The spectra were normalized to an equal total area and peak aligned using icoshift [bib_ref] icoshift: A versatile tool for the rapid alignment of 1D NMR spectra, Savorani [/bib_ref]. Signals from ethanol contamination (3.65-3.69 ppm) were removed from the spectra together with those of lipid residuals at 1.60, 2.05, and 2.27 ppm. Preprocessing of the spectra was performed in MATLAB 7.8.0 (The Mathworks, Inc., USA). In addition to principal component analysis (PCA), partial least squares (PLS) regression and PLS discriminant analysis (PLS-DA) [bib_ref] PLS-regression: a basic tool of chemometrics, Wold [/bib_ref] were used to model the relationship between the MR spectra and tumor/patient characteristics (tissue composition, GS, serum PSA (sPSA), tumor volume, age and pT-stage). In order to avoid overfitting, double cross-validation was performed [bib_ref] Assessment of PLSDA cross validation, Westerhuis [/bib_ref]. A PLS model was built on training samples (80% of the data set) and used to predict the status of independent test samples (the remaining 20%). The optimal number of LVs (latent variables) to use in the model was determined by cross-validation of the training data and applied independently to the test data. Both the inner and outer loops of the double cross-validation procedure were repeated 20 times with different randomly chosen training and test sets, and the average results are presented. As several samples from each patient were analyzed, spectra from one patient were put in either the training or the test set. The variable importance was evaluated by variable importance in projection (VIP) scores [bib_ref] Performance of some variable selection methods when multicollinearity is present, Chong [/bib_ref]. Variables with a VIP score greater than one are generally considered to be important The classification results were validated by permutation testing (n = 1000, significance for p,0.05) [bib_ref] Assessment of PLSDA cross validation, Westerhuis [/bib_ref]. Multivariate analyses were performed in MATLAB using PLS_toolbox 6.2.1 (Eigenvector Research, Inc., USA).
## Absolute quantification of metabolites by lcmodel
The pulse-acquired spectra were quantified using LCModel [bib_ref] Estimation of metabolite concentrations from localized in vivo proton NMR spectra, Provencher [/bib_ref] [bib_ref] Correlations between in vivo 1H MRS and ex vivo 1H HRMAS metabolite..., Opstad [/bib_ref] based on a novel basis set of 23 metabolites. The basis set of simulated metabolite spectra was generated using NMRSIM (Bruker BioSpin, Germany), and the metabolites were quantified between 4.72 ppm and 20.8 ppm. The baseline was modeled with a cubic spline function with a maximum of two knots, and macromolecules were included in the fitting, simulated with single peaks including prior knowledge of line width, chemical shift, and relative amplitude. Small molecule metabolite and lipid chemical shifts were set as mean values based on an initial assignment of spectra from 10 samples of varying tissue type. For metabolites where some peaks were not clearly resolved in these spectra (GPC, GPE, glucose, and the amino acids), literature values were used [bib_ref] Quantification of choline-and ethanolamine-containing metabolites in human prostate tissues using 1H HR-MAS..., Swanson [/bib_ref] [bib_ref] Statistically Integrated Metabonomic2Proteomic Studies on a Human Prostate Cancer Xenograft Model in..., Rantalainen [/bib_ref] [bib_ref] Proton NMR chemical shifts and coupling constants for brain metabolites, Govindaraju [/bib_ref]. Ethanol, a contaminant in some samples, was included in the basis set for a successful subsequent fitting with the metabolite spectra. The metabolites were quantified according to formate and the concentrations are reported as mmol/kg wet weight. Full relaxation of formate was assured by using results from T1 relaxation measurements performed on six additional tissue samples.
## Statistical analysis of metabolite concentrations
Differences in metabolite concentrations between cancer and normal adjacent tissue, and metabolic differences related to aggressiveness (low grade (GS = 6) vs. high grade (GS$7)) were analyzed by linear mixed models, accounting for the effect of samples originating from the same patient. Individual comparison of samples of GS 6, 7, and 8-9, in addition to differences between samples of GS 3+4 and 4+3 were also tested. Analyses were performed in R (version 2.14.1, R Foundation for Statistical Computing) with the lme4 package. The data were log transformed prior to analysis in order to obtain normally distributed residuals. The Benjamini and Hochberg false discovery rate was used to correct for multiple testing. Adjusted pvalues,0.05 were considered significant.
# Results
## Samples
The PCA score plot of the CPMG spectra (n = 162) revealed four outlying samples. These samples were removed from the data set due to very high lipid concentrations and microscopic evidence of severe inflammation. Of the 158 samples included in this study, 47 were shown to contain only normal prostate tissue components, while 111 samples contained cancer tissue. The average cancer content was 61.8% (range: 10-100%) and 30 cancer samples were defined as low grade (GS 6) while 81 samples were defined as high grade (GS 7-9). Sample and patient characteristics are summarized in [fig_ref] Table 1: Characteristics of patients and prostate tissue samples [/fig_ref]. Representative HR-MAS spectra and the corresponding histopathological image of normal prostate tissue and cancer tissue with different Gleason grades are shown in [fig_ref] Figure 2: Representative HR-MAS spectra and corresponding HES stained prostate tissue samples with different... [/fig_ref].
## Metabolic profiles related to clinical parameters
The metabolic profiles were correlated to tissue composition (percentage of benign glandular tissue: r = 0.67, stroma: r = 0.70, and cancer: r = 0.77) (p,0.001). The metabolic profiles were not significantly correlated to the patient's sPSA level, tumor volume, age or pT-stage (p.0.05).
## Distinguishing cancer and normal adjacent tissue
Multivariate analysis. Based on the metabolic profiles, cancer and normal samples were separated with 86% correct classification using PLS-DA on independent test samples (sensitivity 86.9%, specificity 85.2%, p,0.001). A PLS model correlating the metabolic profiles to GS [fig_ref] Figure 3: Prostate cancer metabolic profiles are correlated to aggressiveness [/fig_ref] , A-B) separates the normal adjacent tissue samples from the cancer tissue samples. The loadings showed decreased levels of citrate, taurine and creatine, and an increase in GPC, PCho, Cho, and glycine in cancer compared to normal tissue.
Absolute quantification by LCModel. The quantified metabolite concentrations in cancer and normal tissue samples (n = 153) are shown in [fig_ref] Table 2: Metabolite concentrations [/fig_ref]. Five spectra were not quantified due to insufficient fitting caused by high lipid signals.
## Distinguishing low grade (gs = 6) and high grade cancer tissue (gs$7); correlation with the gleason system
Multivariate analysis. Metabolic profiles were correlated to GS with a correlation coefficient of r = 0.71 using PLS regression analysis (p,0.001) [fig_ref] Figure 3: Prostate cancer metabolic profiles are correlated to aggressiveness [/fig_ref]. When analyzing only the cancer samples, the metabolic profiles were correlated to GS with a correlation coefficient of r = 0.45 (p,0.001) [fig_ref] Figure 3: Prostate cancer metabolic profiles are correlated to aggressiveness [/fig_ref] , C-D). When dividing the samples into normal, high grade (GS$7) and low grade (GS = 6), correct classification by PLS-DA was 85.8% (sensitivity 89.3%, specificity 82.3%), 77.4% (sensitivity 84.4%, specificity 70.5%), and 65.8% (sensitivity 64.1%, specificity 67.6%), respectively.
Absolute quantification by LCModel. The concentrations of spermine and citrate were shown to be significantly different between low grade and high grade cancers, while no significant differences were detected for the other metabolites. The concentrations and statistical results for the significant metabolites are summarized in [fig_ref] Table 3: Metabolite concentrations [/fig_ref]. For further examination of the metabolite concentrations related to aggressiveness, metabolic differences between samples of GS 6, 7, and 8-9 were analyzed individually [fig_ref] Table 3: Metabolite concentrations [/fig_ref]. No significant differences between GS 7 and GS 8-9 were detected for any of the metabolites. In addition, no significant differences in metabolite concentrations were found between samples of GS 3+4 and 4+3 (p.0.05). The correlations between GS and the concentrations of spermine and citrate were r = 20.36 and r = 20.43, respectively.
The clinically relevant CCP/C ratio was significantly increased in high grade compared to low grade cancer samples [fig_ref] Table 3: Metabolite concentrations [/fig_ref]. In addition, a trend of different GPC/PCho ratios between low and high grade cancer samples was detected (p = 0.08). When examining metabolite concentrations related to aggressiveness, the percentages of benign glandular, stroma, and cancer tissue were included in the linear mixed models in order to correct for differences in tissue composition. However, none of the tissue types had a significant contribution to the statistical models (p.0.05), and the results are presented without correction for tissue composition.
# Discussion
In this study performed using prostate tissue with high cancer content, we have shown the possibility to separate low grade from high grade prostate cancer using metabolic profiling. Decreased concentrations of citrate and spermine were shown to be valid MR tissue biomarkers for prostate cancer aggressiveness, and the metabolic profiles were significantly correlated to the GS showing that aggressive cancers have an altered metabolism compared to indolent cancer. Surprisingly, the choline containing components were not increasing with GS, indicating that spermine and citrate are the main contributors to the clinically applied CCP/C ratio which increases with GS. In addition, this study confirms the separation between cancer and normal tissue, and the HR-MAS metabolic profiles were successfully separated with 86.0% correct classification.
Many prostate cancer patients diagnosed with indolent disease (GS 6) are eligible for inclusion in active surveillance programs. It is therefore desirable to separate this group from patients with higher grade cancers. Citrate concentrations could separate samples with GS 6 from both GS 7 and 8-9, while the difference in spermine concentrations was only significant between GS 6 and GS 8-9. Interestingly, none of the metabolites was significantly different between samples with GS 7 and GS 8-9, indicating that samples with GS 7 (intermediate risk patients) have a metabolic pattern similar to higher grade cancers. This finding supports the consensus that only patients with GS#6 should be included in active surveillance programs. Patients with GS 4+3 have worse prognosis than those with GS 3+4, however this study could not separate these clinically relevant subgroups.
Normal prostate epithelial cells produce and accumulate a large amount of citrate which is secreted as a major component of the prostatic fluid. Compared to normal tissue, decreased levels of citrate are previously observed in prostate cancer tissue by ex vivo MRS [bib_ref] Quantitative analysis of prostate metabolites using 1H HR-MAS spectroscopy, Swanson [/bib_ref]. Our study confirms and extends these findings by demonstrating a significant negative correlation with GS, and significant differences between low grade and high grade cancer tissue, between samples of GS 6 and GS 7, and between GS 6 and GS 8-9. This supports the highly clinically relevant hypothesis that the citrate concentration can distinguish between aggressive and indolent prostate cancer.
Our results confirm previous in vivo and ex vivo MRS studies showing that a decrease in polyamines is associated with prostate cancer [bib_ref] Proton HR-MAS spectroscopy and quantitative pathologic analysis of MRI/3D-MRSI-targeted postsurgical prostate tissues, Swanson [/bib_ref] [bib_ref] Non-destructive quantitation of spermine in human prostate tissue samples using HRMAS 1H..., Cheng [/bib_ref] [bib_ref] Detection of prostate cancer with MR spectroscopic imaging: an expanded paradigm incorporating..., Shukla-Dave [/bib_ref] [bib_ref] Proton MR spectroscopy of prostatic tissue focused on the detection of spermine,..., Van Der Graaf [/bib_ref]. Additionally, the very low putrescine concentration in our study confirms that the polyamine peak predominantly consists of spermine. Due to the significantly lower concentration of spermine in high grade compared to low grade tissue, we propose spermine as a discriminative MR biomarker for prostate cancer aggressiveness, and a focus to this should be considered using the CCP/C ratio in MRSI examinations. Today, spermine cannot be fully separated from the choline peak using MRSI, but due to rapid technological developments already in progress and higher field strengths (7T) making separation possible [bib_ref] Detection of fully refocused polyamine spins in prostate cancer at 7 T, Klomp [/bib_ref] , polyamines and especially spermine are potential biomarkers in clinical practice.
Surprisingly, there were no significant differences between high grade and low grade prostate cancer in any of the quantified choline-or ethanolamine-containing metabolites (Eth, PE and GPE). Previous ex vivo studies have demonstrated significant correlations between GS and choline and total choline [bib_ref] High resolution magic angle spinning NMR spectroscopy for metabolic assessment of cancer..., Van Asten [/bib_ref] , and significantly higher concentrations of GPC in high grade (GS$4+3) compared to low grade (GS#3+4) cancers [bib_ref] Correlation of phospholipid metabolites with prostate cancer pathologic grade, proliferative status and..., Keshari [/bib_ref] , which is not in accordance with our findings. We found a trend towards significance for the GPC/PCho ratio (p = 0.0832), which indicates a change in the choline-containing metabolites associated with increased aggressiveness, however not detected when examining the metabolites individually. Due to contradictory findings of choline metabolism also in other types of cancers [bib_ref] Glycerophosphocholine (GPC) is a poorly understood biomarker in breast cancer, Moestue [/bib_ref] , the choline metabolism related to cancer aggressiveness evidently needs further evaluation.
Previous in vivo MRSI studies have concluded a trend towards a correlation between the CCP/C ratio and prostate cancer aggressiveness [bib_ref] In Vivo Assessment of Prostate Cancer Aggressiveness Using Magnetic Resonance Spectroscopic Imaging..., Kobus [/bib_ref] [bib_ref] Correlation of Proton MR Spectroscopic Imaging with Gleason Score Based on Step-Section..., Zakian [/bib_ref] , and our study showed a highly significant difference in the CCP/C ratio between low and high grade cancers. Our findings on the individual metabolites, however, indicate that the decreased CCP/C ratio observed in vivo is mainly resulting from decreased citrate levels.
Although there was a correlation between the metabolic profiles and tissue composition, correction for tissue composition in the analysis of individual metabolite concentrations was not significant. This indicates that the metabolic differences between high and low grade prostate cancer samples are present independently of tissue composition. It is however likely that samples with lower cancer content would require statistical methods correcting for tissue composition. A strength of this study is the inclusion of patients from the whole range of clinical stages, including patients with highly aggressive cancers. A limitation is however that the low grade tissue material (GS 6) was mainly acquired from patients having more aggressive tumors in the vicinity, and this may have induced metabolic perturbation in our low grade material. A sample cohort including more samples from patients with pure low grade cancer may provide even clearer metabolic differences between low and high grade cancers.
# Conclusion
Based on metabolic profiling of human prostate cancer samples this study shows that low and high grade prostate cancer tissue can be distinguished by the concentrations of spermine, citrate and the CCP/C ratio. In the future, by analyzing larger patient cohorts, concentration cut-off values can be determined for spermine and citrate, and models based on the metabolic profiles can become tools for assessing prostate cancer aggressiveness. HR-MAS is feasible as a diagnostic supplementary tool for evaluating transrectal ultrasound guided biopsies, providing metabolic profiles that can predict tumor aggressiveness. Ultimately, the translation from ex vivo measurements in tissue samples to a true non-invasive in vivo examination, rendered possible by improvements in MR technology, will be the main future goal. Thus, our results demonstrate the value of MRS in clinical treatment planning and as a tool for follow-up of patients included in active surveillance programs.
[fig] Figure 1: The prostate sample harvesting method after radical prostatectomy. (A) The two HES-stained sections adjacent to the tissue slice. (B) To localize the cancer and normal areas, micrographs of the two HES stained histological sections adjacent to the removed tissue slice were fused with a photograph of the frozen tissue slice. The regions of interest were marked and transferred to a transparency sheet to be used as a map for guiding sample extraction. (C) Cylindrical samples (3 mm diameter) for HR-MAS were excised from regions with normal tissue and cancer tissue with different Gleason grades. The Gleason grade and the percentages of benign glandular tissue, stroma and cancer tissue were verified by analyzing a 4 mm cryosection from each extracted sample. The figure is adapted from reference[36]. doi:10.1371/journal.pone.0062375.g001 [/fig]
[fig] Figure 2: Representative HR-MAS spectra and corresponding HES stained prostate tissue samples with different Gleason grades. Visual inspection of the spectra show decreased levels of polyamines (predominately spermine) and citrate, and increased levels of GPC, PCho, and Cho with increasing tumor grade. doi:10.1371/journal.pone.0062375.g002 Biomarkers for Prostate Cancer Aggressiveness PLOS ONE | www.plosone.org [/fig]
[fig] Figure 3: Prostate cancer metabolic profiles are correlated to aggressiveness. (A) PLS scores and (B) loadings of LV1 and LV2 from PLS regression correlating the metabolic profiles to GS with a correlation coefficient r = 0.71. The cancer samples are separated from the normal samples in the score plot, with the loadings showing metabolic alterations related to malignancy. Samples with GS 9 are almost completely separated from normal adjacent samples in the score plot, while some samples with a lower score overlap with the normal ones. The PLSDA model explains 48.2% of the x-variance and 53.7% of the y-variance (C) PLS scores and (D) the corresponding loading profile of LV1 from PLS regression of the cancer samples only, correlating the metabolic profiles to GS with a correlation coefficient r = 0.45. The resulting model explains 20.0% of the x-variance and 27.4% of the y-variance of the data. The loadings in (B) and (D) are colored according to their VIP score. S-ino; scyllo-inositol. doi:10.1371/journal.pone.0062375.g003 [/fig]
[table] Table 1: Characteristics of patients and prostate tissue samples. [/table]
[table] Table 2: Metabolite concentrations (mmol/kg) in cancer and normal prostate tissue samples.Concentrations are reported as mmol/kg wet weight. * p,0.05.Cramér Rao lower bound (CRLB, LCmodel uncertainty measure) lower than 20% of the concentration for more than 90% of the samples, which is acceptable for quantification[37,38]. Higher CRLB values are the result of near or actual absence of signals in some samples. b P-values from Linear mixed models corrected for multiple testing by Benjamini-Hochberg correction. doi:10.1371/journal.pone.0062375.t002 [/table]
[table] Table 3: Metabolite concentrations (mmol/kg) and ratios in low grade (GS = 6) and high grade (GS$7) prostate cancer samples and comparison between different GSs.Concentrations are reported as mmol/kg wet weight. a P-values from Linear mixed models corrected for multiple testing by Benjamini-Hochberg correction; * p,0.05. doi:10.1371/journal.pone.0062375.t003 [/table]
|
The cost-effectiveness of hospital-based telephone coaching for people with type 2 diabetes: a 10 year modelling analysis
Background: Type 2 diabetes (T2DM) is a burdensome condition for individuals to live with and an increasingly costly condition for health services to treat. Cost-effective treatment strategies are required to delay the onset and slow the progression of diabetes related complications. The Diabetes Telephone Coaching Study (DTCS) demonstrated that telephone coaching is an intervention that may improve the risk factor status and diabetes management practices of people with T2DM. Measuring the cost effectiveness of this intervention is important to inform funding decisions that may facilitate the translation of this research into clinical practice. The purpose of this study is to assess the cost-effectiveness of telephone coaching, compared to usual diabetes care, in participants with poorly controlled T2DM. Methods: A cost utility analysis was undertaken using the United Kingdom Prospective Diabetes Study (UKPDS) Outcomes Model to extrapolate outcomes collected at 6 months in the DTCS over a 10 year time horizon. The intervention's impact on life expectancy, quality-adjusted life expectancy (QALE) and costs was estimated. Costs were reported from a health system perspective. A 5 % discount rate was applied to all future costs and effects. One-way sensitivity analyses were conducted to reflect uncertainty surrounding key input parameters. Results: The intervention dominated the control condition in the base-case analysis, contributing to cost savings of $3327 per participant, along with non-significant improvements in QALE (0.2 QALE) and life expectancy (0.3 years). Conclusions: The cost of delivering the telephone coaching intervention continuously, for 10 years, was fully recovered through cost savings and a trend towards net health benefits. Findings of cost savings and net health benefits are rare and should prove attractive to decision makers who will determine whether this intervention is implemented into clinical practice.Trial registration: ACTRN12609000075280
# Background
Considerable economic burden is imposed by type 2 diabetes (T2DM), which is increasing in prevalence. Interventions that improve risk factor status and clinical guideline adherence may prevent complications and reduce the healthcare costs associated with T2DM. Given the wide array of interventions for the management of T2DM, decisions to fund and implement these should be informed by estimates of both efficacy and costeffectiveness. This ensures that patients are provided with treatments that represent the optimal use of scarce resources.
There is growing interest in telephone coaching interventions for people with T2DM. As suggested by the Diabetes Telephone Coaching Study (DTCS) [bib_ref] The effect of hospital-based telephone coaching on glycaemic control and adherence to..., Varney [/bib_ref] , these interventions may to improve the risk factor status and diabetes management practices of people with T2DM. The DTCS recruited 94 participants with poorly controlled T2DM (HbA1C greater than 7 %) from the Diabetes Clinic at St Vincent's Hospital Melbourne, an Australian tertiary hospital. Participants were randomised to usual care plus telephone coaching, or usual care alone for 6 months. Follow up occurred at 6 months (the end of the intervention period) and at 12 months (6 months after withdrawal of the intervention).
Diabetes coaching in this study was defined as the regular provision of telephone advice and coaching that addressed lifestyle modification, adherence to treatment schedules, goal setting and barriers to change. Specifically, monthly coaching sessions were delivered by a dietitian. Participants were encouraged to make changes to their diet and exercise habits; to discuss specific medication changes with their general practitioner (GP), and to adhere to the recommended schedule for foot checks, eye checks and vaccinations. Relevant goals were agreed upon at each coaching session and progress towards goal attainment was reviewed at subsequent coaching sessions. If goals were not achieved, barriers to goal attainment were identified and a plan that addressed these barriers was agreed. New goals were set as required. This process was repeated throughout the intervention.
The primary outcome, HbA1C at 6 months, was significantly lower among the intervention group compared to the controls, −0.8 %, 95 % confidence interval (CI) (−1.2 to −0.3) [bib_ref] The effect of hospital-based telephone coaching on glycaemic control and adherence to..., Varney [/bib_ref]. Other parameters that improved at 6 months included fasting glucose, diastolic blood pressure, physical activity and adherence to diabetes management practices. However, improvements observed at 6 months were not sustained at 12 months.
Although the DTCS did not show sustained benefits upon withdrawal of the coaching, numerous trials have indicated that the provision of ongoing follow-up and support facilitates the longer-term maintenance of intervention gains [bib_ref] Preventing weight regain after weight loss, Perri [/bib_ref] [bib_ref] Fit and Strong!: bolstering maintenance of physical activity among older adults with..., Hughes [/bib_ref] [bib_ref] Participants' perspective on maintaining behaviour change: a qualitative study within the European..., Penn [/bib_ref] [bib_ref] A self-regulation program for maintenance of weight loss, Wing [/bib_ref] [bib_ref] Effects of four maintenance programs on the long-term management of obesity, Perri [/bib_ref] [bib_ref] The COACH program produces sustained improvements in cardiovascular risk factors and adherence..., Jelinek [/bib_ref] [bib_ref] Nurse case management to improve glycemic control in diabetic patients in a..., Aubert [/bib_ref] [bib_ref] A 3-year randomized trial of lifestyle intervention for cardiovascular risk reduction in..., Eriksson [/bib_ref] [bib_ref] Long-term effects of a lifestyle intervention on weight and cardiovascular risk factors..., Wing [/bib_ref]. These trials strongly support the notion that if the telephone coaching was delivered on an ongoing basis, improvements observed at 6 months in the DTCS are likely to be maintained.
Extrapolating from the results of the DTCS, the present analysis sought to assess the cost-effectiveness of telephone coaching for patients with T2DM. While many telephone coaching trials have speculated regarding the potential cost-effectiveness of these interventions, few have measured changes in resource use [bib_ref] Cost-effectiveness of a telephone-delivered intervention for physical activity and diet, Graves [/bib_ref] , and fewer still have assessed cost-effectiveness . This is the first Australian study to assess the cost-effectiveness of telephone coaching in a population exclusively with T2DM. Measuring costs in an Australian context is important due to international differences in healthcare costs [bib_ref] U.S. health care spending in an international context, Reinhardt [/bib_ref] [bib_ref] Review of the cost of diabetes complications in Australia, Ray [/bib_ref]. Importantly, estimates of cost-effectiveness are relevant to funding decisions that facilitate the translation of research evidence into clinical practice.
# Methods
A cost utility analysis was undertaken to compare telephone coaching with usual care. Six-month outcome data from the DTCS were applied to the UKPDS Outcomes Model in order to predict marginal changes in risks of clinical events (myocardial infarction [MI], coronary heart disease [CHD], stroke, congestive heart failure [CHF], amputation, renal failure and blindness), years lived, quality-adjusted life years (QALYs) lived and costs. The analysis took a health system perspective, considering direct healthcare costs met by the Victorian State and Commonwealth Governments.
It was assumed that intervention group participants received telephone coaching for 10 years, with intervention costs maintained during each year that participants were predicted to survive. Although other telephone coaching trials have observed improved glycaemic control with 12 months of intervention [bib_ref] Nurse case management to improve glycemic control in diabetic patients in a..., Aubert [/bib_ref] [bib_ref] A telephone-delivered intervention to improve glycemic control in type 2 diabetic patients, Oh [/bib_ref] [bib_ref] Adherence to diabetes control recommendations: impact of nurse telephone calls, Kim [/bib_ref] [bib_ref] A brief, regular, proactive telephone "coaching" intervention for diabetes: rationale, description, and..., Sacco [/bib_ref] [bib_ref] Insulin adjustment by a diabetes nurse educator improves glucose control in insulin-requiring..., Thompson [/bib_ref] [bib_ref] A telephone-delivered intervention for patients with NIDDM. Effect on coronary risk factors, Kirkman [/bib_ref] [bib_ref] Pro-active call center treatment support (PACCTS) to improve glucose control in type..., Young [/bib_ref] , and the maintenance literature indicates that the provision of ongoing follow-up and support facilitates the longerterm maintenance of intervention gains [bib_ref] Preventing weight regain after weight loss, Perri [/bib_ref] [bib_ref] Fit and Strong!: bolstering maintenance of physical activity among older adults with..., Hughes [/bib_ref] [bib_ref] Participants' perspective on maintaining behaviour change: a qualitative study within the European..., Penn [/bib_ref] [bib_ref] A self-regulation program for maintenance of weight loss, Wing [/bib_ref] [bib_ref] Effects of four maintenance programs on the long-term management of obesity, Perri [/bib_ref] [bib_ref] The COACH program produces sustained improvements in cardiovascular risk factors and adherence..., Jelinek [/bib_ref] [bib_ref] Nurse case management to improve glycemic control in diabetic patients in a..., Aubert [/bib_ref] [bib_ref] A 3-year randomized trial of lifestyle intervention for cardiovascular risk reduction in..., Eriksson [/bib_ref] [bib_ref] Long-term effects of a lifestyle intervention on weight and cardiovascular risk factors..., Wing [/bib_ref] , the true effect of continuously delivering this intervention remains uncertain. Consequently, conservative assumptions were made concerning the impact of the intervention on HbA1C. Rather than assuming that HbA1C values observed at 6 months in the DTCS were maintained throughout the modelled time horizon, HbA1C values in each simulation year were predicted by the UKPDS model. Sensitivity analyses were also conducted to account for this uncertainty.
In the base-case analysis, a 5 % discount rate was applied to all future costs and benefits. This rate was varied in the sensitivity analyses to reflect uncertainty. A 10 year time horizon was chosen for the base-case analysis. This was varied in the sensitivity analyses to two, five and 15 years. The primary outcome was an incremental cost-effectiveness ratio (ICER), expressed as a cost per QALY saved.
The UKPDS Outcomes Model is a probabilistic, discrete-time computer simulation model that uses algorithms based on UKPDS data to predict the development of seven diabetes-related complications (MI, CHD, stroke, CHF, amputation, renal failure and blindness) and death. The model enables economic evaluations of interventions that affect risk factors in people with T2DM [bib_ref] A model to estimate the lifetime health outcomes of patients with type..., Clarke [/bib_ref]. In the present analysis, model subjects comprised participants of the DTCS, who entered the model with characteristics based on levels at the end of the six month intervention period. Missing data at 6 months were imputed using the last observation carried forward method, with values observed at baseline used to impute missing values at 6 months. The model also demands data concerning the risk factor status of participants at diagnosis of T2DM. This information was not available to investigators, therefore, it was assumed that these levels were the same as those recorded at the participant's baseline assessment in the DTCS. The model ran in one year cycles, for which the risks of complications and death were predicted. Predictions were made based on each participant's six month characteristics and risk factors that the model changed with time. The model accounted for event-related dependencies, whereby the presence of one complication (such as CHD) increased the likelihood of another (such as CHF) and furthermore increased the risk of death. Participants continued through the model for 10 cycles, or until death.
Key model inputs are summarised in the Additional file 1. Health utility values were updated following each model cycle and used to calculate QALYs at the end of the simulation period. Multiple complications were assumed to have an additive effect on quality of life. The health utility values assigned to participants were based on UKPDS data [bib_ref] Estimating utility values for health states of type 2 diabetic patients using..., Clarke [/bib_ref].
Costs were reported in 2012/13 Australian dollars. Costs were deflated to their net present value using the Health Price Index. The model applied acute and ongoing costs to events predicted to develop in the simulation period. These costs were sourced from Australian data [bib_ref] Estimating the cost of complications of diabetes in Australia using administrative health-care..., Clarke [/bib_ref].
A cost was also applied to participants without diabetesrelated complications. This cost reflected diabetes-related costs incurred by DTCS participants between baseline and 6 months of the study, and thus considered the cost of medications, general practitioner presentations, St Vincent's Hospital outpatient appointments, St Vincent's Hospital emergency department presentations and St Vincent's Hospital inpatient admissions. This six monthly cost was multiplied by a factor of two to estimate annual costs.
To account for the cost of the telephone coaching intervention, an annual discounted cost was applied to intervention group participants 'post-hoc'. This reflected staffing and telephone call costs and was added to the cost of intervention group participants during each simulation year they were predicted to survive. One-way sensitivity analyses were conducted to reflect uncertainty surrounding key input parameters [fig_ref] Table 1: Parameters varied in sensitivity analyses [/fig_ref].
All procedures followed in this study complied with requirements of the St Vincent's Hospital Human Research Ethics Committee.
# Results
The groups were balanced at entry into the model with the exception of HbA1C levels, these being lower in the intervention group, 7.8 % versus 8.7 %, p = 0.003 (reflecting the efficacy of the intervention delivered in the DTCS). In addition, intervention group participants were less commonly Asian/Indian and more commonly Caucasian. The groups differed in the number of years since they had suffered a stroke [fig_ref] Table 2: Characteristics of the simulated population [/fig_ref]. This difference reflected a finding from the DTCS showing that fewer intervention group participants had previously suffered a stroke, nil versus 8 (17 %). Based on data collected between baseline and 6 months of the DTCS, annual costs were applied to participants in each group to reflect the annual cost of treating participants without diabetesrelated complications. The mean (95 % CI) costs applied to intervention and control group participants were $6091 (2183-9998) and $3107 (2530-3683), respectively. To reflect the cost of delivering the telephone coaching intervention, a cost of $1286 was applied to intervention group participants during each simulation year they were predicted to survive.
Over 10 years, the model predicted that the intervention would dominate the comparator, contributing to
## Hba1c
Assumed that HbA1C at 6 months in the DTCS was maintained for one, two and five simulation years.
## Stroke
Assumed that no participants had a past history of stroke.
net health benefits at a lower cost. The intervention contributed to savings of over $3300 per participant and an incremental gain of 0.2 QALYs. Ten year discounted costs were $59,790 and $63,117 among intervention versus control group participants, respectively, while 4.88 and 4.68 discounted QALYs were lived by the two groups. The intervention contributed to an incremental gain in life expectancy of 0.3 years [fig_ref] Table 3: Findings from the base-case analysis [/fig_ref]. The model predicted that the between-group difference in HbA1C at entry into the model reduced over time . There was a trend toward higher health utility scores and lower annual treatment costs among intervention group participants [fig_ref] Figure 2 a: Predicted change in mean [/fig_ref]. [fig_ref] Table 4: Mean [/fig_ref] summarises the cumulative incidence of first events among participants in each group over 10 years. The 10 year risk of any complication was lower in the intervention group, 32 % versus 38 %. The risk of death was also lower among intervention group participants, All results presented as mean (95 % CI) unless otherwise specified. P values in bold < 0.05 and considered statistically significant 32 intervention group participants predicted to survive for 10 years compared to 30 controls. The intervention dominated the control condition under most conditions tested in sensitivity analyses. The largest savings were observed when the treatment costs of participants without complications were adjusted to reflect the lower limit of the 95 % CI surrounding this value [fig_ref] Table 5: Findings from the sensitivity analyses [/fig_ref].
# Discussion
The results of these analyses suggest that the cost of investing in telephone coaching would be fully recovered through cost savings over 10 years. Treatment costs were $3327 lower among intervention group participants. Savings were driven by lower costs associated with treating diabetes-related complications; the cost of treating these was almost $12,000 lower per intervention group participant. Intervention group participants also gained an additional 0.20 QALYs and 0.3 years of life over 10 years. Like cost savings, improvements in QALE and life expectancy were driven by reductions in the risk of complications. The 10 year risk of MI, CHF, any complication and death was lower among intervention group participants, with risk reductions of 24, 20, 13 and 16 % observed. Given that the DTCS was powered for the primary endpoint of change in HbA1C, it is likely that a much larger sample size would be required to demonstrate statistical significance for the economic analysis. Nevertheless, an intervention which would save over $3000 per patient over 10 years would result in substantial cost reductions across the health care system, even if the clinical endpoints were neutral.
Predicted cost savings and net health benefits were apparent despite conservative assumptions concerning the intervention's cost and its impact on glycaemic control. For instance, rather than assuming that HbA1C levels at 6 months in the DTCS were sustained in subsequent simulation years, trends in HbA1C were predicted by the model. Consequently, glycaemic control was predicted to deteriorate in both groups over time. Costs were also applied conservatively, with higher annual treatment costs applied to the intervention group to account for the cost of the telephone coaching intervention and other treatment costs that were higher in this group during the trial. Having applied these conservative assumptions, confidence in the validity of this study's findings is further enhanced.
Also enhancing confidence in the validity were results showing that predictions of cost savings were robust to most conditions tested in the sensitivity analyses. The
## Hba1c (%)
Year Intervention group Control Change in mean (95 % CI) HbA1C over 10 years greatest cost savings were observed when cost of treating participants without diabetes-related complications was adjusted to reflect the lower limit of the 95 % CI surrounding this value. However, cost savings disappeared when past history of stroke was controlled for, suggesting that this chance imbalance between the groups may have biased findings in favour of the intervention group. However, at a cost of $4365 per QALY, the intervention was considered highly cost-effective under this condition and therefore, should still prove attractive to decision makers considering whether this intervention should be implemented into routine clinical practice. Only one Australian study was identified as having assessed the cost-effectiveness of a telephone delivered, behaviour change counselling intervention in people with T2DM. The analysis drew upon data from a randomised controlled trial (RCT) which found that a 12 month telephone coaching intervention contributed to significant improvements in diet but not physical activity in participants with T2DM or hypertension [bib_ref] Telephone-delivered interventions for physical activity and dietary behavior change: an updated systematic..., Goode [/bib_ref]. Modelled over 10 years and compared with usual care, the intervention was not cost-effective, however, compared with existing practice, the intervention was considered cost-effective at a cost of $29,375 per QALY gained [bib_ref] Cost-effectiveness of a telephone-delivered intervention for physical activity and diet, Graves [/bib_ref]. Investigators in this study differentiated existing practice from usual care, noting that participants receiving usual care received more intervention (telephone calls for data collection, verbal feedback on dietary and exercise behaviour and written education material) than was typical under existing practice conditions.
The present economic analysis might be differentiated from that of Graves and colleagues in a number of respects. For instance, the RCT on which Graves and colleagues' economic analysis was based, recruited participants with either T2DM or hypertension [bib_ref] Telephone-delivered interventions for physical activity and dietary behavior change: an updated systematic..., Goode [/bib_ref]. Therefore, projections of costs and effects do not relate specifically to people with T2DM. Furthermore, the study extrapolated outcomes observed at 12 months in the RCT (namely the intervention's impact on physical activity) to predict cost-effectiveness over 10 years [bib_ref] Telephone-delivered interventions for physical activity and dietary behavior change: an updated systematic..., Goode [/bib_ref]. However, physical activity is less reliable as a marker of long-term outcomes in T2DM than HbA1C. Whereas prospective RCTs demonstrate a cause and effect relationship between HbA1C, morbidity and mortality (key drivers of costs and effects in people with T2DM), evidence concerning the impact of physical activity on such endpoints comes from epidemiological and cohort studies [bib_ref] Physical activity in relation to cardiovascular disease and total mortality among men..., Tanasescu [/bib_ref] [bib_ref] Physical activity and risk for cardiovascular events in diabetic women, Hu [/bib_ref] [bib_ref] Exercise capacity and body composition as predictors of mortality among men with..., Church [/bib_ref] [bib_ref] Low cardiorespiratory fitness and physical inactivity as predictors of mortality in men..., Wei [/bib_ref]. Therefore, the present economic analysis may provide a more reliable estimate concerning the cost-effectiveness of telephone coaching in people with T2DM.
No other telephone coaching trials were identified as having contributed to both cost savings and net health benefits in people with T2DM. However, comparison of findings from the present economic analysis with other telephone coaching studies is difficult, firstly, because most were conducted in other countries and secondly, because of methodological issues that limit the validity and generalisability of their findings. For instance, one study expressed the ICER as a cost per unit change in a surrogate endpoint [bib_ref] Cost effectiveness of a telephone intervention to promote dilated fundus examination in..., Schechter [/bib_ref] , another measured only costs [bib_ref] A randomized trial of a telephone care-management strategy, Wennberg [/bib_ref] and several conducted only within-trial economic analyses, failing to project outcomes over a sufficient time horizon to facilitate valid comparison with findings from this economic analysis [bib_ref] Cost effectiveness of a telephone intervention to promote dilated fundus examination in..., Schechter [/bib_ref] [bib_ref] A randomized trial of a telephone care-management strategy, Wennberg [/bib_ref] [bib_ref] Cost-effectiveness of automated telephone self-management support with nurse care management among patients..., Handley [/bib_ref] [bib_ref] Intervention costs and cost-effectiveness of a successful telephonic intervention to promote diabetes..., Schechter [/bib_ref] [bib_ref] Costs and effects of a telephonic diabetes selfmanagement support intervention using health..., Schechter [/bib_ref]. Comparison with results from other countries is invalid owing to international differences in health systems and healthcare costs [bib_ref] U.S. health care spending in an international context, Reinhardt [/bib_ref] [bib_ref] Review of the cost of diabetes complications in Australia, Ray [/bib_ref]. Therefore, this economic analysis makes a valuable contribution to knowledge concerning the costeffectiveness of telephone coaching in Australians with T2DM.
As with all modelling analyses, a degree of uncertainty surrounds predictions obtained through the extrapolation of data from a short-term clinical trial that never empirically assessed the intervention's impact on survival, event rates, costs or QALE. For instance, confounding may have been present due to the age difference between the intervention and comparator groups, but having randomised the groups in the original study, this is unlikely to have changed the conclusion that the DTCS would likely be highly cost-effective. Longer, prospective RCTs are required to validate predictions obtained in this study. Simulation models provide a parsimonious solution to the absence of such long-term prospective data and despite their limitations, are widely used to extrapolate outcomes beyond the conclusion of clinical trials. Having applied conservative assumptions, the best available simulation model and extensive sensitivity analyses, the validity of findings from this study might be enhanced. As our evidence is indirect, caution should be taken in interpreting the results.
Limitations also relate to the UKPDS Outcomes Model. Previous studies have indicated that this model over-estimates event rates and mortality risk in populations dissimilar to the one on which it was developed [bib_ref] Use of the UKPDS outcomes model to predict all-cause mortality in U.S...., Song [/bib_ref] [bib_ref] External validation of the ukpds outcomes model equations (UKPDS 68),and the UKPDS..., Mcewan [/bib_ref]. The model has not been validated for use in an Australian population which is multicultural. Furthermore, the model predicts only a limited range of complications and predicts only first, not subsequent events [bib_ref] A model to estimate the lifetime health outcomes of patients with type..., Clarke [/bib_ref]. However, given that no other simulation models have been validated for Australian populations, this model was considered the best available for the purpose of this economic analysis [bib_ref] An Australian cardiovascular risk equation for type 2 diabetes: the Fremantle Diabetes..., Davis [/bib_ref].
Other limitations relate to the measurement of costs and effects. Consistent with a health system perspective, only direct diabetes-related costs were considered. Therefore, societal costs (to individuals and carers through lost time, income and productivity) were not captured. In terms of effects, utility weights applied in this study were not determined empirically, but were instead sourced from the literature. It is likely that these values would differ from those that would be obtained had DTCS participants been surveyed directly.
Findings from this analysis should be considered transferable to Australians with long-standing, T2DM that is sub-optimally controlled. This population is substantial; self-reported data from 2007 to 08 have indicated that 3.8 % of Australians (787,500 people) are affected by T2DMand observational data have indicated that poor glycaemic control is common among Australians with T2DM [bib_ref] Glycemic control from 1988 to 2000 among U.S. adults diagnosed with type..., Koro [/bib_ref] [bib_ref] Patterns of glycaemic control in Australian primary care (NEFRON 8), Macisaac [/bib_ref] [bib_ref] Glucose, lipid, and blood pressure control in australian adults with type 2..., Kemp [/bib_ref] [bib_ref] Diabetes guidelines: easier to preach than to practise?, Bryant [/bib_ref].
# Conclusions
Interest in diabetes coaching interventions is growing, with numerous such studies listed on the Australian New Zealand Clinical Trials Registry, many of which are collecting real-time, cost-effectiveness data. The RCT on which the present economic analysis was based found that adding a six month telephone coaching intervention to the usual care regimen of participants with poorly controlled T2DM led to improvements in glycaemic control and a range of other parameters. This economic analysis has shown that under conditions of the basecase analysis and most sensitivity analyses, the intervention would contribute to net health benefits and cost savings. In assessing cost-effectiveness, this study has extended findings from the existing telephone coaching literature. Two sensitivity analyses did not predict cost savings, instead predicting that the intervention would be highly cost-effective at a cost of less than $10,000 per QALY. It has previously been stated that dominant interventions and interventions that cost less than $10,000 per QALY should only be ignored if 'decision-makers have very serious reservations about the evidence base or are facing insurmountable problems in relation to stakeholder acceptability or feasibility of implementation'. Findings from this study support the need for a longer, prospective multi-centre trial of telephone coaching to confirm both the clinical and economic benefits prior to implementation into routine clinical practice. Future research should also consider alternative coaching delivery methods, using online and mobile interactive tools.
[fig] Figure 2 a: Predicted change in mean (95 % CI) health utility over 10 years, and b Predicted change in mean (95% CI) cumulative costs over 10 years [/fig]
[table] Table 1: Parameters varied in sensitivity analyses [/table]
[table] Table 2: Characteristics of the simulated population [/table]
[table] Table 3: Findings from the base-case analysis [/table]
[table] Table 4: Mean (95 % CI) cumulative incidence of first events over 10 years [/table]
[table] Table 5: Findings from the sensitivity analyses [/table]
|
Cortical gyrification is abnormal in children with prenatal alcohol exposure
A B S T R A C TObjectives: Prenatal alcohol exposure (PAE) adversely affects early brain development. Previous studies have shown a wide range of structural and functional abnormalities in children and adolescents with PAE. The current study adds to the existing literature specifically on cortical development by examining cortical gyrification in a large sample of children with PAE compared to controls. Relationships between cortical development and intellectual functioning are also examined. Experimental design: Included were 92 children with PAE and 83 controls ages 9-16 from four sites in the Collaborative Initiative on FASD (CIFASD). All PAE participants had documented heavy PAE. All underwent a formal evaluation of physical anomalies and dysmorphic facial features. MRI data were collected using modified matched protocols on three platforms (Siemens, GE, and Philips). Cortical gyrification was examined using a semi-automated procedure. Principal observations: Whole brain group comparisons using Monte Carlo z-simulation for multiple comparisons showed significantly lower cortical gyrification across a large proportion of the cerebral cortex amongst PAE compared to controls. Whole brain comparisons and ROI based analyses showed strong positive correlations between cortical gyrification and IQ (i.e. less developed cortex was associated with lower IQ). Conclusions: Abnormalities in cortical development were seen across the brain in children with PAE compared to controls. Cortical gyrification and IQ were strongly correlated, suggesting that examining mechanisms by which alcohol disrupts cortical formation may yield clinically relevant insights and potential directions for early intervention.
# Introduction
Prenatal alcohol exposure (PAE) can cause a range of abnormalities including facial dysmorphology, growth deficiency, microcephaly, brain alterations, and neurocognitive deficits (reduced IQ scores, executive functioning impairments, etc.). Clinically, the effects of PAE manifest along a range of outcomes commonly referred to as fetal alcohol spectrum disorders (FASD). Although facial dysmorphology and growth deficiency represent overt exposure-related outcomes, subtle brain alterations and neurodevelopmental delays are a less-apparent but devastating set of outcomes. Neuroimaging has made significant contributions to research in FASD, particularly in regard to brain function and structure. Recently, several studies have demonstrated functional connectivity disruptions in PAE, [bib_ref] Default mode network dysfunction in adults with prenatal alcohol exposure, Santhanam [/bib_ref] [bib_ref] Global functional connectivity abnormalities in children with fetal alcohol spectrum disorders, Wozniak [/bib_ref] [bib_ref] Inter-hemispheric functional connectivity disruption in children with prenatal alcohol exposure, Wozniak [/bib_ref]. A larger body of work has shown a range of structural brain abnormalities [bib_ref] Brain dysmorphology in individuals with severe prenatal alcohol exposure, Archibald [/bib_ref] [bib_ref] Fetal alcohol spectrum disorders: extending the range of structural defects, Jones [/bib_ref] [bib_ref] Extensive deep gray matter volume reductions in children and adolescents with fetal..., Nardelli [/bib_ref] [bib_ref] Regional brain volume reductions relate to facial dysmorphology and neurocognitive function in..., Roussotte [/bib_ref] [bib_ref] Development of cortical and subcortical brain structures in childhood and adolescence: a..., Sowell [/bib_ref] [bib_ref] Magnetic resonance imaging of brain anomalies in fetal alcohol syndrome, Swayze [/bib_ref] ; for reviews, see [bib_ref] Fetal alcohol spectrum disorders: recent neuroimaging findings, Moore [/bib_ref].
Only recently has it become feasible to examine complex cortical structure in detail using MRI. Recently, a study examining young children showed a significant reduction in cortical folding (cortical gyrification) amongst participants with PAE compared to healthy controls [bib_ref] A study of cortical morphology in children with fetal alcohol spectrum disorders, De Guio [/bib_ref]. Abnormalities in cortical development were seen in several regions of interest (ROI) and broadly across the brain in measures that included sulcal index, sulcal depth, and fold opening. A second study examining cortical gyrification in adolescents with PAE showed reduced cortical folding in PAE compared to matched controls [bib_ref] Atypical cortical gyrification in adolescents with histories of heavy prenatal alcohol exposure, Infante [/bib_ref]. In this study, the differences were seen primarily in the bilateral insula and bilateral visual cortices on the both the medial and lateral aspects of the occipital lobe. Studies examining related measures of cortical development, such as cortical surface area, have also found evidence for altered development in those with PAE (Alcohol Related Neurodevelopment Disorder in this case) compared to controls [bib_ref] Cortical morphology in children with alcohol-related neurodevelopmental disorder, Rajaprakash [/bib_ref].
The current study sought to add to the existing literature by generating a large sample size from the Collaborative Initiative on Fetal Alcohol Spectrum Disorders (CIFASD) multi-site study. We hypothesized that this larger sample size and correspondent increase in statistical power might reveal more widespread abnormalities in cortical structure than previous smaller studies -primarily because analyses of this type necessitate multiple comparison corrections in order to conduct ROI or vertex-wise analyses. We also sought to test the robustness of cortical gyrification measurements across multiple samples on different MRI scanners.
Because a number of processes that are critical to cortical development and gyrification occur during early to mid-gestation, we expected that PAE would be associated with significant abnormalities in cortical structure when examined during childhood and adolescence. We hypothesized that cortices would be smoother (less mature) in those with PAE compared to controls and that abnormalities in gyrification would be associated with lower global cognitive functioning. The advent of new tools that facilitate quantification of cortical gyrification using standard MRI images and semi-automated processing streams will ultimately allow investigators to ask important questions about relationships between early insults to cortical development and cognition in populations with neurodevelopmental disorders such as FASD.
# Methods
## Participants
Participants were enrolled in the study as part of CIFASD, a multisite investigation of brain, neurocognitive, and physical developmental anomalies in FASD. Detailed information about the CIFASD project is available in a separate publication [bib_ref] Toward a neurobehavioral profile of fetal alcohol spectrum disorders, Mattson [/bib_ref] and at www. cifasd.org. For the current study, participants were recruited from four CIFASD sites (Los Angeles, San Diego, Minneapolis, and Atlanta) between 2012 and 2014. PAE histories were obtained through retrospective maternal report or social service, legal, or medical records. Control participants were recruited with flyers, mailings to control participants of previous non-CIFASD studies, online advertisements, and referrals from participants with FASD. Advertisements and flyers were placed in neighborhoods and online locations chosen to maximize the ethnic, racial, and socioeconomic diversity of the control participants so as to best match the participants with FASD. Control participants were screened by telephone, as were participants with FASD.
Participants were included in the PAE group if there was a history of heavy PAE (> 13 drinks/week or > 4 drinks on any one occasion during pregnancy) or when such exposure was suspected in a child with an FAS diagnosis. In some cases, detailed history about exposure amounts or patterns of exposure was unattainable and decisions about inclusion or exclusion were made on the available evidence. Children were considered to have heavy PAE if mothers were known to have alcoholism or were known to abuse alcohol during pregnancy. In all cases, alcohol was the predominant substance of abuse. Participants were included in the non-exposed control group if there was a reliable history of minimal (< 1 drink/week, never > 2 drinks on any one occasion) or no exposure during pregnancy.
Participants (PAE and controls) were evaluated using a standardized examination conducted by a member of the CIFASD Dysmorphology Core (KLJ) who was blinded to group. Based on criteria outlined previously [bib_ref] Accuracy of the diagnosis of physical features of fetal alcohol syndrome by..., Jones [/bib_ref] [bib_ref] Toward a neurobehavioral profile of fetal alcohol spectrum disorders, Mattson [/bib_ref] , the evaluation resulted in a determination of 1) Fetal Alcohol Syndrome (FAS); 2) non-FAS; or 3) a "deferred" status due to some criteria being met, but not enough to diagnose FAS. The CIFASD approach to diagnosis does not include partial-FAS or Alcohol Related Neurodevelopmental Disorder (ARND). FAS was diagnosed on the basis of two or more of the following key facial features: thin vermillion border, smooth philtrum, and short palpebral fissure lengthtogether with either microcephaly (occipital-frontal circumference ≤ 10%) or growth deficiency (height or weight ≤ 10%) or both. The deferred status was applied when an individual had A.) one key dysmorphic facial feature as described above or B.) microcephaly and growth deficiency, or C.) microcephaly or growth deficiency plus one additional minor non-facial physical malformation (railroad track ear, hockey stick palmar crease, etc.). A significant number of individuals without PAE (i.e. controls) received a "deferred" classificationhighlighting the fact that the presence of one dysmorphic feature is relatively common and not diagnostic in and of itself .
Exclusion criteria for all subjects included another developmental disorder, very low birthweight (< 1500 g), other medical condition affecting the brain (e.g. Epilepsy), severe psychiatric disability that would prevent participation (e.g. psychosis or mania), substance use by the participant, English as a second language, international adoption after age 5, or contraindications to MRI scanning. Traumatic brain injury (including head injury with brief loss of consciousness) was treated as an exclusion criterion. No participant had a loss of consciousness > 2 min. Participants were not excluded for autistic symptoms because these are common in individuals with FASD [bib_ref] Attention deficits and autistic spectrum problems in children exposed to alcohol during..., Aronson [/bib_ref] [bib_ref] Autism in fetal alcohol syndrome: a report of six cases, Nanson [/bib_ref] [bib_ref] Autism characteristics in children with fetal alcohol spectrum disorders, Stevens [/bib_ref] , but those who met diagnostic criteria for an autism-spectrum disorder were excluded.
Control participants were excluded for parent-reported history of PAE and for diagnosed psychiatric conditions. Parents or caregivers of all enrolled participants were administered the Diagnostic Interview Schedule for Children-IV (C-DISC-IV; [bib_ref] NIMH diagnostic interview schedule for children version IV (NIMH DISC-IV): description, differences..., Shaffer [/bib_ref]. Because the study applied pre-enrollment telephone screening for psychiatric disorders, the C-DISC-IV data for enrolled participants revealed only minimal parent-reported symptoms in the control group (2 had ADHD symptoms, 7 had Oppositional Defiant symptoms, 4 had Conduct Disorder symptoms, and one had depressive symptoms; no controls had anxiety disorder symptoms or other major psychiatric symptoms).
Psychiatric co-morbidity was not an exclusion criterion for participants with PAE because it is well-recognized that co-morbidity is an extremely common feature of FASD [bib_ref] Neuropsychiatric implications and long-term consequences of fetal alcohol spectrum disorders, Streissguth [/bib_ref]. Based on the C-DISC-IV data, 62 participants in the PAE group had ADHD symptoms, 40 had Oppositional Defiant symptoms, 15 had Conduct Disorder symptoms, 8 had anxiety disorder symptoms, and 4 had depressive disorder symptoms.
Participants were ages 9-16 at the time of MRI scanning. The vast majority of participants completed the neurocognitive evaluation and MRI on the same day. In a few cases, they were separated by a few days or weeks. A total of 175 participants (92 with PAE & 83 Controls) met inclusion criteria. [fig_ref] Table 1: Demographic characteristics of participants included in analyses [/fig_ref] contains the demographics for the participants who were included in the analyses after eliminating those with excessive movement and aberrant processing. A total of 7 participants were excluded from the analysis due to aberrant FreeSurfer processing, mostly due to the effects of excessive motion during scanning. The excluded participants were as follows: 5 male PAE (1 from Atlanta, 2 from Minnesota, 2 from San Diego), 1 male Control (Minnesota), and 1 female Control (San Diego) (see Results Section 3.2 for a more complete description).
All participants underwent an Institutional Review Board (IRB)approved informed consent process involving a parent or guardian as well as a separate assent process with the child. All study procedures were approved by the IRBs at each of the four sites. Participants were compensated for their time.
## Evaluations
Neuropsychological testing was conducted during one or two sessions by trained research assistants who were blind to participant group. Quality control methods included a video review of test administration procedures and a detailed scoring check for every 10th administration. From a larger battery of neuropsychological measures administered in CIFASD, only IQ is examined here (Differential Ability Scales -Second Edition (DAS-II) [bib_ref] Differential ability scales, Elliott [/bib_ref].
Demographic and historical data were acquired on all CIFASD participants. Substance exposure histories, Puberty Development Scale [bib_ref] A self-report measure of pubertal status: reliability, validity, and initial norms, Petersen [/bib_ref] , racial and ethnic background, and socioeconomic status (SES), measure by the Hollingshead Four Factor Index of Social Statusare examined here. These data are contained in [fig_ref] Table 1: Demographic characteristics of participants included in analyses [/fig_ref].
## Mri acquisition procedures
MRI data were acquired at four sites on scanners from three vendors: Children's Hospital of Los Angeles (Philips Achieva); University of California -San Diego (General Electric MR750); University of Minnesota and Emory University (both Siemens Tim Trio). Acquisition sequences were modeled on protocols developed for multi-site imaging by the Pediatric Imaging Neurocognition and Genetics (PING) group (Table 2) (http://ping.chd.ucsd.edu). The sequence included high resolution T 1 -weighted images, a T 2 -weighted set, 30-direction DTI, and gradient-echo EPI scans for resting-state fMRI. The acquisition parameters in [fig_ref] Table 2: MRI sequence and parameters [/fig_ref] are just those for data examined in the current set of analyses (T 1 ). Participants were not sedated for the MRI scan nor were their usual medications modified.
## Mri processing
2.4.1. T 1 cortical parcellation.
Cortical parcellation of the T 1 volume was performed using FreeSurfer version 5.3.0 (surfer.nmr.mgh.harvard.edu) [bib_ref] Cortical surface-based analysis, Dale [/bib_ref]. Processing included removal of non-brain tissue, automated Talairach transformation, segmentation, intensity normalization, tessellation of the gray matter/white matter boundary, topology correction, and surface deformation. Data were visually inspected by a trained operator to ensure accuracy. In the case of significantly aberrant FreeSurfer processing (typically caused by motion artifact), participant data were not manually edited but, instead, the data were excluded from the analyses.
## Local gyrification index
The Local Gyrification Index (LGI) is a validated add-on metric to FreeSurfer [bib_ref] A surface-based approach to quantify local cortical gyrification, Schaer [/bib_ref]. Briefly, a smoothed outer brain surface map that does not follow the convexities of the cortex is first defined by FreeSurfer. A second map of the pial surface closely follows the folds of the cortex. Overlapping 25 mm circular regions of interest (ROIs) are then defined on the smoothed outer surface map. These ROIs are then paired with matching 25 mm pial surface ROIs.
LGI is computed at each vertex as the ratio of "buried" cortex to the smoothed outer surface [bib_ref] A surface-based approach to quantify local cortical gyrification, Schaer [/bib_ref].
LGI can range between 1 and 5. An LGI of 5 indicates that there is 5 times more cortex contained within the sulci than the amount of cortex on the outer surface (representing a deeply folded region); in contrast, an LGI of 1 represents a totally smooth region of cortex [bib_ref] How to measure cortical folding from MR images: a step-by-step tutorial to..., Schaer [/bib_ref].
LGI is an improvement over twodimensional cortical folding models such as the Gyrification Index (GI) (simply the ratio between the white matter/gray matter boundary and the pial boundary of the brain determined by 2-D coronal sections [bib_ref] The human pattern of gyrification in the cerebral cortex, Zilles [/bib_ref] because it takes into account the three-dimensional nature of the cortical surface [bib_ref] A surface-based approach to quantify local cortical gyrification, Schaer [/bib_ref] and allows for regionally-specific measurements.
# Statistical analysis
Analyses were carried out with the MATLAB Statistics Toolbox (MATLAB, 2015), IBM SPSS Version 22, and FreeSurfer version 5.3.0. Subject characteristics were analyzed using chi-square or independent samples t-tests as shown in [fig_ref] Table 1: Demographic characteristics of participants included in analyses [/fig_ref].
Whole brain LGI comparisons across groups were performed with FreeSurfer. Because the LGI computation itself is relatively smooth [bib_ref] How to measure cortical folding from MR images: a step-by-step tutorial to..., Schaer [/bib_ref] , no additional smoothing was applied. General linear model (GLM) analyses were conducted using FreeSurfer. Separate GLM analyses were used for right and left hemispheres. Group differences between the PAE and control groups were tested, statistically controlling for two factors: study site and sex. In addition, age and total intracranial volume (TIV) were entered as covariates to control for potential confounding. Prior to analysis, the two covariates (age and TIV) were normalized by subtracting each value by the mean and dividing by the standard deviation (creating z-scores). The analyses controlled for TIV because PAE is known to be associated with belowaverage head circumference and brain volumes [bib_ref] Regional brain volume reductions relate to facial dysmorphology and neurocognitive function in..., Roussotte [/bib_ref]. Group (PAE vs. control) was the independent variable. These analyses were performed using a manually created design matrix.
FreeSurfer offers two types of analyses, both based on General Linear Models (GLM): a Different Offset -Different Slope (DODS) model and Different Offset -Same Slope (DOSS) model. DODS includes more regressors and is thus a less powerful, more conservative approach than DOSS, which is a more liberal model that assumes similar slopes for the covariates. We chose to use a hybrid model in between DODS and DOSS to maximize power and to best adhere to the data. Each linear model incorporates two parameters: an offset (or intercept) and a slope. In this case, the offset/intercept was the LGI at a covariate of 0 (z-scored age) and is measured in the same units as LGI (mm). The slope reflects the relationship between the two factors (in this case LGI and agemeasured in mm/year). The model was specified to allow for different offsets (i.e. y-intercept differences) amongst all factors; however the slope was constrained across study sites because there was no evidence of systematic differences in the relationship between age and LGI by site (nor would one assume there to be a difference). Results from each GLM were corrected for multiple comparisons with a two tailed Monte Carlo simulation implemented in FreeSurfer [bib_ref] Smoothing and cluster thresholding for cortical surface-based group analysis of fMRI data, Hagler [/bib_ref] using a cluster-wise forming threshold of p < 0.05 and 10,000 random permutations. Results were visualized by overlaying significant clusters on top of an inflated cortical surface in the visualization tool Freeview.
Additionally, two separate Pearson product-moment correlations were performed to examine the relationship between LGI and IQ. First, for each cluster derived from the prior analysis of the group difference (PAE vs. control) in LGI, mean LGI was computed per participant. Pearson product-moment correlations were performed to assess the relationship between mean LGI and IQ. In a separate analysis designed to examine the spatial pattern of the relationship between IQ and LGI, a whole brain vertex-by-vertex analysis of LGI by IQ Pearson productmoment correlation was performed using the Query Design Estimate Contrast (QDEC) interface tool from FreeSurfer. GLMs were run separately for right and left hemispheres. IQ was normalized by transforming to a z-score and entered as a covariate. In these analyses, no other covariates (age or TIV) or factors (gender, study site, diagnosis) were included in the GLM. For multiple comparison control, a two-tailed false discovery rate (FDR) correction was implemented [bib_ref] Thresholding of statistical maps in functional neuroimaging using the false discovery rate, Genovese [/bib_ref] using a corrected p-value of q = 0.05. Results were visualized in QDEC.
# Results
## Subject characteristics
As shown in [fig_ref] Table 1: Demographic characteristics of participants included in analyses [/fig_ref] , the PAE and control groups did not differ in sex, race, ethnicity, handedness, socioeconomic status (SES), distribution across study sites, or total intracranial volume. It is worth noting that the lack of difference in SES (some studies find lower SES in alcohol-exposed individuals) was likely due to the high proportion of PAE group having been adopted by families with higher SES.
By chance, the control group was older than the PAE group by approximately one year; as a result, subsequent analyses controlled for age in addition to other potential confounds. In addition, a small significant difference in Puberty Development Scale score [fig_ref] Table 1: Demographic characteristics of participants included in analyses [/fig_ref] was found. A set of exploratory analyses were conducted to determine if puberty status explained a significant amount of variance in LGIindependently of age. Ultimately, a set of partial correlations revealed that age, but not puberty status, explained a significant amount of variance in LGI. As a result, age was included in the statistical models but puberty status was not included.
Although there was no overall group difference in racial makeup, there were significantly more controls that identified as Asian compared to those with PAE. As expected, participants in the PAE group had significantly lower IQ compared to controls (14 points, or nearly one standard deviation lower) as well as significantly higher incidence of microcephaly. A one-way ANOVA revealed a modest IQ difference across study sites [F(3, 171) = 3.55, p < 0.05] which reflects slight differences in the populations. The mean IQs were: Atlanta = 89.9; Los Angeles = 94.9; San Diego = 95.7; Minnesota = 100.9.
In addition to alcohol, prenatal exposure data for other toxins/ substances were acquired. Exposure was reported as follows: amphetamines (1 control, 12 PAEs), cocaine (1 control, 21 PAEs), marijuana (1 control, 26 PAEs), tobacco (5 controls, 45 PAEs), caffeine (42 controls, 23 PAEs), hallucinogen (1 PAE), heroin (1 PAE), painkillers/opioids (2 PAEs), and tranquilizers (5 PAEs).
## Motion and data quality
All FreeSurfer automated segmentation and parcellation results were visually inspected for accuracy/artifact. Seven participants were excluded from the analysis due to aberrant FreeSurfer processing, mostly due to the effects of excessive motion during scanning. The excluded participants were as follows: 5 PAE and 2 Controls; 6 males and 1 female; 1 from Atlanta; 0 from Los Angeles; 3 from Minnesota; and 3 from San Diego. No significant difference was found between the included and excluded participants in terms of sex
## Cortical gyrification in pae vs. controls
Controlling for sex, site/scanner, age, and total TIV, the PAE group showed significantly lower LGI across large regions of cortex compared to the control group (cluster forming threshold was set to p < 0.05 and clusters were corrected for multiple comparisons with a Monte Carlo procedure as described earlier). Clusters are illustrated in . Five large clusters of significantly different LGI were evident [fig_ref] Table 3: Cluster summary [/fig_ref] : three in the left hemisphere [all p < 0.01; sizes 26,684 and 3720, and 731mm 2 ], and two in the right hemisphere [p < 0.001; sizes 22,912, and 8780mm 2 ]. Cluster #1 had a peak vertex located within the left postcentral gyrus with the cluster covering portions of the left supramarginal, cuneus, precuneus, paracentral lobule, posterior cingulate, isthmus cingulate, and bank of the superior temporal, middle temporal, inferior temporal, inferior parietal, superior parietal, lateral occipital, and superior frontal gyri. Cluster #2 had a peak vertex within the left rostral middle frontal gyrus with the clusters comprising portions of the left frontal pole, pars opercularis, pars orbitalis, pars triangularis, and medial orbitofrontal gyrus. Cluster #3 was a very small cluster with a peak vertex within the left precentral gyrus. Cluster #4 had a peak vertex within the right postcentral gyrus, and encompassed portions of the cuneus, pericalcarine, paracentral lobule, posterior cingulate, caudal anterior cingulate, isthmus cingulate, rostral anterior cingulate, supramarginal, and medial orbital frontal, superior frontal, lateral occipital, superior parietal, inferior parietal, postcentral, and precentral gyri. Finally, Cluster #5 had a peak vertex located within the right rostral middle frontal gyrus, and included portions of the pars opercularis, pars triangularis, insula, and caudal middle frontal, superior frontal, and, superior temporal gyri.
## Cluster-wise lgi by iq correlational analysis
After extracting average LGI for each cluster from the group comparison, correlational analyses were performed. Pearson productmoment correlations were computed to assess the relationships between average LGI in each of the five clusters and IQ score (see . In the left hemisphere, there was a positive correlation between LGI and IQ in cluster 1 (peak vertex within the left postcentral gyrus), [all participants: r = 0.35, n = 175, p < 0.00001; controls: r = 0.29, n = 83, p < 0.01; PAE: r = 0.27, n = 92, p < 0.01]; cluster 2 (peak vertex within the left rostral middle frontal gyrus), [all participants: r = 0.38, n = 175, p < 0.00001; controls: r = 0.29, n = 83, p < 0.01; PAE: r = 0.33, n = 92, p < 0.01]; and cluster 3 (peak vertex within the left precentral gyrus), [all participants: r = 0.29, n = 175, p < 0.0001; controls: r = 0.17, n = 83, p = 0.12; PAE: r = 0.29, n = 92, p < 0.01]. In the right hemisphere, there was also a positive correlation in cluster 4 (peak vertex within the right postcentral gyrus), [all participants: r = 0.30, n = 175, p < 0.0001; controls: r = 0.23, n = 83, p < 0.05; PAE: r = 0.25, n = 92, p < 0.05]; and cluster 5 (peak vertex within the right rostral middle frontal gyrus), [all participants: r = 0.40, n = 175, . Cortical gyrification group comparison between PAE and control groups. Inflated cortical convolution maps showing clusters after thresholding the uncorrected data and correcting for multiple comparisons, (cluster form threshold, p < 0.05; clusters for multiple comparisons, p < 0.05) of significant reduction in gyrification amongst participants with prenatal alcohol relative to healthy controls. Cluster numbers correspond to those found in [fig_ref] Table 3: Cluster summary [/fig_ref]. T.J. p < 0.00001; controls: r = 0.43, n = 83, p < 0.0001; PAE: r = 0.25, n = 92, p < 0.05]. For illustration, a representative scatter plot of cluster 5 is included in to provide a visualization of these data.
Overall, there was a strong positive correlation between LGI and IQ score bilaterally. Lower LGI (smoother cortex) was associated with lower IQ scores.
## Vertex-wise lgi by iq correlational analysis
An alternative approach to examining/visualizing associations with surface data such as LGI is to examine it at the level of individual vertices (points) rather than whole clusters. QDEC was used to compute Pearson product-moment correlations between IQ and LGI at each vertex. For these analyses, only IQ was included as a covariate and no factors were controlled because the intention was to show an overall LGI by IQ correlation vertex by vertex. A false discovery rate was chosen to limit false positives to a corrected p-value of q = 0.05. Several cortical areas show a strong positive correlation between LGI and IQparticularly the lateral aspects of the frontal, temporal, and parietal lobes bilaterally. Additionally, there were significant findings on the medial aspects bilaterally. Significant correlations were more widespread on the lateral aspects in both hemispheres. The findings on the medial aspects were more localized to the cuneus, precuneus, and the superior and inferior frontal regions. Findings are illustrated in [fig_ref] Figure 3: Vertex-wise IQ by LGI correlations [/fig_ref] and B. Additional information regarding the findings can be found in .
# Discussion
This large-scale examination of cortical development and cognition in children with PAE adds to the understanding of the origins of cognitive deficits in FASD and suggests that further investigation of complex cortical morphometry is warranted. We applied three-dimensional surface MRI techniques to measure cortical gyrification in children and adolescents with PAE compared to controls. Controlling for age, brain volume, sex, and study site, the PAE group showed significantly smoother cortices bilaterally compared to controls. Collectively, the significant clusters spanned 48.7% of the right hemisphere, and 47.6% of the left hemisphere. These large effects were more pronounced and widespread than previous findings reported in the Correlation summary.
A. Pearson product moment correlations following extraction of average LGI by participant within each significant cluster [fig_ref] Table 3: Cluster summary [/fig_ref] at the population (n = 175), PAE (n = 92) and Control (n = 83) levels. Note: R = right hemisphere, L = left hemisphere; Con = Control group, PAE = Prenatal Alcohol Exposure group, Pop. = population. p < 0.05 = ⁎ , p < 0.01 = ⁎⁎ , p < 0.001 = ⁎⁎⁎ , p < 0.0001 = ⁎⁎⁎⁎ , p < 0.00001 = ⁎⁎⁎⁎⁎ .
## Fig. 2. mean
LGI Cluster 5 and IQ correlation. Representative scatter plot of cluster 5 corresponding to that found in [fig_ref] Table 3: Cluster summary [/fig_ref]. Participants with PAE are indicated by a blue "x", and controls are indicated by a red "o". The solid green line represents the correlational fit to the entire dataset (i.e. both PAE and Control groups). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
T.J. literature [bib_ref] Atypical cortical gyrification in adolescents with histories of heavy prenatal alcohol exposure, Infante [/bib_ref] [bib_ref] Cortical morphology in children with alcohol-related neurodevelopmental disorder, Rajaprakash [/bib_ref]. A likely explanation for the size of the effects in the current data is the relatively large sample size which resulted in greater statistical power to detect effects in a set of vertex-wise analyses. Although the specific mechanisms responsible for cortical malformations in PAE are not yet known, it is worth noting that there is a known association with microcephaly (which also occurs in PAE). Previous studies have shown that individuals with microcephaly have significantly smoother cortices than those with average brain volumes [bib_ref] Simplified gyral pattern in severe developmental microcephalies? New insights from allometric modeling..., Germanaud [/bib_ref] [bib_ref] Larger is twistier: spectral analysis of gyrification (SPANGY) applied to adult brain..., Germanaud [/bib_ref] [bib_ref] Brain size and folding of the human cerebral cortex, Toro [/bib_ref] suggesting that early insults may sometimes contribute to both brain anomalies. In the current study, we did not observe significantly lower brain volume in PAE compared to controls, perhaps because we had a relatively small number of full FAS diagnoses in the sample. Nonetheless, the LGI analyses were all controlled for total intracranial volume, increasing confidence in the results showing that those with PAE have cortical smoothing that is more extreme than is accounted for by their brain sizes.
The complex cortical folding that was measured here and in other studies of children and adolescents is the result of processes that begin very early in gestation but continue into the second and third trimester in humans. Cell proliferation in the germinal ventricular zone occurs as early as week 4 [bib_ref] Development of the human cerebral cortex: boulder committee revisited, Bystron [/bib_ref] and continues into the second trimester for 8 to 16 weeks. Cells gradually migrate to the sub-plate, which will eventually become the cerebral cortex [bib_ref] Neuroimaging studies of normal brain development and their relevance for understanding childhood..., Marsh [/bib_ref]. Next, migration and synaptogenesis occur, with neurons extending axons and forming synapses. Connections within the cortex as well as between cortex and sub-cortical regions, brainstem, cerebellum, and spinal cord are established [bib_ref] Assessment of cortical maturation with prenatal MRI. Part I: normal cortical maturation, Fogliarini [/bib_ref]. The majority of cortical folding (the actual formation of sulci and gyri) occurs between gestational week 20 and 35 [bib_ref] Fetal MRI: normal gestational landmarks for cerebral biometry, gyration and myelination, Garel [/bib_ref] , and the brain transitions from lissencephalic ('smooth brain') to convoluted. Lastly, myelination occurs [bib_ref] Sequence of central nervous system myelination in human infancy. II. Patterns of..., Kinney [/bib_ref] [bib_ref] Myelination in the developing human brain: biochemical correlates, Kinney [/bib_ref]. Ultimately, the purpose of cortical folding is to maximize cortical surface area and ensure maximum information processing efficiency. Gestational timing is not only an important aspect of cortical gyrification development, but it is also especially relevant to understanding developmental insults like PAE.
Critical periods, during which there is increased vulnerability to insult, have been identified for a range of developmental processes during gestation. PAE in mice on the 7th day of gestation produces brain alterations akin to those seen in children with FASD as well as facial dysmorphia [bib_ref] Sequence of developmental alterations following acute ethanol exposure in mice: craniofacial features..., Sulik [/bib_ref]. On the 7th day of mouse gestation, cellular proliferation and migration are occurring in the forebrain, midbrain, and hindbrain [bib_ref] Critical periods of vulnerability for the developing nervous system: evidence from humans..., Rice [/bib_ref]. In a firsttrimester equivalent study, alcohol exposure between gestational days 20 and 32 in pigtail macaques also produces an FAS-like syndrome. As the literature has generally shown [bib_ref] Critical periods for prenatal alcohol exposure: evidence from animal and human studies, Coles [/bib_ref] , first trimester human alcohol exposure impacts a number of important developmental processes that can profoundly disrupt later development.
The first trimester is clearly one critical period for human cortical development because cellular proliferation and migration are occurring. Disruptions in cellular proliferation lead directly to abnormalities in migration [bib_ref] Critical periods of vulnerability for the developing nervous system: evidence from humans..., Rice [/bib_ref]. Early rodent models indeed demonstrated significant abnormalities in neuronal migration resulting from PAE [bib_ref] Migration of cortical neurons is altered by gestational exposure to ethanol, Miller [/bib_ref]. In its extreme form, disrupted neuronal migration is associated with severe cortical abnormalities including lissencephaly, or 'smooth brain' [bib_ref] Normal development of brain circuits, Tau [/bib_ref] a condition in which the absence of cortical gyrification results in severe intellectual impairment or even non-viability of the individual [bib_ref] Smooth, rough and upside-down neocortical development, Olson [/bib_ref]. As indicated previously, however, cortical development can also be disrupted in the second and third trimesters via alcohol's effects on neuronal migration and other processes involved in folding.
Beyond the most extreme case of absent cortical gyrification, there is evidence of associations between a wide range of cortical abnormalities and cognition. For example, previous research has shown that cortical folding affects later functional development [bib_ref] Primary cortical folding in the human newborn: an early marker of later..., Dubois [/bib_ref] in a manner that affects intelligence. Polymicrogyria (an abnormality in neuronal migration that results in an increase in the number of gyri) has been observed in a 16 year old female with PAE who met full FAS diagnostic criteria [bib_ref] Polymicrogyria in fetal alcohol syndrome, Reinhardt [/bib_ref]. and colleagues suggested that FAS should, in fact, be considered as part of the differential diagnosis process in cases where polymicrogyria has been identified. Cortical abnormalities such as polymicrogyria are associated with general neurologic disruption including seizures, hypotonia, and problems with speech and swallowing [bib_ref] Bilateral generalized polymicrogyria (BGP) a distinct syndrome of cortical malformation, Chang [/bib_ref]. An animal model of FAS showed polymicrogyria in the frontal lobe, pachygyria (reduced number of gyri) in the parietal lobe, and agyria (absence of gyri) in the occipital lobe of infant pigtail macaques exposed prenatally to alcohol [bib_ref] Measures of alcohol damage in utero in the pigtailed macaque (Macaca nemestrina), Clarren [/bib_ref]. Of course, abnormal cortical gyrification is not specific to FASD and has been shown to occur in several clinical populations including schizophrenia [bib_ref] Aberrant cortical gyrification in schizophrenia: a surface-based morphometry study, Palaniyappan [/bib_ref] , 22q11.2 deletion syndrome [bib_ref] Congenital heart disease affects local gyrification in 22q11.2 deletion syndrome, Schaer [/bib_ref] , autism spectrum disorder [bib_ref] Cortical gyrification in autistic and Asperger disorders: a preliminary magnetic resonance imaging..., Jou [/bib_ref] , Williams syndrome [bib_ref] Abnormal cortical complexity and thickness profiles mapped in Williams syndrome, Thompson [/bib_ref] , and mental retardation [bib_ref] Reduced cortical folding in mental retardation, Zhang [/bib_ref] amongst others.
In addition to finding significantly reduced cortical gyrification in children with PAE, we demonstrated a clear relationship between cortical gyrification development and global cognitive functioning (IQ). The association was in the anticipated direction, such that smoother cortex (reflecting developmental abnormality) was strongly associated with lower IQ. Alcohol-related disruptions to critical early neurodevelopmental processes including cell proliferation, neuronal migration, and cortical folding would be expected to have a generalized, long-term impact on the individual's functioning because they are "foundational" disturbances. Overall, we did not observe a regional T.J. pattern to the cortical abnormalities nor did we find that global cognitive functioning was associated with gyrification in a regional pattern. Clearly, the literature suggests that PAE is associated with a wide range of possible insults to brain development and it is likely that the clinical variation seen in FASD is a result of varying combinations of insults and timing of insults during early development. Cortical gyrification is but one of those insults and, as such, represents a relatively blunt indicator of neurodevelopmental insult. As mentioned previously, cortical folding/gyrification is an evolutionary adaptation to maximize the amount of cortex within the available space of the cranium [bib_ref] How the cortex gets its folds: an inside-out, connectivity-driven model for the..., Mota [/bib_ref] [bib_ref] Why does cerebral cortex fissure and fold -Springer, Welker [/bib_ref] and is thought to directly contribute to increased efficiency of information processing by optimizing the length and the layout of intra-cortical connections (Goldman- [bib_ref] Prenatal formation of cortical input and development of cytoarchitectonic compartments in the..., Goldman-Rakic [/bib_ref] [bib_ref] Topographical variation of the human primary cortices: implications for neuroimaging, brain mapping,..., Rademacher [/bib_ref] [bib_ref] Specification of cerebral cortical areas, Rakic [/bib_ref] [bib_ref] Mechanical model of brain convolutional development, Richman [/bib_ref] [bib_ref] Brain atlases -a new research tool, Roland [/bib_ref] [bib_ref] A tension-based theory of morphogenesis and compact wiring in the central nervous..., Van Essen [/bib_ref] [bib_ref] Area V5 of the human brain: evidence from a combined study using..., Watson [/bib_ref] ; for a review, see [bib_ref] Development of cortical folding during evolution and ontogeny, Zilles [/bib_ref]. Cortical gyrification is inherently linked to cognitive functioning, as suggested by non-human primate studies [bib_ref] The primate neocortex in comparative perspective using magnetic resonance imaging, Rilling [/bib_ref] [bib_ref] Gyrification in the cerebral cortex of primates, Zilles [/bib_ref] as well as human studies [bib_ref] Mapping the relationship between cortical convolution and intelligence: effects of gender, Luders [/bib_ref]. It is worth noting that Luders and colleagues found more modest associations between cortical gyrification and intelligence and the effects were more localized than we observed in PAE. This difference may be due simply to the wider range of gyrification and wider range of IQ in our sample (which contained both typically-developing children and those with PAE).
One limitation of the current study, faced by many studies of FASD, is the relative lack of detail about alcohol exposure in most cases. This prevented us from conducting analyses examining the relationship between exposure amounts/timing and cortical development. In this type of neuroimaging study, an added limitation arises from lost data for some participants. In some cases, detailed cortical measurements were not obtained because of processing difficultiesmost often because of motion artifact (which occurs in both PAE and control groups). We did evaluate motion as a potential confound, and we determined that there were not significant differences in motion between the PAE and control groups. Nonetheless, the lost data do temper the generalizability of the study slightly.
By chance, our groups ended up with a significant age differencewith the PAE group being younger than the controls by about one year. We evaluated the impact of age as well as puberty on our outcome measures and, ultimately, included age as a factor in all of our statistical models. Although statistical control is not a perfect substitute for precisely matched groups, our analyses suggests that these statistical controls allow us to interpret our results with confidence in this case. Another potential limitation of a multi-site study such as this one relates to systematic measurement differences across sites. We did observe a difference in LGI across sites, which could be due to hardware differences, minor differences in acquisition sequences, or population differences across the sites. We took a conservative approach and controlled for site in all of our statistical models. Overall, the data seem to suggest that semi-automated cortical characterization of this type is relatively robust in a clinical population such as FASD across multiple scanner sites. A further methodological consideration pertaining to the current study is the lack of a definitive significance threshold for analyses such as the vertex-wise analyses presented here. Our approach was a conservative one in which we applied a commonly-used false discovery rate correction in one case and a commonly-used Monte Carlo simulation technique in another. Because of the size of the effects seen here and the conservative nature of the corrections, we have confidence in the conclusions. Lastly, although our study included a relatively large sample size, breaking down the FASD group into subgroups (i.e. FAS, non-FAS, and deferred) was not possible because of the resulting small group sizes. Therefore, we were unable to determine if there is a relationship between FASD severity and cortical gyrification. CIFASD plans to continue to acquire new participants with FASD who will undergo neuroimaging, so it will be possible to address these questions in the future.
# Conclusion
The current study examined cortical gyrification in PAE, and explored relationships with cognitive functioning. The data showed robust effects of smoother cortices in PAE compared to controls. These data add to the existing literature showing cortical gyrification abnormalities in PAE and provide a clear example of how alcohol interferes with basic neurodevelopmental processes that ultimately have large downstream effects. One of these effects is illustrated in the strong associations seen between cortical gyrification and IQ. Ultimately, measures of cortical gyrification may be useful as indices of the overall level of neurodevelopmental abnormality in individuals with FASD.
## Funding and compliance with ethical standards
Funding This study was funded by the National Institute on Alcohol Abuse and Alcoholism (NIAAA). The following support was utilized in this work: NIAAA U01AA017122 (PI: ERS); NIAAA U01AA014834 (PI: SNM); U24AA014811 (PI: EPR); U24AA014815 (PI: KLJ); U24AA014818 (PI: Barnett); support from the Minnesota Supercomputing Institute.
[fig] Figure 3: Vertex-wise IQ by LGI correlations. (A-B): Inflated vertex-wise cortical convolution maps showing correlations between IQ and LGI amongst all participants in the left (A) and right hemisphere (B). An FDR of q < 0.05 was employed to correct for multiple comparisons. Color codes are in negative logarithms ("-log") and represent strength of significance with warm colors for positive associations and cool colors for negative associations. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.) [/fig]
[table] Table 1: Demographic characteristics of participants included in analyses. Socioeconomic Status; via the Hollingshead Score (if two caretakers take average). PDSA = Puberty Development Scale Average. FAS = Fetal Alcohol Syndrome. ARND = Alcohol-Related Neurodevelopmental Disorder. ARBD = Alcohol-Related Brain Damage. a Height or weight ≤ 10%ile. b Head circumference ≤ 10%ile. c At least two of the following: Palpebral fissure length ≤ 10%ile, thin vermillion border, smooth philtrum (4 or 5 on lipometer scale). [/table]
[table] Table 2: MRI sequence and parameters. [/table]
[table] Table 3: Cluster summary. Clusters showing differences between the PAE and Control groups controlling for study site, sex, age, and total intracranial volume (TIV) (cluster forming threshold, p < 0.05; clusters for multiple comparisons, p < 0.05). [/table]
|
Green Tea, Intermittent Sprinting Exercise, and Fat Oxidation
Fat oxidation has been shown to increase after short term green tea extract (GTE) ingestion and after one bout of intermittent sprinting exercise (ISE). Whether combining the two will result in greater fat oxidation after ISE is undetermined. The aim of the current study was to investigate the combined effect of short term GTE and a single session of ISE upon post-exercise fat oxidation. Fourteen women consumed three GTE or placebo capsules the day before and one capsule 90 min before a 20-min ISE cycling protocol followed by 1 h of resting recovery. Fat oxidation was calculated using indirect calorimetry. There was a significant increase in fat oxidation post-exercise compared to at rest in the placebo condition (p < 0.01). After GTE ingestion, however, at rest and post-exercise, fat oxidation was significantly greater (p < 0.05) than that after placebo. Plasma glycerol levels at rest and 15 min during post-exercise were significantly higher (p < 0.05) after GTE consumption compared to placebo. Compared to placebo, plasma catecholamines increased significantly after GTE consumption and 20 min after ISE (p < 0.05). Acute GTE ingestion significantly increased fat oxidation under resting and post-exercise conditions when compared to placebo.
# Introduction
The increasing prevalence of overweight and obesity is associated with numerous cardiovascular and metabolic diseases [bib_ref] Mechanisms and effects of green tea on cardiovascular health, Basu [/bib_ref] [bib_ref] Green tea consumption and mortality due to cardiovascular disease, cancer, and all..., Kuriyama [/bib_ref] [bib_ref] Effects of green tea and EGCG on cardiovascular and metabolic health, Wolfram [/bib_ref]. Weight reduction treatments designed to induce fat loss include steady state exercise, appetite suppressants, and lipase inhibitors [bib_ref] The effects of epigallocatechin-3-gallate on thermogenesis and fat oxidation in obese men:..., Boschmann [/bib_ref] [bib_ref] Recent findings of green tea extract AR25 (Exolise) and its activity for..., Chantre [/bib_ref]. Energy restricted diets have also been used in an attempt to lower the body fat of obese individuals [bib_ref] Medicare's search for effective obesity treatments: Diets are not the answer, Mann [/bib_ref]. The minimal long-term effectiveness of these methods has focused attention on other fat loss strategies such as green tea ingestion [bib_ref] The effects of green tea on weight loss and weight maintenance: A..., Hursel [/bib_ref] and participation in intermittent sprinting exercise (ISE) [bib_ref] High-intensity intermittent exercise and fat loss, Boutcher [/bib_ref].
Green tea extract (GTE) is produced from the leaves of camellia sinensis [bib_ref] Mechanisms and effects of green tea on cardiovascular health, Basu [/bib_ref] and contains catechins which are the predominant form of polyphenols. The major catechins are epigallocatechin gallate (EGCG), epigallocatechin (EGC), epicatechin gallate (ECG), and epicatechin (EC) [bib_ref] Simultaneous determination of twelve tea catechins by high-performance liquid chromatography with electrochemical..., Sano [/bib_ref]. EGCG is the most pharmacologically active of the catechins which typically accounts for approximately 50% of the catechin content of green tea [bib_ref] Simultaneous determination of twelve tea catechins by high-performance liquid chromatography with electrochemical..., Sano [/bib_ref]. Green tea typically contains a small amount of caffeine estimated to be about three to five percent of its dry weight [bib_ref] Effects of encapsulated green tea and Guarana extracts containing a mixture of..., Berube-Parent [/bib_ref]. Short term ingestion of tea catechins, typically one to two days before testing, has been shown to increase fat oxidation, particularly during the postprandial period, as indicated by a reduced respiratory exchange ratio (RER) during indirect calorimetry [bib_ref] The effects of epigallocatechin-3-gallate on thermogenesis and fat oxidation in obese men:..., Boschmann [/bib_ref] [bib_ref] Efficacy of a green tea extract rich in catechin polyphenols and caffeine..., Dulloo [/bib_ref] [bib_ref] Epigallocatechin gallate attenuates diet-induced obesity in mice by decreasing energy absorption and..., Klaus [/bib_ref] [bib_ref] Oolong tea increases energy metabolism in Japanese females, Komatsu [/bib_ref] [bib_ref] Oolong tea increases metabolic rate and fat oxidation in men, Rumpler [/bib_ref]. It has been suggested that the catechins in green tea increase fat oxidation through inhibition of catechol-O-methyltransferase, an enzyme that degrades norepinephrine, thereby prolonging the action of sympathetically released norepinephrine [bib_ref] Tea, obesity, and diabetes, Kao [/bib_ref] [bib_ref] Enzymology of methylation of tea catechins and inhibition of catechol-O-methyltransferase by (´)-epigallocatechin..., Lu [/bib_ref] [bib_ref] Anti-obesity effects of green tea: From bedside to bench, Wolfram [/bib_ref].
Steady state aerobic exercise results in small increases in plasma catecholamines [bib_ref] Plasma testerone and catecholamine response to physcial exercise of different intensities in..., Jezova [/bib_ref] [bib_ref] Gratas-Delamarche, A. Catecholamines and the effects of exercise, training and gender, Zouhal [/bib_ref]. In contrast, significantly elevated epinephrine and norepinephrine levels during 20 min of ISE in trained and untrained young women have been found [bib_ref] Metabolic response of trained and untrained women during high-intensity intermittent cycle exercise, Trapp [/bib_ref]. The catecholamine response is an important feature of ISE as catecholamines have been shown to drive lipolysis and are largely responsible for fat release from both subcutaneous and intramuscular fat stores [bib_ref] Role of beta-adrenergic receptors in mobilization of energy sources in exercising dogs, Issekutz [/bib_ref]. The significant catecholamine response to ISE may underlie the ability of regular ISE to induce greater fat loss than that occurring after regular steady state exercise. For example, Trapp et al. [bib_ref] The effects of high-intensity intermittent exercise training on fat loss and fasting..., Trapp [/bib_ref] compared steady state cycle exercise and ISE using a bout of ISE lasting 8 s with a 12 s recovery (20 min total) for three times per week for 15 weeks. Fat mass significantly decreased in the ISE condition compared to no reduction in subcutaneous fat after steady state exercise. A similar protocol with overweight males also resulted in significant decreases in total and visceral fat [bib_ref] The effect of high-intensity intermittent exercise on body composition of overweight young..., Heydari [/bib_ref]. Given the important role of the neurotransmitter norepinephrine in the control of fat oxidation, it is conceivable that GTE, by inhibiting norepinephrine breakdown, may enhance fat oxidation after ISE. Another possible contributor to ISE-induced exercise recovery fat oxidation, however, could be the post exercise oxygen uptake (EPOC) in excess of that required at rest [bib_ref] The relationship between aerobic fitness and recovery from high intensity intermittent exercise, Tomlin [/bib_ref].
Although the combined effect of short term GTE and acute ISE on fat oxidation during post-exercise has not been examined one study has investigated the ability of GTE to elevate fat oxidation during 30 min of steady state aerobic exercise. Venables et al. [bib_ref] Green tea extract ingestion, fat oxidation, and glucose tolerance in healthy humans, Venables [/bib_ref] found that short term GTE ingestion increased fat oxidation by 17% compared to placebo during a 30-min continuous bout of moderate intensity cycling exercise. In contrast, it was found that one day of GTE ingestion had no effect on markers of lipolysis or fat oxidation during moderate intensity, cycling exercise, however, seven days of GTE supplementation increased markers of lipolysis but had no effect on fat oxidation [bib_ref] No effect of 1 or 7 d of green tea extract ingestion..., Randell [/bib_ref]. As studies were similar in their design and participant characteristics, authors suggested that the addition of caffeine in the more recent study could have suppressed fat oxidation as there is a negative correlation between lactate level and fat oxidation [bib_ref] Relation between plasma lactate concentration and fat oxidation rates over a wide..., Achten [/bib_ref].
As these results were obtained during moderate intensity aerobic exercise, it appears that the effect of short term ingestion of GTE on fat oxidation, before and after ISE, has not been investigated. Therefore, the aim of this study was to examine the effect of short term GTE ingestion and one bout of ISE on fat oxidation of untrained women. It was hypothesized that the combination of GTE and ISE, compared to ISE alone, would result in significantly greater fat oxidation during the post-exercise period.
## Experimental section
## Participants
Fourteen untrained females (age: 21.5˘0.5 years; body mass: 65.7˘1.8 kg; BMI: 24.3˘0.4 kg/m 2 ; maximal oxygen consumption (V O 2max ): 32.1˘1.7 mL/kg/min) were recruited to participate in the study. All women were healthy judging from response to a general health survey and all gave written consent for participation. Exclusion criteria included women who were pregnant and regular caffeine (ě2 cups coffee/day) or green tea drinkers (ě2 cups tea/day). The study was approved by a university human research ethics committee.
## Preliminary testing
After an overnight fast of approximately 10 h participants attended the laboratory between 7:00 and 9:00 am where baseline anthropometric measurements, fasting blood sugar level and lipid profile assessment, a VO 2max test, maximum power output assessment, and induction to ISE training was carried out. Anthropometric measures were height, body mass, and BMI. A 23-gauge butterfly needle was inserted into an antecubital vein and blood samples were placed in 10 mL EDTA and 4 mL heparin sodium vacutainers (Becton Dickinson, Plymouth, UK). Blood lipid levels were assessed immediately (Cholestech LDX, Hayward, CA, USA). The laboratory was maintained at a constant ambient air temperature of 22˝C to 23˝C.
An electrically braked, computer-controlled Monark 839E ergometer (Monark, Vansbro, Sweden) linked to a ParvoMedics TrueMax 2400 metabolic cart (ParvoMedics, Sandy, UT, USA) was used to assess VO 2max . Participants were required to maintain a cycling cadence of approximately 70 revolutions per minute (RPM) that was displayed on a computer screen and monitored by an experimenter. Each participant completed a 3-min warm-up at 30 watts (W) after which load was increased by 15 W per min until exhaustion. Heart rate (HR) was recorded continuously using a Polar Watch S810i (Polar Electro, Kempele, Finland). VO 2max was considered to be maximal when at least two of the following conditions were achieved: (1) leveling of VO 2 even with an increase in workload;
(2) a RER ě 1.10; and (3) reduced pedaling speed despite encouragement. Range of time duration for the VO 2max test was 8 min 58 s to 12 min 15 s. Participants then practiced the ISE protocol for 10 min [bib_ref] Multiple-sprint work: Methodological, physiological, and experimental issues, Glaister [/bib_ref] on a Monark 839E ergometer which consisted of repeated bouts of 8-s sprint cycling at 60% of maximum power output and 12 s recovery at 20% of maximum power output. Exercise was accompanied by a tape which prompted the start and finish of sprint and recovery and the pedal cadence (110 RPM) during the sprint and recovery (40 RPM) phases.
## General study design
A double-blinded crossover design involving women completing two exercise sessions with either GTE or placebo was used. A wash-out period of four weeks separated sessions. To standardize testing both exercise sessions were performed during the luteal phase of consecutive menstrual cycles at the same time of day. Determination of luteal phase was achieved through questioning each participant before arranging each testing session. All women reported that they did not use oral contraceptives.
## Diet and capsule content
Participants recorded a food diary for 3 days prior to the first session and were requested to follow the same diet before the second session. The day before each exercise session, participants ingested one capsule containing either GTE or cellulose with breakfast, lunch, and dinner. Order of GTE and placebo was counter-balanced. After approximately 10 h of overnight fasting, the fourth capsule was consumed on the morning of the next day, 50 min before baseline, and 90 min before exercise [bib_ref] Green tea extract ingestion, fat oxidation, and glucose tolerance in healthy humans, Venables [/bib_ref]. The three GTE capsules consumed the day before exercise contained a total of 562.5 mg polyphenols and 375 mg, whereas the one GTE capsule consumed on the exercise day, 90 min before exercise, contained 187.5 mg polyphenols and 125 mg EGCG. Green tea catechins (EGCG, EGC, and EC) have been shown to peak in the blood between 1.3 h and 1.6 h [bib_ref] Pharmacokinetics of tea catechins after ingestion of green tea and (´)-epigallocatechin-3-gallate by..., Lee [/bib_ref]. The GTE capsule (GNC, Pittsburgh, PA, USA) contained 250 mg of camellia sinesis extract (187.5 mg polyphenols, 125 mg EGCG). Each GTE capsule contained 20 mg of caffeine. The placebo capsule contained 500 mg cellulose. All participants were reminded by text message to ingest GTE or placebo capsules the day before and on the day of exercise and all reported that they had ingested the capsules.
## Experimental protocol
The second and third exercise sessions consisted of three phases: (1) rest, (2) ISE, and (3) post-exercise. Participants reported to the laboratory between 7:00 and 9:00 am after the approximate 10-h fast, having ingested the fourth GTE capsule 50 min previously. A 22-gauge cannula (Becton Dickinson, Plymouth, UK) was inserted into each participant's forearm antecubital vein and a 3-way stopcock (Becton Dickinson, Plymouth, UK) was used for repeated blood sampling. The 3-way stopcock line was kept patent by flushing with 0.9% isotonic saline (Pfizer, New York, NY, USA). A resting blood sample was taken at least 30 min after insertion of the cannula. Blood samples, which were centrifuged for 10 min at 3000 RPM after rest, exercise, and post-exercise, were collected in a vacutaneur containing EDTA. Blood plasma was extracted and stored at´86˝C for catecholamine and glycerol analysis.
## Rest
During the first 10 min of rest each participant laid on a plinth and rested whilst no data were collected. Data were then collected for the next 40 min. The first 10 min of the 40-min data collection period were eliminated and the remaining 30 min of data were screened so that data affected by behavior such as coughing and posture change could be excluded. VO 2 , VCO 2 , energy expenditure (EE), and RER were measured for 30 min using a TrueOne 2400 Canopy system (ParvoMedics, Sandy, UT, USA). EE was calculated by using the Weir equation [bib_ref] New methods for calculating metabolic rate with special reference to protein metabolism, Weir [/bib_ref]. HR was recorded continuously using the Polar watch. At the end of the rest period blood lactate levels were assessed (Accutrend Lactate monitor, Roche, Germany). Support for the validity of resting metabolic rate calculated by the Parvomedics system has been previously provided [bib_ref] Assessing validity and reliability of resting metabolic rate in six gas analysis..., Cooper [/bib_ref].
## Interval sprinting exercise and post exercise
After a 5-min warm-up at 30 W, participants immediately completed 20 min of ISE on a Monark Ergomedic 839E ergometer linked to the metabolic cart. Pedal cadence was paced at 110 RPM during the sprint phase and 40 RPM during the recovery phase. Pedal resistance for the sprint phase was calculated as 60% of each participant's maximal power output (W max ) determined from prior testing and 20% for the recovery phase. Pedal cadence was monitored throughout exercise and recorded for both phases at minutes 5, 10, 15, and 20. In the 20-min ISE session, sixty 8-s/12-s bouts totaling 8 min of sprinting and 12 min of easy pedaling recovery were performed followed by a 5-min cool-down at 30 W. RPM for sprinting recorded every 5 min, whereas HR was recorded every beat and then averaged each 15 s. Gas sampling was averaged every 15 s throughout the whole 20 min of exercise. Blood was sampled at 7 min, 14 min, and 20 min during the exercise recovery pedaling period and blood samples from min 7 and 20 were assayed immediately for lactate. Lactate was also analysed during post-exercise at min 30 and 75. Glycerol levels were measured using blood collected before exercise; 7 min, 14 min, and 20 min into exercise; and 15 min, 30 min, 45 min, 60 min, and 75 min post-exercise. Epinephrine and norepinephrine levels were assessed using blood collected before exercise; after completing 20 min of ISE; and 20 min post-exercise. RER was averaged at 5-min stages. RER was analysed during the 30 min to 75 min period post-exercise because with changing blood lactate concentrations (e.g., immediately after exercise) the bicarbonate concentration is also changing which results in CO 2 production that influences RER without necessarily representing the true quotient [bib_ref] Effect of high-intensity interval exercise on lipid oxidation during postexercise recovery, Malatesta [/bib_ref]. During the 30 min to 75 min period post-exercise lactate levels were stable. Rating of perceived exertion (RPE) was recorded every 5 min during exercise [bib_ref] Psychological basis of perceived exertion, Borg [/bib_ref].
After exercise, participants rested for 15 min allowing time for gas and volume calibration before undertaking the 1-h post-exercise period under the ventilated canopy. The 1-h post-exercise period started at minute 15 after exercise and finished at minute 75 [fig_ref] Figure 1: Diagrammatic representation of the study design [/fig_ref]. HR was monitored continuously and blood was sampled every 15 min during minutes 15 to 75 during post-exercise and was assessed for lactate at 30 min and 75 min post-exercise [fig_ref] Figure 1: Diagrammatic representation of the study design [/fig_ref]. [fig_ref] Figure 1: Diagrammatic representation of the study design [/fig_ref] illustrates the time line of the study.
## Blood variables and oxidation rates
Glycerol was measured using the Free Glycerol Determination (FG0100) reagent assay kit (Sigma Aldrich) and Glycerol Standard (G7793). The degree of enzymatic turnover of the substrate was determined by dual wavelength absorbance measurement at 450 nM and 540 nM. The coefficient of variation (CV) for glycerol was 7.8%. Norepinephrine and epinephrine were measured using mass spectrometry with a 5973N Mass Selective Detector, coupled to a 6890N gas chromatograph, and an SGE Forte BPX5ˆ0.25 IDˆ0.25 micron column [bib_ref] Measurement of norepinephrine and 3,4-dihydroxyphenylglycol in urine and plasma for the diagnosis..., Duncan [/bib_ref]. Accuracy and precision were determined by analysis of spiked serum samples at low, medium, and high nM concentrations of norepinephrine and epinephrine in triplicate on 7 separate days. Serum norepinephrine recoveries at 20, 100, and 500 nM were above 95% and inter-day average CV was 4.28%; similarly, epinephrine recoveries at 2 nM, 10 nM, 50 nM were above 99% and inter-day average CV 5.88% (n = 21 for each concentration level). Average intra-day CV was 2.02% for norepinephrine and 2.26% for epinephrine. Average intra-day recoveries of 98.6% and 102% were obtained for norepinephrine and epinephrine respectively.
Fat and carbohydrate oxidation rates (g/min) were calculated using the following Equations [bib_ref] Table of nonprotein respiratory quotient: An update, Peronnet [/bib_ref] :
[formula] F atoxidation p1.695ˆVO 2 q´p1.701ˆVCO 2 q (1) Carbohydrateoxidation p4.585ˆVO 2 q´p3.226ˆVCO 2 q(2) [/formula]
When assessing fat oxidation rate only the values at rest and during minutes 30 to 75 during post-exercise were used because RER does not represent substrate utilization when blood and tissue lactate concentrations are changing [bib_ref] Effect of high-intensity interval exercise on lipid oxidation during postexercise recovery, Malatesta [/bib_ref]. Lactate levels were assessed using whole blood obtained from the indwelling cannula.
Nutrients 2015, epinephrine in triplicate on 7 separate days. Serum norepinephrine recoveries at 20, 100, and 500 nM were above 95% and inter-day average CV was 4.28%; similarly, epinephrine recoveries at 2 nM, 10 nM, 50 nM were above 99% and inter-day average CV 5.88% (n = 21 for each concentration level). Average intra-day CV was 2.02% for norepinephrine and 2.26% for epinephrine. Average intra-day recoveries of 98.6% and 102% were obtained for norepinephrine and epinephrine respectively. Fat and carbohydrate oxidation rates (g/min) were calculated using the following Equations [bib_ref] Table of nonprotein respiratory quotient: An update, Peronnet [/bib_ref] :
[formula] Fat oxidation (1.695 × V̇O2) − (1.701 × V̇CO2) Carbohydrate oxidation (4.585 × V̇O2) − (3.226 × V̇CO2) [/formula]
When assessing fat oxidation rate only the values at rest and during minutes 30 to 75 during post-exercise were used because RER does not represent substrate utilization when blood and tissue lactate concentrations are changing [bib_ref] Effect of high-intensity interval exercise on lipid oxidation during postexercise recovery, Malatesta [/bib_ref]. Lactate levels were assessed using whole blood obtained from the indwelling cannula. Lactate assessed at * 1 , * 2 , * 4 , * 6 , * 9 ; catecholamines at * 1 , * 4 , * 5 , and glycerol at * 1-* 9 .
# Statistical analysis
Data were analyzed using SPSS 20.0 (SPSS Inc., Chicago, IL, USA). A two-factor (time × condition) repeated measures ANOVA was used to compare differences across time and condition. The Mauchly sphericity test was used to test for homogeneity of covariance for within subject factors. The Greenhouse-Geisser test was used to correct for non-homogenous values. When repeated measures ANOVA interactions were significant, adjusted Bonferroni post hoc tests were also performed. Data are presented as mean ± SEM and significance was set at p < 0.05.
# Results
## Blood testing
Total cholesterol (4.26 ± 0.2 mmol/L), high density lipoprotein (1.45 ± 0.1 mmol/L), low density lipoprotein (2.76 ± 0.3 mmol/L), triglyceride (0.72 ± 0.1 mmol/L), and blood glucose levels (4.8 ± 0.4 mmol/L) were in the normal range for women of this age. Lactate assessed at * 1 , * 2 , * 4 , * 6 , * 9 ; catecholamines at * 1 , * 4 , * 5 , and glycerol at * 1´*9 .
# Statistical analysis
Data were analyzed using SPSS 20.0 (SPSS Inc., Chicago, IL, USA). A two-factor (timeˆcondition) repeated measures ANOVA was used to compare differences across time and condition. The Mauchly sphericity test was used to test for homogeneity of covariance for within subject factors. The Greenhouse-Geisser test was used to correct for non-homogenous values. When repeated measures ANOVA interactions were significant, adjusted Bonferroni post hoc tests were also performed. Data are presented as mean˘SEM and significance was set at p < 0.05.
# Results
## Blood testing
Total cholesterol (4.26˘0.2 mmol/L), high density lipoprotein (1.45˘0.1 mmol/L), low density lipoprotein (2.76˘0.3 mmol/L), triglyceride (0.72˘0.1 mmol/L), and blood glucose levels (4.8˘0.4 mmol/L) were in the normal range for women of this age.
## Workload and exercise intensity
There were no significant differences in mean power output, RPE, lactate levels, RPM , and HR levels between the GTE and placebo trials [fig_ref] Table 2: Response at rest and during and after intermittent sprinting exercise in the... [/fig_ref]. . Mean power output, rating of perceived exertion, and lactate response to the sprinting and recovery components of the intermittent sprinting exercise for the placebo and green tea conditions (mean and SEM).
## Lactate
Blood lactate levels were similar at rest during the GTE (1.76˘0.17 mmol/L) and placebo (1.67˘0.13 mmol/L) conditions. Lactate levels were significantly increased, p < 0.05, during exercise in both conditions . During post-exercise (30 min and 75 min) lactate levels were similar for both the GTE (2.5˘0.24 mmol/L; 2.1˘0.15 mmol/L) and placebo (2.5˘0.25 mmol/L; 2.0˘0.20 mmol/L) conditions.
## Respiratory exchange ratio (rer)
During the resting period, before exercise, RER was significantly decreased, p < 0.05, after GTE ingestion compared to placebo [fig_ref] Table 2: Response at rest and during and after intermittent sprinting exercise in the... [/fig_ref]. There was also a significant condition main effect (p < 0.01) with RER being significantly decreased throughout the 60 min post-exercise period in the GTE compared to the placebo condition [fig_ref] Table 2: Response at rest and during and after intermittent sprinting exercise in the... [/fig_ref].
## Fat oxidation
During the resting period, before exercise, fat oxidation significantly increased by 24% (p < 0.01) after GTE ingestion (0.059˘0.004 g/min) compared to placebo (0.045˘0.005 g/min; . In the 30 min to 75 min period during post-exercise there was a significant condition main effect (p < 0.01). Fat oxidation rate was significantly increased by 29% (0.069˘0.006 g/min) from 35 min to 75 min during post-exercise in the GTE condition compared to the fat oxidation response during placebo (0.049˘0.004 g/min; .
Nutrients 2015, 7 8
## Respiratory exchange ratio (rer)
During the resting period, before exercise, RER was significantly decreased, p < 0.05, after GTE ingestion compared to placebo [fig_ref] Table 2: Response at rest and during and after intermittent sprinting exercise in the... [/fig_ref]. There was also a significant condition main effect (p < 0.01) with RER being significantly decreased throughout the 60 min post-exercise period in the GTE compared to the placebo condition [fig_ref] Table 2: Response at rest and during and after intermittent sprinting exercise in the... [/fig_ref].
## Fat oxidation
During the resting period, before exercise, fat oxidation significantly increased by 24% (p < 0.01) after GTE ingestion (0.059 ± 0.004 g/min) compared to placebo (0.045 ± 0.005 g/min; . In the 30 min to 75 min period during post-exercise there was a significant condition main effect (p < 0.01). Fat oxidation rate was significantly increased by 29% (0.069 ± 0.006 g/min) from 35 min to 75 min during post-exercise in the GTE condition compared to the fat oxidation response during placebo (0.049 ± 0.004 g/min; .
## Figure 2.
Fat oxidation at rest and after intermittent sprinting exercise with either placebo or green tea extract (GTE) ingestion. * Significantly greater compared to placebo, p < 0.05.
## ̇o 2, ̇o 2, and energy expenditure
Oxygen consumption was significantly higher (p < 0.01) during post-exercise in the GTE condition, whereas VĊ O2 levels were similar [fig_ref] Table 2: Response at rest and during and after intermittent sprinting exercise in the... [/fig_ref]. Energy expenditure (EE) was similar throughout testing in the GTE and placebo conditions [fig_ref] Table 2: Response at rest and during and after intermittent sprinting exercise in the... [/fig_ref].
## Glycerol and catecholamine levels
At rest and at minute 15 during post-exercise GTE ingestion brought about significantly higher (p < 0.05) plasma glycerol concentrations compared to placebo [fig_ref] Figure 3: Glycerol levels at rest, during, and after intermittent sprinting exercise with either... [/fig_ref]. . Fat oxidation at rest and after intermittent sprinting exercise with either placebo or green tea extract (GTE) ingestion. * Significantly greater compared to placebo, p < 0.05.
## V o 2 , v co 2 , and energy expenditure
Oxygen consumption was significantly higher (p < 0.01) during post-exercise in the GTE condition, whereas VCO 2 levels were similar [fig_ref] Table 2: Response at rest and during and after intermittent sprinting exercise in the... [/fig_ref]. Energy expenditure (EE) was similar throughout testing in the GTE and placebo conditions [fig_ref] Table 2: Response at rest and during and after intermittent sprinting exercise in the... [/fig_ref].
## Glycerol and catecholamine levels
At rest and at minute 15 during post-exercise GTE ingestion brought about significantly higher (p < 0.05) plasma glycerol concentrations compared to placebo [fig_ref] Figure 3: Glycerol levels at rest, during, and after intermittent sprinting exercise with either... [/fig_ref]. Nutrients 2015, 7 9 [fig_ref] Figure 3: Glycerol levels at rest, during, and after intermittent sprinting exercise with either... [/fig_ref]. Glycerol levels at rest, during, and after intermittent sprinting exercise with either placebo or green tea extract (GTE) ingestion. * Significantly greater compared to placebo, p < 0.05.
Epinephrine plasma levels were significantly increased (p < 0.05) during exercise after GTE compared to the placebo condition [fig_ref] Figure 4: Epinephrine and norepinephrine levels at rest, during, and after intermittent sprinting exercise... [/fig_ref]. GTE ingestion also brought about significantly higher (p < 0.05) plasma norepinephrine concentrations compared to placebo at 15 min during post-exercise [fig_ref] Figure 4: Epinephrine and norepinephrine levels at rest, during, and after intermittent sprinting exercise... [/fig_ref].
# Discussion
The combined effect of one bout of intermittent sprinting exercise (ISE) and short term ingestion of green tea extract (GTE) on fat oxidation of untrained women was examined. During pre-exercise rest GTE ingestion significantly increased fat oxidation. Fat oxidation levels were significantly higher throughout minutes 30 to 75 during post-exercise. Also plasma glycerol levels at rest and after ISE were significantly higher after GTE consumption compared to the placebo condition. Plasma epinephrine levels showed a significant increase during ISE compared to placebo, whereas norepinephrine levels were significantly higher 15 min during post-exercise after GTE ingestion.
The ability of short term GTE ingestion to enhance fat oxidation at rest has been previously demonstrated [bib_ref] The effects of epigallocatechin-3-gallate on thermogenesis and fat oxidation in obese men:..., Boschmann [/bib_ref] [bib_ref] Efficacy of a green tea extract rich in catechin polyphenols and caffeine..., Dulloo [/bib_ref] [bib_ref] Epigallocatechin gallate attenuates diet-induced obesity in mice by decreasing energy absorption and..., Klaus [/bib_ref] [bib_ref] Oolong tea increases energy metabolism in Japanese females, Komatsu [/bib_ref] [bib_ref] Oolong tea increases metabolic rate and fat oxidation in men, Rumpler [/bib_ref]. Dulloo et al. [bib_ref] Efficacy of a green tea extract rich in catechin polyphenols and caffeine..., Dulloo [/bib_ref] showed that GTE compared to placebo increased resting fat oxidation by 9.9%, whereas Rumpler et al. [bib_ref] Oolong tea increases metabolic rate and fat oxidation in men, Rumpler [/bib_ref] found that ingestion of oolong tea increased fat oxidation by 12%. Some studies have found an elevation in resting energy expenditure (EE) after short term GTE ingestion [bib_ref] Effects of encapsulated green tea and Guarana extracts containing a mixture of..., Berube-Parent [/bib_ref] [bib_ref] Oolong tea increases metabolic rate and fat oxidation in men, Rumpler [/bib_ref]. Women in the present study demonstrated a 24% increase in fat oxidation during rest after short term GTE consumption , however, we did not find the elevation in resting EE after GTE consumption that has been demonstrated by others [bib_ref] Effects of encapsulated green tea and Guarana extracts containing a mixture of..., Berube-Parent [/bib_ref] [bib_ref] Oolong tea increases metabolic rate and fat oxidation in men, Rumpler [/bib_ref]. Lack of elevation in resting EE was probably because GTE consumption appears to exert more of an effect on postprandial EE. For example, Dulloo et al. [bib_ref] Efficacy of a green tea extract rich in catechin polyphenols and caffeine..., Dulloo [/bib_ref] found increases in EE during a 24-h period during which three meals were consumed but no EE increase was found during sleep. The majority of studies examining GTE have recruited male participants [bib_ref] Effects of encapsulated green tea and Guarana extracts containing a mixture of..., Berube-Parent [/bib_ref] [bib_ref] Efficacy of a green tea extract rich in catechin polyphenols and caffeine..., Dulloo [/bib_ref] [bib_ref] Oolong tea increases metabolic rate and fat oxidation in men, Rumpler [/bib_ref] [bib_ref] Green tea extract ingestion, fat oxidation, and glucose tolerance in healthy humans, Venables [/bib_ref] [bib_ref] Effect of moderate intakes of different tea catechins and caffeine on acute..., Gregersen [/bib_ref] thus the present results demonstrate that GTE Epinephrine plasma levels were significantly increased (p < 0.05) during exercise after GTE compared to the placebo condition [fig_ref] Figure 4: Epinephrine and norepinephrine levels at rest, during, and after intermittent sprinting exercise... [/fig_ref]. GTE ingestion also brought about significantly higher (p < 0.05) plasma norepinephrine concentrations compared to placebo at 15 min during post-exercise [fig_ref] Figure 4: Epinephrine and norepinephrine levels at rest, during, and after intermittent sprinting exercise... [/fig_ref].
# Discussion
The combined effect of one bout of intermittent sprinting exercise (ISE) and short term ingestion of green tea extract (GTE) on fat oxidation of untrained women was examined. During pre-exercise rest GTE ingestion significantly increased fat oxidation. Fat oxidation levels were significantly higher throughout minutes 30 to 75 during post-exercise. Also plasma glycerol levels at rest and after ISE were significantly higher after GTE consumption compared to the placebo condition. Plasma epinephrine levels showed a significant increase during ISE compared to placebo, whereas norepinephrine levels were significantly higher 15 min during post-exercise after GTE ingestion.
The ability of short term GTE ingestion to enhance fat oxidation at rest has been previously demonstrated [bib_ref] The effects of epigallocatechin-3-gallate on thermogenesis and fat oxidation in obese men:..., Boschmann [/bib_ref] [bib_ref] Efficacy of a green tea extract rich in catechin polyphenols and caffeine..., Dulloo [/bib_ref] [bib_ref] Epigallocatechin gallate attenuates diet-induced obesity in mice by decreasing energy absorption and..., Klaus [/bib_ref] [bib_ref] Oolong tea increases energy metabolism in Japanese females, Komatsu [/bib_ref] [bib_ref] Oolong tea increases metabolic rate and fat oxidation in men, Rumpler [/bib_ref]. Dulloo et al. [bib_ref] Efficacy of a green tea extract rich in catechin polyphenols and caffeine..., Dulloo [/bib_ref] showed that GTE compared to placebo increased resting fat oxidation by 9.9%, whereas Rumpler et al. [bib_ref] Oolong tea increases metabolic rate and fat oxidation in men, Rumpler [/bib_ref] found that ingestion of oolong tea increased fat oxidation by 12%. Some studies have found an elevation in resting energy expenditure (EE) after short term GTE ingestion [bib_ref] Effects of encapsulated green tea and Guarana extracts containing a mixture of..., Berube-Parent [/bib_ref] [bib_ref] Oolong tea increases metabolic rate and fat oxidation in men, Rumpler [/bib_ref]. Women in the present study demonstrated a 24% increase in fat oxidation during rest after short term GTE consumption , however, we did not find the elevation in resting EE after GTE consumption that has been demonstrated by others [bib_ref] Effects of encapsulated green tea and Guarana extracts containing a mixture of..., Berube-Parent [/bib_ref] [bib_ref] Oolong tea increases metabolic rate and fat oxidation in men, Rumpler [/bib_ref]. Lack of elevation in resting EE was probably because GTE consumption appears to exert more of an effect on postprandial EE. For example, Dulloo et al. [bib_ref] Efficacy of a green tea extract rich in catechin polyphenols and caffeine..., Dulloo [/bib_ref] found increases in EE during a 24-h period during which three meals were consumed but no EE increase was found during sleep. The majority of studies examining GTE have recruited male participants [bib_ref] Effects of encapsulated green tea and Guarana extracts containing a mixture of..., Berube-Parent [/bib_ref] [bib_ref] Efficacy of a green tea extract rich in catechin polyphenols and caffeine..., Dulloo [/bib_ref] [bib_ref] Oolong tea increases metabolic rate and fat oxidation in men, Rumpler [/bib_ref] [bib_ref] Green tea extract ingestion, fat oxidation, and glucose tolerance in healthy humans, Venables [/bib_ref] [bib_ref] Effect of moderate intakes of different tea catechins and caffeine on acute..., Gregersen [/bib_ref] thus the present results demonstrate that GTE consumption also significantly elevates resting fat oxidation of females. With that said the blood levels of the different catechins contained in the GTE were not assessed, thus, a limitation of the study is that it cannot be shown that GTE directly influenced fat oxidation. Also another study limitation concerns the assessment of active nutrient content contained in the green tea capsules which was not independently analyzed. Also reminding participants to consume their capsules by text message and not using a more objective method to verify capsule ingestion was another limitation. Finally, a four-week wash-out period separated sessions to standardize testing during the luteal phase of consecutive menstrual cycles at the same time of day. This design controlled for menstrual cycle influences on fat oxidation but could have allowed for physical fitness changes.
Nutrients 2015, it cannot be shown that GTE directly influenced fat oxidation. Also another study limitation concerns the assessment of active nutrient content contained in the green tea capsules which was not independently analyzed. Also reminding participants to consume their capsules by text message and not using a more objective method to verify capsule ingestion was another limitation. Finally, a four-week wash-out period separated sessions to standardize testing during the luteal phase of consecutive menstrual cycles at the same time of day. This design controlled for menstrual cycle influences on fat oxidation but could have allowed for physical fitness changes. [fig_ref] Figure 4: Epinephrine and norepinephrine levels at rest, during, and after intermittent sprinting exercise... [/fig_ref]. Epinephrine and norepinephrine levels at rest, during, and after intermittent sprinting exercise with either placebo or green tea extract (GTE) ingestion. * Significantly greater compared to placebo, p < 0.05.
The higher plasma glycerol levels at rest and during post-exercise indicate that GTE compared to placebo results in enhanced markers of lipolysis. These results support prior research that manipulated blood free fatty acid levels by administering either nicotinic acid or heparin during a hard 90-min bout of moderate and high-intensity exercise [bib_ref] Skeletal muscle fat metabolism after exercise in humans: Influence of fat availability, Kimber [/bib_ref]. Results showed that during 6 h of exercise recovery plasma-derived free fatty acids was the major fuel source driving enhanced fat oxidation. Although plasma free fatty acid availability changed significantly, no marked change in intramuscular triglyceride (IMTG) concentration was detected. Therefore, after a bout of moderate and high-intensity exercise that resulted in blood lactate levels in excess of 5 mmol/L free fatty acids were found to drive fat oxidation. For shorter, higher intensity exercise such as intermittent sprinting, however, IMTG stores are thought to make more of an influence on whole body fat oxidation as subcutaneous adipose fat stores do not contribute significantly to high-intensity sprinting exercise [bib_ref] Gratas-Delamarche, A. Catecholamines and the effects of exercise, training and gender, Zouhal [/bib_ref]. The sprinting component (fast pedaling) of ISE appears to be mainly fuelled by creatinine-phosphate and anaerobic glycolysis [bib_ref] Multiple sprint work: Physiological responses, mechanisms of fatigue and the influence of..., Glaister [/bib_ref]. Oxygen bound to myoglobin also appears to make a small contribution to energy production during sprinting [bib_ref] Multiple sprint work: Physiological responses, mechanisms of fatigue and the influence of..., Glaister [/bib_ref] , however, the major role of aerobic metabolism appears to be the resynthesis of creatinine-phosphate during the recovery sprint periods [bib_ref] Multiple sprint work: Physiological responses, mechanisms of fatigue and the influence of..., Glaister [/bib_ref]. During the slow pedaling recovery component enhanced lipid utilization is also believed to occur [bib_ref] Effect of exercise intensity, duration and mode on post-exercise oxygen consumption, Borsheim [/bib_ref] [bib_ref] Effect of work and recovery duration on skeletal muscle oxygenation and fuel..., Christmass [/bib_ref]. The brief recovery periods of ISE are thought to prevent complete glycogen resynthesis, therefore, the glycogen depletion accompanying continuous sprinting is thought to impede glycolysis resulting in increased oxidation of IMTG [bib_ref] Effect of work and recovery duration on skeletal muscle oxygenation and fuel..., Christmass [/bib_ref]. Therefore, enhanced lipid oxidation could occur with a higher participation of lipolysis to the aerobic component during exercise recovery. Although an increase in fat oxidation was observed during exercise after short term GTE ingestion compared to placebo these values are not valid indicators of mitochondrial O 2 consumption and CO 2 production because VCO 2 assessed by indirect calorimetry is influenced by bicarbonate pool depletion [bib_ref] Effect of high-intensity interval exercise on lipid oxidation during postexercise recovery, Malatesta [/bib_ref]. Consequently, using the O 2 and CO 2 response during sprinting exercise to reflect nutrient usage typically overestimates carbohydrate and underestimates fat oxidation [bib_ref] Effect of high-intensity interval exercise on lipid oxidation during postexercise recovery, Malatesta [/bib_ref]. Blood lactate concentrations were not changing during minutes 30 to 75 during exercise recovery, thus fat oxidation rate was assessed during this period. That VCO 2 levels were similar and consistent during this period during exercise recovery for both GTE and placebo conditions also support the notion that the bicarbonate pool was stable [bib_ref] Effect of high-intensity interval exercise on lipid oxidation during postexercise recovery, Malatesta [/bib_ref].
Venables et al. [bib_ref] Green tea extract ingestion, fat oxidation, and glucose tolerance in healthy humans, Venables [/bib_ref] investigated the acute combined effects of GTE and steady state exercise (30 min cycling at 60% VO 2max ) using O 2 and CO 2 response to establish fat oxidation levels. GTE significantly elevated fat oxidation during steady state exercise by 17% relative to placebo. Results of the Venables et al. [bib_ref] Green tea extract ingestion, fat oxidation, and glucose tolerance in healthy humans, Venables [/bib_ref] study show that GTE ingestion can further enhance markers of lipolysis during steady state exercise even though exercise alone also brought about an increase in lipolysis. The effect of endurance training and longer term GTE supplementation on fat oxidation during exercise has also been examined in a 10-week intervention with untrained males [bib_ref] Effect of endurance training supplemented with green tea extract on substrate metabolism..., Ichinose [/bib_ref]. Results indicated that regular GTE ingestion, together with moderate intensity aerobic exercise, increased the proportion of whole body fat utilization during exercise although body mass was not significantly reduced. Also it has been shown that short-term consumption of EGCG compared to placebo increased maximal oxygen uptake in both males and females [bib_ref] Epigallocatechin-3-gallate increases maximal oxygen uptake in adult humans, Richards [/bib_ref]. The effect of short and longer GTE ingestion on cycling endurance appears to be negligible, however, as both 6 day [bib_ref] The effects of EGCG on fat oxidation and endurance performance in male..., Dean [/bib_ref] and 3 week ingestion [bib_ref] No effects of three-week consumption of a green tea extract on time..., Eichenberger [/bib_ref] of GTE did not result in an increase in time trial performance.
Following steady state exercise a significant increase in fat oxidation rate compared to the pre-exercise fasting state has been found [bib_ref] Lipid oxidation in fit young adults during postexercise recovery, Kuo [/bib_ref]. During intense sprinting exercise glycogen stores suffer greater depletion than steady state exercise [bib_ref] Mechanical external work and recovery at preferred walking speed in obese subjects, Malatesta [/bib_ref]. Thus, the post-exercise period after ISE should demonstrate enhanced lipid oxidation so that remaining carbohydrates can be utilized for glycogen resynthesis [bib_ref] Effect of high-intensity interval exercise on lipid oxidation during postexercise recovery, Malatesta [/bib_ref]. Results of the present study support this notion as in the placebo condition greater fat oxidation occurred during the last 30 min of post-exercise. McGarvey et al. [bib_ref] Excess post-exercise oxygen consumption following continuous and interval cycling exercise, Mcgarvey [/bib_ref] have also shown that when total work was similar, intermittent compared to steady state exercise, resulted in significantly greater fat oxidation during a 2-h post-exercise period. Muscle glycogen response to ISE, however, was not assessed which is also a limitation of the current study. Results of the present study also show that when ISE is accompanied by short term GTE ingestion then fat oxidation is increased throughout post-exercise. During post-exercise the average increase in fat oxidation was 29% after GTE ingestion compared to placebo. The monitoring of metabolic response over a 1-h post-exercise period, however, is a limitation of this study as it is feasible that post-exercise oxygen consumption could continue to occur many hours after exercise [bib_ref] The relationship between aerobic fitness and recovery from high intensity intermittent exercise, Tomlin [/bib_ref]. Thus, studies monitoring metabolic and hormonal response to ISE for an extended period are needed.
The effect of gender on the oxidation of triglycerides during ISE recovery does not appear to have been examined, however, a number of studies have assessed fat oxidation response of males and females during aerobic exercise. Unfortunately, gender affects are unclear because of conflicting results. The majority of studies that assessed markers of lipolytic rate during moderate-intensity endurance exercise using microdialysis probes or isotope tracers, found that lipolytic rate markers in females were larger than in males [bib_ref] Adrenergic regulation of lipolysis in situ at rest and during exercise, Arner [/bib_ref] [bib_ref] Substrate utilization during endurance exercise in men and women after endurance training, Carter [/bib_ref]. Others, however, found that the lipolytic response to exercise was similar for both genders [bib_ref] Leg free fatty acid kinetics during exercise in men and women, Burguera [/bib_ref]. Also studies that used indirect calorimetry to assess substrate oxidation have demonstrated that women oxidize more fat and less carbohydrate than men during aerobic exercise [bib_ref] Plasma FFA responses to prolonged walking in untrained men and women, Blatchford [/bib_ref]. The reasons for these equivocal results is unclear but may involve differences in body composition, aerobic fitness levels, and exercise modality as these factors have been found to influence the rate of lipolysis and fat oxidation during endurance exercise [bib_ref] Oxidation of nonplasma fatty acids during exercise is increased in women with..., Horowitz [/bib_ref] [bib_ref] Lipid metabolism during endurance exercise, Horowitz [/bib_ref]. Interestingly, catecholamine levels increase more in males than females during high-intensity exercise [bib_ref] Fuel metabolism in men and women during and after long-duration exercise, Horton [/bib_ref] which is likely a result of the larger male muscle mass which generates more power and enables men to typically work at a higher intensity than females. Consequently, future research is required to examine the effect of short term green tea ingestion and acute ISE on fat oxidation response of males.
The green tea capsules ingested by participants contained 20 mg of caffeine which was reported by the manufacturer. That the amount of caffeine in the capsules was not verified is a limitation of the study. The effect of caffeine on resting energy expenditure has been examined and studies have shown that a single oral dose of more than 100 mg caffeine is needed to elicit a significant increase in thermogenic response. Also to increase energy expenditure a 600 mg to 1000 mg caffeine dose per day appears to be necessary [bib_ref] Effects of caffeine on energy metabolism, heart rate, and methylxanthine metabolism in..., Bracco [/bib_ref] [bib_ref] Normal caffeine consumption: Influence on thermogenesis and daily energy expenditure in lean..., Dulloo [/bib_ref]. Thus, it is feasible that the small amount of caffeine ingested by participants in the current study is unlikely to have had a significant effect on their resting and post-exercise fat oxidation levels.
It was predicted that greater fat oxidation would occur after short term ingestion of GTE during the post-exercise period. The increase in post-exercise fat oxidation being driven by enhanced norepinephrine and epinephrine release during ISE causing an increased accumulation of circulatory sulfo-conjugated catecholamines [bib_ref] Sulfonation and molecular action, Strott [/bib_ref]. In contrast to the short half-life of catecholamines (1-3 min) the half-life of sulfo-conjugated catecholamines has been estimated to be 3-4 h [bib_ref] Metabolic response to green tea extract during rest and moderate-intensity exercise, Hodgson [/bib_ref]. It is thought that green tea catechins increase fat oxidation through inhibition of catechol-O-methyltransferase, the enzyme that degrades norepinephrine, thereby prolonging adrenergic drive [bib_ref] Tea, obesity, and diabetes, Kao [/bib_ref] [bib_ref] Enzymology of methylation of tea catechins and inhibition of catechol-O-methyltransferase by (´)-epigallocatechin..., Lu [/bib_ref] [bib_ref] Anti-obesity effects of green tea: From bedside to bench, Wolfram [/bib_ref]. As blood catechin, caffeine, and flavonoid levels were not assessed, however, it is unclear to what extent these nutrients found in green tea affected fat oxidation. The increased norepinephrine and glycerol levels during post-exercise after the GTE ingestion supports the notion that catechins may have made some contribution towards the increase in fat oxidation. Future research should attempt to identify the contribution of catechins, caffeine, and flavonoids to the green tea fat oxidation affect. That epinephrine levels were also elevated after GTE consumption during exercise also suggests that GTE may increase adrenergic drive. However, no relationship was found between adrenergic drive and blood catecholamine level during a 60-min bout of acute aerobic exercise performed at 56% of maximal oxygen uptake after 7 days of GTE ingestion [bib_ref] Metabolic response to green tea extract during rest and moderate-intensity exercise, Hodgson [/bib_ref]. In a recent review Hodgson et al. [bib_ref] The effect of green tea on fat oxidation at rest and during..., Hodgson [/bib_ref] have pointed out that the inhibition of catechol-O-methyltransferase hypothesis has little in vivo support and they suggest that changes in the expression of fat metabolism genes could be brought about with chronic exercise training. The expression of fat metabolism genes could include the upregulation of skeletal muscle fat metabolism enzyme gene and down regulation of hepatic adipogenic gene expression [bib_ref] The effect of green tea on fat oxidation at rest and during..., Hodgson [/bib_ref]. As it is unlikely that the combination of short term ingestion of GTE and one bout of ISE could bring about transcriptional activity changes more studies have to be carried out in order to identify the mechanisms underlying the short term GTE ingestion and acute ISE fat oxidation effect.
Fifteen weeks of ISE resulted in a significantly greater reduction of subcutaneous fat compared to 15 weeks of steady state cycle exercise [bib_ref] The effects of high-intensity intermittent exercise training on fat loss and fasting..., Trapp [/bib_ref]. A similar protocol with overweight males also resulted in significant decreases in total and visceral fat [bib_ref] The effect of high-intensity intermittent exercise on body composition of overweight young..., Heydari [/bib_ref] and in overweight females a significant decrease in total body fat and central abdominal fat was found [bib_ref] The effect of a lifestyle intervention on metabolic health in young women, Dunn [/bib_ref]. Consequently, the significant elevation in fat oxidation caused by one bout of ISE together with GTE ingestion suggests that repeated use of this combination may have the potential to reduce fat mass of overweight females.
# Conclusions
In conclusion, it was found that short term green tea ingestion significantly elevated fat oxidation and plasma glycerol levels before and post intermittent sprinting exercise. Epinephrine levels were elevated during exercise and norepinephrine levels were increased post-exercise after green tea ingestion.
[fig] Figure 1: Diagrammatic representation of the study design. * indicates blood collection. [/fig]
[fig] Figure 3: Glycerol levels at rest, during, and after intermittent sprinting exercise with either placebo or green tea extract (GTE) ingestion. * Significantly greater compared to placebo, p < 0.05. [/fig]
[fig] Figure 4: Epinephrine and norepinephrine levels at rest, during, and after intermittent sprinting exercise with either placebo or green tea extract (GTE) ingestion. * Significantly greater compared to placebo, p < 0.05.The higher plasma glycerol levels at rest and during post-exercise indicate that GTE compared to placebo results in enhanced markers of lipolysis. These results support prior research [/fig]
[table] Table 2: Response at rest and during and after intermittent sprinting exercise in the green tea and placebo conditions (mean and SEM). [/table]
|
Pediatric Veno-Veno Extracorporeal Membrane Oxygenation Rescue From Carbon Monoxide Poisoning
Background: Carbon monoxide poisoning affects approximately 5000children per year and can be challenging to diagnose and treat (Pediatr Emerg Med Pract. 2016;13:1-24). It is in the differential diagnosis of a patient presented with altered consciousness. Patients may look quite "pink" and well perfused, but are often in serious distress. We present the first case in the literature of carbon monoxide poisoning treated with the use of veno-veno extracorporeal membrane oxygenation (ECMO).Case: We report the case of a 10-year-old patient who had carbon monoxide poisoning (carboxyhemoglobin of 18%). She was treated with hydroxocobalamin at 70 mg/kg and was being prepared to transfer to a facility that offered hyperbaric therapy when she suffered a cardiac arrest requiring cardiopulmonary resuscitation. After 11 minutes of resuscitation, she had return of spontaneous circulation and an echocardiogram showed reasonable cardiac function. She was judged too unstable for ambulance transport and the ECMO team was called. Veno-veno ECMO was placed via a single right internal jugular dual-lumen catheter with fluoroscopy in the cardiac catheterization laboratory. There was a rapid improvement in carboxyhemoglobin level, and the ECMO therapy was weaned the next day. The patient eventually made a full recovery.Conclusions: This is the first time that veno-veno ECMO has been reported for the emergent treatment of carbon monoxide intoxication. If emergency physicians are treating such a patient and cannot administer hyperbaric oxygen therapy, ECMO represents a valuable alternative that is not commonly thought of in this situation before.
C arbon monoxide poisoning affects approximately 5000 children per year in the United States and can be challenging to diagnose and treat. [bib_ref] Carbon monoxide poisoning in children: diagnosis and management in the emergency department, Macnow [/bib_ref] The mainstay of treatment is oxygen therapy and supportive care, or hyperbaric oxygen therapy in severe cases. Severe intoxication can be fatal, however, even with these treatments. We present a case of carbon monoxide poisoning treated with the use of veno-veno (VV) extracorporeal membrane oxygenation (ECMO).
## Case
A 10-year-old, 40-kg African American girl was brought to the emergency department by first responders with carbon monoxide poisoning in her home due to a faulty furnace vent. She and a sibling were discovered by family and the patient was brought to our facility. The carbon monoxide meter reading was 500 ppm inside the home, and the period of exposure was unknown. Her younger brother was taken to another hospital but was pronounced dead on arrival. The time to get the patient from home to the emergency department was unclear. The patient was unresponsive upon arrival to the pediatric emergency department with a temporary airway in place; her blood pressure was 98/61 mm Hg; heart rate, 121 beats/min; respirations, 32 breaths/min, and a pulse oximeter saturation of 98%. The initial arterial blood gas (on 100% via ball-valve mask) showed a pH level of 7.06, partial pressure of carbon dioxide (pCO 2 ) of 56, partial pressure of oxygen (pO 2 ) of 517, bicarbonate of 13, base excess of −15, hemoglobin of 14.5, oxyhemoglobin of 82%, and carboxyhemoglobin of 18%. The child was endotracheally intubated and started on hydroxocobalamin at 70 mg/kg. She had cool clammy peripheral extremities. She was treated with volume infusion. The decision was made to transfer the patient to a facility to receive hyperbaric oxygen treatment, and the hydroxocobalamin was discontinued. The child, however, suffered from a cardiac arrest immediately before transfer. Pediatric advanced life support resuscitation was instituted for 11 minutes with return of spontaneous circulation. The patient was placed on an epinephrine infusion. The repeat arterial blood gas revealed a pH level of 7.03, pCO 2 of 63 mm Hg, pO 2 of 462 mm Hg, bicarbonate of 14 mEq/L, and base excess of −15 mEq/L. Because of the arrest, she was not deemed stable for transfer, and despite excellent pO 2 , the ECMO team was called to the bedside. A rapid bedside echocardiogram was performed revealing normal heart function, and therefore, a swift multidisciplinary decision was made to place VV ECMO in the cardiac catheterization laboratory. A summary of her blood gas measurements is in [fig_ref] TABLE 1: Table of Arterial Blood Gas Measurements CarboxyHgb [/fig_ref].
## Procedural details
The patient was brought emergently to the cardiac catheterization laboratory where she was prepped and draped in the usual sterile fashion and remained sedated under the care of a pediatric anesthesiologist. Venous access was obtained via the right internal jugular vein, and a 0.035 inch Amplatz Super-Stiff wire was advanced to the inferior vena cava guided by and confirmed with fluoroscopy. Progressive skin/soft tissue dilation was performed and a 23F catheter (7.7-mm diameter) bicaval cannula was advanced to the intrahepatic inferior vena cava, and the wire was removed after cannula position was confirmed with fluoroscopy. Wet-to-wet connections were made and VV ECMO circulation was commenced with a Maquet Cardiohelp integrated oxygenator/rotary blood pump.
## Outcome
After ECMO initiation, her blood gas revealed a pH level of 7.20, pCO 2 of 28 mm Hg, pO 2 of 424 mm Hg, and carboxyhemoglobin of 6%. The patient was transferred to the pediatric intensive care unit where the blood gas showed a pH level of 7.36, pCO 2 of 19 mm Hg, pO 2 of 430 mm Hg, base excess of −13 mEq/L, and carboxyhemoglobin of 2.7%. The patient's carboxyhemoglobin dropped to 0% after 21 hours of VV ECMO, and the ECMO was weaned off after a total of 24 hours of support.
Previous cases have used venoarterial ECMO to provide complete cardiopulmonary support, but this has complications particularly in small patients and may require an arterial cutdown. Veno-veno ECMO in this case was accomplished with a single jugular venous cannula to be used for therapy. The VV ECMO resulted in reduction of the carboxyhemoglobin and a rapid clinical stabilization. We cannot rule out spontaneous improvement without ECMO, but given her peri-cardiac arrest status, the likely cause of stabilization and improvement was ECMO.
The cannula was removed at bedside with placement of a single purse-string suture. An initial head magnetic resonance imaging showed typical carbon monoxide neurological injury. After a total of 21 days, the patient was discharged to a pediatric rehabilitation facility. She recovered complete physical and neurological function and is in regular classes for her grade and has not been admitted since.
# Discussion
Carbon monoxide binds to hemoglobin, myoglobin, and intracellular cytochromes with a higher affinity than oxygen and hence significant exposure results in poisoning of the blood and cardiac muscle. [bib_ref] Carbon monoxide poisoning: pathogenesis, management, and future directions of therapy, Rose [/bib_ref] [bib_ref] Clinical policy: critical issues in the evaluation and management of adult patients..., Wolf [/bib_ref] Lactic acidosis from poor tissue perfusion results, despite the often "rosy" pink complexion of the poisoning patient. Carboxyhemoglobin is a bright red color similar to oxyhemoglobin. Hyperbaric oxygen treatment is often considered in severe cases to prevent delayed neurological sequelae and results in displacement of carbon monoxide from hemoglobin by shifting the equilibrium to oxyhemoglobin instead. [bib_ref] Hyperbaric oxygen therapy is associated with lower short-and long-term mortality in patients..., Huang [/bib_ref] [bib_ref] Should hyperbaric oxygen be used in acute carbon monoxide poisoning?, Hui-Jun [/bib_ref] Usually, mechanical ventilation with 100% oxygenation is not sufficient to displace the tightly bound carbon monoxide.
Extracorporeal membrane oxygenation has been used in rare cases, but all prior reports involve venoarterial ECMO. [bib_ref] Extracorporeal support in an adult with severe carbon monoxide poisoning and shock..., Mccunn [/bib_ref] [bib_ref] Successful treatment of severe carbon monoxide poisoning and refractory shock using extracorporeal..., Teerapuncharoen [/bib_ref] [bib_ref] Extracorporeal membrane oxygenation (ECMO) for Severe Toxicological Exposures: review of the Toxicology..., Wang [/bib_ref] [bib_ref] Treatment of acute carbon monoxide poisoning with extracorporeal membrane trioxygenation, Yin [/bib_ref] Venoarterial ECMO involves a large central vein and artery and is associated with greater complexity and potentially higher risk of vascular complications, particularly in an emergent setting. Venoarterial ECMO is often needed because of cardiac collapse and severe ventricular dysfunction, which was not noted in this young patient.
Veno-veno ECMO can be performed with a single large (23F-31F catheter) venous cannula, as in the current report, or with 2 femoral venous cannulas. It is important to note that if the patient is unable to be moved to the cardiac catheterization laboratory, the placement of VV ECMO can be done at bedside particularly with imaging such as echocardiography. By using VV ECMO, we were able to rapidly remove carbon monoxide in a patient after arrest who was not a candidate for transfer to a facility with hyperbaric treatment. The ability to treat the patient without the necessity of a large arterial cannula and its attendant complications as well as removal of the venous dual-lumen cannula at bedside with a stitch represent significant advantages to this approach over those described previously.
# Conclusions
The use of short-term VV ECMO has not been reported before and may represent an option in the care of these critically ill patients if other standard of care options is not available. It is applicable to pediatric and adult patients, but it requires the specialized team to be notified as soon as possible to allow for rapid evaluation and treatment.
[table] TABLE 1: Table of Arterial Blood Gas Measurements CarboxyHgb; carboxyhemoglobin; Sat, saturation. FIGURE 1. Enhancement of bilateral globus pallidus. Emergency Care • Volume 36, Number 10, October 2020 ECMO Rescue From Carbon Monoxide Poisoning [/table]
|
Late evolution of arrhythmogenic cardiomyopathy in patients with initial presentation as idiopathic ventricular fibrillation
# Introduction
Arrhythmogenic cardiomyopathy (ACM), also known as arrhythmogenic right ventricular dysplasia/cardiomyopathy, is an inheritable heart muscle disorder in which sudden cardiac death due to ventricular fibrillation (VF) may occur unexpectedly as the first manifestation of the disease. [bib_ref] Impact of genotype on clinical course in arrhythmogenic right ventricular dysplasia/cardiomyopathyassociated mutation..., Bhonsale [/bib_ref] This event is usually preceded by a long preclinical phase. [bib_ref] Arrhythmogenic right ventricular cardiomyopathy, Corrado [/bib_ref] Thus, prior to this event, the disease may frequently go unnoticed owing to the absence of relevant symptoms. This early presymptomatic stage has been defined as "concealed stage," which does not necessarily mean absence of disease; in the absence of symptoms, criteria for ACM diagnosis may even be already present. [bib_ref] Arrhythmogenic right ventricular cardiomyopathy, Corrado [/bib_ref] Since 2010, ACM diagnosis is made according to revised Task Force Criteria (2010 TFC) obtained by international consensus. [bib_ref] Diagnosis of arrhythmogenic right ventricular cardiomyopathy/dysplasia, Marcus [/bib_ref] Fulfillment of these 2010 TFC (ie, presence of 2 major, 1 major plus 2 minor, or 4 minor criteria) is required for "definite" ACM diagnosis. Although fulfillment of TFC is required for definite ACM diagnosis, in the early disease stage it is conceivable that VF may occur in the presence of most but not all criteria. These cases are often regarded as idiopathic VF (IVF).
In contrast to VF due to ACM, in IVF all known cardiac, respiratory, metabolic, and toxicologic etiologies should have been excluded by the clinical evaluation available at the time of the arrhythmic event. [bib_ref] Idiopathic ventricular fibrillation, Visser [/bib_ref] However, in our experience, some survivors of initially unexplained cardiac arrest may develop "definite" ACM years after the index event, 5 either by progression of the disease or by improvement of diagnostic tools.
We present 2 cases of ACM patients with no or minimal disease at the time of their initial cardiac arrest episode. These episodes occurred in 1991 and 1995, thus before availability of the 2010 TFC. This means that ACM diagnosis was initially based on the less sensitive TFC published in 1994 6 in 1 case, and on clinical and electrocardiogram (ECG) characteristics in the other.
## Case reports case 1
A 24-year-old man collapsed twice in 1 year while playing in a soccer match. The first time, in 1991, he regained consciousness shortly after the start of resuscitation. Clinical evaluation was inconclusive. The second time, in 1992, he received 2 external defibrillator shocks because of VF. Resuscitation was successful and he was admitted for diagnostic evaluation. ECG showed sinus rhythm with intraventricular conduction delay, including prolonged terminal activation duration (70 ms), and J-point elevation in the inferior leads. History, Holter monitoring, and exercise test were unremarkable, and no ventricular extrasystoles
## Key teaching points
Idiopathic ventricular fibrillation is a diagnosis by exclusion. Absence of diagnosis does not mean absence of disease.
Reevaluation of diagnosis during follow-up is important, as diagnosis of genetic disease can impact treatment and familial risk stratification.
Diagnosis of arrhythmogenic cardiomyopathy in a later stage in patients initially presented with idiopathic ventricular fibrillation may be due to improvement of diagnostic tools, progression of disease, or misinterpretation of initially available data. were reported. Family history was negative for sudden cardiac death. Echocardiogram showed normal cardiac function and biventricular dimensions; however, a local abnormality under the tricuspid valve was noted. Transesophageal echocardiogram was performed, in which this abnormality was characterized as prolapse of a tricuspid valve leaflet. During electrophysiologic study, a prolonged HV interval of 60-70 ms was measured. Right ventricular (RV) stimulation induced polymorphic ventricular tachycardia (VT) starting as monomorphic VT with left bundle branch block morphology and superior axis. Other diagnostic tests, including coronary angiogram, myocardial biopsy, and ergonovine provocation test, were normal. An implantable cardioverter-defibrillator (ICD) was implanted and he was discharged. In the absence of an obvious etiology, IVF was the initial diagnosis. Cardiac magnetic resonance imaging (CMR) and signal-averaged ECG were not performed.
Ten years after the initial presentation, negative T waves were recorded in V 1 -V 3. Thus, he went on to fulfill definite ACM diagnosis criteria. On follow-up echocardiogram, there were no signs of tricuspid valve prolapse anymore, but careful reevaluation of imaging showed subtricuspid dyskinesia. After 16 years, RV dilatation was noted (echocardiogram 2008: dilated, hypokinetic RV; parasternal long-axis RV outflow tract [RVOT]: 32 mm, parasternal short-axis RVOT: 35 mm). Consecutive molecular genetic testing revealed an unclassified variant in the DSG2 gene and a p.Leu729del mutation in the sodium channel gene SCN5A. Familial cosegregation supported pathogenicity of this SCN5A mutation [fig_ref] Figure 2: Pedigree of the p [/fig_ref]. During 25 years of follow-up, he remained free of ventricular tachyarrhythmias.
In this case, fulfillment of 2010 TFC in follow-up was owing to disease progression (eg, progression of precordial T-wave inversions), misinterpretation of clinical evaluation (eg, wrong interpretation of subtricuspid dyskinesia as tricuspid valve prolapse), and incomplete clinical evaluation (no CMR was performed).
## Case 2
In 1995, a 25-year-old man collapsed during light chores while at work in a café. He was resuscitated and admitted to the intensive care unit. His history was unremarkable and there was no sudden death in his family. Electrocardiography showed sinus rhythm with negative T waves in leads V 4 -V 6 and low voltage in extremity leads. Echocardiogram, coronary angiogram, and myocardial biopsy were normal. The left ventricular ejection fraction was 62% by echocardiography. During electrophysiologic study, RVOT stimulation showed inducibility of a monomorphic VT with left bundle branch block morphology and vertical axis. An ICD was implanted and he was discharged. Thus, at initial evaluation, he fulfilled 2 minor TFC for ACM diagnosis, indicating IVF as initial diagnosis.
# Discussion
## Absence of diagnosis does not mean absence of disease
We present 2 cases with an initial presentation of nonfatal cardiac arrest due to underlying VF. Since only nonspecific minor abnormalities were identified at the time of presentation, these cases were diagnosed as IVF. Many years after the initial presentation, TFC fulfillment led to a definite diagnosis of ACM [fig_ref] Table 1: Overview of Task Force Criteria fulfilled per case ACM 5 arrhythmogenic cardiomyopathy [/fig_ref].
The presented cases suggest that IVF or "concealed" ACM does not necessarily mean absence of disease, and periodic reevaluation of suspected ACM cases may lead to a definite diagnosis. Diagnosing ACM in former IVF patients is important, as it provides additional management options such as pharmacologic therapy and lifestyle interventions. [bib_ref] Arrhythmogenic right ventricular cardiomyopathy, Corrado [/bib_ref] More importantly, in family members of patients with a genetic disorder, a correct diagnosis may pave the way for cascade screening as a first step toward arrhythmic risk stratification.
## Reevaluation of diagnosis
IVF is a diagnosis by exclusion in which initial evaluation does not identify an underlying cause. Diagnosing a specific disease, such as ACM, during follow-up may be a consequence of (1) improvement in diagnostic tools, (2) disease progression, or (3) misinterpretation of initially available data. These are common challenges in clinical practice, and a combination of these is illustrated in this report; our cases may be part of a larger patient population that is not diagnosed at the time of clinical presentation.
## Improvement in diagnostic tools
The ability of diagnostic tests to detect electrical and structural abnormalities in the initial evaluation of patients suspected of ACM varies considerably. Qualitative assessment of RV angiogram relies on clinical experience, diagnostic sensitivity of myocardial biopsy is limited by the segmental nature of disease, [bib_ref] Diagnosis of arrhythmogenic right ventricular cardiomyopathy/dysplasia, Marcus [/bib_ref] and echocardiography's performance is significantly less compared to CMR in detecting abnormalities fulfilling the 2010 TFC. [bib_ref] The diagnostic performance of imaging methods in ARVC using the 2010 task..., Borgquist [/bib_ref] In both these presented cases, CMR was not performed, which may have impacted our ability to fully appreciate the structural phenotype in these individuals. The improved availability in recent years of CMR has made the detection of ACM with a subtler phenotype possible. Based on this information and our clinical experience, we strongly suggest a comprehensive CMR be performed in the work-up of a patient with presumed IVF and used for evaluation of ACM diagnosis in suspected cases.
## Progressive disease
ACM is a progressive disease, [bib_ref] Evaluation of structural progression in arrhythmogenic right ventricular dysplasia/cardiomyopathy, Mast [/bib_ref] and its disease course in desmosomal mutation carriers is well studied. 9 A long-lasting asymptomatic phase precedes the clinical phase in most ACM patients, in which electrical abnormalities tend to precede detectable structural changes. Moreover, presence of both an electrical and structural substrate identifies patients at high risk for arrhythmic events. [bib_ref] Incremental value of cardiac magnetic resonance imaging in arrhythmic risk stratification of..., Te Riele [/bib_ref] Accordingly, diagnosis of ACM according to the 2010 TFC is based on electrical, structural, genetic, and histologic abnormalities 3 and requires extensive diagnostic testing and periodic reevaluation in suspected cases. In case 1, right precordial T-wave inversion and echocardiographic RV dilatation after 10 years of follow-up are indicators of disease progression, facilitating ACM diagnosis.
## Misinterpretation of initially available data
Our knowledge of the underlying disease process and associated pathogenic mutations has facilitated diagnosis of more subtle and heterogeneous clinical presentations of ACM. Retrospectively, subtle abnormalities could be misinterpreted during initial evaluation, as is illustrated by both our cases.
In both patients, pathogenic mutations were found in nondesmosomal genes. The involvement of nondesmosomal genes in ACM has become clearer in recent years. For example, overlap of desmosomal and sodium channel disease (eg, ACM and Brugada syndrome) is an emerging area of interest (as described in case 1) [bib_ref] Relationship between arrhythmogenic right ventricular cardiomyopathy and brugada syndrome, Corrado [/bib_ref] ; the phospholamban gene mutation found in case 2 is associated with both an arrhythmogenic and dilated cardiomyopathy phenotype. [bib_ref] Phospholamban R14del mutation in patients diagnosed with dilated cardiomyopathy or arrhythmogenic right..., Van Der Zwaag [/bib_ref] In case 2, increased awareness for biventricular/leftdominant forms in the new 2010 TFC and the availability of molecular genetic testing facilitated the diagnosis during follow-up. In case 1, already at baseline, abnormal wall motion in the subtricuspid region was noted, which was falsely interpreted as tricuspid valve prolapse. Since the subtricuspid region is a hotspot region for early structural changes in ACM, [bib_ref] Mutation-positive arrhythmogenic right ventricular dysplasia/cardiomyopathy, Te Riele [/bib_ref] overt ACM was likely already present, albeit not recognized.
# Conclusion
VF may be the first clinical manifestation of ACM, as shown in this report. Appropriate diagnosis at later disease stages may be due to (1) improvement in diagnostic tools, (2) disease progression, or (3) misinterpretation of initially available data. In a VF survivor, a specific ACM diagnosis may be clinically less relevant for the index patient. However, in family members, a correct diagnosis may pave the way for cascade screening as a first step toward arrhythmic risk stratification. To detect affected family members before occurrence of ventricular arrhythmias and sudden death, periodic reevaluation of suspected ACM cases during follow-up is important.
[fig] KEYWORDS: Arrhythmogenic cardiomyopathy; Diagnostic testing; Disease progression; Electrophysiology; Idiopathic ventricular fibrillation; Ventricular arrhythmia (Heart Rhythm Case Reports 2019;5:25-30) [/fig]
[fig] Figure 1 A: : A 12-lead electrocardiogram (ECG) while off antiarrhythmic drugs, showing sinus rhythm, QRS right axis deviation, QRS width 140 ms, and prolonged terminal activation duration in V 2 (70 ms). Clear J-point and ST elevation in II, III, and aVF. B: ECG after 10 years of follow-up, showing negative T-waves in V 1 -V 3 . [/fig]
[fig] Figure 2: Pedigree of the p.Leu729del mutation carrier (case 1) reveals co-segregation of the variant with the arrhythmogenic cardiomyopathy phenotype with reduced penetrance. ARVD/C 5 arrhythmogenic right ventricular dysplasia/cardiomyopathy; Asx 5 asymptomatic; .TAD 5 prolonged terminal activation duration. Symbols: square 5 male; circle 5 female; 1 5 SCN5A mutation carrier; (1) 5 obligate SCN5A mutation carrier; solid 5 clinical ARVD/C diagnosis; gray 5 clinical symptoms and/or borderline ARVD/C diagnosis; empty 5 negative phenotype for ARVD/C. *No abnormalities on comprehensive cardiac evaluation including 12-lead electrocardiogram, Holter monitoring, and cardiac imaging. (From Te Riele and colleagues. 14 ) [/fig]
[fig] Figure 3 A: : A 12-lead low-voltage electrocardiogram while off antiarrhythmic drugs showing normal sinus rhythm and QRS right axis deviation. Negative T waves in left precordial recordings and I, II, and aVF. B: Transthoracic echocardiogram during follow-up showing a dilated right ventricle with subtricuspid aneurysm (arrow). [/fig]
[table] Table 1: Overview of Task Force Criteria fulfilled per case ACM 5 arrhythmogenic cardiomyopathy; LBBB 5 left bundle branch block; PLAX 5 parasternal long axis; RV 5 right ventricle; RVOT 5 right ventricular outflow tract; TAD 5 terminal activation duration; TFC 5 Task Force Criteria; VT 5 ventricular tachycardia. [/table]
|
Accuracy of electrocardiographic criteria for atrial enlargement: validation with cardiovascular magnetic resonance
Background: Anatomic atrial enlargement is associated with significant morbidity and mortality. However, atrial enlargement may not correlate with clinical measures such as electrocardiographic (ECG) criteria. Past studies correlating ECG criteria with anatomic measures mainly used inferior M-mode or two-dimensional echocardiographic data. We sought to determine the accuracy of the ECG to predict anatomic atrial enlargement as determined by volumetric cardiovascular magnetic resonance (CMR).Methods: ECG criteria for left (LAE) and right atrial enlargement (RAE) were compared to CMR atrial volume index measurements for 275 consecutive subjects referred for CMR (67% males, 51 ± 14 years). ECG criteria for LAE and RAE were assessed by an expert observer blinded to CMR data. Atrial volume index was computed using the biplane area-length method.Results:The prevalence of CMR LAE and RAE was 28% and 11%, respectively, and by any ECG criteria was 82% and 5%, respectively. Though nonspecific, the presence of at least one ECG criteria for LAE was 90% sensitive for CMR LAE. The individual criteria P mitrale, P wave axis < 30°, and negative P terminal force in V1 (NPTF-V1) > 0.04s·mm were 88-99% specific although not sensitive for CMR LAE. ECG was insensitive but 96-100% specific for CMR RAE.Conclusion:The presence of at least one ECG criteria for LAE is sensitive but not specific for anatomic LAE. Individual criteria for LAE, including P mitrale, P wave axis < 30°, or NPTF-V1 > 0.04s·mm are highly specific, though not sensitive. ECG is highly specific but insensitive for RAE. Individual ECG P wave changes do not reliably both detect and predict anatomic atrial enlargement.
# Introduction
Atrial enlargement is a marker of increased cardiovascular events. Anatomic left atrial (LA) enlargement (LAE) is a marker of left ventricular (LV) diastolic dysfunction [bib_ref] Left atrial volume as a morphophysiologic expression of left ventricular diastolic dysfunction..., Tsang [/bib_ref] and is associated with an abnormal stress test in subjects with known or suspected coronary artery disease [bib_ref] Predictive value of normal left atrial volume in stress echocardiography, Alsaileek [/bib_ref]. In addition, it is a predictor for the development of atrial fibrillation [bib_ref] Can P wave parameters obtained from 12-lead surface electrocardiogram be a predictor..., Altunkeser [/bib_ref] [bib_ref] Left atrial volume: important risk marker of incident atrial fibrillation in 1655..., Tsang [/bib_ref] [bib_ref] Risks for atrial fibrillation and congestive heart failure in patients >/= 65..., Tsang [/bib_ref] , congestive heart failure [bib_ref] Risks for atrial fibrillation and congestive heart failure in patients >/= 65..., Tsang [/bib_ref] , stroke [bib_ref] Left atrial volume in the prediction of first ischemic stroke in an..., Barnes [/bib_ref] , increased cardiac mortality [bib_ref] Left atrial volume in the prediction of first ischemic stroke in an..., Barnes [/bib_ref] [bib_ref] Left atrial size is the major predictor of cardiac death and overall..., Modena [/bib_ref] , incidence and survival after myocardial infarction [bib_ref] Left atrial volume: a powerful predictor of survival after acute myocardial infarction, Moller [/bib_ref] [bib_ref] Prediction of cardiovascular outcomes with left atrial size: is volume superior to..., Tsang [/bib_ref] [bib_ref] The relationship between renal function and cardiac structure, function, and prognosis after..., Verma [/bib_ref] , and combined cardiovascular events [bib_ref] Prediction of cardiovascular outcomes with left atrial size: is volume superior to..., Tsang [/bib_ref] [bib_ref] Prediction of risk for first age-related cardiovascular events in an elderly population:..., Tsang [/bib_ref]. Right atrial (RA) enlargement (RAE) is associated with increased risk for congestive heart failure [bib_ref] The value of the electrocardiogram and chest X-ray for confirming or refuting..., Fonseca [/bib_ref] and increased mortality in patients with primary pulmonary hypertension [bib_ref] Echocardiographic predictors of adverse outcomes in primary pulmonary hypertension, Raymond [/bib_ref].
The electrocardiogram (ECG) is used ubiquitously in clinical practice to evaluate patients with cardiac disease. ECG criteria for atrial enlargement have shown poor sensitivity and moderate specificity for detecting enlargement as defined by two-dimensional echocardiography parasternal long axis (PLAX) and four-chamber (4 Ch) measurements [bib_ref] Electrocardiographic detection of left atrial enlargement. Correlation of P wave with left..., Chirife [/bib_ref] [bib_ref] Diagnostic accuracy of the resting electrocardiogram in detection and estimation of left..., Hazen [/bib_ref] [bib_ref] Electrocardiographic diagnosis of left atrial enlargement. Role of the P terminal force..., Hopkins [/bib_ref] [bib_ref] Sensitivity and specificity of commonly used electrocardiographic criteria for left atrial enlargement..., Munuswamy [/bib_ref] [bib_ref] Reassessment of electrovectorcardiographic signs of left atrial enlargement, Perosio [/bib_ref] [bib_ref] ECG criteria for right atrial enlargement, Reeves [/bib_ref] [bib_ref] Comparison of left atrial size and pulmonary capillary pressure with P wave..., Rubler [/bib_ref] [bib_ref] Left atrial enlargement: an electrocardiographic misnomer? An electrocardiographic-echocardiographic study, Van Dam [/bib_ref] [bib_ref] Left atrial enlargement. Echocardiographic assessment of electrocardiographic criteria, Waggoner [/bib_ref]. However, echocardiographic measures are subject to error due to variability in acquisition of appropriately aligned images, limitations in acoustic windows, and other causes of technical variation. In addition, because the LA is not a sphere with a constant radius [bib_ref] Best method in clinical practice and in research studies to determine left..., Lester [/bib_ref] , uni-or two-dimensional (2-D) echocardiographic measurements may not reflect true chamber size. Cardiovascular magnetic resonance (CMR) imaging can provide images in standardized planes, thereby providing more precise volumetric assessment of cardiac chamber size. We sought to determine the accuracy of the ECG for detection of atrial enlargement (AE) using the gold standard of volumetric CMR.
# Methods
## Study population
The study population consisted of 275 consecutive subjects referred for CMR either for clinical (n= 255) or research (n = 20) purposes, between February 2001 and July 2004. Clinical CMR indications included evaluation for the following: left ventricular function (n = 111), pulmonary veins (n = 74), coronary arteries (n = 47), perfusion and/or viability (n = 43), cardiomyopathy (n = 38), intracardiac shunt (n = 4), valvular function (n = 3), pericardium (n = 2), myocarditis (n = 1), and cardiac mass (n = 1). Many subjects were referred for more than one indication. Subjects were excluded if no sinus rhythm ECG was available (see below). Age, gender, height, weight, as well as history of atrial fibrillation (AF) and hypertension were recorded. Research subjects gave informed consent according to a protocol approved by the hospital institutional review board.
## Electrocardiography
A standard 12-lead ECG was performed on the same day of the CMR in all subjects in sinus rhythm (n = 201). An ECG in sinus rhythm, no more than 35 days from the CMR scan and prior to any pulmonary vein ablation procedures, was obtained for subjects with AF (n = 74). ECG LAE was defined by any one of the following: 1) P wave in any lead > 0.11s, 2) Notched P wave with interpeak duration > 0.04s (P mitrale), 3) P wave axis < 30°, 4) Area of negative P terminal force in lead V1 (NPTF-V1) > 0.04s·mm, or 5) Positive P terminal force in aVL (PPTF-aVL) > 0.5 mm [bib_ref] Diagnostic accuracy of the resting electrocardiogram in detection and estimation of left..., Hazen [/bib_ref] [bib_ref] Sensitivity and specificity of commonly used electrocardiographic criteria for left atrial enlargement..., Munuswamy [/bib_ref] [bib_ref] Left atrial enlargement. Echocardiographic assessment of electrocardiographic criteria, Waggoner [/bib_ref]. ECG RAE included 1) P wave in inferior leads II, III, aVF > 2.5 mm or 2) Positive P wave in V1 > 1.5 mm [bib_ref] In Electrocardiography in Clinical Practice, Chou [/bib_ref] [bib_ref] Atrial abnormalities, Friedman [/bib_ref]. All ECG determinations were made by one experienced observer (MEJ) blinded to other results.
## Cmr technique
CMR was performed on a 1.5T whole-body scanner (Gyroscan NT, Philips Medical Systems, NL). Among subjects with AF, the CMR scan was performed prior to any pulmonary vein isolation procedure. Participants were imaged in the supine position with a phased-array fiveelement cardiac synergy coil for radiofrequency signal reception. Localizing scans were followed by free-breathing axial spin-echo (repetition time 800 ms, echo time 20 ms, field of view 300 mm, matrix size 192 × 512, slice thickness 6 mm, 0.5 mm gap) and end-expiratory, breathhold, ECG-gated SSFP acquisitions (temporal resolution 33 ms, repetition time 3.1 ms, echo time 1.6 ms, flip angle 60 degrees, field of view 400 mm, matrix size 208 × 256, slice thickness 10 mm, gap = 0).
# Image analysis
CMR analyses were performed using dedicated software (EasyVision 5.1, Philips Medical Systems, Best, NL). LA length was measured from the posterior wall to the plane of the mitral annulus, parallel to the long-axis of the heart, in the two-chamber (2 Ch) and four-chamber (4 Ch) orientations on cine SSFP acquisitions at maximum atrial diastole [fig_ref] Figure 1: CMR images with representative measurements at atrial end-diastole [/fig_ref] , B). Maximum atrial diastole was defined as the image immediately preceding the opening of the mitral and tricuspid valves. Similarly, RA length was measured from the posterior wall of the RA to the plane of the tricuspid annulus in the 4 Ch orientation [fig_ref] Figure 1: CMR images with representative measurements at atrial end-diastole [/fig_ref]. LA area in apical 2 Ch and 4 Ch orientation were planimetered by tracing the endocardial border in maximum atrial diastole, excluding the confluence of the pulmonary veins and LA appendage [fig_ref] Figure 1: CMR images with representative measurements at atrial end-diastole [/fig_ref]. Similarly, the 4 Ch RA area was planimetered, excluding the confluence of the vena cavae and the atrial appendage [fig_ref] Figure 1: CMR images with representative measurements at atrial end-diastole [/fig_ref]. The borders of the atria were delimited at the planes of the AV annulus and the junctions of venous inflow. Atrial volumes were calculated according to the biplane area-length method, then indexed for body surface area (BSA) . Enlarged CMR volume index was defined to be ≥ 55 ml/m 2 , two standard deviations (SD) above the mean of normal, healthy subjects in published population studies [31-36]. CMR analyses were performed by two experienced observers (CWT, TDO) blinded to other results.
# Statistical analysis
Continuous data are presented as mean ± SD. Categorical data are presented as counts and percentages. The difference in continuous measures between subject subgroups was assessed using Student's t test. The difference in categorical measures between subject subgroups and the prevalence of AE among the subject subgroups were assessed using Fisher's exact test. Sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) were used to assess the test performance for ECG LAE and RAE. The accuracy of ECG LAE and RAE were computed individually and in combination. The independent relationships of hypertension and CMR LAE to ECG LAE were further evaluated with logistic regression. A p-value of ≤ 0.05 was considered statistically significant. All statistical analyses were performed using SAS for Windows (v.9.1, SAS Institute, Cary, NC).
# Results
## Study population
The 275 study subjects consisted of 184 (67%) males, 91 females (age 51 ± 14 years; range 18 to 83 years) who were consecutively referred for clinical or research CMR study, including 27% subjects with a history of AF and 34% with a history of hypertension. For the entire group, CMR LA and RA volumes were 97 ± 38 ml and 79 ± 32 ml, respec-CMR images with representative measurements at atrial end-diastole [fig_ref] Table 1: Prevalence of Atrial Enlargement by CMR and ECGCMR [/fig_ref]. Normotensive subjects and those with a history of hypertension had a similar prevalence of CMR LAE (30%, p = 1.0). Compared to subjects in sinus rhythm, subjects with a history of AF had a significantly higher prevalence of CMR LAE and but not CMR RAE .
## Accuracy of ecg for anatomic atrial enlargement
The accuracy of ECG LAE is presented in [fig_ref] Table 2: Accuracy of ECG criteria for Left and Right Atrial Enlargement as Detected... [/fig_ref]. The presence of at least one ECG LAE criteria had a sensitivity of 90% with a NPV of 84% and specificity of 21% for detection of CMR LAE. When ECG criteria were analyzed independently, P wave > 0.11s offered the highest sensitivity for CMR LAE, with a NPV of 85%. Three individual ECG criteria were very specific for detecting CMR LAE. The presence of P mitrale, P wave axis < 30°, and NPTF-V1 > 0.04s·mm were 99%, 90%, and 88% specific, respectively, for detecting CMR LAE, although these criteria were not sensitive. The presence of P mitrale had a high PPV of 86% for CMR LAE.
The ECG was insensitive but highly specific for CMR RAE [fig_ref] Table 2: Accuracy of ECG criteria for Left and Right Atrial Enlargement as Detected... [/fig_ref]. The presence of P wave in leads II, III, or aVF > 2.5 mm or positive P wave in V1 > 1.5 mm had specificities of 100% and 96%, respectively, for diagnosing CMR RAE. Both criteria had a NPV of 90%, indicating low rates of false-negative results.
## Association of hypertension with ecg lae
# Discussion
The commonly used ECG criteria for atrial enlargement have been compared to anatomic atrial size by M-mode or 2-D echocardiography [bib_ref] Electrocardiographic detection of left atrial enlargement. Correlation of P wave with left..., Chirife [/bib_ref] [bib_ref] Diagnostic accuracy of the resting electrocardiogram in detection and estimation of left..., Hazen [/bib_ref] [bib_ref] Electrocardiographic diagnosis of left atrial enlargement. Role of the P terminal force..., Hopkins [/bib_ref] [bib_ref] Sensitivity and specificity of commonly used electrocardiographic criteria for left atrial enlargement..., Munuswamy [/bib_ref] [bib_ref] Reassessment of electrovectorcardiographic signs of left atrial enlargement, Perosio [/bib_ref] [bib_ref] ECG criteria for right atrial enlargement, Reeves [/bib_ref] [bib_ref] Comparison of left atrial size and pulmonary capillary pressure with P wave..., Rubler [/bib_ref] [bib_ref] Left atrial enlargement: an electrocardiographic misnomer? An electrocardiographic-echocardiographic study, Van Dam [/bib_ref] [bib_ref] Left atrial enlargement. Echocardiographic assessment of electrocardiographic criteria, Waggoner [/bib_ref] , which have inherent disadvantages including limited acoustic windows and spatial resolution, geometric assumptions made for estimation of atrial size, and moderate interobserver variability of measurements. The goal of this study was to characterize the accuracy of commonly used ECG atrial enlargement criteria against a gold standard of volumetric CMR measurements.
The presence of any ECG LAE criteria was 90% sensitive with a NPV of 84%, but had poor specificity of only 21% for CMR LAE. Using a combination of ECG LAE criteria is thus highly sensitive for detection of anatomic LAE, with a low false negative rate. This study indicates that commonly used ECG LAE criteria are more sensitive for anatomic LAE than the 6-69% sensitivity previously reported using echocardiographic standards [bib_ref] Diagnostic accuracy of the resting electrocardiogram in detection and estimation of left..., Hazen [/bib_ref] [bib_ref] Sensitivity and specificity of commonly used electrocardiographic criteria for left atrial enlargement..., Munuswamy [/bib_ref] [bib_ref] Left atrial enlargement: an electrocardiographic misnomer? An electrocardiographic-echocardiographic study, Van Dam [/bib_ref] [bib_ref] Left atrial enlargement. Echocardiographic assessment of electrocardiographic criteria, Waggoner [/bib_ref]. The criteria of P mitrale, P wave axis < 30°, and terminal P wave in V1 > 0.04s·mm were each highly specific although not sensitive for CMR LAE. In our CMR study, we found higher specificities of P mitrale and terminal P wave in V1 > 0.04s·mm for anatomic LAE than those obtained by similar atrial bi-plane volume analysis using 2-D echocardiography . In addition, the high specificity of the criteria of P wave axis < 30° to detect LAE has not been previously described to our knowledge. Compared with prior M-mode echocardiographic studies [bib_ref] Diagnostic accuracy of the resting electrocardiogram in detection and estimation of left..., Hazen [/bib_ref] [bib_ref] Sensitivity and specificity of commonly used electrocardiographic criteria for left atrial enlargement..., Munuswamy [/bib_ref] [bib_ref] Left atrial enlargement. Echocardiographic assessment of electrocardiographic criteria, Waggoner [/bib_ref] , we found higher sensitivity but lower specificity for P > 0.11s, and similar excellent specificity for P mitrale and NPTF-V1 > 0.04s·mm to detect CMR LAE. The overall increased sensitivity of ECG LAE compared to ECG RAE may be explained by a larger number of criteria used, as well as high prevalence of the very sensitive criteria P > 0.11s.
Our results indicate that ECG criteria are insensitive but highly specific to detect CMR RAE, and are in agreement with analyses using echocardiographic volumetric RAE [39]. We found similar sensitivity but greatly improved specificity and both positive and negative predictive values of ECG for CMR RAE compared to 2-D echocardiographic standards . No individual ECG criteria had both high sensitivity and specificity for predicting anatomic RAE. Though many of the commonly used individual ECG LAE and RAE criteria have high specificity for anatomic AE, their prevalence is rare even in our referral population.
Prevalence of CMR atrial enlargement (AE) among subjects in sinus rhythm (SR) and with atrial fibrillation (AF) Prevalence of CMR atrial enlargement (AE) among subjects in sinus rhythm (SR) and with atrial fibrillation (AF). Subjects with AF have a significantly higher prevalence of left (LAE) but not right atrial enlargement (RAE), compared with subjects in SR. Abnormalties in atrial conduction resulting from hypertension, elevated pulmonary capillary wedge pressure, or intrinsic conduction defects are associated with abnormalities in the P wave [37,46,47] and may be independent of atrial size. Thus, it is not surprising that in our study, ECG LAE criteria were sensitive but not specific. Since multiple etiologies may account for similar conduction disturbances in the P wave, the manifestation of such ECG criteria may be more accurately described as indicating atrial "abnormality" rather than "enlargement".
In logistic regression analysis, we demonstrated that CMR LAE was associated with ECG LAE independently of hypertension. However, a limitation of our study was that we were not able to adjust for other factors which may affect P wave morphology, such as the use of sodium channel blocking medications, as this information was unavailable. In addition, since this study was noninvasive, we were not able to measure atrial pressures, which may also affect the P wave [46].
# Conclusion
Our results further validate the strengths and weaknesses of commonly used ECG LAE and RAE criteria against the gold standard of anatomic measurements. Compared with volumetric CMR measures, none of the commonly used ECG LAE or RAE criteria provided high accuracy for detecting anatomic LAE or RAE. High sensitivity was achieved only with lower specificity and vice versa. The presence of any ECG LAE criteria and the individual criterion of P > 0.11s were highly sensitive for CMR LAE, while the criteria P mitrale, P wave axis < 30°, and NPTF-V1 > 0.04s·mm were very specific, though poorly sensitive, for detecting CMR LAE. ECG RAE criteria were highly specific, though insensitive, for detecting anatomic RAE. The limited specificity of the most sensitive ECG criteria for anatomic atrial enlargement is consistent with the fact that P wave perturbations reflect changes in atrial conduction, which may be due to a variety of etiologies. Thus, the term atrial "abnormality" may be more appropriate than "enlargement" to describe such ECG findings.
[fig] Figure 1: CMR images with representative measurements at atrial end-diastole. (A) Two-chamber (2 Ch) view with left atrial (LA) length measurement. (B) Apical four-chamber (4 Ch) view with LA and right atrial (RA) 4 Ch length measurement. (C) 2 Ch view with LA endocardial border tracings. (D) 4 Ch view with LA and RA endocardial border tracings. P: Posterior. L: Left. respectively. The prevalence of CMR AE and ECG AE is presented in [/fig]
[table] Table 1: Prevalence of Atrial Enlargement by CMR and ECGCMR: Cardiovascular magnetic resonance. ECG: Electrocardiography. LAE: Left atrial enlargement. RAE: Right atrial enlargement. P mitrale: Notched P wave with interpeak duration > 0.04s. NPTF-V1 > 0.04s·mm: Area of negative P terminal force in lead V1 > 0.04s·mm. PPTF-aVL: Positive P terminal force in lead aVL. Inferior P waves: P wave in any of the leads II, III, or aVF. [/table]
[table] Table 2: Accuracy of ECG criteria for Left and Right Atrial Enlargement as Detected by CMRCMR: Cardiovascular magnetic resonance. ECG: Electrocardiography. PPV: Positive predictive value. NPV: Negative predictive value. LAE: Left atrial enlargement. RAE: Right atrial enlargement. P mitrale: Notched P wave with interpeak duration > 0.04s. NPTF-V1 > 0.04s·mm: Area of negative P terminal force in lead V1 > 0.04s·mm. PPTF-aVL: Positive P terminal force in lead aVL. Inferior P waves: P wave in any of the leads II, III, or aVF. [/table]
|
Multiple structure alignment and consensus identification for proteins
Background: An algorithm is presented to compute a multiple structure alignment for a set of proteins and to generate a consensus (pseudo) protein which captures common substructures present in the given proteins. The algorithm represents each protein as a sequence of triples of coordinates of the alpha-carbon atoms along the backbone. It then computes iteratively a sequence of transformation matrices (i.e., translations and rotations) to align the proteins in space and generate the consensus. The algorithm is a heuristic in that it computes an approximation to the optimal alignment that minimizes the sum of the pairwise distances between the consensus and the transformed proteins. Results: Experimental results show that the algorithm converges quite rapidly and generates consensus structures that are visually similar to the input proteins. A comparison with other coordinate-based alignment algorithms (MAMMOTH and MATT) shows that the proposed algorithm is competitive in terms of speed and the sizes of the conserved regions discovered in an extensive benchmark dataset derived from the HOMSTRAD and SABmark databases. The algorithm has been implemented in C++ and can be downloaded from the project's web page. Alternatively, the algorithm can be used via a web server which makes it possible to align protein structures by uploading files from local disk or by downloading protein data from the RCSB Protein Data Bank.Conclusions: An algorithm is presented to compute a multiple structure alignment for a set of proteins, together with their consensus structure. Experimental results show its effectiveness in terms of the quality of the alignment and computational cost.
# Background
This paper presents an algorithm to compute a multiple structure alignment for a set of proteins and to generate a consensus structure. The algorithm is called MAPSCI, which stands for Multiple Alignment of Protein Structures and Consensus Identification. MAPSCI addresses the problem of global structure alignment, which has also been considered by CE-MC [bib_ref] A new algorithm for the alignment of multiple protein structures using Monte..., Guda [/bib_ref] , MAMMOTH [bib_ref] A new progressive-iterative algorithm for multiple structure alignment, Lupyan [/bib_ref] , and MATT [bib_ref] Matt: Local Flexibility Aids Protein Multiple Structure Alignment, Menke [/bib_ref]. Specifically, MAPSCI computes an approximation to the multiple structure alignment that minimizes the so-called Sum-of-Consensus distance (SCdistance), i.e. the sum of the pairwise distances between the consensus structure and each protein in the set (see the Methods section for the precise definition of SC-distance). Our experiments show that MAPSCI converges quite rapidly and produces alignments that compare favorably with the alignments produced by MAM-MOTH and MATT. The consensus structures generated by MAPSCI are visually quite similar to the input proteins. Although the consensus structures are not real proteins, they could be used, for instance, as templates to perform fast searches through protein structure databases, such as the Protein Data Dank [bib_ref] The Protein Data Bank, Berman [/bib_ref] , to identify structurally similar proteins.
MAPSCI has similar structure to the algorithm of Ye and Janardan [bib_ref] Approximate multiple protein structure alignment using the Sum-of-Pairs distance, Ye [/bib_ref]. However, MAPSCI works directly on the coordinates of the C a atoms and produces true alignments; by contrast, the algorithm in [bib_ref] Approximate multiple protein structure alignment using the Sum-of-Pairs distance, Ye [/bib_ref] requires that the backbone vectors be translated to the origin, hence information about the relative positions of the C a atoms in R 3 is lost and as a result the algorithm does not generate true alignments. The Methods section presents the mathematical and algorithmic framework of MAPSCI and provides the complete details where the two algorithms differ significantly; when there is an overlap the reader is referred to publication [bib_ref] Approximate multiple protein structure alignment using the Sum-of-Pairs distance, Ye [/bib_ref]. Implementation MAPSCI represents the input proteins and the consensus as sequences of triples of coordinates of the alphacarbon (or C a ) atoms along the backbone. It then computes a correspondence between the coordinate triples of the C a atoms in the different protein structures by choosing one of the proteins as the initial consensus and applying an algorithm that is analogous to the center-star method for multiple sequence alignment. Next, MAPSCI derives a set of translation and rotation matrices that are optimal for the computed correspondence and uses these to align the structures in space via rigid motions and obtain the new consensus. The process is repeated until the change in SC-distance is less than a prescribed threshold. This iterative process is well-defined as it is shown in the Methods section that the SC-distance is non-increasing from one iteration to the next. The computation of the optimal translations and rotations and the new consensus is itself an iterative process that both uses the current consensus and generates simultaneously a new one. [fig_ref] Table 1: Algorithm MAPSCI [/fig_ref] summarizes the algorithm in pseudocode form. The various steps in the pseudocode are described in more detail in the Methods section. The algorithm has been implemented in C++ and can be used standalone or run remotely via a web-based interface. The source code of the implementation is available for download from the project's website (see the Availability section). The implementation is organized as a library of algorithms and simple data structures that can be integrated in other projects. Examples of using the library within a C++ program are given in the README file of the source code distribution. The iterative process described above employs pairwise structure alignment as an intermediate step and the parameters that control the execution of the multiple alignment algorithm are the parameters for the underlying pairwise alignment algorithm. The current implementation uses the pairwise alignment algorithm described in [bib_ref] Pairwise protein structure alignment based on an orientation-independent backbone representation, Ye [/bib_ref] ; however, other algorithms for pairwise structure alignment can be used instead.
# Results
## Web server
MAPSCI has been incorporated into a web server for remote access over the Internet (see [fig_ref] Figure 1: is a user-specified threshold [/fig_ref]. This tool allows for protein structures to be uploaded from files on the local disk or retrieved from the Protein Data Bank (PDB) [bib_ref] The Protein Data Bank, Berman [/bib_ref] by specifying their PDB ids. The results from the alignment are annotated in the standard NBRF/PIR format, which can be previewed online via the Jalview applet [bib_ref] Jalview Version 2 -a multiple sequence alignment editor and analysis workbench, Waterhouse [/bib_ref]. Integrated with the server is the molecular viewer applet Chemis 3D, which allows for visualization of the aligned protein structures.
The web server offers a simple interface that allows for remote access from within other software. gives an example of using the programming language Python to retrieve the transformed coordinates (in PDB format) for the multiple alignment of the structures from the HOMSTRAD CUB family. Additional examples and the complete set of options for remote access can be found at the server web page (see the Availability section).
## Comparison
As discussed earlier, there are many algorithms for multiple structure alignment. In general, it is difficult to make comparisons among them, since they operate under different sets of assumptions and problem formulations. We compare MAPSCI to two recent algorithms -MAMMOTH [bib_ref] A new progressive-iterative algorithm for multiple structure alignment, Lupyan [/bib_ref] and MATT [bib_ref] Matt: Local Flexibility Aids Protein Multiple Structure Alignment, Menke [/bib_ref] which also work with coordinate triples, but employ a different objective function. Our experiments show that MAPSCI is competitive in terms of the sizes of the so-called conserved regions and runs significantly faster than the other two algorithms, hence can potentially scale to much larger datasets.
The comparison is based on two benchmark datasets. The first dataset is compiled from the HOMSTRAD database [bib_ref] HOMSTRAD: a database of protein structure alignments for homologous families, Mizuguchi [/bib_ref] , which is a curated database of structurebased alignments for homologous protein families and is considered the "gold" standard. The benchmark dataset consists of the 232 HOMSTRAD families that have at least 4 structures. The second dataset consists of the superfamily set in the SABmark database [bib_ref] SABmark -A benchmark for sequence alignment that covers the entire known fold..., Vanwalle [/bib_ref] (version 1.65). It contains 425 families with low to intermediate sequence similarity. The metrics considered in the comparison are the strict core (or just core) and the core RMSD. This follows the experimental setup in [bib_ref] A new progressive-iterative algorithm for multiple structure alignment, Lupyan [/bib_ref] where strict core is defined as "the set of positions with 100% conservation, and within 4.0 Å of each other in the final structural alignment in 3D". A similar metric is discussed in [bib_ref] Alignment of multiple protein structures based on sequence and structure features, Madhusudhanm [/bib_ref] and [bib_ref] Comparison of performance in successive CASP experiments, Venclovas [/bib_ref]. The results are summarized in [fig_ref] Figure 2: HOMSTRAD dataset comparison [/fig_ref] , which show the pairwise comparisons (MAPSCI, MAMMOTH), (MAPSCI, MATT) in terms of the core size (expressed in percent of the length of the shortest protein) and the core RMSD. [fig_ref] Table 3: Benchmark datasets performance [/fig_ref] provides a comparison of the average core size and average core RMSD for the three methods on the benchmark datasets.
In general, it is difficult to compare two algorithms based on these two metrics (larger cores tend to have larger RMSD). However, on the HOMSTRAD dataset MAPSCI outperformed MAMMOTH in 45% of the test cases and MATT in 59% of the test cases by computing alignments with both larger cores and smaller core RMSD. (MAMMOTH and MATT were better than MAPSCI on both metrics combined in 6% and 5% of the test cases, respectively). MAPSCI computed cores for all 232 test cases, while MAMMOTH failed to compute a core for one family (bowman), and MATT failed to compute a core for three families (asp, lipocalin, and tln).
On the SABmark dataset MAPSCI computed larger cores with better RMSD in 39% of the test cases when compared with MAMMOTH and in 37% of the test cases against against MATT. (MAMMOTH and MATT were better than MAPSCI on the two metrics combined in 15% and 26% of the test cases, respectively.) MATT was the most robust of the three algorithms and failed to compute a core in only five test cases; MAPSCI failed on 40 families and MAMMOTH failed on 31 families.
MAPSCI took only 151 seconds to align the 425 families in the SABmark dataset and 85 seconds to align the families in the HOMSTRAD dataset. MAMMOTH took 1100 seconds on the SABmark dataset and 649 seconds on the HOMSTRAD dataset. By contrast,
[formula] P 0 0 from { } P i i K 1 . i 0. SC 0 ∞. [/formula]
## Do
3. if i = 0 then compute pairwise structure alignment between P i 0 and every P j .
4. else use standard dynamic programming to align P i 0 with every P j . 5. i i + 1.
6. Compute correspondence i from the above alignments (either pairwise or dynamic programming) using center-star-like method. 7. Compute optimal translation matrix T j i and optimal rotation matrix R j i iteratively (Theorems 2 and 3). Transform P j by R j i and T j i for every j to obtain multiple structure alignment ℳ i . SC i SC(ℳ i ).
8. Post-process ℳ i by removing all columns consisting of only gaps.
9. Compute new consensus structure P i 0 from ℳ i by Theorem 1. MATT took several hours to process the two datasets. shows the actual time taken by MAPSCI for all families in the benchmark dataset in terms of the total number of residues per family. The algorithm converges very quickly and can potentially scale to large datasets. The machine used for all experiments reported in the paper runs Ubuntu Linux 8.04 and has 4 GB of RAM with Intel®Core™2 Quad CPU Q9550 @ 2.83 GHz. MAMMOTH and MATT were run with their default parameter settings.
[formula] 10. Until SC i SC i SC i 1 1 .//h [/formula]
# Methods
In this section, we provide the mathematical and algorithmic framework underlying MAPSCI. As mentioned earlier MAPSCI shares common elements with the algorithm in [bib_ref] Approximate multiple protein structure alignment using the Sum-of-Pairs distance, Ye [/bib_ref] , and therefore, we follow the same general outline. However, we only present the full details when there are significant differences and refer the reader to [bib_ref] Approximate multiple protein structure alignment using the Sum-of-Pairs distance, Ye [/bib_ref] when there is an overlap.
## Multiple structure alignment: problem formulation
Let {P 1 , P 2 , ..., P k } be the given set of K proteins and let l i be the number of C a atoms along the backbone of protein P i . We represent P i as a sequence of coordinate triples
[formula] u x y z j i j i j i j i ( , , ), 1 ≤ j ≤ l i , that represent the [/formula]
coordinates of the jth C a atom of P i along the backbone.
(As is customary [bib_ref] Protein Structure Comparison by Alignment of Distance Matrices, Holm [/bib_ref] [bib_ref] Hierarchical protein structure superposition using both secondary structure and atomic representation, Singh [/bib_ref] , we consider only the backbone, not the amino acid residues themselves.) Let P 0 = u 1 0 , ..., u l 0 0 denote the consensus structure, of length l 0 . A correspondence of the K proteins in and the consensus structure P 0 can be represented as a matrix H = (
[formula] h ij ) 0 ≤ i ≤ K,1 ≤ j ≤ L , for some L ≥ max 0 ≤ i ≤ K {l i }, [/formula]
where h ij is either a coordinate triple belonging to the ith protein or a gap. Distances between coordinate triples are based on the squared distance between them in R 3 . The distance between a coordinate triple and a gap is called a gap penalty, and is denoted by r.
The results reported in this paper use 16.0 for the value of the gap penalty.
Let Under the multiple structure alignment we define the distance between the consensus structure P 0 and protein P j as D P P d g g
[formula] G i = (H i -T i )R i = (H i -e × t i )R i , for i >0, where R i R 3 × 3 is some rotation matrix, T i = e × t i is the translation matrix, e R L ×j j L ( , )( , ) 0 02 1 [/formula]
, where d(·, ·) denotes the following distance function: .
[formula] u v [/formula]
The distance between P 0 and P j can be represented compactly as D P P G G j j F ( , ) || || 0 0 2 , where ||·|| F denotes the Frobenius norm, with the additional convention that the squared difference between a coordinate triple and a gap is r 2 . The total distance of the K proteins to the consensus structure, called the Sum-of-Consensus distance, or SC-distance, is then defined as (1) Remote access to the server import urllib2 url = "http://www.geom-comp.umn.edu/mapsci/align.cgi?wsget=pdb&rcsb=1sfp+1spp:A+1spp:B" server = urllib2.urlopen(url) output = file("alignment.zip", 'wb') output.write(server.read()) output.close() server.close() An example of using the programming language Python to retrieve the transformed coordinates (in PDB format) for the multiple alignment of the structures from the HOMSTRAD CUB family. Additional examples and the complete set of options for remote access can be found at the server web page (see the Availability section). Statistics for the performance of the three methods on the benchmark datasets. The subscripts in the Average Core RMSD columns indicate how many values were used in computing the statistics, since the algorithms failed to compute a core for some of the data sets. For the Average Core (%) columns all reported values were used and therefore n = 232 and n = 425 for the HOMSTRAD and SABmark datasets, respectively.
[formula] SC D P P G G j j K j F j K ( , ) || || . [/formula]
Intuitively, the SC-distance measures how well the consensus structure represents the given set of K proteins. A similar distance function is used in [bib_ref] Finding the consensus shape of a protein family, Chew [/bib_ref] , where each protein is represented as a set of vectors in R 4 .
We can now define the multiple structure alignment problem as follows:
## Multiple structure alignment problem
Given a set {P 1 , P 2 , ..., P K } of protein structures, compute a transformation (i.e., rotation and translation) for each protein, and generate a consensus structure P 0 , such that the resulting multiple structure alignment has minimum SC-distance as defined in Equation (1).
In the next section, we present a heuristic for this problem. Our algorithm approximates the global minimum of the SC-distance by iterative refinement of an initial multiple structure alignment and converges to a local minimum.
Step I: Choice of the initial consensus structure
We consider four choices for initial consensus structure: (i) median protein, i.e. the protein of median length; (ii) center protein, i.e. the protein that minimizes the sum of the pairwise distances to all the other proteins; (iii) the minmax protein, i.e. the protein with the smallest maximum pairwise distance; and (iv) maxcore protein, i.e. the protein that generates the largest initial core. (The first three choices for initial consensus are considered in [bib_ref] Approximate multiple protein structure alignment using the Sum-of-Pairs distance, Ye [/bib_ref] The experimental results in [fig_ref] Figure 5: Consensus choice comparison [/fig_ref] indicate that MAPSCI is quite robust in terms of the choice of initial consensus. However, the data suggests that the median protein occasionally leads to alignments with very low core size, and therefore is the least reliable choice. The other three choices seem to work well in practice, although they are more expensive computationally. The results reported in the Comparison section use the maxcore protein as the initial consensus.
Step II: Compute an initial correspondence After we determine the consensus structure P 0 in Step I, the K -1 pairwise structure alignments between P 0 and P i ≠ P 0 , for i = 1, ..., K, are computed using the algorithm in [bib_ref] Pairwise protein structure alignment based on an orientation-independent backbone representation, Ye [/bib_ref]. (Other pairwise structure alignment algorithms could also be used instead.) The K -1 pairwise structure are combined in Line 6 of the algorithm [fig_ref] Table 1: Algorithm MAPSCI [/fig_ref] using the center-star-like method described in [bib_ref] Approximate multiple protein structure alignment using the Sum-of-Pairs distance, Ye [/bib_ref].
Step III: Compute optimal rotation and translation matrices and consensus structure Given a correspondence H = ( h ij ) the objective is to find the rotation and translation matrices R j and T j , for j = 1, ..., K, and the consensus structure J , such that the sum of the pairwise alignment distances between J and each (transformed) P j is minimum; i.e. we wish to minimize
[formula] S J H T R j j j F j K || ( ) || . 2 1(2) [/formula]
Direct minimization of S over J , and the T j 's and R j 's seems difficult. Instead, we propose an iterative procedure for minimizing S. Within each iteration, the minimization of S is carried out in two stages that are interleaved: (1) computation of the optimal J for given R j 's and T j 's, and (2) computation of the optimal R j 's and T j 's for a given J .
## Computation of the optimal consensus structure
First, we show how to compute the consensus structure, given the rotation and translation matrices R j 's and T j 's, as stated in the following theorem: Proof. For each j, we consider two distinct cases for J j : either it is a coordinate triple, x, or a gap. If J j is a gap, then the sum of the distances between J and each protein P j along the jth column is |I n |r 2 , where r is the gap penalty. If J j is a coordinate triple, x, then the sum of the distances between J and each protein P j along the
[formula] jth column is | | || || I h x g i j i I n 2 2 , which is minimized, for x x h j i j i I I n n 1 | | . Therefore, if | | | | || || I I h x n g i j j i I n 2 2 2 [/formula]
, then the optimal choice for J j is the coordinate triple x j ; otherwise, the optimal choice for J j is a gap.
## Computation of the optimal translation matrix
In this section, we show how to compute the optimal translation matrix T i , for each i, for a given consensus structure J . From Eq. (2), it is clear that the optimal T i and T j , for i ≠ j are independent of each other. Hence, in the following, we focus on the computation of T i , for a specific i. The translation matrix T i can be
[formula] decomposed as T i = e × t i , where t i R 1 × 3 is the translation vector. [/formula]
As mentioned earlier, the transformation of a gap remains a gap. Hence the computation of the translation and rotation matrices is independent of the mismatches (i.e., where at least one of the two elements being compared is a gap). We can thus simplify the computation by removing all mismatches in the alignment between the consensus structure J and the ith protein P i .
Let A R n × 3 and B R n × 3 consist of the coordinate triples from the consensus structure and the ith protein, respectively, after removing the mismatches. (Here n is the number of matches between the consensus structure and the ith protein, i.e., comparison of two non-gaps). Without loss of generality, assume e T A = [0, 0, 0], i.e., the coordinate triples in the consensus protein are centered at the origin. The optimal translation vector is the one that matches the centroids of the coordinate triple vectors from A and B as stated in the following theorem:
Theorem 2. Let A and B be defined as above. Assume that e T A = [0, 0, 0]. Then for any rotation matrix R i , the optimal translation vector t i for minimizing More details can be found in [bib_ref] Least-square estimation of transformation parameters between two point patterns, Umeyama [/bib_ref].
[formula] S A B T R A B e t R i i iF i iF || ( ) || || ( ) || [/formula]
## Computation of the optimal rotation matrix
Next, consider the rotation matrix R i . We can assume that the coordinate triple vectors from both A and B are centered at the origin. It follows that
[formula] S A BR A BR A BR A A i i i T i T || || (( ) ( )) ( ) ( 2 2 trace trace trace A A BR B B T i T ) ( ) . trace [/formula]
Hence the minimum of S i is obtained when trace (A T BR i ) is maximized.
Let the Singular Value Decomposition (SVD)of A T B be UΣV T , where U and V are orthogonal and Σ is diagonal.
Theorem 3. The optimal rotation matrix R i that minimizes S i = ||A -BR i || 2 is given by R i = UWV T , where W = diag(1, 1, 1), if det(UV T ) = 1, and W = diag(1, 1, -1), if det(UV T ) = -1.
More details can be found in [bib_ref] Least-square estimation of transformation parameters between two point patterns, Umeyama [/bib_ref].
## Convergence of the algorithm
In this section, we show that MAPSCI converges, by showing that the SC-distance is non-increasing from one iteration to the next. Recall that from Eq. (1), Line 4 in MAPSCI decreases the distance between the consensus structure and each of the K proteins, since the dynamic programming produces an alignment with minimum cost. By the property of the center-star-like method, Line 6 leaves unchanged the distance between the consensus structure and each of the K proteins. By Theorems 2 and 3, the transformations computed in Line 7 do not increase the distance between the consensus structure and the jth protein, for each j. It is clear that Line 8 does not change the pairwise distance, since the cost for aligning two gaps is zero. Finally, by Theorem 1, Line 9 does not increase the sum of the pairwise distances from the consensus structure to the other proteins. Hence, the SC-distance is non-increasing, and the algorithm converges.
[formula] SC D P P H H T R j j K j j j F j K ( , ) || ( ) || . [/formula]
# Complexity analysis
Let n be the maximum length of the K proteins. Then the overall running time of the algorithm is O(K 2 n 2 ). (If we choose the initial consensus structure as the protein of median length, the running time is O(Kn 2 + K 2 n).)
The run time analysis is similar to that of the algorithm in [bib_ref] Approximate multiple protein structure alignment using the Sum-of-Pairs distance, Ye [/bib_ref].
# Conclusions
We have presented an algorithm, called MAPSCI, to compute a multiple structure alignment for a set of proteins, together with their consensus structure. The algorithm represents the input proteins and the consensus as sequences of coordinate triples and computes an approximation to the optimal multiple structure alignment that minimizes the sum of the pairwise distances between the consensus and each input protein. Experimental results on a benchmark datasets derived from the HOMSTRAD and SABmark databases show that the algorithm compares favorably with existing algorithms for multiple structure alignment (MAMMOTH and MATT).
## Availability and requirements
- Project name: MAPSCI - Project home page: http://www.geom-comp. umn.edu/mapsci - Operating system(s): Platform-independent - Programming language: C++ - License: Free BSD
[fig] Figure 1: is a user-specified threshold (currently set at 0.0001) Web server screenshots. Screenshots from the web server: main page (top left), results page (bottom left), structure view (top right), sequence view (bottom right). [/fig]
[fig] 1: is a vector with 1 in each entry, and t i R 1 × 3 is a translation vector. (The transformation of a gap remains a gap.) Note that P 0 remains unchanged, i.e. G 0 = H 0 . [/fig]
[fig] Figure 2: HOMSTRAD dataset comparison. Comparison based on the strict core metric (expressed in percent of the size of the shortest protein) and the strict core RMSD on the HOMSTRAD dataset. [/fig]
[fig] Figure 3, Figure 4: SABmark dataset comparison. Comparison based on the strict core metric (expressed in percent of the size of the shortest protein) and the strict core RMSD on the SABmark dataset. Execution time. The actual execution time of MAPSCI for all families in the benchmark datasets plotted in terms of the total number of residues per family. [/fig]
[fig] Figure 5: Consensus choice comparison. Comparison between the sizes of the aligned cores for different choices of initial consensus protein. [/fig]
[table] Table 3: Benchmark datasets performance [/table]
|
Health Needs and Their Relationship with Life Expectancy in People with and without Intellectual Disabilities in England
# Introduction
Addressing the burden of health inequalities is now a global priority. Strategies to reduce these inequalities tend to focus on the most vulnerable, such as people living with disabilities or in areas of social deprivation. Particularly at risk are those with intellectual disabilities (also known as learning disabilities in the UK) owing to a combination of genetic, social and behavioural factors [bib_ref] Multiple physical and mental health comorbidity in adults with intellectual disabilities: Population-based..., Cooper [/bib_ref]. Whilst there are measures in place to reduce health inequalities in this population, such as annual health checks [bib_ref] The impact of health checks for people with intellectual disabilities: An updated..., Robertson [/bib_ref] , mortality data suggest that the situation has not improved, despite some deaths being potentially avoidable [bib_ref] Mortality disparities and deprivation among people with intellectual disabilities in England: 2000-2019, Tyrer [/bib_ref] [bib_ref] Cause of death and potentially avoidable deaths in Australian adults with intellectual..., Trollor [/bib_ref].
One of the challenges to reducing inequalities among people living with intellectual disabilities is that they are more likely than the general population to have severe health needs, including epilepsy, cerebral palsy and eating/feeding difficulties, which are known to shorten life expectancy [bib_ref] An evaluation of the effectiveness of annual health checks and quality of..., Carey [/bib_ref]. Although not always life-limiting if managed well, they are relatively rare in the general population and so tend not to feature in population-level policy initiatives. Thus far, their individual contribution to life expectancy has not been formally investigated, but it is important to do so because this contribution may be over-inflated or seen as an inevitable consequence of having intellectual disabilities without seeking to improve health outcomes and/or quality of life for the individuals affected.
The aim of the current study was to investigate specific health needs and quantify their contribution to life expectancy in people with intellectual disabilities and to compare these findings with a cohort of individuals without intellectual disabilities. A further aim was to investigate people without any of the specified health needs to determine if loss in life years for people with intellectual disabilities remained.
# Materials and methods
## Data sources
This study followed the Reporting of studies Conducted using Observational Routinelycollected health Data (RECORD) checklist [bib_ref] The REporting of studies Conducted using Observational Routinely-collected health Data (RECORD) statement, Benchimol [/bib_ref] (see [fig_ref] Table 1: Baseline and follow-up characteristics of the study population by intellectual disability and... [/fig_ref]. We used the Clinical Practice Research Datalink (CPRD GOLD), linked (person-level) with hospital episode statistics (HES) and death registrations from the Office for National Statistics (approved study protocol number 19_267). Details of the study population have been described in a previous work [bib_ref] Mortality disparities and deprivation among people with intellectual disabilities in England: 2000-2019, Tyrer [/bib_ref] , with the exception of 23 additional individuals identified, after an amendment to the original protocol, with Cockayne and Angelman syndrome; details of these 23 individuals were received in August 2021 (due to COVID-19 delays; please see the data flow diagram in Supplementaryfor the initial extract and the study population used for the current study). Briefly, the CPRD is an electronic health record primary care research database which is broadly representative of the national population in terms of age, gender and ethnicity [bib_ref] Data resource profile: Clinical Practice Research Datalink (CPRD), Herrett [/bib_ref]. Only GP surgeries in England that consented to their data being linked with hospital episode statistics (HES) and death data (approximately 75% of CPRD surgeries in England) were included in this study.
## Sample population
Initial inclusion criteria for the broader programme of work on which this study was based were: registered at the GP surgery at any point between 1 January 2000 and 29 September 2019 and 10 years old or older to account for delays in reporting of diagnoses of intellectual disability in children [bib_ref] Mortality among a cohort of persons with an intellectual disability in New..., Florio [/bib_ref]. A random sample of 980,586 people without intellectual disabilities (initially 1 million prior to exclusions; see Supplementary Figure S1) was used for the comparison group with the same eligibility criteria (but without a diagnosis of intellectual disability). For this study, data were further restricted to the 2017-2019 observation period such that people entered the study on 1 January 2017 if this was after the original date of cohort entry or were excluded if they were last seen or died before this date. The final population comprised 7794 individuals with intellectual disabilities and 176,807 individuals without intellectual disabilities (n = 440 of whom changed status within the observation window at their first intellectual disability diagnosis).
## Definition of intellectual disabilities and health needs
Diagnostic codes (Read codes and International Classification of Diseases (ICD)-10 codes) for intellectual disabilities and health needs are reported in the Supplementary Material [fig_ref] Table 2: Additional years expected to live at age 10, 20 and 40 years... [/fig_ref]. These were based on a combination of previous literature [bib_ref] Mortality disparities and deprivation among people with intellectual disabilities in England: 2000-2019, Tyrer [/bib_ref] , free text searching of diagnostic code descriptions and clinical opinion (RK, SKG, RM). The initial choice of health needs was based on the literature in this area [bib_ref] Health characteristics and consultation patterns of people with intellectual disability: A cross-sectional..., Carey [/bib_ref] and discussions with carers and people living with intellectual disabilities as being sufficiently severe to affect life expectancy. These were: epilepsy; incontinence (urinary or faecal); severe visual loss; severe hearing impairment; severe mobility difficulties; cerebral palsy and feeding via a percutaneous endoscopic gastrostomy (PEG) tube (i.e., as a measure of severe eating/feeding difficulties). To avoid inclusion of shorter-term health needs that had resolved over time and/or been misdiagnosed in childhood, such as epilepsy [bib_ref] Incidence and Prevalence of Childhood Epilepsy: A Nationwide Cohort Study, Aaberg [/bib_ref] , health needs were defined as being present only if their most recent diagnosis was within 10 years of cohort entry. The exception to this was cerebral palsy, which was defined by a diagnosis ever being present given that it is a life-long condition from birth/early infancy [bib_ref] Survival and mortality in cerebral palsy: Observations to the sixth decade from..., Blair [/bib_ref].
# Statistical methods
The date of entry into the cohort was defined as the latest date according to the person and practice's characteristics: the beginning (i.e., 1 January 2017) of the observation window; the date of registration with the GP practice; the date the practice was defined as being up to standard (using the CPRD's own quality indicators); or the date the individual turned 10 years old (to align with the eligibility criteria). Because there are known delays in reporting intellectual disability diagnoses [bib_ref] Immortal time bias for life-long conditions in retrospective observational studies using electronic..., Tyrer [/bib_ref] and to avoid conditioning on the future, an intellectual disability status was treated as an age-dependent covariate such that people with intellectual disabilities contributed to the comparison cohort prior to their first diagnosis. Health needs were also treated as age dependent, and individuals contributed to both the presence and absence of health need at different ages if they were diagnosed with a new health need during the observation period. The date of exit was defined as: the date of the last CPRD update (29 September 2019); the date of death; the date of the end of the calendar period; the date of the last practice update or the date of transfer out of practice, whichever was first. The cohort was also sub-divided into individuals without any health needs at baseline or follow-up to assess whether life expectancy was similar between people with and without intellectual disabilities (i.e., excess mortality could be explained by the health needs).
The methodology for the life expectancy work used in this study has been described in detail elsewhere. Life expectancy and 95% confidence intervals (CIs) were compared for people with and without intellectual disabilities and by the presence/absence of each health need using flexible parametric models with intellectual disability and health need status treated as age-varying covariates (and an interaction term fitted). Knots were placed according to the event distribution in the intellectual disability group for greater statistical precision. All models used 5 knots (including the boundary knots; 4 degrees of freedom (df); 3df for age-varying effects) with the exception of PEG feeding, which used 4 knots (3df; 2df for age-varying effects) owing to the small sample size. [fig_ref] Table 1: Baseline and follow-up characteristics of the study population by intellectual disability and... [/fig_ref] shows the characteristics of the study population over the observation period. The characteristics of the population with each individual health need are shown in . In comparison to the rest of the population, people with intellectual disabilities were generally younger (median age 33 vs. 43 years) and more were male (57.1% vs. 49.0%). There were also more white individuals (77.0% vs. 67.5%), although this partly reflects more complete recording of ethnicity in hospital settings (only 14.0% vs. [bib_ref] Health characteristics and consultation patterns of people with intellectual disability: A cross-sectional..., Carey [/bib_ref].6% had missing data because more people with LD were hospitalised and had their ethnicity recorded). Most individuals (73.4%) with intellectual disabilities had no cause identified: the most common genetic/chromosomal condition reported was Down syndrome (10.9% of the individuals). People with intellectual disabilities had a substantially higher proportion of all of the health needs under investigation compared to those without intellectual disabilities, as is reflected in the greater proportion without any health needs at baseline and follow-up (53.6% vs. 90.3%; intellectual disability vs. no intellectual disability).
# Results
## Baseline characteristics
The largest differences between people with and without intellectual disabilities were observed for cerebral palsy, which was ≈58 times more prevalent during the 2.7-year observation window (i.e., at baseline or follow-up). Epilepsy, severe visual loss, severe mobility difficulties and PEG feeding were ≈12-22 times more prevalent; and incontinence and severe health impairment were ≈2-4 times more prevalent. The most common severe health need in people with intellectual disabilities was epilepsy, which was present in 18.7% of the individuals at baseline. For people without intellectual disabilities, incontinence was the most common health need, present in 3.8% of the individuals at baseline. Perhaps the most striking finding from the figures is that life expectancy was substantially higher across the board in people with neither intellectual disabilities nor specified health need. At 10 years of age, these individuals could expect to live between 72.2 (absence of severe health impairment) and 74.3 additional years (all of the health needs absent). At the same age, children with intellectual disabilities but without each specified health need lost ≈15-22% of life years compared to this first group, living, on average, an additional 57-62 years. Those with intellectual disabilities but without any of the health needs under investigation needs lost ≈12% of life years, living on average 8.9 years shorter than those with neither intellectual disabilities nor health needs.
## Life expectancy
We can see that the most severe of the health needs, regardless of intellectual disability status, were PEG feeding and severe mobility difficulties. Ten-year-old children with a PEG feeding tube could expect to live only an additional 23.0 years (95% CI 17.1-31.0) if they had intellectual disabilities and 25.7 years (95% CI 18.3-36.0) if they did not have intellectual disabilities, representing a loss in life years of ≈65-68% compared to those with neither condition. Similarly, children with severe mobility difficulties lost ≈41-44% of life years, living an additional 41-43 years only compared to the almost 73 years in those with neither condition. The disadvantages for individuals with PEG feeding tubes and severe mobility difficulties continued to be observed in adulthood [fig_ref] Table 2: Additional years expected to live at age 10, 20 and 40 years... [/fig_ref].
Of the remaining health needs, people with epilepsy had shorter life expectancy overall. Confidence overlapped at 10 years but, subsequently, having intellectual disability in addition to epilepsy incurred additional life expectancy disadvantages (see [fig_ref] Table 2: Additional years expected to live at age 10, 20 and 40 years... [/fig_ref] , with a loss in life years of ≈38%. Severe visual loss or incontinence was about equivalent to having intellectual disabilities without the health need in terms of life expectancy, but people with both intellectual disabilities and incontinence were again further disadvantaged, with a loss in life years of ≈34%. Conversely, we did not find an effect of severe hearing impairment on life expectancy. The effect of cerebral palsy on life expectancy was harder to determine, owing to small numbers, but those with cerebral palsy and intellectual disabilities had the shortest life expectancy compared with the those without cerebral palsy or with cerebral palsy but without intellectual disabilities, with a loss in life years of ≈43%. All of these findings were relatively consistent across the age range [fig_ref] Table 2: Additional years expected to live at age 10, 20 and 40 years... [/fig_ref].
# Discussion
This work deepens our understanding of health inequalities in people with intellectual disabilities. By reaffirming that severe health needs make a significant contribution to the mortality disparities that people with intellectual disabilities are known to experience, our findings also reveal that they only partially explain these. After restricting the study population to those without health needs, life expectancy remained shorter for those with intellectual disabilities, with a loss in life years of 12%. Of those with the specified health needs, life expectancy was generally further shortened if intellectual disability was also present, suggesting combined disadvantages.
# Strengths and limitations
To the best of our knowledge, this is the first time that life expectancy has been explored by health needs in people with and without intellectual disabilities. The utilisation of flexible parametric methods to estimate life expectancy is also a novel component and supports previous methodological findings that borrowing strength from larger covariate samples can be an effective way of increasing statistical precision for small samples. However, we recognise that life expectancy is only a crude measure of health inequalities that does not encapsulate other social determinants of health, such as deprivation, or factors that may contribute towards inequalities, such as access to and quality of healthcare provision. We are also unable to comment on other equally important health indicators, including quality of life and well-being.
As with all electronic health record data of this nature which rely on Read code and ICD diagnoses, we are unable to capture variability in the severity of health needs between people with intellectual disabilities and the general population. A particular concern is incontinence, which is likely to be less severe in the general population if it occurred and was resolved during certain life events, such as post-pregnancy [bib_ref] Cumulative incidence of urinary incontinence and associated factors during pregnancy and after..., Chang [/bib_ref] ; it is noteworthy that almost three-quarters (73.3%) of the people in the general population with incontinence were female, compared with only half (53.2%) in the intellectual disability population . Moreover, both incontinence and PEG feeding may be indicative of additional comorbidities (e.g., frailty or dysphagia) rather than directly causing mortality [bib_ref] Is There an Association between Urinary Incontinence and Mortality? A Retrospective Cohort..., Matta [/bib_ref] [bib_ref] Functional dysphagia therapy and PEG treatment in a clinical geriatric setting, Becker [/bib_ref]. Our findings nonetheless support their relationship (even if indirect) with life expectancy. Another limitation of GP health record data is that they do not provide complete information on the severity of intellectual disabilities; we are also likely to have missed people with mild intellectual disabilities who do not have significant support needs and may also be more vulnerable to abuse, discrimination and high-risk behaviours. We also recognise that many of these health needs co-occur and that they are more likely to do so if the individual also has intellectual disabilities, as is reflected in the larger proportion of individuals with at least one health need (46% vs. 10%) and high prevalence of co-occurring health needs, particularly for individuals with cerebral palsy (53% of individuals with intellectual disabilities also had epilepsy; 56% had severe mobility difficulties) and PEG feeding tubes (55% had cerebral palsy; 65% epilepsy and 70% severe mobility difficulties) . This descriptive study does not look in more depth at the co-occurring health needs, nor does it adjust for additional clinically relevant comorbidities, such as dementia, or other social determinants of health which all contribute to the mortality disadvantages that people with intellectual disabilities experience [bib_ref] Mortality, predictors and causes among people with intellectual disabilities: A systematic narrative..., Tyrer [/bib_ref]. Such issues could be explored further using propensity score methodologies or multiple logistic regression/time-to-event analyses, which are recommended to further develop the work described here.
This study took place before the COVID-19 pandemic, during which people with intellectual disabilities have been adversely affected owing to increased risk of transmission (e.g., through residential homes and community-based support) and increased risk of respiratory deaths [bib_ref] Improved survival in Down syndrome over the last 60 years and the..., Glasson [/bib_ref] [bib_ref] Causes of death in remote symptomatic epilepsy, Day [/bib_ref]. People with severe health needs have also been disproportionately affected by COVID-19 [bib_ref] Hospital-acquired SARS-CoV-2 infection in the UK's first COVID-19 pandemic wave, Read [/bib_ref]. In the current climate, the recommendations made here are, therefore, likely to be more relevant.
## Comparison with existing literature
The prevalence of severe health needs found at baseline in this study is largely similar to that found in previous research carried out in the UK and internationally. The prevalence of epilepsy was 18.7% (vs. 1.1% in the general population), which corresponds with previous population-based studies from England (18.5% vs. 0.7% (matched age/gender/practice population sample) [bib_ref] Health characteristics and consultation patterns of people with intellectual disability: A cross-sectional..., Carey [/bib_ref] and Scotland (18.8% vs. 0.8% [bib_ref] Multiple physical and mental health comorbidity in adults with intellectual disabilities: Population-based..., Cooper [/bib_ref]. The prevalence of incontinence in the intellectual disability population (13.3%) was lower than previous UK estimates using the CPRD (20.5% [bib_ref] Health characteristics and consultation patterns of people with intellectual disability: A cross-sectional..., Carey [/bib_ref] , which we attribute to the exclusion of 'H/O incontinence' and incontinence diagnoses within 10 years of cohort entry for the current study. The baseline prevalence of 3.8% found in the general population is at the lower end of the estimates of international figures of 3-18% for severe incontinence (urinary only) in adult women (about half of this for men) [bib_ref] Epidemiology and natural history of urinary incontinence, Hunskaar [/bib_ref] , given that many do not seek support from a healthcare provider [bib_ref] Bridging the gap: Determinants of undiagnosed or untreated urinary incontinence in women, Duralde [/bib_ref]. The prevalence of PEG feeding (1.9% vs. 0.1%) also falls within the 5-year incidence rate (1.3%) of PEG procedures in England based on 17,000 per year [bib_ref] Percutaneous endoscopic gastrostomy: A prospective audit of the impact of guidelines in..., Sanders [/bib_ref].
In our study, prevalence of severe visual impairment among people with intellectual disabilities (13.0%) was lower than previous estimates in the Netherlands for visual impairment and blindness (13.8% and 5.0%, respectively) but, of the latter population, 40.6% were undiagnosed prior to study commencement [bib_ref] Prevalence of visual impairment in adults with intellectual disabilities in the Netherlands:..., Van Splunder [/bib_ref]. We did not find a relationship between severe hearing impairment and life expectancy, which differs from previous (albeit not statistically significant) work in the general population [bib_ref] Association of Hearing Impairment and Mortality in the National Health and Nutrition..., Contrera [/bib_ref].
The prevalence of cerebral palsy we reported here (8.4% vs. 0.1%) can be interpreted using information from the random general population sample. Given that this sample was drawn from 6.2 million individuals (Supplementaryand that 0.5% of the general population sample had intellectual disabilities, and assuming a representative random sample draw, the prevalence of cerebral palsy in our study was approximately 1.42 per 1000 population, which is similar to birth estimates reported of 2.11 per 1000 population [bib_ref] An update on the prevalence of cerebral palsy: A systematic review and..., Oskoui [/bib_ref] conditional on surviving to 10 years. We would also expect there to be 2276 (i.e., 6.2 times as many) people with cerebral palsy in the entire population sample, which would equate to 29% of people with cerebral palsy having intellectual disabilities. This is within the range reported in the literature, which cites 22-40% of individuals with cerebral palsy having cognitive impairment (IQ < 70) [bib_ref] The epidemiology of cerebral palsy: Incidence, impairments and risk factors, Odding [/bib_ref] , although some studies report figures closer to onehalf [bib_ref] Intellectual disability in cerebral palsy: A population-based retrospective study, Reid [/bib_ref]. It is worth emphasising, however, that many-if not most-people with cerebral palsy do not have intellectual disabilities and it is important to make sure that their needs are adequately met. The prevalence of severe mobility difficulties (10.5%) in people with intellectual disabilities is similar to previous estimates of 9.2% for being 'non-mobile' [bib_ref] Diabetes in adults with intellectual disability: Prevalence and associated demographic, lifestyle, independence..., Tyrer [/bib_ref] , but it is hard to determine the prevalence in the general population owing to the recognised variation in thresholds for reporting mobility disabilities [bib_ref] Variation in Thresholds for Reporting Mobility Disability Between National Population Subgroups and..., Melzer [/bib_ref].
## Recommendations
Given that this descriptive study did not seek to control for other contributing factors, we are nonetheless able to make some broad recommendations. First, it is clear from our findings that health needs are a significant problem for people with intellectual disabilities and that, with the exception of severe hearing impairments, they play a key role in shortening life expectancy. More effective management and treatment of these health needs, including regular assessment of associated care requirements, have the potential to improve outcomes and quality of life for those affected. Many of these health needs are relatively rare in the general population, so the development of tailored care pathways for people with intellectual disabilities, based on national guidelines and policies where available, is likely to be a priority. Such pathways may include monitoring medication (e.g., epilepsy-with a focus on epilepsy syndromes and tuberous sclerosis), provision of specialist support (e.g., visual impairment and hearing impairment), communication plans (e.g., cerebral palsy), pain management (e.g., severe mobility difficulties), prevention strategies (e.g., incontinence) and oral care (e.g., PEG feeding). All pathways should include mechanisms for the provision of coordinated care between health, social care and voluntary services so that unnecessary burden is not placed on carers. They should also be adequately flexible to allow for individuals' differences and needs.
# Conclusions
We conclude that differential life expectancy in people living with intellectual disabilities compared to the general population is not wholly attributable to increased prevalence of severe health needs. Our findings highlight the need to continue to find ways to improve health outcomes and quality of life for people living with intellectual disabilities so that they can be supported to lead long and fulfilling lives.
Supplementary Materials: The following supplementary information can be downloaded at https: //www.mdpi.com/article/10.3390/ijerph19116602/s1, Table S1: RECORD checklist; [fig_ref] Table 2: Additional years expected to live at age 10, 20 and 40 years... [/fig_ref] : Diagnostic and classification codes for intellectual disabilities, ethnicity and severe health needs investigated;: Data flow diagram of individuals included in the study population from original extracted data; : Characteristics of individuals with specific health needs under investigation. Funding: This research was funded from a Baily Thomas Doctoral Fellowship award (TRUST/VC/AC/ SG/5366-8393). The funders had no role in study design, data collection and analysis, the decision to publish or the preparation of the manuscript. The study is based in part on data from the CPRD GOLD database obtained under licence from the UK Medicines and Healthcare products Regulatory Agency. The provision of CPRD and linked data was through Leicester Real World Evidence (LRWE) Unit, which is funded by University of Leicester, National Institute for Health Research (NIHR) Applied Research Collaboration (ARC) East Midlands and Leicester NIHR Biomedical Research Centre. The interpretation and conclusions contained in this article are those of the authors alone and not necessarily those of the LRWE Unit, the NHS, the NIHR or the Department of Health and Social Care.
Institutional Review Board Statement: The CPRD has Health Research Authority (HRA) approval for all studies using anonymised data for observational research (East Midlands Research Ethics Committee No. 05/MRE04/87). Research using the CPRD is also subject to regulatory approval from the UK Medicines and Healthcare products Regulatory Agency (MHRA) Independent Scientific Advisory Committee (ISAC); the approved protocol reference no. for this study is: 19/267RA3.
# Informed consent statement:
The CPRD has Health Research Authority (HRA) approval for all studies using anonymised data for observational research (East Midlands Research Ethics Committee No. 05/MRE04/87). Consent was not obtained for the use of anonymised data-approval for the use of the CPRD (data controller) was through the UK MHRA Independent Scientific Advisory Committee (ISAC) (approved protocol no. 19/267RA3).
# Data availability statement:
Data for this study were obtained from the Clinical Practice Research Datalink (CPRD), provided by the UK MRHA. The authors' licence for using these data does not allow sharing of raw data with third parties. Information about access to CPRD data is available here: https://www.cprd.com/research-applications (accessed on 20 May 2022). Researchers should contact the ISAC Secretariat at [email protected] for further details.
[fig] Figure 1a -: g shows the life expectancy estimates and percentage of life years lost (compared with the general population without health needs), for the severe health needs under investigation, by presence/absence of the health need and intellectual disability status. The final figure (Figure 1h) shows the life expectancy estimates for people without any of the health needs under investigation.Table 2also presents the exact life expectancy estimates (with 95% CI) at 10, 20 and 40 years old. [/fig]
[fig] Figure 1: (a-h) Life expectancy from 10 years of age by presence/absence of intellectual disability and health needs. [/fig]
[fig] Author: Contributions: Conceptualisation, F.T. and M.J.R.; data curation, F.T. and M.J.R.; methodology, F.T., M.J.R., R.M., R.K. and S.K.G.; validation, M.J.R.; formal analysis, F.T.; investigation, F.T., M.J.R., R.M., R.K., H.K. and S.K.G.; resources, F.T. and M.J.R.; writing-original draft preparation, F.T.; writing-review and editing, F.T., M.J.R., R.M., R.K., H.K. and S.K.G.; visualisation, F.T. and M.J.R.; supervision, M.J.R. and R.M.; project administration, F.T. and M.J.R.; funding acquisition, F.T. and M.J.R. All authors have read and agreed to the published version of the manuscript. [/fig]
[table] Table 1: Baseline and follow-up characteristics of the study population by intellectual disability and health need status. 1 n = 440 individuals moved from no intellectual disability to intellectual disability sample at first diagnosis during observation window. 2 n = 831 (10.7%) with phenylketonuria (not defined as a specific syndrome for this study). 3 PEG: percutaneous endoscopic gastrostomy. [/table]
[table] Table 2: Additional years expected to live at age 10, 20 and 40 years by individual health need status and absence of health needs. [/table]
|
The Impacts of Prenatal Mental Health Issues on Birth Outcomes during the COVID-19 Pandemic: A Scoping Review
# Introduction
Adverse birth outcomes, including low birth weight, preterm birth and small for gestational age, are the leading causes of infant morbidity and mortality [bib_ref] Antenatal Depressive Symptoms and Adverse Birth Outcomes in Hanoi, Ngo [/bib_ref]. Worldwide, 75% of neonatal mortality and morbidity is due to adverse birth outcomes [bib_ref] Determinants of Adverse Birth Outcome in Sub-Saharan Africa: Analysis of Recent Demographic..., Tamirat [/bib_ref]. Prenatal depression and anxiety increase the risk of adverse birth outcomes, such as preterm birth, low birth weight and small for gestational age [bib_ref] Elevated Depression and Anxiety Symptoms among Pregnant Individuals during the COVID-19 Pandemic, Lebel [/bib_ref]. According to Ding et al., women who have experienced maternal anxiety are 1.50 times more at risk of preterm birth and 1.76 times more likely to have low birth weight infants [bib_ref] Maternal Anxiety during Pregnancy and Adverse Birth Outcomes: A Systematic Review and..., Ding [/bib_ref].March 2020, the World Health Organization (WHO) declared the coronavirus disease 2019 (COVID-19) outbreak as a public health emergency of international concern. As of 11 January 2021, the WHO reported that the cumulative infection cases worldwide have reached nearly 309 million people, with a total of more than 5 million deaths. Studies have shown that infectious disease outbreaks raise the risk of developing depression and anxiety [bib_ref] Elevated Depression and Anxiety Symptoms among Pregnant Individuals during the COVID-19 Pandemic, Lebel [/bib_ref]. A survey conducted in Belgium among 5866 participants reported that the prevalence of depressive symptoms among pregnant women during the COVID-19
Int. J. Environ. Res. Public Health 2022, [bib_ref] Lower Birth Weight of Dutch Neonates Who Were in Utero at the..., Smits [/bib_ref] , 7670 2 of 10 crisis was 25.3%, and this number had increased dramatically from 8.0% to 11.1% before the pandemic [bib_ref] Mental Health Status of Pregnant and Breastfeeding Women during the COVID-19 Pandemic-A..., Ceulemans [/bib_ref] [bib_ref] Investigating the Influence of Maternal Cortisol and Emotional State during Pregnancy on..., Hompes [/bib_ref]. Such stressful events and conditions may also have negative impacts on birth outcomes, including preterm birth and low birth weight, and this could be due to stress-related physiological pathways during disasters [bib_ref] Disaster-Related Prenatal Maternal Stress Influences Birth Outcomes: Project Ice Storm, Dancause [/bib_ref].
The severity of the COVID-19 pandemic is likely to be an added burden to prenatal mental health problems and, thus, increase the risk of adverse birth outcomes. In order to address potential burdens, it is important to explore the potential impacts of prenatal mental health issues on birth outcomes during the COVID-19 pandemic. However, since the declaration of the COVID-19 pandemic, there have been limited reviews focused on prenatal mental health and its impact on adverse birth outcomes during the COVID-19 pandemic.
# Materials and methods
## Research aim
Purpose: This scoping review was developed to map the current literature linking prenatal mental health issues and adverse birth outcomes during the COVID-19 pandemic and to inform the relevant research and policy development.
Guiding questions: What are the impacts of prenatal mental health issues on birth outcomes during the COVID-19 pandemic?
Objective: To assess the published literature related to the impacts of prenatal mental health issues, which include symptoms of stress, anxiety and depression during pregnancy, on the adverse birth outcomes of preterm birth, low birth weight, small for gestational age, birth defects and macrosomia during the COVID-19 pandemic.
# Methods
## Search strategy
An expert health sciences librarian (SC) from the University of Alberta developed our search strategy and executed the final search. Seven databases were searched, including PROSPERO, Cochrane Library (CDSR and Central Register of Controlled Trials), OVID Medline, OVID EMBASE, OVID PsycInfo, EBSCO CINAHL and SCOPUS. The search was conducted using controlled vocabulary (e.g., MeSH, Emtree, etc.) and keywords representing the concepts "COVID19" and "mental health" and "birth outcomes". The search included variations of search filters from the John W. Scott Health Sciences Library Search Filters. All searches were conducted in July 2021 and adjusted appropriately for different databases. The scope was limited to articles from late 2019 to the present to capture the COVID-19 pandemic period. No other limits were applied. Our search identified a total of 894 records. All records were exported to the COVIDENCE systematic review program. We removed 252 duplicates, leaving 642 unique records. The detailed search strategies and results from each database are available in the Supplementary Materials.
## Screening
Only peer-reviewed and published articles were included in this review. All titles and abstracts were screened for relevance by two independent reviewers (TZ and HZ). Articles that potentially discussed COVID-19 in relation to mental health, pregnant women who gave birth during the COVID-19 pandemic and birth outcomes were kept for further full-text screening. The title and abstract screening process identified 636 irrelevant records and left 6 records that met all inclusion criteria for full-text reviews.
The full-text screening was conducted by two independent reviewers (TZ and HZ). Articles that clearly discussed the targeted content were included for data extraction. Four studies were excluded at this stage, as they did not focus on our outcomes of interest, leaving two records for data extraction. Snowball searching of the final two eligible studies was also conducted, but no additional eligible studies were discovered. [fig_ref] Figure 1: PRISMA flow diagram. [/fig_ref] presents the PRISMA flow diagram to show our search strategy and results. [fig_ref] Table 1: Study characteristics of the included studies [/fig_ref] presents the PRISMA-S checklist of this scoping review. studies was also conducted, but no additional eligible studies were discovered. [fig_ref] Figure 1: PRISMA flow diagram. [/fig_ref] presents the PRISMA flow diagram to show our search strategy and results. [fig_ref] Table 1: Study characteristics of the included studies [/fig_ref] presents the PRISMA-S checklist of this scoping review.
## Data extraction
Data extraction was performed following the Joanna Briggs Institute (JBI) recommendations to capture the data. In the data extraction spreadsheet, the information extracted included authors, year of publication, country of origin, study aims, study design, population and sample size, methods, outcomes and key findings that related to the scoping review questions.
# Results
We identified two relevant articles (see [fig_ref] Table 1: Study characteristics of the included studies [/fig_ref]. Preis et al. conducted a study in 2020 using a prospective cohort design [bib_ref] Adverse Perinatal Outcomes Predicted by Prenatal Maternal Stress Among U.S. Women at..., Preis [/bib_ref]. They focused on pregnant women across the USA aged 18 years and above. Wdowiak et al. conducted a study in Poland in 2020, where longitudinal data was used, with the target population being pregnant women who received care from medical facilities in the city of Lublin [bib_ref] Effect of Excessive Body Weight and Emotional Disorders on the Course of..., Wdowiak [/bib_ref]. Both articles found that, during the COVID-19 pandemic, there was an association between prenatal mental health disorders and adverse birth outcomes, but they approached the research from different perspectives. Whether or not such associations differed from the pre-pandemic period was not mentioned. In addition, macrosomia and birth defects were not found in this review.
## Data extraction
Data extraction was performed following the Joanna Briggs Institute (JBI) recommendations to capture the data. In the data extraction spreadsheet, the information extracted included authors, year of publication, country of origin, study aims, study design, population and sample size, methods, outcomes and key findings that related to the scoping review questions.
# Results
We identified two relevant articles (see [fig_ref] Table 1: Study characteristics of the included studies [/fig_ref]. Preis et al. conducted a study in 2020 using a prospective cohort design [bib_ref] Adverse Perinatal Outcomes Predicted by Prenatal Maternal Stress Among U.S. Women at..., Preis [/bib_ref]. They focused on pregnant women across the USA aged 18 years and above. Wdowiak et al. conducted a study in Poland in 2020, where longitudinal data was used, with the target population being pregnant women who received care from medical facilities in the city of Lublin [bib_ref] Effect of Excessive Body Weight and Emotional Disorders on the Course of..., Wdowiak [/bib_ref]. Both articles found that, during the COVID-19 pandemic, there was an association between prenatal mental health disorders and adverse birth outcomes, but they approached the research from different perspectives. Whether or not such associations differed from the pre-pandemic period was not mentioned. In addition, macrosomia and birth defects were not found in this review.
Preis et al. investigated whether experiencing stress during pregnancy at the beginning of the pandemic was associated with a higher prevalence of adverse prenatal outcomes (see [fig_ref] Table 1: Study characteristics of the included studies [/fig_ref] [bib_ref] Adverse Perinatal Outcomes Predicted by Prenatal Maternal Stress Among U.S. Women at..., Preis [/bib_ref]. This study showed that preterm birth was predicted by elevated prenatal maternal stress [bib_ref] Adverse Perinatal Outcomes Predicted by Prenatal Maternal Stress Among U.S. Women at..., Preis [/bib_ref]. In this study, prenatal maternal stress was measured using the Revised Prenatal Distress Questionnaire [bib_ref] Adverse Perinatal Outcomes Predicted by Prenatal Maternal Stress Among U.S. Women at..., Preis [/bib_ref]. The risk of preterm birth increased by 40% with experiences of prenatal maternal stress [bib_ref] Adverse Perinatal Outcomes Predicted by Prenatal Maternal Stress Among U.S. Women at..., Preis [/bib_ref]. Overall, 1% of the variance in preterm birth and 3% in small for gestational age were attributed to elevated stress over and above other relevant maternal sociodemographic and medical characteristics [bib_ref] Adverse Perinatal Outcomes Predicted by Prenatal Maternal Stress Among U.S. Women at..., Preis [/bib_ref]. Preparedness stress was found to be another factor that affected birth outcomes [bib_ref] Adverse Perinatal Outcomes Predicted by Prenatal Maternal Stress Among U.S. Women at..., Preis [/bib_ref]. Preparedness stress refers to stressful feelings of not being prepared for birth or postpartum, as a result of the pandemic. It was measured using the Pandemic-Related Pregnancy Stress Scale [bib_ref] Adverse Perinatal Outcomes Predicted by Prenatal Maternal Stress Among U.S. Women at..., Preis [/bib_ref]. This study shows that the risk of delivering a small gestational age infant was 66% higher among women who reported higher preparedness stress [bib_ref] Adverse Perinatal Outcomes Predicted by Prenatal Maternal Stress Among U.S. Women at..., Preis [/bib_ref].
## Longitudinal study
The severity of depression was significantly higher during than before the pandemic. Women who experienced the COVID-19 pandemic on average had 0.87 week less on gestational age at delivery (p-value: 0.001) and their infants on average were 184.72 g lighter (p-value: 0.01) compared to those before the pandemic. Pre-pregnancy BMI and severity of depression in week 32 are associated with lower birth weight.
The higher the pre-pregnancy BMI and the severity of depression, the lower the gestational age at delivery;
During the pandemic, pre-pregnancy BMI was positively correlated with level of depression in weeks 10-13 (r = 0.526, p < 0.001) and week 32 (r = 0.539, p < 0.001).
The odds of lower APGAR (5 or 6) were significantly higher during COVID-19 pandemic than before it (OR:3.18. p-value: 0.009).
Women with higher pre-pregnancy BMI and higher severity of depression during pregnancy have higher odds of lower APGAR.
Wdowiak et al. focused on evaluating the effect of the COVID-19 pandemic on depression and whether the course of pregnancy and infant well-being would be affected by depression and the COVID-19 pandemic (see [fig_ref] Table 1: Study characteristics of the included studies [/fig_ref] [bib_ref] Effect of Excessive Body Weight and Emotional Disorders on the Course of..., Wdowiak [/bib_ref]. The study used the Beck Depression Inventory (the second edition) twice during the pregnancy (first between weeks 11 and 14, and then on week 32) among the same population to assess the depressive symptoms [bib_ref] Effect of Excessive Body Weight and Emotional Disorders on the Course of..., Wdowiak [/bib_ref]. The study showed that the severity of depression during weeks 10-13 (i.e., first trimester) and week 32 (i.e., third trimester) were significantly different from the depression severity in the second trimester, which also correlated with the COVID-19 pandemic [bib_ref] Effect of Excessive Body Weight and Emotional Disorders on the Course of..., Wdowiak [/bib_ref]. Additionally, the severity of depression was also dependent on the interaction between the pre-pregnancy Body Mass Index (BMI) and the COVID-19 pandemic [bib_ref] Effect of Excessive Body Weight and Emotional Disorders on the Course of..., Wdowiak [/bib_ref]. Moreover, the depression severity was negatively associated with gestational age and birth weight [bib_ref] Effect of Excessive Body Weight and Emotional Disorders on the Course of..., Wdowiak [/bib_ref]. More severe depression and higher pre-pregnancy BMI were associated with lower gestational age at birth and lower birth weight [bib_ref] Effect of Excessive Body Weight and Emotional Disorders on the Course of..., Wdowiak [/bib_ref]. Wdowiak et al. reported that the average birth weight and gestational age at birth were lower during the COVID-19 pandemic compared to the pre-pandemic measures: women who experienced the COVID-19 pandemic, on average, had 0.87 weeks less gestational age at delivery (p = 0.001), and their infants, on average, were 184.7 g lighter (p = 0.01), compared to infants born before the pandemic [bib_ref] Effect of Excessive Body Weight and Emotional Disorders on the Course of..., Wdowiak [/bib_ref].
# Discussion
This scoping review assessed the state of the literature on prenatal mental health disorders' impacts on adverse birth outcomes during the COVID-19 pandemic. After reviewing the identified articles, we found that very few studies explored the association between prenatal mental health issues and adverse birth outcomes during the COVID-19 pandemic. This limited literature may be due to timing. The pandemic began early in 2020, but for prenatal mental health issues to develop and to study the impacts on adverse birth outcomes require a relatively long follow-up period. Additionally, in order to understand the unique impacts of the pandemic, longitudinal data are often needed for comparison with trends in the previous years. As the onset of the pandemic occurred relatively quickly, there might be little time to collect the longitudinal data; thus, longitudinal studies might not have been a feasible choice to assess the association.
Although limited, the published literature has demonstrated that prenatal mental health issues may have a negative impact on birth outcomes during the COVID-19 pandemic. Both articles highlighted that pregnant women experienced prenatal mental health issues, including stress, depressive symptoms and anxiety during the COVID-19 pandemic. Moreover, both articles reported that experiencing such mental health issues increased the risk of developing adverse birth outcomes, including preterm birth, low birth weight and small for gestational age [bib_ref] Adverse Perinatal Outcomes Predicted by Prenatal Maternal Stress Among U.S. Women at..., Preis [/bib_ref] [bib_ref] Effect of Excessive Body Weight and Emotional Disorders on the Course of..., Wdowiak [/bib_ref].
## Prenatal mental health issues during the covid-19 pandemic
Pregnant women are likely to experience heightened mental health issues during any stressful events, including the COVID-19 pandemic. As the two studies reported, the average levels of prenatal maternal stress and depression have been higher during the pandemic compared to the levels before the pandemic [bib_ref] Adverse Perinatal Outcomes Predicted by Prenatal Maternal Stress Among U.S. Women at..., Preis [/bib_ref] [bib_ref] Effect of Excessive Body Weight and Emotional Disorders on the Course of..., Wdowiak [/bib_ref]. Similarly, a study conducted in Poland reported that the pandemic and related restrictions exaggerated the prevalence of prenatal anxiety from 15% pre-COVID-19 pandemic to 38% during the pandemic [bib_ref] COVID-19 Pandemic-Related Anxiety in Pregnant Women, Nowacka [/bib_ref]. There are three potential reasons for the exacerbation of prenatal stressors and mental health issues during the pandemic. Firstly, increasing worries and stress could be related to uncertainty regarding the effects of COVID-19 on maternal-fetal health and pregnancy. Women may feel uncertain about the effect of the virus on themselves, their fetuses and their infants [bib_ref] Adverse Perinatal Outcomes Predicted by Prenatal Maternal Stress Among U.S. Women at..., Preis [/bib_ref]. Secondly, higher stress was also likely due to pandemic restrictions. Preis et al. stated that 58.0% of the participants reported canceled or altered appointments due to the pandemic, which compounded uncertainties around fetal-infant health and delivery [bib_ref] Adverse Perinatal Outcomes Predicted by Prenatal Maternal Stress Among U.S. Women at..., Preis [/bib_ref]. Wdowiak et al. also hypothesized that physical distancing guidelines and social isolation would negatively impact mothers' mental health, elevating the depressive symptoms, anxiety, insomnia and stress [bib_ref] Effect of Excessive Body Weight and Emotional Disorders on the Course of..., Wdowiak [/bib_ref]. Furthermore, Preis et al. also reported that 40.5% of their participants experienced a pandemic-related income loss and suggested that pandemic-related financial strain was another factor that negatively affected mothers' health [bib_ref] Adverse Perinatal Outcomes Predicted by Prenatal Maternal Stress Among U.S. Women at..., Preis [/bib_ref]. The above results are consistent with other similar studies. For instance, Thayer and Gildner found that 43% of their participants reported financial stress due to the pandemic, which was significantly associated with the elevated depression scores among the pregnant women in their study [bib_ref] COVID-19-Related Financial Stress Associated with Higher Likelihood of Depression among Pregnant Women..., Thayer [/bib_ref].
## Prenatal mental health issues and adverse birth outcomes
Both articles reported significant associations between prenatal mental health issues and adverse birth outcomes during the COVID-19 crisis. Preis et al. reported that pandemicrelated stress was significantly associated with higher risks of preterm birth and the delivery of small for gestational age infants and also suggested that being infected with COVID-19 was a strong independent predictor of SGA infants, as infected mothers are more likely to experience higher stress [bib_ref] Adverse Perinatal Outcomes Predicted by Prenatal Maternal Stress Among U.S. Women at..., Preis [/bib_ref]. Furthermore, Wdowiak et al. observed that the intensification of prenatal depressive symptoms was positively associated with low-birth-weight infants during the pandemic [bib_ref] Effect of Excessive Body Weight and Emotional Disorders on the Course of..., Wdowiak [/bib_ref]. Such findings are consistent with previous studies, which have suggested that maternal stress exposure is related to preterm birth, as explained by the stress-related hypothesis [bib_ref] Lower Birth Weight of Dutch Neonates Who Were in Utero at the..., Smits [/bib_ref] [bib_ref] The Effects of the World Trade Center Event on Birth Outcomes among..., Lederman [/bib_ref]. However, in Preis's study, they also found that the overall rate of preterm birth was lower than it was among the U.S. population in 2019 (7.1% and 10.2%, respectively) [bib_ref] Adverse Perinatal Outcomes Predicted by Prenatal Maternal Stress Among U.S. Women at..., Preis [/bib_ref]. Preis et al. suggested it could be due to the exclusion of the youths who are under 18 years old in their sample, since younger age is one of the risk factors for preterm birth [bib_ref] Adverse Perinatal Outcomes Predicted by Prenatal Maternal Stress Among U.S. Women at..., Preis [/bib_ref]. Similarly, there was some evidence also suggesting a reduction in adverse birth outcomes during the COVID-19 pandemic. For instance, the overall rate of preterm birth and SGA were lower during the COVID-19 pandemic compared to similar time periods in 2017-2019 in Botswana [bib_ref] Modest Reduction in Adverse Birth Outcomes Following the COVID-19 Lockdown, Caniglia [/bib_ref]. In Ireland, the rate of very low-birth-weight infants also reduced from 8.2 to 2.2 per 1000 live births during the COVID-19 lockdown period, compared to the periods prior to the pandemic [bib_ref] Unprecedented Reduction in Births of Very Low Birthweight (VLBW) and Extremely Low..., Philip [/bib_ref]. This could also be explained by stress-related pathways. Philip et al. suggested that such reductions could contribute to "unparalleled and widespread socioenvironmental alterations to which pregnant women would have responded with appropriate behavioral and lifestyle modifications" [bib_ref] Unprecedented Reduction in Births of Very Low Birthweight (VLBW) and Extremely Low..., Philip [/bib_ref]. For instance, the lockdown restrictions likely resulted in mothers resting at home and not working, in some instances. Women may also have received more partner and family support during their pregnancy compared to the pre-pandemic periods. These may have reduced their stress anxiety and depression levels of being pregnant, especially during the pandemic, and likely resulted in better infant health [bib_ref] Unprecedented Reduction in Births of Very Low Birthweight (VLBW) and Extremely Low..., Philip [/bib_ref]. On the other hand, pandemicrelated stress and anxiety could increase the risk of adverse birth outcomes from increased financial, family or other strains [bib_ref] Adverse Perinatal Outcomes Predicted by Prenatal Maternal Stress Among U.S. Women at..., Preis [/bib_ref]. Therefore, the relationship between prenatal mental health with adverse birth outcomes during the pandemic may be a little more nuanced. Despite the association between prenatal mental health issues and adverse birth outcomes, there is a need to further investigate other potential mediators and impact factors beyond mental health issues. Furthermore, the differences in the associations between pandemic and non-pandemic periods were not mentioned in either of the included articles.
# Limitations
This review has potential limitations. The two reviewed studies are from high-income countries. The results might not be generalizable to low-and middle-income countries or other population groups. Although Preis et al. suggested that African-American women and other women of color are at higher risk of experiencing higher pandemic-related stress and adverse birth outcomes, it remains unknown if the same trend applies to mothers from other historically disadvantaged groups during the COVID-19 pandemic [bib_ref] Adverse Perinatal Outcomes Predicted by Prenatal Maternal Stress Among U.S. Women at..., Preis [/bib_ref]. Furthermore, this review did not observe any changes in the associations between prenatal mental health measures and adverse birth outcomes during the pandemic period compared to pre-pandemic periods. In addition, this review included a study with a relatively small sample size (i.e., 280 participants), so the results might not be generalizable to larger populations.
# Conclusions
This scoping review highlighted that the existing knowledge on the impact of prenatal mental health issues on birth outcomes during the COVID-19 pandemic is not yet wellestablished. Given these early results, which show the negative effects of prenatal mental health issues on birth outcomes during the pandemic, there is an urgent need for continued research exploring and addressing prenatal mental health issues and related adverse birth outcomes during the COVID-19 pandemic.
Supplementary Materials: The following supporting information can be downloaded at https: //www.mdpi.com/article/10.3390/ijerph19137670/s1: [fig_ref] Table 1: Study characteristics of the included studies [/fig_ref] : Detailed search string. Institutional Review Board Statement: Only previously published data were discussed in this study. Ethical approval was not applicable to this article.
## Informed consent statement: not applicable.
# Data availability statement:
No new data were created or analyzed in this study. Data sharing was not applicable to this article.
## Conflicts of interest:
The authors declare no conflict of interest. Describe any online or print source purposefully searched or browsed (e.g., tables of contents, print conference proceedings, web sites), and how this was done.
# Appendix a
N/A Citation searching 5
Indicate whether cited references or citing references were examined, and describe any methods used for locating cited/citing references (e.g., browsing reference lists, using a citation index, setting up email alerts for references citing included studies).
## N/a
Contacts 6 Indicate whether additional studies or data were sought by contacting authors, experts, manufacturers, or others. N/A [fig_ref] Table 1: Study characteristics of the included studies [/fig_ref]. PRISMA-S checklist.
## Section/topic # checklist item location(s) reported
Other methods 7 Describe any additional information sources or search methods used. Methods
## Search strategies
Full search strategies 8 Include the search strategies for each database and information source, copied and pasted exactly as run.
Supplementary Files Limits and restrictions 9
Specify that no limits were used, or describe any limits or restrictions applied to a search (e.g., date or time period, language, study design) and provide justification for their use.
# Methods
## Search filters 10
Indicate whether published search filters were used (as originally designed or modified), and if so, cite the filter(s) used.
# Methods
## Prior work 11
Indicate when search strategies from other literature reviews were adapted or reused for a substantive part or all of the search, citing the previous review(s). Describe the processes and any software used to deduplicate records from multiple database searches and other information sources.
## N/a
# Methods
[fig] Figure 1: PRISMA flow diagram. [/fig]
[fig] Author: Contributions: T.Z., screening, data management, data analysis and manuscript writing; H.Z., screening and manuscript editing; S.M.C., searching, supervision and manuscript editing; G.S.J., supervision and manuscript editing; K.S.D., supervision and manuscript editing; J.Y.L., project administration and manuscript editing; S.S.P., supervision and manuscript editing; F.T., supervision and manuscript editing; B.Z., supervision and manuscript editing and S.S.Y., supervision and manuscript editing. All authors have read and agreed to the published version of the manuscript. [/fig]
[fig] Funding: This research was funded by Canadian Institutes of Health Research, grant number RN419823-439880, Alberta Innovates and National Natural Sciences Foundation of China, grant number 81761128034. [/fig]
[table] Table 1: Study characteristics of the included studies. [/table]
[table] Table A1: PRISMA-S checklist.If databases were searched simultaneously on a single platform, state the name of the platform, listing all of the databases searched. [/table]
|
One case of wrist cyst rupture with nerve and blood vessel compression
Introduction and importance: Tendon sheath cysts are mostly located around the joint capsule and tendon sheath, which often occur in the wrist, ankle, and wrist back(O'Valle et al., 2014;Nguyen et al., 2004 [1, 2]). The palmar side of the wrist is relatively rare, which is often associated with the wrist. Ultrasound and MRI can detect and diagnose early. Case presentation: In this case report, we discussed an elderly woman with palmar carpal tendon sheath cyst and ruptured. She communicated with the outside world through the skin sinus, and at the same time pressed the radial artery and the median nerve to produce obvious clinical symptoms. Clinical discussion: Because of its deep position and close relationship with the surrounding important nerves and vessels, the operation was relatively difficult. In particular, in this case, tendon sheath cyst ruptured to form sinus, so there were some difficulties. Conclusion: Pathological diagnosis was tendon sheath cyst. The analysis of the relationship between the tumor and the surrounding tissue by preoperative MRI and other imaging examinations has important guiding significance for surgery.Case informationClinical dataThe patient, a 67-year-old woman, was admitted to the department with a 45-day history of right wrist tumor rupture and hand numbness. The patient had no history of drug abuse, no family history, no history of smoking and drinking, and had a good mental state. Special physical examination: the palm side of the right wrist radial side visible a tumor, about 20× 15 mm in size. The skin on the surface of the tumor was dull, and a skin sinus was visible in the center. Some scabs were formed on the surface, and gel-freeze-like liquid exudation was visible. The tumor touched softly, adhered closely to the surrounding tissue, and could touch the radial artery pulse. Capillary refill of the five fingers was good, and the blood supply was good. The activities of the right wrist joint and the five fingers were not significantly limited. The muscle strength was grade V, and the right thumb, index finger and middle finger felt numbness(Fig. 1). Auxiliary examination: MRI examination of wrist joint showed oval long T1 and long T2 signals in the radial joint space of the right wrist, and high signal on T2 WI fat compression, about 18 × 11.57 mm in size. The lesions showed isthmus and septum, and the radial artery and median nerve were significantly compressed in the wrist(Fig. 2).This case is reported in line with the SCARE criteria [3].MethodSurgical resectionAfter the preoperative examination was completed, the surgical contraindications were excluded. After preoperative discussion, the left wrist tumor resection was performed under local anesthesia. Surgery operated by the authors. The patient was supine, and the right upper limb was extended and abducted. Local block anesthesia was performed around the tumor with 2 % lidocaine. After the anesthesia was effective, the surgical area was prepped and draped in usual fashion, and sterile sheets were laid. The operation began: 30 × 10 mm ellipse incision was made at the center of the sinus tract formed by the tumor and skin, and the skin of the sinus tract and some dark and light skin were removed, and the tumor was gradually separated and exposed (Fig.
tendon sheath cyst had slight compression on the median nerve of the wrist. The blood flow of nutrient vessels branches were interrupted on the surface of median nerve of wrist . Attention should be paid to the protection of nerves, blood vessels and other important tissues during the operation. The cyst was stripped and removed layer by layer . The cyst size was about 20 × 15 mm. The cyst was a multinodular cyst with yellow jelly-like liquid. The deep part of the cyst was pedicled, which was closely related to the wrist. The cyst was completely separated from the surrounding tissues and completely removed . The wound cavity was washed thoroughly, no obvious active bleeding was found in the wound cavity, and the bleeding was stopped thoroughly. The incision was sutured layer by layer, and the sterile dressing was wrapped well. After the operation, the specimen was sent to pathology. Postoperative hospital treatment, to prevent infection, timely dressing treatment.
## Pathological examination of specimens
General view: 1 cystic tumor, size 13 × 7 × 5 mm, multilocular cystic. Microscopically, the cystic wall was composed of fibrous tissue. Local cystic wall showed glass-like changes, with infiltration of lymphocytes and plasma cells in the cystic wall, and focal vascular dilatation and congestion . Pathological diagnosis: tendon sheath cyst.
The postoperative incision healed well without obvious infection and exudation. The numbness in the hand gradually disappeared on the second day after operation. On the seventh day after operation, there was no numbness in the hand. The radial artery pulse was palpated. The incision was healed now. Patients are satisfied with the current treatment.
# Discussion
This case has the following characteristics: rupture of tendon sheath cyst, compression and adhesion of important blood vessels, and compression of important nerves. Although tendinous cysts are common in clinical practice [bib_ref] Morphological and immunohistochemical evaluation of ganglion cysts. Cross-sectional study of 354 cases, O'valle [/bib_ref] , tendinous cysts on the palmar side of the wrist are relatively rare. In this case, the tendon sheath cyst was resected to achieve the purpose of treatment, and the auxiliary examination played an important role. MRI has a very important guiding significance for the clear diagnosis and surgical treatment. It is because the MRI examination found that the tendon sheath cyst was closely related to the radial artery and the median nerve, which reduced the nerve and blood vessel damage caused by intraoperative operation errors, and avoided the surgical complications caused by this, and played a protective role in the nerve and blood vessel. At the same time, it is found that it is difficult to achieve the healing effect through non-surgical treatment for the ruptured tendon sheath cyst that forms the sinus tract with the outside world. It is often repeatedly ruptured and the wound does not heal. Even during the operation, the skin around the sinus tract should be debridemented, and the skin and subcutaneous tissue around the sinus tract should be removed. Only when the wound is fresh, can the normal healing of the incision after the excision of the cyst be ensured. Tendon sheath cyst is a common disease, and surgical treatment is also a conventional treatment. However, if the above situation occurs, it should still be treated with caution and be alert to serious complications.
At present, there are many statements about its pathogenesis, and it is believed that it may be related to degenerative injury in and around the joint capsule, synovial hernia around tendon sheath, mesenchymal tissue metaplasia, repeated trauma or ligament injury [bib_ref] Imaging of wrist masses, Nguyen [/bib_ref]. Loder et al. [bib_ref] A surface ultrastructure study of ganglia and digital mucous cysts (J), Loder [/bib_ref] compared finger mucinous cyst and tendon sheath cyst and found that although they had different occurrence sites, they had the same performance in optics and electron microscopy, so they believed that they were caused by the same reason. High-frequency color Doppler ultrasound is a non-invasive examination method, which has the advantages of low price, simple operation and no side effects. It has been accepted by patients, especially suitable for the examination and diagnosis of muscles, joints and ligaments [bib_ref] Diagnostic value of high frequency ultrasound in soft tissue mass of foot, Jinlan [/bib_ref]. The principle of highfrequency color Doppler ultrasound in the examination of tendon sheath cysts is mainly due to the different density and acoustic impedance of different tissues in the human body. When ultrasound is performed, ultrasound will pass through the acoustic interface between the two tissues to form different echoes. Doctors can judge the pathological changes of tendons, articular cartilage and ligaments according to this echo performance [bib_ref] Clinical value of high frequency color doppler ultrasound in the diagnosis of..., Ru [/bib_ref]. Although MRI is expensive, it has high tissue resolution and good indications for soft tissue and joint injury, so it has great clinical value [bib_ref] Soft tissue masses in the foot and ankle: characteristics on MR imaging, Woertler [/bib_ref]. Tendon sheath cyst usually needs surgical treatment. Surgical resection of the cyst and release of external pressure factors can quickly reduce edema, numbness, pain and other symptoms. At the same time, the long-term effect is good. Complete resection of the cyst wall is performed by surgery. If there is a pedicle connected with the articular cavity, after the removal of the cyst, the joint capsule needs to be repaired, ligated, repaired and connected with the joint capsule to prevent postoperative recurrence [bib_ref] Treatment report of 7 cases of iliopubic cyst, Jiaxiang [/bib_ref]. In recent years, with the extensive development of arthroscopic surgery, more and more scholars have adopted arthroscopic surgery to remove cysts and relieve compression symptoms, which has achieved good results and low recurrence rate [bib_ref] Medial synovial fold cyst in the hip leading to pectineofoveal impingement, Nakano [/bib_ref] [bib_ref] Hip impingement: beyond femoroacetabular, Bardakos [/bib_ref] [bib_ref] Hip labral cyst caused by psoas impingement, Tey [/bib_ref].
# Conclusion
In summary, palmar carpal ganglion cyst is relatively rare, and because of its deep location, it is often delayed in diagnosis or difficult to find. When it is large or because of pain, swelling, numbness and other corresponding clinical symptoms when pressing surrounding important tissues, it is easy to be found. Symptomatic ganglion cyst often needs surgical resection, and ruptured ganglion cyst appears. It should be operated in time to prevent infection caused by the same articular cavity as the outside world. The final diagnosis of tendon sheath cyst depends on pathological diagnosis. Ultrasonography and MRI have good reference value for diagnosis and surgical treatment.
# Ethical approval
All patients with surgical procedures, surgical methods, possible complications and risks were informed. And the above details have been informed consent of patients and their families. All operations were completed by the Affiliated Hospital of Changchun University of Traditional Chinese Medicine. All procedures were guided and allowed by the Ethics Committee of Affiliated Hospital of Changchun University of Traditional Chinese Medicine (number: CCZYFYLL-SQ-2021).
# Funding
This study was supported by Jilin Province Young and Middle-aged Science and Technology Innovation and Entrepreneurship Outstanding Talent (team) Project (No. 20210509001RQ).
## Guarantor
## Xiangyang leng
## Research registration number
## Cczyfyll-sq-2021
Author contribution
[fig] Figure 1: A: Front view image of wrist cyst. B: Lateral view image of wrist cyst. [/fig]
[fig] Figure 2: A: MRI examination of wrist cysts coronal images. The red arrow showed radial artery compressed by wrist cyst. B: MRI examination of wrist cysts transverse images. The yellow arrow showed the median nerve compressed by wrist cysts, the red arrow showed the radial artery compressed by wrist cysts, and the blue arrow showed the head vein compressed by wrist cysts. C: MRI examination of sagittal images of wrist joint cyst. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.) [/fig]
[fig] Figure 3, Figure 4: A: Spindle excision of necrotic black skin and subcutaneous tissue around sinus tract, separation of surrounding tissue gradually revealed wrist cysts. B: The wrist cyst is adjacent to the head vein, and there is mild adhesion and compression. The blue arrow shows the head vein. C: Wrist joint cyst wall of radial artery wrapped serious adhesion, there is a serious compression. Red arrow shows radial artery, blue arrow shows head vein. D: The median nerve of wrist was slightly compressed by wrist cyst. The nutrient vessels on the surface of median nerve of wrist were disrupted. The green arrow shows the median nerve. E: Vascular and nerve protection after complete excision of wrist cyst. F: Complete excision of wrist cyst, black arrow showing cyst pedicle. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.) The capsule wall is composed of fibrous tissue. Glassy changes of local capsule wall, infiltration of lymphocytes and plasma cells in capsule wall, focal vasodilation and congestion. A: Scale = 200um. B: Scale = 100um. [/fig]
|
Posterior reversible encephalopathy syndrome in a patient with mixed connective tissue disease: a case report
Background: Posterior reversible encephalopathy is a syndrome highly associated with hypertension and cytotoxic therapy. The syndrome typically presents with headache, visual abnormality, seizures and characteristic vasogenic edema on magnetic resonance imaging. The entity warrants a prompt diagnosis to avoid deteriorating consequences. Case presentation: In this report, we describe a 15-year-old Iranian boy who was diagnosed with mixed connective tissue disease, and cyclophosphamide pulse therapy was administered. Three days after the second pulse of cyclophosphamide, when he was receiving prednisolone and hydroxycholoroquine, our patient developed generalized tonic-clonic seizures. Magnetic resonance imaging findings showed high signal intensities in the posterior areas of his brain. After 8 days, the brain magnetic resonance imaging abnormalities were resolved following the control of his blood pressure and antiepileptic treatment. These observations have been indicative of posterior reversible encephalopathy syndrome. Nevertheless, our patient developed uncontrollable respiratory distress and eventually died. Conclusions: To the best of our knowledge, this case is the first report of posterior reversible encephalopathy syndrome in a patient with mixed connective tissue disease. As the patient developed posterior reversible encephalopathy syndrome 3 days after cyclophosphamide pulse therapy to reduce the disease activity, it is hard to accurately determine whether posterior reversible encephalopathy syndrome in this case is a complication of cyclophosphamide or a condition that resulted from the mixed connective tissue disease flare-up.
# Background
Posterior reversible encephalopathy syndrome (PRES) is a life-threatening condition which can be characterized by symmetric involvement of posterior white matter on magnetic resonance imaging (MRI) and neurological impairments such as seizures, altered mental status, headache, and visual disturbances [bib_ref] Posterior reversible encephalopathy syndrome, part 2: controversies surrounding pathophysiology of vasogenic edema, Bartynski [/bib_ref] [bib_ref] Posterior reversible encephalopathy syndrome, part 1: fundamental imaging and clinical features, Bartynski [/bib_ref]. PRES has been reported in different conditions such as hypertensive encephalopathy, eclampsia, thrombotic thrombocytopenia purpura, and rheumatologic disorders [bib_ref] Diffusion-weighted imaging discriminates between cytotoxic and vasogenic edema in a patient with..., Schaefer [/bib_ref] [bib_ref] Quantitative assessment of diffusion abnormalities in posterior reversible encephalopathy syndrome, Provenzale [/bib_ref] [bib_ref] Posterior reversible encephalopathy syndrome: prognostic utility of quantitative diffusionweighted MR images, Covarrubias [/bib_ref].
The mainstay of management of PRES is timely diagnosis and discontinuation of causative agents that may prevent subsequent abnormalities of the central nervous system.
The extensive use of immunosuppressive therapy and the autoimmune nature of rheumatologic diseases may make patients more vulnerable for developing PRES in the course of disease. Nevertheless, to the best of our knowledge, PRES has not been reported as a complication of treatment or a manifestation of disease in patients with mixed connective tissue disease (MCTD).
In this report, we describe a 15-year-old Iranian boy with MCTD who presented with PRES 3 days after cyclophosphamide pulse therapy when he was receiving a high dose of steroids. Our patient was treated with antihypertension and antiepileptic medications and a repeat MRI scan showed no abnormality 8 days later.
## Case presentation
Our patient was a 15-year-old Iranian boy with a 2-year history of skin ulcer compatible to pyoderma gangrenosum. From the onset of his skin problems, he had been receiving a low dose of steroids, which was increased to 1 mg/kg 2 months prior to admission. He was referred to our hospital following development of muscle weakness and severe dyspnea. History-taking revealed a 1-year history of discoloration of his fingers in cold temperatures. A physical examination showed scleroderma-like signs of acrosclerosis and a small mouth orifice with difficulty in opening. Blood tests showed a remarkable elevation of muscle enzymes (creatine phosphokinase [CPK] >3000, aldolase 39.4 and lactate dehydrogenase [LDH] 1510) and electromyogram-nerve conduction (EMG-NCV) tests indicated chronic moderate to severe myopathic process. We performed a muscle biopsy of his left deltoid muscle that revealed multiple necrotic fibers and extensive inflammatory endomysial foci. The laboratory findings showed antinuclear antibodies (ANA) 1:2500 positive, anti-doublestranded (ds) DNA 198 positive, anti-SM >200 positive, anti-SCL-70 >200 positive, anti-centromere >2 positive, anti-U1 RNP 178.4 positive, white blood cells (WBC) (from 3500 to 6900 during the hospitalization), hemoglobin (Hb) 12.3, platelets (PLT) 128,000, and erythrocyte sedimentation rate (ESR) 56 (normal range <30). His anticardiolipin, anti-beta 2 glycoprotein I and lupus anticoagulant antibodies were negative and also his complement 3 (C3) and complement 4 (C4) levels were in normal range. Our patient fulfilled the Alarcon-Segovia diagnostic criteria [bib_ref] Comparative study of 4 diagnosis criteria sets for mixed connective tissue disease..., Amigues [/bib_ref] with positive serology and three of the five clinical criteria especially Raynaud's phenomenon, acrosclerosis, and myositis. Our patient also met the Kasukawa diagnostic criteria [bib_ref] Comparative study of 4 diagnosis criteria sets for mixed connective tissue disease..., Amigues [/bib_ref] with one common symptom of Raynaud's, positive serology, and mixed findings of leukopenia/thrombocytopenia, acrosclerosis, and muscle weakness. A chest X-ray showed diffuse pulmonary infiltration and a computed tomography (CT) scan reported a bronchiolitis obliterans organizing pneumonia (BOOP) reaction. Further tests also showed heart failure (ejection fraction [EF] = 30%) and pulmonary arterial hypertension (pulmonary artery pressure [PAP] = 75 mmHg). Pulmonary embolus was ruled out by CT angiography. According to the criteria, our patient was diagnosed with mixed connective tissue disorder (MCTD), and 1 g intravenous methylprednisolone was administered. Then our patient received 500 mg cyclophosphamide pulse therapy, and was discharged with prednisolone 70 mg daily and hydroxycholoroquine 200 mg daily. Our patient received the second pulse of 500 mg cyclophosphamide 2 weeks later. Three days after the second cyclophosphamide pulse when he was receiving prednisolone 70 mg/day, he developed several generalized tonic-colonic seizures.
After admission to the intensive care unit, he developed another seizure that lasts 3 minutes. At this time, his blood pressure was 170/130. Therefore, phenytoin and antihypertension drugs were prescribed. An MRI scan of our patient revealed high signal intensities on T2weighted images and fluid-attenuated inversion recovery (FLAIR) sequences in the subcortical white matter of the occipital, posterior parietal, and posterior temporal lobes, and the cerebellum [fig_ref] Figure 1: Brain magnetic resonance imaging T2-weighted/fluid-attenuated inversion recovery scan showing high signal intensities... [/fig_ref]. After 8 days, the brain MRI abnormalities had completely been resolved. These observations have been indicative of PRES.
From the first day of his second admission, our patient presented fever, cough and dyspnea and the laboratory tests showed creatinine 2.35 mg/dL, urea 182 mg/dL, WBC 3500, and platelets 75,000. A chest X-ray revealed extensive pleural effusion. In a peripheral blood smear, there was no evidence of thrombotic thrombocytopenia purpura. The serial blood test showed a progressive increase in his creatinine level and decrease in his platelet count. Therefore, we started rituximab 500 mg for 2 weeks to control the disease flare-up. However, rituximab was not efficient and we started intravenous immunoglobulin (IV Ig) for 5 days. IV Ig did not improve our patient and we had to initiate plasma exchange. However, his creatinine level increased and respiratory symptoms became worse. Dialysis prevented the worsening of his condition and relieved our patient's symptoms. The results from bronchoalveolar lavage (BAL) showed infection to cytomegalovirus (CMV), Gram-negative bacillus, and candida. Although our patient received broad-spectrum antibiotics and antifungal agents, his respiratory manifestations were not improved and he was intubated. After 5 days of intubation, our patient developed heart arrest and, following 45 minutes of cardiopulmonary resuscitation (CRP), our patient died.
# Discussion
Posterior reversible encephalopathy syndrome was first introduced by Hinchey and his colleagues in a study of 15 patients [bib_ref] A reversible posterior leukoencephalopathy syndrome, Hinchey [/bib_ref].
PRES can be diagnosed in a patient with reversible neurological manifestations including headache, nausea/ vomiting, visual abnormalities, consciousness impairment, seizure activity, and focal neurologic signs, in conjunction with bilateral involvement of posterior brain areas on magnetic resonance imaging [bib_ref] Posterior reversible encephalopathy syndrome, part 1: fundamental imaging and clinical features, Bartynski [/bib_ref].
The pathophysiology of PRES remains elusive. However, it has been suggested that a compromised cerebrovascular autoregulation due to acute hypertension may play a pivotal role. Accordingly, impaired cerebrovascular regulation may lead to arteriole leakage and cerebral vasogenic edema [bib_ref] Posterior reversible encephalopathy syndrome, part 2: controversies surrounding pathophysiology of vasogenic edema, Bartynski [/bib_ref] [bib_ref] Posterior reversible encephalopathy syndrome, part 1: fundamental imaging and clinical features, Bartynski [/bib_ref]. Although the co-occurrence of hypertension and PRES is remarkable, other possible mechanisms have also been reported to be important in PRES such as disrupted cerebral autoregulatory mechanisms due to autonomic dysfunction [bib_ref] Autonomic dysreflexiainduced reversible posterior leukoencephalopathy syndrome in patients with spinal cord injury:..., Joa [/bib_ref] , breakdown of blood-brain barrier following cytotoxic agents-induced endothelial toxicity, autoimmunity, and sepsis [bib_ref] Posterior reversible encephalopathy syndrome in infection, sepsis, and shock, Bartynski [/bib_ref].
In addition, immunosuppressive agents such as methylprednisolone, dexamethasone, cyclosporine and cyclophosphamide have been reported to be related to PRES [bib_ref] Posterior reversible encephalopathy syndrome: incidence of atypical regions of involvement and imaging..., Mckinney [/bib_ref].
Our patient presented PRES 3 days after cyclophosphamide pulse therapy when he was receiving a high dose of prednisolone in the setting of a MCTD flare-up. On the other hand, clinical signs showed that our patient developed sepsis concomitant with PRES. Therefore, we could not accurately determine the main causative source of PRES among the cyclophosphamide pulse therapy, prednisolone, sepsis, and the flare-up of the underlying disease. Our patient had a 2-year history of taking steroids for his dermatological disorder that may decrease the importance of maintenance prednisolone as the cause of PRES. Most of the reported cases of PRES are suggested to be due to cytotoxic or steroid therapy. In the present case, the diagnosis of PRES a few days following the second pulse of cyclophosphamide may underscore its offensive effects and may suggest the cyclophosphamide pulse as a potential cause. Nevertheless, several reports show presentation of PRES unattributable to therapy in patients with rheumatologic diseases, which suggested it to be a condition resulting from underlying disease and even, in some cases, PRES was the first manifestation of the disease [bib_ref] Posterior reversible encephalopathy syndrome: an emerging disease manifestation in systemic lupus erythematosus, Barber [/bib_ref]. As in our case, PRES was concomitant with sepsis, therefore sepsis may play a role in this condition.
From the standpoint of pathogenesis, MCTD may be strongly suggestive as a cause of PRES. Various studies highlight that endothelial cell damage resulted from different autoantibodies in MCTD [bib_ref] Endothelial cell markers reflecting endothelial cell dysfunction in patients with mixed connective..., Soltesz [/bib_ref]. Along these lines, vasculopathy has been suggested to be a specific feature of MCTD that may lead to pulmonary arterial hypertension, which is responsible for most of the deaths in the late stages of MCTD [bib_ref] Pulmonary vascular manifestations of mixed connective tissue disease, Bull [/bib_ref]. In addition, the dysfunction of the autonomic system has been shown in MCTD [bib_ref] Cardiovascular autonomic function, autoantibodies, and esophageal motor activity in patients with systemic..., Stacher [/bib_ref]. Therefore, PRES may be a manifestation of MCTD To the best of our knowledge, a case of PRES in a patient with mixed connective tissue disease has not yet been reported. The neurological manifestations of MCTD have been believed to be less frequent than findings of other systems. Although the main neurological manifestations of MCTD are trigeminal neuropathy, headaches, and aseptic meningitis, this report suggests PRES as a neurological condition which may occur during the course of MCTD [bib_ref] Mixed connective tissue disease: an overview of clinical manifestations, diagnosis and treatment, Ortega-Hernandez [/bib_ref]. Although there is no difference between our patient and previously reported cases in terms of PRES characteristics, further reports are needed for better understanding of PRES in MCTD.
# Conclusions
To the best of our knowledge, this case is the first report of posterior reversible encephalopathy syndrome in a patient with mixed connective tissue disease. As the patient developed PRES shortly after the cyclophosphamide pulse, during an MCTD flare-up, and concomitant with sepsis, it is hard to determine the accurate cause of PRES in this case. Taken together, MCTD may be strongly suggestive as the cause of PRES for the extensive endothelial dysfunction involved in its pathogenesis.
## Consent
Written informed consent was obtained from the patient's family for publication of this case and any accompanying images. A copy of the written consent is available for review by the Editor-in-Chief of this journal.
Abbreviations CT, computed tomography; MCTD, mixed connective tissue disease; MRI, magnetic resonance imaging; PRES, posterior reversible encephalopathy syndrome.
[fig] Figure 1: Brain magnetic resonance imaging T2-weighted/fluid-attenuated inversion recovery scan showing high signal intensities in a the subcortical white matter of occipital, posterior parietal, and posterior temporal lobes and b the cerebellum. c, d Follow-up brain magnetic resonance imaging T2-weighted/fluid-attenuated inversion recovery scan 8 days after the first imaging showed complete resolution in our case because our patient presented PRES during the disease flare-up. [/fig]
|
System‐level metabolic modeling facilitates unveiling metabolic signature in exceptional longevity
## | introduc ti on
Population aging is an increasingly urgent issue confronting many countries worldwide [bib_ref] Measuring population ageing: An analysis of the Global Burden of Disease Study, Chang [/bib_ref]. As most disabilities and fatal human diseases are age-related [bib_ref] Global, regional, and national disability-adjusted life-years (DALYs) for 359 diseases and injuries..., Dalys [/bib_ref] , understanding the mechanisms of aging will help with the development of therapeutics for aging-related diseases. Metabolic control plays a crucial role in regulating the health span and life span of various organisms, for example, worms [bib_ref] Cell nonautonomous activation of flavin-containing monooxygenase promotes longevity and health span, Leiser [/bib_ref] and primates [bib_ref] Impact of caloric restriction on health and survival in rhesus monkeys from..., Mattison [/bib_ref]. Dysregulated metabolism often leads to premature aging and certain diseases in humans [bib_ref] Metabolic control of longevity, López-Otín [/bib_ref]. In contrast, long-lived people, such as centenarians (CENs), may have "healthy" metabolic profiles that support them in resisting age-and metabolic-related diseases, although the exact mechanism remains elusive [bib_ref] Metabolic control of longevity, López-Otín [/bib_ref].
Human metabolism is a complex network that contains thousands of reactions and metabolites, and systematic identification of metabolic changes in health and diseases remains challenging [bib_ref] Recon3D enables a three-dimensional view of gene variation in human metabolism, Brunk [/bib_ref]. Constraint-based reconstruction and analysis is based on flux balance analysis theory [bib_ref] What is flux balance analysis?, Orth [/bib_ref] and uses different types of constraints, including metabolite availability, nutrient limits, and the most widely available data, gene expression from either microarray or RNA-seq, to build tissue-specific metabolic models [bib_ref] Integration of expression data in genome-scale metabolic network reconstructions, Blazier [/bib_ref] [bib_ref] Using genome-scale models to predict biological capabilities, O'brien [/bib_ref] [bib_ref] Fast reconstruction of compact context-specific metabolic network models, Vlassis [/bib_ref] [bib_ref] Reconstruction of genomescale metabolic models for 126 human tissues using mCADRE, Wang [/bib_ref] and perform metabolic modeling [bib_ref] Quantitative and logic modelling of molecular and gene networks, Le Novere [/bib_ref] [bib_ref] Modeling cancer metabolism on a genome scale, Yizhak [/bib_ref]. Many COBRA methods have been developed to perform the metabolic modeling, and most of them were merged into the COBRA toolbox, a desktop software suite of interoperable COBRA methods [bib_ref] Quantitative prediction of cellular metabolism with constraint-based models: The COBRA Toolbox, Becker [/bib_ref] [bib_ref] Creation and analysis of biochemical constraint-based models using the COBRA, Heirendt [/bib_ref] [bib_ref] Quantitative prediction of cellular metabolism with constraint-based models: The COBRA Toolbox v2.0, Schellenberger [/bib_ref]. COBRA methods have been widely used for modeling cellular metabolism [bib_ref] Towards the routine use of in silico screenings for drug discovery using..., Bintener [/bib_ref] [bib_ref] Creation and analysis of biochemical constraint-based models using the COBRA, Heirendt [/bib_ref] [bib_ref] Network context and selection in the evolution to enzyme specificity, Nam [/bib_ref] , and discovering disease mechanisms [bib_ref] Large-scale in silico modeling of metabolic interactions between cell types in the..., Lewis [/bib_ref] [bib_ref] Systems biology in hepatology: Approaches and applications, Mardinoglu [/bib_ref] , targets [bib_ref] Genomescale metabolic modeling of glioblastoma reveals promising targets for drug development, Larsson [/bib_ref] , and drug candidates [bib_ref] Towards the routine use of in silico screenings for drug discovery using..., Bintener [/bib_ref].
A major challenge in previous metabolic modeling studies is the "low accuracy" in predicting metabolic fluxes in human cells, largely due to the fact that they considered merely qualitative data, rather than quantitative information. In the most commonly used methods [bib_ref] Reconstruction of genome-scale active metabolic networks for 69 human cell types and..., Agren [/bib_ref] [bib_ref] Integration of expression data in genome-scale metabolic network reconstructions, Blazier [/bib_ref] [bib_ref] Identifying and targeting cancer-specific metabolism with network-based drug target prediction, Pacheco [/bib_ref] [bib_ref] Network-based prediction of human tissue-specific metabolism, Shlomi [/bib_ref] [bib_ref] Fast reconstruction of compact context-specific metabolic network models, Vlassis [/bib_ref] [bib_ref] Reconstruction of genomescale metabolic models for 126 human tissues using mCADRE, Wang [/bib_ref] , quantitative gene expression or proteomics data need to be translated into qualitative values [bib_ref] Using genome-scale models to predict biological capabilities, O'brien [/bib_ref]. Such kind of doing inevitably leads to the loss of most of the quantitative information and thus introduces biases in predicting metabolic fluxes in human cells.
In this work, we present a systems biology approach to quantitatively integrate omics (i.e., transcriptome and proteome) data and kinetome information into genome-wide precision metabolic modeling (GPMM), aiming to accurately identify metabolic changes in human health and diseases. To benchmark its performance, we applied GPMM and other methods commonly used for metabolic modeling on the same transcriptome data from the NCI-60 cell lines [bib_ref] Cell Miner: A web-based suite of genomic and pharmacologic tools to explore..., Reinhold [/bib_ref] to compare the predicted metabolic fluxes with the experimentally measured values. GPMM robustly and reliably predicted the experimentally measured fluxes and significantly outperformed the existing methods. We then applied GPMM to study the metabolism of a Chinese centenarian cohort to understand why CENs can delay or avoid many serious age-related diseases. We found that elevated fatty acid oxidation (FAO) is the most significant metabolic feature in the CENs. Further serum metabolomic data showed that the decreased serum fatty acid concentration was the most significant feature in the CENs, supporting our observations from metabolic modeling results. Our study suggested a new signature in exceptional longevity.
## | re sults
## | developing genome-wide precision metabolic modeling method
In the present study, we developed a genome-wide precision metabolic modeling (GPMM) method to address the "low accuracy" challenge. The method quantitatively integrates the enzyme kinetics information from knowledge bases and the enzyme levels from transcriptome and proteome data into metabolic models and Figures S1-S3). Specifically, we first curated the generic human metabolic model [bib_ref] Recon3D enables a three-dimensional view of gene variation in human metabolism, Brunk [/bib_ref] and set the upper bounds for the main nutrient uptake rates in blood using information from the literature. To reduce noise from reactions without enzyme kinetics information (the turnover number, Kcats), we then constructed a reduced Recon 3D model to maximize the number of reactions with Kcats and minimize the number of reactions without Kcats . We next quantitatively integrated the enzyme kinetic parameters and gene expression levels to constrain the upper and lower bounds of each reaction (Figures S1-S3). According to Michaelis-Menten kinetics, the upper bound of a reaction is the product of the concentration and turnover number (Kcat) of its enzyme. Finally, we used flux variability analysis (FVA) to obtain individual models by removing the reactions with zero flux and then performed Markov Chain Monte Carlo (MCMC) sampling to identify metabolic changes and key regulators.
Notably, GPMM enabled several in silico analyses that were not included in the state-of-art COBRA methods toolbox (COBRA toolbox v3.0), and hence can be broadly applicable in metabolic K E Y W O R D S aging, GPMM, longevity, metabolic modeling, omics integration, systems biology engineering and therapy and . These include not only genome-wide single and combinatorial knock-in and knock-out analysis, but also quantitative inhibition and activation analysis to identify key regulators for target discovery and . In addition, GPMM can be applied to conduct personalized metabolic modeling for precision medicine and .
## | gpmm robustly and precisely captured the experimentally measured fluxes
Since the input transcriptome may carry out noise from the experimental or mapping procedures, we analyzed the robustness of GPMM to demonstrate whether GPMM has the ability to tolerate gene expression noise. We first constructed a noise-induced transcriptome by adding random values (viz. artificial noise) into the original expression data (viz. the genuine expression values) of each gene. Specifically, we induced 1%, 5%, and 10% noise into the gene expression data of NCI-60 cell lines to construct noise-induced transcriptomes. Then, we performed the metabolic modeling on the genuine and the noise-induced expression datasets, and then compared the obtained flux results between both datasets. If a method is robust, this method should be able to tolerate a certain extent of noise on quantified gene expression; then, the correlation (measured by R 2 or R-squared) between the genuine and noise-induced samples should approach 1. The results showed that the R-squared in each cell line is larger than 0.98 under either 1%, 5%, or 10% noise . To further investigate the robustness of GPMM, we next performed multiple random sampling to investigate whether GPMM is still robust at different sampling times . We induced 5% gene expression noise to the H460 (one of NCI-60 cell lines) for 100 times and performed metabolic modeling using GPMM. We obtained an average R 2 of 0.984 between the genuine and the noise-induced samples. Some important fluxes in cancer cells, such as ATP production and lactate secretion, were also consistent among these 100 simulations . These results indicated that GPMM is a robust method. F I G U R E 1 Flowchart of genome-wide precision metabolic modeling. A generic human metabolism model (Recon 3D) was first curated from the literature, and transcriptome data were then used to estimate enzyme abundance using a steady-state mathematical model. Next, a reducing model was constructed, and the upper bound of each reaction was calculated using the product of the concentration and turnover number (Kcat) of its enzyme. Finally, flux variability analysis (FVA) was performed to reconstruct individual models, and Markov Chain Monte Carlo (MCMC) sampling was used to detect metabolic differences and key regulators
To benchmark the performance of GPMM, we chose to utilize the transcriptome and metabolic flux (uptake and secretion) data of the NCI-60 cell lines ;. Four other methods commonly used for metabolic modeling, GIMME [bib_ref] Integration of expression data in genome-scale metabolic network reconstructions, Blazier [/bib_ref] , Fastcore [bib_ref] Fast reconstruction of compact context-specific metabolic network models, Vlassis [/bib_ref] , rFASTCORMICS [bib_ref] Identifying and targeting cancer-specific metabolism with network-based drug target prediction, Pacheco [/bib_ref] , and ecModel [bib_ref] An atlas of human metabolism, Robinson [/bib_ref] [bib_ref] Improving the phenotype predictions of a yeast genome-scale metabolic model by incorporating..., Sánchez [/bib_ref] , were chosen as comparisons. We applied each of the four methods to the transcriptome data of the NCI-60 cells and compared the computationally calculated metabolic fluxes with the reported experimental measurements. The results showed that the GPMM method had a much higher correlation between the predicted and experimental values , R 2 = 0.72, p = 2.3e-106) than GIMME , R 2 = 0.011), Fastcore The Warburg effect, indicated by an increase in lactate secretion, is one of the most important cancer hallmarks in NCI-60 cells. We thus compared the predicted lactate secretion fluxes with the experimental values, and found that GPMM Comparisons between predicted metabolic fluxes and experimentally measured lactate fluxes in NCI-60 cells using GPMM (e), GIMME (f), Fastcore (g), rFASTCORMICS (h) and ecModel (i). GPMM, Fastcore, and rFASTCORMICS had R 2 values of 0.86, 0.088 and 0.33, respectively, whereas GIMME failed to predict lactate secretion. ecModel has the ability to predict lactate secretion, but the magnitude of the predicted fluxes is different from the experimental values. Note: the ecModel reconstruction and the flux detection were derived from Zenodo (https://doi.org/10.5281/zenodo.3577466), and only 11 ecModels are available
[formula] (a) (b) (c) (d) (e) (f) (g) (h) (i) [/formula]
well-predicted the secretion of lactate in NCI-60 cells (R 2 = 0.86, p = 2.2e-24) . For other four methods, neither GIMME nor Fastcore could predict lactate secretion . One of the upto-date methods, rFASTCORMICS, returned reasonable results with the R 2 of 0.33 . However, the other up-to-date method, ecModel, fails to quantitatively predict lactate secretion , although its overall prediction performance is reasonable . These results showed that GPMM can precisely capture the experimentally measured fluxes and significantly outperformed the existing methods.
## | metabolic modeling reveals elevated fatty acid oxidation as the most significant metabolic feature in centenarians
To shed light on the metabolic characteristics of the CENs to better understand why these individuals are able to delay or avoid many serious age-related diseases that afflict the normal population [bib_ref] Morbidity profiles of centenarians: Survivors, delayers, and escapers, Evert [/bib_ref] , we aimed to study the metabolism of longevity in a CEN cohort sampled from Hainan Province, China. The cohort included 76 CENs, 54 centenarian-children (F1s), and 41 spouses of centenarian-children (F1SPs , whose RNA-sequencing data were reported in our previous study [bib_ref] Transcriptome evidence reveals enhanced autophagy-lysosomal function in centenarians, Xiao [/bib_ref].
We next applied GPMM to study the metabolic features of longevity in this cohort. In total, we developed 171 individual GPMM the CENs and younger controls (viz. F1SPs) and adjusted for age and gender effect, we obtained 343 upregulated and 90 downregulated fluxes. We observed that the overall CEN flux signature was slightly negatively correlated with the aging effect (r = −0.15, p = 7.1e-12) , suggesting that the CENs contain some signatures that are different from the ones associated with age.
The most striking signature in all four metabolic processes consistently indicated that long-chain fatty acid beta-oxidation ( Consistent with these observations, total cellular ATP production capacity was also significantly enhanced (p = 0.032) in the CENs [fig_ref] 6: DA i = # Up regulated fluxes − # down regulated fluxes... [/fig_ref].
## Ta b l e 1 overall population attributes of the hainan centenarian cohort
## | serum metabolomics supports the metabolic modeling observations
Because a higher systemic FAO leads to higher uptake and con- are likely heritable. Intriguingly, among the upregulated metabolites, the most significant ones were bile acids, a group of metabolites for fatty acid absorption . These results suggest that the decreased serum fatty acid concentration was the most significant feature in our centenarian metabolomics data. This observation also explains our previous epidemiological survey, which found that total cholesterol is decreased in the CENs compared with F1SPs [bib_ref] Assessment of the health status of centenarians in the south of China:..., He [/bib_ref]. A similar result was also obtained by analyzing the clinical data of the same longevity cohort studied here .
## | discuss ion
Identifying metabolic signatures in centenarians is important for healthy aging. Constraint-based reconstruction and analysis (COBRA) is a promising method that can capture metabolic signatures in health and diseases, but has been a long-standing challenge to quantitatively predict molecular phenotypes [bib_ref] Large-scale in silico modeling of metabolic interactions between cell types in the..., Lewis [/bib_ref] [bib_ref] Using genome-scale models to predict biological capabilities, O'brien [/bib_ref]. In this study, we present a solution to this criti- . As most parameters and datasets are precalculated, GPMM is easy to use, and the only required input is the transcriptome (RNA-sequencing or microarray). Therefore, GPMM is able to systematically analyze cellular metabolic profiles using only the transcriptome and enables broad computational studies on discovering disease mechanisms.
Previous studies indicated that the existing continued methods, such as PRIMEand , are less robust than the methods that utilize discretization workflows, such as rFASTCORMICS [bib_ref] Benchmarking procedures for high-throughput context specific reconstruction algorithms, Pacheco [/bib_ref] [bib_ref] Identifying and targeting cancer-specific metabolism with network-based drug target prediction, Pacheco [/bib_ref]. However, the results showed that our developed quantitative method can robustly and well predict experimentally measured fluxes. The reason may be because we not only translated the transcripts to proteome data using a simple but efficient model, but also used enzymatic parameters to restrain the maximum rate for each reaction. In addition, to avoid the effect of the unconstrained reactions on the metabolic simulation, we also reduced the generic models (i.e., Recon3D) to maximize the number of reactions with
Kcats and minimize the number of reactions without Kcats. Similar to rFASTCORMICS, GPMM can also use the secretion information to improve the model performance by adding the secretion reaction information in the exchange input file. These improvements thus largely overcome the performance and robustness issues of the existing methods. Therefore, the dramatic improvement of GPMM will enable many computational studies to discover biomedical mechanisms.
Utilizing this method, we studied the metabolic profiles of CENs and identified the elevated fatty acid oxidation as the most significant metabolic feature in the CENs. As the input of GPMM is the transcriptome, we investigated the main beta-oxidation genes from Recon 3D [bib_ref] Recon3D enables a three-dimensional view of gene variation in human metabolism, Brunk [/bib_ref] and found that 11 of 14 (78.6%)
FAO-related genes were slightly upregulated in the CENs , including four essential peroxisomal beta-oxidation genes (EHHADH, HSD17B4, ACAA1, and ACOX1) and five key mitochondrial beta-oxidation genes (HADHB, ACAA2, ECHS1, ACADL, and ACADVL) . We also observed that 9 of 14 (64.3%) FAO-related genes were slightly downregulated with aging in F1SP samples . These results support our findings that the elevated fatty acid oxidation as a metabolic signature in the CENs. . Consistently, serum metabolism results also showed that most of down-regulated metabolites in serum are FALs (12 of 20, 60%); and most of FALs with significant differences between the two groups are downregulated (12 of 19, 63%) in F1s .
Interestingly, we also observed that F1s has lower free fatty acid levels (i.e., trans-vaccenic and trans-vaccenic levels) than F1SP . Given that F1s might have a higher probability of long lifespan than F1SPs, these results added further support to our conclusion that the elevated fatty acid oxidation is a signature involved to healthy human aging and longevity.
There are many studies in both model organisms and humans
showing associations between FAO decline and aging [bib_ref] Hepatic lipid metabolism and non-alcoholic fatty liver disease in aging, Gong [/bib_ref] [bib_ref] Reduced whole-body fat oxidation in women and in the elderly, Levadoux [/bib_ref] [bib_ref] Decline in skeletal muscle mitochondrial function with aging in humans, Short [/bib_ref]. However, whether FAO also declines in the CENs, the paradigm of healthy human aging, remains unclear. Interestingly, multiple lines of evidence from our study argues for an enhanced FAO in the CENs, which well explains the previous observations that decreased fatty acid levels, especially the PEs, were found in centenarians' serum [bib_ref] Exceptional human longevity is associated with a specific plasma phenotype of ether..., Pradas [/bib_ref]. In addition, impaired FAO is frequently observed in many age-related diseases, including atherosclerosis [bib_ref] Fatty acid-induced mitochondrial uncoupling elicits inflammasome-independent IL-1α and sterile vascular inflammation in..., Freigang [/bib_ref] and diabetes [bib_ref] Fatty acid synthesis configures the plasma membrane for inflammation in diabetes, Wei [/bib_ref]. Importantly, elevated FAO is reported to be causally associated with metformin-induced longevity in Caenorhabditis elegans . Elevated fatty acid beta-oxidation related genes extend the lifespan of worms . Collectively, these results suggest that the elevated long-chain FAO function in the CENs, at least in female CENs, represents a "healthy" metabolic profile of longevity, which may convey survival advantages to long-lived individuals by reducing lipid accumulation and lowering the risks of common age-related diseases, especially those involved in lipid metabolic disorders.
In summary, we have developed a novel systems biology approach to effectively integrate omics data in the modeling of metabolic mechanisms in human health and disease. This approach dramatically improved the performance over the existing methods.
Our method thus immediately enables many computational studies on discovering disease mechanisms and candidate drug targets, as well as further developments of the algorithms. We applied this method to investigate the metabolic profiles of CENs, and suggested the enhanced fatty acid oxidation as a novel metabolic signature of healthy aging in exceptional longevity.
Nevertheless, there are some limitations that need to be over-
## | knowledgebase curations
To reconstruct genome-wide metabolic models (GEMs) by the GPMM approach, we collected several relevant knowledge bases.
The knowledge-based human metabolic model was obtained from Recon 3D [bib_ref] Recon3D enables a three-dimensional view of gene variation in human metabolism, Brunk [/bib_ref]. The enzyme Kcat values were downloaded from BRENDA. Serum metabolite concentrations were obtained from the Human Metabolome Database (HMDB) [bib_ref] HMDB 4.0: The human metabolome database for, Wishart [/bib_ref].
Next, we manually curated the global human metabolic network of Recon 3D using thermodynamic analysis and the precured Recon 2 model . Because adenosine monophosphate (AMP) cannot be directly changed into adenosine triphosphate (ATP) in any reaction, some reversible reactions, such as FACOAL150 and RE1514M, were curated as irreversible.
The curated Recon 3D model is shown in .
## | setting quantitative upper and lower bounds of biochemical reactions
For each biochemical reaction, the flux of any reaction has the following equation:
where V is the flux of a reaction, V max is the maximum reaction rate according to Michaelis-Menten kinetics, [E] is the enzyme concentration,
and Kcat is the turnover number of the enzyme. The Kcat values of human enzymes were obtained from the BRENDA database. If an enzyme had multiple Kcat records, their median was used. Where experimental data were missing, we used Kcats from other species.
We obtained 2602 Kcat records in the 4352 reactions with an EC number in Recon 3D . Although 42% reactions with an EC number lacked Kcat records, the enzyme abundance percentage was smaller than 10% .
A previously published method, named GIMME [bib_ref] Integration of expression data in genome-scale metabolic network reconstructions, Blazier [/bib_ref] , was used to reduce the Recon 3D model to maximize the number of reactions with Kcats and minimize the number of reactions without Kcats. The GIMME objective functions were set to ATP production and biomass reaction, as described in previous studies [bib_ref] Integration of expression data in genome-scale metabolic network reconstructions, Blazier [/bib_ref] [bib_ref] A systems approach to predict oncometabolites via context-specific genome-scale metabolic networks, Nam [/bib_ref]
## | predicting enzyme abundance using gene expression data
Recently, a simple but efficient mathematical model was proposed to predict protein abundance using gene expression data [bib_ref] Mass-spectrometry-based draft of the human proteome, Wilhelm [/bib_ref]. Changes in enzyme abundance can be determined by the number of proteins synthesized from mRNA minus the number of proteins degraded. In the steady-state, we have following equation:
where E and M are the enzyme and corresponding mRNA abundances, respectively; α is the enzyme synthesis rate from mRNA; and γ is the enzyme degradation rate.
Thus, in the steady-state, we can predict enzyme abundance as follows:
To estimate the ∕ ratio, we downloaded microarray data of 12 normal tissues with GSE7307 (http://www.ncbi.nlm.nih.gov/ geo/query/ acc.cgi?acc=GSE7307) and RNA-seq data of 15 normal tissues from the Human Protein Atlas .
We also obtained the corresponding protein abundance data from the MOPED database [bib_ref] MOPED 2.5-An integrated multi-omics resource: Multi-omics profiling expression database now includes transcriptomics..., Montague [/bib_ref] with the unit of nmol/L. MOPED uses the human body map dataset and estimates protein concentration from protein abundance [bib_ref] MOPED 2.5-An integrated multi-omics resource: Multi-omics profiling expression database now includes transcriptomics..., Montague [/bib_ref].
Thus, the ∕ ratio was estimated using the median ratio of protein/ mRNA across multiple tissues.
The correlation between the transcriptome and proteome is usually quite low (Pearson correlation of 0.4-0.5, . Notably, using the steady-state kinetic method, the Pearson correlation between the predicted proteome and experimental measurements reached 0.8-0.9 , indicating that protein abundance could be correctly estimated using transcriptome data.
## | steps of the metabolic modeling
First, enzyme abundance was predicted using the above men-
[formula] (1) V ≤ V max = [E] × Kcat (2) dE dt = M − E( [/formula]
## | predicting fluxes of nci-60 cells using gpmm
The gene expression dataset (RNA-seq) was downloaded from the CellMiner website (https://disco ver.nci.nih.gov/cellm iner/loadD ownlo ad.do), and the uptake rate for each cell was obtained from the above experimental fluxes. Using the GPMM toolbox mentioned above, and setting ATP production and biomass reaction as the optimized functions, genome-wide precision metabolic modeling of NCI-60 cells was performed to obtain the flux matrix.
The predicted secretion fluxes were then compared with the experimental dataset to evaluate the overall performance of the GPMM.
## | predicting fluxes of nci-60 cells using gimme
The quantitative gene expression of NCI-60 was the first transformed into qualitative present/absent logical values using a Fragments Per Kilobase of transcript per Million mapped reads (FPKM) cutoff = 3.0. Metabolic models were then reconstructed using the "GIMME" function in the Cobra toolbox 3.0 [bib_ref] Creation and analysis of biochemical constraint-based models using the COBRA, Heirendt [/bib_ref]. Next, MCMC sampling was conducted to obtain the distribution of fluxes, and the average flux for each reaction was then calculated to obtain the GIMME-based flux matrix.
## | predicting fluxes of nci-60 cells using fastcore
The consistent Recon 3D model was first constructed using the "fastcc" function in the Fastcore toolbox [bib_ref] Fast reconstruction of compact context-specific metabolic network models, Vlassis [/bib_ref] with an epsilon of 1e-4 using the linear solver of cplex (https://www.ibm. com/analy tics/cplex -optim izer). The metabolic models were then reconstructed using the "fastcore" function in the Fastcore toolbox, and MCMC sampling was conducted using the "ACHRSampler" function in the Cobra toolbox 3.0 [bib_ref] Creation and analysis of biochemical constraint-based models using the COBRA, Heirendt [/bib_ref].
## | predicting fluxes of nci-60 cells using rfastcormics
rFASTCORMICS is an updated version of Fastcore that uses discretization workflows instead of the heuristic thresholds method [bib_ref] Identifying and targeting cancer-specific metabolism with network-based drug target prediction, Pacheco [/bib_ref]. The consistent Recon 3D model was also first constructed using the "fastcc" function in the Fastcore toolbox [bib_ref] Fast reconstruction of compact context-specific metabolic network models, Vlassis [/bib_ref] with an epsilon of 1e-4 using the linear solver of cplex (https://www.ibm.com/analy tics/cplex -optim izer). The metabolic models were then reconstructed using rFASTCORMICS [bib_ref] Identifying and targeting cancer-specific metabolism with network-based drug target prediction, Pacheco [/bib_ref] , and MCMC sampling was conducted using the "ACHRSampler" function in the Cobra toolbox 3.0 [bib_ref] Creation and analysis of biochemical constraint-based models using the COBRA, Heirendt [/bib_ref].
## | predicting fluxes of nci-60 cells using ecmodel
The model reconstruction and flux detection of 11 NCI-60 cell lines were derived from Zenodo (https://doi.org/10.5281/zenodo.3577466). Note: as each constructed ecModel has over 20,000 reactions and has a different model framework from the COBRA toolbox, performing MCMC sampling is difficult. Therefore, the fluxes for each ecModel were estimated using the suggested method "minProSimulation" from Human 1 [bib_ref] An atlas of human metabolism, Robinson [/bib_ref].
## | comparisons of the predicted and experimentally measured fluxes
The predicted secretion fluxes using different methods, that is, GPMM, GIMME, and Fastcore, were compared with the experimental flux datasets as mentioned above. To avoid linear optimization precision error, we removed the fluxes with absolute values smaller than 1e-6 mmol/L/min. Pearson correlations between predicted fluxes and experimental measurements were calculated to compare the performance of the different methods.
## | robustness analysis of gpmm
We first constructed noise-induced transcriptomes by adding random numbers (viz. artificial noise) to the original expression data (or the genuine transcriptome) of each gene. Specifically, 1%, 5%, and 10% noise was induced in each NCI-60 cell line transcriptome. In these processes, the noise inducing is produced from a uniform distribution of [0.99, noise. For example, in the 5% noise translation procedure, if a gene in a cell line has a gene expression (e.g., fpkm) of 1.0, the adding random number of this gene is ranged from −0.05 to 0.05, such as 0.03; then, the noised-induced gene expression is a number ranged from 0.95 to 1.05, such as 1.03. Second, we performed the metabolic modeling on the genuine and the noise-induced transcriptomes and compared the obtained flux results to evaluate GPMM robustness. In addition, to further test whether multiple sampling affects robustness, we also induced 5% noise to the gene expression of H460 cell line 100 times and performed metabolic modeling to determine the stability of GPMM. As shown in , 96% of CENs lived independently (e.g., eating, walking, and talking). Compared with F1SPs, CENs had significantly higher diastolic blood pressure (146.0 vs. 137.9, p = 0.03), similar systolic blood pressure (83.2 vs. 86.1 p = 0.19), slightly lower blood glucose (5.98 vs. 6.70, p = 0.14, t-test), lower total cholesterol (4.68 vs. 5.43 p = 0.009, t-test), and lower low-density lipoprotein cholesterol (2.45 vs. 2.97, p = 0.043, t-test). These results are also consistent with our previous studies, where levels of risk factors for cardiovascular diseases, including blood glucose, triglyceride, and total cholesterol, were significantly lower in the CENs than those of the general older population from the same province, and the diagnoses of type 2 diabetes mellitus, hypertriglyceridemia, and hypertension were lower in the CENs than Chinese national levels [bib_ref] Assessment of the health status of centenarians in the south of China:..., He [/bib_ref].
## | the chinese centenarians study
The relatively healthy status of CENs suggests that they can serve as a good model for healthy aging studies.
For transcriptome analysis, peripheral blood samples were treated with red blood cell lysis buffer (Tiangen Biotech) and then centrifuged at 1800 g for 10 min to isolate white blood cells. For metabolomics and proteomics measurements, peripheral blood samples were allowed to clot at room temperature for 30 min and then centrifuged for 10 min at 1500 g to extract the serum.
## | genome-wide precision modeling of the metabolism of centenarians
## | metabolic modeling
Gene expression (FPKM) levels in 170 individuals from the Hainan longevity cohort, including 76 CENs, 52 centenarian-children, and 42 spouses of centenarian-children (F1SPs), were derived from our previously published data [bib_ref] Transcriptome evidence reveals enhanced autophagy-lysosomal function in centenarians, Xiao [/bib_ref]. The upper bounds of the white blood cell metabolite uptake rates were separated into three categories: nutrient uptake for energy production, cofactors, and iron/oxygen uptake . The nutrient uptake rates, including those of glucose, l-glutamine, and fatty acids, were derived from published literature . The essential amino acid uptake rates were set to a small number, whereas those of cofactors, iron, oxygen, and primers for glycogen synthesis were set to unlimited . After preparing the gene expression and nutrient uptake rates, genome-wide precision metabolic modeling was conducted using our developed GPMM toolbox.
## | identification of metabolic flux profiles of centenarians
As the CENs (aged 98-108) were older than the controls (45-75), we used two linear models to distinguish centenarian signatures and aging effects. Model l was applied to CEN and F1SP samples to determine unfiltered centenarian signature:
Model 2 was applied to F1SP samples to determine aging effects:
After the unfiltered centenarian signatures and aging effect were determined, we obtained the actual centenarian signatures by excluding the overlapping fluxes between the unfiltered centenarian signature and the age effect. Therefore, upregulated fluxes in CENs were defined as fluxes with p < 0.05 and beta >0 in model 1 but not significant or beta <0 in model 2. Downregulated fluxes were defined vice versa.
## | identifying significant metabolic subsystems using differential abundance scores
The differential abundance (DA) score was calculated using previously published methods [bib_ref] An Integrated metabolic atlas of clear cell renal cell carcinoma, Hakimi [/bib_ref]. For each metabolic subsystem, the ith DA score (DA i ) was calculated as follows:
To obtain the significance of DA scores, we used a "bootstrap without replacement" method to calculate p-values. Briefly, we first randomly shuffled the sample labels 1000 times. Second, for each randomly shuffled label, the corresponding random DA scores were calculated using the above formula. We thus obtained 1000 random (4) Model 1: flux ∼ lm (centenarians + sex) (5) Model 2: flux ∼ lm (age + sex) where DA i is the centenarian DAscore of the ith subsystem and
RandomDAs i is the random DAscore of the ith subsystem. The adjusted p-value was calculated using the false discovery rate (FDR).
## | metabolomic analysis of the centenarians
## | sample preparation
The collected serum samples were thawed on ice. Samples (100 μl)
were extracted with 750 μl of methanol/acetonitrile/water solution (V methanol :V acetonitrile :V water = 2:2:1), with 30 μl of 1 mg/ml l-2chlorophenylalanineas then added as an internal standard, followed by vortexing for 10 s and sonicating for 10 min on ice. After that, the extract was incubated for 1 h at −20°C. Following centrifugation . Ion source gas 1 was 60, ion source gas 2 was 60, curtain gas was set to 30 L/h, source temperature was set to 550°C, and ion spray voltage floating (ISVF) was set to 5500 V or −4500 V in positive or negative modes, respectively. 4.5.3 | Data preprocessing and annotation UPLC-MS raw data (.wiff) were converted to mzXML, with ProteoWizard Peak exaction, identification, integration, alignment, and retention time correction processed with XCMS (R package, v3.2). The preprocessing results generated a data matrix that consisted of retention time (RT), mass-to-charge ratio (m/z) values, and peak intensity. The R package CAMERA was used for peak annotation after XCMS data processing. The in-house MS2 database was applied for metabolite identification, and only the metabolites with MS2 >0.8 remained.
## | identification of metabolomic profiles of centenarians
Using equations to those shown in flux analysis, we also used two linear models to distinguish centenarian signatures and aging effects in the metabolomic data. Model l was applied to CEN and F1SP samples to determine unfiltered centenarian signatures:
Model 2 was applied to F1SP samples to determine aging effects:
Then, the actual centenarian signature was calculated by excluding the overlapping metabolites between the unfiltered centenarian signature and the age effect. Therefore, upregulated metabolites in CENs were defined as the log2 transformed metabolic abundance with p < 0.05 and beta >0 in model 1 but not significant or beta <0 in model 2. Downregulated metabolites were defined vice versa.
## | identifying significant metabolic classes in metabolomic data
Similar to the flux subsystem analysis in Equations (6) and (7). The significance of the metabolic class was also analyzed using the differential abundance (DA) score (for . For each metabolic class, the DA score was calculated as the number of upregulated metabolites minus the downregulated metabolites and then divided by the total number of metabolites in the given class. Similarly, as presented in flux subsystem analysis, we used a "bootstrap without replacement" method to calculate p-values and then adjusted these p-values using the false discovery rate (FDR).
## Ack n owled g em ents
This work was supported jointly and equally by grants from the
[fig] F: I G U R E 2 Benchmark analysis of GPMM. (a) Main applications of the GPMM toolbox and comparison with COBRA Toolbox 3.0. (b) Pearson correlation of fluxes between the noise-induced gene expression and the genuine samples using GPMM in NCI-60 cell lines. (c) The Pearson correlation between 5% noise-induced gene expression and genuine samples in the H460 cell line 100 times. (d) Variations in two important fluxes (ATP production and lactate secretion) in cancer cells after inducing 5% gene expression noise using GPMM. (e-i) [/fig]
[fig] 6: DA i = # Up regulated fluxes − # down regulated fluxes Total reactions in ith subsystem DA scores for each subsystem. Third, the p-value was calculated as follows: [/fig]
|
Dominant mutations in ITPR3 cause Charcot‐Marie‐Tooth disease
Objective: ITPR3, encoding inositol 1,4,5-trisphosphate receptor type 3, was previously reported as a potential candidate disease gene for Charcot-Marie-Tooth neuropathy. Here, we present genetic and functional evidence that ITPR3 is a Charcot-Marie-Tooth disease gene. Methods: Whole-exome sequencing of four affected individuals in an autosomal dominant family and one individual who was the only affected individual in his family was used to identify diseasecausing variants. Skin fibroblasts from two individuals of the autosomal dominant family were analyzed functionally by western blotting, quantitative reverse transcription PCR, and Ca 2+ imaging. Results: Affected individuals in the autosomal dominant family had onset of symmetrical neuropathy with demyelinating and secondary axonal features at around age 30, showing signs of gradual progression with severe distal leg weakness and hand involvement in the proband at age 64. Exome sequencing identified a heterozygous ITPR3 p.Val615-Met variant segregating with the disease. The individual who was the only affected in his family had disease onset at age 4 with demyelinating neuropathy. His condition was progressive, leading to severe muscle atrophy below knees and atrophy of proximal leg and hand muscles by age 16. Trio exome sequencing identified a de novo ITPR3 variant p.Arg2524Cys. Altered Ca 2+ -transients in p.Val615Met patient fibroblasts suggested that the variant has a dominant-negative effect on inositol 1,4,5-trisphosphate receptor type 3 function. Interpretation: Together with two previously identified variants, our report adds further evidence that ITPR3 is a disease-causing gene for CMT and indicates altered Ca 2+ homeostasis in disease pathogenesis.1962
# Introduction
Charcot-Marie-Tooth disease (CMT) is a group of hereditary neuropathies, characterized by progressive distal sensory and motor impairment, which affects 1:2500 individuals.The disease is categorized into demyelinating CMT1, where median motor nerve conduction velocity (NCV) is < 38 m/s and axonal CMT2 where median motor NCV is> 38 m/s but compound muscle action potentials are (CMAP) decreased. Cases with features of both demyelination and axonopathy and NCV in the 30-45 m/s range are sometimes referred to as intermediate CMT.A large number of CMT disease gene discoveries have led to insights of the disease mechanisms and potential therapies.Dominant variants in ITPR3, which encodes the inositol 1,4,5-trisphosphate (IP 3 ) receptor (IP 3 R) type 3, were recently suggested as potential causes of CMT.Linkage analysis combined with exome sequencing revealed a p.Thr1424Met variant that segregated with a CMT phenotype in three patients from a single family.Furthermore, gene panel screening found a p.Met1064Val variant in a single index case for which no additional clinical details were provided.The pathogenicity of these variants was not confirmed by functional studies or by segregation of the variants in additional families.
Humans have three IP 3 R isoforms: IP 3 R1, IP 3 R2, and IP 3 R3. They are homologous in sequence but differ in physiological functions and tissue expression. 6,7 IP 3 is produced after activation of G protein coupled receptors (GPCR), and binds the tetrameric IPRs, which release Ca 2+ from ER into cytoplasm.The resulting elevation of intracellular Ca 2+ concentration has several downstream effects on the cell.The importance of IP 3 R signaling in neurons is underscored by the defects in IP 3 R1 leading to ataxia or Gillespie syndrome.Furthermore, the ER associated degradation pathway of activated IPRs is disrupted by inactivating variants in the genes ERLIN1, ERLIN2, and RNF170, causing hereditary spastic paraplegia and other neurodegenerative diseases. 14,15 IP 3 R3 itself has been implicated in apoptosis control, while alterations in its activity and/or expression levels drive oncogenesis and impact the survival of malignant cells.In this study, we provide confirmatory evidence of the association of ITPR3 with CMT. We introduce a CMT family with autosomal dominant mutation and one case with de novo mutation in ITPR3. In addition, we provide functional evidence of altered Ca 2+ dynamics in patient fibroblasts.
# Methods
## Patients and sequencing
Individuals P1, P2, P3, and P4 gave written informed consent and the ethics review board of HUS Helsinki University Hospital approved the study. Control fibroblast cells were from anonymous donors, who consented to use of the cells in scientific research. The fibroblast cells were collected from skin biopsy and cultured in Dulbecco's Modified Eagle Medium (DMEM), supplemented with 10% FBS (Life Technologies), 1% penicillin/streptomycin (Life Technologies), 1% L-glutamine (Life Technologies), and 0.2% uridine (Sigma). Cells were incubated at 37°C in 5 % CO2.
Exome sequencing for the Finnish family was performed as described previously.Sanger sequencing primers are shown in. Research trio whole-exome sequencing (WES) was done on individual P5 and parents after written informed consent was obtained through an institutional review board-approved research study at the Institute for Genomic Medicine at Columbia University (protocol AAAO8410). DNA was extracted from maternal, paternal, and proband samples, exome sequenced on a HiSeq 2500 or NovaSeq 6000 with the Kapa Biosystem's Library Preparation Kit, and whole-exome captured with NimbleGen SeqCap EZ v.3.0 rapid or v.4.
## Western blot
Fibroblasts were lysed in RIPA buffer (Cell Signaling #9806) and an aliquot containing 10 µg of total protein was boiled, separated in 4-20% Criterion TM TGX TM gels, and transferred to 0.2 µm nitrocellulose using Trans-Blot Turbo (all Bio-Rad). Blocking was done with 10% milk in 0.1% TBS-T. Antibodies were: IP 3 R1 (CT-1) , IP 3 R2 (NT-2) (gift from Dr. David Yule, University of Rochester, NY), IP 3 R3 (BD-Transduction #610312), and b-tubulin (Cell Signaling #2146S). Anti-rabbit and anti-mouse (Jackson Immunore-search#111-035-144 and #115-035-146) secondary antibodies were used. Detection was done by WesternBright TM ECL-spray (Thermo Scientific) and imaging with Molecular Imager ChemiDoc XRS + with ImageLab (Bio-Rad).
Quantitative PCR and siRNA RNA was extracted by NucleoSpinâ RNA extraction kit (Macherey-Nagel #740955) and reverse transcribed by Maxima first strand cDNA synthesis kit (Thermo Fischer). RT-qPCR was performed using DyNAmo Flash SYBR Green qPCR Kit (Thermo Scientific) with specific primers for ITPR1, ITPR2, ITPR3, and GAPDH using Bio-Rad CFX Maestro 1.1 software (Bio-Rad). ITPR3 siRNA knockdown experiments were conducted as describedusing ITPR3 ON-TARGETplus SMARTpool siRNA (Dharmacon #L-006209-00-0005) and ON-TARGETplus Non-targeting siRNA (Dharmacon #D-001810-01-05).
## Fibroblast ca 2+ imaging
We performed fibroblast Ca 2+ imaging by two different methods in two different laboratories: non-ratiometric manual Ca 2+ assay using ATP stimulation (performed in University of Helsinki), and ratiometric automated Ca 2+ assay using ionomycin, thapsigargin, and bradykinin stimulation (performed in KU Leuven).
## Non-ratiometric manual ca 2+ assay
Cells were washed two times with Hank's Balanced Salt Solution (HBSS) (in mM: 130 NaCl, 2.5 KCl, 1.8 CaCl 2 , 1.2 MgCl 2 , 10 HEPES, pH 7.4) and incubated for 50 min in dark in room temperature with 5 µg/ml Fluo-4 AM Ca 2+ indicator. After incubation, cells were washed three times with HBSS. For imaging, coverslips were placed on MatTek glass bottom dish (MatTek #P35G-1.5-14-C) and imaged with Zeiss Axio Observer Z1 inverted phase contrast fluorescence microscope. During experiment, cells were perfused with Multichannel systems PPS2 Peristaltic perfusion system. Before every experiment, cells were allowed to rest for 5 min in the stage under HBSS perfusion. The excitation light was filtered through 494 nm band pass filter and the emission light passed through a 506 nm band pass filter. Emission wavelength was captured by Photometrics Prime BSI sCMOS Camera with ZEISS ZEN 3.1 (blue edition) imaging software. Acquisition protocol lasted a total of 12 minutes under constant perfusion at a rate of 2 ml/min. The cells were first perfused with HBSS containing 1.8 mM Ca 2+ for 3 min and then with HBSS solution containing no Ca 2+ (0-Ca 2+ HBSS) for another 3 min. Then Ca 2+ release from ER was evoked by perfusion with 0-Ca 2+ HBSS containing 80 µmol/L adenosine 5'-triphosphate (ATP) magnesium salt (Sigma-Aldrich #A9187) for 2 min. After ATP-evoked Ca 2+ response, perfusion solution was changed back to 0-Ca 2+ HBSS for 2 min and finally back to 1.8 mmol/L Ca 2+ containing HBSS solution for another 2 min to induce store-operated Ca 2+ entry (SOCE) as a positive control.
The results were analyzed using MatLab (MATLAB R2019b) and RStudio (version 1.2.5033). Regions of interests (cells) were masked from each experiment and the mean pixel intensity was measured at each time point (frame) using a modified version of the previously described MatLab script.The baseline (F 0 ) was selected from the first 0-Ca 2+ period, from frames with stable intensity values. Relative intensities were calculated by first subtracting the baseline value from each frame and then dividing it by the baseline value [DF t /F 0 =(F t -F 0 )/ F 0 ]. To analyze the kinetics of the Ca 2+ response peaks, we created an R-script, which allows automatic analysis of the peaks. The area under the curve (AUC), peak amplitude and time to peak were measured from each cell, and the averages per coverslip were calculated. The scripts used are available online (https://github.com/Julius Ronkko/Ca2-analysis).
## Ratiometric automated ca 2+ assay
Cytosolic Ca 2+ levels of fibroblasts seeded in 96-well plates (Greiner) were monitored using ratiometric fluorescent Ca 2+ indicator dye Fura-2 AM (Eurogentec, Belgium). They were loaded with the Fura2-AM (1 µmol/L) at RT for 30 min in a modified Krebs solution (in mM: 150 NaCl, 5.9 KCl, 1.2 MgCl 2 , 11.6 HEPES (pH 7.3), and 1.5 CaCl 2 ). After loading, cells were rested for 30 min at RT in the absence of Fura-2 AM to allow complete dye de-esterification before proceeding to analysis on a FlexStation 3 microplate reader (Molecular Devices, Sunnyvale, CA, USA). The Ca 2+ indicator was alternately excited at 340 and 380 nm and emission of fluorescence at 510 nm recorded. EGTA was added after 30 seconds in all conditions at a final concentration of 3 mmol/L. After another 60 seconds, cells were exposed to stimuli prepared in Ca 2+ -free modified Krebs solution containing 3 mmol/L EGTA and Ca 2+ transients were monitored for 6 min. Ionomycin and the irreversible SERCA-inhibitor thapsigargin were both added at a final concentration of 10 µmol/L and bradykinin at a final concentration of 50 nmol/L. All traces are shown as the ratio of both emission wavelengths F 340 /F 380 and were smoothened using a running average of 5. For quantification purposes, a baseline value was determined for each measurement as mean fluorescence between 30 sec and 90 sec. Fluorescence ratio F 340 /F 380 was then normalized to the baseline values, and AUC of the peak, the peak amplitude (both analyzed between 90 and 450 sec) and the time to the peak were measured.
# Results
## Autosomal dominant cmt family
The proband of the Finnish family (P1) first came to neurologic investigations at age 38. His first symptoms had 1964 ª 2020 The Authors. Annals of Clinical and Translational Neurology published by Wiley Periodicals LLC on behalf of American Neurological Association started slightly before age 30 with weakened foot dorsiflexion and tendency to foot drop. Around age 33 he also started to experience increased clumsiness in his hands. He was last evaluated at age 64. His symptoms have been slowly progressive with increasing difficulty in walking in rough terrain. He uses supportive insoles but has remained ambulant without external aids. He has severe muscle atrophy in his lower legs, hammer toes, pes cavus, and thenar muscle atrophy. Sensation to light pressure was decreased distally from wrists and in foot soles. Vibration sense was decreased at ankles but present at wrists. Foot dorsiflexion or plantarflexion did not overcome gravity. Deep tendon reflexes were absent. His other diseases were hypothyroidism, hypercholesterolemia, and severe obstructive sleep apnea.
The proband had two brothers, one of whom had no neuropathic symptoms, while the other (P3) had pes cavus and progressive distal muscle weakness and wasting. His diseased father (P4) had been diagnosed with hereditary neuropathy of unknown cause but was otherwise relatively healthy and lived to an age of 93 years. He had remained ambulant until that age. The daughter of the proband (P2) first came to neurologic examination at age 35. She had had hammertoes and other deformities of the small bones of the feet since childhood, which had been operated on first at age 27. Despite this, she was ambulant without aids and able to play sports. She was able to walk on toes and heels. Marked pes cavus and hammertoes were noted, in addition to mild impairment of vibration sense at the right ankle.
Nerve conduction studies (NCS) of P1, P2, and P3 were consistent with demyelinating neuropathy, which were graded at least mild. The reduction in NCV was clearly less than is typical for CMT1A. In addition, there was variable degrees of axonal neuropathy, which tended to worsen with age. Biopsies from P1 had been obtained at age 38. In sural nerve biopsy, a clear hypertrophic neuropathy with prominent onion bulbs was seen. Muscle biopsy from tibialis anterior muscle showed prominent small group atrophy, fiber type grouping, and secondary myopathic change.
## Single affected individual
The proband of the Ashkenazi Jewish family (P5) was developing normally until 4 years of age when he began falling. An evaluation at that time was concerning for pes cavus with hammertoes, and a motor only NCS showed a small tibial CMAP (0.5 mV) with a demyelinating range conduction velocity (20 m/sec). The median CMAP was normal (4.0 mV) but the NCV was 35 m/sec. No temporal dispersion or conduction block was present in either nerve. He was diagnosed with a demyelinating CMT and over the intervening decade he experienced progressive loss of leg strength and sensation in a symmetric distal to proximal gradient. After this evaluation he was not evaluated by a neurologist until 16 years of age when he established care at Columbia University Irving Medical Center (CUIMC).
On neurological examination at age 16, he had pes planus with complete loss of muscle below the knees, and atrophy of proximal leg and hand intrinsic muscles. There was no movement at the ankles, anti-gravity strength at the knees, and mild weakness of hip flexion and the hand intrinsics. Temperature, pin prick, and vibration were severely reduced at the toes and ankles, but remarkably, joint position sense was preserved. All reflexes were absent. Otherwise, general examination, cognitive evaluation, and neurological examination were normal. The proband was the product of a consanguineous union (parents are second cousins). His father is known to have flat feet and mildly reduced distal sensation in his early 40's but no demonstrable weakness. His mother's examination was normal. Of the proband's 11 full siblings, none are suspected by the family to have similar symptoms but have not been formally tested.
Repeat EMG/NCS showed absent motor and sensory responses in the legs. The median and ulnar nerves showed normal CMAP amplitudes and slowed conduction velocities (32 m/sec), with mildly reduced SNAP amplitudes and conduction velocities of 38 m/sec.
## Genetic findings
In P1, we first excluded PMP22 duplication, MFN2 and GJB1 mutations. After this we performed exome sequencing on P1 and P3. We filtered the exome data for (1) shared nonsynonymous changes (excluding in frame insertions/deletions), (2) absence in SNP database of which c.1843G>A (p.Val615Met) in exon 16 of ITPR3 (NM_002224.4) was of interest. We confirmed the segregation of the variant with disease in the family by Sanger sequencing. The other identified variants had no previous studied function in Schwann cells or suggested association with CMT.
In P5, a chromosomal analysis and SNP microarray were within normal limits though the SNP microarray showed long contiguous runs of homozygosity consistent with the history of consanguinity. A CMT gene panel that included sequencing and deletion/duplication analysis of 42 genes identified single missense variants of uncertain significance in each of two autosomal recessive CMT genes (IGHMBP2 and NDRG1) without other coding or copy number variants in those genes. A clinical exome was reported as negative. Trio sequence data were subsequently analyzed with an updated version of the Institute for Genomic Medicine's established trio sequencing frameworkBoth ITPR3 variants are predicted to be deleterious and damaging to protein structure and/or function based on in silico analyses (damaging by SIFT, probably damaging by PolyPhen2). The affected amino acids are highly conserved. We assessed the affected residues using published IP 3 R structures.p.Val615 is located in the armadillo repeat domain (ARM1), just after the IP 3 -binding domain (running from aa 1 to~aa 600). p.Arg2524 is located in the transmembrane domain (TMD) and lies in the channel pore.
## Protein and mrna
Skin fibroblasts of P1 and P2 were available for study. We assessed the levels of IP 3 R proteins and corresponding mRNAs. IP 3 R3 protein level was decreased in the skin fibroblasts of P2 but not in P1, as compared with controls. However, P1 fibroblasts had significantly elevated ITPR3 mRNA level as compared with controls, which suggests a compensatory mRNA upregulation to preserve the normal IP 3 R3 protein level. IP 3 R2 protein level was also decreased in P2, while those of IP 3 R1 were unaffected. Also, ITPR1 and ITPR2 mRNA levels were unchanged in both patient lines.
## Fibroblast ca 2+ flux
Next, we performed siRNA knockdown of ITPR3 in control fibroblasts. The knockdown was confirmed by western blot and led to altered Ca 2+ flux dynamics in response to GPCR agonist ATP, with delayed peak of response but no change in amplitude or area under the curve (AUC) . This experiment confirmed that loss of IP 3 R3 produces a detectable phenotype in fibroblasts.
After this, we analyzed Ca 2+ homeostasis in fibroblasts from P1 and P2 and healthy controls. The response to GPCR agonist ATP, which results in IP 3 signaling and thus opening of IP 3 Rs, was analyzed by manual non-ratiometric assay in the absence of extracellular Ca 2+ . The ATP-evoked Ca 2+ release was altered in both patient fibroblasts. P1 fibroblasts had statistically significant decrease in AUC, while P2 fibroblasts had increased time to peak . We performed additional Ca 2+ -signaling analyses in cell populations using ratiometric automated technique. In these experiments, cells were first exposed to EGTA, an extracellular Ca 2+ buffer. The response to the Ca 2+ ionophore ionomycin, which provides an estimate of the total intracellular Ca 2+ content, was decreased in P1 fibroblasts . The SERCA inhibitor thapsigargin, which gives an estimate of ER Ca 2+ store content, did not cause statistically significant changes in the patient fibroblasts . Finally, the alternative GPCR agonist bradykinin evoked smaller cytosolic Ca 2+ transients in P1 and P2 fibroblasts, the latter having significantly lower peak amplitude . Overall, the results suggest that the p.Val615Met mutation in ITPR3 affects Ca 2+ homeostasis and IP 3 -mediated Ca 2+ release.
# Discussion
This study provides genetic and functional evidence for ITPR3 as a dominant CMT disease gene. We describe two new mutations: p.Val615Met in adult onset and p.Arg2524Cys in childhood onset CMT. The reduction in median motor NCV in these patients was consistent with demyelinating neuropathy, which was also confirmed by nerve biopsy in one patient. However, the magnitude of reduction was less severe than is usually observed in CMT1A (OMIM #118220), which is the most common form of CMT.Our patients' NCV was in the 30-45 m/s range suggesting it should be considered in those with "intermediate CMT". 2 However, in CMT, conduction velocities vary by nerve, disease duration and patient, thus additional patients will be required to more conclusively define whether this is a demyelinating or intermediate CMT. Axonal involvement in our patients tended to become worse with age, which suggests that the axonal degeneration was secondary to demyelination.
Both of the identified mutations affect highly conserved residues. Being located in the central, modulatory region of IP 3 R3, shortly after the ligand-binding domain, 24 the p.Val615Met variant might influence IP 3 R3 activity, for example, through interfering with allosteric regulators.The p.Arg2524Cys variant, which localizes in the channel pore, may affect the channel properties and/or the ion flux directly, thus accounting for the earlier onset, faster progression, and more severe phenotype in this patient. The previously reported variant of unknown significance, p.Thr1424Met, 4 was in a patient who similarly to our patients had moderately decreased median motor NCV of 34.7 m/s. Onset was at age 40, and two additional individuals in the same family were similarly affected.The p.Thr1424Met variant localizes in the armadillo repeat domain 2 near the subunit contact site. It may therefore affect oligomerization of the channel. Finally the variant p.Met1064Val, previously found in one index case of hereditary neuropathy,In the second method (C-E), we measured Ca 2+ in single wells of a 96well plate, using 10 lmol/L ionomycin, 10 lmol/L thapsigargin or 50 nmol/L bradykinin to evoke responses, and compared patient cells to one control line performing five independent experiments in each setting. All stimuli (added after 90 sec, 2nd dotted line) were added in the presence of EGTA (added after 30 sec, 1st dotted line). (C) Traces showed a decrease in ionomycin-induced Ca 2+ -transients for both patients compared to the healthy control, with a significant decrease in peak amplitude for P2. (D) In response to SERCA inhibitor thapsigargin, patient fibroblasts did not display statistically significant decrease in Ca 2+ ER store content compared to the healthy control. (E) In response to bradykinin, we observed a significant decrease in the peak amplitude of the response in P2 fibroblasts. All results are presented as mean AE SEM of independent experiments and statistical comparisons performed with one-way ANOVA. (*P < 0.05, **P < 0.01, ***P < 0.001). dominant missense variants may have different molecular effects on IP 3 R3 function, which also influences disease severity.
Our measurements of Ca 2+ flux demonstrate altered Ca 2+ homeostasis in p.Val615Met patient cells. The primary fibroblasts express ITPR3 and thus are a useful tool to study the effects of the mutation under physiological conditions. The weakness of this model is that it does not account for possible neuron-or Schwann cell specific effects of the mutation. In addition, the results may be influenced by other genetic differences between control and patient cells in addition to the ITPR3 mutation. Treatment with ITPR3 siRNA was used to confirm that IP 3 R3 is active in fibroblasts under normal conditions and to model the effect of loss-of-function of the channel. We found altered GPCR agonist responses in both patient lines. The slowed response in P2 fibroblasts was similar but less pronounced than in siRNA-treated cells. In addition, the P1 cells showed a decreased amplitude of the Ca 2+ response to ATP. The differences in the Ca 2+ responses between the two patient cell lines may be related to reduction in the amounts of IP 3 R2 and IP 3 R3 in P2 cells, which were compensated by ITPR3 mRNA upregulation in P1 cells. In addition, the difference in sex or other possible genetic differences may in part account for the differences between P1 and P2 fibroblasts. Based on these results, the p.Val615Met variant may produce a dominant negative effect on channel function. The effect appears to be subtle, which is consistent with late onset and slowly progressive nature of our patients' phenotype. As ionomycin and thapsigargin-induced Ca 2+ release tended to be lower in patient fibroblasts, it cannot be excluded that IP 3 R3 p.Val615Met is leaky, that is, has an increased likelihood of being open compared with the wild type situation, thereby lowering steady state ER Ca 2+ levels and thus dampening Ca 2+ release in the cytosol upon agonist exposure.
Our results suggest an important role of IP 3 R3 in peripheral nerve maintenance. This is supported by its localization in paranodal regions of rat Schwann cells, where its proximity to another CMT gene product, gap junction protein beta-1 (GJB1), may allow swift propagation of Ca 2+ signals from cell to cell.Abnormal Ca 2+ flux could contribute to altered axonal Ca 2+ microdomains that disturb mitochondrial transport, as has been suggested for dominant mutations in the plasma membrane cation channel transient receptor potential cation channel, subfamily V, member 4 (TRPV4) 28-31 which cause CMT2C. Furthermore, mutations in other components of the IP 3 signaling pathway, for example, FIG4 and SBF2, cause demyelinating CMT,which highlights the importance of this pathway for peripheral myelin maintenance. Finally, IP 3 R3 has important implications for regulation of mitochondrial function and cell death and survival by participating in Ca 2+ transfer between ER and mitochondria,a process which is also dependent on another important CMT gene, mitofusin 2 (MFN2). 38 IP 3 R3 defects may decrease mitochondrial Ca 2+ and predispose to defective bioenergetic or ER membrane function, which have been found in other forms of CMT.In conclusion, our results provide further evidence that ITPR3 is a disease gene for CMT. Additional studies, ideally in neuronal or animal models, will be needed to elucidate the effects of the disease variants on IP 3 R3 function and evaluate the potential of targeting Ca 2+ flux as a therapeutic target in CMT.
## Supporting information
Additional supporting information may be found online in the Supporting Information section at the end of the article.. Sequencing primers used for Sanger sequencing ITPR3 DNA and cDNA. . Primers used for quantitative reverse transcription PCR of ITPR1, ITPR2, ITPR3, and GAPDH. . Filtering of exome sequencing data of P1 and P3 left nine variants. The variants were analyzed further in silico. The variants not found in gnomAD were Sanger sequenced in all family members. The analysis left ITPR3 as a gene of interest. |
Using Microdosing to Induct Patients Into a Long-Acting Injectable Buprenorphine Depot Medication in Low Threshold Community Settings: A Case Study
Healthcare innovation has never been more important as it is now when the world is facing up to the unprecedented challenges brought by the COVID-19 pandemic. Within addictions services in Scotland, the priority has been to tackle our rising drug related death rate by maintaining and improving access to treatment while protecting frontline workers and managing operational challenges as a result of the pandemic. We present here a case study of five patients with opioid use disorder whose treatment represents a confluence of three important Medication Assisted Treatment (MAT) service innovations. The first was a low threshold drop in and outreach MAT service to rapidly and safely initiate opiate replacement therapy (ORT). The second was the provision of a microdosing regimen to enable same day induction to oral buprenorphine while minimizing the risk of precipitated opioid withdrawals and/or treatment disengagement. The third was rapid transitioning to an injectable long-acting buprenorphine depot which reduced unnecessary face to face patient contact and treatment non-adherence. This case study of five patients highlights the valuable role that buprenorphine microdosing can play in making induction to long-acting buprenorphine depot feasible to a broader range of patients, including those on a high dose methadone treatment regime.
# Introduction
Scotland has the highest per capita Drug Related Death (DRD) rate in Europe, approaching that of the United States, with opioids implicated in 86% of cases. The Drug Deaths Task Force (DDTF) (Scottishwas created by the Scottish Government to stem this rising trend. A key DDTF priority has been to support service innovations which improve access and availability of Medication Assisted Treatment (MAT) for People Who Use Drugs (PWUD). Innovative, flexible and responsive MAT has become even more important during the COVID-19 pandemic, as people who use opioids and other drugs have heightened health and social risks increasing their vulnerability to poor outcomes (EMCDDA, 2020).
## Background pwud health and social issues
Scotland has an ageing cohort of older drug users (over 40 years old), who, through long drug use careers, experience accelerated metabolic ageing and an earlier onset of cardiovascular and respiratory disease . In particular, high nicotine and cannabis smoking rates, and the use of inhaled heroin and crack cocaine make PWUD particularly vulnerable to the respiratory and cardiovascular complications of . Many PWUD therefore come into the category of people needing to shield for a prolonged period of time during pandemic peaks which has implications also for their ongoing mental health . Opioids contributed to 86% of DRD in Scotland in 2018-nearly always alongside other drugs and/or alcohol (Scottish.
There is a higher incidence of Hepatitis C Virus (HCV) and Human Immunodeficiency Virus (HIV) among people who inject drugs (PWID)and an increased likelihood of this rising during the pandemic . This is due to a combination of riskier drug use as usual supplies dry up and reduced access to injecting equipment, blood borne virus testing and MAT as already beleaguered addiction servicesface additional challenges from the pandemic such as maintaining continuity of care while protecting frontline workers.
Finally, PWUD often experience greater exclusion and isolation, family separation, unstable housing or homelessness and imprisonment. These factors alongside an increased risk of withdrawals in the absence of access to MAT mean that PWUD would struggle to practice social distancing, self-isolation or shielding advice (EMCDDA, 2020), with significant implications for both their own and public health .
## Mat service innovations during the pandemic
## Low threshold and assertive outreach mat service
In March of 2020, in response to COVID-19 and the first United Kingdom wide lockdown, we initiated an assertive outreach and low threshold drop-in program to enable people with opioid dependence to access evidence based treatment such as buprenorphine and methadone rapidly. In keeping with the literature around low threshold MAT, the service provided same-day treatment entry and prescribing where appropriate, a harm reduction approach, flexibility in terms of appointments, dispensing and re-initiation if a visit was missed. Many of the patients captured by this clinical intervention were older (over 45 years old) with Chronic Obstructive Airway disease (COPD) and other co-morbidities such as HIV or HCV. Many also had unstable housing or were street homeless and some had never been in treatment before.
## Buprenorphine microdosing to enable same day induction onto oral buprenorphine
While the United Kingdom has both methadone and buprenorphine medications as first line options for MAT , there may be an advantage to the latter in terms of its safety profile, although this needs to be balanced against the patient's own preference and consequent concordance. Buprenorphine is a partial μ-opioid agonist, with high receptor affinity and a ceiling effect on respiratory depression. This results in an effective, longacting treatment for opioid dependency which may be safer in those with compromised respiratory function for example, due to COPD and/or poly-substance use. Our older patients already with increased risk of both chronic heart disease as well as COVID-19 may be more vulnerable to the cardiovascular adverse effects associated with high dose methadone such as QTc prolongation .
Due to buprenorphine's strong binding affinity for the μ receptor which supersedes that of the majority of full μ agonists, introducing it in opioid dependent patients who are not in withdrawal can induce this intensely unpleasant state. To avoid this happening, current guidance requires the patient to be in moderate withdrawal (Clinical Opioid Withdrawal Scale greater than 13 (Independent Expert Working Group, 2017)) before taking their first dose. If the patient has been taking short acting opioids, this often means abstinence for 12-24 h and 48-72 h for long-acting drugs such as methadone (Independent Expert Working Group, 2017). When switching from high dose methadone, the requirement is more stringent, with prior tapering to 30 mg or less daily and a cessation of at least 36 h before induction. This is a simple process to understand, but difficult for the patient to do and it is associated with destabilization due to an often prolonged methadone taper.
Microdosing, also known as the Bernese method, is the practice of administering minute doses of buprenorphine to obtain benefit from its action with minimal side effects. It was first described in a case report in 2016and proved the pharmacological hypothesis that administering small amounts of buprenorphine to an opioid dependent person who is comfortable on their drug of choice, does not precipitate opioid withdrawal. Further, due to its relatively long half-life, buprenorphine gradually accumulates at the opioid receptors ultimately replacing the μ-agonist enabling the patient to cease its use. As a result, this method is particularly useful where:
- Patients have failed or refused a conventional induction due to the inability to tolerate moderate withdrawals and/or for whom opioid withdrawals would be harmful for example when presenting with poor physical or mental health or pregnancy (The College of Physicians and Surgeons of Manitoba, 2020)outlines some of these regimens ranging from a 7-11-days induction period. In Canada, where most of these regimens originate, it is common practice to use buprenorphine/ naloxone combinations which are quartered or halved to make up the smaller initial doses. Some of the twice daily regimens involve the patient having a supervised dose earlier in the day and a takeaway for later in the day.
## Transitioning to an injectable long-acting buprenorphine depot
In August 2019, a depot buprenorphine formulation allowing weekly or monthly subcutaneous injections was approved for use by the Scottish Medicines Consortium for the management of OUD (Scottish Medicines Consortium, 2019). Early on in the onset of the pandemic, the Scottish Government and the Victorian Government in Australia were at the forefront in identifying the potential benefits of making depot buprenorphine more readily available for high risk groups (Scottish.
The projected benefits of depot buprenorphine included increased protection of frontline workers and patients seeking MAT from droplet spread of COVID-19 by reducing daily or frequent attendances in pharmacies, enabling people who have been asked by the government to self-isolate or shield to be able to do so, reducing the impact upon patients of pharmacies being closed due to illness or quarantine. Further, depot buprenorphine also negates the risks such as diversion or overdose implicit in allowing larger amounts of takeaway controlled drugs to minimize unnecessary travel, and the efficacy of treatment will no longer be dependent on the patients adherence to daily dosing, resulting in less risk of overdose and withdrawals.
In order to benefit from the buprenorphine depot however, patients need to go through a similar induction processes as for the oral formulation. For example, a public hospital in Victoria, Australia, has launched the first rapid access clinic for depot buprenorphine and suggests that people in need of a transfer from methadone to buprenorphine-based treatment may require admission to a residential withdrawal unit. Certainly, in our setting, places in such units are hard to come by, costly, and have been suspended since the pandemic started. The Scottish Government produced a document recommending the use of depot buprenorphine in prisons to provide effective protection against withdrawals while also protecting staff and patients from exposure to . However, significant numbers of opioid dependent patients in prison are on methadone, and broad acceptance of depot buprenorphine may be limited by the expectation that they should cease their full agonist in order to In microdosing method 1 and 5, the patient tapers down on their full agonist on day 7 to a full stop by day 11. In methods 2,3 & 4, cessation of the full agonist should happen on day 7. go into moderate withdrawals before being given their first injection (Scottish . Similarly, we encountered patients in our clinic who were keen to have depot buprenorphine but would not have tolerated conventional induction. We therefore present a case study of five patients who were identified through our low threshold intervention who were inducted unto a long-acting buprenorphine depot through a microdosing regimen.
## Case definition
This case study consists of consecutive patients attending an assertive outreach service between March and October 2020, with a confirmed history of opioid dependence who wished to commence depot buprenorphine for whom conventional induction precautions were not or unlikely to be tolerated. Further, all the included individuals went through a tailored microdosing bridging schedule onto an adequate sublingual buprenorphine dose up to the day before the depot formulation was administered. Excluded were individuals who transitioned onto depot buprenorphine via conventional means as described in the manufacturers product information (The electronic medicines compendium, 2018), or individuals who completed a microdosing schedule in order to remain on sublingual buprenorphine, even if they opted for the depot later on in their treatment.provides an overview of the five patients seen for microdosing and induction of Buvidal.
## Cases
Case 1 Referred to Here as John (Male, 36 Years Old) had a Long History of IVDU, From the Age of 14 years
He is HIV positive and struggled at different times with alcohol dependence, crack cocaine, heroin and illicit benzodiazepine use. Through the years, John had been on methadone and buprenorphine and managed to stabilize for periods of time, but inevitably struggled with attendances at the pharmacy. He also became criminally involved when intoxicated. He selffunded a naltrexone implant, a treatment modality not offered in Scotland which helped him stay of opioids for around 1 month. He seemed to feel an effect from heroin use which made him wonder if the implant had been inauthentic.
John was on methadone 75 mg daily and was keen to convert to depot buprenorphine so as to cease regular pharmacy pick-ups. We started John on the 14-days regimen with at home microdosing with regular telephone support. John was provided with 15 × 0.4 mg, 9 × 2 mg and 4 × 8 mg sublingual buprenorphine tablets and clearly color-coded instructions. We agreed that John would reduce his methadone to 70 mg at the outset and would then taper down further on his methadone doses once he was on 4 mg of buprenorphine. On his eighth day, we began to taper down by 10 mg daily and he ceased all methadone when he got to 16 mg buprenorphine. John managed the regimen with no issues, and on the 14th day we administered the buprenorphine depot as a 96 mg monthly dose.
## Case 2 & 3 referred to here as derek (male, 45 years old) and eleanor (female, 51 years old)
Derek and Eleanor are a married couple. Eleanor often disengaged from OST when pharmacy attendances interfered with her employment. Derek was entrenched within the local drug-using scene and had been criminally involved. When the pandemic started, the couple was required to shield for three months due to underlying COPD. They were both finding that their substance use was having a significant impact on their respiratory function and wanted to stop. Eleanor opted for depot buprenorphine first while Derek was more dubious. Both were concerned about the risk of precipitated withdrawals, something they had experienced in the past. Eleanor struggled somewhat to understand the microdosing instructions, especially differences in the tablet strengths. We dispensed the 0.4, 2 and 8 mg tablets to her only when they were due to be initiated. With appropriate pandemic related precautions, we administered a 96 mg s/c monthly depot buprenorphine injection at her home once she settled on a 16 mg s/L dose. Eleanor was pleased with the outcome of her treatment and her experience encouraged Derek to undergo the same process. We started his microdosing regime and scheduled Derek's first injection of the same dose on the day of Eleanor's second injection. Follow up reviews of the couple have been positive.
## Case 4 & 5 referred to here as george (male, 42 years old) and harry (male, 46 years old)
George and Harry have been a couple for over 3 years. Both are HIV positive on anti-retroviral medications. George had a much longer history of IV heroin use, and a long and varied treatment experience which included periods on prescribed methadone, buprenorphine and also two periods in an abstinence-based recovery program. George also had a 20-years history of benzodiazepine use, initially through a prescription, but latterly from the illicit market. Harry had a much shorter history of IV heroin use and has always needed George to inject him. Harry had never been on any form of MAT. Both attended for treatment at the same time when George was discharged from his rehabilitation program due to the pandemic. They opted to be seen together and requested methadone. Unfortunately, after two non-fatal overdoses, it was clear that methadone was not reducing their risk.
We discussed buprenorphine and both were concerned that they would struggle with concordance. Also, as George was on 80 mg of methadone, he was concerned that he would not manage the associated withdrawals of conventional induction. We needed to specifically counsel Harry around the reduced efficacy of opiate analgesics should he need to attend the accident and emergency (A&E) department with renal colic which he sometimes suffered. He was reassured however by our standard practice of placing a medical alert in shared records about patients being on depot buprenorphine. In his case, should he attend A&E in acute pain, hospital care staff will recognize that he will need larger doses of opioid analgesics or alternate analgesics altogether.
Both patients stabilized on 24 mg of s/l buprenorphine which translated into 128 mg of monthly the depot which was administered in the clinic. We were particularly concerned about George and Harry's illicit benzodiazepine use. It was unrealistic to expect them not to take some benzodiazepines, especially for George who had a twenty-year history of dependence on these. There is a known risk of benzodiazepines reducing the ceiling effect of buprenorphine, such that combining these with other drugs may result in an overdose. Once this happens, higher naloxone doses would be required due to the high receptor affinity that buprenorphine has. We agreed on a maintenance dose of 20 mg a day of diazepam on an interval dispensing regime to support them in trying to avoid the far more potent illicit benzodiazepines known to be circulating in the local market. Also, they were provided with further naloxone kits and ongoing support. On the latest review, both have fully ceased IV drug use and have managed to avoid illicit benzodiazepines.
# Discussion
As a confluence of three service delivery innovations, this case study is an example of a nimble response to unprecedented challenges to addiction services. Also, to our knowledge, this is the first case study describing the use of a microdosing regimen to induct patients with opioid use disorder unto a long-acting buprenorphine depot. This study is limited by its retrospective case study design, the absence of a comparison group, short duration of follow up, and a lack of objective outcome measures, such as systematic urine results. Furthermore, assessment of withdrawal was by clinician impression and patient self-report as opposed to a formal Clinical Opiate Withdrawal Score (COWS). This is partly as, in a less than ideal setting of a time limited outreach visit, formally completing a COWS can be challenging. Nevertheless, objective measures such as these would have made crosscomparisons across different clinical settings more feasible.
Furthermore, each of the three innovations came with its challenges. For example, while a low threshold assertive outreach model improves access to marginalized groups, robust clinical governance must be in place to ensure the patient's medical and medications history is known before a prescription is initiated and to avoid duplicate prescribing of controlled drugs. Inevitably, there will be times that a systems failure results in delays in MAT initiation and consequently, the possibility of patient disengagement.
While buprenorphine microdosing has clear advantages in over-coming potential delays inherent in traditional induction on the day or patient presentation, it is important to note that it cannot be recommended as an equivalent alternative to current standard practice due to the lack of high level evidence such as randomized controlled trials. There have however been case reports and substantial practical experience with this method in Canada, Germany and locally in the South East and the West of Scotland. The result is a broad range of regimens with no consensus on optimum dosing.
Conventional induction and stabilization unto buprenorphine is attainable within 2-3 days provided the patient is able to tolerate withdrawals. Microdosing can take 7-14 days with the patient continuing to use illicit opiates as required. Microdosing therefore increases immediate accessibility to buprenorphine but prolongs the patient's risk exposure to illicit drug use by several days. Finally, in North America particularly, the use of buprenorphine/naloxone combinations are favored over buprenorphine on its own This is primarily to avoid situations of diversion or misuse of buprenorphine for example through snorting or injecting it. The consequence of this is that the smaller doses within a microdosing regime (less than 2 mg) usually consists of portions of buprenorphine/naloxone tablets. These tablets are used sublingually and so are friable, meaning that the dosing accuracy is likely to be variable. Our strategy has been to use buprenorphine sublingual tablets available in 200 and 400 mcg doses. This has allowed us to provide more accurate dosing and simpler patient instructions.
While the pandemic highlighted the advantages of depot treatment in terms of reducing the risks of exposure to COVID-19, there are definite limitations that need to be considered. These include the need for nursing or medical staff to administer the injection (Scottish Medicines Consortium, 2019), the significantly higher costs (20% higher than oral formulations, 16 times more expensive than methadone solution) (Scottish Public Health Observatory, 2019), the lack of generic products to compensate for potential supply chain interruptions and limited clinician and patient experiences with its use. Further, the consumption of large amounts of potent street benzodiazepines, a particularly worrying issue among PWUD in Scotland, reduces the threshold of the ceiling effect of buprenorphine, removing the protection patients normally have against respiratory depression (Independent Expert Working Group, 2017). Once administered, the depot injection dose cannot be removed and the long duration of action of buprenorphine magnified by its prolonged release formulation will limit the effectiveness of naloxone. Unfortunately, the extent to which this scenario is likely to increase patient risk is as yet poorly understood.
Buprenorphine makes up 19% of MAT for opioid use disorder in Scotland, with the remainder being primarily methadone (Scottish Public Health Observatory, 2019). Clinical experience in Scotland is that patients tend to opt for methadone, possibly as this is what they are more familiar with. During the pandemic, and within the context of our outreach model of care, it was often clinically safer to encourage the use of buprenorphine. It may be that with the introduction of the depot, more flexibility in induction through microdosing and increased patient education as to its favorable safety profile, buprenorphine may become increasingly more common.
What we have been able to demonstrate is a range of clinical scenarios where microdosing has been effective in inducting patients onto depot buprenorphine enabling them to gain from the benefits of this treatment at a crucial time. We have also been able to administer depot buprenorphine injections to patients in their homes, supporting them to adhere to government advice on shielding. Notably, some of our patients sometimes found the microdosing regimen confusing and starter packs or pre-prepared dosette boxes of buprenorphine tablets used in some settingscould be a helpful addition to our practice.
Issues which need to be better understood include a cost benefit and sustainability analysis based on a larger number of cases. Specifically, will investment in the more expensive depot buprenorphine injection reduce the available resources to provide care for the growing number of people who use drugs needing treatment? Further, what are the implications for the patient when their treatment is distilled into a monthly visit for an injection? Indeed, with an eye on the adaptations we undertook to provide ongoing care for patients during the pandemic, it is also important to evaluate what the essential components of safe and high-quality MAT actually are. In other words, what aspects of our practice in initiating buprenorphine and methadone must be continued for the safety of our patients, and what aspects simply represent organizational dogmatism?
This last point relates to the need to develop the quality of the evidence base around microdosing. At what point do we acknowledge the successful application of clinical expertize over many years in the application of buprenorphine microdosing, almost a naturalistic clinical trial, rather than insisting on interventional randomized controlled trials? Perhaps if it is randomized controlled trials which are required, the possibility of using microdosing as a means to induct unto the relatively newly developed range of long-acting buprenorphine depot injections may provide the necessary impetus.
# Conclusion
The COVID-19 pandemic is challenging health systems throughout the world and forcing addictions services to be agile and innovative to meet the needs of PWUD while also protecting them and frontline workers from viral transmission. This study demonstrates the utility of using microdosing to facilitate the induction of patients onto depot buprenorphine in situations where conventional methods are impractical or not tolerated. Certainly, microdosing may be a more affordable and acceptable alternative to an inpatient detoxification or subjecting patients on high dose methadone to unpleasant withdrawals as is currently practiced in some settings. The lack of published literature on buprenorphine microdosing undertaken in a range of different settings is a barrier to its more widespread adoption. We propose an international collaboration to collate clinical experience and case report data and produce definitive best practice guidelines in the mainstream use of buprenorphine microdosing.
## Permission to reuse and copyright
Figures, tables, and images will be published under a Creative Commons CC-BY license and permission must be obtained for use of copyrighted material from other sources (including republished/adapted/modified/partial figures and images from the internet). It is the responsibility of the authors to acquire the licenses, to follow any citation instructions requested by thirdparty rights holders, and cover any supplementary charges. institutional requirements. The patients/participants provided their written informed consent to participate in this study AUTHOR CONTRIBUTION JW: Conceptualization, Methodology, Writing-Original draft preparation, AB: Conceptualization, Validation, Writing-Reviewing and Editing, LG: Methodology, Validation, Writing-Reviewing and Editing, CL: Investigation, Writing-Reviewing and Editing |
Prognostic Value and Clinicopathological Differences of HIFs in Colorectal Cancer: Evidence from Meta-Analysis
Background: The prognostic value of HIFs in colorectal cancer was evaluated in a large number of studies, but the conclusions were inconclusive. Meanwhile, clinicopathologic differences of HIF-1a and HIF-2a were rarely compared in recent studies.Methodology: Identical search strategies were used to search relevant literatures in the PubMed and Web of Science databases. The prognostic significances and clinicopathological differences of HIFs in CRC were analyzed.Principal Findings: A total of 23studies comprising 2984 CRC patients met the inclusion criteria. The results indicated that overexpressed HIFs were significantly associated with increase of mortality risk, including overall survival (OS) (HR 2.06 95%CI 1.55-2.74) and disease free survival (HR 2.84, 95%CI 1.87-4.31). Subgroup analysis revealed that both overexpressed HIF-1a and HIF-2a had correlations with worse prognosis. The pooled HRs were 2.01 (95% CI: 1.55-2.6) and 2.07(95% CI: 1.01-4.26). Further subgroup analysis on HIF-1a was performed by study location, number of patients, quality score and cutoff value. The results showed that HIF-1a overexpression was significantly associated with poor OS, particularly in Asian countries (HR 2.3, 95% CI: 1.74-3.01), while not in European or other countries. In addition, overexpression of HIF-1a was closely related with these clinicopathological features, including Dukes' stages (OR 0.39, 95% CI: 0.17-0.89), UICC stages (OR 0.42 95% CI: 0.3-0.59), depth of invasion (OR 0.71, 95% CI: 0.51-0.99), lymphnode status (OR 0.49, 95% CI: 0.32-0.73) and metastasis (OR 0.29, 95% CI: 0.11-0.81). While overexpression of HIF-2a was only associated with grade of differentiation (OR 0.48, 95% CI: 0.29-0.81).Conclusions: This study showed that both HIF-1a and HIF-2a overexpression were associated with an unfavorable prognosis. HIF-1a overexpression seemed to be associated with worse prognosis in Asian countries. Additionally, HIF-1a and HIF-2a indicated distinct clinicopathologic features.
# Introduction
Colorectal cancer is the third most common malignancy worldwide, and one of the leading causes of cancer-related deaths [bib_ref] Cancer statistics, Siegel [/bib_ref]. An increasing trend in the incidence of this carcinoma has been noticed in the Asian nations. Despite recent therapeutic advances, its 5-year survival rate is still pessimistic due to its recurrence and drug resistance [bib_ref] Cancer statistics, Jemal [/bib_ref]. Growing evidence suggests that hypoxia plays a pivotal role in disease progression and therapy resistance in most solid tumors, including colorectal cancer [bib_ref] Hypoxia-inducible factor-1alpha expression in the gastric carcinogenesis sequence and its prognostic role..., Griffiths [/bib_ref] [bib_ref] HIF1A overexpression is associated with poor prognosis in a cohort of 731..., Baba [/bib_ref]. Rapid oxygen consumption and aberrant tumor angiogenesis and blood flow result in a hypoxic tumor environment. Owing to the fundamental importance of oxygen for metabolism and survival, cells have evolved intricate response mechanisms to respond to hypoxia. The most important regulators mediating the primary transcriptional responses to hypoxic stress are hypoxia-inducible factors (HIFs). Given that hypoxia promotes tumor progression and therapy resistance, HIFs are expected to be useful biomarkers associated with progress disease and poor prognosis in CRC. Increased expression of HIFs has also been observed in a broad range of human cancer cell types, and has been associated with poor prognosis in many cases [bib_ref] Prognostic impact of hypoxia-inducible factors 1alpha and 2alpha in colorectal cancer patients:..., Yoshimura [/bib_ref] , but the prognostic value of HIFs for CRC patients is inconclusive.
HIFs are heterodimers composed of an inducible a-subunit (HIF-1a, HIF-2a or HIF-3a), and a constitutive HIF-1b subunit (also known as aryl hydrocarbon receptor nuclear translocator or ARNT), which together form the HIF-1, HIF-2 and HIF-3 transcriptional complexes, respectively. Of the three HIF family members, HIF-1 and HIF-2 are the most well-characterized. HIF-1a and HIF-2a are usually detected to measure tumor oxygen levels because the HIF-1b subunit is constitutive. HIF-1a and HIF-2a have 48% amino acid sequence identity and similar protein structures, but distinct target genes and mechanisms of regulation. HIF-1a preferentially induces glycolytic pathway, whereas HIF-2a regulates genes involved in tumor growth, cell cycle and maintaining stem cell pluripotency [bib_ref] Differential roles of hypoxia-inducible factor 1alpha (HIF-1alpha) and HIF-2alpha in hypoxic gene..., Hu [/bib_ref]. Thus, HIF1a and HIF2a can promote highly divergent, even opposing, outcomes, which results in distinct clinicopathologic features and prognosis. Multiple xenograft tumour models also support the hypothesis that HIF1a and HIF2a play different roles in tumor progression by regulating both shared and unique target genes [bib_ref] HIF-2alpha deletion promotes Kras-driven lung tumor development, Mazumdar [/bib_ref]. However, clinicopathologic and prognostic differences of HIF1a and HIF2a in CRC were rarely compared in recent studies.
Therefore, we made a meta-analysis from eligible studies to investigate the relationship between HIF expression and prognosis of CRC patients. Meanwhile, we performed a subgroup analysis to assess the roles of HIF-1a and HIF-2a in clinicopathologic features and prognosis of CRC.
# Materials and methods
## Identification and eligibility of relevant studies
We searched literature from PubMed, WanFang and Web of Science databases using the terms: ''HIF'', ''colorectal neoplasms'', ''colorectal Cancer'', ''colon cancer'' ''rectal cancer'', ''prognosis'' with all possible combinations. Bibliographies, review articles and other pertinent studies were searched manually for additional eligible studies.
The inclusion criteria for eligibility of a study in the metaanalysis were as follows: (1) evaluating HIF expression in the human CRC tissues; (2) assessing the relationships between HIFs expression with CRC clinicopathologic features or prognosis; [bib_ref] Hypoxia-inducible factor-1alpha expression in the gastric carcinogenesis sequence and its prognostic role..., Griffiths [/bib_ref] articles written in English or Chinese; (4) sufficient information provided to estimate hazard ratio (HR) about overall survival (OS) or disease free survival (DFS), or to estimate odds ratio (OR) about clinicopathologic features. In addition, letters, reviews, conference abstracts, and case reports were not in the scope of our analysis because of the limited data. Overlapping articles were also excluded from this meta-analysis, only the most recent or the most complete study was involved in the analysis.
## Data extraction and management
Two investigators (Xin He and Wenjie Xia) reviewed each eligible study independently and extracted data from all the publications meeting the inclusion criteria. Controversial problems were arbitrated by the third investigator (Jinhong Xu). The following information was collected from each study: the first author's name, year of publication, country of origin, number of patients, gender of patients, HIF isoforms, source and dilution of antibody, cut-off value, tumor characteristics, condition of adjuvant therapy and survival data.
# Methodological assessment
Newcastle-Ottawa quality assessment scale (NOS) was used to assess the quality of each study [bib_ref] Critical evaluation of the Newcastle-Ottawa scale for the assessment of the quality..., Stang [/bib_ref]. The score assessed eight items of methodology, categorized into three dimensions including selection, comparability, and outcome. A maximum of 1 score was awarded for each item with the exception of the item related to comparability that allowed the assignment of two scores. A total of 0 and 9 scores were respectively designated as lowest and highest quality, and the studies with 6 scores or more were graded as the high quality ones in the scale. The scores provided by two researchers were compared and a consensus value for each item was achieved.
# Statistical methods
For the pooled analysis of the impact of HIF expression on survival outcome, HRs and its 95% CI were used. If these Heterogeneity in between-study was assessed by Chi-square based Q statistical test [bib_ref] Systematic reviews on rehabilitation interventions, Handoll [/bib_ref]. And the I 2 statistic to quantify the proportion of the total variation, which is due to inter-study heterogeneity rather than sampling error and is measured from 0% to 100% [bib_ref] Uncertainty in heterogeneity estimates in meta-analyses, Ioannidis [/bib_ref]. Higher values indicate a greater degree of heterogeneity. When the studies were found to be homogeneous(with P.0.10 for the Q test), the pooled ORs and HRs estimate of each study were calculated by the fixed-effects model (the Mantel-Haenszel method) [bib_ref] Statistical aspects of the analysis of data from retrospective studies of disease, Mantel [/bib_ref]. Otherwise, we chose the random-effects model (the DerSimonian and Laird method) [bib_ref] Measuring inconsistency in meta-analyses, Higgins [/bib_ref]. We assessed the possibility of publication bias by visually assessing a funnel plot for asymmetry and by quantitatively performing Egger's test. Publication bias was indicated when p value of Egger's test ,0.05. The meta-analysis was performed using STATA version 12.0 software (Stata Corporation, Collage Station, Texas, USA). All the P values were for a two-side test and considered statistically significant when p,0.05.
# Results
## Description of studies
As shown in [fig_ref] Figure 1: Flow diagram of study selection procedure [/fig_ref] , 227 published records were identified from a search of the above databases using the search strategy as described above. After exclusion of the studies that were out of the scope of our systematic review, a total of 23 eligible studies were included in the final meta-analysis [bib_ref] HIF1A overexpression is associated with poor prognosis in a cohort of 731..., Baba [/bib_ref] [bib_ref] Prognostic impact of hypoxia-inducible factors 1alpha and 2alpha in colorectal cancer patients:..., Yoshimura [/bib_ref] [bib_ref] Low expression of prolyl hydroxylase 2 is associated with tumor grade and..., Xie [/bib_ref] [bib_ref] Carbonic anhydrase IX, hypoxia-inducible factor-1alpha, ezrin and glucose transporter-1 as predictors of..., Korkeila [/bib_ref] [bib_ref] Expression of hypoxia-inducible factor 1alpha predicts clinical outcome after preoperative hyperthermo-chemoradiotherapy for..., Shioya [/bib_ref] [bib_ref] Pretreatment HIF-1alpha and GLUT-1 expressions do not correlate with outcome after preoperative..., Havelund [/bib_ref] [bib_ref] Clinicopathological significance of p53, hypoxia-inducible factor 1alpha, and vascular endothelial growth factor..., Kwon [/bib_ref] [bib_ref] Clinical significance of CD133 and hypoxia inducible factor-1alpha gene expression in rectal..., Saigusa [/bib_ref] [bib_ref] Clinicopathologic significance of HIF-1alpha, CXCR4, and VEGF expression in colon cancer, Wu [/bib_ref] [bib_ref] Hypoxia-inducible factor-1alpha modulates the down-regulation of the homeodomain protein CDX2 in colorectal..., Zheng [/bib_ref] [bib_ref] Expression of HIF-1alpha and VEGF in colorectal cancer: association with clinical outcomes..., Cao [/bib_ref] [bib_ref] Expression of delta-like ligand 4 (Dll4) and markers of hypoxia in colon..., Jubb [/bib_ref] [bib_ref] Expression patterns of hypoxic markers at the invasive margin of colorectal cancers..., Rajaganeshan [/bib_ref] [bib_ref] Combined analysis of hypoxia-inducible factor 1 alpha and metallothionein indicates an aggressive..., Schmitz [/bib_ref] [bib_ref] Hypoxiainducible factor-1alpha and -2alpha are expressed in most rectal cancers but only..., Rasheed [/bib_ref] [bib_ref] Stromal expression of hypoxia regulated proteins is an adverse prognostic factor in..., Cleven [/bib_ref] [bib_ref] Clinical significance of immunohistochemical expression of hypoxia-inducible factor-1alpha as a prognostic marker..., Lu [/bib_ref] [bib_ref] Hypoxia, angiogenesis and apoptosis markers in locally advanced rectal cancer, Theodoropoulos [/bib_ref] [bib_ref] Expression of hypoxia-inducible factor-1alpha is associated with tumor vascularization in human colorectal..., Kuwai [/bib_ref] [bib_ref] EPAS1 mRNA in plasma from colorectal cancer patients is associated with poor..., Mohammed [/bib_ref]. Of these 23 publications, 20 studies assessed the relationships between HIF-1 expression with CRC clinicopathologic features or prognosis, while 6 studies evaluated the association of HIF-2 expression and CRC pathological features or prognosis. The clinical features of these 23 included studies were summarized in [fig_ref] Table 1: Characteristics of studies included in the meta-analysis [/fig_ref]. These studies were published from 2003 to 2013, and total 2984 CRC patients were enrolled. Sample sizes ranged from 30 to 731 patients (mean 130). 14 of these studies enrolled less than 100 patients and 9 studies included more than 100 patients. 6 of these studies evaluated patients from China, 5 from Japan, 3 from England, others from America, Korea, Finland, Germany, Austrialia, Holand and Greece. 19 of these studies got 6 scores or more in methodological assessment, which meant they had high qualities. Impact of HIFs expression on overall survival and disease free survival of colorectal cancer
The meta-analysis was performed on 15 studies assessing the association of HIFs expression with OS. The pooled HR was 2.06 (95%CI 1.55-2.74; I 2 69.1%) [fig_ref] Figure 2: Forrest plot of Hazard ratio [/fig_ref]. Nine studies evaluating the correlation of HIFs expression with DFS were all about HIF-1a. The pooled HR was 2.84 (95%CI 1.87-4.31; I 2 41%) [fig_ref] Figure 2: Forrest plot of Hazard ratio [/fig_ref]. It suggested that overexpressed HIF was significantly associated with increase of mortality risk. In addition, sensitive analysis was performed. We removed one study at a time and evaluated the rest, pooled HR of HIFs overexpression on OS ranged from 1.98(95% CI: 1.5-2.61) to 2.28(95% CI: 1.74-2.98) [fig_ref] Table 2: HRs [/fig_ref] , and combined HR of HIFs overexpression on DFS ranged from 2.34(95% CI: 1.68-3.26)to 3.22(95% CI: 2.08-4.99) [fig_ref] Table 3: HRs [/fig_ref]. We also performed subgroup analysis about association of HIFs expression with OS by HIF isoforms, the results showed that both HIF-1a and HIF-2a were associated with worse prognosis. The pooled HR was 2.01 (95% CI: 1.55-2.6, I 2 33.1%) and 2.07(95% CI: 1.01-4.26, I 2 86.1%) respectively [fig_ref] Figure 2: Forrest plot of Hazard ratio [/fig_ref]. Subgroup analysis about association of different subcellular localization of HIFs expression with OS was per-formed, and the results showed that the correlation was not changed no matter where HIF located in (nucleus or cytoplasm). The pooled HR was 2.456 (95% CI: 1.694-3.561, I 2 49.2%) and 2.049(95% CI: 1.519-2.764, I 2 0%), respectively [fig_ref] Figure 3: Forrest plot of Hazard ratio [/fig_ref].
Moreover, further subgroup analysis on HIF-1a was performed by study location, number of patients, antibody dilution, cut-off value. Subgroup analysis indicated a significant relation between HIF-1a overexpression and OS was exhibited in Asian countries (HR 2.3, 95% CI: 1.74-3.01, I 2 0%). Other factors comprising number of patients, antibody dilution and cut-off value did not alter the significant OS of overexpressed HIF-1a [fig_ref] Table 4: Stratified analysis of pooled hazard ratios for colorectal cancer patients with overexpressed... [/fig_ref].
## Correlation of hifs expression with clinicopathological parameters
The meta-analysis was also assessed the correlation between HIF-1a expression and clinicopathological characteristics of CRC. As shown in [fig_ref] Table 5: HIF-1a and HIF-2a expression and clinicopathological features for colorectal cancer [/fig_ref] Furthermore, there was no significant association between HIF-1a expression with grade of differentiation. The pooled OR was 0.97 (95% CI: 0.67-1.39). .
In addition, we evaluated the correlation between HIF-2a overexpression with clinicopathological characteristics of CRC. The result showed that overexpression of HIF-2a was significantly associated with grade of differentiation (OR 0.48, 95% CI: 0.29-0.81). There was no significant association between HIF-2a expression with Dukes' stages, depth of invasion and lymphnode status. The pooled OR was 0.91(95% CI: 0.20-4.17), 0.38 (95% CI: 0.04-3.80), and 0.95 (95% CI: 0.428-2.16), respectively [fig_ref] Table 5: HIF-1a and HIF-2a expression and clinicopathological features for colorectal cancer [/fig_ref].
## Publication bias
Egger's test indicated that there was no evidence of significant publication bias after assessing the funnel plot [fig_ref] Figure 1: Flow diagram of study selection procedure [/fig_ref] for the studies included in our meta-analysis.
# Discussion
Hypoxia has been recognized as a common feature of solid tumors and a negative prognostic factor for response to treatment and survival of cancer patients. In 1993, Höckel reported that cervix cancer patients with hypoxic tumors (median pO 2 ,10 mmHg) had a significantly lower overall and recurrence-free survival [bib_ref] Intratumoral pO2 predicts survival in advanced cancer of the uterine cervix, Hockel [/bib_ref]. Since then, hypoxia has been found to indicate a highly aggressive disease phenotype associated with poor prognosis in many cancers, including brain, breast, prostate, pancreas, cervix, bladder and ovary [bib_ref] Expression of hypoxia-inducible factor-1 alpha in oligodendrogliomas: its impact on prognosis and..., Birner [/bib_ref] [bib_ref] Levels of hypoxia-inducible factor-1alpha independently predict prognosis in patients with lymph node..., Bos [/bib_ref] [bib_ref] Prognostic significance of HIF-1 alpha overexpression in human pancreatic cancer, Shibaji [/bib_ref] [bib_ref] Hypoxia-inducible factor 1 alpha expression correlates with angiogenesis and unfavorable prognosis in..., Theodoropoulos [/bib_ref]. HIFs are the best characterized markers mediating transcriptional responses to hypoxic stress and expected to be unfavorable prognostic indicators. Hypoxia and consequently HIF activation is regarded as an important stimulus of CRC angiogenesis. HIF binds to the HRE in the VEGF promoter region, leading to up-regulation of VEGF transcription and the formation of new blood vessels [bib_ref] Hypoxia-inducible factor 1alpha protein expression is controlled by oxygen-regulated ubiquitination that is..., Sutter [/bib_ref]. Surprisingly, both HIF-1 and HIF-2 can function as tumor suppressors in certain cancers [bib_ref] The expression of hypoxia-inducible factor 1alpha is a favorable independent prognostic factor..., Lidgren [/bib_ref]. Many studies were also performed to assess the prognostic value of HIF for CRC patients, but the conclusions were also inconclusive. On the other hand, HIF-1 and HIF-2 have distinct target genes, but few studies compared the clinicopathologic and prognostic differences between HIF-1 and HIF-2.
This meta-analysis aimed to examine the association between HIFs expression and the prognosis of CRC patients, and assess the roles of HIF-1a and HIF-2a in clinicopathologic features. Our analysis combined the outcomes of 23 studies comprising 2984 CRC patients, indicating that overexpressed HIF was significantly associated with increase of mortality risk, including OS (2.06 95%CI 1.55-2.74; Z = 4.95; P = 0.000) and DFS (2.84,, 95%CI 1.87-4.31; Z = 4.92; P = 0.000). Additionally, the results of sensitivity analysis showed that the association was not changed after removing any study. Subgroup analysis revealed that both overexpressed HIF-1a and HIF-2a were associated with worse prognosis in CRC. On the basis of different HIF isoforms, further subgroup analysis was performed by study location, number of patients, antibody dilution, cut-off value. HIF-1a overexpression was significantly associated with poor OS in Asian countries (HR 2.3, 95% CI: 1.74-3.01, Z = 5.76, P = 0.000), while not in European or other countries. It indicated that HIF-1a overexpression seemed to be associated with disease progress and unfavorable prognosis in Asian CRC patients. Other factors did not alter the significant OS of overexpressed HIF-1a. In addition, significant correlations were observed between HIF-1a overexpression with clinicopathological features including Dukes' stages, UICC stages, depth of invasion, lymphnode status and metastasis. Our results concurred with previous study that HIF-1a expression had a significant inverse correlation in T1 and T2 CRC. On the other hand, overexpression of HIF-2a was significantly associated with grade of differentiation. Thus, HIF-1 and HIF-2 indicate distinct clinicopathologic features.
In this meta-analysis, we had dealt with highly significant heterogeneity among the 23 studies. Although we used random effects models to analyze the data, it did not identify the source of heterogeneity. Thus, we performed stratified analysis according to study location, number of patients, cut-off value. When the analysis on OS was performed without consideration of other factors, heterogeneity was detected (I 2 69.1% P = 0.000). While when the studies included were classified into three groups according to evaluation criterion (percentage, staining and percentage plus staining), the heterogeneity was not detected (I 2 0% P = 0.621, I 2 0% P = 0.624, I 2 0% P = 0.727). Therefore, the heterogeneity in this study could be explained by the evaluation standards. Meanwhile, there were some limitations in this metaanalysis. First, the study included in our meta-analysis was restricted only to articles published in English or Chinese, which probably provided additional bias. Second, HRs calculated from data or extracted from survival curves might be less reliable than direct analysis of variance. Third, the sample size in European studies was not big enough so that the difference of HIF-1expression on survival was not significant.
In summary, we showed that both overexpressed HIF-1a and HIF-2a were significantly associated with worse prognosis in CRC. Subgroup analysis indicated that HIF-1a overexpression was associated with progress disease and unfavorable prognosis in Asian CRC patients. Significant correlations were observed between HIF-1a overexpression with Dukes' stages, UICC stages, depth of invasion, lymphnode status and metastasis, but there was no significant association between overexpressed HIF-1a with grade of differentiation. While overexpressed HIF-2a was only associated with differentiation. Large, well-designed prospective studies are required to investigate the precise prognostic significance and clinicopathologic differences of HIFs expression. [fig_ref] Figure 1: Flow diagram of study selection procedure [/fig_ref] Egger's publication bias plot showed no publication bias for studies regarding overexpressed HIF-1a and overall survival (OS) in the meta-analysis: the relationship between the effect size of individual studies (HR, vertical axis) and the precision of the study estimate (standard error, horizontal axis).
## Supporting information
[fig] Figure 1: Flow diagram of study selection procedure. doi:10.1371/journal.pone.0080337.g001 [/fig]
[fig] Figure 2: Forrest plot of Hazard ratio (HR) for the association of different HIF isoforms expression with overall survival (OS) and disease free survival (DFS). A. HRs with corresponding 95% CIs of the HIFs expression with OS. B. HRs with corresponding 95% CIs of the HIFs expression with DFS. HR.1 implied worse survival for the group with increased HIFs/negative expression and overexpressed HIFs was significantly with the worse prognosis of CRC patients. doi:10.1371/journal.pone.0080337.g002 [/fig]
[fig] Figure 3: Forrest plot of Hazard ratio (HR) for the association of HIF in different subcellular localization with overall survival (OS). doi:10.1371/journal.pone.0080337.g003 [/fig]
[fig] Figure: S2 Egger's publication bias plot showed no publication bias for studies regarding overexpressed HIF-1a and disease free survival (DFS) in the metaanalysis. (TIF) S3 Egger's publication bias plot showed no publication bias for studies regarding overexpressed HIF-2a and overall survival (OS) in the meta-analysis. (TIF) Checklist S1 PRISMA Checklist. (DOC) Author Contributions [/fig]
[table] Table 1: Characteristics of studies included in the meta-analysis. In brief, if the trials offered the data such as log-rank test p values, number of total events. The number of aberrant HIF expression and number of preserved HIF expression were extracted to allow estimation of the HR and its 95% CI. If only Kaplan Meier graphs were published, Kaplan-Meier curves were read by Engauge Digitizer version4.1 (http://digitizer.sourceforge. net/). Time-to-event data from the Kaplan-Meier curves was extracted and HR and its 95% CI were calculated via SPSS16.0. Odds ratios (ORs) and their 95%CIs were combined to evaluate the association between HIF expression and clinicopathological factors, such as differentiation grade, Dukes' stages, depth of invasion, lymphnode status and metastasis. An observed HR.1 implies worse survival for the group with overexpressed/negative HIF expression. An observed OR,1 implies unfavorable parameters for the group with overexpressed/negative HIF expression. The impact of overexpressed/negative HIF expression on survival or clinicopathological factors was considered to be statistically significant if the 95%CI did not overlap with 1. [/table]
[table] Table 2: HRs (95% CI) of sensitivity analysis for HIFs overexpression on OS. [/table]
[table] Table 3: HRs (95% CI) of sensitivity analysis for HIFs overexpression on DFS. [/table]
[table] Table 4: Stratified analysis of pooled hazard ratios for colorectal cancer patients with overexpressed HIF-1a. [/table]
[table] Table 5: HIF-1a and HIF-2a expression and clinicopathological features for colorectal cancer. REM, random-effects model; FEM, fixed-effects model; OR, odds ratio; CI, confidence interval. doi:10.1371/journal.pone.0080337.t005 [/table]
|
Reirradiation of pulmonary artery intimal sarcoma: A case report
This is an open access article under the terms of the Creat ive Commo ns Attri butio n-NonCo mmercial License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited and is not used for commercial purposes.AbstractWe report here the case of a patient with 4 years long-term survival after treatment with surgery, chemotherapy, and radiotherapy with good local control. This case highlights the possible role of radiation therapy in this tumor.K E Y W O R D Scase report, helical tomotherapy, intensity modulated RT, pulmonary artery sarcoma, reirradiation
# | introduction
Intimal sarcoma of the pulmonary artery is a rare tumor with poor oncological outcomes. We report here the case of a patient with 4 years long-term survival after treatment with surgery, chemotherapy, and radiotherapy with good local control. This case highlights the possible role of radiation therapy in this tumor.
Primary pulmonary artery sarcomas (PAS) are the most frequent sarcoma of the great arteries. Most PAS are derived from the embryologic bulbus cordis area, mainly from the pulmonary artery. Intimal sarcoma of the pulmonary artery (PAIS) is a rare tumor with an estimated incidence between 0.001% and 0.003%, 1 which is likely an underestimate due to frequent misdiagnosis and late detection during surgery or autopsy.Median overall survival (OS) of approximately 17 months is reported in the literature for PAIS patients.Surgery remains the mainstay of management for patients with PAIS, even though the use of various chemotherapy agents has been reported.To date, the role of radiotherapy remains unclear.Below, we report the case of a woman with long-term survival from PAIS who was initially treated with surgical resection and postoperative chemotherapy, which was successfully treated with repeated (fractionated) courses of radiotherapy.
## | case report
The patient, a 62-year-old woman, reported a history of pulmonary hypertension since March 2010 with findings of a high D-dimer, shortness of breath, and chest pain. In April 2014, during follow-up for her pulmonary hypertension, chest computed tomography (CT) revealed the presence of a solid mass originating from the main pulmonary artery with a maximal extension of 37 mm and involvement of the pulmonary valve but without signs of extravascular invasion. The subsequent 18-fluorodeoxyglucose positron emission tomography (18 FDG PET/CT) confirmed chest CT findings with pathologic uptake at the pulmonary trunk. In May 2014, she underwent pulmonary endarterectomy (PEA) and paratracheal lymphadenectomy. The definitive histological diagnosis was poorly differentiated mesenchymal tumor with strong fibroblastic differentiation at | 1343
[formula] CHIOLA et AL. [/formula]
immunohistochemistry, consistent with pulmonary arterial intimal sarcoma (KI67 30%, G3, mitotic index >20, necrosis <50%); negative nodes (0/6) but positive margins (right and left margin of pulmonary artery and the margin of ventricular). Due to the negative CT re-evaluation 1 month after surgery, the patient received four cycles of adjuvant chemotherapy combining adriamycin and ifosfamide between June and September 2014. The forth chemotherapy cycle was complicated by neurological toxicity and a diagnosis of encephalopathy, which resolved completely. From October to December 2014, the patient underwent adjuvant radiation treatment, which was delivered by Hi-Art helical tomotherapy and daily image guided radiotherapy (IGRT). A CT simulation scan with 2.5 mm slices was acquired, and the following volumes were identified: clinical target volume (CTV), including the surgical areas of pulmonary artery; planning target volume (PTV), defined adding 5 mm to CTV; organs at risk (OARs), such as spinal cord, esophagus, heart, and right and left lungs. The prescribed dose was 60 Gy in 30 fractions, five fractions per week . The radiotherapy treatment was well tolerated with no relevant toxicity. At the end of the radiotherapy, the patient started a regular clinical and radiological follow-up. In May 2016, 19 months after external beam radiotherapy (EBRT), a chest CT revealed a solid nodule of 7 × 6 mm located nearby the inferior wall of the pulmonary artery descending branch. The lesion was suspicious for recurrence, but was too small to be defined. Close radiological follow-up with repeated chest CTs in July 2016, October 2016, and January 2017 revealed stability in nodule size. In July 2017, the CT scan revealed a pulmonary descending artery filling defect; however, the defect increased in size three times compared to its original size. No evidence of distant metastases was observed. From September to November 2017, the patient underwent second-line chemotherapy (Docetaxel and Gemcitabina-4 cycles). This regimen was interrupted because of disease progression at CT re-evaluation (lesion of 30 mm in maximum size that involved the whole lumen of the inferior right pulmonary artery). Particle radiations treatment was evaluated at Protontherapy Center, Pavia, but the site of the lesion was not compatible and photon reirradiation concomitant to pazopanib was proposed. This second radiotherapy treatment (reRT) was delivered in February 2018 by VMAT technique with LINAC (VARIAN Trilogy). The re-RT dose prescribed was 24 Gy in 4 fractions, once a week, with energy of 6 MV. The CTV included F I G U R E 1 Dose distribution of helical tomotherapy plan the macroscopic disease plus a 5 mm margin . Like the first one, this EBRT treatment was well tolerated. An early CT performed 1 month after reRT revealed a partial response with maximum lesion size reduction from 3 to 2 cm. In June 2018, a further CT scan showed stable disease. In view of the lack of change and reasonable tolerability, the patient will continue Pazopanib treatment.
# | discussion
Primary pulmonary artery sarcomas (PAS) are the most frequent sarcomas of the great arteries characterized by local growth, with only a slight ability to metastasize. 2 They typically affect middle-aged people, particularly women, with a mean age at diagnosis of 48 years.Our patient had an older age of onset than is generally reported in literature. The aim of this paper is to report our experiences with a patient who had long-term (50 month) survival with PAS, considering that this disease has a poor prognosis and a median OS of 17 months.Diagnosis of PAS is difficult and often delayed due to the nonspecific nature of the symptoms. It is often confused with pulmonary thromboembolism and is therefore treated inappropriately with prolonged anticoagulation or thrombolysis.Common symptoms at the time of presentation are progressive dyspnea, cough, chest pain, and weight loss. The patient described here presented with all the above symptoms but had been diagnosed with pulmonary hypertension 4 years prior. This misdiagnosis is quite common as PAS starts with nonspecific symptoms and it can be easily confused with pulmonary embolism; however, the onset of the symptoms is usually more gradual with PAS than pulmonary embolism.In our patient, this missed diagnosis led to a delay in therapy that could have affected outcomes. Surgical treatment remains the mainstay of management for patients with PAIS and can include pulmonary endarterectomy, lobectomy, or pneumonectomy depending on mono or bilateral disease.The correct surgical approach must be evaluated individually, according to the tumor presentation, the presence of pulmonary hypertension, and the patient's clinical condition.Surgery can prolong survival and often improve symptoms but margin status, like in this case, is rarely clear. For this reason, the patient underwent adjuvant treatment.Chemotherapy is normally given postoperatively, although cases of improved outcome with neoadjuvant chemotherapy have been described. It is not clear if adjuvant chemo and radiotherapy bring any improvement, but some data reported an increased OS in patients who received multimodality therapy vs patients who received one single therapy 9 and our patient received both chemo and radiotherapy. Several perioperative agents have been reported in the literature, including anthracyclines, ifosfamide, gemcitabine, taxanes, platinums, and immunotherapy. Anthracyclines, either alone or in combination, are the most commonly used agents. In our case, first-line chemotherapy was adriamycin and ifosfamide.The role of postoperative radiotherapy remains unclear.Chemotherapy regimens are normally followed by intensity modulated RT (IMRT) at a radical dose of 60 Gy in 30 fractions to surgical bed as happened in our patient.Helical tomotherapy allows the delivery of higher doses to target volumes along with better sparing of normal tissues and to perform daily IGRT permitting us to correct any possible setup errors at every application, increasing the precision of the treatment and reducing interfraction changes.In our case, the patient had a disease-free survival and was symptom-free for 19-months after first EBRT. At recurrence, the patient was clearly refractory to the second chemotherapy therapeutic regimens and presented with disease progression. For this reason, a second RT course concomitant to new targeted therapy was proposed. This re-RT was delivered with a weekly fractionation in order to exploit a radiobiological rationale designed to increase the therapeutic index.The only targeted agent approved for use in soft tissue sarcoma at present is the tyrosine kinase inhibitor pazopanib, based on the results of the PALETTE trial.Pazopanib should be considered in patients with intimal sarcoma of the pulmonary artery that is unresectable or recurrent after surgery or cytotoxic chemotherapy.This case suggests that the combination of RT and pazopanib is safe and well tolerated and that disease recurrence may well respond precociously (reevaluation after 1 month) and effectively to RT.
# | conclusion
In this case, we can confirm the beneficial role of postoperative radiotherapy and the safety of reirradiation on residual recurrence. Early diagnosis could be an essential prerequisite to optimal management of PAS, especially by a multidisciplinary team. Being a rare and radioresistant cancer, more studies are still necessary to better understand the role of adjuvant treatment and the role and timing of RT, which, as this case suggests, can be delivered more than once when disease recurs.
CONFLICT OF INTERESTNone declared.AUTHOR CONTRIBUTIONSIL, LB: contributed equally to this work, wrote the paper, and reviewed final manuscript; EMV, MG: review and editing of final manuscript, provided insight into how results relate to medical physic; RC: review and editing of final manuscript, provided insight into how results relate to radiation oncology.INFORMED CONSENTConsent was obtained from patient for the publication of this report and any accompanying images.ORCIDLiliana Belgioiahttps://orcid.org/0000-0003-0508-6344 |
Impact of Interspecific Hybridization between Crops and Weedy Relatives on the Evolution of Flowering Time in Weedy Phenotypes
Background: Like conventional crops, some GM cultivars may readily hybridize with their wild or weedy relatives. The progressive introgression of transgenes into wild or weedy populations thus appears inevitable, and we are now faced with the challenge of determining the possible evolutionary effects of these transgenes. The aim of this study was to gain insight into the impact of interspecific hybridization between transgenic plants and weedy relatives on the evolution of the weedy phenotype.Methodology/Principal Findings: Experimental populations of weedy birdseed rape (Brassica rapa) and transgenic rapeseed (B. napus) were grown under glasshouse conditions. Hybridization opportunities with transgenic plants and phenotypic traits (including phenological, morphological and reproductive traits) were measured for each weedy individual. We show that weedy individuals that flowered later and for longer periods were more likely to receive transgenic pollen from crops and weed6crop hybrids. Because stem diameter is correlated with flowering time, plants with wider stems were also more likely to be pollinated by transgenic plants. We also show that the weedy plants with the highest probability of hybridization had the lowest fecundity.Conclusion/Significance: Our results suggest that weeds flowering late and for long periods are less fit because they have a higher probability of hybridizing with crops or weed6crop hybrids. This may result in counter-selection against this subset of weed phenotypes, and a shorter earlier flowering period. It is noteworthy that this potential evolution in flowering time does not depend on the presence of the transgene in the crop. Evolution in flowering time may even be counter-balanced by positive selection acting on the transgene if the latter was positively associated with maternal genes promoting late flowering and long flowering periods. Unfortunately, we could not verify this association in the present experiment.
# Introduction
When transgenic plants were initially developed, most plant evolutionary biologists and geneticists considered spontaneous hybridization between species to be rare and of little importance in terms of evolution. This view extended to both crops and their wild or weedy relatives [bib_ref] Current knowledge of gene flow in plants: implications for transgene flow, Ellstrand [/bib_ref] , but has now radically changed. More than twenty years of gene-flow research has shown that interspecific hybridization is very common in some groups of vascular plants [bib_ref] Distribution of spontaneous plant hybrids, Ellstrand [/bib_ref] [bib_ref] Plant hybridization, Rieseberg [/bib_ref] and may be of considerable evolutionary significance. Hybridization may occasionally result in the extinction of a population [bib_ref] Current knowledge of gene flow in plants: implications for transgene flow, Ellstrand [/bib_ref] [bib_ref] The evolution of California's wild radish has resulted in the extinction of..., Hedge [/bib_ref] , may trigger the evolution of plant invasiveness [bib_ref] Hybridization as a stimulus for the evolution of invasiveness in plants, Ellstrand [/bib_ref] , or initiate speciation [bib_ref] Plant Invasions, Interspecific Hybridization and the Evolution of New Plant Taxa, Abbott [/bib_ref] [bib_ref] Hybrid origins of plant species, Rieseberg [/bib_ref]. A substantial body of evidence [bib_ref] Gene flow and introgression from domesticated plants into their wild relatives, Ellstrand [/bib_ref] [bib_ref] Transgene introgression from genetically modified crops to their wild relatives, Stewart [/bib_ref] has now accumulated, demonstrating the high potential for interspecific hybridization between agricultural crops and their wild or weedy relatives. Transgenic crops are no exception, and empirical studies have provided evidence of transgene dispersal from GM crops to their weedy relatives [bib_ref] Pollen flow between herbicide-resistant Brassica napus is the cause of multiple-resistant B-napus..., Hall [/bib_ref] [bib_ref] Transgenic DNA introgressed into traditional maize landraces in Oaxaca, Mexico, Quist [/bib_ref] [bib_ref] Transgenic Brassica napus fields and Brassica rapa weeds in Quebec: sympatry and..., Simard [/bib_ref] [bib_ref] Unwanted Transgenes Re-Discovered in Oaxacan Maize, Snow [/bib_ref] [bib_ref] Transgenes in Mexican maize: molecular evidence and methodological considerations for GMO detection..., Pineyro-Nelson [/bib_ref].
Many factors have been shown to influence the rate of hybrid formation between crops and their wild or weedy relatives. Population effects such as the local densities of the parental types and their relative frequencies, have been demonstrated in several cases [bib_ref] Transgenic Brassica napus fields and Brassica rapa weeds in Quebec: sympatry and..., Simard [/bib_ref] [bib_ref] Competition affects gene flow from oilseed rape (female) to Brassica rapa (male), Johannessen [/bib_ref] [bib_ref] Competition affects the production of first backcross offspring on F-1-hybrids, Brassica napus..., Johannessen [/bib_ref] [bib_ref] Long-term introgression of crop genes into wild sunflower populations, Linder [/bib_ref] [bib_ref] Male fitness of oilseed rape (Brassica napus), weedy B-rapa and their F-1..., Pertl [/bib_ref] [bib_ref] Impact of ecological factors on the initial invasion of Bt transgenes into..., Vacher [/bib_ref]. Mating system differences at the individual level due to, for example, selfing rates and apomixis, have also been found to affect hybridization rates [bib_ref] Within-population variation in hybridisation and transgene transfer between wild Brassica rapa and..., Pallett [/bib_ref]. Moreover, several studies have shown that overlap in the flowering periods of crop and weed plants affect opportunities for hybridization [bib_ref] Male fitness of oilseed rape (Brassica napus), weedy B-rapa and their F-1..., Pertl [/bib_ref] [bib_ref] Fecundity selection in a sunflower crop-wild study: Can ecological data predict crop..., Cummings [/bib_ref].
The aim of this study is to gain insight into the impact of hybridization with transgenic crops on the evolution of the weedy relatives by [bib_ref] Current knowledge of gene flow in plants: implications for transgene flow, Ellstrand [/bib_ref] verifying that hybridization opportunities for weedy plants depend on their phenotypic traits (including flowering phenology), (2) measuring the relative fitness of hybridizing weeds, and (3) searching for associations between the transgenic trait and the phenotypic traits increasing hybridization opportunities in the offspring of weedy plants.
We studied hybridization opportunities, phenotypic traits (including phenological, morphological and reproductive traits) and offspring phenotype of weedy individuals in experimental plant populations cultivated under glasshouse conditions. Experimental populations were composed of weeds (birdseed rape, Brassica rapa L., AA, 2n = 20) and transgenic plants in a 1:1 ratio. Transgenic plants were crop plants of the Brassica genus (rapeseed, Brassica napus L. ssp oleifera, AACC, 2n = 38), F1 hybrids between B. rapa and B. napus, or first-generation backcrosses. Crop plants were all homozygous for the Btcry1Ac transgene from Bacillus thuringiensis (Bt) [bib_ref] How Bacillus thuringiensis has evolved specific toxins to colonize the insect world, Maagd [/bib_ref] , F1 hybrids were all hemizygous and first-generation backcrosses and consisted of an equal mixture of hemizygotes and null homozygotes. Hybridization opportunities for each weedy individual was calculated as the expected proportion of pollen received from transgenic plants (PPR) based on the observed flowering schedules.
This experimental system was ideal for addressing the question of interest in this study, for three reasons. First, despite barriers to interspecific mating such as apomixis [bib_ref] Within-population variation in hybridisation and transgene transfer between wild Brassica rapa and..., Pallett [/bib_ref] or preferential exclusion of hybrid zygotes [bib_ref] Preferential exclusion of hybrids in mixed pollinations between oilseed rape (Brassica napus)..., Hauser [/bib_ref] , numerous studies [bib_ref] Hybridisation within Brassica and allied genera: evaluation of potential for transgene escape, Fitzjohn [/bib_ref] have shown that B. napus and B. rapa readily hybridize under controlled conditions, but also in the field. Spontaneous hybridization has, for instance, been reported in weedy populations of B. rapa growing in agricultural crops [bib_ref] Transgenic Brassica napus fields and Brassica rapa weeds in Quebec: sympatry and..., Simard [/bib_ref] [bib_ref] Hybridization between transgenic Brassica napus L. and its wild relatives: Brassica rapa..., Warwick [/bib_ref] [bib_ref] Do escaped transgenes persist in nature? The case of an herbicide resistance..., Warwick [/bib_ref] and in natural populations of B. rapa occurring near waterways [bib_ref] Hybridization between Brassica napus and B-rapa on a national scale in the..., Wilkinson [/bib_ref]. Second, flowering time has been extensively studied in B. rapa [bib_ref] Male fitness of oilseed rape (Brassica napus), weedy B-rapa and their F-1..., Pertl [/bib_ref] [bib_ref] Genetic variation in flowering time induces phenological assortative mating: Quantitative genetic methods..., Weis [/bib_ref] [bib_ref] Phenological assortative mating in flowering plants: the nature and consequences of its..., Weis [/bib_ref] , and temporal clines in phenotypic traits have been observed. For example, time to first flowering has been shown to be positively correlated with stem height and stem diameter [bib_ref] Genetic variation in flowering time induces phenological assortative mating: Quantitative genetic methods..., Weis [/bib_ref] [bib_ref] Genetics of Brassica campestris. 1. Genetic constraints on evolution of life-history characters, Dorn [/bib_ref] [bib_ref] Rapid evolution of flowering time by an annual plant in response to..., Franks [/bib_ref]. Third, transgenic lines of B. napus containing a green fluorescent protein (GFP) gene associated with the Bt transgene have been constructed [bib_ref] Green fluorescent protein as a marker for expression of a second gene..., Harper [/bib_ref] [bib_ref] Removal of a cryptic intron and subcellular localization of green fluorescent protein..., Haseloff [/bib_ref]. The presence of the Bt transgene in the offspring of weedy plants can therefore be inferred by exposing the plants to UV light [bib_ref] Green fluorescent protein as a marker for expression of a second gene..., Harper [/bib_ref] [bib_ref] Expression of GFP and Bt transgenes in Brassica napus and hybridization with..., Halfhill [/bib_ref].
# Results
(1) Relationship between hybridization opportunities for weedy individuals, their flowering phenology, and their morphology As expected from previous results [bib_ref] Male fitness of oilseed rape (Brassica napus), weedy B-rapa and their F-1..., Pertl [/bib_ref] , weeds flowered earlier than transgenic crops and hybrids [fig_ref] Figure 1: Phenology of transgenic and weedy plants [/fig_ref] , with the F1 hybrids flowering the latest. Correspondingly, the expected proportion of crosses between weeds and F1 plants was lower than that for crop or backcross plants [fig_ref] Table 2: Effects of phenological traits on the expected proportion of pollen received from... [/fig_ref]. Moreover, PPR (log transformed) increased with the time to first flower and the duration of flowering in weedy individuals (see overall slopes in [fig_ref] Table 2: Effects of phenological traits on the expected proportion of pollen received from... [/fig_ref]. The overall slope for the interaction between the two phenological traits [fig_ref] Table 2: Effects of phenological traits on the expected proportion of pollen received from... [/fig_ref] was close to zero and did not qualitatively modify these effects. However, significant interactions [fig_ref] Table 3: Mixed linear model for the effects of transgenic type and weedy plant... [/fig_ref] indicated that the effects of phenology of weedy plants on PPR depended on transgenic type (crop, F1 hybrid or first-generation backcross). The regression coefficients and their 95% confidence limits indicated that a longer time to flowering and a longer flowering duration increased PPR more for F1 hybrids than for crops or firstgeneration backcrosses (see within-type slopes in [fig_ref] Table 2: Effects of phenological traits on the expected proportion of pollen received from... [/fig_ref]. Thus, weedy individuals flowering later and for longer periods were more likely to receive transgenic pollen, particularly if the transgenic donors were first-generation crop x weed F1 hybrids.
As expected from the results of previous studies [bib_ref] Genetic variation in flowering time induces phenological assortative mating: Quantitative genetic methods..., Weis [/bib_ref] [bib_ref] Genetics of Brassica campestris. 1. Genetic constraints on evolution of life-history characters, Dorn [/bib_ref] , we observed temporal clines in the morphological traits under study. Time to first flower was positively correlated with stem diameter (r s = 0.31, P,0.001) and stem height (r s = 0.18, P,0.05). These correlations indicate that the opportunity for hybridization may not be random, and may instead depend on the morphology of the weed. We found a significant, single effect of stem diameter on PPR (F 1,105 = 5.0, P,0.05). The overall slope was positive and its 95% confidence interval did not include zero (slope = 0.05, CL = (0.01, 0.09)), indicating that plants with large stems on the day of the first flower were more likely to hybridize with transgenic plants. No such effect was detected for stem height, either as a single effect (F 1,104 = 0.05, P = 0.82) or in interaction with transgenic type (F 2,104 = 0.87, P = 0.42).
(2) Relative fitness of hybridizing weeds For any given weedy plant in the experimental populations, the total number of filled seeds decreased significantly with PPR (see overall slopes in [fig_ref] Table 4: Effect of the expected proportion of pollen received from transgenic plants [/fig_ref] and the significant effect of PPR in [fig_ref] Table 5: Mixed linear model for the effects of transgenic type and expected proportion... [/fig_ref]. We observed no significant interaction between PPR and transgenic type [fig_ref] Table 5: Mixed linear model for the effects of transgenic type and expected proportion... [/fig_ref] , indicating that this decrease in fecundity with PPR was not dependant on transgenic type (crop, F1 hybrid or first-generation backcross). This decrease in fecundity was observed despite the positive correlation between PPR and total flower production within weedy plants (r s = 0.37, P,0.001). An alternative analysis (not shown), including transgenic type as fixed effect and phenological traits of weeds (time to first flower or flowering duration) as covariates also predicted the total number of filled seeds. We found a significant effect of the time to first flower on the number of seeds, in interaction with transgenic type (F 2,107 = 3.66, P,0.05). However, all the 95% confidence intervals of the regression coefficients for each transgenic type included zero, making further interpretation impossible. Flowering duration was significant as a single effect (F 1,105 = 37.4, P,0.001). The overall slope was negative and its 95% confidence interval did not include zero (slope = 221.3, CL = (228.77, 214.51)), indicating that weedy plants with longer flowering times produced fewer seeds. Thus, the weedy plants with the highest probability of being pollinated by Bt-transgenic plants were those with the lowest fecundity [fig_ref] Figure 2: Decrease in fecundity with PPR [/fig_ref].
(3) Associations between the transgenic trait and the phenotypic traits increasing hybridization opportunities in the offspring of weedy individuals An analysis of offspring phenotype showed that time to first flower in weedy mother plants had a significant effect on the average time to first flower of their offspring (F 1,104 = 7.48, P,0.05). Transgenic type (crop, F1 hybrid or first-generation . Phenotypic traits studied in weedy mother plants (M) and their offspring (O). In contrast, offspring stem diameter was not affected by maternal diameter (F 1,104 = 1.80, P = 0.18) or maternal time to first flower (F 1,104 = 2.03, P = 0.15). These results confirm that late-flowering plants tend to produce late-flowering offspring [bib_ref] Genetic variation in flowering time induces phenological assortative mating: Quantitative genetic methods..., Weis [/bib_ref]. Because lateflowering plants were also more likely to receive transgenic pollen, we therefore expected to find more transgenic offspring in the offspring of late-flowering weedy mothers and an association between the transgenic trait and time to first flower in the offspring generation.
Contrary to expectation, we found no evidence to suggest that weedy individuals with higher PPR produced more transgenic offspring. A total of 1648 seedlings, obtained from 126 weedy plants, were scored under UV light for the Bt-GFP construct. Only Correlations with the proportion of Bt-GFP+ seedlings were also weak and non significant for all other maternal traits measured. Thus, variation in the probability of weedy mother plants being pollinated by transgenic donors did not translate into variation in the proportion of Bt-seedlings in their offspring.
Because of the very low proportions of Bt-GFP+ seedlings, we could not study the associations between the transgenic trait and the phenotypic traits increasing hybridization opportunities in the offspring of weedy plants. Among the 1654 seedlings scored under UV light, 1048 reached the first flower stage and were measured. Unfortunately, only nine of these plants were Bt-GFP+, and seven of these nine plants were half sibs (the nine plants were produced by only three weedy mothers). The 31 remaining Bt-GFP+ seedlings did not reach the first flower stage. There were, therefore, clearly too few Bt-GFP+ plants to compare the phenotypic characteristics of Bt-GFP+ and Bt-GFPoffspring.
# Discussion
The aim of our experiment was to assess the impact of interspecific hybridization between weedy B. rapa and transgenic B. napus on the evolution of the weedy phenotype. This was done by identifying the phenotypic traits increasing hybridization opportunities for weedy individuals, searching for associations between thesephenotypic traits and the transgenic trait in the offspring of weedy mothers and evaluating the relative fitness of hybridizing weeds. Our results show that weedy individuals that flowered later and for longer periods were more likely to receive transgenic pollen from crops and weed6crop hybrids. Because stem diameter is correlated with flowering time [bib_ref] Genetic variation in flowering time induces phenological assortative mating: Quantitative genetic methods..., Weis [/bib_ref] [bib_ref] Genetics of Brassica campestris. 1. Genetic constraints on evolution of life-history characters, Dorn [/bib_ref] , plants with wider stems were also more likely to be pollinated by transgenic plants. Our results suggest that the transgene and maternal genes promoting late flowering, long flowering periods and stem thickening may be preferentially associated in the offspring of weedy mothers. However, although time to first flower is a heritable trait in B. rapa [bib_ref] Genetic variation in flowering time induces phenological assortative mating: Quantitative genetic methods..., Weis [/bib_ref] , our experiment did not confirm the gametic association between the transgene and genes promoting late-flowering in the offspring of hybridized weedy plants. Indeed, given the very small numbers of Bt-GFP+ seedlings recovered from the experimental populations, we could not study the association between the transgenic trait and other phenotypic traits in weed plant offspring.
We also found that the weedy plants with the highest probability of hybridization produced fewer seeds, despite producing larger numbers of flowers. The most straightforward interpretation of this result is that fecundity was reduced by hybrid crosses. Controlled crosses between the weedy and transgenic plants used in the experiment (unpublished results) and several previous studies [bib_ref] Fitness of hybrids between rapeseed (Brassica napus) and wild Brassica rapa in..., Allainguillaume [/bib_ref] [bib_ref] Fitness of F-1 hybrids between weedy Brassica rapa and oilseed rape (B-napus), Hauser [/bib_ref] have indeed shown that crops and weed6crop hybrids have lower siring success than do weeds. Therefore, our experiment suggests that maternal weeds that flowered late and for long periods are less fit, because they have a higher probability of hybridizing with GM crop plants or hybrids. This may result in counter-selection against this subset of weed phenotypes, and a shorter earlier flowering period. It is noteworthy that this potential evolution in flowering time does not depend on the presence of the Bt transgene in the crop, and may even be counter-balanced by positive selection acting on the transgene if the latter was positively associated with maternal genes promoting late flowering and long flowering periods. Recent experiments indeed indicate that the Bt transgene does not induce any fitness costs in hybrids between transgenic B. napus and weedy relatives [bib_ref] Growth, productivity, and competitiveness of introgressed weedy Brassica rapa hybrids selected for..., Halfhill [/bib_ref] [bib_ref] Fitness and maternal effects in hybrids formed between transgenic oilseed rape (Brassica..., Kun [/bib_ref]. It may therefore convey a selective advantage under insect herbivore pressure [bib_ref] Increased fitness of transgenic insecticidal rapeseed under insect selection pressure, Stewart [/bib_ref].
In conclusion, our analyses show that phenological differences between weedy birdseed rape and transgenic rapeseed are likely to alter the phenotypic structure of weed populations, by promoting interspecific hybridization in only a subset of weedy plants with specific phenotypes and by altering the fitness of hybridizing weeds. Unfortunately, we could not verify the non-random association between the transgenic trait and other phenotypic traits in the offspring of weedy populations because of the very low rate of transgene introgression.
# Materials and methods
## Experimental design
Nine populations, each composed of 15 Brassica rapa plants and 15 of one of three types of transgenic plants (see below) were sown as seeds and then grown from germination until death in a glasshouse at the University of California, Irvine. The nine populations were divided into three blocks, with each transgenic type replicated once per block. Plants were grown in individual ConetainerH (3.8621 cm) pots filled with a 75/25 mixture of potting soil and sand. Before planting, seeds were vernalized on wet filter paper at 4uC for 5 days. Pots were spaced 7.6 cm apart and were watered every day until 90% stopped producing flowers. An equal amount of 10:10:10 NKP liquid fertilizer was applied to each pot on the sowing date.
The three types of transgenic plants were: Bt-transgenic B. napus crop plants, Bt-transgenic B. napus 6 B. rapa F 1 hybrids, and firstgeneration backcrosses (B. rapa 6F 1 hybrids). Over 20 unique seed and 20 unique pollen parents were used to produce each of the three types. B. rapa plants served as seed parents for the F1 and backcross types. B. napus were all homozygous for the Bt-GFP insertion, whereas the F1 plants were all hemizygous. The backcross generation was expected to consist of an equal mixture of hemizygotes and null homozygotes for the insertion.
B. rapa seeds were obtained from over 400 mature plants in a population at Back Bay, near Irvine, California [bib_ref] A steep cline in flowering time for Brassica rapa in southern California:..., Franke [/bib_ref]. Transgenic B. napus plants were derived from spring rapeseed lines (variety Westar, supplied by Dr. Neal Stewart, University of Tennessee). In addition to the Btcry1Ac gene from Bacillus thuringiensis (Bt) [bib_ref] How Bacillus thuringiensis has evolved specific toxins to colonize the insect world, Maagd [/bib_ref] , these lines contained a green fluorescent protein (GFP) gene (mGFP5er) under the control of the cauliflower mosaic virus 35S promoter and a nopaline synthase terminator cassette [bib_ref] Green fluorescent protein as a marker for expression of a second gene..., Harper [/bib_ref] [bib_ref] Removal of a cryptic intron and subcellular localization of green fluorescent protein..., Haseloff [/bib_ref]. The fate of the Bt transgene could therefore be inferred by exposing the offspring to UV light [bib_ref] Green fluorescent protein as a marker for expression of a second gene..., Harper [/bib_ref] [bib_ref] Expression of GFP and Bt transgenes in Brassica napus and hybridization with..., Halfhill [/bib_ref]. Each of the nine populations had its own brush, and new brushes were used for each pollination session. This hand-pollination procedure was chosen to approximate the behaviour of a bumble bee in a patch of oilseed rape. Bumblebees tend to visit many plants successively and rarely revisit the plants [bib_ref] A comparison of bumblebees' movements in uniform and aggregated distributions of their..., Cresswell [/bib_ref]. They deposit most of the pollen from a source plant on immediate neighbours [bib_ref] Predicted pollen dispersal by honey-bees and three species of bumble-bees foraging on..., Cresswell [/bib_ref].
We did not keep track of the random sequences of plants generated for each experimental population on each pollination day so we used observed flowering schedules to calculate the expected proportion of pollen received from transgenic plants (PPR) for each weedy plant. On each pollination day, the probability of a weedy plant receiving pollen from a transgenic plant was assumed to be proportional to the number of transgenic plants in flower in the experimental population. Over the entire flowering period:
[formula] PPR ij~X d h ijd X jd [/formula]
where PPR ij is the expected proportion of flowers crossed with a transgenic plant for weedy plant i from population j, h ijd is the proportion of flowers open on pollination day d for the weedy plant i from population j, and X jd is the proportion of plants in flower that were of the transgenic type on pollination day d in population j. The proportion X jd was calculated by excluding the focal plant i, since B. rapa is known to be largely self-incompatible.
In addition to phenological traits, several morphological and reproductive traits were assessed. On the day of first flower, we recorded basal stem diameter and stem height. Dry siliques were collected once the plants had died. The aggregate mass of filled seeds was determined for each plant by separating these seeds from the lighter, aborted seeds, using an air-flow system. We selected five seeds per plant at random and weighed them, to estimate the total number of seeds per plant. We confirmed the accuracy of these measures by counting and weighing all the seeds for 47 plants spanning the range of seed masses.
Finally, for each weedy plant of the nine experimental populations described above, 14 randomly chosen seeds were sown and grown until the day of the first flower. If a mother plant had less than 14 seeds in total, all were sown. Growing conditions were identical to those for the parental generation. Each seedling was scored for fluorescence under high-intensity UV light, at the four-leaf stage. At this stage, the petioles and main nerves of the A. Mean seed production for the transgenic type treatments. B. The influence of PPR. ''Slopes'' are the coefficients for the effect of each trait on seed production. The ''overall'' slope indicates the effect across all transgenic types. The within-type slopes were obtained from the mixed linera model presented in [fig_ref] Table 5: Mixed linear model for the effects of transgenic type and expected proportion... [/fig_ref] leaves of transgenic plants displayed fluorescence [bib_ref] Expression of GFP and Bt transgenes in Brassica napus and hybridization with..., Halfhill [/bib_ref]. This made it possible to determine the proportion of Bt-GFP+ seedlings for each mother plant. To investigate the association between the transgenic trait and phenotypic traits in the offspring, time to first flower was recorded for each seedling and, on the day of the first flower, basal stem diameter was measured.
# Statistical analysis
We performed all statistical analyses with SAS/STATH software. Plants that died during the experiment were excluded from the analysis and the final data set contained 117 weedy plants.
We first investigated how phenological traits affected the chances of interspecific hybridization between Bt-trangenic plants and weeds. We used a mixed linear model (SAS, Procedure MIXED), with transgenic type (crop, F1 hybrid or first-generation backcross) as the fixed treatment effect, phenological traits of weeds (time to first flower, flowering duration and total number of flowers) as covariates, and block and treatment6block interaction as random effects. The response variable was the proportion of flowers receiving pollen from Bt-transgenic plants (PPR). The response variable was log-transformed to increase its normality (Kolmogorov-Smirnov goodness-of-fit; SAS, Procedure UNIVAR-IATE). If a factor was not significant as a single effect or in interaction with other factors, it was eliminated from the model and the analysis was rerun. We continued until there was no further improvement in residual maximum likelihood.
We then investigated how morphological traits affected the chances of hybridization. Temporal phenotypic clines were assessed by correlating morphological traits of weeds (with time to first flower (Spearman's rank correlation test; SAS, Procedure CORR). A mixed linear approach (SAS, Procedure MIXED) was then used to determine whether the morphological traits changing with time to first flower had a significant effect on PPR. As above, transgenic type (crop, F1 hybrid or first-generation backcross) was treated as a fixed treatment effect, morphological traits were covariates and block and treatment6block interaction were treated as random effects.
We used the mixed linear approach (SAS, Procedure MIXED) with block and treatment x block interactions as random effects, to investigate whether the phenological and morphological traits which were found to favour hybridization of weedy mothers were transmitted to their offspring. In this model, transgenic type (crop, F1 hybrid or first-generation backcross) was treated as a fixed effect, the maternal trait as a covariate and the average offspring phenotypic trait as the response variable. The normality of the response variables was checked (Kolmogorov-Smirnov goodnessof-fit; SAS, Procedure UNIVARIATE), and data was transformed as necessary.
Finally we investigated the relationship between opportunities for hybridization and fecundity in weeds. We used the mixed linear approach (SAS, Procedure MIXED) with transgenic type (crop, F1 hybrid or first-generation backcross) as the fixed treatment effect, PPR as the covariate and block and treatment6 block interaction as random effects. The response variable was the total number of filled seeds. Its normality was checked with a Kolmogorov-Smirnov goodness-of-fit test (SAS, Procedure UNI-VARIATE).
We then checked that the mother plants with the highest expected probability of receiving transgenic pollen (PPR) also had the highest proportion of Bt-GFP+ seedlings. The proportion of Bt-GFP+ seedlings did not follow a normal distribution (Kolmogorov-Smirnov goodness-of-fit; SAS, Procedure UNIVARIATE) and could not be transformed. We therefore checked the effects of transgenic type, PPR and block separately, in non parametric oneway ANOVAs (SAS, Proc NPAR1WAY, Kruskal-Wallis test). The correlation between PPR and the proportion of Bt-GFP+ seedlings was assessed using Spearman's rank correlation test (SAS, Proc CORR). [fig_ref] Table 5: Mixed linear model for the effects of transgenic type and expected proportion... [/fig_ref]. doi:10.1371/journal.pone.0014649.g002
[fig] 38: seedlings, produced by 17 weedy mothers, scored positively. None of them was sired by the pollen of F1 hybrids. The proportions of fluorescent seedlings were equal to 0.0460.01 for populations with crop plants, 0.060.0 for populations with F1 hybrids and 0.0260.01 for populations with first-generation backcrosses (x 2 = 17.55, df = 2, P,0.001). Significant differences were observed between replicates for the proportion of positive scores (x 2 = 15.76, df = 2, P,0.001). The proportion of Bt-GFP+ seedlings was not correlated with PPR in populations with crop plants (r s = 20.16, P = 0.33) nor with backcross plants (r s = 0.23, P = 0.17). [/fig]
[fig] Figure 1: Phenology of transgenic and weedy plants. Phenology of weedy plants (WT; hatched bars) and their Bt-transgenic relatives (CR, F1 or BC; white bars). For each combination, three mixed populations of 30 plants were monitored. Bars represent are the mean numbers of opened flowers per population for each day of observation, with standard errors. WT: weedy plants of B. rapa; CR: Bt-crop plants of B. napus; F1: F1 hybrids between WT and CR; BC: Bt-plants from the backcross of F1 on WT. Arrows indicate the date at which 50% of the flowers had been produced. doi:10.1371/journal.pone.0014649.g001 [/fig]
[fig] Figure 2: Decrease in fecundity with PPR. Total number of filled seeds (TNS) produced by weedy individuals as a function of the expected proportion of pollen (PPR) received from transgenic plants (CR: Bt-crop plants of B. napus; F1: F1 hybrids between weedy plants and CR; BC: Bt-plants from the backcross of F1 on weedy plants). The grey line corresponds to the regression line across all transgenic types (TNS = 2303.15 PPR +272.69); its slope is the overall slope given in [/fig]
[table] Table 2: Effects of phenological traits on the expected proportion of pollen received from transgenic plants (PPR) of weedy B. rapa in mixed populations including transgenic B. napus crop and crop-weed hybrids.A. The mean PPR for the transgenic type treatments. B. The influence of weedy traits. ''Slopes'' are the coefficients for the effect of each trait on PPR. The ''overall'' slope indicates the effect across all transgenic types. The within-type slopes were obtained from the mixed linear model presented inTable 3, they indicate the relationships for crop, F1 and backcross migrants. Coefficients that do not include zero in their 95% confidence interval are shown in bold typeface. doi:10.1371/journal.pone.0014649.t002 [/table]
[table] Table 3: Mixed linear model for the effects of transgenic type and weedy plant phenological traits on the expected proportion of pollen received from transgenic plants (PPR). [/table]
[table] Table 4: Effect of the expected proportion of pollen received from transgenic plants (PPR) on the total number of filled seeds produced by weedy B. rapa in mixed populations including transgenic B. napus crop and crop-weed hybrids. [/table]
[table] Table 5: Mixed linear model for the effects of transgenic type and expected proportion of pollen received from transgenic plants (PPR) on the total number of filled seeds. 22 residual log likelihood = 1240.8. Akaike's information criterion = 1244.8. doi:10.1371/journal.pone.0014649.t005 [/table]
|
Universal thermal climate index associations with mortality, hospital admissions, and road accidents in Bavaria
When meteorological conditions deviate from the optimal range for human well-being, the risks of illness, injury, and death increase, and such impacts are feared in particular with more frequent and intense extreme weather conditions resulting from climate change. Thermal indices, such as the universal thermal climate index (UTCI), can better assess human weatherrelated stresses by integrating multiple weather components. This paper quantifies and compares the seasonal and spatial association of UTCI with mortality, morbidity, and road accidents in the federal state of Bavaria, Germany. Linear regression was applied to seasonally associate daily 56 million hospital admissions and 2.5 million death counts as well as approximately 930,000 road accidents and 1.7 million people injured (2002)(2003)(2004)(2005)(2006)(2007)(2008)(2009)(2010)(2011)(2012)(2013)(2014)(2015) with spatially interpolated same day-and lagged-(up to 14 days) average UTCI values. Additional linear regressions were performed stratifying by age, gender, region, and district. UTCI effects were clear in all three health outcomes studied: Increased UTCI resulted in immediate (1-2 days) rises in morbidity and even more strongly in mortality in summer, and lagged (up to 14 days) decreases in fall, winter, and spring. The strongest UTCI effects were found for road accidents where increasing UTCI led to immediate decreases in daily road accidents in winter but pronounced increases in all other seasons. Differences in UTCI effects were observed e.g. between in warmer north-western regions (Franconia, more districts with heat stress-related mortality, but hospital admissions for lung, heart and external reasons decreasing with summer heat stress), the touristic alpine regions in the south (immediate effect of increasing UTCI on road accidents in summer), and the colder south-eastern regions (increasing hospital admissions for lung, heart and external reasons in winter with UTCI). Districts with high percentages of elderly suffered from higher morbidity and mortality, particularly in winter. The influences of UTCI as well as the spatial and temporal patterns of this influence call for improved infrastructure planning and resource allocation in the health sector.Data Availability Statement:The data used in our study are from the "Research Data Centers of the Federal Statistical Office and the Statistical Offices of the Federal States" (RDC) which can be regarded as a third-party "non-commercial, stateowned" organization. Access to the data is granted after an application referring to conditions under which data are provided for scientific purposes. For this, direct contact is required through the website https://www.forschungsdatenzentrum.de/en/ contact. extends to mortality and road accidents. Losses will intensify in absence of climate change adaptation and mitigation practices[3].Among climate variables, air temperature remains the most studied predictor of morbidity. Higher temperatures were related to more emergency cases in summer [4], more nervous, circulatory and respiratory diseases [5], strokes [6], trauma [7-10], injuries[11,12], and preterm births[13]. Heatwave days witnessed more emergency department admissions and higher mortality[14]due to heat strokes, sunstrokes, and fluid disorders[15]. Mortality rates increased on hot days [16], rapidly per degree above thresholds [17] and injury-related mortality increased with temperature in the US[18]. The reduction of mortality after heat-related mortality peaks, the so-called "harvesting effect", disappeared in cases of extreme heat[16]. On the other hand, cold increased respiratory and circulatory diseases [5] as well as unintentional injuries[11]and led to high mortality levels in France[16]. Mortality due to cold spells was especially high among elderly, respiratory patients, and the less educated[19]. Consequently, temperature influences on morbidity and mortality differ between cold and warm seasons. More trauma patients, a higher proportion of young patients[20], and more orthopedic trauma consultation [10] occurred in summer than in winter in the US and the UK. Wider diurnal temperature ranges in cold seasons were associated with more patients with chronic respiratory diseases, but less in hot seasons[21]. Most interestingly exposure to heat in warm seasons had no impact on hospital admissions for cardiovascular and respiratory reasons in Spanish cities but was associated with higher mortality risks. In contrast, cold exposure was associated with more hospitalization, but lower mortality risks due to cardiovascular and respiratory reasons during cold seasons[22]. Thus, it is necessary to consider seasonal variations to understand the impact of weather conditions for health services[23]. In addition to temperature, other meteorological parameters for predicting morbidity and mortality are solar radiation, humidity, wind speed, precipitation, and foehn[4,6,8,24,25]. Following extreme weather events, there is an increased demand for emergency services and an increased mortality risk[2,26]. Road accidents as well as associated injuries and fatalities have been influenced by both high and low temperatures [27-30], precipitation [31, 32], sunshine and wind speed [33], as well as by sandstorms[34]. Higher intensity of weather conditions increased road accidents[30,35], particularly when temperatures were below freezing in Germany[36].Thermal indices provide an excellent way to predict the demand for medical services due to adverse weather conditions more efficiently than direct weather variables[37,38]. Indices, such as the Universal Thermal Climate Index (UTCI), have been recommended as physiologically relevant indices for biometeorology and climate impact studies[39]. UTCI better represented the physiological response of the human body, and was more sensitive to heat stress changes than other thermal indices[40]. Consequently, UTCI has been proposed as a basis for constructing heat warning systems[41], leading to assessments of its spatial and temporal variation[42][43][44]. An increase in UTCI has been observed for Europe over the last decade, with the south being more prone to heat stress than the north[45].Various correlations between UTCI and morbidity/mortality have been addressed in recent studies. UTCI performed well in estimating the occupational heat stress in mines in Iran [46], the intensity of summer excess mortality in the Czech Republic [47], as well as in Europe, particularly in France during the heatwave of summer 2003[45]. The impact of UTCI on mortality varied between warm and cold regions in Poland [48] and between rural and urban areas in the Czech Republic[37].Within Germany, Bavaria is expected to suffer an increase in mortality rate due to climate change and population aging [49]. Temperature extremes induced higher mortality rates among the elderly [50] and higher cardiovascular mortality [51]. Also, ambulance activity in Munich was affected by temperature, humidity, sunshine, and precipitation[4], and severe PLOS ONE UTCI effects on mortality, hospital admissions, and road accidents in Bavaria PLOS ONE | https://doi.
# Introduction
Climate change is altering the frequency, intensity, spatial extent, duration, and timing of extremes. This affects morbidity, particularly among vulnerable populations, and a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 trauma by foehn winds. However, the influence of weather conditions on traffic accidents has not yet been studied at all. To the best of our knowledge, there have been no studies addressing and comparing the UTCI associations with mortality, morbidity, and traffic accidents.
We hypothesize that the influence of heat stress as assessed by UTCI on both morbidity and road accidents is comparable to the established influence of heat stress on mortality across seasons. We expect the spatial variation in these effects to be related to population characteristics. Therefore, the following questions are addressed (1) How do mortality, morbidity, and road accidents respond to variation in UTCI?Does such an influence differ with seasons, age, sex, and regions within Bavaria?
By combining daily UTCI averages and daily counts of hospital admissions, death cases, and traffic accidents, this report quantifies the impact of UTCI on the population of Bavaria in order to mitigate severe impacts of weather extremes, improve the resilience and preparedness of health care systems, and reduce casualties.
# Materials and methods
## Morbidity, mortality, and road accidents
Bavarian hospital admissions and mortalities for the period 1995-2015 as well as road accidents for 2002-2015 were provided after anonymization by the Research Data Centers of the Federal Statistical Office and the Statistical Offices of the Federal States . Access to this data and can be arranged through their website after strict procedures for ethical and privacyrelated reasons. For further details https://www.forschungsdatenzentrum.de/en/request. Hospital admissions and mortality data comprised date, age, sex, and ICD-10 code, while the road accidents data included the number of people involved in each accident. A total of 56,028,368 hospital admissions, 2,557,651 mortalities, and 930,861 road accidents involving 1,753,980 injured persons were aggregated for this study (see. The three datasets may partially overlap; however, it is not possible to identify the overlapping cases. The daily counts of hospital admissions were itemized by sex, age (child: <18, adult: 18-70, senior: >70), diagnoses (heart, lungs, external), and death. Similarly, daily mortality counts were itemized by sex, age, and diagnoses. "External" here refers to injuries due to external causes such as falling, machinery, or fire.
It is important to note the spatial variation of population and age structure among Bavarian regions and districts. Therefore, the daily counts were calculated for the whole of Bavaria, for its seven regions, namely Lower-Bavaria/Niederbayern (NB), Upper-Bavaria/Oberbayern (OB), Upper-Franconia/Oberfranken (OF), Upper-Palatinate/Oberpfalz (OPf), Swabia/Schwaben (Sch), Lower-Franconia/Unterfranken (UF), Middle-Franconia/Mittelfranken (MF) (see S1 , and for its 96 districts (i.e. Landkreise).
## Utci
To determine the UTCI, the meteorological variables solar radiation, relative humidity, wind speed, and air temperature are required. The German Meteorological Service (DWD)provided hourly measurements for 44 stations throughout Bavaria between 1995-2015. To obtain precise UTCI values, detailed information on topography, human activities, and clothing in a physiological model is required. Since this information is not available, hourly UTCI values at individual stations were approximated by 6 th order polynomial regression functions. Daily grids of UTCI at 200m resolution were then produced based on the daily averaged values of UTCI and integrated nested Laplace approximation (INLA) models. From this analysis, mean daily UTCI values were extracted for Bavaria, each region, and district.
## Plos one
# Statistical methods
Multiple normal linear regression was used for association analyses of daily hospital admissions, death, and road accidents with UTCI, spatial and seasonal predictors. Lagged effects of UTCI up to 14 days were considered as potential predictors in a model for expected daily cases E(Y) (hospital admissions, deaths, or road accidents), which controlled for the year, official state holiday, and day of the week as shown below:
[formula] E Y ð Þ ¼ a þ b y � y þ b H � H þ X Saturday d¼Monday b d � dþ X lag¼14 lag¼0 b UTCI lag � UTCI lag ;ð1Þ [/formula]
where α is the intercept, Y the year, β y the change in E(Y) per year, H an indicator with value 1 for holidays and 0 otherwise, β H the difference in counts between holidays and non-holidays, d the day of the week, β d the difference in counts between day d and the reference day, UTCI lag is the UTCI on lag days previously, and β UTCI_lag is the change in counts for a unit increase in UTCI. Friday was chosen as the reference weekday for hospital admissions since weekdays at the start of the week may be more compromised / biased by planned surgeries or examinations. For road accidents, the choice was moved to Thursday since Mondays and Fridays witness different levels of transportation activity on the road due to weekend commuters and weekend activities. For mortality, Wednesday was chosen as the reference day at the middle of the weekdays, and to highlight the different situation compared to the weekends. Inclusion of the lagged UTCI effects as fixed effects in the model removed autocorrelation, justifying the assumption of independent errors as required for the models. Large sample sizes enabled the assumption of a Normal distribution by the central limit theorem, which was confirmed by residual Normal quantile plots. Separate models were constructed for meteorological seasons (DJF winter, MAM spring, JJA summer, SON fall), for age, sex, and diagnosis subgroups, and for Bavaria, its seven regions, and 96 districts. The total number of analyses comprised 84 for Bavaria, 588 for regions, and 8064 for districts. Daily case numbers and UTCI varied between regions, districts, subgroups, and seasons (see also , and hence were standardized by subtracting means and then dividing by the respective standard deviations (sd). Therefore, when interpreting the modelled effects for a target group, season, and spatial unit, the sd of the respective daily values of mortality, morbidity, and accidents has to be considered as the unit of change (see. For example, a β UTCI value of 0.1 means that a change of one standard deviation in UTCI within a particular season and region (or district) was associated with a positive increase in the daily count of cases of the target group with a magnitude of 10% of its standard deviation in the respective season and region. Explanatory examples in each figure caption help with this calculation. The set of predictors for each model was chosen by minimization of the Bayesian Information Criterion (BIC). Variance inflation factor values were mostly below 2 indicating low multicollinearity in the final models. All computations were performed in the R statistical software package (version 4.0.3) and all comparisons were tested at the two-sided 0.05 level of significance.
# Results
## Seasonal utci effects on daily morbidity, mortality, and road accidents
In summer, higher UTCI was consistently associated with increasing daily numbers of mortality, hospital admissions, and road accidents. The strength of immediate UTCI effects UTCI effects on daily cases of mortality, hospital admissions, and road accidents for each subgroup and season in Bavaria. The horizontal axis represents the lag in days, and the vertical axis represents the effect estimate, either positive (red) or negative (blue). Crosses are lacking when the effect of UTCI was not significant for a particular lag in the corresponding model. Effect estimates are expressed in proportion of the standard deviation of daily number of cases within the corresponding subgroup (given in the upper right corner of each panel) when UTCI changes by one standard deviation of its daily value for the corresponding season. The numbers in the gray facet titles on the right are the standard deviations of UTCI within each season for the whole of Bavaria. The absence of the crosses indicates that the effect of UTCI was not significant for a particular lag in the corresponding model. Explanatory example: For winter, the UTCI sd is 6.2˚C, the hospital admissions sd is 3143.5, and the estimated UTCI effect is -0.0513. This indicates a decrease in daily hospital admissions in Bavaria by -0.0513 � 3143.5 = 161 when the value of UTCI increases by 6.2˚C.
https://doi.org/10.1371/journal.pone.0259086.g003
## Plos one
(lag 0), however, differed substantially per 4.6˚C UTCI increase: Daily hospital admissions increased by~+0.04 sd, mortality by~0.20 sd, and road accidents by~+0.40 sd. Correspondingly, the category death within hospital admissions was characterized by larger UTCI effects (0.12 sd) than total admissions (0.04 sd).
In a few other cases amplifying (positive) UTCI effects were derived for seasons other than summer, namely for road accidents in spring and fall, children hospital admissions in fall as well as mortality due to external reasons in spring and fall. Here again, the UTCI effect size in spring was larger for road accidents than for external mortality.
Model results also pointed to temporal changes in UTCI effect based on the lag structure. In summer, an increase in mortality occurred up to three days after the increase in UTCI, but later there was a decrease within the second week. In the other seasons, however, no immediate effect was detected, but a decreasing effect of UTCI was apparent about three days after the UTCI change which lasted up to two weeks. Nevertheless, there were two exceptions: death among children was not influenced by UTCI fluctuations, and death due to external causes increased immediately after an increase in UTCI in spring, summer, and fall.
Higher summer UTCI values caused an immediate to three days lagged increase in Bavarian daily hospital admissions, but a decrease after one week in fall and after two weeks in winter and spring. The effect was consistent among sex and age groups, except for children's hospital admissions, which increased immediately after an increase in UTCI in fall. No effect was detected on heart or external admission categories. However, lung-related admissions in spring decreased a couple of days after a UTCI increase. Remarkably, UTCI had an immediate positive effect on death within hospital admissions only in summer, and a negative lagged effect in winter, spring, and summer.
In winter, an increase in UTCI was associated with an immediate reduction in road accidents without any lagged effects. In other seasons, the association between UTCI and road accidents was always positive, very strong, and immediate in spring and summer and decreased over time up to one week in summer and fall, and up to two weeks in spring.
## Spatial variation of utci effects in summer and winter
To examine the spatial variability of UTCI effects on the daily cases of morbidity, mortality, and road accidents, the effect estimates were compared between regions and districts. In order to tighten the results, we only discuss the summer and winter seasons in more detail (Figs 4 and 5), the corresponding results for spring and fall are in the supplement (S2 and S3 Figs).
In winter, there were no apparent differences in the UTCI effect structure on mortality among regions and their districts. Concerning morbidity, increases in UTCI were associated with immediate rises in hospital admissions due to heart, lungs, and external reasons in the three south-eastern (partly higher elevated and thus colder) regions of Bavaria, namely OB, NB, and OPf, but lagged reductions in MF and UF (i.e. the warmer regions of Bavaria) and some districts of the remaining regions. Additionally, increases in UTCI led to immediate reductions in road accidents in all regions except for UF, but delayed increases in all regions except NB and OPf.
Summerly UTCI effects on mortality were more frequent in two regions of Franconia (MF, UF), the relatively warmer north-western part of Bavaria, where more districts revealed significant UTCI effects. Quite strikingly, summer hospital admissions due to heart, lungs, and external causes were negatively correlated with UTCI change in OF, MF, and UF for 1 to 14 days lags structures. However, no effect was detected for all other regions, except NB where admissions due to external injuries were positively correlated with UTCI change. Additionally, deaths among hospital admissions decreased in the second week after UTCI increased in NB, MF, and Sch. Road accidents strongly increased with higher UTCI in summer in OB and Sch, the two main touristic regions in the south of Bavaria comprising alpine forelands and the alpine region.
For the two transitional seasons, spring and fall, some regional patterns in UTCI were apparent: Like in summer, the immediate increase in road accidents in spring with UTCI was high in OB, NB and Sch, which could be associated with the increase in deaths due to external causes in these three regions (S2 During fall, and despite the significant effect of UTCI on hospital admissions for all age groups for Bavaria, children's hospital admissions were not affected by UTCI in MF and UF (S3 Similarly, the elderly were not affected in NB. In contrast, while
## Plos one
UTCI had no effect on heart, lungs, or external hospital admissions, a significant negative effect appeared only in OF, MF, UF, and Sch for heart and lungs, and a positive effect for external cases in NB. The positive immediate effect of UTCI change on road accidents in Bavaria was limited to OB, NB, Sch, and reversed to a negative effect in OF and few other districts.
## Calendar effects
Daily hospital admissions and mortality increased across Bavaria during the 1995-2015 study period, while daily road accidents decreased. Overall, annual increases in daily hospital
## Plos one
admissions ranged from 2.6% of sd in the spring to 3% in winter, corresponding to 75 and 92.2 additional daily cases per year, respectively. Total daily death counts in Bavaria increased during the study period, but not consistently across districts and for subgroups. Remarkably, the positive trend in mortality was the least in winter. While mortality increased for the elderly and for lungs cases, it decreased among adults, children, heart, and external cases (results for subgroups not shown). Daily road accidents and the resulting injuries decreased in all seasons, with the strongest decreases in fall, and the smallest decreases in winter.
Holidays significantly reduced the number of hospital admissions and road accidents but had no significant effect on deaths except in spring. Total daily hospital admissions were reduced especially in winter holidays (-105% of sd), followed by spring (-45% sd) and the reducing effect was smallest in summer and fall (-23% sd). The effect on children having Year and holiday effects on daily mortality, hospital admissions, and road accidents for each season, expressed as a proportion of the standard deviation of the respective daily number of cases within Bavaria. The absence of a column means that the effect of the year or holiday was not significant in the corresponding model. Explanatory example: For winter hospital admissions, the sd is 3143.5 (see
## Plos one
school holidays was comparably high during summer (-42%), but lower in spring (-28%), and the least in fall (-9%). We observed only a marginal reduction of mortality during the spring holidays. Road accidents decreased during holidays by 90%, 78%, 42%, and 25% of the respective daily sds in winter, summer, spring, and fall.
Hospital admissions were the highest on Mondays, and declined until Fridays, while Saturdays and Sundays witnessed the least daily cases. This pattern was spatially consistent, and across sex and age groups (results for subgroups not shown). The effect size was highest in summer and fall, and smallest in winter, slightly lower for children and the elderly, and mostly limited to Mondays and weekends for heart, lungs, external causes, and death within hospital admissions. Mortality in Bavaria generally dropped during the weekends, especially on Sundays. Children mortality was not affected by the day of the week, while death due to external injuries increased on Mondays. Road accidents increased on Fridays in all seasons, and on
## Plos one
Mondays in winter. However, Sundays had the least number of accidents followed by Saturdays.
# Discussion
While few studies have separately examined the effects of weather on morbidity, mortality, and road accidents in Bavaria, this is the first study to compare the effects and consider the integrative effect of UTCI on these three public health outcomes. In addition, we examined variations of UTCI effects in space (districts, regions, and Bavaria), with season, and specific to different population groups. The results clearly indicated that UTCI had a significant effect on all three public health outcomes after adjusting for calendar effects.
## Utci effect
The significant UTCI effect on hospital admissions, mortality, and road accidents in Bavaria extended to all regions, districts, and subgroups except children mortality which was not influenced by UTCI. This might be due to the low daily death cases among children, especially since an increase in children's hospital and emergency admissions particularly in summer has been reported by other studies.
Mortality and morbidity in Bavaria were similarly affected by UTCI as previously reported for temperatures, i.e. higher UTCI during summer caused an immediate increase in hospital admissions and mortality. The observed increase in hospital admissions agrees with the previous findings of higher ambulance activity in the Munich area with higher temperatures. Similar increases in mortality with high UTCI have been reported for France, Czech Republic, and Bangladesh. In fall, winter and spring, however, lower UTCI increased hospital admissions and mortality in Bavaria with a lag of one to two weeks. Such cold-related mortality has been equally reported for the Czech Republicand Greece.
The unique result of our study is that for both promoting UTCI effects on mortality and morbidity in summer as well as mitigating UTCI effects in winter, the effects were clearly stronger for mortality than for morbidity by up to a factor of 5. Here, our study supports recent results for 52 Spanish citieswhere heat exposure increased the risk of cardiovascular and respiratory mortality, but not hospital admissions.
The second important point of our study is a deeper look into the spatial patterns of such UTCI effects. Generally, the eastern districts of Bavaria have higher hospital admission rates per 100,000 inhabitants than the western ones, probably not due to more hospital beds (see S4 Fig for no apparent east-west differences), but likely due to a higher share of elderly people in the (north-) eastern districts of Bavaria, especially the districts of Wunsiedel and Hof, and in Garmisch-Partenkirchen in the south. It is well known that this senior part of the population is particularly vulnerable to heat/cold stress [49], probably resulting in higher hospital admissions and mortality in these districts at all times of the year, but especially in winter. These above-mentioned districts are characterized by lower UTCI values in winter, likely contributing to this spatial pattern in winter. A similar spatial variability of the UTCI effect on mortality has been recognized when comparing cool and warm cities in Poland .
Systematic reviews showed that heat-related morbidity and mortality differ by sex. Yet, no clear distinct differences between female and male vulnerability to UTCI were detected in this study.
Although there was a lack of a significant UTCI effect on heart, external, and lung hospital admissions for Bavaria, season-specific spatial variations became evident for these subgroups in summer and winter when considering the seven regions and the 96 districts, comparable to the study on cardiovascular-and respiratory-related hospital admissions in 52 Spanish cities. For the (south-) eastern regions (OB, NB, and OPf), higher UTCI in summer had no effect on hospital admissions and lower UTCI in winter was associated with fewer hospital admissions. In contrast, for the (north-) western regions (OF, MF, and UF) higher UTCI in summer reduced hospital admissions and lower UTCI in winter was associated with increased hospital admissions. These east-west contrasts within Bavaria are striking since the UTCI heat effect on hospital admissions in summer was smaller in the comparably hotter regionsand in winter the UTCI cold effect was even reverse in the comparably cooler regions of Bavaria. Such region-specific variations in seasonal heat/cold stress effects on hospital admissions should be addressed when adjusting the best prevention and adaptation policies across the whole federal state of Bavaria accordingly.
Higher UTCI considerably increased road accidents in all seasons except for winter when more road accidents happen immediately after a decrease in UTCI. This can be well explained by bad driving conditions due to fog, precipitation, and the drop of temperatures below freezing level in winter, constituting major reasons for road accidents in Germany. The increase in road accidents following high UTCI values in spring, summer, and fall coincide with a higher mortality (see. This may be explained by a greater demand for leisure seeking activities and transportation. The immediate UTCI effect on road accidents in summer was about twice as strong as on mortality. Remarkably, during summer, road accidents increased in touristic districts especially in the alpine region in the south. The high number of road accidents in larger cities in all seasons was probably due to the high traffic density, which also peaks in summer with more non-local drivers. Since immediate amplifying UTCI effects were comparably strong in three seasons (spring through fall), awareness-raising activities and additional safety measures should target these regions before and during the peak periods.
## Confounding effects
All UTCI models accounted for other calendar-related effects (year, holidays, day of the week) on mortality, hospital admission, and road accidents. Since a deeper understanding of these confounding effects may support adaptation and policy measures, we will briefly discuss selected results in this respect.
The Bavarian population increased by 850,000 inhabitants over two decades, reaching 12.8 million in 2015. The corresponding increase in daily hospital admissions was concentrated in the elderly, children, and males, most likely due to a higher proportion of elderly in the German population being especially vulnerable to heat stress . The increasing trend of death due to lung problems might be related to air pollution, as well as lung cancer and chronic obstructive pulmonary disease (COPD), especially among the increasing number of smoking women. This highlights the importance of including the effect of air pollution in future studies, especially in the north-eastern districts, which are open to pollution sources from the east and commonly suffer blocking situations. Most Bavarian districts and regions had similar year effects on hospital admissions, mortality, and road accidents, with a few interesting exceptions. The districts with the highest population density, the cities of Munich, Nurnberg, and Augsburg, had a negative trend in total mortality, while most of the remaining districts had a positive trend, mirroring quite likely the concentration of the mobile, younger working population where the jobs are. Additionally, the observed discrepancies in mortality trends may be influenced by variations in birth rates, migration of young people to large cities, and investments in infrastructure. The observed reduction in daily road accidents over the study period could be associated with the technological improvement of safety measures, and the continuous efforts within the European Union to reduce casualties on the road. There was only one district (Landshut) with a positive trend in road accidents in both summer and spring without any obvious explanation. Such deviations from the general patterns may be associated with the demographic characteristics, employment situation, number of hospital beds in these districts, and how they change over time. Factors like this should be addressed on the district level to resolve deficiencies and extract good policies on how to adapt to UTCI changes.
Holidays had almost no effect on mortality but witnessed fewer hospital admissions in all seasons, especially in winter, and for all regions, age, and sex groups. This evident winter reduction may be related to fewer outdoor activities. The reduction of mortality within hospital admissions during holidays was always less than the corresponding reduction in the total number of admissions. This finding may correspond to results of a meta-analysis reporting that the mortality rate is higher among patients admitted during holidays and weekends. Seniors mostly do not depend on (school) holidays anymore, thus should not have holiday effects as e.g. shown by an absence of holiday effect on elderly emergency hospitalization in Munich. However, in our study elderly hospital admissions did not deviate from the general pattern of the remaining age groups in Bavaria, and not even when considering the city of Munich. This discrepancy may be caused by the different years included in the emergency hospitalization study in Munich (2014-2018). The overall reducing effect of holidays was especially strong in winter and summer, the two seasons with a larger proportion of holidays in Bavaria.
Day of the week effects basically mirroring human behavior and habits were most prominent for hospital admissions, followed by road accidents, and least for mortality. Planned hospital admissions and a catch-up effect after the weekend perfectly explain the reduced hospital admissions on the weekend and the respective increase during weekdays, especially stronger at the beginning of the week. Weekends, especially Sundays had the least number of road accidents due to the reduced road activity in these days, whereas increased numbers on Mondays and Fridays may be related to weekly commuters. The excess in road accidents on Fridays, and Saturdays compared to Sundays might be due to alcohol consumption, risk-seeking, and leisure drives at night among young drivers.
## Future research and limitations
Our study confirmed that admission to hospitals, mortality, and road accidents were clearly associated with UTCI changes, with sizes of immediate effects in summer increasing in this order. Thus, the established influence of heat stress/UTCI extends to our health outcomes. In the next step, comparisons between prediction models which use thermal indices, particularly UTCI, and temperature should be carried out. In winter, the influence of precipitation might interfere with the UTCI influence. For example, it is evident that precipitation affects hospital admissions in the United States, while precipitation influences road accidents worldwide. Particularly in the case of Munich, precipitation had a negative effect on emergency department visits in fall, and the number of female patients was negatively correlated with hail warnings and maximum precipitation intensity. The simultaneous influence of precipitation and UTCI on mortality, morbidity, and road accidents is not addressed in this paper due to its complexity. However, this interaction calls for dedicated future research.
We expected UTCI effects to differ spatially within Bavaria and with subgroup characteristics. Therefore the specific patterns in the significant effects of UTCI on morbidity, mortality, and road accidents, that become obvious when conducting separate tests per season and agegender subpopulation at different spatial scales, are extremely important. These results highlight the importance of planning climate change adaptation and mitigation efforts in both the local and regional context. According to literature, people who live in regions of moderate climate show higher sensitivity to weather extremes. By applying this concept to Bavaria, inhabitants of warmer areas in Franconia and colder alpine regions are supposed to be less sensitive to UTCI variations in summer and winter, respectively. Therefore, our first, but not consistent results on spatial differences in UTCI effects should be intensified in future studies. Then, also the possibly amplified effect of UTCI in urban areas that might contribute to the increasing number of hospital admissions and mortality, especially among the elderly, should be addressed. Future research should also consider the possible influence of heat islands in large cities, and the moderating influence of vegetation and water bodies on the heat stressrelated morbidity and mortality which has been recognized in other geographical locations.
The correlations reported between UTCI and the daily counts of Hospital admissions, death, and road accidents should be interpreted with care. Previous studies have demonstrated the causal connection between weather conditions, particularly heat-stress, and mortality in Bavarian large cities . Additionally, the observed correlations are highly influenced by the public behavior patterns. Specific ranges of UTCI at specific seasons may cause higher or lower levels of activity, use of transportation, interaction with the environment and other individuals, and stress which all eventually accumulate in causing the observed fluctuations in morbidity, mortality, and road accidents numbers. The available data does not contain enough details to attribute each case to a specific cause. However, our results indicate the surplus in cases which occur following UTCI fluctuations regardless of their nature of a direct or indirect link. This is particularly important for improving preparedness in the healthcare sector. Spring UTCI effect on the daily mortality, hospital admissions, and road accidents for each subgroup. The horizontal axis represents the lag in days, and the vertical axis represents the effect estimate. It is expressed in proportion of the standard deviations of daily number of cases within the corresponding subgroup when UTCI changes by one standard deviation of its daily value. The black circles represent those effects for each region. The colored dots represent those effects for districts within the region. The absence of the points means that the effect of UTCI was not significant for a particular lag in the corresponding model. (TIF) S3 Fall UTCI effect on the daily mortality, hospital admissions, and road accidents for each subgroup. The horizontal axis represents the lag in days, and the vertical axis represents the effect estimate. It is expressed in proportion of the standard deviations of daily number of cases within the corresponding subgroup when UTCI changes by one standard deviation of its daily value. The black circles represent those effects for each region. The colored dots represent those effects for districts within the region. The absence of the points means that the effect of UTCI was not significant for a particular lag in the corresponding model. (TIF) S4 Bavaria. The number of hospital beds per 1000 inhabitants averaged over the study period 1995-2015. Border shapefiles were provided with written permission by the Bundesamt für Kartographie und Geodäsie under the license (CC BY 4.0). (TIF) S1
## Supporting information
|
The relationship between IL-6 levels and the angiographic severity of coronary artery disease following percutaneous coronary intervention in acute coronary syndrome patients
Background: The present investigation was developed for the exploration of the association between IL-6 levels and acute coronary syndrome (ACS) findings upon angiographic evaluation.Methods: A retrospective review of 346 patients suffering from chest discomfort that underwent coronary angiography was performed. The SYNergy between Percutaneous Coronary Intervention with TAXus and cardiac surgery (SYNTAX) score (SS) and SS II were used to gauge ACS severity, with ACS patients being stratified into two groups based on an SS value of 22 and the median SS II value. Associations between IL-6 levels and SS or SS II values were assessed through Spearman's correlation analyses, and independent predictors of intermediate-high SS or high SS II were identified via a multivariate logistic regression approach. A receiver operating characteristic (ROC) curve was employed to explore of the predictive value of IL-6 levels.Results: IL-6 was positively correlated with both SS (r = 0.479, P < 0.001) and SS II (r = 0.305, P < 0.001). Moreover, IL-6 levels were independently predictive of intermediate-high SS and high SS II values. ROC curves further demonstrated that IL-6 was able to predict intermediate-high SS and high SS II, with area under the curve (AUC) values of 0.806 and 0.624, respectively.Conclusion: IL-6 levels are closely linked to the extent of coronary artery disease in ACS patients undergoing percutaneous coronary intervention. IL-6 levels may thus serve as a valuable and non-invasive biomarker of high-risk ACS patients.
treatment and revascularization approaches have improved the prognosis of ACS patients, many patients nonetheless experience unsatisfactory adverse cardiovascular outcomes [bib_ref] Acute myocardial infarction, Anderson [/bib_ref]. At present consensus criteria pertaining to understanding ACS pathogenesis are lacking, and defining the optimal treatment for this condition is thus an important global public health goal [bib_ref] Acute coronary events, Arbab-Zadeh [/bib_ref].
A growing body of evidence suggests that inflammatory activity plays a maladaptive role in the context of ACS pathophysiology, triggering initiation and progression of atherosclerosis, driving plaque destabilization and degradation, and responding to myocardial necrosis [bib_ref] Progress and challenges in translating the biology of atherosclerosis, Libby [/bib_ref] [bib_ref] Interleukin-6 and the risk of adverse outcomes in patients after an acute..., Fanola [/bib_ref] [bib_ref] Targeting cardiovascular inflammation: next steps in clinical translation, Lawler [/bib_ref]. Consistently, many studies have reported an association between cardiovascular disease (CVD) and increased levels of biomarkers indicative of inflammation [bib_ref] Interleukin-6 and the risk of adverse outcomes in patients after an acute..., Fanola [/bib_ref] [bib_ref] Association between IL-6 and the extent of coronary atherosclerosis in the veterans..., Saremi [/bib_ref] [bib_ref] Inflammatory biomarkers interleukin-6 and c-reactive protein and outcomes in stable coronary heart..., Held [/bib_ref]. In particular, the pro-inflammatory cytokine interleukin-6 (IL-6), which is primarily produced by macrophages and T cells, has been noted as a key driver of plaque destabilization, atheroprogression, and the production of high-sensitivity C-reactive protein (hs-CRP), leading to the development and progression of clinical atherosclerosis [bib_ref] Interleukin-6 and the risk of adverse outcomes in patients after an acute..., Fanola [/bib_ref] [bib_ref] Impact of interleukin-6 on plaque development and morphology in experimental atherosclerosis, Schieffer [/bib_ref] [bib_ref] Interleukin 6 inhibition and coronary artery disease in a high-risk population: a..., Bacchiega [/bib_ref] [bib_ref] IL-6 in diabetes and cardiovascular complications, Qu [/bib_ref]. Links between higher IL-6 levels and an elevated risk of cardiovascular events among otherwise healthy individuals are well documented [bib_ref] Plasma concentration of interleukin-6 and the risk of future myocardial infarction among..., Ridker [/bib_ref] [bib_ref] C-reactive protein and other markers of inflammation in the prediction of cardiovascular..., Ridker [/bib_ref] , and IL-6 has also been proposed to be a predictor of coronary artery disease (CAD) severity and associated mortality among ACS patients [bib_ref] Interleukin-6 and the risk of adverse outcomes in patients after an acute..., Fanola [/bib_ref] [bib_ref] Prognostic value of IL6 in young adults presenting with acute coronary syndrome, Mowafy [/bib_ref] [bib_ref] Association between serum interleukin-6 concentration and mortality in patients with coronary artery..., Su [/bib_ref]. Additionally, levels of IL-6 have been reported to be associated with plaque burden as defined by intracoronary imaging [bib_ref] Trans-myocardial blood Interleukin-6 levels relate to intracoronary imaging defined features of plaque..., Bambrough [/bib_ref].
In previous studies, the SYNergy between Percutaneous Coronary Intervention with TAXus and cardiac surgery (SYNTAX) score (SS), which is frequently used to quantify CAD degree and severity, has been shown to predict prognostic outcomes in stable CAD and ACS patients [bib_ref] Prognostic value of the SYNTAX score in patients with acute coronary syndromes..., Palmerini [/bib_ref] [bib_ref] The Syntax score predicts peri-procedural myocardial necrosis during percutaneous coronary intervention, Van Gaal [/bib_ref] [bib_ref] Value of the SYNTAX score in patients treated by primary percutaneous coronary..., Magro [/bib_ref]. In recent years, the SS II indicator has been expanded to take individual clinical characteristics into account, reportedly achieving higher accuracy rates in the prognostic assessment of ACS patients [bib_ref] Anatomical and clinical characteristics to guide decision making between coronary artery bypass..., Farooq [/bib_ref] [bib_ref] Usefulness of the SYNTAX score II to predict 1-year outcome in patients..., Wang [/bib_ref].
The relationships between IL-6 levels and both SS and SS II values, however, are poorly documented. As such, this study was designed to explore the link between IL-6 levels and ACS severity as measured using the SS and SS II indicators at the time of admission.
# Methods
## Study population
For this study, patients suffering from chest pain that were admitted to the Division of cardiology and underwent coronary angiography (CAG) between January 2021 to August 2021 were enrolled. The criteria for diagnosing ACS (including STEMI, NSTEMI, and UA patients) were based on the standard recommended by the ESC guidelines [bib_ref] ESC Guidelines for the management of acute coronary syndromes in patients presenting..., Hamm [/bib_ref]. Patients were excluded if they had already undergone percutaneous coronary intervention (PCI) or coronary artery bypass grafting surgery (CABG), or if they exhibited malignancies, autoimmune disease, severe hepatic or renal failure, or infectious or inflammatory disease. In total, 346 patients were included in the final study, with these patients being classified into an ACS group and a stable angina pectoris (SAP) group (individuals with diseased vessels exhibiting > 50% luminal narrowing). This study was consistent with the Declaration of Heksinki and was approved by the Institutional Review Board of Yijishan Hospital Affiliated of Wannan Medical College.
## Patient's characteristics
The hospital electronic database was used to document all patient demographic and clinical characteristics. Fasting blood samples were obtained from the peripheral veins of all patients before PCI to assess hematologic indices, hs-CRP levels, biochemical parameters, and IL-6 concentrations using standard approaches in our hospital's clinical laboratory. IL-6 concentrations were measured using an enzyme-linked immunosorbent assay (Human IL-6 ELISA Kit, Fine Test, Wuhan, China). Transthoracic echocardiography was conducted prior to angiography. The Cockcroft-Gault equation was utilized to calculate the estimated glomerular filtration rate (eGFR) for each patient.
# Coronary angiographic analysis
All patients underwent CAG via the radial approach at admission. CAG was performed by two expert interventional cardiologists who were blinded to patient clinical information. SS calculations were based on the coronary artery with a ≥ 50% luminal narrowing in a vessel ≥ 1.5 mm and were performed using the SS calculator (www. synta xscore. com, version 2.1). Furthermore, SS II values were established based on these SS values, the presence of left main coronary artery disease, peripheral arterial disease (PAD), chronic obstructive pulmonary disease (COPD), female sex, eGFR, and left ventricular ejection fraction (LVEF) [bib_ref] Assessment of relationship between C-reactive protein to albumin ratio and coronary artery..., Çağdaş [/bib_ref].
# Statistical analysis
The Kolmogorov-Smirnov approach was employed to assess whether continuous data were normally distributed. Normally distributed quantitative data are given as mean ± standard deviation (SD), while outcomes that were non-normally distributed are expressed as medians with the interquartile range. These outcomes were compared via Student's t-tests or Mann-Whitney U-tests. Categorical variables are given as numbers (percentages) and were compared using chi-squared tests or Fisher's exact assessment. Parameters significant (P < 0.1) in initial univariate analyses were incorporated into a multivariate logistic regression analysis designed to identify independent predictors of intermediate-high SS and high SS II values. Receiver operating characteristic (ROC) curves were utilized to demonstrate the ability of IL-6 levels to predict these two outcomes. SPSS 23.0 was used for statistical analyses, and P < 0.05 was considered as statistically significant difference.
# Results
## Study participants
The demographic and clinical characteristics of ACS (n = 201) and SAP (n = 145) group patients were initially evaluated [fig_ref] Table 1: Demographic, clinical and biochemical characteristics between ACS and SAP group ACS, acute... [/fig_ref]. Overall, patients in the ACS group were mostly male and exhibited higher creatinine and fibrinogen levels (P < 0.05) relative to those of patients in the SAP group. Moreover, these ACS patients exhibited significant increases in white blood cell (WBC), neutrophil (NEUT), and platelet counts as well as neutrophil to the lymphocyte ratio (NLR) values compared to SAP patients (P < 0.05). Furthermore, IL-6 and hs-CRP concentrations were significantly greater in ACS group patients as compared to those in SAP group patients (P < 0.01), while apolipoprotein A1 (apoA1), high-density lipoprotein cholesterol (HDL-c), albumin, and LVEF were all reduced in these ACS patients (P < 0.05). No differences in age, body mass index (BMI), or other characteristics were observed when comparing these two groups. Additionally, no significant relationship between IL-6 levels and SS values in SAP patients were observed via Spearman's correlation analyses (r = 0.128, P > 0.05; Additional file 1: .
## The relationship between il-6 levels and an intermediate-high syntax score
ACS patients were separated based on SS values cited in prior studies [bib_ref] The SYNTAX Score: an angiographic tool grading the complexity of coronary artery..., Sianos [/bib_ref] , with one group of patients with low SS values (SS ≤ 22, n = 168) and one group with intermediate-high SS values (SS > 22, n = 33). The demographic, clinical, biochemical, and angiographic parameters of ACS patients in these two groups are compiled in [fig_ref] Table 2: Demographic, clinical, biochemical and angiographic characteristics in low and intermediate-high SYNTAX score [/fig_ref]. IL-6 (P < 0.001) and hs-CRP (P < 0.001) levels in the intermediate-high SS group were significantly elevated in comparison to those in the low SS group, and the SS II and residual SS (rSS) values for cases in the intermediate-high SS group were also greater compared to those in the low SS group (P < 0.05). Spearman's correlation analyses revealed IL-6 levels and SS values to be significantly positively associated with one another (r = 0.479, P < 0.001; . Additionally, IL-6 levels were not correlated with post-PCI troponin values (r = 0.107, P > 0.05; Additional file 1: .
## Identification of independent predictors of intermediate-high syntax score
Multivariable logistic regression analyses were next conducted to explore independent predictors of intermediate-high SS by analyzing all variables that exhibited significant predictive value (P < 0.1) in univariate assessment, including, IL-6, hs-CRP, and albumin. The outcomes of this analysis revealed that serum IL-6 levels (odds ratio [OR] = 1.081, 95% confidence interval [CI]: 1.036-1.128, P < 0.001) were independent predictors of intermediate-high SS [fig_ref] Table 4: Independent predictors of intermediate-high SYNTAX score [/fig_ref]. Consistently, an ROC curve analysis for IL-6 yielded an AUC of 0.806 (95% [fig_ref] Table 3: Demographic, clinical, biochemical and angiographic characteristics in low and high SYNTAX score... [/fig_ref]. Additionally, no differences in rSS, history of hypertension or diabetes, or other characteristics were observed when comparing these two groups. IL-6 levels were also positively associated with SS II (r = 0.305, P < 0.001; . After univariate analyses of associated parameters, several variables, including IL-6, hs-CRP, hemoglobin, and NLR, were then incorporated into a multivariable logistic regression analysis, which revealed that IL-6 levels (OR = 1.082, 95% CI: 1.025-1.143, P < 0.01) and hemoglobin levels (OR = 0.948, 95% CI: 0.924-0.972, P < 0.001) were independent predictors of high SS II values [fig_ref] Table 4: Independent predictors of intermediate-high SYNTAX score [/fig_ref]. Consistently, ROC curves demonstrated the value of IL-6 as a predictor of high SS II (AUC = 0.624, P < 0.01; . An IL-6 > concentration greater than 12.9 pg/ml was predictive of high SS II, with sensitivity and specificity of 30.69% and 92.00%, respectively.
# Discussion
ACS remains the most prominent threat to global public health, despite advances in revascularization techniques and antithrombotic therapy [bib_ref] Heart disease and stroke statistics-2018 update: a report from the American Heart..., Benjamin [/bib_ref] [bib_ref] Acute myocardial infarction, Anderson [/bib_ref]. It is thus essential that approaches to reliably predicting ACS severity be developed in order to guide the prevention, diagnosis and treatment of this debilitating disease. Herein, we found that IL-6 levels were positively correlated with ACS severity as measured by SS and SS II, with IL-6 levels additionally offering value as an independent predictor of intermediate-high SS and high SS II values. A growing body of evidence suggests that inflammation is a key driver of ACS onset, progression, and patient prognosis [bib_ref] Progress and challenges in translating the biology of atherosclerosis, Libby [/bib_ref] [bib_ref] Interleukin-6 and the risk of adverse outcomes in patients after an acute..., Fanola [/bib_ref] [bib_ref] Targeting cardiovascular inflammation: next steps in clinical translation, Lawler [/bib_ref]. Likewise, cytokines, regarded as the "messengers" of inflammatory response, have been implicated in the pathogenesis of atherosclerosis and CAD [bib_ref] Interleukin-6 and the risk of adverse outcomes in patients after an acute..., Fanola [/bib_ref] [bib_ref] Association between IL-6 and the extent of coronary atherosclerosis in the veterans..., Saremi [/bib_ref] [bib_ref] Serum cytokine profile in relation to the severity of coronary artery disease, Min [/bib_ref]. Of note, IL-6, which is primarily derived from mononuclear cells, can contribute to CAD initiation and progression through several mechanisms. Not only does IL-6 mainly initiate the production of hepatic CRP, resulting in increased blood viscosity and platelets numbers; it also accelerates the deposition of fibrinogen [bib_ref] Association between serum interleukin-6 concentration and mortality in patients with coronary artery..., Su [/bib_ref]. IL-6 can further stimulate macrophages to phagocytose lipids, thereby driving foam cell formation [bib_ref] Relationship among IL-6, LDL cholesterol and lipid peroxidation, Lubrano [/bib_ref]. There is also evidence to suggest that IL-6 is capable of activating the hypothalamic-pituitary-adrenal axis and accelerating insulin resistance [bib_ref] Interleukin-6 response to insulin-induced hypoglycemia is associated with hypothalamic-pituitary-adrenal axis activation, Drummond [/bib_ref].
In prior studies, higher IL-6 levels in healthy males were associated with future myocardial infarction incidence [bib_ref] Plasma concentration of interleukin-6 and the risk of future myocardial infarction among..., Ridker [/bib_ref]. Additionally, Ikeda et al. demonstrated that ACS patients exhibit substantially higher levels of circulating IL-6 as compared to stable angina patients [bib_ref] Interleukin-6 and acute coronary syndrome, Ikeda [/bib_ref]. Moreover, CAD patients exhibit higher serum IL-6 concentrations as compared to controls [bib_ref] Plasma levels of C1q/TNF-related protein 1 and interleukin 6 in patients with..., Tang [/bib_ref]. One previous study reported that no statistically significant differences in IL-6 levels were observed in the blood of ACS patients taken from the coronary sinus in comparison to blood taken from a peripheral vein, supporting the concept of a systemic rather than a local vascular inflammation contributing to the development of atherosclerosis [bib_ref] Time course of systemic markers of inflammation in patients presenting with acute..., Brueckmann [/bib_ref]. Herein, we similarly observed higher serum IL-6 The SS is a valuable tool that can guide appropriate revascularization planning by aiding in the detection of high-risk ACS cases, and it is also linked to the complexity of atherosclerotic lesions. Furthermore, SS values can effectively predict adverse cardiovascular event risk [bib_ref] Prognostic value of the SYNTAX score in patients with acute coronary syndromes..., Palmerini [/bib_ref] [bib_ref] The Syntax score predicts peri-procedural myocardial necrosis during percutaneous coronary intervention, Van Gaal [/bib_ref] [bib_ref] Value of the SYNTAX score in patients treated by primary percutaneous coronary..., Magro [/bib_ref]. Additionally, several studies have demonstrated that a correlation exists between IL-6 levels and the severity of coronary stenoses and mortality [bib_ref] Interleukin-6 and the risk of adverse outcomes in patients after an acute..., Fanola [/bib_ref] [bib_ref] Association between serum interleukin-6 concentration and mortality in patients with coronary artery..., Su [/bib_ref]. Indeed, there have also been prior reports of a positive association between IL-6 levels and the severity of CAD as assessed based on the Gensini score, which is one of the most common scoring systems for quantifying CAD severity [bib_ref] Serum cytokine profile in relation to the severity of coronary artery disease, Min [/bib_ref] [bib_ref] Serum cytokine tumor necrosis factor-alpha and interleukin-6 associated with the severity of..., Gotsman [/bib_ref]. These prior results suggested a likely correlation between serum IL-6 levels and ACS severity as measured using SS and SS II values. Consistently, we found that IL-6 concentrations were independently predictive of intermediate-high SS values, with which they were positively correlated. Given that a range of clinicopathological variables can influence patient prognosis, the SS II scoring system was developed as a more reliable predictor of cardiovascular events among ACS patients by taking these variables into consideration [bib_ref] Anatomical and clinical characteristics to guide decision making between coronary artery bypass..., Farooq [/bib_ref] [bib_ref] Usefulness of the SYNTAX score II to predict 1-year outcome in patients..., Wang [/bib_ref]. We detected a positive relationship between IL-6 levels and SS II values. Moreover, we identified hemoglobin concentrations and IL-6 levels to be independent predictors of high SS II. This relationship between hemoglobin and SS II values was in accordance with a previous study [bib_ref] Assessment of relationship between C-reactive protein to albumin ratio and coronary artery..., Çağdaş [/bib_ref]. Other reports have also found that factors, including older age, decreased LVEF, and eGFR, can increase SS II values [bib_ref] Age as a modulator of inflammatory cardiovascular risk factors, Anuurad [/bib_ref] [bib_ref] Inflammatory cytokines as biomarkers in heart failure, Ueland [/bib_ref] [bib_ref] The progress of inflammation and oxidative stress in patients with chronic kidney..., Xu [/bib_ref]. Patients with anemia more frequently present with advanced age, higher creatinine levels, and LVEF dysfunction [bib_ref] Relationship between baseline haemoglobin and major bleeding complications in acute coronary syndromes, Bassand [/bib_ref] [bib_ref] Impact of anaemia on long-term outcomes in patients treated with firstand second-generation..., Wańha [/bib_ref]. Therefore, the association between decreased hemoglobin and a high SS may be attributable to older age, reduced eGFR, and lower LVEF in these patients.
# Limitations
There are certain limitations to the present study. For one, a high degree of variability with respect to the IL-6 levels of ACS patients was observed, suggesting that measurements taken at a single time point may be insufficient to reliably capture the extent of potential inflammatory activity. Besides that, we examined the association between the IL-6 and SS or SS II, however, it is difficult to make causal inferences due to the nature of the cross-sectional design. A mendelian randomization study may be further designed to explore this association. Additionally, this was a retrospective single-center analysis without the potential to avoid selection bias. Several confounding factors might have affected the results even after the adjusted analysis. Furthermore, data regarding SS values < 30 and IL-6 levels may be skewed, thus weakening the observed correlations. Lastly, this was a retrospective analysis with relatively few samples, underscoring the need for future large-scale prospective analyses designed to validate and expand upon these results.
[fig] Figure 1 AFigure 2 A: Correlation between IL-6 levels and SS in ACS patients; B Correlation between IL-6 levels and SS II in ACS patients. IL-6, interleukin 6; SS, SYNergy between Percutaneous Coronary Intervention with TAXus and cardiac surgery (SYNTAX) score; SS II, SYNTAX score II; The receiver operating characteristic (ROC) cure of IL-6 in predicting SS > 22 in ACS patients; B ROC curve of IL-6 in predicting SS II > 25.4 in ACS patients. AUC, area under the curve; SS, SYNergy between Percutaneous Coronary Intervention with TAXus and cardiac surgery (SYNTAX) score; SS II, SYNTAX score II concentrations in ACS patients relative to those in the SAP group. [/fig]
[table] Table 1: Demographic, clinical and biochemical characteristics between ACS and SAP group ACS, acute coronary syndrome; SAP, stable angina pectoris; BMI, Body Mass Index; WBC, white blood cell; NEUT, neutrophil; LYM, lymphocyte; NLR, the neutrophil to the lymphocyte ratio; RDW, red cell distribution width; IL-6, interleukin 6; hs-CRP, high-sensitivity C-reactive protein; TC, total cholesterol; HDL-c, high-density lipoprotein cholesterol; LDL-c, low-density lipoprotein cholesterol; apoB, apolipoprotein B; apoA1, apolipoprotein A1; apoB/apoA1, the apoB to the apoA1 ratio; Lp(a), lipoprotein a; eGFR, estimated glomerular filtration rate; LVEF, left ventricular ejection fraction [/table]
[table] Table 2: Demographic, clinical, biochemical and angiographic characteristics in low and intermediate-high SYNTAX score (SS) group BMI, Body Mass Index; WBC, white blood cell; NEUT, neutrophil; LYM, lymphocyte; NLR, the neutrophil to the lymphocyte ratio; RDW, red cell distribution width; IL-6, interleukin 6; hs-CRP, high-sensitivity C-reactive protein; TC, total cholesterol; HDL-c, high-density lipoprotein cholesterol; LDL-c, lowdensity lipoprotein cholesterol; apoB, apolipoprotein B; apoA1, apolipoprotein A1; apoB/apoA1, the apoB to the apoA1 ratio; eGFR, estimated glomerular filtration rate; LVEF, left ventricular ejection fraction; SYNTAX, SYNergy between Percutaneous Coronary Intervention with TAXus and cardiac surgery; SS, SYNTAX score; SS II, SYNTAX score II; rSS, residual SYNTAX score [/table]
[table] Table 3: Demographic, clinical, biochemical and angiographic characteristics in low and high SYNTAX score II group BMI, Body Mass Index; WBC, white blood cell; NEUT, neutrophil; LYM, lymphocyte; NLR, the neutrophil to the lymphocyte ratio; RDW, red cell distribution width; IL-6, interleukin 6; hs-CRP, high-sensitivity C-reactive protein; TC, total cholesterol; HDL-c, high-density lipoprotein cholesterol; LDL-c, lowdensity lipoprotein cholesterol; apoB, apolipoprotein B; apoA1, apolipoprotein A1; apoB/apoA1, the apoB to the apoA1 ratio; eGFR, estimated glomerular filtration rate; LVEF, left ventricular ejection fraction; SYNTAX, SYNergy between Percutaneous Coronary Intervention with TAXus and cardiac surgery; SS, SYNTAX score; SS II, SYNTAX score II [/table]
[table] Table 4: Independent predictors of intermediate-high SYNTAX score (> 22) and high SYNTAX score II (> 25.4)IL-6, interleukin 6; hs-CRP, high-sensitivity C-reactive protein; NLR, the neutrophil to the lymphocyte ratio; OR, odds ratio; SYNTAX, SYNergy between Percutaneous Coronary Intervention with TAXus and cardiac surgery [/table]
|
Bone Infarcts and Tumorigenesis—Is There a Connection? A Mini-Mapping Review
Citation: Konarski, W.; Poboży, T.; Hordowicz, M.;Śliwczyński, A.; Kotela, I.; Krakowiak, J.; Kotela, A.
# Introduction
Bone infarct, also known as aseptic or avascular necrosis of the bone (AVN), is characterized by osteocytes and bone marrow element death that results from inadequate blood supply, which causes local ischemia [bib_ref] Isolated bone infarct of the calcaneus after fracture, Bui-Mansfield [/bib_ref]. The disease most often affects bone epiphyses but may also affect metaphyses and diaphyses. Bone epiphyses are more vulnerable to necrosis due to the lack of connection between the bone epiphyses and local blood vessels and the consequent lack of collateral circulation leads to bone ischemia. Disturbed circulation and ischemia lead to the necrosis of osteocytes and damage to the bone structure. Given that no pathogenic microorganisms are involved in this process, it is also known as sterile bone necrosis.
Sterile bone necrosis encompasses nearly forty different conditions. Nonetheless, they all are characterized by similar anatomopathological lesions and clinical courses [bib_ref] Osteonecrosis of the femoral head: Pathophysiology and current concepts of treatment, Petek [/bib_ref]. The most common types of bone necrosis are shown in . . The most common types of bone necrosis [bib_ref] Osteonecrosis of the femoral head: Pathophysiology and current concepts of treatment, Petek [/bib_ref] [bib_ref] Freiberg's Infarction as the First Clinical Presentation of Sneddon Syndrome, Samanta [/bib_ref] [bib_ref] Arthroscopic Resection of Symptomatic Tibial Tubercle Ossicles for Recalcitrant Osgood-Schlatter Disease Using..., Mcdonough [/bib_ref] [bib_ref] Legg-Calvé-Perthes Disease, Tai [/bib_ref].
## Disease name bones affected by the disease
Scheuermann's disease vertebral body border plates Haglund's syndrome exostosis of the posterior calcaneal tuberosity Mueller-Weiss syndrome tarsal navicular bone Freiberg disease 2nd and 3rd metatarsal head Osgood Schlatter disease patellar tendon insertion on the anterior tibial tuberosity Legg-Calvé-Perthes disease femoral head
## Bone infarcts-clinical picture and diagnosis
The primary symptom of bone infarcts is a pain in the affected bone area [bib_ref] Guidelines for clinical diagnosis and treatment of osteonecrosis of the femoral head..., Zhao [/bib_ref]. Initially, the pain intensifies during physical activity and disappears when it is discontinued but may be present at rest in more chronic cases. In some cases, pain may be accompanied by reduced mobility in the illness joint [bib_ref] Vascular evaluation after cervical hip fractures in children: A case series of..., Juréus [/bib_ref].
The patient's history and imaging findings are usually unambiguous enough to establish an AVN diagnosis and rule out other causes of joint and bone pain. X-ray allows for diagnosis confirmation in advanced cases but may be useful in the initial differential diagnosis [bib_ref] Guidelines for clinical diagnosis and treatment of osteonecrosis of the femoral head..., Zhao [/bib_ref] [bib_ref] Vascular evaluation after cervical hip fractures in children: A case series of..., Juréus [/bib_ref]. Bone scintigraphy offers the advantage of detecting abnormalities present at the earlier stages of the disease thana classic X-ray. The affected bone tissue presents with "donut-like" changes ('cold' in 'hot'). Nonetheless, its use is limited because of a specificity lower than MRI. MRI has both higher sensitivity and specificity than X-ray and allows for identifying early signs of the disease [bib_ref] Guidelines for clinical diagnosis and treatment of osteonecrosis of the femoral head..., Zhao [/bib_ref]. Still, positron emission tomography (PET) is superior to MRI as a more sensitive imaging method, detecting early changes and allowing for the prediction of disease prognosis [bib_ref] F-18 fluoride positron emission tomography of the hip for osteonecrosis, Dasa [/bib_ref] [bib_ref] Identifying Patients Who Will Most Benefit from Single Photon Emission Computerized Tomography..., Fan [/bib_ref]. Its use is limited by low availability in comparison with the aforementioned methods.
## Treatment and management
The main aim of bone infarct treatment is the prevention of further loss of bone mass. The choice of intervention depends on the severity of the bone damage. Pharmacological treatment is mainly symptomatic [bib_ref] Nontraumatic Osteonecrosis of the femoral head: Where do we stand today?: A..., Mont [/bib_ref] [bib_ref] Vascular evaluation after cervical hip fractures in children: A case series of..., Juréus [/bib_ref]. Studies regarding nonsurgical treatment included small groups, and their quality is low-therefore, they might be regarded as experimental [bib_ref] Nontraumatic Osteonecrosis of the femoral head: Where do we stand today?: A..., Mont [/bib_ref]. The medications commonly used in the treatment of AVN are listed in . . Medications used in the treatment of AVN [bib_ref] Guidelines for clinical diagnosis and treatment of osteonecrosis of the femoral head..., Zhao [/bib_ref] [bib_ref] Nontraumatic Osteonecrosis of the femoral head: Where do we stand today?: A..., Mont [/bib_ref] [bib_ref] Vascular evaluation after cervical hip fractures in children: A case series of..., Juréus [/bib_ref].
## Classes of drugs examples role in osteonecrosis management
Non-steroidal anti-inflammatory drugs Ibuprofen or naproxen Help relieve pain and inflammation associated with AVN Most patients who suffer from AVN will seek help in the advanced stages of the disease when the symptoms begin to interfere with their activities. In such patients, surgical treatment might be considered [bib_ref] Treatment for avascular necrosis of bone in people with sickle cell disease, Martí-Carvajal [/bib_ref]. Selected surgical procedures used to treat AVN are presented in . . Surgical procedures used to treat AVN [bib_ref] Management of femoral head osteonecrosis: Current concepts, Tripathy [/bib_ref] [bib_ref] Guidelines for clinical diagnosis and treatment of osteonecrosis of the femoral head..., Zhao [/bib_ref] [bib_ref] Nontraumatic Osteonecrosis of the femoral head: Where do we stand today?: A..., Mont [/bib_ref] [bib_ref] Management of osteonecrosis of the femoral head in children with sickle cell..., Mallet [/bib_ref] [bib_ref] Treatment for avascular necrosis of bone in people with sickle cell disease, Martí-Carvajal [/bib_ref].
## Treatment characteristics
## Spinal decompression
During this procedure, the surgeon removes part of the inner layer of bone. In addition to reducing pain, this treatment has the effect of stimulating osteogenesis and neovascularization.
## Bone graft (transplant)
The procedure helps to strengthen the area of bone affected by the lesions. During the procedure, some healthy bone taken from another part of the body is used.
## Bone osteotomy
During this procedure, a bone wedge above or below the stressed joint is removed-This helps to shift weight away from the damaged bone. Changing the shape of the bone may allow the joint replacement surgery to be pushed back.
## Joint replacement (alloplasty)
This treatment is used when other treatments do not help; it involves replacing the damaged parts of the joint with plastic or metal parts.
## Possible associations between tumorigenesis and bone infarct
Several patient cases were published in the past linking osteonecrosis to malignant diseases. The first case of an infarct-associated sarcoma was described in 1960. This was followed by several other reports, though most describe single-patient cases or relatively small case series. Bone infarcts were portrayed as the primary cause of the tumorigenesis or secondary to it [bib_ref] Case report 120, probable bone infarcts in long bones secondary to pheochromocytoma, Hernandez [/bib_ref] [bib_ref] A p53 gene mutation in malignant fibrous histiocytoma associated with bone infarction, Yamamoto [/bib_ref]. Infarct-related sarcomas are extremely rare even when compared with bone malignancies secondary to Paget's disease or radiation [bib_ref] Radiation and pagetic osteogenic sarcomas, Healey [/bib_ref] [bib_ref] Prognosis of radiation-induced bone sarcoma is similar to primary osteosarcoma, Shaheen [/bib_ref]. Diagnosis and management are challenging, given the disease's rarity and the paucity of available data to guide clinicians [bib_ref] Bone infarct-associated osteosarcoma, Torres [/bib_ref] [bib_ref] Infarct-associated bone sarcomas, Domson [/bib_ref]. In addition, little is known about the pathogenesis of the disease, risk factors for the malignant transformation of the necrotic tissue, and its natural course. The last overview of the published literature, by . Therefore, we reviewed and synthesized the data on infarct-associated tumors from the last decade, intending to identify if there have been any advancements in the treatment of the disease, diagnosis, and pathogenesis, including risk factors for the development of malignancy in necrotic bone.
# Materials and methods
A mapping review of the available literature was performed. Such studies aim to screen the available literature systematically, prepare an overview of open data, and identify knowledge gaps [bib_ref] A typology of reviews: An analysis of 14 review types and associated..., Grant [/bib_ref]. Currently, no guidelines apply specifically to mapping reviews. We adopted a systematic search strategy to maximize the chances of identifying relevant papers. However, questions used in mapping reviews are less detailed and broader than in the case of systematic reviews, and the data is summarized narratively without grading its quality.
Two independent reviewers screened and selected relevant publications by reviewing their abstracts and titles. The papers were then checked against the research questions. Due to the two reviewers working in parallel, we did not remove duplicates before screening to maximize the chances of identifying all relevant papers. A narrative synthesis of the results was then conducted.
## Research questions
What studies are available discussing a link between cancerogenesis and bone infarcts? Which malignancies may develop in the region affected by AVN? What is known about the pathomechanisms of tumorigenesis in necrotic bone? What are the radiologic findings indicative of an infarct-related malignancy? What management is preferred in these patients? What are the outcomes of treatment and survival rates in such patients?
## Search strategy and eligibility criteria
We searched the PubMed, Google Scholar, and Cochrane databases using search phrases with keywords related to tumorigenesis, cancer, and synonyms. The results were limited to the last ten years (2012-2022). We excluded all papers in languages other than Polish, English, and Spanish. The search was performed on the 21 July 2022. Gray literature and the references of articles included in the review were also checked to identify other papers meeting the inclusion criteria.
The complete search phrases for PubMed were the following:
1. "bone infarct*" AND (cancer OR "cancer treatment" OR "cancer patient*" OR radiotherapy OR leukemia OR neoplasm* OR carcinogenesis OR tumorigenesis OR sarcoma OR osteosarcoma) 2.
("avascular necrosis" OR AVN) AND (cancer OR "cancer treatment" OR "cancer patient*" OR radiotherapy OR leukemia OR neoplasm* OR carcinogenesis OR tumorigenesis OR sarcoma OR osteosarcoma).
The search strategy for Google Scholar and the Cochrane database are described in the Supplementary File S1.
The full inclusion and exclusion criteria are listed in [fig_ref] Table 4: Inclusion and exclusion criteria [/fig_ref].
# Results
The search yielded a total of 233 results. Among these, eight met the inclusion criteria [fig_ref] Table 5: Overview of studies. [/fig_ref]. The gray literature search allowed for the identification of one other case report. A flow diagram with the reasons for the exclusion of the screened publications in Pubmed is shown in [fig_ref] Figure 1: Review of the literature flow diagram of PubMed publications [/fig_ref]. In the Cochrane database, 212 results met the search criteria, including 6 Cochrane reviews and 206 trials, none of which met the inclusion criteria. In the Google Search, a total of 14,615 results were found; three additional papers were identified. No new papers were found through the screening of the references.
## Animal and preclin
# Results
The search yielded a total of 233 results. Among these, eight m [fig_ref] Table 5: Overview of studies. [/fig_ref]. The gray literature search allowed for the identification A flow diagram with the reasons for the exclusion of the screened is shown in [fig_ref] Figure 1: Review of the literature flow diagram of PubMed publications [/fig_ref]. In the Cochrane database, 212 results m including 6 Cochrane reviews and 206 trials, none of which met the Google Search, a total of 14,615 results were found; three identified. No new papers were found through the screening of th
## Research question 1: what studies discuss a link between cancer infarcts?
Among the twelve studies meeting the inclusion criteria [ reports [bib_ref] Bone infarct transformation into undifferentiated pleomorphic sarcoma in sickle cell disease: A..., Alhamdan [/bib_ref] [bib_ref] Low-grade central osteosarcoma arising from bone infarct, Endo [/bib_ref] [bib_ref] Low-Grade Central Osteosarcoma Arising from Bone Infarct: A Case, Goel [/bib_ref] [bib_ref] First report of myxofibrosarcoma of bone arising at a bone infarct, Kayser [/bib_ref] [bib_ref] Chondroblastic Osteosarcoma Arising in a Bone Infarct in a Patient with a..., Lewin [/bib_ref] [bib_ref] A Rare Case of an Osteolytic Bone-infarct-associated Osteosarcoma: Case Report with Radiographic..., Mcdonald [/bib_ref] [bib_ref] Chronic knee pain in an 80-year-old woman, Robbin [/bib_ref] [bib_ref] Secondary multifocal osteosarcoma developing on the background of bone infarct, Sivrioglu [/bib_ref]. Three were case series-one focused prim [bib_ref] Bone Infarct-Associated Osteosarcoma: Epidemiologic and Survival Trends, Laranga [/bib_ref] , another on radiographic findings [bib_ref] Infarct-Associated Bone Sarcomas: Multimodality Imaging Findings, Stacy [/bib_ref]. The remaining on secondary osteosarcomas from a single institution, of which one previous bone infarct [bib_ref] Cetintas sk2, muhammed sadik bilgen ms. Secondary osteosarcomas diagnosed in a single..., Yalcinkaya [/bib_ref]. An overview of these studies is presen
## Research question 1: what studies discuss a link between cancerogenesis and bone infarcts?
Among the twelve studies meeting the inclusion criteria [bib_ref] Bone infarct transformation into undifferentiated pleomorphic sarcoma in sickle cell disease: A..., Alhamdan [/bib_ref] [bib_ref] Low-grade central osteosarcoma arising from bone infarct, Endo [/bib_ref] [bib_ref] Low-Grade Central Osteosarcoma Arising from Bone Infarct: A Case, Goel [/bib_ref] [bib_ref] First report of myxofibrosarcoma of bone arising at a bone infarct, Kayser [/bib_ref] [bib_ref] Bone Infarct-Associated Osteosarcoma: Epidemiologic and Survival Trends, Laranga [/bib_ref] [bib_ref] Chondroblastic Osteosarcoma Arising in a Bone Infarct in a Patient with a..., Lewin [/bib_ref] [bib_ref] A Rare Case of an Osteolytic Bone-infarct-associated Osteosarcoma: Case Report with Radiographic..., Mcdonald [/bib_ref] [bib_ref] Chronic knee pain in an 80-year-old woman, Robbin [/bib_ref] [bib_ref] Secondary multifocal osteosarcoma developing on the background of bone infarct, Sivrioglu [/bib_ref] [bib_ref] Infarct-Associated Bone Sarcomas: Multimodality Imaging Findings, Stacy [/bib_ref] [bib_ref] Cetintas sk2, muhammed sadik bilgen ms. Secondary osteosarcomas diagnosed in a single..., Yalcinkaya [/bib_ref] , nine were case reports [bib_ref] Bone infarct transformation into undifferentiated pleomorphic sarcoma in sickle cell disease: A..., Alhamdan [/bib_ref] [bib_ref] Low-grade central osteosarcoma arising from bone infarct, Endo [/bib_ref] [bib_ref] Low-Grade Central Osteosarcoma Arising from Bone Infarct: A Case, Goel [/bib_ref] [bib_ref] First report of myxofibrosarcoma of bone arising at a bone infarct, Kayser [/bib_ref] [bib_ref] Chondroblastic Osteosarcoma Arising in a Bone Infarct in a Patient with a..., Lewin [/bib_ref] [bib_ref] A Rare Case of an Osteolytic Bone-infarct-associated Osteosarcoma: Case Report with Radiographic..., Mcdonald [/bib_ref] [bib_ref] Chronic knee pain in an 80-year-old woman, Robbin [/bib_ref] [bib_ref] Secondary multifocal osteosarcoma developing on the background of bone infarct, Sivrioglu [/bib_ref]. Three were case series-one focused primarily on epidemiology [bib_ref] Bone Infarct-Associated Osteosarcoma: Epidemiologic and Survival Trends, Laranga [/bib_ref] , another on radiographic findings [bib_ref] Infarct-Associated Bone Sarcomas: Multimodality Imaging Findings, Stacy [/bib_ref]. The remaining one was a case series of secondary osteosarcomas from a single institution, of which one had developed from a previous bone infarct [bib_ref] Cetintas sk2, muhammed sadik bilgen ms. Secondary osteosarcomas diagnosed in a single..., Yalcinkaya [/bib_ref]. An overview of these studies is presented in [fig_ref] Table 5: Overview of studies. [/fig_ref]. A patient with sickle cell anemia was diagnosed initially with AVN. He refused to undergo THR; therefore, it was managed conservatively. Three years later, he presented with increasing pain in the left thigh. On radiographs, multiple bone infarcts were detected in both femurs, as well as in shoulders and hip joints. There were no metastases found in the chest CT and bone scintigraphy. He underwent a proximal femur resection with prosthetic reconstruction. When preparing the report, the patient was still receiving ChT.
## Mfh proximal femur
Endo M., 2012 [bib_ref] Low-grade central osteosarcoma arising from bone infarct, Endo [/bib_ref] Case report 23-year-old female A patient with no known risk factors for AVN was followed up for 13 years using X-ray due to an infarct-like lesion in her right humerus. It was initially assessed as a benign lesion and was detected accidentally during examination for other causes. At 36-years old, the mass begun to protrude through the bone and was palpable. She underwent a joint replacement surgery and tumor resection. At a 4-year follow-up, she did not have any signs of disease progression or recurrence.
## Mfh humerus
Goel R., 2018 [bib_ref] Low-Grade Central Osteosarcoma Arising from Bone Infarct: A Case, Goel [/bib_ref] Case report 65-year-old female Patient with multiple risk factors for AVN presented with a restriction in flexion in the at-risk knee. Physical examination revealed a 10 × 12 cm soft tissue mass in the lower right thigh. Bone biopsy findings and expression of MDM2 and CDK4 was indicative of low-grade osteosarcoma. [bib_ref] Chondroblastic Osteosarcoma Arising in a Bone Infarct in a Patient with a..., Lewin [/bib_ref] Case report 29-year-old male Patient with bone infarct related to corticosteroid use due to Hodgins's lymphoma in childhood. At the time of writing the paper, he was still in treatment with ChT before potential limb-sparing surgery. He will undergo evaluation and qualification for a surgical procedure after ChT.
Chondroblastic osteosarcoma Distal femur Patient presented with pain in his right humerus, which increased after hearing a crack while dressing. He underwent X-ray, which showed a minimally displaced humeral fracture and a sclerotic lesion. His pain worsened over the next 2 weeks, and follow-up examination with MRI and X-ray showed a lytic lesion with periosteal reaction and a mature bone infarct. Patient did not respond well to chemotherapy and underwent an amputation. After two months, he developed metastases in lungs and lumber spine. He died 7 months after the initial presentation.
## Research question 2: which malignancies may develop in the region affected by avn?
Different histological types of secondary sarcomas were found to arise from bone infarcts in the papers eligible for this review. These include malignant fibrous histiocytoma (MFH), fibrosarcoma, myxofibrosarcoma, and osteosarcoma [bib_ref] Bone infarct transformation into undifferentiated pleomorphic sarcoma in sickle cell disease: A..., Alhamdan [/bib_ref] [bib_ref] Low-grade central osteosarcoma arising from bone infarct, Endo [/bib_ref] [bib_ref] Low-Grade Central Osteosarcoma Arising from Bone Infarct: A Case, Goel [/bib_ref] [bib_ref] First report of myxofibrosarcoma of bone arising at a bone infarct, Kayser [/bib_ref] [bib_ref] Bone Infarct-Associated Osteosarcoma: Epidemiologic and Survival Trends, Laranga [/bib_ref] [bib_ref] Chondroblastic Osteosarcoma Arising in a Bone Infarct in a Patient with a..., Lewin [/bib_ref] [bib_ref] A Rare Case of an Osteolytic Bone-infarct-associated Osteosarcoma: Case Report with Radiographic..., Mcdonald [/bib_ref] [bib_ref] Chronic knee pain in an 80-year-old woman, Robbin [/bib_ref] [bib_ref] Secondary multifocal osteosarcoma developing on the background of bone infarct, Sivrioglu [/bib_ref] [bib_ref] Infarct-Associated Bone Sarcomas: Multimodality Imaging Findings, Stacy [/bib_ref] [bib_ref] Cetintas sk2, muhammed sadik bilgen ms. Secondary osteosarcomas diagnosed in a single..., Yalcinkaya [/bib_ref]. Multifocal lesions, or the asynchronous development of malignancy in more than one bone in a single patient, were also described [bib_ref] Infarct-Associated Bone Sarcomas: Multimodality Imaging Findings, Stacy [/bib_ref]. Nonetheless, given that most studies were short case series or single case descriptions, the list of infarct-related tumors might not be complete.
Though it is a rare primary bone tumor, MFH was the most common type of malignant tumor secondary to bone infarction. The World Health Organization recently reclassified it as pleomorphic undifferentiated sarcoma. Immunohistological staining should be negative for S100 protein, cytokeratin, desmin, and muscle-specific actin [bib_ref] Low-grade central osteosarcoma arising from bone infarct, Endo [/bib_ref] [bib_ref] Chronic knee pain in an 80-year-old woman, Robbin [/bib_ref]. Some authors suggest that such a high prevalence of MFH in association with AVN indicates that there might be a causal relation [bib_ref] Low-grade central osteosarcoma arising from bone infarct, Endo [/bib_ref] [bib_ref] Chronic knee pain in an 80-year-old woman, Robbin [/bib_ref]. MFH is most prevalent in patients in the 6th decade of life and older and is twice as common in men as in women. Most cases are diagnosed at the disseminated stage, and the survival rates are relatively poor [bib_ref] Chronic knee pain in an 80-year-old woman, Robbin [/bib_ref]. MFH was also found in 7 out of 9 patients described in the Stacy et al. case series [bib_ref] Infarct-Associated Bone Sarcomas: Multimodality Imaging Findings, Stacy [/bib_ref].
Only one case of myxofibrosarcoma, which developed within a bone infarct, was described [bib_ref] First report of myxofibrosarcoma of bone arising at a bone infarct, Kayser [/bib_ref]. The diagnosis was confirmed with immunostaining for S-100, SOX10, and MDM2, which excluded tumors originating from the neural crest and liposarcoma [bib_ref] First report of myxofibrosarcoma of bone arising at a bone infarct, Kayser [/bib_ref].
Myfoxibrosarcoma arises mainly from the soft tissue and is characterized by spindle cells in a gelatinous (myxoid) background. The authors suggest that this kind of tumor, similarly to MFH, has an affinity for areas of bone affected by osteonecrosis [bib_ref] First report of myxofibrosarcoma of bone arising at a bone infarct, Kayser [/bib_ref]. Another case of rare, low-grade sarcoma developing in a previous bone infarct was described by Endo et al. This kind of tumor accounts for around 1% of all osteosarcomas [bib_ref] Low-grade central osteosarcoma arising from bone infarct, Endo [/bib_ref].
## Research question 3: what is known about the pathomechanisms of tumorigenesis in necrotic bone?
Bone infarcts might affect the epiphysis, diaphysis, and metaphysis of long bones [bib_ref] Secondary multifocal osteosarcoma developing on the background of bone infarct, Sivrioglu [/bib_ref]. Epiphyseal infarcts are usually symptomatic at an early stage. The remaining two might cause pain initially but frequently are found incidentally during an imaging study ordered for some unrelated concern [bib_ref] Low-Grade Central Osteosarcoma Arising from Bone Infarct: A Case, Goel [/bib_ref]. Given that the malignant transformation of cells is a process that requires considerable time, most secondary tumors occur in the metaphyseal and diaphyseal areas [bib_ref] Low-Grade Central Osteosarcoma Arising from Bone Infarct: A Case, Goel [/bib_ref] [bib_ref] Secondary multifocal osteosarcoma developing on the background of bone infarct, Sivrioglu [/bib_ref]. Tumors appear most frequently in the proximal tibia and distal femur [bib_ref] Infarct-Associated Bone Sarcomas: Multimodality Imaging Findings, Stacy [/bib_ref]. However, one of the included papers, by Endo et al. and McDonald, described a case of the malignant transformation of infarcts in the humeral bone [bib_ref] Low-grade central osteosarcoma arising from bone infarct, Endo [/bib_ref] [bib_ref] A Rare Case of an Osteolytic Bone-infarct-associated Osteosarcoma: Case Report with Radiographic..., Mcdonald [/bib_ref].
Little is known about the pathways of carcinogenesis in necrotic bone. It was previously suggested that transformation occurs due to local inflammation and reparative processes caused by the infarct [bib_ref] Cetintas sk2, muhammed sadik bilgen ms. Secondary osteosarcomas diagnosed in a single..., Yalcinkaya [/bib_ref]. When inflammatory processes become chronic, the likelihood of the malignant transformation of the reparative cells increases. Some histologic types, such as fibrosarcomas and MFH, affect areas where another lesion was already present; these include bone infarcts [bib_ref] First report of myxofibrosarcoma of bone arising at a bone infarct, Kayser [/bib_ref] [bib_ref] Cetintas sk2, muhammed sadik bilgen ms. Secondary osteosarcomas diagnosed in a single..., Yalcinkaya [/bib_ref]. The margin of the infarct, where these processes are intensive, might be more prone to cancerous transformation. Mutations in the p53 gene were proposed to play a role by one of the authors [bib_ref] Cetintas sk2, muhammed sadik bilgen ms. Secondary osteosarcomas diagnosed in a single..., Yalcinkaya [/bib_ref]. Mutations in the p53 gene were also found in the Endo et al. MFH case [bib_ref] Low-grade central osteosarcoma arising from bone infarct, Endo [/bib_ref]. The authors of other reports did not perform a molecular analysis of the cancerous tissue [bib_ref] Low-Grade Central Osteosarcoma Arising from Bone Infarct: A Case, Goel [/bib_ref] [bib_ref] First report of myxofibrosarcoma of bone arising at a bone infarct, Kayser [/bib_ref] [bib_ref] Bone Infarct-Associated Osteosarcoma: Epidemiologic and Survival Trends, Laranga [/bib_ref] [bib_ref] Secondary multifocal osteosarcoma developing on the background of bone infarct, Sivrioglu [/bib_ref] [bib_ref] Infarct-Associated Bone Sarcomas: Multimodality Imaging Findings, Stacy [/bib_ref] [bib_ref] Cetintas sk2, muhammed sadik bilgen ms. Secondary osteosarcomas diagnosed in a single..., Yalcinkaya [/bib_ref].
Stacy et al. stated in their manuscript that MFHs penetrate the bone cortex easier than primary tumors because they arise from the periphery of the centrally located infarct. The avascular, necrotic infarct area makes it relatively immune to tumor penetration [bib_ref] Infarct-Associated Bone Sarcomas: Multimodality Imaging Findings, Stacy [/bib_ref]. Nonetheless, no study to date has identified pathomechanisms to support these hypotheses. Alhamdan et al. suggested that patients with sickle cell trait (SCT) (heterozygous, with fewer symptoms) might have more propensity to develop malignancies in infarcts than homozygotes (i.e., those with sickle cell disease (SCD)). The reparative processes might be more intense in the former group, thus increasing the chance of carcinogenesis [bib_ref] Bone infarct transformation into undifferentiated pleomorphic sarcoma in sickle cell disease: A..., Alhamdan [/bib_ref]. Though the hypothesis is interesting, it was based solely on five cases reported in the literature to date (four with the SCT and one in a SCD patient. Furthermore, given the rarity of the disease and limited data available, it would be difficult to determine if these conditions were genuinely increasing the chances of the malignant transformation of infarcted tissue or if they develop by chance [bib_ref] Bone infarct transformation into undifferentiated pleomorphic sarcoma in sickle cell disease: A..., Alhamdan [/bib_ref] [bib_ref] Chondroblastic Osteosarcoma Arising in a Bone Infarct in a Patient with a..., Lewin [/bib_ref].
## Research question 4: what radiologic findings are indicative of infarct-related malignancy?
Imaging studies are an essential step in the diagnosis of bone tumors. Infarct-associated osteosarcomas are primarily located in the lower limb, especially around the knee joint, less frequently in the proximal femur. Pain, usually around the knee or hip joints, is the leading symptom of bone infarcts and tumors [bib_ref] Bone infarct transformation into undifferentiated pleomorphic sarcoma in sickle cell disease: A..., Alhamdan [/bib_ref] [bib_ref] Bone Infarct-Associated Osteosarcoma: Epidemiologic and Survival Trends, Laranga [/bib_ref] [bib_ref] Chronic knee pain in an 80-year-old woman, Robbin [/bib_ref] [bib_ref] Infarct-Associated Bone Sarcomas: Multimodality Imaging Findings, Stacy [/bib_ref]. X-ray is one of the primary imaging techniques used to determine the cause of pain in this location [bib_ref] Secondary multifocal osteosarcoma developing on the background of bone infarct, Sivrioglu [/bib_ref]. MRI and CT might also be employed in the initial diagnostic process [bib_ref] Chronic knee pain in an 80-year-old woman, Robbin [/bib_ref]. The common finding in infarct-associated sarcomas is the presence of a well-defined sclerotic band with or without a central lucency, which corresponds to a mature bone infarct [bib_ref] A Rare Case of an Osteolytic Bone-infarct-associated Osteosarcoma: Case Report with Radiographic..., Mcdonald [/bib_ref] [bib_ref] Chronic knee pain in an 80-year-old woman, Robbin [/bib_ref]. Nonetheless, the necrotic bone might also appear as a mixed pattern with lucencies and sclerotic changes, with no easily definable margins. Tumor radiological characteristics are described in more detail below.
## Radiographs
On X-ray, an infarcted area is visible in most cases, with a soft-tissue mass adjacent to it. Larger osteosarcomas may displace muscle shadows; on X-ray, they may present with or without aggressive features (including periosteal reaction and cortex disruption) [bib_ref] Secondary multifocal osteosarcoma developing on the background of bone infarct, Sivrioglu [/bib_ref] [bib_ref] Infarct-Associated Bone Sarcomas: Multimodality Imaging Findings, Stacy [/bib_ref]. In some cases, features of an upcoming fracture may also be observed. [bib_ref] Bone infarct transformation into undifferentiated pleomorphic sarcoma in sickle cell disease: A..., Alhamdan [/bib_ref] Infarct-related MFHs, on the other hand, might be overlooked and diagnosed during a follow-up examination or only after a pathological fracture of the bone due to the underlying malignancy [bib_ref] A Rare Case of an Osteolytic Bone-infarct-associated Osteosarcoma: Case Report with Radiographic..., Mcdonald [/bib_ref] [bib_ref] Infarct-Associated Bone Sarcomas: Multimodality Imaging Findings, Stacy [/bib_ref]. They tend to grow eccentrically. A soft-tissue mass might not be visible on radiographs, though discrete cortical thinning and subtle osteolysis signs might indicate MFH's presence [bib_ref] Infarct-Associated Bone Sarcomas: Multimodality Imaging Findings, Stacy [/bib_ref]. In Stacy et al.'s case series, only three out of seven cases could be diagnosed based on the initially taken radiographs [bib_ref] Infarct-Associated Bone Sarcomas: Multimodality Imaging Findings, Stacy [/bib_ref]. In one case of myxofibrosarcoma, the primary finding was that extensive bone destruction by a soft-tissue mass, with periosteal reaction and some signs of mineralization in the epicenter, is indicative of a previous bone infarct [bib_ref] First report of myxofibrosarcoma of bone arising at a bone infarct, Kayser [/bib_ref].
## Computed tomography
Computed tomography (CT) results are also different in the case of osteosarcomas and MFHs. In osteosarcomas, soft-tissue ossification is hyperdense to the skeletal muscle. As the mass penetrates the cortex, signs of cortical destruction are also present. Hyperdense calcifications may also be present within the tumor [bib_ref] Bone infarct transformation into undifferentiated pleomorphic sarcoma in sickle cell disease: A..., Alhamdan [/bib_ref] [bib_ref] Low-grade central osteosarcoma arising from bone infarct, Endo [/bib_ref] [bib_ref] Infarct-Associated Bone Sarcomas: Multimodality Imaging Findings, Stacy [/bib_ref]. A hyperdense osteosclerotic band may separate osteosarcoma from the infarct [bib_ref] Low-grade central osteosarcoma arising from bone infarct, Endo [/bib_ref] [bib_ref] Chronic knee pain in an 80-year-old woman, Robbin [/bib_ref]. For MFHs, a slightly hypodense appearance, in comparison to muscle tissue, is typical [bib_ref] Infarct-Associated Bone Sarcomas: Multimodality Imaging Findings, Stacy [/bib_ref]. A hyperdense, calcified infarct margin might be present, but only residues might be visualized in some cases [bib_ref] Low-grade central osteosarcoma arising from bone infarct, Endo [/bib_ref] [bib_ref] Infarct-Associated Bone Sarcomas: Multimodality Imaging Findings, Stacy [/bib_ref]. In proximity to the tumor, focal osteolysis is also a common finding. Contrast enhancement in MFH and myxofibrosarcoma was moderate; the lesion appeared slightly hyperdense in comparison to the skeletal muscle. A CT scan may also reveal the foci of newer, asynchronous bone infarcts near the tumor or in other bones. For other types, no information about contrast-enhanced CTs was available [bib_ref] Infarct-Associated Bone Sarcomas: Multimodality Imaging Findings, Stacy [/bib_ref].
## Magnetic resonance imaging (mri)
Magnetic resonance is useful in imaging soft-tissue masses, interference of the tumor with surrounding blood vessels and nerves, and other adjacent structures. It enables both measurements of the tumor or the determination of multifocal lesions and the confirmation of the presence of a bone infarct. T1-weighted images of MFHs reveal a mass with mostly homogenous, intermediate, or low signal intensity (similar to cartilage or muscle tissue) [bib_ref] Chronic knee pain in an 80-year-old woman, Robbin [/bib_ref] [bib_ref] Infarct-Associated Bone Sarcomas: Multimodality Imaging Findings, Stacy [/bib_ref]. The margin of a previous bone infarct (with apparent disruptions) was observed in all cases in the series by Stacy et al. The presence of areas of high signal intensity on T1 images, suggesting intratumor bleeding, was found in only one patient. On gadolinium-enhanced T1-weighted images, the tumor signal increased moderately in a non-homogenous manner. On T2-weighted images and fat-suppressed T2 images, MFHs were characterized by high signal intensity [bib_ref] Infarct-Associated Bone Sarcomas: Multimodality Imaging Findings, Stacy [/bib_ref]. In addition, periosteal reaction, cortical destruction, and endosteal resorption might be seen in MFHs [bib_ref] A Rare Case of an Osteolytic Bone-infarct-associated Osteosarcoma: Case Report with Radiographic..., Mcdonald [/bib_ref] [bib_ref] Chronic knee pain in an 80-year-old woman, Robbin [/bib_ref] [bib_ref] Infarct-Associated Bone Sarcomas: Multimodality Imaging Findings, Stacy [/bib_ref]. In the case of sarcomas, a mixed pattern of contrast enhancement is typically found (moderate to high), suggesting intra-tumor necrosis, infarct margins, and focal low-vascularity [bib_ref] Infarct-Associated Bone Sarcomas: Multimodality Imaging Findings, Stacy [/bib_ref].
## Scintigraphy with technetium 99-labeled methylene diphosphonate
The role of scintigraphy is mostly the differentiation of bone infarcts (low radionuclide uptake) from active tumors (high uptake) and the determination of whether the disease is localized or has already disseminated [bib_ref] Bone infarct transformation into undifferentiated pleomorphic sarcoma in sickle cell disease: A..., Alhamdan [/bib_ref] [bib_ref] A Rare Case of an Osteolytic Bone-infarct-associated Osteosarcoma: Case Report with Radiographic..., Mcdonald [/bib_ref] [bib_ref] Infarct-Associated Bone Sarcomas: Multimodality Imaging Findings, Stacy [/bib_ref]. Skeletal scintigraphy might also help detect multifocal lesions [bib_ref] First report of myxofibrosarcoma of bone arising at a bone infarct, Kayser [/bib_ref] [bib_ref] Bone Infarct-Associated Osteosarcoma: Epidemiologic and Survival Trends, Laranga [/bib_ref]. Most sarcomas secondary to bone infarct present an intense radionuclide uptake, regardless of histologic type [bib_ref] Bone infarct transformation into undifferentiated pleomorphic sarcoma in sickle cell disease: A..., Alhamdan [/bib_ref] [bib_ref] Chronic knee pain in an 80-year-old woman, Robbin [/bib_ref] [bib_ref] Infarct-Associated Bone Sarcomas: Multimodality Imaging Findings, Stacy [/bib_ref]. A central portion with a decreased radiotracer uptake might be observed if the tumor arises from an area of extensive bone infarct [bib_ref] Bone infarct transformation into undifferentiated pleomorphic sarcoma in sickle cell disease: A..., Alhamdan [/bib_ref] [bib_ref] Infarct-Associated Bone Sarcomas: Multimodality Imaging Findings, Stacy [/bib_ref]. Nonetheless, its use in diagnosis requires further research.
## Research question 5: what management is preferred in these patients?
Sarcomas are rare tumors, and those that arise from a bone infarct are even rarer. Most cases of sarcomas secondary to bone infarcts require multimodal treatment. Usually, a combination of chemotherapy (ChT) and surgical management is proposed [bib_ref] Bone infarct transformation into undifferentiated pleomorphic sarcoma in sickle cell disease: A..., Alhamdan [/bib_ref] [bib_ref] Low-Grade Central Osteosarcoma Arising from Bone Infarct: A Case, Goel [/bib_ref] [bib_ref] A Rare Case of an Osteolytic Bone-infarct-associated Osteosarcoma: Case Report with Radiographic..., Mcdonald [/bib_ref] [bib_ref] Chronic knee pain in an 80-year-old woman, Robbin [/bib_ref] [bib_ref] Infarct-Associated Bone Sarcomas: Multimodality Imaging Findings, Stacy [/bib_ref]. Neoadjuvant chemotherapy might be used if limb-preserving surgery is planned [bib_ref] Bone Infarct-Associated Osteosarcoma: Epidemiologic and Survival Trends, Laranga [/bib_ref]. Surgical management might include the amputation of a wide resection with endoprosthesis implantation. It may also serve as limb-salvage surgery in patients with a pathological fracture in the preoperative period. Nonetheless, such extensive resection and use of sizeable prosthetic material may provoke an infection. Such a risk must be carefully weighed against possible benefits, especially when immunocompromising ChT is administered.
Spazzioli et al. claimed that adjuvant and neoadjuvant ChT might increase survival rates. Nonetheless, Laranga et al. have shown that the difference in 5-year overall survival between patients treated with ChT vs. surgery was not statistically significant (71% vs. 50%; p = 0.4773), though the survival rates were higher in that study [bib_ref] Bone Infarct-Associated Osteosarcoma: Epidemiologic and Survival Trends, Laranga [/bib_ref].
## Research question 6: what are the outcomes of treatment and survival time in such patients?
Osteosarcomas secondary to bone infarct carry a poor prognosis. Most patients remain symptom-free for a long period, and the diagnosis is established at stage III or IV [bib_ref] Low-grade central osteosarcoma arising from bone infarct, Endo [/bib_ref] [bib_ref] Low-Grade Central Osteosarcoma Arising from Bone Infarct: A Case, Goel [/bib_ref] [bib_ref] First report of myxofibrosarcoma of bone arising at a bone infarct, Kayser [/bib_ref] [bib_ref] A Rare Case of an Osteolytic Bone-infarct-associated Osteosarcoma: Case Report with Radiographic..., Mcdonald [/bib_ref]. The mean survival time in Laranga's case series (cohort 1) was 74 months (95% CI, 12 not reached) for patients treated with ChT and surgery, whereas with surgery alone, it was only 20 months (95% CI, 8 not reached) [bib_ref] Low-Grade Central Osteosarcoma Arising from Bone Infarct: A Case, Goel [/bib_ref] , and it was longer than in the historical cohort (cohort 2) from the literature (12 months; p = 0.0247). Mortality hazard in Laranga's cohort 1 remained lower even when adjusting for ChT administration and age when compared to the historical cohort 2, though the CI remained wide (HR: 0.333; 95% CI: 0.11-1.03). Secondary metastases occurred in 40% of patients within nine months after the initial diagnosis on average [bib_ref] Bone Infarct-Associated Osteosarcoma: Epidemiologic and Survival Trends, Laranga [/bib_ref]. In another case series by Stacy et al., consisting of osteosarcomas and MFH, patients without metastases had a better prognosis; they lived 11 months to 25 years after the initial diagnosis. On the other hand, patients who develop metastases died (on average) within eight months after their discovery [bib_ref] Infarct-Associated Bone Sarcomas: Multimodality Imaging Findings, Stacy [/bib_ref]. Spazzioli et al. also calculated OS for 12 cases described in the literature. The OS at 12 months and 36 months was 54% (95% CI 26-76) and 36% (95% CI 25-76), respectively. These differences might be explained by the fact that the cases included in the calculation were different than in Laranga et al.'s paper [bib_ref] Bone Infarct-Associated Osteosarcoma: Epidemiologic and Survival Trends, Laranga [/bib_ref].
There is no long-term follow-up information about the single myxofibrosarcoma case reported by Kayser et al. [bib_ref] First report of myxofibrosarcoma of bone arising at a bone infarct, Kayser [/bib_ref]. Endo et al. have described a matter of low-grade osteosarcoma. The latter was followed up for 17 years. The patient underwent surgery 13 years after the initial diagnosis, and four years past surgery, there were no signs of disease progression [bib_ref] Low-grade central osteosarcoma arising from bone infarct, Endo [/bib_ref]. Nonetheless, the authors of that paper admitted that a clinical approach consisting of regular X-rays could cause potential harm, and that low-grade osteosarcomas could be diagnosed earlier with other imaging techniques, such as CT or MRI. Therefore, there was a risk that there was a diagnostic delay in that case, and that a misdiagnosed lesion could have progressed to high-grade spindle cell sarcoma, which occurs in 1 out of 5 cases [bib_ref] Low-grade central osteosarcoma arising from bone infarct, Endo [/bib_ref]. Nonetheless, more data is necessary to determine the prognosis of these tumors arising from a bone infarct.
# Discussion
Ischemia of the bone is a relatively common finding in orthopedic practice. Many conditions and medications were shown to increase the risk of bone infarcts, including conditions that promote the formation of blood clots (e.g., sickle cell anemia, pregnancy, Cushing disease), steroids, alcohol abuse, diving, radiation therapy, and others [bib_ref] Guidelines for clinical diagnosis and treatment of osteonecrosis of the femoral head..., Zhao [/bib_ref] [bib_ref] Nontraumatic Osteonecrosis of the femoral head: Where do we stand today?: A..., Mont [/bib_ref] [bib_ref] Bone infarct-associated osteosarcoma, Torres [/bib_ref]. In fact, in roughly 70% of cases, the cause of bone infarct remains unknown [bib_ref] Infarct-associated bone sarcomas, Domson [/bib_ref] [bib_ref] Low-grade central osteosarcoma arising from bone infarct, Endo [/bib_ref] [bib_ref] Secondary multifocal osteosarcoma developing on the background of bone infarct, Sivrioglu [/bib_ref] [bib_ref] Infarct-Associated Bone Sarcomas: Multimodality Imaging Findings, Stacy [/bib_ref] [bib_ref] Cetintas sk2, muhammed sadik bilgen ms. Secondary osteosarcomas diagnosed in a single..., Yalcinkaya [/bib_ref]. Not all infarcts are symptomatic, and patients remain symptom-free for a long time, which is especially true for diaphyseal and metaphyseal AVN [bib_ref] Bone infarct-associated osteosarcoma, Torres [/bib_ref] [bib_ref] Low-grade central osteosarcoma arising from bone infarct, Endo [/bib_ref]. That results in diagnostic and treatment delays and poorer clinical outcomes [bib_ref] Bone Infarct-Associated Osteosarcoma: Epidemiologic and Survival Trends, Laranga [/bib_ref].
Available epidemiological studies demonstrate that less than 1% of malignant tumors affect the bone. In the USA, only 3300 cases of primary bone tumors are diagnosed each year [bib_ref] Low-grade central osteosarcoma arising from bone infarct, Endo [/bib_ref]. Secondary bone tumors, including infarct-associated tumors, are even rarer, making up to 0.6-1% of all sarcomas of the bone [bib_ref] Bone Infarct-Associated Osteosarcoma: Epidemiologic and Survival Trends, Laranga [/bib_ref] [bib_ref] Infarct-Associated Bone Sarcomas: Multimodality Imaging Findings, Stacy [/bib_ref] [bib_ref] Leiomyosarcoma of bone arising in association with a bone infarct, Petra [/bib_ref]. The first cases of infarctassociated sarcomas were described by . In the literature, less than 150 cases of AVN-related tumors were described, inducing over 120 MFH cases [bib_ref] Chronic knee pain in an 80-year-old woman, Robbin [/bib_ref] [bib_ref] Infarct-Associated Bone Sarcomas: Multimodality Imaging Findings, Stacy [/bib_ref]. Most patients affected were male and in their fifth or sixth decade of life [bib_ref] Bone infarct transformation into undifferentiated pleomorphic sarcoma in sickle cell disease: A..., Alhamdan [/bib_ref] [bib_ref] A Rare Case of an Osteolytic Bone-infarct-associated Osteosarcoma: Case Report with Radiographic..., Mcdonald [/bib_ref] [bib_ref] Chronic knee pain in an 80-year-old woman, Robbin [/bib_ref] [bib_ref] Infarct-Associated Bone Sarcomas: Multimodality Imaging Findings, Stacy [/bib_ref]. A majority of these tumors were located in the lower limbs, especially the tibia and femur, though two case reports concerned humoral lesions [bib_ref] Bone infarct-associated osteosarcoma, Torres [/bib_ref] [bib_ref] Infarct-associated bone sarcomas, Domson [/bib_ref] [bib_ref] First report of myxofibrosarcoma of bone arising at a bone infarct, Kayser [/bib_ref] [bib_ref] A Rare Case of an Osteolytic Bone-infarct-associated Osteosarcoma: Case Report with Radiographic..., Mcdonald [/bib_ref]. Many cases came from small case series or single-patient case descriptions [bib_ref] Bone infarct-associated osteosarcoma, Torres [/bib_ref] [bib_ref] Infarct-associated bone sarcomas, Domson [/bib_ref] [bib_ref] A typology of reviews: An analysis of 14 review types and associated..., Grant [/bib_ref] [bib_ref] Bone infarct transformation into undifferentiated pleomorphic sarcoma in sickle cell disease: A..., Alhamdan [/bib_ref] [bib_ref] Low-grade central osteosarcoma arising from bone infarct, Endo [/bib_ref] [bib_ref] Low-Grade Central Osteosarcoma Arising from Bone Infarct: A Case, Goel [/bib_ref] [bib_ref] First report of myxofibrosarcoma of bone arising at a bone infarct, Kayser [/bib_ref] [bib_ref] Bone Infarct-Associated Osteosarcoma: Epidemiologic and Survival Trends, Laranga [/bib_ref] [bib_ref] Chondroblastic Osteosarcoma Arising in a Bone Infarct in a Patient with a..., Lewin [/bib_ref] [bib_ref] A Rare Case of an Osteolytic Bone-infarct-associated Osteosarcoma: Case Report with Radiographic..., Mcdonald [/bib_ref] [bib_ref] Chronic knee pain in an 80-year-old woman, Robbin [/bib_ref] [bib_ref] Secondary multifocal osteosarcoma developing on the background of bone infarct, Sivrioglu [/bib_ref] [bib_ref] Infarct-Associated Bone Sarcomas: Multimodality Imaging Findings, Stacy [/bib_ref] [bib_ref] Cetintas sk2, muhammed sadik bilgen ms. Secondary osteosarcomas diagnosed in a single..., Yalcinkaya [/bib_ref]. It should be underlined that silent infarcts, which are more likely to be a background for carcinogenesis and tumor development, do not present with noticeable symptoms [bib_ref] Low-grade central osteosarcoma arising from bone infarct, Endo [/bib_ref]. It is possible that a growing tumor would obscure radiological signs of the past infarct. Therefore, the true epidemiology of AVN-associated tumors is likely to be underestimated [bib_ref] Infarct-Associated Bone Sarcomas: Multimodality Imaging Findings, Stacy [/bib_ref]. On the contrary, epiphyseal bone infarcts are more likely to be discovered early, since they are usually symptomatic. Tumorigenesis, being a lengthy process, is unlikely to occur in such cases, and as a result, symptomatic infarcts are less likely to be a background for secondary malignancies [bib_ref] Infarct-associated bone sarcomas, Domson [/bib_ref] [bib_ref] A Rare Case of an Osteolytic Bone-infarct-associated Osteosarcoma: Case Report with Radiographic..., Mcdonald [/bib_ref] [bib_ref] Infarct-Associated Bone Sarcomas: Multimodality Imaging Findings, Stacy [/bib_ref] [bib_ref] Prognostic factors in high-grade osteosarcoma of the extremities or trunk: An analysis..., Bielack [/bib_ref].
The pathophysiology of the malignant transformation of necrotic bone tissue remains a field of hypotheses and assumptions that still need to be investigated in depth. It has been hypothesized that the reparative process and chronic inflammation lead to the transformation into sarcoma [bib_ref] Bone infarct transformation into undifferentiated pleomorphic sarcoma in sickle cell disease: A..., Alhamdan [/bib_ref] [bib_ref] Low-grade central osteosarcoma arising from bone infarct, Endo [/bib_ref] [bib_ref] Cetintas sk2, muhammed sadik bilgen ms. Secondary osteosarcomas diagnosed in a single..., Yalcinkaya [/bib_ref]. In general, the development of sarcoma in the AVN of the bones is a relatively slow process. One study of Caisson workers demonstrated that sarcoma developed 17-22 years after quitting their job. A similar timeframe applied to a case described by Endo et al., wherein the malignancy was detected 13 years after the infarct was visualized on an X-ray performed for to another reason. The close follow-up of that patient led to the detection of malignancy relatively early, before it transformed into high-grade osteosarcoma [bib_ref] Low-grade central osteosarcoma arising from bone infarct, Endo [/bib_ref].
Nonetheless, the exact time interval between bone infarct and the development of bone sarcoma is unknown, since AVN timing is usually impossible to determine. In all of the cases described by Stacy et al., all patients were diagnosed when the symptoms related to the malignancy itself became apparent, and only subtle signs of previous AVN indicated that it was secondary to infarction. The situation was the same in all but one case [bib_ref] Infarct-Associated Bone Sarcomas: Multimodality Imaging Findings, Stacy [/bib_ref]. In addition, none of the studies included in our review have proven that there is a connection between any condition, medication, or other risk factors promoting oncogenesis in patients who suffered a bone infarct [bib_ref] Bone infarct transformation into undifferentiated pleomorphic sarcoma in sickle cell disease: A..., Alhamdan [/bib_ref] [bib_ref] Low-grade central osteosarcoma arising from bone infarct, Endo [/bib_ref] [bib_ref] Low-Grade Central Osteosarcoma Arising from Bone Infarct: A Case, Goel [/bib_ref] [bib_ref] First report of myxofibrosarcoma of bone arising at a bone infarct, Kayser [/bib_ref] [bib_ref] A Rare Case of an Osteolytic Bone-infarct-associated Osteosarcoma: Case Report with Radiographic..., Mcdonald [/bib_ref] [bib_ref] Secondary multifocal osteosarcoma developing on the background of bone infarct, Sivrioglu [/bib_ref]. It is yet to be established if such tumors develop by chance or if some populations need closer follow-up to allow for early detection and treatment, which could improve the long-term prognosis. Alhamdan et al. have made a suggestion that SCT patients are more likely to develop secondary malignancies in infarcted areas. Still, there is no proven connection between these two, but this concept warrants further investigation [bib_ref] Bone infarct transformation into undifferentiated pleomorphic sarcoma in sickle cell disease: A..., Alhamdan [/bib_ref].
Imaging is key to diagnosis and determining signs of a previous bone infarct [bib_ref] Low-grade central osteosarcoma arising from bone infarct, Endo [/bib_ref] [bib_ref] Infarct-Associated Bone Sarcomas: Multimodality Imaging Findings, Stacy [/bib_ref]. A radiologist assessing the tumor might see a well-defined, calcified band representing the previous infarct's margins. A mixed pattern of lucencies and sclerotic changes found in the tumor's proximity might also indicate that the disease is secondary to osteonecrosis [bib_ref] Infarct-Associated Bone Sarcomas: Multimodality Imaging Findings, Stacy [/bib_ref]. Such changes are observed both in classic radiographs, MRIs, and CT [bib_ref] Infarct-associated bone sarcomas, Domson [/bib_ref] [bib_ref] A typology of reviews: An analysis of 14 review types and associated..., Grant [/bib_ref] [bib_ref] Bone infarct transformation into undifferentiated pleomorphic sarcoma in sickle cell disease: A..., Alhamdan [/bib_ref] [bib_ref] Low-grade central osteosarcoma arising from bone infarct, Endo [/bib_ref] [bib_ref] Low-Grade Central Osteosarcoma Arising from Bone Infarct: A Case, Goel [/bib_ref] [bib_ref] First report of myxofibrosarcoma of bone arising at a bone infarct, Kayser [/bib_ref] [bib_ref] Bone Infarct-Associated Osteosarcoma: Epidemiologic and Survival Trends, Laranga [/bib_ref] [bib_ref] Chondroblastic Osteosarcoma Arising in a Bone Infarct in a Patient with a..., Lewin [/bib_ref] [bib_ref] A Rare Case of an Osteolytic Bone-infarct-associated Osteosarcoma: Case Report with Radiographic..., Mcdonald [/bib_ref] [bib_ref] Chronic knee pain in an 80-year-old woman, Robbin [/bib_ref] [bib_ref] Secondary multifocal osteosarcoma developing on the background of bone infarct, Sivrioglu [/bib_ref] [bib_ref] Infarct-Associated Bone Sarcomas: Multimodality Imaging Findings, Stacy [/bib_ref] [bib_ref] Cetintas sk2, muhammed sadik bilgen ms. Secondary osteosarcomas diagnosed in a single..., Yalcinkaya [/bib_ref]. Nonetheless, an X-ray is usually insufficient to assess the tumor properly, and MRI and CT are indispensable. For staging and the determination of the presence of metastases or multifocal lesions, scintigraphy and SPECT might be considered [bib_ref] A Rare Case of an Osteolytic Bone-infarct-associated Osteosarcoma: Case Report with Radiographic..., Mcdonald [/bib_ref] [bib_ref] Infarct-Associated Bone Sarcomas: Multimodality Imaging Findings, Stacy [/bib_ref].
More recently, molecular markers and genomic analyses have helped determine the histological characteristics of infarct-associated tumors [bib_ref] Low-grade central osteosarcoma arising from bone infarct, Endo [/bib_ref] [bib_ref] First report of myxofibrosarcoma of bone arising at a bone infarct, Kayser [/bib_ref] [bib_ref] A Rare Case of an Osteolytic Bone-infarct-associated Osteosarcoma: Case Report with Radiographic..., Mcdonald [/bib_ref]. Immunohistochemical stains are especially useful in rare histological tumors. Myxofibrosarcomas express CD34 protein, whereas low-grade osteosarcomas express CDK4, MDM2, and mutations of the p53 protein. INI1 was retained in osteosarcoma [bib_ref] Low-grade central osteosarcoma arising from bone infarct, Endo [/bib_ref] [bib_ref] First report of myxofibrosarcoma of bone arising at a bone infarct, Kayser [/bib_ref] [bib_ref] A Rare Case of an Osteolytic Bone-infarct-associated Osteosarcoma: Case Report with Radiographic..., Mcdonald [/bib_ref]. Most of the included studies, however, did not share results of immunohistological or molecular examination.
Secondary osteosarcomas typically develop in individuals over 50 years of age, which translates to challenges in the management of the disease [bib_ref] Bone Infarct-Associated Osteosarcoma: Epidemiologic and Survival Trends, Laranga [/bib_ref]. The overall prognosis is unfavorable in patients with sarcomas arising at the bone infarct site. Past reports have suggested that aggressive and multimodal treatment might result in 2-year survival rates comparable to primary osteosarcomas (60-70% vs. 50-80%, respectively) [bib_ref] Infarct-associated bone sarcomas, Domson [/bib_ref] [bib_ref] Infarct-Associated Bone Sarcomas: Multimodality Imaging Findings, Stacy [/bib_ref] [bib_ref] Cetintas sk2, muhammed sadik bilgen ms. Secondary osteosarcomas diagnosed in a single..., Yalcinkaya [/bib_ref]. Unfortunately, not all patients can complete a high-dose course of ChT, due to their comorbidities, and some instead receive a shorter, low-dose regimen or none [bib_ref] Prognosis of radiation-induced bone sarcoma is similar to primary osteosarcoma, Shaheen [/bib_ref] [bib_ref] Bone Infarct-Associated Osteosarcoma: Epidemiologic and Survival Trends, Laranga [/bib_ref]. Some patients might not respond to such management either [bib_ref] A Rare Case of an Osteolytic Bone-infarct-associated Osteosarcoma: Case Report with Radiographic..., Mcdonald [/bib_ref]. Therefore, overall-survival rates are much lower. In a review from 1992 by Torres and Kyriakos, only 22% of patients were alive five years after MFH diagnosis [bib_ref] Bone infarct-associated osteosarcoma, Torres [/bib_ref]. Domos et al. reported that 46% of their patients died within seven months of the diagnosis, and taking together their data and other cases published until 2004, 57% of patients were dead within 19.2 months [bib_ref] Infarct-associated bone sarcomas, Domson [/bib_ref].
Conversely, in the recently published case series by Stacy et al., patients who had localized disease survived 11 months to 25 years, and, in a review prepared by Laranga et al., five-year survival reached 62% (CI 28-84), and the median survival was 74 months, compared to 12 months in the cases described previously in the literature. The results were worse for patients treated with surgery only (50% lived for five years post-diagnosis and, for the case reports described previously, 50% lived for two years) [bib_ref] Bone Infarct-Associated Osteosarcoma: Epidemiologic and Survival Trends, Laranga [/bib_ref] [bib_ref] Infarct-Associated Bone Sarcomas: Multimodality Imaging Findings, Stacy [/bib_ref]. These data suggest that surgery combined with adjuvant treatment and even surgery alone offer better results than in the past decades. However, given the small case numbers, definite conclusions cannot be drawn [bib_ref] Bone infarct-associated osteosarcoma, Torres [/bib_ref] [bib_ref] Infarct-associated bone sarcomas, Domson [/bib_ref] [bib_ref] Bone Infarct-Associated Osteosarcoma: Epidemiologic and Survival Trends, Laranga [/bib_ref] [bib_ref] Infarct-Associated Bone Sarcomas: Multimodality Imaging Findings, Stacy [/bib_ref]. To date, there is no standardization of treatment for this type of malignancy. This might be important, especially for patients with metastases, who still have the worst prognosis, with reported survival rates at 1-year post-discovery close to 0% [bib_ref] Bone infarct-associated osteosarcoma, Torres [/bib_ref] [bib_ref] Infarct-associated bone sarcomas, Domson [/bib_ref] [bib_ref] Infarct-Associated Bone Sarcomas: Multimodality Imaging Findings, Stacy [/bib_ref].
# Strengths and limitations
Though this review mapped out recent literature regarding bone-infarct-associated malignancies, several limitations apply to our study. Mapping reviews provide a surfacelevel synthesis of available data and do not involve extensive and critical analysis of the main topic. This was not a full systematic review, though we applied a similar methodology in our search. We did not conduct a meta-analysis, but we were aware that this was not possible given the type of studies we expected to find. The number of analyzed cases was small, but this was also true for previously published reviews describing secondary sarcomas [bib_ref] Infarct-associated bone sarcomas, Domson [/bib_ref] [bib_ref] Low-Grade Central Osteosarcoma Arising from Bone Infarct: A Case, Goel [/bib_ref] [bib_ref] Bone Infarct-Associated Osteosarcoma: Epidemiologic and Survival Trends, Laranga [/bib_ref]. We have also limited the timing of publication to the last ten years because we aimed to identify any new epidemiological data and treatment outcomes changes in comparison with previous works. As a result of that restriction, fewer papers were included. On the other hand, this was the first review to focus on radiological findings and their description in a synthetic manner, as well as to discuss the hypothesized pathomechanisms of tumor development in a necrotic bone. There is a possibility that some of the papers might not have been identified, as we limited our search to one database; however, we do not think that our conclusions would change much, given the exploratory nature of this review and the fact that most papers were case reports or retrospective case series.
# Conclusions
Sarcomas arising at the site of AVN are an infrequent clinical entity, with less than 200 cases described worldwide. To date, there have been no prospective studies to collect data allowing for standardization in disease management. MFH accounts for around 60% of infarct-associated tumors, though other types of osteosarcomas, including spindle-cell osteosarcoma and myxofibrosarcoma, were described. Usually, determining the presence of a previous infarct can be observed on a routine X-ray. Nonetheless, MRI/CT is essential for the assessment of the tumor. The identification of certain histological types may warrant immunostaining. To date, there is no standardization of treatment, but recent data might suggest that the overall survival improves with adjuvant ChT in localized diseases. However, overall survival is relatively low if the patient is disqualified from aggressive treatment. To date, most patients with disseminated disease did not live more than a year after the detection of metastases. More research is needed to answer questions about the pathomechanisms of the malignant transformation of bone infarcts and identify the possible risk factors that increase the chances of secondary malignancy development. That could lead to decreased mortality related to such tumors, due to early detection at a localized stage. Institutional Review Board Statement: Not applicable.
## Informed consent statement: not applicable.
Data Availability Statement: Additional data are available from the corresponding author upon reasonable request.
## Conflicts of interest:
The authors declare no conflict of interest.
[fig] Figure 1: Review of the literature flow diagram of PubMed publications. [/fig]
[fig] A 59 -: year-old patient with bilateral infarcts of the femur and tibia and osteolytic lesion in the right distal femur. He presented with pain in the right knee. Patient had no systemic disease or other risk factors for bone infarct. The lesion was treated primarily with bone grafting and curettage. After confirmation of the malignant nature of the lesion, he was referred to the author's institution. He underwent ChT and RxT. HE died 4 years after surgery due to lung metastasis.Osteosarcoma Distal femur AVN-avascular necrosis; ChT-chemotherapy; CT-computed tomography, GTC-giant cell tumor MFH-Malignant fibrous histiocytoma; MRI-magnetic resonance; OS-overall survival; ROM-range of motion, RxT-radiotherapy. [/fig]
[fig] Supplementary: Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/ijerph19159282/s1. Author Contributions: Conceptualization, W.K., T.P., A.K., A.Ś., I.K. and J.K.; methodology, W.K.; formal analysis, W.K., T.P., A.K., A.Ś., I.K., M.H. and J.K.; investigation, W.K.; data curation, T.P., W.K. and A.K.; writing-original draft preparation, W.K.; writing-review and editing, W.K., T.P. and A.K.; visualization, W.K.; supervision, W.K., T.P., A.K., A.Ś., I.K., M.H. and J.K. All authors have read and agreed to the published version of the manuscript. Funding: Publication was funded by the Medical University of Lodz, Department of Social Medicine (project No. 503/6-029-01/503-61-001-19-00). [/fig]
[table] Table 4: Inclusion and exclusion criteria. [/table]
[table] Table 5: Overview of studies. [/table]
|
Advancements in stem cells treatment of skeletal muscle wasting
† These authors have contributed equally to this work.Muscular dystrophies (MDs) are a heterogeneous group of inherited disorders, in which progressive muscle wasting and weakness is often associated with exhaustion of muscle regeneration potential. Although physiological properties of skeletal muscle tissue are now well known, no treatments are effective for these diseases. Muscle regeneration was attempted by means transplantation of myogenic cells (from myoblast to embryonic stem cells) and also by interfering with the malignant processes that originate in pathological tissues, such as uncontrolled fibrosis and inflammation. Taking into account the advances in the isolation of new subpopulation of stem cells and in the creation of artificial stem cell niches, we discuss how these emerging technologies offer great promises for therapeutic approaches to muscle diseases and muscle wasting associated with aging.
# Introduction
Skeletal muscle is a highly complex system formed by thousands of contractile units called muscle fibers. Each muscle fiber is limited by a plasma membrane called sarcolemma and by a basal lamina, that are surrounded by an extra cellular matrix constituted of connective tissue [bib_ref] The formation of skeletal muscle: from somite to limb, Buckingham [/bib_ref]. Muscle remodeling occurs throughout the entire life although at different rate considering the developmental stages. Starting from embryo until childhood, protein synthesis is upregulated and satellite cells (SCs) actively develop new muscle fibers while in adults cellular turnover is strongly reduced [bib_ref] Activity-dependent signaling pathways controlling muscle diversity and plasticity, Schiaffino [/bib_ref]. In response to exogenous stimuli or to biological factors such as age or nutrition, the muscle increases its size, the amount of contractile proteins and consequently force production. The regulation of muscle cell size is a tightly regulated phenomena, and it is a balance between muscle proliferation and degradation of preexisting proteins. Uncontrolled events, often associated with diseases, lead to hypertrophy or atrophy, respectively [bib_ref] Signaling in muscle atrophy and hypertrophy, Sandri [/bib_ref]. The complex hierarchy of events that triggers muscle remodeling is often unbalanced in muscular diseases. For instance, Duchenne muscular dystrophy (DMD), the most frequent among all the dystrophies, is characterized by a rapid atrophy in youth, muscular wasting and inability to walk in adolescence and premature death for cardiorespiratory failure by the age of 30. As the genetic nature of these pathologies leads to uncontrolled fiber degeneration, different treatments were proven to delay the progression of the diseases. The main goal was to retard the atrophy and replace diseased muscle with new healthy and functional muscle fibers by using myogenic stem cells [bib_ref] The immune system and the repair of skeletal muscle, Brunelli [/bib_ref]. Unfortunately, the use of stem cell in regenerative medicine is limited by the poor engraftment and persistence of transplanted cells and the risk of neoplastic formation [bib_ref] Improving cell engraftment with tissue engineering, Suuronen [/bib_ref] [bib_ref] Exploiting extracellular matrix-stem cell interactions: a review of natural materials for therapeutic..., Kuraitis [/bib_ref]. Due to these findings, other aspects were deeply investigated to increase the survival of injected stem cells into pathological muscle. Modulation of the inflammatory reaction is a key step for stem cell transplantation [bib_ref] Enhanced migration and fusion of donor myoblasts in dystrophic and normal host..., Smythe [/bib_ref] : myeloid cells [bib_ref] Endurance exercise causes interaction among stress hormones, cytokines, neutrophil dynamics, and muscle..., Suzuki [/bib_ref] [bib_ref] Splicing mutation in dysferlin produces limb-girdle muscular dystrophy with inflammation, Mcnally [/bib_ref] , macrophages [bib_ref] A nitric oxide synthase transgene ameliorates muscular dystrophy in mdx mice, Wehling [/bib_ref] [bib_ref] Shifts in macrophage phenotypes and macrophage competition for arginine metabolism affect the..., Villalta [/bib_ref] , neutrophils [bib_ref] Reduced necrosis of dystrophic muscle by depletion of host neutrophils, or blocking..., Hodgetts [/bib_ref] and eosinophils [bib_ref] Eosinophilia of dystrophin-deficient muscle is promoted by perforinmediated cytotoxicity by T cell..., Cai [/bib_ref] actively contribute to development of pathogenesis in several myophaties but only macrophages and sometimes eosinophils play a role in muscle regeneration [bib_ref] Regulatory interactions between muscle and the immune system during muscle regeneration, Tidball [/bib_ref]. Pathological conditions modify the microenvironment of stem cells (the so-called niche) preventing the activation of resident stem cells and reducing the success of exogenous cell therapies. Tissue engineering technologies may create a novel in vitro niche allowing the maintenance and propagation of SCs and enhancing their muscular potential.
In this review, we will describe the efforts that are necessary to design a successful therapeutic approach for muscular diseases, relating to find a functional stem cell population, to identify feasible matrix/polymer to engineer stem cells' niche and to modulate secondary-but relevant-effects of impaired muscle regeneration, as fibrosis and inflammation.
## Myogenic stem cells embryonic stem cells (escs)
# Introduction to escs
Embryonic stem cells (ESCs) are pluripotent cells derived from the early embryo that are characterized by the ability to proliferate over prolonged periods of culture remaining undifferentiated and maintaining a stable karyotype [bib_ref] Derivation and spontaneous differentiation of human embryonic stem cells, Amit [/bib_ref] [bib_ref] Characterization and differentiation of human embryonic stem cells, Carpenter [/bib_ref] [bib_ref] Characterization and culture of human embryonic stem cells, Hoffman [/bib_ref]. ESCs differentiate into cells forming all 3 embryonic germ layers, and are characterized by self-renewal, immortality, and pluripotency [bib_ref] Human embryonic stem cells and gene therapy, Strulovici [/bib_ref]. As ESCs possess the potential to differentiate into all normal tissues, the ability to derive and maintain these cells in culture opened the possibility to have an unlimited supply of differentiated cells to replace pathological tissues [bib_ref] Generation, culture, and differentiation of human embryonic stem cells for therapeutic applications, Moon [/bib_ref] [bib_ref] The derivation of clinicalgrade human embryonic stem cell lines, Skottman [/bib_ref].
## Markers of escs
Cell origins are often defined by one or more cell-surface markers and intracellular epitopes unique to that particular cell type. hESCs are maintained in culture on feeder layers of heterologous cells and then differentiated into specific cell lineages [bib_ref] Induction of pluripotent stem cells from mouse embryonic and adult fibroblast cultures..., Takahashi [/bib_ref] [bib_ref] Generation of pluripotent stem cells from adult human testis, Conrad [/bib_ref]. Stage-specific embryonic antigen citation(SSEA) markers are used to distinguish early stages of cell development and to denote pluripotency: hESCs express SSEA-3 and -4 during pluripotency and only SSEA-1 upon differentiation [bib_ref] Comparative analysis of cell surface antigens expressed by cell lines derived from..., Andrews [/bib_ref] [bib_ref] Neural progenitors from human embryonic stem cells, Reubinoff [/bib_ref]. Nanog is a NK-2-type homeodomain gene encoding for a transcription factor that is critically involved in the self-renewal of stem cells. In 2005, Lin's group demonstrated that the tumor suppressor p53 binds to the promoter of Nanog, stimulating p53 [bib_ref] p53 induces differentiation of mouse embryonic stem cells by suppressing Nanog expression, Lin [/bib_ref]. Octamer-binding transcription factor 4 (Oct-4) down-regulation is observed in differentiating cells [bib_ref] A POU-domain transcription factor in early stem cells and germ cells of..., Rosner [/bib_ref]. It was suggested that only Oct-4 was necessary for the maintenance of pluripotency, but its expression level governed three cell fates once differentiation occurs. Similarly, published that the catalytic component of telomerase, telomerase reverse transcriptase (hTERT), was expressed in undifferentiated cells and down-regulated upon differentiation [bib_ref] Feeder-free growth of undifferentiated human embryonic stem cells, Xu [/bib_ref].
## Limits of escs
Although the attentions that received, scientific and medical issues need to be addressed before hESCs can be considered safe for clinical applications [bib_ref] The biological and ethical basis of the use of human embryonic stem..., Leist [/bib_ref]. The American federal government severely restricted access and use of hESCs in 2001 but they were largely overturned by the Obama administration. Many organizations and countries have already banned reproductive cloning of human beings. As this procedure can be used to generate stem cells for therapeutic purposes, in countries where this type of cloning is legal, such as Australia and the United Kingdom, the created embryos must be destroyed within 14 days. Guidelines in using ESCs were proposed by the International Society of Stem Cell Research citation (http://www. isscr.org/guidelines/index.htm).
## Myogenic potential of escs
Several lineages (blood, cardiac muscle and endothelial cells) were obtained by in vitro differentiation of ESCs, however for skeletal muscle several drawbacks arose, especially for the difficulty to identify a temporal expression of myogenic regulatory factors [bib_ref] Muscle cell differentiation of embryonic stem cells reflects myogenesis in vivo: developmentally..., Rohwedel [/bib_ref]. This way, in 2005 Bhagavati et al. cocultured ESCs derived from normal mice with a preparation from mouse muscle enriched for myogenic stem and precursor cells. They transplanted ESCs into dystrophic mdx mice but unfortunately newly-formed muscle was occasionally seen [bib_ref] Generation of skeletal muscle from transplanted embryonic stem cells in dystrophic mice, Bhagavati [/bib_ref]. Similarly, Barberi et al. described a stroma-free induction system to derive mesenchymal precursors and skeletal myoblast from hESCs. Following in vitro maturation, these cells were injected into tibialis anterior of immunodeficient scid mice and it was observed a long-term myoblast engraftment and the lack of teratomas [bib_ref] Derivation of engraftable skeletal myoblasts from human embryonic stem cells, Barberi [/bib_ref]. As it was suggested that the lack of myogenic differentiation of ESCs was due to the impairment of myogenic signals in the mesoderm [bib_ref] Functional skeletal muscle regeneration from differentiating embryonic stem cells, Darabi [/bib_ref] , and paired box 7 (Pax7) during early mesoderm development and obtained several early embryonic skeletal myogenic progenitors [bib_ref] The therapeutic potential of embryonic and adult stem cells for skeletal muscle..., Darabi [/bib_ref] [bib_ref] Assessment of the myogenic stem cell compartment following transplantation of Pax3/Pax7-induced embryonic..., Darabi [/bib_ref]. These cells were also implanted into mdx mice and gave rise to large numbers of skeletal muscle fibers and SCs, so that muscle force was ameliorated [bib_ref] Engraftment of embryonic stem cell-derived myogenic progenitors in a dominant model of..., Darabi [/bib_ref] [bib_ref] Assessment of the myogenic stem cell compartment following transplantation of Pax3/Pax7-induced embryonic..., Darabi [/bib_ref]. More recently, Sakurai et al. described that the elimination of bone morphogenetic protein 4 (BMP4) from serum-free ESC cultures together with the implementation of lithium chloride (LiCl) allowed the differentiation of these cells to myogenic progenitors cells. hESCs-derived progenitors showed a notable capacity of differentiation into skeletal muscle cells [bib_ref] Bidirectional induction toward paraxial mesodermal derivatives from mouse ES cells in chemically..., Sakurai [/bib_ref].
## Induced pluripotent stem cells (ipscs)
## Generation of ipscs
Recent advances in the understanding of ESC biology included the identification of several master regulators of ESC pluripotency and differentiation [bib_ref] Induction of pluripotent stem cells from mouse embryonic and adult fibroblast cultures..., Takahashi [/bib_ref]. Intensive study of ESC growth conditions has not yet produced a complete picture of the unique transcriptional and epigenetic state that is responsible for pluripotency and self-renewal in ESCs. Yamanaka's group identified four factors (Oct3/4, Klf4, Sox2, and c-myc) whose expression is sufficient to produce cells similar to ESCs, called induced pluripotent stem cells (iPSCs). The same factors were used to reprogram human fibroblasts to an ESC-like pluripotent state.
## The new era of ipscs
Now that embryonic tissue is no longer required to make a pluripotent cell, investigators have the ability to create tissuebased models of human disease based on cells derived from individual patients [bib_ref] Induced pluripotent stem cells generated from patients with ALS can be differentiated..., Dimos [/bib_ref] [bib_ref] Reprogramming of human somatic cells to pluripotency with defined factors, Park [/bib_ref] [bib_ref] Induced pluripotent stem cells from a spinal muscular atrophy patient, Ebert [/bib_ref] [bib_ref] Parkinson's disease patient-derived induced pluripotent stem cells free of viral reprogramming factors, Soldner [/bib_ref]. Accordingly, iPSCs were efficiently used in murine models of sickle cells anemia and Parkinson's disease. Even if these cells were showed to be suitable for cell therapy, it has to be yet demonstrated the possibility to generate human iPSCs without introduction of DNA into the genome (to avoid oncogenic potential of undifferentiated iPSCs following the unsafe reintroduction of these genes), to ameliorate the efficiency of manipulation of human iPSCs and the capacity to obtain any desired cell types.
## Ipscs and human disease
Since the work of Yamanaka was published, reprogramming of cells provided a realistic way not only to obtain lines from patients with incurable pathologies to investigate disease mechanisms and drug screening but to generate sufficient numbers of patientspecific pluripotent stem cells [bib_ref] Drug screening for ALS using patient-specific induced pluripotent stem cells, Egawa [/bib_ref]. The generation of patient-specific iPSCs has the advantage of avoiding many of the ethical concerns associated with the use of embryonic or foetal material, and have no risk of immune rejection. Many cell types like motor neurons [bib_ref] Induced pluripotent stem cells generated from patients with ALS can be differentiated..., Dimos [/bib_ref] , hepatocytes [bib_ref] Efficient generation of hepatocyte-like cells from human induced pluripotent stem cells, Song [/bib_ref] , pancreatic insulin producing cells , hematopoietic cells [bib_ref] Treatment of sickle cell anemia mouse model with iPS cells generated from..., Hanna [/bib_ref] , retinal cells [bib_ref] Protective effects of human iPS-derived retinal pigment epithelium cell transplantation in the..., Carr [/bib_ref] , cardiomyocyte [bib_ref] Cardiomyocyte differentiation of human induced pluripotent stem cells, Zwi [/bib_ref] and mesenchymal stem cells [bib_ref] Functional mesenchymal stem cells derived from human induced pluripotent stem cells attenuate..., Lian [/bib_ref] , have been successfully derived from human iPSCs. Nelson et al. reported the use of iPSCs for myocardial repair in animal models of acute myocardial infarction [bib_ref] Repair of acute myocardial infarction by human stemness factors induced pluripotent stem..., Nelson [/bib_ref] while Ye used iPSCs in different hematological disorders [bib_ref] Induced pluripotent stem cells offer new approach to therapy in thalassemia and..., Ye [/bib_ref].
## Myogenic potential of ipscs
As described above, to be used for clinical applications, iPSCs need to be generated in large amount in safety; this way, protocols to isolate and characterize these cells were largely improved. Mizuno et al. identified iPS-derived satellite-like cells by means the expression of the SM/C-2.6 antibody [bib_ref] Generation of skeletal muscle stem/progenitor cells from murine induced pluripotent stem cells, Mizuno [/bib_ref] while Darabi purified PDGFαR+Flk−1− murine iPS cells that expressed the myogenic factor Pax7 [bib_ref] Functional skeletal muscle regeneration from differentiating embryonic stem cells, Darabi [/bib_ref] [bib_ref] Assessment of the myogenic stem cell compartment following transplantation of Pax3/Pax7-induced embryonic..., Darabi [/bib_ref]. In fact, the group of Perlingeiro recently isolated large quantity of Pax7+ human iPSCs (and ESCs) that, transplanted into dystrophic mice, engrafted well producing high amount of dystrophin and replenishing the satellite cell compartment [bib_ref] Human ES-and iPS-derived myogenic progenitors restore DYSTROPHIN and improve contractility upon transplantation..., Darabi [/bib_ref]. Similarly, the expression of MyoD and Myf5 allowed the purification of myogenic iPS cells [bib_ref] Inducible cassette exchange: a rapid and efficient system enabling conditional gene expression..., Iacovino [/bib_ref]. Filareto et al. successfully obtained iPSCs from fibroblast of dystrophin/utrophin double knockout mice and engineered them with the micro-dystrophin gene. Injected into dystrophic mice, these cells engrafted well and improved muscle strength [bib_ref] An ex vivo gene therapy approach to treat muscular dystrophy using inducible..., Filareto [/bib_ref]. In parallel, Tedesco et al. generated mesoangioblast/mesenchymal-like cells from iPSCs of healthy and dystrophic patients: these cells were also modified to express constitutively the MyoD gene. Transplanted into model mice of LGMD-2A, iPSCs cells ameliorated their dystrophic phenotype [bib_ref] Transplantation of genetically corrected human iPSC-derived progenitors in mice with limb-girdle muscular..., Tedesco [/bib_ref].
## Satellite cells (scs)
SCs are small progenitor cells originating from somites that lie between the basement membrane and sarcolemma of individual muscle fibers [bib_ref] Muscle stem cells in development, regeneration, and disease, Shi [/bib_ref] [bib_ref] Skeletal muscle stem cell birth and properties, Sambasivan [/bib_ref]. SCs are normally present in healthy adult mammalian muscle as quiescent cells and are characterized by the expression of Pax7, that is fundamental for their maintenance and self-renewal, and by the absence of Myogenic differentiation 1 (MyoD) and myogenin, that conversely are necessary for myogenic differentiation. Once activated in response to specific stimuli such as oxidative stress, SCs up-regulate the expression of Myf5 to start their proliferation so that they differentiate into new myofibers, driven by specific factors such as MyoD, myogenin and myosin heavy chain [bib_ref] Expression of myosin isoforms during notexin-induced regeneration of rat soleus muscles, Whalen [/bib_ref]. Since the work of Montarras and colleagues [bib_ref] Direct isolation of satellite cells for skeletal muscle regeneration, Montarras [/bib_ref] , different techniques for SCs isolation were assessed. Sacco et al.
derived SCs from transplantation of one intact myofiber and demonstrated that once transplanted into dystrophic mice, SCs proliferated and contributed to form new muscle fibers [bib_ref] Selfrenewal and expansion of single transplanted muscle stem cells, Sacco [/bib_ref]. Cerletti et al. isolated the skeletal muscle precursors (SMPs): injected into animal models, these SC-like cells restored dystrophin expression and, more importantly, were positioned into the SC niche, where they regulated the subsequent rounds of injury and repair [bib_ref] Highly efficient, functional engraftment of skeletal muscle stem cells in dystrophic muscles, Cerletti [/bib_ref]. Similarly, the muscle sidepopulation cells (mSP) isolated by Tanaka et al. engrafted into host SC niche, giving rise both to SCs and myonuclear population [bib_ref] Syndecan-4-expressing muscle progenitor cells in the SP engraft as satellite cells during..., Tanaka [/bib_ref]. Autologous transplantation of genetically corrected SCs into patients suffering from muscular diseases could be our ideal approach [bib_ref] Stem cell based therapies to treat muscular dystrophy, Price [/bib_ref] : unfortunately, it was demonstrated that the growth of SCs in vitro significantly reduced their in vivo myogenic potential, rendering their transplantation an inefficient technique [bib_ref] Results of a triple blind clinical study of myoblast transplantations without immunosuppressive..., Tremblay [/bib_ref] [bib_ref] Myoblast transfer in the treatment of Duchenne's muscular dystrophy, Mendell [/bib_ref] [bib_ref] The fate of individual myoblasts after transplantation into muscles of DMD patients, Gussoni [/bib_ref]. To overcome these problems, several studies investigated the SC niches, as described in detail in section Satellite cells niche.
## Muscle-derived stem cells (mdscs)
Besides SCs, muscle-derived stem cells (MDSCs) were isolated within the muscle, with the capacity of self-renewal and mesodermal differentiation. Sarig et al. identified a subpopulation of MyoD+ stem cells that formed muscle fibers but also osteogenic and adipogenic cells [bib_ref] Regeneration and transdifferentiation potential of muscle-derived stem cells propagated as myospheres, Sarig [/bib_ref]. Tamaki et al. purified a subpopulation of CD34-CD45-cells that proliferated into myogenic, vasculogenic and neural cell lineages [bib_ref] Clonal multipotency of skeletal muscle-derived stem cells between mesodermal and ectodermal lineage, Tamaki [/bib_ref]. Sca−1+CD34+ stem cells purified from murine muscle differentiated into myogenic and multimyeloid lineages in vitro and regenerated muscle in vivo [bib_ref] Intraarterial injection of muscle-derived CD34(+)Sca-1(+) stem cells restores dystrophin in mdx mice, Torrente [/bib_ref]. showed that muscle-derived stem cells positive for desmin and vimentin differentiated in vitro into skeletal muscle fibers and neurons [bib_ref] Isolation and culture of human muscle-derived stem cells able to differentiate into..., Alessandri [/bib_ref]. Notably, Rouger et al. identified early myogenic progenitors that originated from SC niche, the MuStem cells; transplanted into Golden retriever muscular dystrophy (GRMD) dogs, these cells allowed the re-expression of dystrophin [bib_ref] Systemic delivery of allogenic muscle stem cells induces long-term muscle repair and..., Rouger [/bib_ref].
## Mesenchymal stem cells (mscs)
Mesenchymal stem cells (MSCs) are clonogenic and adherent cells, isolated from adult and foetal bone marrow and from other tissues and organs [bib_ref] Mesenchymal stem cells: isolation and therapeutics, Alhadlaq [/bib_ref] [bib_ref] Mesenchymal stem cells: progress toward promise, Blanc [/bib_ref] [bib_ref] Mesenchymal stem cells: isolation, in vitro expansion and characterization, Beyer Nardi [/bib_ref] : they are able to differentiate into several lineages [bib_ref] Prospective identification of myogenic endothelial cells in human skeletal muscle, Zheng [/bib_ref] [bib_ref] Differentiation potential of multipotent progenitor cells derived from war-traumatized muscle tissue, Nesti [/bib_ref]. As MSCs were identified into muscle tissue biopsies, it was suggested that skeletal muscle could be an important source of MSCs for therapeutic interventions [bib_ref] Potential therapeutic applications of muscle-derived mesenchymal stem and progenitor cells, Jackson [/bib_ref]. Transplanted into DMD patients, MSCs fused with host fibers and enhanced the activity of endogenous stem cells through the secretion of trophic factors [bib_ref] Mesenchymal stem cells as anti-inflammatories: implications for treatment of Duchenne muscular dystrophy, Ichim [/bib_ref]. Interestingly, De Bari et al. described the in vitro myogenic potential of MSCs isolated from adult human synovial membrane [bib_ref] Multipotent mesenchymal stem cells from adult human synovial membrane, De Bari [/bib_ref]. Following injection into dystrophic mice, these cells formed new myofibers, re-expressed the dystrophin and contributed to SCs replenishment [bib_ref] Skeletal muscle repair by adult human mesenchymal stem cells from synovial membrane, De Bari [/bib_ref]. showed that MSCs from umbilical cord blood differentiated into skeletal muscle, expressing late myogenic markers as MyoD [bib_ref] Skeletal myogenic differentiation of mesenchymal stem cells isolated from human umbilical cord..., Gang [/bib_ref]. Riordan et al. described that hematopoietic precursors present in the bone marrow were protected from inflammatory damage by MSCs [bib_ref] Cord blood in regenerative medicine: do we need immune suppression?, Riordan [/bib_ref] while Nemeth et al. demonstrated that MSCs can modulate the activity of macrophages and consequently inhibit inflammatory processes [bib_ref] Bone marrow stromal cells attenuate sepsis via prostaglandin E(2)-dependent reprogramming of host..., Nemeth [/bib_ref]. The capacity of MSCs to modulate inflammation could be an important feature in the perspective of cell therapy in dystrophic patients as inflammation is a prominent component of the disease (as reviewed in detail in section Inflammation and repair mechanisms in skeletal muscle). Following these evidences, MSCs injection were proven to reduce inflammation in animal models for several human diseases, such as autoimmune arthritis and diabetes [bib_ref] Immunomodulatory function of bone marrow-derived mesenchymal stem cells in experimental autoimmune type..., Fiorina [/bib_ref] [bib_ref] Mesenchymal stem cells protect NOD mice from diabetes by inducing regulatory T..., Madec [/bib_ref] , multiple sclerosis [bib_ref] Adipose-derived mesenchymal stem cells ameliorate chronic experimental autoimmune encephalomyelitis, Constantin [/bib_ref] [bib_ref] Allogeneic mesenchymal stem cells for treatment of experimental autoimmune encephalomyelitis, Rafei [/bib_ref] , lupus [bib_ref] Transplantation of human bone marrow mesenchymal stem cell ameliorates the autoimmune pathogenesis..., Zhou [/bib_ref] , rheumatoid arthritis [bib_ref] Lysophosphatidic acid mediates migration of human mesenchymal stem cells stimulated by synovial..., Song [/bib_ref] and autoimmune encephalomyelitis [bib_ref] Mesenchymal stromal cells ameliorate experimental autoimmune encephalomyelitis by inhibiting CD4 Th17 T..., Rafei [/bib_ref]. Although all these encouraging results, several problems need to be solved. First of all, more efforts are needed to elucidate the origin of MSCs; moreover, protocols for isolation of the cells and their expansion in vivo have to be standardized.
## Muscle-derived cd133+ stem cells
Torrente et al. isolated stem cells from human normal and DMD biopsies expressing the glycoprotein CD133. CD133+ stem cells co-expressed CD34, CD45, and kinase insert domain receptor (KDR) and differentiated into muscle [bib_ref] Autologous transplantation of muscle-derived CD133+ stem cells in Duchenne muscle patients, Torrente [/bib_ref]. Moreover, Negroni et al. found that muscle-derived CD133+ stem cells co-expressed the satellite cell marker CD56 and eventually formed myosin heavy chain (MyHC)+ multinucleated myotubes [bib_ref] In vivo myogenic potential of human CD133+ muscle-derived stem cells: a quantitative..., Negroni [/bib_ref]. As Phase I clinic trial demonstrated that infusion of these cells was safe and feasible [bib_ref] Autologous transplantation of muscle-derived CD133+ stem cells in Duchenne muscle patients, Torrente [/bib_ref] , muscle-derived dystrophic CD133+ stem cells were engineered to express a shorter but still functional dystrophin. Transplanted into dystrophic mice, CD133+ stem cells allowed the expression of dystrophin and the formation of new myofibers, improving murine muscular force. Interestingly, some of injected CD133+ stem cells were identified beneath the basal lamina, in SC-like position, thus expressing M-Cadherin [bib_ref] Restoration of human dystrophin following transplantation of exon-skipping-engineered DMD patient stem cells..., Benchaouir [/bib_ref].
## Mesoangioblasts
Physically associated with the embryonic dorsal aorta in avian and mammalian species, mesoangioblasts are multipotent progenitors of mesodermal tissues, expressing α-smooth muscle actin (SMA) and retaining myogenic capacity [bib_ref] TGFbeta/BMP activate the smooth muscle/bone differentiation programs in mesoangioblasts, Tagliafico [/bib_ref]. Cossu et al. engineered these cells with human microdystrophin and demonstrated that they improved muscle function after injection into GRMDs [bib_ref] New therapies for Duchenne muscular dystrophy: challenges, prospects and clinical trials, Cossu [/bib_ref]. In order to ameliorate their ability of migration, mesoangioblasts were exposed to Stromal cell-derived factor (SDF)-1 and tumor necrosis factor (TNF)-α so that, following transplantation into α-sarcoglycan KO mice, the large majority of α-sarcoglycan-expressing myofibers was reconstituted [bib_ref] Complete repair of dystrophic skeletal muscle by mesoangioblasts with enhanced migration ability, Galvez [/bib_ref]. Similarly, Tedesco et al. transduced mdxderived mesoangioblasts with a vector carrying the entire human dystrophin genetic locus. Injected into scid/mdx mice, these cells formed several muscle fibers expressing dystrophin and replenished the SC compartments [bib_ref] Stem cell-mediated transfer of a human artificial chromosome ameliorates muscular dystrophy, Tedesco [/bib_ref]. More recently, Cossu's group obtained mesoangioblasts from iPSCs of LGMD-2D patients that rescued the expression of α-sarcoglycans in dystrophic mice [bib_ref] Transplantation of genetically corrected human iPSC-derived progenitors in mice with limb-girdle muscular..., Tedesco [/bib_ref]. According to these evidences, mesoangioblasts seemed to be feasible to treat MDs and they are currently being utilized in a phase I/II clinical trial (EudraCT no. 2011-000176-33).
## Artificial stem cell niche satellite cells niche
SCs behavior is influenced by factors that are secreted by myofibers. SDF-1 can bind to receptor CXCR4 on the surface of SC activating a migratory response [bib_ref] Isolation of adult mouse myogenic progenitors: functional heterogeneity of cells within and..., Sherwood [/bib_ref] [bib_ref] Embryonic stem cell-derived microvesicles reprogram hematopoietic progenitors: evidence for horizontal transfer of..., Ratajczak [/bib_ref] while M-cadherin enhance the adhesion of SC to myofibers allowing their fusion [bib_ref] Expression pattern of M-cadherin in normal, denervated, and regenerating mouse muscles, Irintchev [/bib_ref]. Interestingly, SCs can regulate their own quiescence and self-renewal according to the expression of ligands for the Notch receptor family [bib_ref] The regulation of Notch signaling controls satellite cell activation and cell fate..., Conboy [/bib_ref] [bib_ref] Notchmediated restoration of regenerative potential to aged muscle, Conboy [/bib_ref] [bib_ref] Asymmetric selfrenewal and commitment of satellite stem cells in muscle, Kuang [/bib_ref]. Like other stem cells, SCs can proliferate in a asymmetric manner, giving rise to one stem cell and one differentiated cell; and in a symmetric manner, originating two daughter cells retaining full stem cell potential [bib_ref] Asymmetric and symmetric stemcell divisions in development and cancer, Morrison [/bib_ref]. Asymmetric self-renewal is preferred in quiescient conditions while the other is typical in case of injury or disease. Each tissue-specific stem cell is located inside anatomically-defined microenvironment, called niche, surrounded by extracellular matrix (ECM) composed of a network of fibrillar proteins, growth factors, chemokines, cytokines and proteins that are present on the surface of neighboring cells. According to the interactions with these components, the cell choose self-renewal or a pathway of differentiation, following specific stimuli [bib_ref] Synthetic biomaterials as instructive extracellular microenvironments for morphogenesis in tissue engineering, Lutolf [/bib_ref] [bib_ref] A home away from home: challenges and opportunities in engineering in vitro..., Cosgrove [/bib_ref]. SCs reside in the niches that are positioned in a compartment between the myofiber plasma membrane and the basal lamina that surrounds the myofiber so that in the apical part of the niche they receive the signals from the myofibers while on their basal surface they are influenced by basal lamina signals [bib_ref] Niche regulation of muscle satellite cell self-renewal and differentiation, Kuang [/bib_ref]. SCs express several molecules to interact with the basal lamina and all its components (collagen, laminin, fibronectin) [bib_ref] The alpha7beta1 integrin in muscle development and disease, Burkin [/bib_ref]. Conversely, the proteoglycan components of the basal lamina bind growth factors secreted by SCs such as basic Fibroblast Growth Factors (bFGF), and Insulin-like growth factor 1 (IGF-1) that regulate SC survival and proliferation [bib_ref] Skeletal muscle stem cells express anti-apoptotic ErbB receptors during activation from quiescence, Golding [/bib_ref] [bib_ref] Wnt7a activates the planar cell polarity pathway to drive the symmetric expansion..., Grand [/bib_ref]. Other factors derived from cells that are not proximal to the niches or from the systemic circulation can influence SCs, such as myostatin, and wingless-type MMTV integration site family, member 3a (Wnt3a) [bib_ref] Myostatin negatively regulates satellite cell activation and self-renewal, Mccroskery [/bib_ref] [bib_ref] Intrinsic changes and extrinsic influences of myogenic stem cell function during aging, Brack [/bib_ref]. These extrinsic factors play a fundamental role in aging, when the regenerative capacity of skeletal muscle declines [bib_ref] Phagocytosis of necrotic muscle in muscle isografts is influenced by the strain,..., Grounds [/bib_ref] : for example, increased levels of circulating Wnt3a allowed the activation of β-catenin pathway in SCs, so that muscle regeneration is reduced and fibrosis is enhanced [bib_ref] A temporal switch from notch to Wnt signaling in muscle stem cells..., Brack [/bib_ref]. The incredible complexity of niche regulation is the reason why, after removal from their in vivo localization, SCs-and other adult stem cells-rapidly lost their myogenic ability [bib_ref] High-resolution video monitoring of hematopoietic stem cells cultured in single-cell arrays identifies..., Dykstra [/bib_ref] so that they cannot be used in clinical trials [bib_ref] Cell based therapy for Duchenne muscular dystrophy, Farini [/bib_ref]. As Kuang and collaborators suggested, the balance among the signals deriving from the various components of the niche is necessary to maintain the myogenic potential of the SCs [bib_ref] Niche regulation of muscle satellite cell self-renewal and differentiation, Kuang [/bib_ref].
Recent studies have focused on imitate the regulatory machinery of the in vivo SC niche, as a powerful tool to control stem cell function. Three dimensional (3-D) matrices are the model system that mimics the in vivo microenvironment, allowing the investigation of these physiologic events [bib_ref] Taking cell-matrix adhesions to the third dimension, Cukierman [/bib_ref] [bib_ref] Cell culture: biology's new dimension, Abbott [/bib_ref]. They can derive from cells or tissues while others can be composed of ECM proteins. Natural ECMs can be formed by various protein fibrils and fibers interwoven within a hydrated network of glycosaminoglycan chains, providing a structural scaffold. Fibrils, pores, elastin and collagen can be present and alter the biophysical properties of ECMs. Moreover, artificial synthetic materials were produced with similar structure. Polyethylene glycol (PEG)-based hydrogels were used for the maintenance of SCs in vitro [bib_ref] Synthetic biomaterials as instructive extracellular microenvironments for morphogenesis in tissue engineering, Lutolf [/bib_ref] while, recently, Kloxin and colleagues developed PEG hydrogels that controlled matrix stiffness without toxicity to cells [bib_ref] Photodegradable hydrogels for dynamic tuning of physical and chemical properties, Kloxin [/bib_ref]. As these matrices were able to alter biophysical properties in a non-invasive manner, they were used to investigate the progression of biophysical changes associated with muscle fibrosis or disease [bib_ref] Myotubes differentiate optimally on substrates with tissue-like stiffness: pathological implications for soft..., Engler [/bib_ref]. Moreover, Lutolf et al. demonstrated that PEG hydrogels were suitable for single-stem cell clonal assays and resistant to non-specific cell adhesion mediated by protein adsorption [bib_ref] Perturbation of single hematopoietic stem cell fates in artificial niches, Lutolf [/bib_ref]. However, further studies are necessary to define exactly all the components that constitute the microenvironment of the SCs and the molecular steps that regulate the transition between SCs quiescence and proliferation.
## Tem cell fate in vitro
In vitro stem-cell colture is carried out on flat coated with different substrates like collagen or laminin, on feeder-cell layers and within hydrogels synthetized from ECM components (for example collagen or Matrigel). Most frequently culture of stem cells was performed on rigid polystyrene tissue-culture plastic exposing cells to soluble factors in liquid media [bib_ref] Designing materials to direct stem-cell fate, Lutolf [/bib_ref].
These culture conditions are far from resemble the in vivo condition, where cells live in close proximity to each other and in contact with the ECM. Recently, 3D niche are still being explored and should be considered. Blau's group are studying the twodimensional (2D) biomaterial culture systems deconstructing the niche and identifying and assessing the effects of individual niche components on stem-cell fate [bib_ref] Designing materials to direct stem-cell fate, Lutolf [/bib_ref]. Normally, the effects of cell-cell interactions are studied by coculturing; this strategy makes it difficult to discriminate the role of particular molecules.
In vivo, secreted growth factors and cytokines are mostly tethered to ECM components like proteoglycans. At the same time, receptor ligands are presented to stem cells surface and to nearby support cells. In both cases, molecule immobilization probably has the critical role of increasing protein stability, promoting persistent signaling and inducing receptor clustering [bib_ref] Simulations of cell-surface integrin binding to nanoscale-clustered adhesion ligands, Irvine [/bib_ref]. A covalent binding of fibroblast growth factor 2 (FGF2) to a synthetic polymer stabilized the growth factor and increased its potency 100-fold relative to FGF2 in solution. Similarly, the epidermal growth factor (EGF) covalently tethered to a biomaterial scaffold, was shown to be more effective than its soluble counterpart in inducing mesenchymal stem cells differentiation and preventing Fas-ligand-induced death [bib_ref] Tethered epidermal growth factor provides a survival advantage to mesenchymal stem cells, Fan [/bib_ref]. Natural and synthetic matrices can be used to create cell-culture substrates with known elastic modulus providing diffusion of soluble molecules to the basal surface and the apical one, and can be used to test the relevance of homeostatic and disease related matrix stiffness to stem-cell behavior. Soluble factors in culture media used in combination with the tissue-culture matrix affect cell fate. Human MSCs expressed genes consistent with differentiation into distinct tissue-specific cell types when exposed to polyacrylamide gels with a range of stiffness typical of brain, muscle and bone [bib_ref] Matrix elasticity directs stem cell lineage specification, Engler [/bib_ref]. The effects of the physical properties of culture substrate on stem-cell fate are fully appreciated, culture platforms based on soft biomaterials are likely to replace, rigid, tissue-culture plastic. Within the niche, cells dialog with the surrounding ECM during development and in adulthood [bib_ref] Role of cell shape in growth control, Folkman [/bib_ref]. Although some of these effects are probably due to alterations in the adhesive interactions and crosstalk between the ECM and the cell as they work to define each other, there is ample evidence suggesting that physical control of cell shape alone can act as a potent regulator of cell signaling and fate determination [bib_ref] Mechanotransduction in development: a growing role for contractility, Wozniak [/bib_ref].
## Stem cell fate in vivo
Biomaterials technologies offer great opportunities to control the stem cell fate in vivo, especially in case of tissue damage. Two main modes of application have been proposed: one in which biomaterials are used as carriers for introducing stem cells into damaged, diseased or aged tissue, and one in which biomaterials are used to augment endogenous stem-cell function [bib_ref] Designing materials to direct stem-cell fate, Lutolf [/bib_ref]. In regenerative medicine, stem cell transplantation has some limitations: survival and engraftment of transplanted stem cells and the disrupted biological environment characterized by abundant cell and tissue necrosis. Biomaterials have to be designed to act as carriers for local delivery of stem cells, supporting cells or molecular niche cues. Biomaterials may improve the effect of stem cell transplantation; they may be used as multifunctional stem-cell microenvironments. They have to increase the delivering and enhancing the viability of the cells, to function as support in order to increase the numbers of the cells and stimulate the function of endogenous stem cells. Moreover, biomaterials can deliver diffusible cytokines in order to promote the mobilization of endogenous cells involved in repair, to enhance survival and to stimulate self-renewal and expansion of the transplanted cells. Materials would enhance tissue regeneration, tissue function and overcome the adverse effects of disease or ageing [bib_ref] Rejuvenation of aged progenitor cells by exposure to a young systemic environment, Conboy [/bib_ref] [bib_ref] Therapeutic targeting of a stem cell niche, Adams [/bib_ref]. Therefore, they could permit local and specific delivery of bioactive niche components able to inhibit and stimulate molecules and drugs that have to increase the number and the functions of transplanted stem cells. In order to obtain these benefits in vivo, materials have to be achieved by forming a scaffold that deliver biomolecules near the stem-cell niche or by targeted delivery of soluble microparticles or as carriers of such bioactive niche components [bib_ref] Therapeutic targeting of a stem cell niche, Adams [/bib_ref]. Recently, Rothenfluh et al. isolated polymer nanoparticles, sufficiently small to enter the matrix of the targeted tissues; then, they modified them with a biomolecular ligand for matrix binding. This way, the modified the matrix into a source of nanoparticles [bib_ref] Precise engineering of targeted nanoparticles by using self-assembled biointegrated block copolymers, Gu [/bib_ref] [bib_ref] Biofunctional polymer nanoparticles for intra-articular targeting and retention in cartilage, Rothenfluh [/bib_ref]. Similarly, Gu and co-workers modified existing nanoparticles so that they were used for differential delivery and controlled release of drugs [bib_ref] Precise engineering of targeted nanoparticles by using self-assembled biointegrated block copolymers, Gu [/bib_ref]. Biomaterials aim is not only to create materials to control spatially and temporally the components of the niche but also to study microenvironmental regulation of stem cell proliferation and fate [bib_ref] Rejuvenation of aged progenitor cells by exposure to a young systemic environment, Conboy [/bib_ref]. Artificial niches could incorporate appropriate "homing" signals that would attract endogenous stem cells and localize them by means of known cell-cell or cell-matrix adhesive interactions. Biomaterial research is focused on create artificial niche where cells could to be exposed to tethered signals that control stem-cell function and expansion by self-renewal division.
## Muscle pathophysiology
## Muscle fibrosis
Following injury, a cascade of events starts to repair damaged tissues. First of all, inflammatory cells phagocyte the cell debris and secrete growth factors and cytokines that allow the proliferation of other cell types in the site of injury, as described in details below (see section Inflammation and repair mechanisms in skeletal muscle). Then, SCs start to proliferate and differentiate, a process which ultimately ends with the formation of new muscle fibers. Unfortunately, in muscular pathologies, the deficiency of structural proteins leads to continuous cycles of myofiber degeneration and regeneration, so that the damaged muscle fibers cannot be replaced by new fibers, causing myofiber degeneration, inflammation and fibrosis [bib_ref] Strength at the extracellular matrix-muscle interface, Grounds [/bib_ref] [bib_ref] Regulation and dysregulation of fibrosis in skeletal muscle, Serrano [/bib_ref]. In particular, the inflammatory cells eliminate the basement membranes of necrotic fibers that cannot be used to build the new fibers: this condition leads to abnormal muscle fiber arrangement in dystrophic muscles. Due to the chronic persistence of inflammatory cells, dystrophic muscles are characterized by higher concentration of growth factors and cytokines, that induce the massive proliferation and activation of fibroblasts. Their activity causes the accumulation of fibrotic elements that are responsible for uncontrolled events such as remodeling of the basal lamina and formation of collagenous tissues [bib_ref] Regulation and dysregulation of fibrosis in skeletal muscle, Serrano [/bib_ref]. Normally, the events of muscle regeneration are tightly controlled by the interplay among different molecules. Insulin-like growth factor (IGF) is a key element in controlling tissue activity: it binds to cell surface receptors and to IGF-binding proteins, exerting a fundamental role in modulating myofibroblast and SCs proliferation. The matrix metallo-proteases (MMPs) have the function to degrade the ECM and to recruit inflammatory and myogenic cells in the site of injury while Sca-1 inhibits myoblast proliferation, preserving the progenitor cells [bib_ref] Regulation and dysregulation of fibrosis in skeletal muscle, Serrano [/bib_ref]. Transforming growth factor (TGF)-β is highly expressed in regenerating muscle and it is a key regulator of fibrosis' development [bib_ref] Temporal and spatial mRNA expression patterns of TGF-beta1, 2, 3 and TbetaRI,..., Zhou [/bib_ref] ; often, it functions in synergy with connective tissue growth factor (CTGF), inducing fibrosis and promoting dedifferentiation of myoblasts [bib_ref] Skeletal muscle cells express the profibrotic cytokine connective tissue growth factor (CTGF/CCN2),..., Vial [/bib_ref]. CTGF binds to IGF-binding proteins and it is associated with fibrotic remodeling.
In the case of MDs, especially in DMD, membranes lacking the members of the dystroglycan complex are vulnerable to mechanical and oxidative stress. Due to myofiber breakdown, myofibroblasts remained activated: these phenomena are associated with altered production of ECM components and the accumulation of these molecules that lead to muscle cell necrosis and fibrosis [bib_ref] The role of fibrosis in Duchenne muscular dystrophy, Klingler [/bib_ref]. Fibrosis development was considered a progressive and irreversible pathologic phenomenon, but recent advances in knowledge of its development steps render this pathological feature amenable for clinical treatments. A better understanding of the factors that participate in fibrosis may help identify pharmacological targets capable of attenuating the progression of untreatable muscular diseases.
## Muscular hypertrophy and atrophy: two opposites of the same phenomenon
Skeletal muscle is the most abundant tissue in mammals and muscle remodeling occurs throughout the entire life. A fine regulated pathway determines the balance between new protein accumulation and degradation of pre-existing ones [bib_ref] Signaling in muscle atrophy and hypertrophy, Sandri [/bib_ref]. Different stimuli, originated by functional overload or aging, can modulate this pathway causing a shift in this balance toward one side. Besides of physiological conditions, this pathway is influenced by lots of inherited and acquired disorders such as MDs, cancer cachexia and commons drugs as glucocorticoids [bib_ref] Cellular mechanisms and local progenitor activation to regulate skeletal muscle mass, Cassano [/bib_ref]. Among signals that can produce hypertrophy, IGF1 pathway is one of the best characterized. IGF-1Ec is expressed in response to mechanical stimuli and cellular damage and promotes both proliferation and differentiation of satellite cells, while in adult myofibers it increases DNA content per myofiber and can influence myosin phenotype [bib_ref] Mechanical load increases muscle IGF-I and androgen receptor mRNA concentrations in humans, Bamman [/bib_ref]. The binding of IGF-1 to its receptor IGF1R, triggers the activation of several kinases including phosphatidylinositol-3-kinase (PI3K), the consequent production of PIP3 recruits protein kinase B (AKT). AKT plays a central role in muscle remodeling: it acts by either activating positive signal (mTor) or blocking negative pathway (Myostatin, apoptotic cascade, GSK3β). A trophy results from degradation of both myofiber number and protein contents, through calpain system, lysosomal and the ubiquitin-proteasoma pathways [bib_ref] Altered expression of eukaryotic initiation factor 2B in skeletal muscle during sepsis, Voisin [/bib_ref] [bib_ref] Ubiquitin conjugation by the N-end rule pathway and mRNAs for its components..., Lecker [/bib_ref]. Two genes were found up-regulated in atrophy models: muscle-specific ubiquitin ligase atrogin-1 (MAFbx) and muscle RING-finger protein-1 (MURF1); further studies showed that they were ubiquitin-ligase expressed only in skeletal and cardiac muscle [bib_ref] Identification of ubiquitin ligases required for skeletal muscle atrophy, Bodine [/bib_ref]. Another important factor is nuclear factor kappa-lightchain-enhancer of activated B cells (NF-kB) which is involved in inflammatory pathway leading to TNF-α and INF-γ expression and it can induce the degradation of MyoD. Moreover, knock out of myostatin, a member of the TGF-β family, can lead to an enormous enlargement of skeletal muscle mass [bib_ref] Regulation of skeletal muscle mass in mice by a new TGF-beta superfamily..., Mcpherron [/bib_ref]. Myostatin is in fact the most important negative regulatory element of fiber synthesis and it is strictly regulated during myogenesis thanks to the presence of E-boxes, MEF2 and GRE binding sites [bib_ref] The myostatin gene is a downstream target gene of basic helix-loop-helix transcription..., Spiller [/bib_ref]. In particular, myostatin is synthesized as a precursor, that is processed by furin proteases to generate a dimer composed by an N-terminal pro-peptide, bound to biologically active C-terminal fragment. When the pro-peptide is cleaved, myostatin is activated and interact with several proteins, such as follistatin. Interestingly, mice without the expression of this protein have a reduced body mass [bib_ref] Multiple defects and perinatal death in mice deficient in follistatin, Matzuk [/bib_ref] while follistatin forced expression leads to muscular hyper-growth [bib_ref] Transgenic expression of a myostatin inhibitor derived from follistatin increases skeletal muscle..., Nakatani [/bib_ref]. To test whether lack of myostatin could ameliorate the symptoms of muscular diseases, Whittemore et al. demonstrated that in wild type mice the blocking of the protein increased muscle mass [bib_ref] Inhibition of myostatin in adult mice increases skeletal muscle mass and strength, Whittemore [/bib_ref] while showed that this condition in mdx mice improved myofibers size and muscular force [bib_ref] Functional improvement of dystrophic muscle by myostatin blockade, Bogdanovich [/bib_ref]. According to these studies, Wagner et al. described a phase I/II clinical trial of MYO-029 (a neutralizing antibody to myostatin) in dystrophic patients. This trial did not demonstrate any improvement in muscle strength, but no side effects were assessed, except for hypersensitivity skin reactions. This trial was originally designed to test safety so that a bigger cohort of patient or different choice of samples are required to detect arrest of disease progression or minimal improvements in strength.
Furthermore, results could be explained by the fact that the patients were selected at late stage of the disease when the regenerative response is exhausted and the myostatin substrate was eliminated [bib_ref] A phase I/IItrial of MYO-029 in adult subjects with muscular dystrophy, Wagner [/bib_ref]. Similar studies were conducted also with animal models of other muscular diseases but opposite results were obtained [bib_ref] Elimination of myostatin does not combat muscular dystrophy in dy mice but..., Li [/bib_ref] [bib_ref] Muscular atrophy of caveolin-3-deficient mice is rescued by myostatin inhibition, Ohsawa [/bib_ref]. Further works demonstrated that myostatin not only downregulates the expression of several myogenic genes [bib_ref] The regulation and action of myostatin as a negative regulator of muscle..., Amthor [/bib_ref] [bib_ref] Myostatin signals through Pax7 to regulate satellite cell self-renewal, Mcfarlane [/bib_ref] but efficiently inhibits the proliferation of muscle progenitor cells [bib_ref] Myostatin, a negative regulator of muscle growth, functions by inhibiting myoblast proliferation, Thomas [/bib_ref]. The complexity of mechanism involving muscle growth and regeneration is further increased by the discovery of microRNA. Recently, skeletal muscle specific microRNAs able to interact with master regulatory genes in muscle development were found . As an example, miRNA206 can influence satellite cells behavior by modulating Pax3 and MET transforming gene (cMet) [bib_ref] A mutation creating a potential illegitimate microRNA target site in the myostatin..., Clop [/bib_ref] [bib_ref] MicroRNA-206 is overexpressed in the diaphragm but not the hindlimb muscle of..., Mccarthy [/bib_ref].
## Inflammation and repair mechanisms in skeletal muscle
Injuries affecting skeletal muscle determine the activation of the immune system and activate a cascade of events that are required to clean cellular debris and to allow the replacement of lost fibers with new ones. Furthermore, immune cells promote regeneration through the release of growth factors [bib_ref] The immune system and the repair of skeletal muscle, Brunelli [/bib_ref]. After acute muscular damage neutrophils rapidly appear, followed by phagocytic macrophages which continue to increase in numbers until about 2 days post-injury. A second population of macrophages develops at about 4 days postinjury and it is characterized by a non-phagocytic phenotype [bib_ref] Regulatory interactions between muscle and the immune system during muscle regeneration, Tidball [/bib_ref]. In parallel, myogenic precursors start to proliferate and differentiate by recapitulating developmental steps. Firstly, response to injury is mediated by Th1 cytokines (INFγ and TNFα) which trigger the activation of classic M1 proinflammatory macrophages [bib_ref] Monocyte and macrophage heterogeneity, Gordon [/bib_ref]. At a second stage, a population of M2 anti-inflammatory macrophages is predominant thanks to Th2 cytokines stimulation, such as interleukin (IL)−4, −10, −13. This phenotype-switch is required to stop inflammation and to permit the differentiation and fusion of satellite cells. This process is strictly regulated and several signals are known to be involved [bib_ref] Differential effects of apoptotic versus lysed cells on macrophage production of cytokines:..., Fadok [/bib_ref] [bib_ref] Inflammatory monocytes recruited after skeletal muscle injury switch into antiinflammatory macrophages to..., Arnold [/bib_ref] but further studies are needed to better understand each phase. In MDs, skeletal muscles are subjected to chronic injuries that maintain a continue activation of the immune system. In fact, inflammatory infiltrates consisting of both macrophages and lymphocytes are present and elevated serum cytokines levels are detectable. Furthermore, a partial adaptive response to treatment with corticosteroid supports a role for the immune system in exacerbating muscular wasting [bib_ref] Low-dose prednisolone treatment in Duchenne and Becker muscular dystrophy, Backman [/bib_ref]. Progressive MDs like DMD are characterized by an initial phase that recapitulates the event observed in acute injury and repair. A second phase is dominated by chronic inflammation which triggers fibrosis deposition and atrophy. In fact in adult mdx mice a transition from M2a macrophages to M2c macrophages occurs in an attempt to control M1 cytolitic macrophages and to promote muscle regeneration through the release of IL-10 and IL-4 [bib_ref] Alternative activation of macrophages, Gordon [/bib_ref] [bib_ref] IL-4 acts as a myoblast recruitment factor during mammalian muscle growth, Horsley [/bib_ref]. M2 macrophages may also partecipate in activation of cytotoxic T-cells (which promote muscle damage through perforin-mediate process) and promote muscle fibrosis through arginase metabolism of arginine [bib_ref] Shifts in macrophage phenotypes and macrophage competition for arginine metabolism affect the..., Villalta [/bib_ref] [bib_ref] Regulatory interactions between muscle and the immune system during muscle regeneration, Tidball [/bib_ref]. The importance of modulating immune system cells was proven in different animal model of MDs, for example depletion of macrophages from mdx mice resulted in reduced muscle membrane lysis [bib_ref] Dystrophin protects the sarcolemma from stresses developed during muscle contraction, Petrof [/bib_ref]. Furthermore, nonsteroidal anti-inflammatory drug (NSAID) treatment was effective both in ameliorating muscle morphology and reducing macrophage infiltration [bib_ref] Inflammation in muscular dystrophy and the beneficial effects of non-steroidal anti-inflammatory drugs, Serra [/bib_ref] and anti-oxidant drugs (N-acetylcysteine) in mdx mice reduced necrosis by regulating TNF-α level (De Senzi Moraes [bib_ref] N-acetylcysteine treatment reduces TNF-alpha levels and myonecrosis in diaphragm muscle of mdx..., Pinto [/bib_ref]. Recently an important role for acquired immunity in DMD pathogenesis has been pointed out by [bib_ref] Myoblast transfer in the treatment of Duchenne's muscular dystrophy, Mendell [/bib_ref] [bib_ref] Cancerous stem cells can arise from pediatric brain tumors, Hemmati [/bib_ref] [bib_ref] Anti-dystrophin T cell responses in Duchenne muscular dystrophy: prevalence and a glucocorticoid..., Flanigan [/bib_ref] opening new perspectives in treatment of MDs.
# Conclusions
Skeletal muscle emerged as a promising tissue source for stem and progenitor cells that can be used in a variety of therapeutic applications. Skeletal muscle constitutes around one third of body weight in a healthy subjects [bib_ref] Regenerative medicine for the musculoskeletal system based on muscle-derived stem cells, Gates [/bib_ref]. Muscle has an high capacity to repair itself after injury; this characteristic suggests that it serves as a reservoir for cells that participate in tissue regeneration processes [bib_ref] Muscle-derived stem cells for tissue engineering and regenerative therapy, Usas [/bib_ref]. Several works described the ability of different muscle-derived stem cell populations to differentiate into multiple cell types, including osteoblasts, adipocytes, chondrocytes, myoblasts and endothelial cells. In addition, these cells showed regenerative, anti-inflammatory and anti-apoptotic properties. Each of these cell types is characterized primarily on the basis of their in vitro characteristics after they have been isolated from the body. In vivo they exhibited the capacity to migrate through different tissues where they are exposed to different extracellular and environmental signals. While rudimentary models were developed to describe the in vivo relationship among these stem cell www.frontiersin.org February 2014 | Volume 5 | Article 48 | 7 populations, substantial additional studies are needed to refine and verify these relationships. New approaches using organisms genetically modified and transgenic mouse models proposed the importance of the microenvironment-like the niche and the extrinsic factors-to be a key component in stem cell regulation. Particularly, significant progress has been made in understanding how satellite cells can act as tissue-specific adult stem cells in skeletal muscle. In the same time, many studies investigated the satellite cell properties in term of efficacy after in vivo transplantation using novel approaches such as non-invasive bioluminescence imaging. These tools provided information for assessing not only satellite cell function but, in general, stem cell function. Investigations on the molecular nature of stem cell niche signals on in vivo models and short-term cultures of isolated myofibers, are now on-going. Bioengineering offers significant tools for the development of strategies to mimic biochemical and biophysical features of the in vivo niche microenvironment [bib_ref] Designing materials to direct stem-cell fate, Lutolf [/bib_ref]. We hope that the synthesis of biomaterials, micro-fabrication technology and stem cell biology will provide systems potentially innovative to better understand how stem cell fate is controlled. The analysis of the niche and the dynamic responses of stem cells to welldefined artificial microenvironments, might give us the possibility to understand the role of specific niche components and niche architecture in regulating fundamental cellular mechanisms such as cellular division, self-renewal, and differentiation in vitro and in vivo. Development of biomaterials able to re-create an in vitro SCs niche could give rise to novel insights into understanding the molecular cues, critical for the in vitro maintenance and expansion of muscle stem cells. Above all, these in vitro systems can well lead to the generation of adequate numbers of stem cells and the ability to control their differentiation in order to maximize their utility, not only as cell-based therapeutics for tissue regeneration and replacement, but also as the control of inflammation after muscle damage [bib_ref] A home away from home: challenges and opportunities in engineering in vitro..., Cosgrove [/bib_ref]. In conclusion, all these considerations will be important not only to better characterize satellite cell biology and therapeutic approaches to treat muscle diseases and aging-related muscle wasting, but also to give necessary information for the study of adult tissue-specific stem cells.
# Author contributions
Mirella Meregalli and Yvan Torrente designed the approach, Andrea Farini and Clementina Sitzia wrote the manuscript.
www.frontiersin.orgFebruary 2014 | Volume 5 | Article 48 | 5
ACKNOWLEDGMENTSThis work was supported by Associazione La Nostra Famiglia Fondo DMD Gli Amici di Emanuele, Associazione Amici del Centro Dino Ferrari, Ministry of Health (RF-2009-1547384). |
Structural characterization of the virulence factor Sda1 nuclease from Streptococcus pyogenes
Infection by Group A Streptococcus pyogenes (GAS)is a leading cause of severe invasive disease in humans, including streptococcal toxic shock syndrome and necrotizing fasciitis. GAS infections lead to nearly 163,000 annual deaths worldwide. Hypervirulent strains of S. pyogenes have evolved a plethora of virulence factors that aid in disease--by promoting bacterial adhesion to host cells, subsequent invasion of deeper tissues and blocking the immune system's attempts to eradicate the infection. Expression and secretion of the extracellular nuclease Sda1 is advantageous for promoting bacterial dissemination throughout the host organism, and evasion of the host's innate immune response. Here we present two crystal structures of Sda1, as well as biochemical studies to address key structural features and surface residues involved in DNA binding and catalysis. In the active site, Asn211 is observed to directly chelate a hydrated divalent metal ion and Arg124, on the putative substrate binding loop, likely stabilizes the transition state during phosphodiester bond cleavage. These structures provide a foundation for rational drug design of small molecule inhibitors to be used in prevention of invasive streptococcal disease.
# Introduction
The Group A Streptococcus pyogenes (GAS), a strictly human pathogen, is a leading cause of over 700 million cases per year of superficial infections (pharyngitis and impetigo), as well as greater than 650,000 cases of invasive infections [bib_ref] The global burden of group A streptococcal diseases, Carapetis [/bib_ref]. Examples of invasive infection include cellitis, bacteraemia, streptococcal toxic shock syndrome (STSS) and necrotizing fasciitis, and GAS infections have been estimated to be the primary cause of nearly 163,000 deaths annually. The 1980s witnessed an explosive resurgence of in-vasive streptococcal infections worldwide, which has been attributed largely to the emergence of a highly virulent GAS subclone with serotype M1T1 [bib_ref] Rise and persistence of global M1T1 clone of Streptococcus pyogenes, Aziz [/bib_ref]. This clone differs from other, less pathogenic GAS strains in that it contains three prophages--M1T1.X, M1T1.Y and M1T1.Z.
M1T1.Y is similar to the phage 307.3 of the M1-SF370 strain, but M1T1.X and M1T1.Z, encoding virulence factors SpeA2 and Sda1, respectively, are unique to the M1T1 subclone [bib_ref] Mosaic prophages with horizontally acquired genes account for the emergence and diversification..., Aziz [/bib_ref]. The acquisition of multiple prophages allows bacteria to exchange virulence factors, toxins and gain antibiotic resistance, and is an evolutionarily beneficial process by which bacteria adapt to a changing environment [bib_ref] Transfer of scarlet fever-associated elements into the group A Streptococcus M1T1 clone, Ben Zakour [/bib_ref] [bib_ref] Emergence of scarlet fever Streptococcus pyogenes emm12 clones in Hong Kong is..., Davies [/bib_ref].
GAS has developed a vast arsenal of virulence factors that work synergistically to promote efficient bacterial infection of the host organism. These factors play a variety of different roles during infection, including adhesion to host cells, invasion and spread to deeper tissues, toxins to break down those tissues and evasion of the host immune response [bib_ref] Molecular insight into invasive group A streptococcal disease, Cole [/bib_ref]. One of the two virulence factors distinguishing the M1T1 strain, SpeA2 (Streptococcal pyrogenic exotoxin A2), is a superantigen whose function involves simultaneous engagement of major histocompatibility complexclass II molecules and T-cell receptor [bib_ref] TCR recognition of peptide/MHC class II complexes and superantigens, Sundberg [/bib_ref]. This interaction may cause a catastrophic overproduction of cytokines and is directly correlated with severe forms of STSS [bib_ref] Host variation in cytokine responses to superantigens determine the severity of invasive..., Norrby-Teglund [/bib_ref] [bib_ref] An immunogenetic and molecular basis for differences in outcomes of invasive group..., Kotb [/bib_ref].
Sda1, identified as a homologue of SdaD, displays potent sequence-nonspecific nuclease activity on DNA substrates in a divalent cation-dependent manner [bib_ref] Post-proteomic identification of a novel phage-encoded streptodornase, Sda1, in invasive M1T1 Streptococcus..., Aziz [/bib_ref] [bib_ref] DNase Sda1 allows invasive M1T1 Group A Streptococcus to prevent TLR9-dependent recognition, Uchiyama [/bib_ref]. Overexpression of secreted Sda1 occurs during phase-shifts to hypervirulent invasive streptococcal infection [bib_ref] Invasive M1T1 group A Streptococcus undergoes a phase-shift in vivo to prevent..., Aziz [/bib_ref] [bib_ref] DNase Sda1 provides selection pressure for a switch to invasive group A..., Walker [/bib_ref] and is thought to play a role in evasion of the host's innate immune response in a variety of different ways. Sda1 is thought to be involved primarily in degradation of the DNA component of chromatin-rich neutrophil extracellular traps (NETs), and the similar DNA-based extracellular traps extruded by macrophages [bib_ref] DNase Sda1 allows invasive M1T1 Group A Streptococcus to prevent TLR9-dependent recognition, Uchiyama [/bib_ref]. The secreted nuclease activity of Sda1 provides a protective effect against bacterial killing by activated neutrophils and macrophages. Additionally, it has been recently reported that Sda1 may also provide a protective effect by avoidance of TLR9-mediated recognition Nucleic Acids of unmethylated CpG-rich bacterial DNA, and subsequent cytokine overproduction [bib_ref] DNase Sda1 allows invasive M1T1 Group A Streptococcus to prevent TLR9-dependent recognition, Uchiyama [/bib_ref]. Therefore, the presence of Sda1 may represent an evolutionarily advantageous adaptation that aids in severe streptococcal invasion and disease.
In order to gain a better understanding of Sda1 function as a virulence factor in S. pyogenes infection, catalytically inactive forms (general base His188 replaced with glycine) of the recombinant nuclease were expressed and purified for structural characterization. Two X-ray crystal structures of Sda1(H188G) were determined--the catalytic core with a bound zinc ion in the active site (1.95Å) and the other included a partial structure of the C-terminal region unique to Sda1 (1.90Å). Detailed structural comparison to other structurally related bacterial nucleases led to identification of surface residues contributing to catalytic activity, which were tested using an imidazole chemical rescue strategy. Asn211 was discovered to directly chelate the divalent metal cation in the active site, and His188 serves as the general base. On the putative DNA substrate binding loop, Arg124 likely stabilizes the transition state intermediate during cleavage of the phosphodiester bond. Analysis of the Sda1 active site reveals that the canonical hydrogen bonding network that exists in related nucleases is not conserved and have been functionally replaced by a different web of interactions in this nuclease. This study provides a deeper understanding of the relationship between enzyme function and invasive streptococcal disease.
# Materials and methods
## Cloning, expression and purification of sda1 for biochemical assays
Sequences encoding Sda1 [fig_ref] Figure 1: Sequence, domain structure and activity of Sda1 [/fig_ref] and B) were cloned into the NotI and BamHI restriction sites of the pGEXM (16) expression vector, fusing Sda1 to the C-terminus of glutathione-S-transferase (GST). Due to historical difficulties with solubly expressing wild-type nucleases in Escherichia coli, all constructs contained a putatively deactivating mutation of the general base (His188) to glycine. The resulting constructs were transformed into Rosetta2 (DE3) cells (Novagen) and expressed in LB medium. Cultures were grown at 37 - C, with shaking at 275 rpm, to an OD 600nm of 0.6-0.8, at which point the temperature was decreased to 18 - C. Protein expression was induced by addition of isopropyl -D-1-thiogalactopyranoside to a final concentration of 0.4 mM and continued overnight. Cells were pelleted by centrifugation and lysed by sonication in 25 mM Tris pH 8, 500 mM NaCl. The resulting lysate was clarified by further centrifugation and the soluble Sda1 was bound in-batch to glutathione sepharose 4B resin (GE Healthcare) at 4 - C. The N-terminal GST fusion tag was removed by onresin Tobacco Etch Virus protease cleavage. Soluble, cleaved Sda1 proteins were concentrated and further purified by size exclusion chromatography. For detailed information on purification and crystallization of proteins, see the Supplementary Materials.
## Generation of sda1 mutations and deletion variants
Several loop deletion variants and point mutations on the surface of Sda1 (residues Glu38-Glu390) were gen-erated on the background of the H188G mutant, using QuikChange mutagenesis (Agilent). The resulting variants were expressed and purified in small scale as GST-fusion proteins, as described above for the H188G mutant alone. The mutants were concentrated to 1-4 mg/ml in a storage buffer containing 25 mM Tris pH 8, 100 mM NaCl. All mutants behaved indistinguishably from Sda1 (H188G) in size exclusion chromatography experiments.
## Plasmid conversion nuclease activity assays
The supercoiled population of the pBluescript SK(+) plasmid was purified by gel extraction used as the substrate for the plasmid conversion assay. Sda1 proteins lacking His188 were diluted to 20 nM in 20 mM Tris pH 7, 0.1 mM MgCl 2 and incubated with 15 ng/l of supercoiled pBluescript SK(+) plasmid for 40 min, with or without 30 mM imidazole (pH 7). The reactions were quenched by addition of loading dye containing EDTA. Samples were run on an 0.8% (w/v) agarose gel, dissolved in 1X Tris-Acetate-EDTA buffer. DNA species in the gel were visualized by ethidum bromide staining, scanned with a Typhoon fluorescence imager and analysed using ImageQuant TL. Nuclease activity assays were performed in triplicate.
# Results
## Sequence analysis of sda1 domain structure
Sequence analysis of Sda1 from the M1T1 strain of S. pyogenes shows a high degree of similarity to the SdaD nuclease from the M49 strain [bib_ref] Post-proteomic identification of a novel phage-encoded streptodornase, Sda1, in invasive M1T1 Streptococcus..., Aziz [/bib_ref]. BLAST sequence alignment [bib_ref] Gapped BLAST and PSI-BLAST: a new generation of protein database search programs, Altschul [/bib_ref] also reveals that this nuclease is related to DNA/RNA nonspecific endonucleases (streptodornases) from other Streptococcus species, including Nuclease A (GBS NucA) from S. agalactiae [bib_ref] Nuclease A (Gbs0661), an extracellular nuclease of Streptococcus agalactiae, attacks the neutrophil..., Derre-Bobillot [/bib_ref] , DNA-entry nuclease EndA from S. pneumoniae [bib_ref] Genetic and structural characterization of endA. A membrane-bound nuclease required for transformation..., Puyet [/bib_ref] and another phage-encoded nuclease from S. pyogenes known as Spd1 [bib_ref] The structural characterization of a prophage-encoded extracellular DNase from Streptococcus pyogenes, Korczynska [/bib_ref]. However, strong sequence similarity exists only for the catalytic core of the enzyme (approximately Tyr55-Ser294) [fig_ref] Figure 1: Sequence, domain structure and activity of Sda1 [/fig_ref]. The C-terminal region (Thr295-Glu390) of the protein displays no similarity to any other proteins but streptodornases similar to SdaD from strain M49. The function of this domain remains undetermined.
Hydropathy plots of the Sda1 protein sequence suggest that this nuclease contains a putative transmembrane ␣helix including residues Leu8-Phe21 (21) [fig_ref] Figure 1: Sequence, domain structure and activity of Sda1 [/fig_ref]. Since previously published studies have shown secreted Sda1 in the supernatents of cultured M1T1 S. pyogenes clinical isolates [bib_ref] Post-proteomic identification of a novel phage-encoded streptodornase, Sda1, in invasive M1T1 Streptococcus..., Aziz [/bib_ref] , it seems likely that this nuclease may contain a signal peptide for extracellular extrusion and peptide cleavage. Analysis of the N-terminal region by the SignalP server (22) reveals a possible cleavage site between Ala37 and Glu38, which would yield a mature protein containing residues Glu38-Glu390. Secondary structure prediction suggests that the boundaries of the ordered catalytic domain may extend from Lys60 through Asp285 (23,24) (Supplementary [fig_ref] Figure 1: Sequence, domain structure and activity of Sda1 [/fig_ref]. Though the C-terminal region appears to be largely unstructured, there is one predicted ␣helix (Glu373-Thr378) and three small putative -strands (Thr340-Ala344, Gln348-Trp353 and Val369-Met371). Surface residues mutated in this study are marked by asterisks: green, orange or red based on having minimal, moderate or severe effects on activity, respectively. Disordered residues of the C-terminal domain are marked in light blue. (B) Sda1 construct designs used in this study, with the beginning and ending residues of Sda1 indicated. (C-D) Nuclease activity assays for the different construct lengths of Sda1(H188G), as measured using the imidazole chemical rescue strategy. The DNA substrate is a supercoiled plasmid, which forms an open circle when nicked (blue bar), and is linearized upon accumulation of double-strand breaks (red bar). The linear form is further degraded to a variety of smaller fragments, as visualized by a lower molecular weight smear (green bar). All reactions were performed in triplicate, and the resulting relative product populations were calculated using ImageQuant TL. The indicated error bars represent the standard deviation for each calculation.
Nucleic Acids To explore the roles of the N-and C-terminal regions and the nuclease activity of Sda1, three constructs of varying lengths were generated for bacterial expression: the mature secreted protein (Glu38-Glu390, henceforth referred to as Sda1), a C-terminal truncation (Glu38-Ser294, Sda1. C) and an N-terminal truncation (Gly59-Glu390, Sda1. N) [fig_ref] Figure 1: Sequence, domain structure and activity of Sda1 [/fig_ref]. Historically, wild-type endonucleases are difficult to express in E. coli, due to toxicity resulting in degradation of the bacterial genome. Expression of recombinant Sda1 in E. coli yielded only very small quantities of active protein--sufficient for limited biochemical assays, but not for structural characterization [bib_ref] Post-proteomic identification of a novel phage-encoded streptodornase, Sda1, in invasive M1T1 Streptococcus..., Aziz [/bib_ref]. Therefore, inactive variants of Sda1, containing a mutation of the putative general base His188 to glycine, were created for each construct length. The resulting constructs expressed solubly and were observed to be well-behaved monodisperse populations, as assayed by size-exclusion chromatography (Supplementary [fig_ref] Figure 2: Structural characterization of Sda1 [/fig_ref]. Interestingly, the Sda1 construct representing the mature protein (40.1 kDa) exhibited a decreased retention time, relative to the chicken ovalbumin standard protein (44 kDa), suggesting the possibility of dimerization. In contrast, the Sda1. C construct lacking the C-terminal region (29.2 kDa) exhibited an increased retention time, as compared to the horse myoglobin (17 kDa). The drastic change in retention time cannot be accounted for by simply reducing the protein size by 10.9 kDa. Taken together, these observations suggest that the mature protein could be forming a dimer in solution, possibly mediated by the Cterminal region.
## Evaluation of sda1 oligomeric state by small-angle neutron scattering (sans)
To determine whether the C-terminal domain of Sda1 might play a role in dimerization, the solution scattering behaviours of Sda1 and Sda1. C constructs were analysed using SANS. Comparison of the Guinier plots for Sda1 and Sda1. C suggests that the proteins are monodisperse and display no aggregation at the protein concentrations used in the SANS studies (Supplementary [fig_ref] Figure 3: Structure of EGFPX-Sda1 [/fig_ref]. The intensity at zero scattering angle (I o ) was used to determine molecular mass. I o analysis yielded molecular weights of 39,883 Da and 32,384 Da for Sda1 (actual MW = 40,169 D) and Sda1. C (actual MW = 29,297), respectively (Supplementary [fig_ref] Table 1: Data collection and refinement statistics EGFPX-Sda1 [/fig_ref] and Supplementary [fig_ref] Figure 3: Structure of EGFPX-Sda1 [/fig_ref]. Therefore, the SANS data indicates that both constructs of Sda1 exist in a monomeric state in solution.
## Assessment of the contributions of the sda1 n-and c-termini to nuclease activity
Constructs of Sda1(H188G) were tested for nuclease activity in a plasmid conversion assay, using an imidazole chemical rescue strategy [bib_ref] Structural insights into catalytic and substrate binding mechanisms of the strategic EndA..., Moon [/bib_ref] [bib_ref] Chemical rescue of active site mutants of S. pneumoniae surface endonuclease EndA..., Midon [/bib_ref]. This strategy is based on the principle that exogenously added imidazole may effectively substitute for the histidine (His188) serving as the general base in the active site of Sda1 [bib_ref] S)-Mandelate dehydrogenase from Pseudomonas putida: mutations of the catalytic base histidine-274 and..., Lehoux [/bib_ref]. Imidazole could extract a proton from a nearby water molecule in a similar fashion to a histidine side chain, activating the water for a nucleophilic attack on the scissile phosphate [bib_ref] Structural insights into the mechanism of nuclease A, a betabeta alpha metal..., Ghosh [/bib_ref]. In this assay, Sda1 constructs are presented with a supercoiled plasmid substrate, and the subsequent conversion from this state to either nicked open circle or linearized forms, or even further degradation to smaller fragments, is measured. The Sda1(H188G) mutant has nearly undetectable activity in the absence of imidazole, and its nuclease activity is readily rescued by imidazole addition [fig_ref] Figure 1: Sequence, domain structure and activity of Sda1 [/fig_ref] and D). Comparison of the nuclease activity levels of the mature Sda1 versus Sda1. C shows that removing the C-terminal domain has only a minimal effect on substrate conversion. Both constructs completely transformed the supercoiled substrate to a different form, but the C-terminal truncation variant displayed decreased conversion to the more degraded fragments, possibly indicating a slight decrease in the overall catalytic rate [fig_ref] Figure 1: Sequence, domain structure and activity of Sda1 [/fig_ref] and D). Conversely, removing the first 20 residues of the Sda1. N construct severely abrogated its nuclease activity to barely detectable levels.
## Structural characterization of sda1 using x-ray crystallography
In order to gain an understanding of the structure/function relationship for Sda1, different constructs of the nuclease were screened for crystallization potential. The crystal structure of Sda1. C(H188G) was determined at 1.95Å resolution [fig_ref] Table 1: Data collection and refinement statistics EGFPX-Sda1 [/fig_ref] and contained four molecules in the asymmetric unit. This structure reveals a globular fold for the catalytic domain, wherein a central antiparallel -sheet, largely open to solvent on the 'back' face of the sheet, is flanked on the 'front' face by the ␣-metal finger motif (␣-Me) that forms the active centre [fig_ref] Figure 2: Structural characterization of Sda1 [/fig_ref]. The secondary structural elements comprising the ␣ motif are -strands 10 and 11, and ␣-helix E. Like in other ␣-Me nucleases [bib_ref] Structural insights into catalytic and substrate binding mechanisms of the strategic EndA..., Moon [/bib_ref] [bib_ref] Structural insights into the mechanism of nuclease A, a betabeta alpha metal..., Ghosh [/bib_ref] [bib_ref] Structural characterization of the virulence factor nuclease A from Streptococcus agalactiae, Moon [/bib_ref] , the catalytic site is very compact, with its breadth and depth determined by residues in the ␣ motif and by -strands 12-14 in the central sheet, respectively. The Sda1. C(H188G) active site contains a bound divalent, partially hydrated zinc ion in the active site [fig_ref] Figure 2: Structural characterization of Sda1 [/fig_ref]. The zinc ion is buried within the compact active site and is directly coordinated by Asn211 alone. Other nearby residues within the ␣ motif coordinate three water molecules hydrating the metal (Glu225, Ser187 and the backbone carbonyls of Asn211 and Gly188). The position of the zinc ion is consistent with the pentahydrated catalytic metal ions observed in other ␣-Me nuclease apoprotein structures, as is the observation that only a single asparagine side chain from the protein coordinates the metal [bib_ref] Structural insights into catalytic and substrate binding mechanisms of the strategic EndA..., Moon [/bib_ref] [bib_ref] Structural insights into the mechanism of nuclease A, a betabeta alpha metal..., Ghosh [/bib_ref] [bib_ref] Structural characterization of the virulence factor nuclease A from Streptococcus agalactiae, Moon [/bib_ref]. Canonical octahedral geometry for the metal is disrupted in this structure, due to bifurcated coordination of the ion by the C-terminal carboxyl group of a neighbouring molecule in the crystal lattice (Supplementary [fig_ref] Figure 4: Contribution of Sda1 loop regions to nuclease activity [/fig_ref] and B). This interaction is likely a crystallographic artefact, since truncation of the nuclease at this location is an artificial modification used to enhance crystallization.
Though residues Glu38-Asn45 of the N-terminus are disordered, residues Trp46-Glu58 are visible in the electron density and provide some insight into the decrease in nuclease activity in their absence [fig_ref] Figure 2: Structural characterization of Sda1 [/fig_ref]. Trp46 partially stacks over Tyr54, which begins a -strand (-strand 1, Tyr54-Lys60) that extends the central -sheet. Tyr55 on this strand forms a stacking interaction with Tyr223 on ␣-
[formula] b R work = | |F obs | -|F calc | | / |F obs |, [/formula]
where R free is calculated for a randomly chosen 5% of reflections, which were not used for structure refinement and R work is calculated for the remaining reflections.
helix E. Additionally, there is a putative hydrogen bond between the backbone carbonyl of Asn53 and the backbone amide of Gly220 on the 'finger loop'. The additional Nterminal residues cover a small hydrophobic patch comprised of residues Gly219, Gly220, Val273 and Val275. It is possible that the presence of the N-terminal residues in this region may stabilize the central -sheet, the 'finger loop' and the position of ␣-helix E, aiding in assembly of correct active site architecture for catalysis. Comparison of each of the four Sda1. C(H188G) molecules in the asymmetric unit reveals only slight structural variation (RMS deviation of <0.2Å between each molecule in the asymmetric unit, Supplementary [fig_ref] Figure 4: Contribution of Sda1 loop regions to nuclease activity [/fig_ref]. Insertions (Ins.) I and II are disordered to some extent in two of the Sda1. C(H188G) molecules (molecules A and C). Ins.II displays a small amount of 'hinge' flexibility at the base of its 'pseudoknot' structure, which shifts the position of that loop slightly, relative to the rest of the catalytic domain. Intriguingly, the putative DNA substrate binding loop [fig_ref] Figure 2: Structural characterization of Sda1 [/fig_ref] is ordered in all four molecules of Sda1. C(H188G) in the asymmetric unit, though the N-terminal portion of the loop shows four slightly different conformations in each molecule (Supplementary [fig_ref] Figure 4: Contribution of Sda1 loop regions to nuclease activity [/fig_ref]. These conformational variations may be influenced by packing interactions with neighbouring molecules in the crystal.
Constructs containing the C-terminal domain did not yield usable crystals without the aid of a 'fixed-arm' fusion protein carrier [bib_ref] A synergistic approach to protein crystallization: combination of a fixed-arm carrier with..., Moon [/bib_ref]. Sda1. N(H188G), when used in conjunction as a fixed-arm fusion downstream of EGFP (31), generated diffraction-quality crystals that provided a 1.90 A resolution structure [fig_ref] Table 1: Data collection and refinement statistics EGFPX-Sda1 [/fig_ref] and [fig_ref] Figure 3: Structure of EGFPX-Sda1 [/fig_ref] and B). The EGFP fusion carrier protein is located proximal to the back face of the nuclease, with the three alanine 'fixed-arm' linker residues doubling back on the Sda1 N-terminus (Gly59) to accommodate the fusion [fig_ref] Figure 3: Structure of EGFPX-Sda1 [/fig_ref]. Superposition of the EGFPX-Sda1. N(H188G) and Sda1. C(H188G) structures shows a high degree of structural similarity between the two crystal forms [fig_ref] Figure 3: Structure of EGFPX-Sda1 [/fig_ref]. Unsurprisingly, the configuration of the N-terminus of Sda1. C(H188G) differs from that observed in the EGFP fusion and is likely influenced by the limited flexibility of the linker.
The C-terminal domain of EGFPX-Sda1(H188G), comprised residues Thr295-Glu390, is largely disordered, which is consistent with the lack of predicted secondary structure. However, small portions of the region are visible in the electron density, wedged between the EGFP and the Sda1 catalytic domain [fig_ref] Figure 3: Structure of EGFPX-Sda1 [/fig_ref]. Residues Asp350-Trp353 and His382-Thr385 create a small, two-stranded, antiparallel -sheet (-strands 15 and 16) and Glu373-Arg379 form a short ␣-helix (␣-helix F). Though the degree of thermal motion is higher for this C-terminal domain, the amino acid sequence could be determined by the electron density. It should be noted that the connectivity for this domain could not be determined, since more than 50 amino acid residues are disordered between the C-terminal end of the catalytic domain (Ser294) and the first ordered residues of the C-terminal domain (Ser349), in addition to the nearby location of Ser294 from a neighbouring symmetry-related molecule (Supplementary [fig_ref] Figure 5: Substitution mutagenesis survey of surface residue involvement in Sda1 catalysis [/fig_ref].
One intriguing aspect of ␣-Me nucleases is that their active sites have evolved to bind a hydrated divalent metal, yet the active site in this structure of EGFPX-Sda1. N(H188G) contains only solvent molecules, despite the presence of divalent magnesium ions in the crystallization condition. The crystallization condition also contained a high concentration of sulfate ions, and a few such ions are visible near the active centre in this crystal structure. Since sulfate ions could mimic the negative charge presented by the phosphate backbone of the DNA, the location of these sulfate ions could suggest the likely binding position of the substrate. The region of Sda1 that is commensurate with the DNA substrate binding loop (Tyr123-Pro131) in other ␣-Me nucleases [bib_ref] Structural insights into catalytic and substrate binding mechanisms of the strategic EndA..., Moon [/bib_ref] [bib_ref] Structural characterization of the virulence factor nuclease A from Streptococcus agalactiae, Moon [/bib_ref] is disordered in this structure, which suggests a certain amount of flexibility in the absence of a bound substrate.
## Structural comparison of sda1 with other ␣-me nucleases
Analysis of Sda1. C(H188G) and the nuclease moiety of the EGFPX-Sda1. N(H188G) fusion protein using the Dali server (32) reveals that Sda1 shows the highest degree of structural similarity to GBS NucA from S. agalactiae, EndA from S. pneumoniae and Spd1 encoded by prophage SF370.1 . Values for structurebased sequence identity ranged from 23 to 26%. Z-scores were higher for the Sda1. C(H188G), as compared to the EGFPX-Sda1. N(H188G). Sda1 also shows similarity to SM nuclease from Serratia marcescens, NucA from Anabaena sp., CSP-6 from C. elegans and human mitochondrial EndoG. However, the extent of similarity is lower, as indicated by markedly decreased Z-scores and sequence identity, and increased RMS deviation (Supplementary Table S2).
Structural superposition of Sda1. C(H188G) with the catalytic domain of Spd1 (PDB ID code 2XGR (20), Supplementary , another prophage-encoded ␣-Me nuclease, reveals that the two nucleases share a similar core architecture, yet display some striking differences [fig_ref] Figure 4: Contribution of Sda1 loop regions to nuclease activity [/fig_ref] -C and Supplementary [fig_ref] Figure 6: Functionally important residues for Sda1 [/fig_ref]. The disposition of the N-terminus differs between Sda1 and Spd1, with that of , 'side' (E) and 'front' (F) views, relative to the central -sheet. Magenta asterisk and yellow or red dashed circles mark regions of structural differences between the three nucleases and Sda1. Orange asterisk marks the Spd1 -strand that is structurally but not sequentially homologous to -strand 1 in Sda1. Secondary structural elements are labelled using Sda1 nomenclature, as described in [fig_ref] Figure 1: Sequence, domain structure and activity of Sda1 [/fig_ref] Sda1 beginning on the 'back' face of the enzyme, running -strand 1 alongside the central -sheet, then traversing the entire length of that face to connect to Ins.I [fig_ref] Figure 4: Contribution of Sda1 loop regions to nuclease activity [/fig_ref]. In contrast, the N-terminus of Spd1 begins on the 'front' face of the enzyme, then skirts the side of the central -sheet in a -strand that forms a small, two-stranded antiparallel sheet with another strand from a C-terminal region that is not conserved in Sda1 [fig_ref] Figure 4: Contribution of Sda1 loop regions to nuclease activity [/fig_ref] and C), before it too traverses the 'back' face to the other side of the sheet. Interestingly, -strand 1 that proved to be critical for Sda1 nuclease activity is structurally homologous with a short -strand (residues Lys219-Ile222) from Spd1, though the connectivity is entirely dissimilar and the sequence is not conserved [fig_ref] Figure 4: Contribution of Sda1 loop regions to nuclease activity [/fig_ref] , orange asterisk). Conservation of a -strand at this position underscores the critical importance of stabiliz-ing the central -sheet. On the 'front' face of these enzymes, the architecture of the ␣ motif is very similar, though the 'finger loop' extrusion from the long ␣-helix E varies in length and structure between the two nucleases [fig_ref] Figure 4: Contribution of Sda1 loop regions to nuclease activity [/fig_ref]. Additionally, the return of the 'finger loop' extrusion to the ␣-helix appears to cause a mild distortion at that position in Spd1, which is not observed for Sda1. One critical difference between Sda1 and Spd1 is the length and structure of the putative substrate binding loop. This loop is ordered in the structure of Sda1. C(H188G), and disordered in the structure of EGFPX-Sda1. N(H188G), which implies flexibility in this region in the absence of a DNA substrate. This loop is considerably longer in Spd1 than in Sda1 (Supplementary [fig_ref] Figure 6: Functionally important residues for Sda1 [/fig_ref] and is disordered over a larger area. The substrate binding loop is also disordered in EndA [bib_ref] Structural insights into catalytic and substrate binding mechanisms of the strategic EndA..., Moon [/bib_ref] , and certain structures of GBS NucA [bib_ref] Structural characterization of the virulence factor nuclease A from Streptococcus agalactiae, Moon [/bib_ref]. Given its length in all known ␣-Me nucleases, this loop could play a substantial role in substrate binding for these enzymes, by determining the length, breadth and charge distribution of the DNA binding cleft.
Structural superposition of Sda1. C(H188G) with the structures of GBS NucA [bib_ref] TCR recognition of peptide/MHC class II complexes and superantigens, Sundberg [/bib_ref] and EndA (25) reveals some intriguing structural differences [fig_ref] Figure 4: Contribution of Sda1 loop regions to nuclease activity [/fig_ref] -G and Supplementary . First, the random coil traversing the 'back' face of the enzyme (residues Lys60-Tyr72 in Sda1) displays different conformations in each enzyme [fig_ref] Figure 4: Contribution of Sda1 loop regions to nuclease activity [/fig_ref] , red circle). Slight variations in the conformation of this coil could alter the solvent-exposed landscape on this 'face' of the nuclease. Second, the length of the loop between -strands 4 and 5 in Sda1varies between the three nucleases. This loop is longest in EndA, shortest in GBS-NucA and is an intermediate length in Sda1 closer to that of GBS NucA [fig_ref] Figure 4: Contribution of Sda1 loop regions to nuclease activity [/fig_ref]. The N-terminal end of -strand 4 in Sda1 is slightly splayed outward from that of -strand 5 in the plane of the central -sheet, as compared to EndA and GBS NucA, which wedges open the sheet at that position. There is also a significant variation in the conformation of the N-termini of these nucleases [fig_ref] Figure 4: Contribution of Sda1 loop regions to nuclease activity [/fig_ref]. GBS NucA and EndA display a high degree of similarity in their N-termini, with an ␣-helix traversing the 'shoulder' of the molecule from the 'front' face toward the 'back' face. GBS NucA and EndA also have a small two-stranded antiparallel -sheet on that 'shoulder', which places the second -strand of this sheet (Ile69-Leu72 of GBS NucA and Ala77-Val80 of EndA) in a position roughly similar to that of -strand 1 in Sda1. As seen for Spd1, there are differences in the sequences and structures of the 'finger loop' extrusion from the long ␣-helix (␣-helix E for Sda1) in the ␣ motif, though its length appears to be conserved between Sda1, GBS NucA and EndA [fig_ref] Figure 4: Contribution of Sda1 loop regions to nuclease activity [/fig_ref] and Supplementary [fig_ref] Figure 6: Functionally important residues for Sda1 [/fig_ref]. Additionally, there is a slight variation in the C-terminal end of this helix. In GBS NucA and Enda, this helix is slightly bent over the length of the helix, downstream of the 'finger loop', which ultimately shifts the end of the helix by 2.7Å, as compared to the same location in Sda1. The structure of the DNA substrate binding loop differs between Sda1 and GBS NucA. For GBS NucA, this loop (referred to as the S-loop) contains a short ␣-helix. Given the similarity of loop length and sequence in this region, EndA has been hypothesized to have a homologous ␣-helix at this position [bib_ref] Structural insights into catalytic and substrate binding mechanisms of the strategic EndA..., Moon [/bib_ref] [bib_ref] Structural characterization of the virulence factor nuclease A from Streptococcus agalactiae, Moon [/bib_ref]. In Sda1, however, this loop is slightly shorter, has little sequence conservation beyond a positively charged residue near the N-terminal end of the loop (Arg124). Subsequently, there is no S-loop ␣-helix observed for Sda1. The C-termini of the GBS NucA and EndA catalytic domains occupy very similar positions, but which differ from that of the Sda1 catalytic domain [fig_ref] Figure 4: Contribution of Sda1 loop regions to nuclease activity [/fig_ref].
Sda1 contains three insertion regions on the 'front' face of the enzyme, which are unique structural features as compared to Spd1, GBS NucA or EndA [fig_ref] Figure 4: Contribution of Sda1 loop regions to nuclease activity [/fig_ref] , F, and Supplementary [fig_ref] Figure 6: Functionally important residues for Sda1 [/fig_ref]. Ins.I (Tyr72-Ser92) is an insertion between -strands 1 and 4 in Sda1, and consists of a two small -strands forming a short, antiparallel -sheet, with a short ␣-helix in the loop between them. The position of this loop could likely affect the splaying of the end of -strand 4, along the bottom of the central -sheet. Ins.II (Tyr134-Ser164), an extrusion between the C-terminal end of the substrate binding loop and the beginning of -strand 8, forms an intriguing 'pseudoknot' type of structure. strands 6 and 7 form a two-stranded antiparallel -sheet, the loop between the two strands contains the short ␣-helix C and several random coil residues that flatten back over the short -sheet. This region displays no structural similarity to any known protein, as surveyed by the Dali Server [bib_ref] Dali server: conservation mapping in 3D, Holm [/bib_ref]. In GBS NucA and EndA, the substrate binding loop flows smoothly to the following -strand [fig_ref] Figure 4: Contribution of Sda1 loop regions to nuclease activity [/fig_ref]. Finally, Ins.III (Ser170-Arg179), which connects -strands 8 and 9 in Sda1, is up to seven residues longer than in GBS NucA or EndA. The confluence of Insertions II and III in Sda1 drastically alters the landscape of that particular region, as compared to Spd1, GBS NucA or EndA.
## Contributions of sda1 loop regions to nuclease activity
In order to determine if any of the insertions in Sda1 contributes to its nuclease activity, each insertion was deleted ( Ins.I: Tyr72-Val91, connected with glycine;
Ins.II: Tyr134-Ser164; Ins.III: Ile171-Arg179, connected with glycine) and subsequently tested in the plasmid conversion assay, using the imidazole chemical-rescue strategy. The 'finger loop' extrusion from ␣-helix E was also deleted ( Loop: Asn214-Gly219), similarly to what had previously been done in EndA [bib_ref] Structural insights into catalytic and substrate binding mechanisms of the strategic EndA..., Moon [/bib_ref] , such that the helical structure was not disrupted. None of the deletions had any apparent effect on protein folding, since each deletion variant expressed solubly and behaved as a uniform population, well-resolved from the void volume peak during size exclusion chromatography (Supplementary . In the plasmid conversion assay, deletion of Ins.I had a limited effect on nuclease activity, while the Ins.III had a more pronounced effect [fig_ref] Figure 4: Contribution of Sda1 loop regions to nuclease activity [/fig_ref]. Surprisingly, the 'finger loop' deletion variant had nearly undetectable levels of nuclease activity, and deletion of Ins.II yielded catalytically inactive protein. These results were unexpected, given that Ins.II lies a considerable distance (∼18Å) from the catalytic centre and that a similar 'finger loop' deletion in EndA had no effect on either DNA substrate binding or catalysis [bib_ref] Structural insights into catalytic and substrate binding mechanisms of the strategic EndA..., Moon [/bib_ref].
## Surveying sda1 surface residues for involvement in catalysis
Analysis of DNA substrate binding by Sda1 was attempted, by co-crystallization, soaking of apoprotein crystals, electrophoretic mobility shift assays and fluorescence polarization anisotropy, all of which were unsuccessful. It is likely that the lack of sequence specificity and low binding affinity both contribute to difficulties in obtaining detailed information pertaining to DNA substrate interactions. Therefore, the crystal structure of the nuclease VVN from Vibrio vulnificus was used as a model for DNA binding to Sda1. VVN is currently the only ␣ motif-containing nuclease with a DNA substrate bound in a crystal structure [bib_ref] DNA binding and cleavage by the periplasmic nuclease Vvn: a novel structure..., Li [/bib_ref] [bib_ref] Structural basis for sequence-dependent DNA cleavage by nonspecific endonucleases, Wang [/bib_ref]. Though the amino acid sequence and core protein architecture display no similarity to that of Sda1, it is possible to superimpose the two structures, based solely on the ␣ motif [fig_ref] Figure 5: Substitution mutagenesis survey of surface residue involvement in Sda1 catalysis [/fig_ref]. Using this method, the asparagine residues (Asn211 in Sda1 and Asn127 in VVN) that chelate the hydrated divalent metal, and the metals themselves, align well. Therefore, it is likely that a duplex DNA substrate would bind the ␣ motif of Sda1 in a similar orientation [fig_ref] Figure 5: Substitution mutagenesis survey of surface residue involvement in Sda1 catalysis [/fig_ref].
Using this model, it was possible to perform a structural survey for residues that may contribute to DNA binding and/or catalysis [fig_ref] Figure 5: Substitution mutagenesis survey of surface residue involvement in Sda1 catalysis [/fig_ref]. These residues were mutated using site-directed mutagenesis, on the background of the H188G mutant. Therefore, the effects of the surface alterations could be assayed using the imidazole chemical rescue strategy [fig_ref] Figure 5: Substitution mutagenesis survey of surface residue involvement in Sda1 catalysis [/fig_ref] and E). The K126A, K130A, K168A, R217A and K218A mutants were nearly indistinguishable from the H188G mutant alone, while R179A, R186K and Q222A displayed slight, but statistically significant decreases in activity. The S187G, R208A, Q210A and R124A mutants showed a more severe effect on catalysis, and the R124A/K126A double mutant exhibited significantly lower activity than either of the two individual mutants combined. The N211A mutant, which would negate binding of the divalent metal in the active site, produced a catalytically inactive enzyme. Though not entirely inactive, R186A had barely detectable activity levels.
Nucleic Acids
# Discussion
Previously published reports have established that Sda1 serves as a virulence factor for invasive strains of S. pyrogenes, degrading NETs, and allowing bacterial dissemination throughout the host organism [bib_ref] DNase expression allows the pathogen group A Streptococcus to escape killing in..., Buchanan [/bib_ref]. In this study, we have attempted to elucidate the mechanism of nuclease activity by Sda1, through structural and biochemical methods. We present two X-ray crystal structures of an inactivated Sda1 variant (general base, His188, replaced with glycine), which yields insight into the putative catalytic mechanism. These structures also provide a limited structural portrait of the C-terminal domain [fig_ref] Figure 3: Structure of EGFPX-Sda1 [/fig_ref] and B) and underscore the importance of an N-terminal -strand for stabilization of the central -sheet. Detailed analyses of these structures, compared with those of related nucleases, were the foundation of structure-driven mutagenesis experiments that broaden our understanding of nuclease function.
Sda1 is a ␣-metal finger nuclease that contains a DRGH-like catalytic motif similar to those of EndA and GBS NucA, and prophage-encoded Spd1 from other Streptococcus species [bib_ref] Sustained active site rigidity during synthesis by human DNA polymerase mu, Moon [/bib_ref] [bib_ref] The structural characterization of a prophage-encoded extracellular DNase from Streptococcus pyogenes, Korczynska [/bib_ref] [bib_ref] Structural insights into catalytic and substrate binding mechanisms of the strategic EndA..., Moon [/bib_ref]. However, in Sda1, this motif has evolved as DRSH, with Ser187 replacing the glycine residue. Despite this alteration in the ␣ motif, Sda1 is expected to utilize a similar catalytic mechanism as has been described for other structurally related DRGH nucleases [bib_ref] Structural insights into catalytic and substrate binding mechanisms of the strategic EndA..., Moon [/bib_ref] [bib_ref] Structural insights into the mechanism of nuclease A, a betabeta alpha metal..., Ghosh [/bib_ref]. Substitution of the general base, His188, allows for soluble overexpression of the nuclease in an E. coli expression system and provides an effective means of rescuing nuclease activity by addition of exogenous imidazole [bib_ref] Structural insights into catalytic and substrate binding mechanisms of the strategic EndA..., Moon [/bib_ref] [bib_ref] Chemical rescue of active site mutants of S. pneumoniae surface endonuclease EndA..., Midon [/bib_ref]. This imidazole chemical rescue strategy identified a few residues that contribute to catalysis by Sda1. As expected, alanine substitution of Asn211, the only residue directly chelating the active site metal, produced a catalytically inactive enzyme. Similarly, mutations along -strand 10 (S187G and R186A) had severely deleterious effects on catalysis. A conservative mutation of Arg186 to lysine maintained the positive charge in this area and had only a minimal effect on nuclease activity.
On the putative substrate binding loop, mutation of two positively charged residues, Arg124 and Lys126, yielded interesting results. When mutated individually, K126A had only a minimal effect, while R124A severely diminished activity. Mutating both residues simultaneously produced a synergistic decrease in activity that surpassed R124A alone. Since deletion of the 'finger loop' drastically decreased catalysis by Sda1, residues Arg217 and Lys218 were also mutated to determine the extent of their individual contribution to the loss of activity. However, K218A was nearly indistinguishable from the H188G mutant alone, and R217A displayed only a slight decrease in catalysis. Arg179 on Ins.III was mutated in a similar fashion and did produce a decrease in plasmid conversion. However, mutation of this residue alone cannot account for the loss of activity exhibited by the Ins.III mutant (66.4% versus 27.2% supercoiled substrate conversion for R179A and Ins.III, respectively).
Gln210 lies deeper within the active site and may serve as a structural substitute for Gln174 in GBS NucA [fig_ref] Figure 6: Functionally important residues for Sda1 [/fig_ref] and Gln186 in EndA, which is not conserved in Sda1 (Gly206). The Q210A mutation has a moderate effect Secondary structural elements of the ␣ motifs for Sda1 (-strands, green; ␣-helix, blue; coils, wheat, labelled as in [fig_ref] Figure 1: Sequence, domain structure and activity of Sda1 [/fig_ref] and GBS NucA (PDB ID code 4QH0 [bib_ref] Structural characterization of the virulence factor nuclease A from Streptococcus agalactiae, Moon [/bib_ref] , grey) are superimposed. Amino acid side chains for each protein are drawn in stick (Sda1, pink; GBS NucA, grey). Metal coordination and putative hydrogen bonding interactions are marked as solid and dashed black lines, respectively. (B) Stereo diagram of putative interactions that could exist between Sda1 surface residues (amino acids in stick, purple) and a hypothetical bound DNA molecule (modeled from VVN crystal structure PDB ID code 2IVK (34)). The DNA (brown) is drawn as a transparent cartoon. Insertions I-III are coloured as previously described (Ins.I in cyan, Ins.II in yellow, Ins.III in orange, 'finger loop' in magenta). Location of -strand 10 in the ␣ motif is marked in black. on Sda1 catalysis. Gln222 and Asn226 lie on the solventaccessible side of ␣-helix E and may be capable of interacting with the DNA. Interestingly, Q222A displayed a decrease in catalytic activity, while N226A was considerably more active than the H188G alone. This mutant converted more of the supercoiled substrate to degraded product forms (63.6% versus 41.8% for H188G alone). Attempts to influence DNA binding and catalysis by mutating positively charged residues along the DNA binding cleft met with only limited success. The K168A mutant was nearly indistinguishable from the H188G alone, and K130A displayed only a slight decrease in activity. R208A was moderately inhibited, to a similar extent of the Q210A mutant (34.7% substrate conversion for R208A and 38.9% substrate conversion for Q210A).
Taken together, the results from the mutagenesis experiments are intriguing, and highlight essential differences between Sda1 and other related streptococcal nucleases. In each of the enzymes, only a single amino acid residue directly chelates the divalent metal ion and several neighboring active site residues putatively hydrogen bond with the water molecules comprising the hydration sphere around the metal [bib_ref] Structural insights into catalytic and substrate binding mechanisms of the strategic EndA..., Moon [/bib_ref] [bib_ref] Structural characterization of the virulence factor nuclease A from Streptococcus agalactiae, Moon [/bib_ref]. The same is true for Sda1, where Asn211 chelates the metal, while Ser187, Glu225 and some backbone carbonyls coordinate the water molecules [fig_ref] Figure 6: Functionally important residues for Sda1 [/fig_ref]. These residues comprise the 'first shell' interactions around the hydrated metal cluster and represent the catalytic cen-tre. In EndA and GBS NucA, a 'second shell' of residues creates a complex hydrogen bonding network that appears to stabilize the first shell. In Sda1, however, many of the residues involved in this network are not conserved, and different hydrogen bonding interactions appear to stabilize the first shell residues. For example, in GBS NucA, His142 [fig_ref] Figure 6: Functionally important residues for Sda1 [/fig_ref] and Supplementary [fig_ref] Figure 6: Functionally important residues for Sda1 [/fig_ref] putatively hydrogen bonds with the side chain of Asp145, which then hydrogen bonds with Asn179, the chelating amino acid. However, in Sda1, the structural equivalent of the histidine is Tyr182, which is rotated away from the active site. Asp185 is conserved, but now mediates stabilizing interactions with Arg208 and Asn211. Gln180 on the 'finger loop' of GBS NucA was found to play a role in second shell interactions, however, this residue is not conserved in Sda1, and is replaced with valine (Val212). One of the most interesting differences between these enzymes is that the structural equivalent of Gln174 in GBS NucA is Gly206 in Sda1. Gln174 in GBS NucA and Gln186 in EndA coordinate one of the water molecules hydrating the metal ion, in the first shell. In Sda1, the side chain of Ser187 coordinates the equivalent water, and Gln210 may serve as a structural surrogate for Gln174, indirectly stabilizing the position of Ser187 [fig_ref] Figure 6: Functionally important residues for Sda1 [/fig_ref]. As might be expected, mutation of first shell residues (N211A and S187G) appears to have a more deleterious effect than alterations of second shell residues (Q210A and R208A) [fig_ref] Figure 5: Substitution mutagenesis survey of surface residue involvement in Sda1 catalysis [/fig_ref] and E).
Though DNA substrate binding to Sda1 could not be directly observed, it is possible to glean some information from the mutagenesis studies. For example, modeling the DNA substrate from VVN onto the Sda1 structure reveals a potential clash between protein and DNA, at the confluence of the 'finger loop' and Ins.III [fig_ref] Figure 5: Substitution mutagenesis survey of surface residue involvement in Sda1 catalysis [/fig_ref]. It seems logical that the 'finger loop' might alter its conformation to permit binding of a DNA substrate. Given their proximity to the DNA, it was surprising that mutation of Arg217 and Lys218 on the 'finger loop' had negligible effect on nuclease activity. However, in a different loop conformation, these residues may lie farther from the DNA substrate. Lys130 and Lys168 were also more distal from the DNA binding cleft, so lack of involvement in DNA substrate binding is perhaps unsurprising.
Interpreting the roles of Gln222 and Asn226 is more difficult. Both are positioned to make contact with the DNA. Given their position, it was hypothesized that altering them to smaller side chains might create a more open cleft in which the substrate might bind. Since the N226A mutation generated a more active enzyme, such a hypothesis might be accurate. However, Q222A exhibited decreased activity. It is possible that this side chain may adopt a different rotamer in the presence of a DNA substrate, bringing it closer to the first shell residues and allowing it to stabilize the conformation of Glu225.
Arg124 on the putative substrate binding loop currently lies just outside of hydrogen bonding range of the DNA substrate and the catalytic centre. However, with a slightly different rotamer conformation, it could be precisely placed to hydrogen bond with the scissile phosphate. Such placement would be consistent with a role in transition state stabilization, as hypothesized for the structurally homologous residues in other ␣-Me nucleases (Arg127 in EndA (25), Arg116 in GBS NucA [bib_ref] Structural characterization of the virulence factor nuclease A from Streptococcus agalactiae, Moon [/bib_ref] and Arg93 from Anabaena sp. NucA [bib_ref] Structural insights into the mechanism of nuclease A, a betabeta alpha metal..., Ghosh [/bib_ref].
The structural and biochemical data presented in this study provide a foundation for better understanding the activities of secreted sequence/structure-nonspecific nucleases as virulence factors during invasive Streptococcus sp. infection. These nucleases represent novel therapeutic targets that attack the virulence factor independently of the invading bacteria. Inhibiting these enzymes would maximize bacterial destruction by the innate immune response and minimize the effect of either the infection or the resulting treatment on the host. This work reinforces the catalytic roles of many conserved residues (Arg186, Asn211 and Arg124), while revealing different effects on nuclease activity by other nonconserved residues (Arg179, Ser187, Gln222, Asn226) and loop regions, such as the 'finger loop' and Ins. regions II and III. Ongoing and future studies will focus on delineating the roles of specific surface residues in substrate specificity differences--endo-versus exonucleolytic activity on deoxyribo-versus ribonucleotide containing substrates. As such, obtaining structures of these enzymes in complex with a variety of substrates would aid in this endeavor. Such structures would significantly contribute to the growing database of information on these and related virulence factors, and could facilitate rational drug design of small molecule inhibitors--targeting DNA binding and possible conformational changes.
[fig] Figure 1: Sequence, domain structure and activity of Sda1. (A) Protein sequence of Sda1. The putative transmembrane ␣-helix is boxed in purple, with the likely cleavage site immediately upstream of Glu38 (pink). The 'Finger Loop' and Insertions (Ins.) I, II and III are boxed in magenta, cyan, yellow and orange, respectively. The putative DNA substrate binding loop is shown in a red box. Secondary structural elements are indicated by green arrows (-strands, labelled alphabetically) or blue rectangles (␣-helices, labelled numerically). The location of the general base (His188) is shown in red. [/fig]
[fig] Figure 2: Structural characterization of Sda1. C. (A-B) Cartoon diagram of Sda1. C(H188G), viewing the 'front' (A) and 'back' (B) faces of the Sda1 catalytic domain (180 • y-axis rotation, as indicated by black arrows), with key structural features coloured and labelled as in Figure 1A. (C) Diagram of interactions involving ␣ motif (-strands in green, ␣-helix in blue) and the divalent zinc ion (purple). Water molecules (red spheres) filling out the hydration sphere are shown. Direct chelation of the metal ion is depicted as solid black lines, and putative hydrogen bonding interactions as dashed black lines. Amino acid side chains are shown in yellow sticks, with interactions from the C-terminus of the neighbouring molecule in cyan. 2Fo-Fc electron density is shown in grey, contoured at 1. (D) Interactions involving the N-terminus of Sda1. C(H188G). Nterminal resides Trp46-Asn53 are shown in cyan, and -strand 1 (Asn54-Lys60) in cartoon in magenta. Putative hydrogen bonding between the backbone carbonyl of Asn53 and the backbone amide of Gly220 is marked with a dashed black line. All structural figures were created using PyMOL (http://www.pymol.org). [/fig]
[fig] Figure 3: Structure of EGFPX-Sda1. N(H188G) and comparison to Sda1. C(H188G). (A) Cartoon diagram of EGFPX-Sda1. N(H188G). The EGFP fusion carrier protein (yellow), with its fluorphore (cyan). Sda1. N(H188G) is shown in a 'side' view, looking down the plane of the central -sheet, with the 'front' and 'back' faces to the right and left, respectively. The C-terminal domain is wedged between the EGFP and the Sda1 catalytic domain (top, centre). (B) Cartoon diagram of the ordered regions of the C-terminal domain, with the ends of each fragment labelled. (C-D) Superposition of EGFPX-Sda1. N(H188G) (grey) with molecule B of Sda1. C(H188G), coloured as indicated in Figure 1A. 'Front' (C) and 'back' (D) faces of the superimposed catalytic domains are shown, with the location of -strand 1 marked (green asterisk). Relative locations of the N-termini of the different constructs are shown, with EGFPX-Sda1. N(H188G) in grey and Sda1. C in wheat. The C-terminal domain has been excluded from panels C and D for simplicity. [/fig]
[fig] Figure 4: Contribution of Sda1 loop regions to nuclease activity. (A-C) Superposition of the Sda1 catalytic domain (blue) with that of SF370.1 prophageencoded Spd1 (PDB ID code 2XGR (20), green), showing 'back' (A), 'side' (B) and 'front' (C) views, relative to the central -sheet. (D-F) Superposition of Sda1 catalytic domain with that of EndA from S. pnuemoniae (PDB ID code 3OWV (25), purple) and GBS NucA from S. agalactiae (PDB ID code 4QH0 (29), orange), showing 'back' (D) [/fig]
[fig] Figure 5: Substitution mutagenesis survey of surface residue involvement in Sda1 catalysis. (A) Structural superposition of Sda1. C(H188G) with VVN from Vibrio vulnificus (PDB ID codes 2IVK (34) and 1OUP (33)), based on the ␣ motif (Asp184-Gly188 and Leu203-Val212 from Sda1, blue; Ile76-Ala80 and Leu119-Gly128 from VVN, brown). The position of the asparagine (Asn211 for Sda1 and Asn127 for VVN) chelating the zinc (Sda1, purple) or calcium (VVN, green) is shown. (B) Putative position of DNA substrate (16 base pair duplex bound to VVN, PDB ID code 2IVK (34)) binding to Sda1 catalytic domain (light blue surface). The cleavable strand (orange) and its complement (brown) are drawn as a cartoon, with the scissile phosphate in green. The 'finger loop' (magenta), substrate binding loop (red), Insertions I, II and III (cyan, yellow and light orange, respectively) and the catalytic residues on -strand 10 (black) of the ␣ motif are shown, relative to the DNA substrate. (C) Positions of Sda1 surface residues mutated in this study are shown in stick. Residues hypothesized to play a role in catalysis are coloured black, while those with a putative role in DNA binding are coloured blue. Orientation of the molecule is as shown in panel B, but with the DNA removed. (D) Plasmid conversion nuclease activity assay for Sda1 and Sda1 mutants. (E) Conversion of the supercoiled plasmid substrate from D to either nicked open circle (blue bar), linearized (red bar) or degraded forms (green bar).Each assay was performed in triplicate, and the error bars for each population represent the standard deviation. [/fig]
[fig] Figure 6: Functionally important residues for Sda1. (A) Stereo diagram of the hydrogen bonding network surrounding the hydrated metal cluster (zinc, purple sphere; water molecules, red spheres) in Sda1. C(H188G). [/fig]
[table] Table 1: Data collection and refinement statistics EGFPX-Sda1. N(H188G) Sda1. C (H188G) Values in parentheses refer to the highest resolution shell. [/table]
|
Predicting Glycemic Index and Glycemic Load from Macronutrients to Accelerate Development of Foods and Beverages with Lower Glucose Responses
Low glycemic index (GI) and/or low glycemic load (GL) are associated with decreased risks of type-2 diabetes and cardiovascular disease. It is therefore relevant to consider GI and GL in the early phases of the development of packaged foods and beverages. This paper proposes a model that predicts GI and GL from macronutrient composition, by quantifying both the impact of glycemic carbohydrates and the GI-lowering effects of nutrients such as proteins, fats and fibers. The precision of the model is illustrated using data on 42 breakfast cereals. The predictions of GI (r = 0.90, median residual = 2.0) and GL (r = 0.96, median residual = 0.40 g) compete well with the precision of the underlying in-vivo data (Standard Error SE = 3.5 for GI). This model can guide product development towards lowering GI and GL, before final confirmation by in vivo testing.
# Introduction
Carbohydrates should account for 50% of total energy intake in a normal diet, but free sugars should account for less than 10%, as recommended by the World Health Organization, to prevent both obesity and dental caries. In order to achieve this public health target, consumers will need to make significant changes to their diets and providers of packaged food and beverages will need to replace free sugars with alternatives that are truly supportive of health benefits. Sugars have many functionalities that go beyond providing sweetness. Therefore, complete removal or partial reduction of sugar is not simple and various routes for nutritious sugar replacements have been proposed. The first approach consists of carefully selecting alternative glycemic carbohydrates in order to avoid replacing low glycemic sugars such as lactose by high glycemic carbohydrates such as maltodextrin. The second approach consists in replacing sugars and rapidly digestible glycemic carbohydrates by slowly digestible carbohydrates, non-glycemic carbohydrates such as dietary fibers, and particularly β-glucans. Further considering non-glycemic nutrients such as proteins and fats is important since they further modulate postprandial glycemic response.
Foods and beverages with low glycemic index (GI) and low glycemic load (GL) are considered beneficial, based on evidence coming from meta-analyses of both observational studies, as well as from randomized clinical trials. Benefits include the management of diabetes, the prevention of diseases such as type-2 diabetes, and cardiovascular disease, including coronary heart disease.
To estimate the GI/GL of a meal or a diet, one can use the model of the Food and Agriculture Organization (FAO). This model predicts the GI of a diet as a simple function of its constituting products by using average GI and GL values of hundreds of common foods and beverages available in generic tables. This simple model is unfortunately not always reliable; for example, it failed to predict GI of composite breakfast meals. The FAO-model has two inherent limitations. First, the GIs of the constituting products are averages coming from tables that do not necessarily reflect the specificities of the actual products (e.g., depending on the local sources of ingredients and processes, the reported GIs of white bread made from wheat flour varies between 59 and 89). Second, the model does not properly take into account the GI-lowering effect of the non-glycemic components in the meal (e.g., the butter on the bread).
This work presents a new model that builds on the logic of the FAO-model, but that overcomes its two inherent limitations. First, in order to overcome the issue of product specificities, and in order to be useful for both complete meals and single products, it models GI as a function of nutrients, not as a function of products (with GI = 100 for pure glucose). Second, it both deterministically quantifies the effects of glycemic carbohydrates and empirically estimates the GI-lowering effects of other macronutrients, such as proteins, fats and fibers.
Published data, including data on honey, pasta, bread and milk, allowed us to set up the shape of the model combining deterministic and empirical aspects. Sixty in-house in-vivo trials allowed the estimation of empirical coefficients of the model. These 60 trials covered various product categories (i.e., infant cereals, cereal bars, biscuits and dairy beverages) and a wide range of recipes, leading to GI spreading between 15 and 95. Consequently, the correlation between prediction and observation was very large (r = 0.97, p-value < 0.01, n = 60). In order to challenge the model, it has been applied on a single product category that was not part of the model setup. The application presented includes correlation plots for 42 breakfast cereals, as well as the corresponding Bland-Altman difference plots.
# Materials and methods
## Products
Forty-two breakfast cereals made from various grains (i.e., wheat, oat, mixtures of grains), with or without inclusions (e.g., chocolate, fruits, nuts, honey), have been characterized for their macronutrient composition using standard analytical methods. The type of process and composition of these 42 starch-based products is presented . These products contained glycemic carbohydrates accounting for 57 to 82 g/100 g and the proportion of Rapidly Digestible Starch ranged between 65% for granola and 88% for extruded products.
## In-vivo testing
The in-vivo testing to measure GI was performed according to international standards. Fourteen studies, all approved by the Human Research Ethics Committee of the University of Sydney, tested the 42 products with 2 to 5 products per study protocol. The constant test protocol enabled the combination of data from all 14 studies to perform secondary analyses. Serving sizes were calculated to deliver 50 g glycemic carbohydrates. Test products were compared with a reference of 50 g glucose, which was assessed in triplicate in each study. Within each study, 10 healthy subjects consumed all test products and the reference in a crossover design, with one sample per day, under fasting conditions, with at least one-day washout between two test days. Subjects tested the dry products along with a glass of 250 mL water. Two fasting blood samples (t = −5 min and t = 0 min) were obtained, and after the second fasting blood sample a test product was consumed and postprandial glycaemia was monitored for 120 min, 60, 90, 120 min). A total of 140 subjects (61 women, 79 men) were enrolled in these 14 studies with age ranging between 19.8 and 52.3 years (mean = 29.6, SD = 8.6) and BMI ranging between 18.2 and 24.9 kg/m 2 (mean = 22.0, SD = 2.2). GI and GL (Equations (1a) and (1b)) were determined using commonly accepted equations:
[formula] GI = Incremental [/formula]
## Fao predictive model
The FAO-model predicts GI of a meal as a function of its N constituting foods. If c i is the amount (in grams) of glycemic carbohydrates of the i th food (i = 1 . . . N) and GI i its GI, the model can be written in a synthetic form (Equation (2)):
[formula] GI Meal = N i=1 c i GI i N i=1 c i (2) [/formula]
As an example, using published data, consider a simple test meal composed of only two foods, namely 110 g white bread (c 1 = 50 g glycemic carbohydrates, GI 1 = 88) and 30 g butter (c 2 = 0g, GI 2 = 0). The equation predicts GI Meal = (50 × 88 + 0 × 0)/(50 + 0) = 88, whereas the reported in-vivo data suggest that the addition of butter reduces GI of the white bread from 88 to 67. This example illustrates the inability of the current model to account for the GI-lowering effect of fat (and similarly for proteins or fibers).
## Development of a new model to predict gi
The main idea of the new model is to apply the logic of the FAO-model on macronutrients instead of foods. The model development includes three steps (Equations (3a)-(3c)).
In a first step, the model considers products composed exclusively of water and of N mono/ disaccharides. Let x i be the relative amount (%) of the i th mono/disaccharide (i = 1 . . . N) and GI i its tabulated GI, given as an average of GI-values reported by the University of Sydney. Sugar alcohols that are partially glycemic such as maltitol (GI = 35) or xylitol (GI = 12) should be considered as mono/disaccharides in this context. With these definitions, the synthetic form of the FAO-model remains unchanged (Equation (3a)), while accurately predicting the GI of any mono/disaccharide mix. As a fictive example, a syrup composed of 90 g/100 g water and a mix of 3 g glucose (GI = 100), 2 g fructose (GI = 20) and 5 g maltitol (GI = 35) would yield GI = (3 × 100 + 2 × 20 + 5 × 35)/(3 + 2 + 5) = 51. In-vivo data of such mixes are hardly available, but various published data on products such as honey tend to confirm the accuracy of the model.
[formula] GI = N i=1 x i GI i N i=1 x i (3a) [/formula]
In a second step, the model extends mono/disaccharides to glycemic carbohydrates, including glycemic polysaccharides. Equation (3a) still properly handles polysaccharides such as maltotriose, maltotetraose or more generally maltodextrin of any dextrose equivalent, but not starch. GI of starch varies due to botanical diversity and the fact that processing can affect the availability of glucose as demonstrated by the impact on GI of various ways of cooking pasta, or of cooling cooked rice. Equation (3b) accounts for this change in starch availability by introducing a correcting factor a i to characterize the availability of the glycemic carbohydrates.
[formula] GI = N i=1 x i a i GI i N i=1 x i (3b) [/formula]
As shown in, a i = 1 for all glycemic carbohydrates except for starch for which it is best estimated by the proportion (0 ≤ a i ≤ 1) of Rapidly Digestible Starch.
In the third step, the model quantifies the impact on GI of all non-glycemic nutrients. Any product can be considered as composed of N nutrients (N = m + n) with m glycemic carbohydrates and n other nutrients, with x k the relative amount (%) of the k th nutrient (k = 1 . . . N) and x k = 100%. Equation (3c) introduces the non-glycemic nutrients as diluting factors of GI, along with coefficients b j (j = 1 . . . n) that characterizes their GI-lowering power.
[formula] GI = m i=1 x i a i GI i m i=1 x i + n j=1 x j b j (3c) [/formula]
Non-glycemic nutrients without GI-lowering power (e.g., water) have b j = 0, whereas b j of other nutrients were a-priori unknown and were estimated to be highest for β-glucans, fats and proteins (0.6), followed by other soluble fibers (0.3) insoluble fibers and ashes . These empirical coefficients were estimated by Ordinary Least Square (OLS) regression applied on a heterogeneous dataset of 60 in-vivo studies, not including the 42 breakfast cereals (data not shown). . GI-lowering power (b i ) of common macronutrient as defined by empirical data fitting for starch-based products.
## Gi-lowering macronutrients b i
Carbohydrates As an example, published data point to a GI of 66 for a test product composed of 50% glucose and 50% proteins. This is close to the GI as predicted by Equation (3c) with GI = 50% × 100/(50% + 50% × 0.6) = 63. Other published data report GI = 27 for 100 g of whole milk composed of 88 g water, 4.9 g lactose, 3.3 g fat, 3.1 g protein and 0.7 g minerals. In comparison, Equation (3c) yields a predicted GI of 26 with GI = 4.9 × 47/(4.9 + 88 × 0 + 3.3 × 0.6 + 3.1 × 0.6 + 0.7 × 0.0) = 26. Finally, coming back to the example of the 30 g of butter on 110 g white bread, Equation (3c) predicts a decrease from GI = 88 for plain bread to GI = 65 for bread + butter (vs. GI = 67 in-vivo). Similar decreases have been observed when replacing dairy butter with peanut butter.
Predicted glycemic load (GL) is derived from predicted GI using the standard GL Equation (1b).
# Results and discussion
## Precision of predictions for 42 breakfast cereals
In-vivo GI of 42 breakfast cereals ranged between 50 and 83 (mean = 68, SD = 9.2). The granola and muesli products tested had the lowest GI (50-60), before bars (66-72), whereas flakes and extruded products cover a wider range depending on their composition (61-83). The standard error (SE) of in-vivo data is SE = 3.5. Predicted GI of the same breakfast cereals ranged between 52 and 82 (mean = 68, SD = 8.2). When visualizing observed vs. predicted GI, the correlation is r = 0.90 (p-value < 0.01) and the precision of the prediction given by the median absolute residual (=2.0) is smaller than the in-vivo SE.
Nutrients 2019, 11, x FOR PEER REVIEW 6 of 10 1CD). These plots show that the predictions were not biased and that no simple relationships between the differences and the mean value can be identified. These results illustrate the strength of the new model that combines deterministic modelling to quantify the impact of glycemic nutrients and empirical modelling to quantify the impact of GIlowering nutrients. This hybrid model delivers more accurate predictions than pure empirical models such as the one described by Meynier, who used advanced feature selection techniques to select main effects and interactions of four product characteristics (slowly digested starch, fat, fiber, and rapidly digested starch) to predict the GI of cereal products (r = 0.73, p-value < 0.01).
The proposed model delivers accurate GI and GL prediction for products with high proportion of glycemic nutrients such as the 42 tested breakfast cereals. This is because it first captures the effect of glycemic nutrients in a very simple deterministic way by modelling the GI of a mix of glycemic nutrients as the weighted average of their GIs and second accounts for the GI-lowering effect of other nutrients. Glycemic nutrients include mono/disaccharides and glycemic polysaccharides as well as nutrients such as maltitol and xylitol, which are partially glycemic, and therefore increase glycemic response. Non-glycemic nutrients include fat, fibres, proteins, ashes and water. The provided list of With a standard serving size of 30 grams of product, the in-vivo GL of these 42 breakfast cereals ranged between 9.4 and 20.4 grams (g) of glucose equivalent (mean = 14.7 g, SD = 3.1 g) and their predicted GL range between 9.4 g and 20.2 g (mean = 14.7 g, SD = 2.9 g). When visualizing observed vs. predicted GL, the correlation is r = 0.96 (p-value < 0.01) and the precision of the prediction was high since the median absolute residual was as low as 0.40 g.
Three products featuring specific inclusions (i.e., cashew nuts, almonds, quinoa or cocoa) yielded predictions differing from actual in-vivo data by 9-11 points for GI and by 2.0-2.3 g for GL. These three products are also outside the mean ± 2SD range in the Bland-Altman difference plots.
These plots show that the predictions were not biased and that no simple relationships between the differences and the mean value can be identified.
These results illustrate the strength of the new model that combines deterministic modelling to quantify the impact of glycemic nutrients and empirical modelling to quantify the impact of GI-lowering nutrients. This hybrid model delivers more accurate predictions than pure empirical models such as the one described by Meynier, who used advanced feature selection techniques to select main effects and interactions of four product characteristics (slowly digested starch, fat, fiber, and rapidly digested starch) to predict the GI of cereal products (r = 0.73, p-value < 0.01).
The proposed model delivers accurate GI and GL prediction for products with high proportion of glycemic nutrients such as the 42 tested breakfast cereals. This is because it first captures the effect of glycemic nutrients in a very simple deterministic way by modelling the GI of a mix of glycemic nutrients as the weighted average of their GIs and second accounts for the GI-lowering effect of other nutrients. Glycemic nutrients include mono/disaccharides and glycemic polysaccharides as well as nutrients such as maltitol and xylitol, which are partially glycemic, and therefore increase glycemic response. Non-glycemic nutrients include fat, fibres, proteins, ashes and water. The provided list of nutrients is not exhaustive and can be adapted to any product category, including novel glycemic or non-glycemic nutrients if needed.
## Limitations of the model
Whilst the deliberate simplicity of the model facilitates user experience, it does limit its accuracy. As a case in point, among the 42 breakfast cereals tested, the predicted GL was too high for two products with inclusions of cashew nuts, almonds and quinoa and too low for a product rich in cocoa. These products reveal that the model is not fully capturing the specificities of components that are rich in GI-modulating nutrients such as particular fat types or polyphenols.
More generally, a limitation of the model is that starch digestibility is captured through a single global marker (i.e., coefficient a i ) regardless of the underlying mechanisms. This global marker is well estimated by the fraction of rapidly digestible starch, as suggested by the empirical data used to build the model. But this simple approach is eventually not properly addressing mechanisms such as the modulation of enzyme availability induced by specific ingredients (e.g., quinoa, lentils or chickpeas), the impact of ingredient processing on particle sizes within the same grain (e.g., cracked wheat vs. wheat flour), the impact of process on products comparable in composition (e.g., bread vs. pasta), and the complex interactions occurring among food constituents, especially when the physical format of the product varies (e.g., solid vs. liquid). Such mechanisms might be captured more adequately by alternative in-vitro methodologies that have shown to be highly correlated with in-vivo GI; but such methodologies rely on long and tedious procedures that conflict with the simplicity of the proposed model.
Another potential limitation of the model is that it captures the GI-lowering effect of non-glycemic nutrients in the same way regardless of the underlying mechanism. This simplifies the tool, but limits the accuracy for estimating the GI-lowering power of various nutrients. For example, proteins-similarly to fibers and fats-might increase viscosity and therefore physically protect glycemic nutrients from enzymatic digestion; but in addition, some amino acids are known to trigger the insulin response and therefore to lower the glycemic response. Consequently, the b-coefficient of proteins in the model might strongly vary according to their amino acid profiles leading to two or more protein categories (e.g., low and high insulinogenic effect).
## Potential use of model beyond breakfast cereals
The estimated b-coefficients should be adapted for each product category based on empirical data. As an example, the model was initially developed with total fibers instead of a split into three fiber types (insoluble, soluble and β-glucans). It worked well on tested products that featured fibers of similar type (i.e., constant ratio of soluble to insoluble, no β-glucans), however, the method became inaccurate when comparing products containing a variety of fibers. Splitting total fibers into three nutrients indirectly integrates the modelling of viscosity. This logic could be extended to further splitting β-glucans (and including other fibers such as pectin, glucomannans or acacia gum) according to their effect on viscosity, or to their molecular weight. Similarly, the model could be extended to split fats into low and high viscous fats, or according to their impact on gastric emptying.
The proposed model is scalable and should theoretically become more precise with more detailed nutrient breakdowns. However, the more detailed the input composition, the more difficult it becomes to access this information in a precise manner. Consequently, there is a trade-off for each product category and each application of the model. As an example, the model used back of pack labelling to benchmark prepared composite meals, such as pizzas, enchiladas or burritos. The GI predictions were relevant since structural and microstructural transformations happening during processing, as well as resulting starch availability were well mastered at factory level and the reheating by consumers was expected to have minimal additional impact. In this sense, the model might be more relevant for finished packaged foods than for semi-finished foods (e.g., pizza dough) or homemade meals for which cooking procedures can have a strong impact. The benchmark of prepared composite meals helped to prioritize brands to be improved. Once these products are identified, product developers could systematically vary their processes and ingredient parameters in order to identify settings that minimize estimated GI and GL, whist achieving the ideal sensory profile.
# Conclusions
This work presents a model that provides precise analytical predictions of GI and GL in the case of breakfast cereals. It quantifies both the impact of glycemic nutrients and the GI-lowering effect of other macronutrients. Limitations of the model and potential usage of the model for other product categories are discussed. These analytical predictions are more precise for GL than for GI, which is interesting because GL is a proxy of the physiological glycemic response, taking into account not only the characteristics of macronutrients but also their quantity. The model is therefore particularly useful for products for which the mode of preparation can vary. Indeed, breakfast cereals can be consumed dry or reconstituted with milk (whole, semi-skimmed or skimmed), with or without further additions (e.g., added sucrose). As an example, when adding 125 mL of whole milk on top of the 30 g of cereals, the GL of the 42 breakfast cereals tested was predicted to increase between 0.0 to 1.2 g (mean = 0.56 g, SD = 0.32). The resulting GL is below the simple addition of the two GLs of cereals and whole milk (i.e., +1.7 g for 125 mL), simply because milk protein and milk fat act as GI-lowering agents for the glycemic carbohydrates of the cereals. In such cases, the model allows estimation of GI and GL for any mode of preparation. The model can therefore help the food industry to accelerate the development of breakfast cereals (and potentially other foods and beverages) with lower glucose responses, while taking into account personalized usages and modes of preparation. Prototypes with the highest predicted potential should then be tested in-vivo to confirm the physiological accuracy of these statistical predictions. |
Many important natural products are produced by multidomain nonribosomal peptide synthetases (NRPSs) 1-4 . During synthesis, intermediates are covalently bound to integrated carrier domains and transported to neighboring catalytic domains in an assembly line fashion 5 . Understanding the structural basis for catalysis with NRPSs will facilitate bioengineering to create novel products. Here we describe the structures of two different holo-NRPSs modules, each revealing a distinct step in the catalytic cycle. One structure depicts the carrier domain cofactor bound to the peptide bond-forming condensation domain, whereas a second structure captures the installation of the amino acid onto the cofactor within the adenylation domain. These structures demonstrate that a conformational change within the adenylation domain guides transfer of intermediates between domains. Furthermore, one structure shows that the condensation and adenylation domains simultaneously adopt their catalytic conformations, increasing the overall efficiency in a revised structural cycle. These structures and single-particle electron microscopy analysis demonstrate a highly dynamic domain architecture and provide the foundation for understanding the structural mechanisms that could enable engineering novel NRPSs.
phosphopantetheine cofactor [bib_ref] The ubiquitous carrier protein--a window to metabolite biosynthesis, Mercer [/bib_ref] , an adenylation domain that loads the amino acid substrate onto the PCP cofactor, and a condensation domain that catalyzes peptide bond formation. NRPSs then use a C-terminal thioesterase or reductase domain to catalyze product release. Structures of individual domains 1 provide insight into the NRPS structural mechanism. Interestingly, the adenylation domains have been shown to adopt two catalytic conformations [bib_ref] Conformational dynamics in the acyl-CoA synthetases, adenylation domains of nonribosomal peptide synthetases,..., Gulick [/bib_ref]. First the adenylate-forming conformation activates the amino acid substrate using ATP to form an aminoacyl adenylate and pyrophosphate. A C-terminal subdomain then rotates by ~140° to form the thioester-forming conformation that is used to install the amino acid onto the PCP 7 . These two functional states have been observed in structures of the phenylalanine activating adenylation domain of gramicidin synthetase 8 and the complexes between adenylation and carrier proteins obtained with mechanism-based inhibitors [bib_ref] Structure of PA1221, a nonribosomal peptide synthetase containing adenylation and peptidyl carrier..., Mitchell [/bib_ref] [bib_ref] Structural and functional investigation of the intermolecular interaction between NRPS adenylation and..., Sundlov [/bib_ref]. Once loaded, both the pantetheine and loaded substrate have been shown to interact transiently with the core of the carrier protein [bib_ref] Solution Structure of a Nonribosomal Peptide Synthetase Carrier Protein Loaded with Its..., Goodrich [/bib_ref] [bib_ref] Structure and Substrate Sequestration in the Pyoluteorin Type II Peptidyl Carrier Protein..., Jaremko [/bib_ref]. The structure of SrfA-C, the terminal module from surfactin biosynthesis, contains a condensation-adenylation-PCPthioesterase architecture and is to date the only structure of an intact NRPS module [bib_ref] Crystal Structure of the Termination Module of a Nonribosomal Peptide Synthetase, Tanovic [/bib_ref]. The condensation and adenylation domains share an extensive interface and were proposed to form the core of the module [bib_ref] Crystal Structure of the Termination Module of a Nonribosomal Peptide Synthetase, Tanovic [/bib_ref]. Lacking the pantetheine modification, this apo-structure shows the PCP domain directed towards the condensation domain. The other active sites are 40-60 Å from the pantetheinylation site, indicating that extensive domain rearrangements are required to complete the NRPS catalytic cycle. Movement of the PCP domain, potentially coupled to the adenylation C-terminal subdomain rotation 7 , is necessary for delivery of the peptide intermediates to the different catalytic domains.
We determined structures of two NRPSs with the same architecture as SrfA-C (Extended Data [fig_ref] Figure 1: Ribbon diagrams of complete NRPS modules [/fig_ref] , but with holo-proteins that show functional interactions between the PCP and catalytic domains [fig_ref] Figure 1: Ribbon diagrams of complete NRPS modules [/fig_ref]. First we present two structures of AB3403 from the human pathogen Acinetobacter baumannii (protein annotation ABBFA_003403 in strain AB307-0294) that belongs to an uncharacterized biosynthetic pathway implicated in motility [bib_ref] Genetic analysis of surface motility in Acinetobacter baumannii, Clemmer [/bib_ref] , and biofilm [bib_ref] Whole Transcriptome Analysis of Acinetobacter baumannii Assessed by RNA-Sequencing Reveals Different mRNA..., Rumbo-Feal [/bib_ref] and pellicle [bib_ref] Identification of genes essential for pellicle formation in Acinetobacter baumannii, Giles [/bib_ref] formation. We describe the structures of holo-AB3403 obtained without ligands and also upon crystallization in the presence of Mg- ATP and glycine, which among the proteinogenic amino acids serves as the best substrate (Extended Data [fig_ref] Figure 2: NRPS domain structuresThe condensation domain of AB3403 [/fig_ref]. Second, we present the structure of EntF from Escherichia coli, showing the PCP cofactor covalently trapped with a mechanism-based inhibitor to model thioester formation within the adenylation domain. These results provide views of two distinct steps in the NRPS catalytic cycle and demonstrate how the domain rotation within the adenylation domain mediates the delivery of the PCP between the two catalytic domains.
The structures of AB3403 were determined at 2.7 and 2.9Å resolution (Extended Data . No prior structure exists of an NRPS condensation domain bound to a ligand; the holo-AB3403 protein shows the pantetheine cofactor residing in the active site [fig_ref] Figure 2: NRPS domain structuresThe condensation domain of AB3403 [/fig_ref] and Extended Data [fig_ref] Figure 3: Conformational Dynamics in NRPS modules [/fig_ref]. The two lobes of the condensation domain adopt the closed orientation seen recently in the CDA synthetase condensation domain [bib_ref] Crystal Structures of the First Condensation Domain of CDA Synthetase Suggest Conformational..., Bloudoff [/bib_ref]. Contacts are made between the pantetheine and the helix running from Glu20-Leu30, in particular Tyr26 and Ile27, which forms one wall of the tunnel through which the pantetheine approaches the active site [fig_ref] Figure 2: NRPS domain structuresThe condensation domain of AB3403 [/fig_ref]. Additionally, Tyr37 forms a hydrogen bond with the amide of the cysteamine moiety of the pantetheine cofactor. As the main chain carbonyl of Tyr37 hydrogen bonds to the main chain amide of the catalytic His145, this is a critical interaction to close the two lobes and bring the active histidine into proper position.
holo-AB3403 therefore illustrates the conformation that is adopted to properly deliver the pantetheine of the PCP to the condensation domain. The PCP is rotated ~30° relative to the orientation of the PCP domain of SrfA-C (Extended Data . The AB3403 PCP interface with the condensation domain is composed of residues from helix α2, the helix that follows the pantetheinylation site at Ser1006, and the loops that precede and follow this helix. In particular, residues Phe999 through Tyr1032 face the condensation domain. Leu1007 and Val1010 form a hydrophobic interaction with Leu22 and Ile80 of the condensation domain. The side chain of Lys1011 forms a hydrogen bond with the main chain carbonyl of Gln78. Finally, Val1026, Ala1027, and Ala1030 on the PCP helix α3 form a hydrophobic interaction with Tyr26 and Leu30. Arg344 of the condensation domain, which is positioned on an insertion compared to SrfA-C, interacts with the phosphate from the cofactor.
The AB3403 adenylation domain [fig_ref] Figure 2: NRPS domain structuresThe condensation domain of AB3403 [/fig_ref] is precisely positioned in the the adenylateforming conformation, unlike the adenylation domain of SrfA-C, which is in an open conformation that may be used for substrate binding or release [bib_ref] Nonribosomal peptide synthetases: structures and dynamics, Strieker [/bib_ref]. The lysine of the conserved catalytic A10 motif 7,18 interacts with a phosphate oxygen from AMP and a carboxylate oxygen with glycine and superimposes with the homologous lysine in the gramicidin synthetase domain. In SrfA-C, the homologous lysine is ~12 Å away.
The thioesterase domain of AB3403 is structurally similar to the homologous domains of both SrfA-C and EntF (Extended Data , the latter of which has been characterized by NMR and crystallography in complex with the upstream PCP domain [bib_ref] Dynamic thiolation-thioesterase structure of a non-ribosomal peptide synthetase, Frueh [/bib_ref] [bib_ref] Structural basis for phosphopantetheinyl carrier domain interactions in the terminal module of..., Liu [/bib_ref]. Despite the similarities in domain structure, the thioesterase domain of AB3403 is in a markedly different location compared to SrfA-C [fig_ref] Figure 3: Conformational Dynamics in NRPS modules [/fig_ref]. Interestingly, in this new position the thioesterase domain cradles the back face of the PCP domain. The thioesterase domains of SrfA-C or AB3403 do not make substantial contacts with the other catalytic domains.
We next examined the delivery of the holo-PCP to the adenylation domain in a different NRPS protein. We have previously used targeted mechanism-based inhibitors, harboring a vinylsulfonamide moiety that traps the thioester-forming reaction [bib_ref] A mechanism-based aryl carrier protein/thiolation domain affinity probe, Qiao [/bib_ref] to characterize functional adenylation-PCP didomain interactions [bib_ref] Structure of PA1221, a nonribosomal peptide synthetase containing adenylation and peptidyl carrier..., Mitchell [/bib_ref] [bib_ref] Structural and functional investigation of the intermolecular interaction between NRPS adenylation and..., Sundlov [/bib_ref]. These inhibitors mimic the native aminoacyl adenylate, but contain a Michael acceptor positioned to react with the pantetheine thiol. EntF crystallized only in the presence of the serine adenosine vinylsulfonamide (Ser-AVS) inhibitor [fig_ref] Figure 2: NRPS domain structuresThe condensation domain of AB3403 [/fig_ref] and Extended Data that limits conformational flexibility to promote crystallization. Crystals of the EntF protein diffract to 2.8Å (Extended Data . No electron density was observed for the thioesterase domain although the intact protein was present in the crystal lattice (Extended Data .
The condensation domain of EntF is similar to the closed AB3403 conformation [fig_ref] Figure 2: NRPS domain structuresThe condensation domain of AB3403 [/fig_ref]. The adenylation domain adopts the catalytic thioester-forming conformation of prior adenylation-PCP proteins 9,10 , demonstrating that the conformation is compatible with a full NRPS module. The active site of the EntF adenylation domain identifies conserved residues [fig_ref] Figure 2: NRPS domain structuresThe condensation domain of AB3403 [/fig_ref] that have been shown to play important catalytic roles in other members of this enzyme superfamily 7 . Arg863 interacts with the cofactor phosphate, while Gly864 and Gln865 form one wall of the pantetheine tunnel. Interactions with the nucleotide occur between Asp840 and the ribose hydroxyls, and between Tyr746 and Tyr852 and the adenine ring. The inhibitor serine binds in the binding pocket formed by Asp648, Ser722, and Asp754 [fig_ref] Figure 2: NRPS domain structuresThe condensation domain of AB3403 [/fig_ref].
The lack of density for the thioesterase domain in EntF suggested multiple conformations in the crystal lattice. This is not surprising given the limited interactions in SrfA-C and AB3403 between the thioesterase domains and the other domains. To assess thioesterase conformational mobility, we examined EntF by negative stain electron microscopy followed by classification and averaging of single particle projections (Extended Data . The class averages revealed primarily a tri-lobed density with two neighboring globular densities of similar size attributed to the condensation and adenylation domains and a smaller lobe attributed to the thioesterase domain [fig_ref] Figure 3: Conformational Dynamics in NRPS modules [/fig_ref]. The positioning of the thioesterase domain assumes a surprisingly wide range of distances and angles relative to the other domains.
The large interface of the SrfA-C condensation and adenylation domains 13 , suggested they constitute a catalytic platform, upon which the other domains move. We therefore compared the interfaces of the three NRPS modules [fig_ref] Figure 3: Conformational Dynamics in NRPS modules [/fig_ref]. The interface in AB3403 is 1023 Å 2 , comparable in size to the 1097 Å 2 interface of SrfA-C. In contrast, the interface in EntF is only 780 Å 2 , resulting from the rotation of the adenylation C-terminal subdomain to the thioester-forming conformation.
Additionally, the conformation of the interface is not conserved between all three proteins. Alignment of the structures on the basis of the N-terminal subdomains of the adenylation domain shows that the condensation domain of both AB3403 and EntF differ slightly from each other and more significantly from SrfA-C. In AB3403 and EntF, the condensation domains are rotated by ~25° relative to the adenylation domains. Furthermore, the EntF condensation domain is shifted closer towards the adenylation domain. Structural comparisons suggest that this alternate conformation in EntF may not be compatible with the adenylate-forming conformation. The three different condensation-adenylation domain conformations, the adenylate-forming incompatibility seen in EntF, and the multiple extended and compact conformations seen in the EM data, suggest that the condensationadenylation domain platform may be more dynamic then previously proposed [bib_ref] Crystal Structure of the Termination Module of a Nonribosomal Peptide Synthetase, Tanovic [/bib_ref].
The new structures confirm the hypothesis 7 that the adenylation domain conformational change is a structural mechanism to guide the PCP between active sites in the context of complete NRPS modules. The rotation of the adenylation domain C-terminal subdomain from the adenylate-forming conformation in AB3403 to the thioester-forming conformation of EntF delivers the PCP into the adenylation domain for loading. The recent structure of loaded holo-PCP has shown the interaction of the substrate with the PCP core which may help to promote release of the substrate from the adenylation domain [bib_ref] Solution Structure of a Nonribosomal Peptide Synthetase Carrier Protein Loaded with Its..., Goodrich [/bib_ref]. This interaction also alters the surface electrostatic potential of regions that interact with the neighboring catalytic domains, including α2 and α3, and may influence the PCP delivery to neighboring catalytic domains. Finally, this transfer is further assisted by the linker region that joins the adenylation C-terminal subdomain with the PCP domain, which includes important contacts that are preserved in the adenylate-and thioester-forming conformations [bib_ref] Analysis of the linker region joining the adenylation and carrier protein domains..., Miller [/bib_ref] , as well as the open conformation of SrfA-C.
The basic NRPS catalytic cycle requires that the PCP visits three adjacent catalytic domains in a coordinated manner. The two catalytic conformations of the adenylation domain 7 require that the full cycle has four catalytic structural states . Specifically, (I) the adenylation domain catalyzes amino acid adenylation, (II) the PCP is delivered to the adenylation domain for thioester-formation to load the PCP, (III) the PCP is delivered to the condensation domain to receive the upstream peptide, and finally (IV) the peptide is delivered to a downstream condensation, thioesterase, or reductase domain for release.
Our results show that states I and III are identical and only three distinct conformations are required to accommodate the four catalytic states of the NRPS cycle , yellow). The protein first adopts an adenylate-forming conformation, seen in AB3403, state III, to catalyze amino acid adenylation. Through the domain rotation of the adenylation C-terminal subdomain, the PCP is delivered to the adenylation domain to load the pantetheine cofactor, as seen in the crystal structure of EntF, state II. Return of the PCP to the condensation domain delivers the loaded PCP for receipt of the upstream peptide, state III. Critically, as seen in AB3403, the adenylation domain can activate a second amino acid to prime the system for another cycle. The ability to simultaneously catalyze peptide bond formation and amino acid adenylation at two active sites significantly increases the overall catalytic efficiency and throughput of the NRPS module. Finally, although no structure exists of a full NRPS module with the PCP directed into the thioesterase or other downstream domain in state IV, the structure of AB3403 also offers a new view of the thioesterase domain and suggests the peptide-loaded PCP could be delivered to the downstream thioesterase domain through a simple rotation.
The modular architecture of NRPSs as well as their capacity to catalyze unusual chemistry 23,24 offers the potential for generating novel products through engineering enzyme activity and the combination of heterologous domains. These efforts have been limited by deficiencies in our understanding of the functional interactions between domains and also within active sites. The new views of two essential catalytic states in the NRPS cycle, an appreciation of the greater dynamics of NRPS systems, and the structures of holo-NRPS proteins with relevant ligands, will provide the necessary insights to guide these engineering efforts. In addition, these studies complement the recent visualization of modular polyketide synthases by cryo-electron microscopy [bib_ref] Structure of a modular polyketide synthase, Dutta [/bib_ref] to set the stage for investigations of the structural foundation of even larger, multi-modular biosynthetic proteins.
# Methods
## Expression, purification, and crystallization of ab3403
The human pathogen Acinetobacter baumannii contains an uncharacterized NRPS cluster that has been implicated in motility and biofilm formation; the product of this operon is unknown. This operon contains eight genes. In strain AB307-0294 [bib_ref] Comparative genome sequence analysis of multidrug-resistant Acinetobacter baumannii, Adams [/bib_ref] , from which the NRPS gene was cloned, this operon consists of genes ABBFA_003399 through ABBFA_003406. In the more commonly used ATCC17978 strain, the same genes are encoded by A1S_0119 through A1S_0112. The ABBFA_003403 (designated AB3403 herein) protein sequence is available at Genbank accession # ACJ56070.1.
The gene encoding AB3403 was PCR-amplified from AB307-0294 genomic DNA [bib_ref] Comparative genome sequence analysis of multidrug-resistant Acinetobacter baumannii, Adams [/bib_ref] (courtesy of Dr. Thomas A. Russo, University at Buffalo). The amplified fragment was cloned into the pET15b-TEV expression vector [bib_ref] Tobacco etch virus protease: mechanism of autolysis and rational design of stable..., Kapust [/bib_ref] and confirmed by DNA sequencing. The vector provides a His 5 -tag, linker, and tobacco etch virus (TEV) protease recognition site that, upon treatment with TEV protease, yields a final recombinant product with glycine and histidine preceding the initial methionine residue.
The AB3403 pET15b-TEV construct was transformed into E. coli (BL21-DE3) cells.
Transformed cells were grown in LB media to an OD 600 of 0.6 at 37 °C. Protein expression was induced by addition of 0.5 mM IPTG and cells were incubated overnight at 16 °C. Cells were harvested by centrifugation, flash-frozen in liquid nitrogen, and stored at −80 °C. Selenomethionine labeled protein was generated in M9 minimal media utilizing a metabolic inhibition protocol [bib_ref] Preparation of selenomethionyl proteins for phase determination, Doubliè [/bib_ref]. All purification steps were identical to the native protein.
For purification, cells were resuspended in a buffer containing 50 mM HEPES (pH 7.5), 250 mM NaCl, 10 mM imidazole, 0.2 mM TCEP. Cells were lysed by mechanical disruption (Branson Sonifier) and the resulting lysate was clarified by centrifugation at 235,000 g for 45 min. The cell lysate was passed over a His-trap (GE-Healthcare) immobilized metal ion affinity (IMAC) column and washed with lysis buffer containing 50 mM imidazole. Bound proteins were eluted with the same buffer containing 300 mM imidazole. The protein was incubated with TEV protease and dialyzed against a TEV cleavage buffer (50 mM HEPES (pH 8.0), 250 mM NaCl, 0.2 mM TCEP, and 0.5 mM EDTA) for 16 h at 4 °C. This partially purified protein was then phosphopantetheinylated by incubation with His 6 -tagged nonspecific phosphopantetheinyl transferase Sfp (10 nM), 12.5 mM MgCl 2 , and 1 mM CoA for 60 minutes at 20 °C. The clarified protein was then passed over the Histrap column a second time to remove uncleaved protein, the TEV protease, Sfp, and other contaminating proteins. The holo-AB3403 protein in the column flow-through was pooled, dialyzed against a size exclusion buffer containing 50 mM HEPES (pH 7.5), 150 mM NaCl, 0.2 mM TCEP, and further purified by gel filtration (Superdex200). Protein concentration was assessed after dialysis against a crystallization buffer (25 mM HEPES (pH 7.5), 50 mM NaCl, 0.2 mM TCEP) using an extinction coefficient at 280 nm of 157,570 M −1 cm −1 .
Crystallization conditions for holo-AB3403 were initially identified from a sparse matrix screen at 20 °C. Final crystals for native and SeMet labeled holo-AB3403 were grown at 14 °C by hanging-drop vapor diffusion against 0.75-0.95 M potassium citrate, 0.01-0.025 M glycine, and 0.05 M BisTrisPropane (BTP) (pH 8.0). Highest quality native crystals were obtained using a protein concentration of 5.5 mg/mL with a protein to cocktail ratio of 1.5:1. SeMet protein was crystallized in the same manner with a protein concentration of 7.5 mg/mL and 1:1 protein to cocktail ratio. To obtain crystals in the presence of ligands, the protein was preincubated for 45 min at 4 °C with 2 mM MgCl 2 , and 1.5-fold molar excess of ATP and glycine.
## Structure determination of ab3403
Crystals of holo-AB3403 were cryoprotected by stages utilizing either ethylene glycol or potassium citrate for native and SeMet protein, respectively. The native protein crystals were cryo-protected with cocktails containing [bib_ref] Structural insights into nonribosomal peptide enzymatic assembly lines, Koglin [/bib_ref] Diffraction data were collected on APS beamline 23-IDB. The native data (2.7 Å) was collected using a multi-crystal, multi-dataset strategy using two crystals. A complete low resolution scan was taken for one crystal followed by a higher resolution scan of the best diffracting crystal. A high resolution region of the second crystal was combined with the two scans from the first crystal. The optimal regions were identified with the JBLU-ICE software at the GM/CA beamline. A single peak wavelength dataset (3.35 Å) was collected for SeMet labeled protein. The liganded protein crystal was collected with a single crystal.
Diffraction data were indexed, merged, and scaled using iMOSFLM 29 in space group P4 x 2 1 2. Structure determination was carried out with PHENIX 30 using a combination of experimental SAD phasing and phased molecular replacement. A partial molecular replacement solution was positioned through PHASER with a sculpted (PHENIX sculptor) model derived from PheA (PDB 1AMU) [bib_ref] Structural basis for the activation of phenylalanine in the non-ribosomal biosynthesis of..., Conti [/bib_ref] and CytC1 (PDB 3VNR). Using this partial molecular replacement model, the selenium sites were identified with the SAD data from SeMet labeled crystals. An initial model was produced with PHENIX Autobuild that contained ~65% of the protein molecule, spread across multiple symmetry related molecules. This model was combined into a single protein chain, built and refined iteratively against native data using ARP-WARP 31 , COOT 32 and PHENIX refine.
The final refinements were performed with TLS parameterization 33 with groups consisting of residues :480, 481:862, 863:959, 960:973, 974:1044, and 1054:1318, roughly defining the NRPS domain (or subdomain) boundaries. The protein is complete from residue Asn2 through Pro1319 with two small disordered loops in the adenylation domain at Asn500-Asp501 and Gly627-Gly630. The latter loop is part of the conserved serine/threonine-and glycine-rich P-loop that is involved in binding the triphosphate of the nucleotide 7 .
Additionally, the condensation domain contains electron density for a diacylglycerol lipid molecule that co-purified with the protein and potentially derived from the bacterial membrane during cell disruption. Diffraction and refinement statistics are presented in Extended Data . Experimental electron density of the ligands of both structures are presented in stereo format in Extended Data [fig_ref] Figure 3: Conformational Dynamics in NRPS modules [/fig_ref].
## Purification of entf
The enterobactin biosynthetic cluster of E. coli has been used as a model system in many studies. The full-length EntF, containing the condensation, adenylation, PCP, thioesterase domain architecture, loads serine onto the PCP domain. The condensation domain then recognizes the external carrier protein EntB that has been loaded with 2,3dihydroxybenzoate (DHB) by the activity of the freestanding adenylation domain EntE. The DHB-serine amide is then transferred to the thioesterase domain while two additional cycles of synthesis complete the enterobactin trilactone.
The EntF protein used in this study (Genbank P11454) was described previously [bib_ref] Analysis of the linker region joining the adenylation and carrier protein domains..., Miller [/bib_ref] [bib_ref] Dissection of the EntF condensation domain boundary and active site residues in..., Roche [/bib_ref]. The entf gene was PCR amplified from E. coli JM109 and cloned into a pET15-TEV vector with a N-terminal 5x His-tag and a TEV protease cleavage site [bib_ref] Analysis of the linker region joining the adenylation and carrier protein domains..., Miller [/bib_ref]. The entf vector was transformed into E. coli (BL21-DE3) cells for protein expression. Cells were grown in LB media to an OD 600 of 0.6 at 37 °C prior to protein induction with 1mM IPTG. Cells were grown overnight at 16 °C, and collected by centrifugation. The cell pellets were flash frozen in liquid nitrogen. Selenomethionine labeled EntF was expressed in M9 minimal media as described [bib_ref] Preparation of selenomethionyl proteins for phase determination, Doubliè [/bib_ref].
For purification of both native and SeMet labeled protein, cells were resuspended in lysis buffer containing 50 mM Tris-HCl pH 7.5, 400 mM NaCl, 0.2 mM TCEP, 10% glycerol, and 10 mM imidazole. Cells were lysed via sonication and centrifuged at 235,000 g for 45 minutes. Initial purification was achieved with a His-trap IMAC column. Protein was eluted using lysis buffer with 300 mM imidazole. EntF was incubated with TEV protease overnight at 4°C in a cleavage buffer containing 50 mM Tris pH 7.5, 400 mM NaCl, 0.2 mM TCEP, 10% glycerol, and 0.5 mM EDTA. Although expressed in E. coli, phosphopantetheinylation was assured by the addition of 10 nM Sfp, 1 mM CoA, and 12.5 mM MgCl 2 . The reaction was incubated at room temperature for 1-2 hours. The holo-EntF was run over an IMAC column once more to remove uncleaved protein along with Sfp. A final polishing step was performed with a Superdex 200 16/600 in a final dialysis buffer containing 50 mM EPPS pH 8.0, 150 mM NaCl, 0.2 mM TCEP, 1 mM MgCl 2 , and 10% glycerol. Prior to crystallization, the Ser-AVS inhibitor was added at a concentration 4x that of EntF and allowed to incubate for 2-4 hours at room temperature.
For electron microscopy, native EntF was purified the same as above with the exception that a minimal dialysis buffer was used, which contained 50 mM EPPS pH 8.0, 100 mM NaCl, and 0.2 mM TCEP. No inhibitor was added.
Crystal conditions for the Ser-AVS inhibited EntF were first identified using the Hauptman-Woodward high throughput screen [bib_ref] A deliberate approach to screening for initial crystallization conditions of biological macromolecules, Luft [/bib_ref]. Large diffraction quality native and SeMet crystals were grown using hanging drop vapor diffusion at 20 °C. A crystallization cocktail, consisting of 100 mM BTP pH 7.5, 125-150 mM MgCl 2 , and 22-28% PEG 4000, was diluted 1:1 with the final dialysis buffer. The hanging drops then combined protein at 30 mg/mL and the undiluted crystallization cocktail at a ratio of 1:2. This "batch mimic" limited the differences between the drop and reservoir and has been successful with other protein samples in our lab [bib_ref] Structure determination of the functional domain interaction of a chimeric nonribosomal peptide..., Sundlov [/bib_ref].
## Structure determination of entf
Native EntF crystals were cryoprotected by that addition of 2,3-butanediol directly to the crystallization drop to a final concentration of ~10%. SeMet crystals were cryoprotected similarly except with glycerol to a final concentration of ~20%. Diffraction data was collected on APS beamline 23-IDB using the rastering option to find the optimal spots on both the native the SeMet crystals. Diffraction data were indexed, merged, and scaled using iMOSFLM [bib_ref] iMOSFLM: a new graphical interface for diffraction-image processing with MOSFLM, Battye [/bib_ref] in space group P4 x 2 1 2. Structure determination for the SeMet inflection data was carried out in PHENIX 30 using a PhaserEP MR-SAD with a partial molecular replacement solution that was obtained using a sculpted model (generated with PHENIX sculptor) derived from the P. aeruginosa bidomain Adenylation-PCP protein PA1221 (PDB: [bib_ref] Structure of PA1221, a nonribosomal peptide synthetase containing adenylation and peptidyl carrier..., Mitchell [/bib_ref]. Automated model building with BUCCANEER was used to build ~65% of the structure [bib_ref] The Buccaneer software for automated model building. 1. Tracing protein chains, Cowtan [/bib_ref]. This partial model from the SeMet data was used as a molecular replacement model for the native data, and the remaining portion of the protein was built by hand (excluding the thioesterase domain, which was unresolved and comprises about 19%). This model was built and refined iteratively using COOT 32 and PHENIX refine. TLS refinement [bib_ref] TLS from fundamentals to practice, Urzhumtsev [/bib_ref] In general, the overall quality of the density was weaker for the N-terminal subdomain of the condensation domain, residues 1-186, likely reflecting the higher mobility of this region of the protein. The average B-factors for different regions of the protein (Extended Data support this conclusion.
[formula] 4DG9) [/formula]
## Negative stain em analysis of entf
EntF, purified as described above, was prepared for electron microscopy using the conventional negative staining protocol [bib_ref] Negative Staining and Image Classification -Powerful Tools in Modern Electron Microscopy, Ohi [/bib_ref] , and imaged at room temperature with a Tecnai T12 electron microscope operated at 120 kV using low-dose procedures. Images were recorded at a magnification of 71,138x and a defocus value of ~1.5μm on a Gatan US4000 CCD camera. All images were binned (2 × 2 pixels) to obtain a pixel size of 4.16 Å on the specimen level. Particles were manually excised using e2boxer [part of the EMAN 2 software suite] [bib_ref] EMAN2: an extensible image processing suite for electron microscopy, Tang [/bib_ref]. 2D reference-free alignment and classification of particle projections was performed using ISAC 40 . 17,431 projections of EntF were subjected to ISAC producing 133 classes consistent over two-way matching and accounting for 5,344 particle projections (Extended .
## Synthesis of serine adenosine vinylsulfonamide
Ser-AVS was synthesized using the protocol summarized in (Extended Data . All reactions were performed under an inert atmosphere of dry Ar in oven-dried (150 °C) glassware. 1 H and 13 C NMR spectra were recorded on a Varian 600 MHz spectrometer. Proton chemical shifts are reported in ppm from an internal standard of residual chloroform (7.26 ppm) or methanol (3.31 ppm), and carbon chemical shifts are reported using an internal standard of residual chloroform (77.3 ppm) or methanol (49.1 ppm). Proton chemical data are reported as follows: chemical shift, multiplicity (s = singlet, d = doublet, t = triplet, q = quartet, p = pentet, m = multiplet, br = broad), integration, coupling constant.
High resolution mass spectra were obtained on an Agilent TOF II TOF/MS instrument equipped with either an ESI or APCI interface. TLC analyses were performed on TLC silica gel 60F254 from EMD Chemical Inc., and were visualized with UV light or 10% PMA solution. Purifications were performed by flash chromatography on silica gel (Dynamic Adsorbents, 60A).
Materials-Chemicals, reagents and solvents were purchased from Sigma Aldrich Company, Chem-Impex or Acros Organic Fischer Company, and were used as received. An anhydrous solvent dispensing system (J. C. Meyer) using 2 packed columns of neutral alumina was used for drying THF, Et 2 O, while 2 packed columns of molecular sieves were used to dry DMF and the solvents were dispensed under argon. Compound 1 was purchased from Chem-Impex and used as received. Compound 2 41 and 4 [bib_ref] Structural and functional investigation of the intermolecular interaction between NRPS adenylation and..., Sundlov [/bib_ref] were synthesized according to the reported procedures.
## Tert-butyl(r,e)-4-(2-(n-(tert-butoxycarbonyl)sulfamoyl)vinyl)-2,2-dimethyloxazolidine-3-carboxylate (3)-to
## Kinetic analysis of ab3403
Substrate preference for the adenylation domain of holo-AB3403 was established by the pyrophosphate exchange assay 42 allowing radiolabeled PP i to be incorporated into ATP in the reverse reaction. 1.0 μM holo-AB3403 was added to 2 mM ATP, 0.2 mM NaPPi, 50 mM HEPES (pH 7.5), 100 mM NaCl, 10 mM MgCl 2 , 0.15 μCi 32 P-PPi, and 5 mM substrate. Reactions (100 μl) were incubated for 10 min at 37°C, then quenched with 0.5 ml 1. To determine the apparent kinetic constants for ATP and glycine for the holo-AB3403 adenylation domain, the NADH + consumption assay monitored at OD with full length AB3403. Hydroxylamine was used as a surrogate for the pantetheine in the second partial reaction to displace AMP for use in the coupled consumption assay [bib_ref] A continuous kinetic assay for adenylation enzyme activity and inhibition, Wilson [/bib_ref] approximately 26% sequence identity. The adenylation and PCP domains are more wellconserved, sharing ~35% identity, while the condensation (21%) and thioesterase (25%) domains are less well conserved. Domain boundaries are described in the table below.
## Extended data figure 2. substrate specificity of full-length ab3403
Amino acid specificity of AB3403 was recorded for all 20 proteinogenic amino acids, as well as 4-chlorobenzoate (4CB) and 4-hydroxybenzoate (4HB). Average values and standard deviations are shown for three replicates with each substrate; results were recorded as μmol radiolabeled ATP incorporated per min per mg of enzyme. Apparent kinetic constants are also shown for ATP and glycine.
pocket of AB3403 bound to glycine and AMP. Stereorepresentation of electron density shows the AMP, glycine, and Mg + present in the active site of the adenylation domain. Ligand carbon atoms are in green, carbons of residues within 5Å of inhibitor in gray, nitrogen in blue, oxygen in red, phosphorous in orange, and the Mg + cofactor in purple. (c) Stereorepresentation of the electron density shows the phosphopantethine on Ser1006 covalently attached to the Ser-vinylsulfonamide inhibitor in the active site of the adenylation domain. Inhibitor carbon atoms in green, carbons of residues within 4Å of inhibitor in gray, nitrogen in blue, oxygen in red, phosphorous in orange, sulfur in yellow, and water in light blue.
## Extended data figure 4. comparison of ab3403 and srfa-c pcp-condensation domain interaction
Stereorepresentation illustrating different orientations of the PCP domains of SrfA-C and AB3403 relative to the condensation domains with which they interact. AB3403 is shown with a white condensation domain and a green-cyan PCP. SrfA-C is shown with a yellow condensation domain and a pale blue PCP. The pantetheine of AB3403 is shown bound to Ser1006. The position of Ser1003, mutated to an alanine residue in SrfA-C, is also highlighted.
Extended Data .
# Supplementary material
Refer to Web version on PubMed Central for supplementary material.
[fig] Figure 1: Ribbon diagrams of complete NRPS modules (a) Domain architecture of three structurally characterized termination modules. The protein structures of (b) AB3403, (c) EntF, and (d) SrfA-C are colored with domains colored white (condensation), pink and red (adenylation domain N-and C-terminal subdomains), greencyan (PCP), and blue (thioesterase). The phosphopantetheine moieties of AB3403 and EntF, and inhibitor Ser-AVS are highlighted. [/fig]
[fig] Figure 2: NRPS domain structuresThe condensation domain of AB3403 (white) was aligned with (a) SrfA-C (yellow) and EntF (orange) on the basis of the condensation C-terminal subdomain. The AB3403 PCP is included. (b) The AB3403 condensation domain highlights residues that form the hydrophobic tunnel through which the pantetheine passes. (c) Superposition of adenylation domains of AB3403 (pink and maroon for N-and C-terminal subdomains), SrfA-C (yellow) and gramicidin synthetase, GrsA (cyan), with phenylalanine and AMP molecules of GrsA. The dotted line highlights the alternate position of the catalytic lysines of AB3403 and SrfA-C. (d) The EntF adenylation domain active site shows a covalent linkage from the pantetheine to the Ser-AVS inhibitor. Electron density calculated with coefficients of the form F o -F c generated prior to inclusion of ligands and contoured at 3σ, are shown for the (e)AB3403 condensation, (f) AB3403 adenylation, and (g) EntF adenylation domains. [/fig]
[fig] Figure 3: Conformational Dynamics in NRPS modules (A) Alternate locations of the thioesterase domain SrfA-C and AB3403. (b) Representative EM class averages of EntF. The smaller TE domain is observed in various positions relative to the condensation-adenylation didomain. Overall EntF adopts a variety of extended (top) to compact (bottom) conformations. (c) The interface between the condensation C-terminal subdomain and the adenylation domain is shown for SrfA-C, AB3403, and EntF. The adenylation surface is shown in white, highlighting in red the regions that interact with the condensation domain. The right panel shows this interface, rotated by 90° around the y-axis, with the condensation domain omitted for clarity. [/fig]
|
Can native clonal moso bamboo encroach on adjacent natural forest without human intervention?
Native species are generally thought not to encroach on adjacent natural forest without human intervention. However, the phenomenon that native moso bamboo may encroach on surrounding natural forests by itself occurred in China. To certificate this encroaching process, we employed the transition front approach to monitor the native moso bamboo population dynamics in native Chinese fir and evergreen broadleaved forest bordering moso bamboo forest in Tianmu Mountain Nature Reserve during the period between 2005 and 2014. The results showed that the bamboo front moved toward the Chinese fir/evergreen broadleaved stand with the new bamboo produced yearly. Moso bamboo encroached at a rate of 1.28 m yr −1 in Chinese fir forest and 1.04 m yr −1 in evergreen broadleaved forest, and produced 533/437 new culms hm −2 yr −1 in the encroaching natural Chinese fir/evergreen broadleaved forest. Moso bamboo coverage was increasing while adjacent natural forest area decreasing continuously. These results indicate that native moso bamboo was encroaching adjacent natural forest gradually without human intervention. It should be considered to try to create a management regime that humans could selectively remove culms to decrease encroachment.The invasion of exotic species has become a global problem and received increasing attention 1,2 . Invasive non-native plants can occupy the space and reduce the biodiversity of native ecosystem composed of local species, disrupt nutrient and hydrologic cycles, and modify the disturbance regimes and geomorphology of invaded ecosystems 3,4 . Understanding the mechanisms underlying biological invasions is crucial to evaluating invasions and will benefit for the management and restoration of invaded ecosystems 3,5 . However, almost all of the area expansions reported are about non-native species. The studies that quantify expansion or overabundance of native species are scarce. Some native species became overabundant in their natural distribution range that occupied the space of other native species like exotic species invasion, impacting seriously local ecosystem 6 . These negative effects are often paid less attention when the overabundant species is economically valuable.Bamboo is one of most valuable plants providing with lots of goods and services. Its remarkable growth rate and versatile properties such as renewability, strength and the high number of applications, have made it one of the most important plants 7 . Bamboo forests are important for biodiversity, from providing food and shelter to large animals (e.g. Giant Pandas and Mountain Gorillas) and birds, to soil organisms, insects, and other plants and shrubs that together make up the bamboo forest ecosystems 8 . Moso bamboo (Phyllostachys edulis) is one of the most widespread subtropical bamboos in the world. It is mainly distributed across southern China, including 12 provinces such as Fujian, Jiangxi, Zhejiang, Hunan, etc. Moso bamboo is the fastest growing plant in the world that can grow up to 119 cm in one day and 24 m high in 40 to 50 days 9 . From March to May is the fast growth period for shoots and new culms, while from July to September is the period of rhizome growth and shoot bud division 9 . Moso bamboo was introduced to Japan in 1736 from China 10 . As an exotic species, moso bamboo forest invasion was found 30 years ago in Japan. Moso bamboo has invaded secondary deciduous broad-leaved forests in eastern Japan near Tokyo 11,12 and in central Japan near Kyoto 13 , broadleaved forest, coniferous forest, bush and grassland near Hirasawa and Kofuki of eastern and western Japan 14 . Such phenomena revealed that the moso bamboo has potential invasiveness.
In China, the moso bamboo plantation area has been increasing steadily since 1990, not only for edible shoots but also for mature culms harvesting. Recently, moso bamboo plantations area gained faster development not only for the economic value but also as a significant carbon sink especially in Zhejiang province [bib_ref] Density, storage and spatial distribution of carbon in Phyllostachy pubescens forest, Zhou [/bib_ref]. Moso bamboo forests cover about 3.37 million hm 2 , about 2% of the total Chinese forest area and about 70% of the total Chinese bamboo area [bib_ref] State Forestry Administration, China [/bib_ref]. The rapid development of moso bamboo plantations has assisted local economic situation on the one hand, but on the other hand they may bring about expansion into natural forests. Ding et al. [bib_ref] Monitoring Phyllostachys pubescens stands expansion in National Nature Reserve of Mount Tianmu..., Ding [/bib_ref] reported that the area of moso bamboo forests increased dramatically at the speed of 4.47 hm 2 yr −1 from 1985 to 2003 occupied the space of surrounding natural forests in Tianmu Mountain National Nature Reserve of Zhejiang Province by using remote sensing and ascribe to few human disturbance for the moso bamboo area expansion, but no information is about the encroachment process. Coincidentally, our initial view was that moso bamboo may expand into adjacent natural forest by itself according to our repeatedly field observation before 2005. Similar to exotic species, expanding native moso bamboo can reduce natural diversity by monopolizing resources, changing the species composition or relative abundance of sympatric species, and even altering forest structure [bib_ref] Monitoring Phyllostachys pubescens stands expansion in National Nature Reserve of Mount Tianmu..., Ding [/bib_ref] [bib_ref] Stand structure change of Phyllostachys pubescens forest expansion in Tianmushan National Nature..., Bai [/bib_ref] [bib_ref] Plant species diversity and dynamics in forests invaded by Moso bamboo (Phyllostachys..., Bai [/bib_ref]. However, whether native moso bamboo can expand into natural forest remains a point of considerable debate. One of the fundamental questions in this debate is whether moso bamoo expansion is dependent on human support or only by itself. In order to clarify the bamboo encroachment process, we started in 2005 to set up permanent plots to test the hypothesis that moso bamboo as a native species can expand into surrounding natural forest sites by itself.
The study of invasions of bamboo in Japan has primarily relied on retrospective methods which are inherently biased toward introduced earlier 12 . However, it is not clear whether the expansion of moso bamboo founded in Japan is a result of human activities or its exotic that made it become invasive. A key point to address it is whether native stands of moso bamboo are expanding their space in China by itself. Although some related research has reported that abandoned moso bamboo plantations are quite invasive in Japan [bib_ref] The transition of bamboo groves in Kyoto, Ogura [/bib_ref] [bib_ref] Causal analysis of the invasion of broad-leaved forest by bamboo in Japan, Okutomi [/bib_ref] [bib_ref] Expansion of bamboo forests caused by reduced bamboo-shoot harvest under different natural..., Suzuki [/bib_ref] , the human activities cannot be excluded definitely. Expansion of bamboo using remotely sensed images collected at different times is unable to characterize pre-existing, human-driven factors that may have fostered invasion. Long-term field data documenting how moso bamboo expands into new habitats and impacts native forest communities over time are not available. Especially as a native species in China, whether moso bamboo can encroach on adjacent natural forest eliminating human support is still lack of information at individual and stand level based on long-term continuous observation. Given such limited research, uncertainly remains over the expansion propensity of moso bamboo.
The objectives for this study are as follows: (i) to test the hypothesis that moso bamboo can encroach into the adjacent forests without human activities with solid proof; (ii) to quantify encroachment rates, and iii) to determine whether there were different encroachment rates when expanding into different forest types by moso bamboo. We addressed these objectives by monitoring the location of new bamboo produced in the bamboo edge zone with two adjacent forest types (Chinese fir and evergreen broadleaved forest respectively) every year during the period 2005 to 2014.
# Materials and methods
Site descriptions. The study site is located in the Tianmu Mountain National Nature Reserve (119°23′ 47′ ′ -119°28′ 27′ ′ E, 30°18′ 30′ ′ -30°24′ 55′ ′ N), which belongs to typical bamboo distribution regions in Zhejiang province of southern China. According to forest resource inventories of the reserve, there was roughly 55.1 hm 2 of moso bamboo forests at the time of reserve establishment. This coverage increased to 87.5 hm 2 by 2004. There were no trees or bamboo planted by humans during this period due to nature reserve policies. Bamboo forests have also not been managed via any type of pruning or selective removal by humans, such that dead and falling bamboo culms can be frequently seen throughout the reserve. Before 1956, moso bamboo encroachment on adjacent local forests might have been limited by human harvesting shoot and culm in surrounding forest areas. It is illegal to harvest forest resources in this area since Nature Reserve established in 1956 with the approval of the Government.
Tianmu Mountain National Nature Reserve lies in the northern limit of mid-subtropical zone covering a total area of 4,284 hm 2 . The climate is damp monsoon climate with an annual precipitation of 1390-1870 mm and an annual temperature of 8.8-14.8 °C. The reserve is one of the sites with the richest subtropical higher plant species in China. There are 2,160 species of higher plants. Among them, more than 37 species are named after Tianmu Mountain and 1,200 species of medicinal plants. The vegetation type in this area is very rich, covering evergreen and deciduous broadleaved forest, bamboo forest, coniferous forests, marshes and aquatic vegetation under protection. The soils are red soil (< 600 m), yellow soil (600-1200 m), and brownish yellow soil (> 1200 m).
The moso bamboo population in this reserve is mainly distributed in the broad-leaved forest and Chinese fir forest with a distinct storey layer. The arbor layer is dominated by Cyclobalanopsis glauca, Castanopsis sclerophylla, Schima superba, Cryptomeria fortunei and Cunninghamia lanceolata. The shrub layer is dominated by Camellia fraterna, Symplocos caudata, Rhododendron ovatum, and Lindera glauca. The grass layer is dominated by the herbs of Gramineae, Compositae, Cyperaceae and Dryopteridaceae.
## Experimental design.
To test the hypothesis that moso bamboo can encroach into adjacent forest, we selected two local typical forest types (i.e. Chinese fir/evergreen broadleaved forest) bordering native moso bamboo forest. In November of 2005, we selected six sites (three for each forest type) where are far from roads and tourist attraction and established one permanent plot in each site with a size of 100 m length × 20 m width along the encroachment pathway. It was comprised of three contiguous segments in order: mono moso bamboo, mixed transition areas, where moso bamboo mixed with Chinese fir/evergreen broadleaved, and the Chinese fir/evergreen broadleaved forest. Summary of moso bamboo, Chinese fir, and broadleaved stand characteristics see . The leading edge of the transition areas extends farther into the other forest type each year, with a well-defined line of moso bamboo in 2005. The sites are moderately sloped, approximately 15°, each site is at least 200 m apart. From this study was conducted, the plots were further protected to ensure nobody entering except this study needs. If bamboo forest can expand by itself, the leading edge will move forward. Therefore, this leading edge of 2005 was defined as a line beginning to expand in future years [fig_ref] Figure 1: Dynamic distribution of moso bamboo culms in bordering Chinese fir forest during... [/fig_ref]. The plot was subdivided into 20 subplots of 5 m × 20 m. We defined the subplot No. 1 start from the leading edge toward Chinese fir/evergreen broadleaved forest. As [fig_ref] Figure 1: Dynamic distribution of moso bamboo culms in bordering Chinese fir forest during... [/fig_ref] shows, the pure moso bamboo stand located in subplot E and F. Chinese fir showed in the other subplots except subplot E and F. Pure Chinese fir stand in 2005 located in subplot No. 1-14. Encroachment of bamboo was traced annually from 2006 to 2014. The locations of the arbor trees were measured by total station and mapped in each plot. For each new culm that emerged in front of the leading edge of 2005, we measured the location in the plot using total station, and recorded height, vitality (alive/dead), and the year in which emerged.
To get detailed information on bamboo rapid growth, total height of all bamboo shoots emerged in subplot 1-2 were measured weekly until height growth ceased for more than 2 weeks. The measurements took place between March and June in 2013. Heights were measured to the nearest centimeter with hypsometer, and diameters were measured with a tape with an accuracy of 0.1 cm. The vitality (alive/dead) of each tree and bamboo culm in each subplot was recorded in 2014.
Expansion procedure mapping and expansion speed calculation. As shown in [fig_ref] Figure 2: Calculation method of annual encroachment speed of moso bamboo [/fig_ref] , if the bamboo forest leading edge moves toward the other forest type, the bamboo area will become larger and larger. The current expansion area was defined as the polygon difference between current year's (x + 1) leading edge and last year's (x) as the rest of boundaries fixed in the plot. The shade area is the expansion area at x + 1 year. The annual expansion speed could be determined by dividing the shade area by the width of plot (20 m for this study). If bamboo forest could not expand, the speed is 0, otherwise, it is bigger than 0.
Statistical analyses. The data were analyzed using SPSS 18.0 software. Paired T test was used to determine the statistical significance of two forests on bamboo encroachment rate over the period. One-way ANOVA was performed to test the statistical significance of bamboo encroachment rate on the same forest among different years. Statistical significance for all analyses was set at p < 0.05.
# Results
Dynamic encroachment process of moso bamboo. In 2005, tree stems were abundant in the non-bamboo forests and culms of moso bamboo were not present in subplot 1. From 2006 on, bamboo shoots began to appear. The bamboo front moved into the Chinese fir/evergreen broadleaved stand with new bamboo produced yearly. Therefore, bamboo occupied a broader and broader space that was dominated by Chinese fir/evergreen broadleaved trees before 2006, suggesting that moso bamboo was encroaching gradually. [fig_ref] Figure 1: Dynamic distribution of moso bamboo culms in bordering Chinese fir forest during... [/fig_ref] Stand . There were significant differences between Chinese fir forest and evergreen broadleaved forest in the same year (P < 0.05). The coefficient of variation (CV) of the yearly speeds was 0.36 m yr −1 for Chinese fir stand and 0.38 m yr −1 for evergreen broadleaved forest. There were significant differences among different years in the same stand
# Discussion
Encroachment evidence of native moso bamboo. We proved that the moso bamboo front expanded automatically in both natural Chinese fir forest and evergreen broadleaved forest without human support by ten years permanent plots observation. This expansion is similar to the results from nonnative areas where moso bamboo has been introduced from China 13 . Okutomi et al. [bib_ref] Causal analysis of the invasion of broad-leaved forest by bamboo in Japan, Okutomi [/bib_ref] reported that the area covered by exotic moso bamboo forest had invaded by a factor of 2.7 over last in southwestern parts of Tokyo in Japan. Additionally, Ding et al. [bib_ref] Monitoring Phyllostachys pubescens stands expansion in National Nature Reserve of Mount Tianmu..., Ding [/bib_ref] also used remote photographs to document bamboo expansion rate of 4.47 hm 2 yr −1 by detecting the bamboo area changes between 1985 and 2003 in Tianmu Mountain Nature Reserve in China. But some researchers denied the bamboo expansion naturally as a native species and argued that bamboo encroachment into forest took place must be in the case of requires human management [bib_ref] Expansion of bamboo forests caused by reduced bamboo-shoot harvest under different natural..., Suzuki [/bib_ref]. Based on above considerations, this study was conducted in the sites where eliminated human support and the results of 10 years observation indicated that moso bamboo could encroach on adjacent forests by itself. Coincidentally, this study is similar to the expansion of native woody bamboos (Guadua tagoara) occurred in the Brazilian Atlantic Forest [bib_ref] Bamboo overabundance alters forest structure and dynamics in the Atlantic Forest hotspot, Lima [/bib_ref].
In this study, we found moso bamboo could encroach into natural undisturbed Chinese fir forest and evergreen broadleaved forest and there were significant difference in the encroachment rate between them (Table 2), indicating that bamboo encroachment rate depended on adjacent forest types. The difference in expansion rate . Encroachment speeds of moso bamboo from 2006 to 2014 (m yr −1 ) Note: Data are means with standard error. Different lowercase letters in the same column indicate significant differences between stands in the same year (p < 0.05). Different capital letters in the same row indicate significant differences among years in the same stand (p < 0.05) between Japan and China indicated that bamboo encroachment may be related to many factors, such as the bamboo itself, the soil type, the climate, the neighbor vegetation, and the topographical conditions of the expanded area.
Mechanism of moso bamboo encroachment. Moso bamboo encroaching on the adjacent forest mainly includes underground and aboveground processes. First, the underground rhizomes spread into neighboring forest, which cannot be seen by people. Then several years later bamboo sprouts appear from the underground rhizomes. Once the sprouts emerge from the earth, it will grow into bamboo in short time (only about 60 days) to finish stem elongation, reaching 12-15 m [fig_ref] Figure 4: The rapid height growth of moso bamboo in subplot 1-2 of Chinese... [/fig_ref]. These characters allow bamboo to occupy the upper crown quickly. Moreover, bamboo shoots are highly shade tolerant. There are almost no leaves and branches when bamboo shoots initiate stem elongation. Maternal bamboo provides almost all of the energy and nutrients instead of its photosynthesis. Bamboo shade tolerance is manifest as low photosynthetic rates and rapid early growth. Most invasive species are not shade tolerant, so it is difficult for them to invade into forests with intact canopies or undisturbed communities 20 . However, some shade tolerant tree species, such as the Norway maple can invade intact forests and thus demonstrate the weak resistance to shade tolerant species in some forests [bib_ref] Intact forests provide only weak resistance to a shade-tolerant invasive Norway maple..., Martin [/bib_ref]. The shade tolerance of bamboo shoots is beneficial for them to grow well beneath the dense canopy [bib_ref] The independence of clonal shoot's growth from light availability supports moso bamboo..., Wang [/bib_ref].
We found that the bamboo encroachment speed differed each two years. This finding is consistent with a pattern of rich and poor year. Bamboo has the rich years bearing lots of shoots and growing up into young bamboos, and the poor year changing leaves and producing new rhizomes. In this way, moso bamboo can encroach into adjacent forests intermittently quickly or slowly two-year cycle. This is different from Norway maple, which invades through seed dispersal and seed propagation at low frequency and long distance [bib_ref] Intact forests provide only weak resistance to a shade-tolerant invasive Norway maple..., Martin [/bib_ref]. Although there are probably many factors that contribute to moso bamboo encroachment, we can affirm that this continuous encroachment of bamboo was inherent in local areas. In other words, undisturbed forests were not immune to the encroachment of moso bamboo. Compared with the reserve after establishment, it can be guessed that the increase in moso bamboo area could also occur before the reserve was established, but local people actually harvested the moso and helped maintain it unconsciously to more manageable levels to prevent it from overrunning the other forests. However, the reserve establishment may cause or promote this increase due to nature reserve policies which prohibit people from harvesting bamboo culms and even bamboo shoots. In this sense, outlawing bamboo harvest inside the reserve was a detriment to this system.
## Implications of bamboo encroachment.
Although moso bamboo is native in China, the proved ability of bamboo encroachment similar to exotic species invasion is somewhat alarming. Unconstrained moso bamboo encroachment could lead to the formation of the new bamboo forests, well-known in Eastern Asia 12 . The surrounding natural forest could be drastically disturbed. Such vegetation changes may alter forest floor microclimate with respect to light, temperature, and moisture [bib_ref] Light intensity changes on Cunninghamia lanceolata in mixed stands with different concentrations..., Liu [/bib_ref]. Encroachment also potentially alters forest structure and dynamics and simplifies stand structure 18 , substantially reducing tree and shrub diversity in the forests [bib_ref] Plant species diversity and dynamics in forests invaded by Moso bamboo (Phyllostachys..., Bai [/bib_ref] , modifying soil community structure, and increasing microbial biomass and taxonomic diversity [bib_ref] Bamboo invasion of native broadleaf forest modified soil microbial communities and diversity, Xu [/bib_ref]. The conversion of broadleaved forests to bamboo-dominated forests also may alter soil properties [bib_ref] Impacts of moso bamboo (Phyllostachys pubescens) invasion on dry matter and carbon..., Fukushima [/bib_ref] and reduced ecosystem storage of carbon (C) and nitrogen (N), having important impacts on regional C and N cycles [bib_ref] The effects of Phyllostachys pubescens expansion on soil fertility in National Nature..., Wu [/bib_ref] [bib_ref] Effects of Phyllostachys edulis expansion on carbon storage of evergreen broadleaved forest..., Yang [/bib_ref]. In southern China, there are vast areas dominated by native moso bamboo (> 3.37 million hm 2 ) [bib_ref] State Forestry Administration, China [/bib_ref]. If rapid encroachment of moso bamboo forest into adjacent broadleaved forests continues in these subtropical areas, it would become ecological risk unexpected.
We found that even for protected undisturbed forests, it is difficult to resist the encroachment of bamboo. If this biological expansion is not to be controlled, moso bamboo seems likely to expand its area every year and decreasing the areas of evergreen broadleaved forest and Chinese fir forest. As shown in [fig_ref] Figure 5: Moso bamboo encroaching on Chinese fir forest in subplot 2, forming the... [/fig_ref] , the pure Chinese fir forest transformed to the mixed forest of moso bamboo and Chinese fir dominated by bamboo. Even some trees lost vitality and began to die one by one 5 years after moso bamboo encroaching. It warned us that the protected forests adjacent to moso bamboo in Tianmushan Nature reserve and other protected areas may be dominated by moso bamboo forest and losing their protecting value. These results have implications for the management and control of this encroaching species. Some actively strategies should be applied to decrease encroachment, such as digging a trench around bamboo stands or physically new shoots or culms each year.
[fig] Figure 1: Dynamic distribution of moso bamboo culms in bordering Chinese fir forest during the period 2005 to 2014. All bamboo culms classified by birth year in subplot 1-3 were mapped. All Chinese fir stems in subplot 1-3 were also mapped. The 10 lines of 05 to 14 (abbreviation of 2005, 2006, … … , 2014 ) were the bamboo leading edge of each year. The pattern for evergreen broadleaved forest is not shown because it is very similar to the one for fir shown here. Note: F→ E: Subplot E and F, mono moso bamboo stand; D→ A: Subplot A, B, C and D, transistion zones; 1→ 3: Subplot 1, 2 and 3, Chinese fir stand without bamboo before 2005, presenting new shoots from 2005 to 2014; 4→ 14: Subplot 4, 5, 6,7, 8, 9, 10, 11, 12, 13 and 14, Chinese fir stand without bamboo shoots present before 2014. The encroachment path is from left to right. depicted the dynamic distribution of moso bamboo in adjacent nature Chinese fir forest during the period 2005 to 2014. Encroachment speed of moso bamboo. The encroachment speeds ranged from 0.57 m yr −1 to 1.80 m yr −1 , with the average being 1.28 m yr −1 in Chinese fir forest and 1.04 m yr −1 in evergreen broadleaved forest [/fig]
[fig] Figure 2: Calculation method of annual encroachment speed of moso bamboo. x + 1 line is current year's leading edge and x line is the last year's leading edge. Scientific RepoRts | 6:31504 | DOI: 10.1038/srep31504 (P < 0.05). The mean rate of encroachment on Chinese fir forest and broadleaved forest by bamboo was approximately 1.65 m yr −1 , 1.36 m yr −1 in rich years in which moso bamboo mainly bears lots of shoots and grow up into young bamboos, and 0.81 m yr −1 , 0.64 m yr −1 in poor years in which moso bamboo mainly changes leaves and produces new rhizomes, respectively. Changes of culm density during the encroachment of moso bamboo. As Fig. 3 showed, bamboo density increased gradually not only in the adjacent Chinese fir forest but also in the evergreen broadleaved forest during the period. The moso bamboo produced 533/437 new culms hm −2 yr −1 in the encroaching natural Chinese fir/evergreen broadleaved forest. In subplot 1 the bamboo number increased from 0 inds. hm −2 in 2005 to 3200 inds. hm −2 in 2014 in Chinese fir forest, from 0 inds. hm −2 in 2005 to 2866.7 inds. hm −2 in 2014 in evergreen broadleaved forest, respectively. The density of bamboo in subplot 1 was higher and decreaed in the expanding direction of subplot 2, where the bamboo began to occur in 2010 for Chinese fir forest and in 2011 for evergreen broadleaved forest. By contrast, the bamboo density increased gradually from 0 to 1333.4 inds. hm −2 in subplot 2 and 266.7 inds. hm −2 in subplot 3 of Chinese fir forest, and 1066.7 inds. hm −2 in subplot 2 of evergreen forest in 2014. It meant there was no bamboo in bordering natural forest in 2005. But more and more bamboo appeared in the adjacent forest and moved farther and farther during the period 2006 to 2014, the mixed bamboo forest area becoming larger and larger. Although dead culms of moso bamboo were found in subplot 1, the bamboo forest may regenerae new culms yearly and increase density autonomously. [/fig]
[fig] Figure 3: Dynamics of moso bamboo culms density in Chinese fir stand (a) and evergreen broadleaved stand (b) during the period 2006 to 2014. Subplot numbers are the same as in Fig. 1. Bars correspond to means and error bars are standard errors. [/fig]
[fig] Figure 4: The rapid height growth of moso bamboo in subplot 1-2 of Chinese fir forest in 2013. Symbols correspond to means and bars are standard errors. [/fig]
[fig] Figure 5: Moso bamboo encroaching on Chinese fir forest in subplot 2, forming the mixed forest of moso bamboo and Chinese fir at the Tianmu Mountain Natural Reserve. Some trees lost vitality and began to die just 5 years after moso bamboo encroaching (photo taken in 2013). Scientific RepoRts | 6:31504 | DOI: 10.1038/srep31504 [/fig]
|
Exposure‐Response Characterization of Tofacitinib Efficacy in Moderate to Severe Ulcerative Colitis: Results From Phase II and Phase III Induction and Maintenance Studies
Tofacitinib is an oral small molecule JAK inhibitor for the treatment of ulcerative colitis. Relationships between plasma tofacitinib concentration and efficacy were characterized using exposure-response (E-R) models, with demographic and disease covariates evaluated as potential predictors of efficacy. Data were from phase II and III (OCTAVE Induction 1 and 2) induction studies, and a phase III maintenance study (OCTAVE Sustain). Induction studies included 1,355 patients (tofacitinib 0.5, 3, 10, or 15 mg b.i.d. or placebo). The maintenance study included 592 patients (tofacitinib 5 or 10 mg b.i.d. or placebo). E-R models, including induction patients predicted placebo-adjusted remission rates of 6.4% and 12.7% at week 8 for tofacitinib 5 and 10 mg b.i.d., respectively; corresponding rates in patients without prior tumor necrosis factor inhibitor (TNFi) failure were 12.8% and 20.4%. Estimates to achieve/maintain remission at week 52 of maintenance were 29% and 18% (tofacitinib 5 mg b.i.d.), and 41% and 26% (tofacitinib 10 mg b.i.d.), for patients in remission or not following induction, respectively. During maintenance, patients with prior TNFi failure had lower probability of remission on 5 mg b.i.d. (24.9%) than 10 mg b.i.d. (35.0%). Results indicated tofacitinib 10 mg b.i.d. was an appropriate induction dose but suggested efficacy with 5 mg b.i.d. in patients without prior TNFi failure. Tofacitinib 5 mg b.i.d. was efficacious for maintenance, although patients with prior TNFi failure might see additional benefit on 10 mg b.i.d. Per product labeling, recommended tofacitinib induction dose is 10 mg b.i.d., then maintenance at 5 mg b.i.d. For patients who lose response during maintenance, 10 mg b.i.d. may be considered, limited to the shortest duration.
This study characterized the relationship between tofacitinib exposure and efficacy end points for both induction and maintenance therapy in patients with ulcerative colitis (UC), and identified patient-specific factors that are important determinants of efficacy.
## What does this study add to our knowledge?
Results support the use of tofacitinib 10 mg b.i.d. for induction in patients with moderate to severe UC. However, they also suggest tofacitinib 5 mg b.i.d. might be efficacious for induction in patients without prior tumor necrosis factor inhibitor (TNFi) failure. Although 5 mg b.i.d. was efficacious for maintenance, patients with prior TNFi failure might see additional benefit on 10 mg b.i.d. Patients with lower baseline disease activity were more likely to achieve efficacy end points at the end of maintenance.
## How might this change clinical pharma-cology or translational science?
Ulcerative colitis (UC) is a chronic inflammatory disease localized to the colon, characterized by abdominal symptoms, including diarrhea, abdominal pain, rectal bleeding, and urgency. [bib_ref] Inflammatory bowel disease: clinical aspects and treatments, Fakhoury [/bib_ref] Treatments available for patients with UC include 5-aminosalicylates (5-ASA), corticosteroids, thiopurines, methotrexate, tumor necrosis factor inhibitors (TNFi; e.g., adalimumab, infliximab, and golimumab), integrin inhibitors (e.g., vedolizumab), interleukin-12/23 inhibitors (e.g., ustekinumab), and Janus kinase (JAK) inhibitors (e.g., tofacitinib and filgotinib). [bib_ref] ACG clinical guideline: ulcerative colitis in adults, Rubin [/bib_ref] [bib_ref] Third European evidence-based consensus on diagnosis and management of ulcerative colitis. Part..., Harbord [/bib_ref] [bib_ref] Clinical outcomes with ustekinumab as rescue treatment in therapy-refractory or therapy-intolerant ulcerative..., Ochsenkühn [/bib_ref] [bib_ref] Filgotinib as induction and maintenance therapy for ulcerative colitis (SELECTION): a phase..., Feagan [/bib_ref] The aim of treatment is to induce and maintain remission, with a goal of sustained steroid-free remission. [bib_ref] ACG clinical guideline: ulcerative colitis in adults, Rubin [/bib_ref] Tofacitinib is an oral small molecule JAK inhibitor for the treatment of UC. The efficacy and safety of tofacitinib 10 mg twice daily (b.i.d.) in patients with moderate to severe UC has been demonstrated as induction therapy in 8-week phase II (NCT00787202) and phase III (OCTAVE Induction 1 and 2; NCT01465763 and NCT01458951) studies. [bib_ref] Tofacitinib, an oral Janus kinase inhibitor, in active ulcerative colitis, Sandborn [/bib_ref] [bib_ref] Tofacitinib as induction and maintenance therapy for ulcerative colitis, Sandborn [/bib_ref] Both tofacitinib 5 and 10 mg b.i.d. were investigated further in a 52-week phase III maintenance study (OCTAVE Sustain; NCT01458574). [bib_ref] Tofacitinib as induction and maintenance therapy for ulcerative colitis, Sandborn [/bib_ref] The approved tofacitinib dose for induction therapy is 10 mg b.i.d., followed by maintenance at 5 mg b.i.d.For patients with loss of response during maintenance therapy, tofacitinib 10 mg b.i.d. may be considered and limited to the shortest duration.
Population pharmacokinetics (PKs) of tofacitinib have been characterized for patients with UC using nonlinear mixed-effects modeling of pooled data from phase II/III studies. Tofacitinib PKs were linear, with a dose-proportional increase in exposure. Oral clearance of tofacitinib did not change significantly over time, and dose adjustments for age, weight, sex, race, or baseline disease severity were not required. [bib_ref] Population pharmacokinetics of tofacitinib in patients with moderate to severe ulcerative colitis, Vong [/bib_ref] In the phase II study, baseline disease activity was the most important predictor of efficacy. [bib_ref] Exposure-response characterization of tofacitinib efficacy in moderate to severe ulcerative colitis: results..., Mukherjee [/bib_ref] The aim of this analysis was to characterize the relationship between tofacitinib exposure and efficacy end points for both induction and maintenance therapy in patients with UC, using data from phase II/III studies, and identify patient-specific (demographic and disease) factors that are important determinants of efficacy.
# Methods
## Study design and patients
These post hoc analyses included data from a phase II dose-ranging induction study, two phase III induction studies (OCTAVE Induction 1 and 2), and a phase III maintenance study (OCTAVE Sustain). [bib_ref] Tofacitinib, an oral Janus kinase inhibitor, in active ulcerative colitis, Sandborn [/bib_ref] [bib_ref] Tofacitinib as induction and maintenance therapy for ulcerative colitis, Sandborn [/bib_ref] These studies were approved by the institutional review boards or independent ethics committees for each center. [bib_ref] Tofacitinib, an oral Janus kinase inhibitor, in active ulcerative colitis, Sandborn [/bib_ref] [bib_ref] Tofacitinib as induction and maintenance therapy for ulcerative colitis, Sandborn [/bib_ref] The phase II induction study was an 8-week, randomized, doubleblind, placebo-controlled, parallel-group, multicenter study involving patients with moderate to severe UC. Disease activity was assessed using the Mayo score (range 0-12 points; higher scores indicate higher disease activity; a full description of the Mayo score is provided in the Supplementary Information). [bib_ref] Coated oral 5-aminosalicylic acid therapy for mildly to moderately active ulcerative colitis...., Schroeder [/bib_ref] Moderate to severe UC was defined as a total Mayo score ≥ 6. Patients were randomized to receive tofacitinib 0.5, 3, 10, or 15 mg b.i.d., or placebo. Local read of endoscopy was used for efficacy assessments.
Patients enrolled in OCTAVE Induction 1 or 2 had moderate to severe UC (defined as a total Mayo score ≥ 6, a rectal bleeding subscore ≥ 1, and an endoscopic subscore ≥ 2), and had failed or were intolerant to ≥ 1 prior UC treatments (oral or intravenous corticosteroids, azathioprine/mercaptopurine, or TNFi). Patients were randomized 4:1 to receive tofacitinib 10 mg b.i.d. or placebo. A total of 22 patients in OCTAVE Induction 1 and 2 received tofacitinib 15 mg b.i.d. [bib_ref] Tofacitinib as induction and maintenance therapy for ulcerative colitis, Sandborn [/bib_ref] Stable doses of concomitant oral 5-ASA and oral corticosteroids (≤ 25 mg/day prednisone equivalent) were permitted. Concomitant therapy with TNFi, azathioprine, methotrexate, or 6-mercaptopurine was prohibited. The primary end point was remission at week 8 (defined as a total Mayo score of ≤ 2 points with no individual subscore > 1 point, and a rectal bleeding subscore of 0). The key secondary end point was endoscopic improvement at week 8 (a Mayo endoscopic subscore of 0 or 1; defined as mucosal healing in the original OCTAVE protocols). Local and central reads of endoscopy were performed; and registration end points were based on central reads.
Patients with clinical response (defined as a decrease from induction study baseline total Mayo score of ≥ 3 points and ≥ 30%, plus a decrease in rectal bleeding subscore of ≥ 1 point or an absolute rectal bleeding subscore of 0 or 1) at week 8 of OCTAVE Induction 1 or 2 were eligible to participate in the maintenance study, OCTAVE Sustain. Patients were re-randomized 1:1:1 to receive tofacitinib 5 or 10 mg b.i.d. or placebo. Patients were required to remain on stable doses of their concomitant medications; per protocol, patients were required to taper off corticosteroids. The primary end point was remission at week 52. Key secondary end points were endoscopic improvement at week 52 and sustained steroidfree remission (defined as being in remission, in addition to not requiring any treatment with corticosteroids for ≥ 4 weeks prior to the visit) among patients in remission at baseline (i.e., patients in remission following induction therapy). Local and central reads of endoscopy were performed; end points were based on central reads.
## Pharmacokinetics
Individual PK parameters for exposure index were reported previously. [bib_ref] Population pharmacokinetics of tofacitinib in patients with moderate to severe ulcerative colitis, Vong [/bib_ref] Average and trough concentrations (C avg and C trough , respectively) were derived from the dose-normalized, individual empirical Bayes estimates obtained from the population PK model. Although both C avg and C trough were evaluated in exposure-response (E-R) analyses, only models using C avg as predictor are presented.
## Exposure-response models
The E-R analyses were performed using maximum likelihood methods. SAS (version 9.2; SAS Institute, Cary, NC) was used for data handling and data analysis. Parameters describing E-R relationship (population typical value and residual variance) were estimated and potential covariate effects were investigated as predictors of variability. End points were modeled as a function of exposure.
The binary efficacy end points for induction studies, measured at the end of the study (week 8), were modeled using logistic regression models evaluating either linear or nonlinear (maximum effect (E max ) model) relationships with tofacitinib exposures, as reported previously. [bib_ref] Exposure-response characterization of tofacitinib efficacy in moderate to severe ulcerative colitis: results..., Mukherjee [/bib_ref] Base E-R models were developed first, using pooled central-read data from phase III induction studies (OCTAVE Induction 1 and 2), followed by evaluation of the effect of covariates on model parameters using step-wise covariate modeling with forward selection and backward elimination steps. Separately, E-R analysis of pooled phase II/III induction studies was performed using locally read endoscopy end points. Only locally read endoscopies were performed in the phase II study, according to the regulatory guidance at that time.
Data from the phase II study were included as a wider range of tofacitinib doses were available (tofacitinib 0.5, 3, 10, and 15 mg b.i.d., as well as placebo), to better inform induction efficacy at tofacitinib doses lower than 10 mg b.i.d. (the only tofacitinib dose evaluated in phase III induction studies).
For the maintenance study, clinical end points (based on central reads) at weeks 24 and 52 were modeled using a Markov transition model and a longitudinal logistic regression model using linear or E max link functions. The Markov model described the probabilities of transition from achiever to non-achiever status, and vice versa, for each binary end point over time [fig_ref] Figure 1: Probability of [/fig_ref]. Nonresponder imputation was used. E-R models were first run without covariates using data from OCTAVE Sustain. E-R models with covariates were used to identify significant predictors of efficacy (based on central reads).
For both induction and maintenance studies, covariates included in the full model development were: baseline Mayo score (based on central or local endoscopic reads), extent of baseline disease, prior TNFi failure, prior immunosuppressant use, concomitant oral corticosteroid use, concomitant 5-ASA use, age, body weight, sex, and race (Asian vs. non-Asian). Albumin and C-reactive protein correlated with baseline Mayo score and were therefore not included in this analysis (albumin was also shown not to be independently correlated with efficacy in a previous analysis of these studies). 14 Covariates were evaluated on placebo (intercept) and drug (E max ) effect parameters , and tested using a stepwise (forward and backward) method. Backward models were established using statistically significant (P < 0.05) covariates in the forward process. Final models were established using statistically significant (P < 0.01) covariates in the backward process. For nested models, significant covariates were determined using a variant of the traditional stepwise selection, where decisions on which potential covariate to add or drop at any step and when to terminate the selection, were based on a likelihood ratio test.
# Results patients
Baseline demographic and clinical characteristics for all patients have been reported previously. [bib_ref] Tofacitinib, an oral Janus kinase inhibitor, in active ulcerative colitis, Sandborn [/bib_ref] [bib_ref] Tofacitinib as induction and maintenance therapy for ulcerative colitis, Sandborn [/bib_ref] These analyses included data from 1,355 patients treated with tofacitinib 0.5, 3, 10, or 15 mg b.i.d., or placebo, in the phase II/III induction studies, and 592 patients treated with tofacitinib 5 or 10 mg b.i.d., or placebo, in OCTAVE Sustain.
For all patients included in the E-R analysis, demographics and clinical characteristics by phase and by treatment group are shown in . Overall, at baseline of induction studies, 46.9% of patients had prior TNFi failure and 69.2% of patients had previously used immunosuppressants. Mean total Mayo score, determined by local read of endoscopy, was lower in the phase II induction study (8.2, SD 1.6) than in OCTAVE Induction 1 and 2 (9.0, SD 1.5). Demographics and clinical characteristics for patients in OCTAVE Sustain were similar to those in OCTAVE Induction 1 and 2, except for a lower mean total Mayo score and a greater proportion taking concomitant 5-ASA. Demographics and clinical characteristics for patients without prior TNFi failure in induction studies were similar to those of the overall population [fig_ref] Table 2: Probability estimates to achieve remission or endoscopic improvement at week 8 of... [/fig_ref].
## Exposure-response models for tofacitinib as induction therapy
For E-R analyses using data from induction studies, an E max logistic regression model was selected for all binary end points, as model fits were found to be acceptable for all end points. C avg was used as the predictor, as this was previously shown to be the most relevant drug-exposure measure of tofacitinib efficacy based on the drug's mechanism of action, temporal response across diseases, and examination of results in rheumatoid arthritis. [bib_ref] Model-informed development and registration of a once-daily regimen of extended-release tofacitinib, Lamba [/bib_ref] Base model: Evidence of greater efficacy for tofacitinib 10 vs. 5 mg b.i.d. in induction studies. Patients enrolled in OCTAVE Induction 1 or 2 were randomized to receive either tofacitinib 10 mg b.i.d. or placebo, although 22 patients received tofacitinib 15 mg b.i.d. E-R model predictions for the pooled OCTAVE Induction 1 and 2 studies were consistent with observed data and indicated that the probability of remission or endoscopic improvement increased with increasing C avg [fig_ref] Figure 1: Probability of [/fig_ref]. At week 8, C avg values were similar in patients in remission to those not in remission [fig_ref] Figure 2: Observed [/fig_ref]. Probability estimates for achievement of remission or endoscopic improvement by the logistic E max model were 6.5% and 14.8% for placebo, 12.8% and 24.8% for tofacitinib 5 mg b.i.d., and 19.1% and 32.6% for tofacitinib 10 mg b.i.d., respectively [fig_ref] Table 2: Probability estimates to achieve remission or endoscopic improvement at week 8 of... [/fig_ref]. Predicted placebo-adjusted remission and endoscopic improvement rates were 12.7% and 17.8% for tofacitinib 10 mg b.i.d., and 6.4% and 10.1% for tofacitinib 5 mg b.i.d., respectively [fig_ref] Table 2: Probability estimates to achieve remission or endoscopic improvement at week 8 of... [/fig_ref].
Estimated concentration at half-maximum effect from base models for remission and endoscopic improvement were 56.5 and 49.0 ng/mL, respectively. Model-predicted placebo-adjusted remission and endoscopic improvement rates with tofacitinib 10 mg b.i.d. were 12.7% and 17.8%, respectively [fig_ref] Table 2: Probability estimates to achieve remission or endoscopic improvement at week 8 of... [/fig_ref]. The geometric mean C avg at 10 mg b.i.d. (33.6 ng/mL; derived from previously reported PK parameters) [bib_ref] Population pharmacokinetics of tofacitinib in patients with moderate to severe ulcerative colitis, Vong [/bib_ref] corresponded to ED 67 on the E-R curve for remission.
Covariate effects on base exposure-response model parameters for induction efficacy. The step-wise covariate modeling approach was used to evaluate the effect of covariates. In the final model, only prior TNFi failure and baseline Mayo score were significant predictors of remission and endoscopic improvement . Estimated parameters (95% confidence interval (CI)) for remission were: prior TNFi failure on intercept, 1.23 (0.39-2.08); prior TNFi failure on E max , 1.61 (−0.26 to 3.48); and baseline Mayo score on intercept, 0.14 (0.080-0.20). For endoscopic improvement, estimated parameters (95% CI) were: prior TNFi failure on intercept, 0.61 (0.31-0.91); and baseline Mayo score on intercept, 0.30 (0.19-0.40).
Evidence of efficacy for tofacitinib 5 mg b.i.d. as induction therapy in patients without prior TNFi failure. E-R modeling of phase III induction data identified prior TNFi failure status as a significant predictor of efficacy, therefore further E-R analyses were performed to examine if tofacitinib 5 mg b.i.d. could be an effective induction dose in patients without prior TNFi failure. As OCTAVE Induction 1 and 2 only included patients who received tofacitinib 10 or 15 mg b.i.d., or placebo, data from the phase II induction study were included as these data included both lower (tofacitinib 0.5 and 3 mg b.i.d.) and higher (tofacitinib 10 and 15 mg b.i.d.) tofacitinib doses relative to the proposed 5 mg b.i.d. dose. Efficacy end points were based on local reads. Given the objective of this analysis, the dataset included only the pooled subpopulation of patients without prior TNFi failure in phase II/III induction studies. Approximately 80% of the phase II population had no prior TNFi failure, 12 therefore, robust doseranging information was available for this subpopulation.
The subpopulation of patients without prior TNFi failure consisted of 687 TNFi-naïve patients and 33 patients who previously received a TNFi without treatment failure. Of these 720 patients, 712 were included in the final E-R analysis dataset; two patients were excluded due to missing baseline Mayo scores and six patients were excluded due to a baseline Mayo score < 6 (per enrollment criteria). Demographics and clinical characteristics at baseline in phase II and phase III induction studies and the phase III maintenance study, for patients included in these analyses C-reactive protein, mean (SD), mg/L 13.8 (20.9) 11.6 (18.9) 11.6 (18.
## Exposure-response models for tofacitinib as maintenance therapy
A longitudinal E-R analysis of binary efficacy end points at weeks 24 and 52 of OCTAVE Sustain was performed using a binomial transition model with Markov dependence. A logistic regression model was applied as an alternate model to evaluate the sensitivity of results to model assumptions. Consistent with the induction analyses, C avg was used as the predictor for maintenance analyses.
Base model: Evidence of greater efficacy for tofacitinib maintenance therapy in patients in remission or with endoscopic improvement at baseline. E-R model predictions using the Markov transition model (without covariates) were carried out at week 52 by baseline remission or endoscopic improvement (i.e., endoscopic subscore of 0 or 1) status using expected C avg values at tofacitinib 5 and 10 mg b.i.d. doses. At weeks 24 and 52, C avg values were similar in patients in remission to those not in remission [fig_ref] Figure 2: Observed [/fig_ref]. E-R model predictions were consistent with observed data and indicated efficacy in maintenance increased with increasing tofacitinib exposure [fig_ref] Figure 3: Probability of [/fig_ref]. Markov transition model predictions for all efficacy end points were also consistent with observed data at week 24, and from weeks 24-52 [fig_ref] Figure 3: Probability of [/fig_ref].
Probability estimates to achieve or maintain remission, sustained steroid-free remission, or endoscopic improvement at week 52 by Markov transition model are provided in ; parameter estimates for the basic Markov transition model are provided
## Article
in . For both the tofacitinib 5 and 10 mg b.i.d. groups, estimated placebo-adjusted probabilities of achieving/maintaining remission or sustained steroid-free remission at week 52 were greater for patients in remission at baseline than for those not in remission . For patients receiving tofacitinib 5 mg b.i.d., the predicted relative probabilities of achieving/maintaining remission and sustained steroid-free remission at week 52 were 29% and 26%, respectively, for patients in remission at baseline, and 18% and 9%, respectively, for patients not in remission at baseline. The predicted relative probabilities of achieving/maintaining remission and sustained steroid-free remission with tofacitinib 10 mg b.i.d. were 41% and 38%, respectively, for patients in remission at baseline, and 26% and 13%, respectively, for patients not in remission at baseline . Similar findings were seen for the end point of endoscopic improvement . Efficacy in maintenance increased with increasing tofacitinib exposure, with predicted relative increase in remission and endoscopic improvement between tofacitinib 5 and 10 mg b.i.d. ranging from 27-44% (based on the ratio of tofacitinib 10 mg b.i.d.:tofacitinib 5 mg b.i.d. response ranging from 1.27-1.44) in the overall population, based on the geometric mean C avg at these doses . The relative increase in efficacy from tofacitinib 5 to 10 mg b.i.d. was 35-37% for sustained remission, 53-59% for sustained endoscopic improvement, and 33-34% for sustained steroid-free remission.
Covariate effects on base exposure-response model parameters for maintenance efficacy. Models were used to identify covariates that were predictive of response at week 52 of the maintenance study. In the final model, baseline Mayo score, oral corticosteroid use, and age were significant predictors of remission and endoscopic improvement . Estimated parameters (95% CI) for remission were: effect of baseline Mayo score on intercept nonachievers, 0.1609 (0.07861-0.2432); effect of age on intercept for nonachievers, −0.00841 (−0.01428 to −0.00254); and effect of oral corticosteroid use at baseline on E max for achievers, −0.3582 (−0.5804 to −0.1360). For endoscopic improvement, estimated parameters (95% CI) were: effect of baseline Mayo score on intercept for nonachievers, 0.1321 (0.03725-0.2270); effect of age on intercept nonachievers, −0.00977 (−0.01703 to −0.00252); and effect of oral corticosteroid use at baseline on intercept for achievers, 0.6382 (0.09051-1.1858). The effects of selected covariates (corticosteroid use at baseline; induction treatment (placebo or tofacitinib 15 mg b.i.d. (remission and sustained steroid-free remission only)); baseline Mayo score (0, 2, or 10); and age (18 or 80 years)) are shown in . Covariates were not significantly associated with the achievement of remission, sustained steroid-free remission, or endoscopic improvement at week 52, with the exception of a baseline Mayo score of 10, which was highly correlated with achievement of all efficacy end points, especially in baseline nonachievers.
## Article
Evidence of greater efficacy for tofacitinib 10 mg b.i.d., vs. 5 mg b.i.d., as maintenance therapy in patients with prior TNFi failure. In the covariate analyses, prior TNFi failure was not a significant predictor of response for the maintenance study, due to its correlation with baseline Mayo score and baseline remission status, which were included in the model. Therefore, the effect of this important covariate was characterized using a logistic E max model.
The probability of predicting remission, sustained steroid-free remission among patients in remission at baseline, or endoscopic improvement at week 52 of OCTAVE Sustain by prior TNFi failure status was assessed . Among patients without prior TNFi failure in the maintenance study, 310 patients were TNFinaïve and 18 patients had previously received a TNFi without treatment failure.
For patients receiving placebo, the predicted probabilities of achieving each efficacy end point were similar between patients with vs. without prior TNFi failure (4.3-13.3% and 5.3-13.3%, respectively). In contrast, the probabilities of achieving these end points were lower in patients with prior TNFi failure (tofacitinib 5 mg b.i.d., 24.1-28.0%; tofacitinib 10 mg b.i.d., 35.0-41.1%) than in patients without prior TNFi failure (tofacitinib 5 mg b.i.d., 39.4-42.4%; tofacitinib 10 mg b.i.d., 46.8-52.6%; . The relative increase in efficacy between tofacitinib 5 and 10 mg b.i.d. was also assessed . For each efficacy end point, patients with prior TNFi failure had a higher probability of response to tofacitinib 10 vs. 5 mg b.i.d. (ratios: 40.6-46.5%). Although tofacitinib 10 mg b.i.d. was also more efficacious than 5 mg b.i.d. in patients without prior TNFi failure, the relative increases in efficacy were smaller (ratios: 15.5-32.9%).
# Discussion
Previous E-R analyses of the 8-week, phase II induction study characterized the relationship between tofacitinib dose and plasma concentration, and provided the basis for selection of tofacitinib 10 mg b.i.d. in the phase III induction studies (OCTAVE Induction 1 and 2). [bib_ref] Exposure-response characterization of tofacitinib efficacy in moderate to severe ulcerative colitis: results..., Mukherjee [/bib_ref] [bib_ref] Pharmacokinetics and exposure-response of tofacitinib in a Phase 3 maintenance study in..., Mukherjee [/bib_ref] Induction efficacy estimates for remission in patients without prior TNFi failure indicated overall consistency between the phase II study and the phase III program, despite some differences between study populations. Results from our E-R modeling analysis suggest clinically meaningful induction efficacy may be achieved with tofacitinib 5 mg b.i.d. in patients without prior TNFi failure. However, because phase III induction studies evaluated tofacitinib 10 mg b.i.d. only, confirmation of tofacitinib 5 mg b.i.d. induction efficacy in a phase III study may be required. Findings from the E-R analysis support the use of tofacitinib 10 mg b.i.d. as induction therapy for patients with moderate to severe UC.
## Article
Our findings show that tofacitinib 5 mg b.i.d. was efficacious as maintenance therapy, with additional clinical benefit with 10 mg b.i.d. in the overall population. Patients with lower baseline disease activity were more likely to achieve efficacy endpoints after 52-week maintenance treatment. This is consistent with previous E-R analyses of the phase II induction study, which also showed that patients with lower disease activity were more likely to achieve remission after 8 weeks of treatment. 12 E-R analyses of maintenance data also indicate that for patients with prior TNFi failure, additional benefit of tofacitinib 10 mg b.i.d., relative to 5 mg b.i.d., is possible. These findings are consistent with a post hoc analysis which found that for patients in OCTAVE Sustain, treatment effects were generally higher with tofacitinib 10 vs. 5 mg b.i.d., regardless of prior TNFi failure. [bib_ref] Efficacy of tofacitinib in patients with ulcerative colitis by prior tumor necrosis..., Dubinsky [/bib_ref] [bib_ref] Efficacy and safety of tofacitinib in ulcerative colitis based on prior tumor..., Sandborn [/bib_ref] These results are also consistent with a recent analysis of filgotinib ( JAK 1 inhibitor) in patients with UC who were stratified by line of therapy, which reported that filgotinib 200 mg was effective in the induction and maintenance of clinical remission in patients who were naïve to biologic therapies (bio-naïve), as well as those who had previously received biologic therapy (bio-experienced). The proportion of patients achieving clinical remission with filgotinib vs. placebo was numerically higher in the bio-naïve group compared with the bioexperienced group, however, no direct comparisons were carried out between groups. [bib_ref] Efficacy of filgotinib in patients with ulcerative colitis by line of therapy..., Peyrin-Biroulet [/bib_ref] The importance of baseline disease activity and prior treatment experience have also been demonstrated with other therapies for UC. An E-R analysis of vedolizumab (integrin inhibitor) in patients with inflammatory bowel disease (including UC) reported that patients who had lower baseline disease activity (as measured by fecal calprotectin concentration) and no prior TNFi experience had a higher probability of achieving clinical remission, regardless of treatment (vedolizumab or placebo), and prior TNFi experience was identified as the covariate with the greatest impact on clinical outcome rates. [bib_ref] Exposure-efficacy relationships for vedolizumab induction therapy in patients with ulcerative colitis or..., Rosario [/bib_ref] In addition, an E-R analysis of ustekinumab (interleukin-12/23 inhibitor) in patients with UC reported that the probability of achieving remission or endoscopic improvement Probability estimates to achieve efficacy end points at week 52 of OCTAVE Sustain, predicted probability of response (difference from placebo), and relative efficacy of tofacitinib 5 and 10 mg b.i.d. by baseline status using the Markov transition model ARTICLE after 8 weeks of induction therapy was higher in patients with lower vs. higher baseline Mayo scores, and in those without vs. with prior biologic treatment failure. [bib_ref] Population pharmacokinetics and exposure-response modeling analyses of ustekinumab in adults with moderately..., Xu [/bib_ref] The previous findings with other UC therapies demonstrate the importance of baseline disease activity and prior treatment as covariates of efficacy. Along with the results of the current analysis, this suggests that patients who have previously received and/or failed TNFi or other biologic therapies, and induction nonresponders, may respond differently to treatment, and therefore may need an alternative therapeutic strategy with respect to treatment options and/or dosing. C avg was used as the predictor of efficacy in E-R models, based in part on the observation that C avg was a better predictor than C trough in E-R models for the induction studies. Additionally, the half-life of tofacitinib is short (~ 2.5 hours) [bib_ref] Effect of CP-690,550, an orally active Janus kinase inhibitor, on renal function..., Lawendy [/bib_ref] compared with the half-lives of biologics (days to weeks) [bib_ref] Clinical pharmacokinetics and pharmacodynamics of infliximab in the treatment of inflammatory bowel..., Hemperly [/bib_ref] [bib_ref] Efficacy, pharmacokinetic, and safety assessment of adalimumab, a fully human anti-tumor necrosis..., Weisman [/bib_ref] [bib_ref] Golimumab pharmacokinetics after repeated subcutaneous and intravenous administrations in patients with rheumatoid..., Zhuang [/bib_ref] ; therefore, variability associated with C trough is higher compared with C avg . Use of C avg as the predictor is consistent with E-R model predictions for tofacitinib in rheumatoid arthritis. [bib_ref] Model-informed development and registration of a once-daily regimen of extended-release tofacitinib, Lamba [/bib_ref] Plasma concentrations of tofacitinib have been shown to be stable long-term in patients with UC, providing there is no change in dose. [bib_ref] Population pharmacokinetics of tofacitinib in patients with moderate to severe ulcerative colitis, Vong [/bib_ref] In contrast, biologic therapies can be susceptible to clearance mechanisms that result in low drug exposure and subsequent loss of response in some patients. [bib_ref] Exposure-response characterization of tofacitinib efficacy in moderate to severe ulcerative colitis: results..., Mukherjee [/bib_ref] Trough plasma concentrations have been shown to be useful as predictors of efficacy for TNFi in patients with inflammatory bowel disease. [bib_ref] Predicting the response to infliximab from trough serum levels, Rutgeerts [/bib_ref] [bib_ref] Infliximab trough levels may predict sustained response to infliximab in patients with..., Bortlik [/bib_ref] [bib_ref] Association between pharmacokinetics of adalimumab and mucosal healing in patients with inflammatory..., Roblin [/bib_ref] [bib_ref] Pharmacokinetics and exposure-response relationship of golimumab in patients with moderately-to-severely active ulcerative..., Adedokun [/bib_ref] Therapeutic drug monitoring of trough plasma concentrations can inform on changes to treatment dose that may be required to compensate for loss of response due to drug clearance. [bib_ref] Therapeutic drug monitoring for current and investigational inflammatory bowel disease treatments, Lee [/bib_ref] However, in contrast to biologic therapies, 30,31 the PK profile of tofacitinib in patients with UC indicates that plasma concentrations of tofacitinib are unaffected by colonic inflammation, and patient characteristics (e.g., age, body weight, sex, race, or baseline disease severity) do not impact tofacitinib exposure, [bib_ref] Population pharmacokinetics of tofacitinib in patients with moderate to severe ulcerative colitis, Vong [/bib_ref] suggesting that therapeutic drug monitoring is unlikely to be of clinical value in patients receiving tofacitinib. [bib_ref] Therapeutic drug monitoring for current and investigational inflammatory bowel disease treatments, Lee [/bib_ref] A limitation of this analysis was the small sample size in the phase II study, and a lack of clinical data at different doses in phase III studies. For example, in OCTAVE Sustain, data were only available for tofacitinib 5 and 10 mg b.i.d. This may have limited the ability to make predictions for tofacitinib 5 mg b.i.d. as maintenance therapy; however, this may have been mitigated by the wide range of exposures reported in patients who received tofacitinib 5 or 10 mg b.i.d. in OCTAVE Sustain, and the fact that E-R models predict there would be suboptimal/clinically inadequate efficacy for maintenance doses < 5 mg b.i.d.
E-R characterization of efficacy end points during induction indicated that tofacitinib 10 mg b.i.d. was an appropriate induction dose, although clinically meaningful induction efficacy may be achieved with tofacitinib 5 mg b.i.d. in patients without prior TNFi failure. Tofacitinib 5 mg b.i.d. was efficacious as maintenance Probability predicted by logistic E max model for remission, sustained steroid-free remission among patients who were in remission at baseline, or endoscopic improvement at week 52, and relative efficacy of tofacitinib 5 and 10 mg b.i.d.
[fig] ARTICLE: In patients without prior TNFi failure, model-predicted E-R relationships of remission in the phase II/III induction studies were shown to adequately describe the observed data (Figure 2). Model predictions (95% CI) for the proportion of patients without prior TNFi failure in remission, based on pooled phase II/III data, were 12.7%(7.5-17.8) for placebo, 25.4% (17.8-33.0) for tofacitinib 5 mg b.i.d., and 33.1% (29.0-37.2) for tofacitinib 10 mg b.i.d.; with placebo-adjusted effects of 12.8% (3.1-22.4) and 20.4% (13.5-27.3) for the 5 and 10 mg b.i.d. groups, respectively. [/fig]
[fig] Figure 1: Probability of (a) remission and (b) endoscopic improvement at week 8 in OCTAVE Induction 1 and 2. The solid lines represent model-predicted probability and the shaded areas represent the 95% CI. Observed probabilities by dose (placebo, tofacitinib 10 mg b.i.d., and tofacitinib 15 mg b.i.d.; black symbols) are plotted at the geometric mean of individual C avg values for tofacitinib 10 mg b.i.d. (33.6 ng/mL) and 15 mg b.i.d. (50.4 ng/mL), error bars represent 95% CI. C avg , average concentration during dosing interval; CI, confidence interval. [/fig]
[fig] Figure 2: Observed (symbols) and model predicted (solid line and shaded area) remission rate based on E-R analysis of pooled phase II and phase III induction data in the subpopulation of patients without prior TNFi failure. Pooled data are displayed separately for phase II and phase III. The solid lines represent model-predicted probability and the shaded areas represent the 95% CI. The vertical dashed lines indicate median C avg for tofacitinib 5 mg b.i.d. (16.8 ng/mL), derived from the dose-normalized, individual empirical Bayes estimates obtained from the population PK model. Observed probabilities by dose are plotted at the geometric mean of individual C avg values at tofacitinib 0.5 mg b.i.d. (1.68 ng/mL), 3 mg b.i.d. (10.08 ng/mL), 10 mg b.i.d. (33.6 ng/mL), and 15 mg b.i.d. (50.4 ng/mL), error bars represent 95% CI. C avg , average concentration during dosing interval; CI, confidence interval; E-R, exposure-response; PK, pharmacokinetic; TNFi, tumor necrosis factor inhibitor. [/fig]
[fig] Figure 3: Probability of (a) remission, (b) sustained steroid-free remission, and (c) endoscopic improvement at week 52 by baseline status in OCTAVE Sustain, using the basic Markov transition model. The solid lines represent model-predicted probability, the shaded area represents 95% prediction interval and the error bars represent 95% CI. The symbols represent the observed ratio for tofacitinib 5 mg b.i.d., tofacitinib 10 mg b.i.d., and placebo. Typical C avg values were 16.8 and 33.6 ng/mL for the tofacitinib 5 and 10 mg b.i.d. groups, respectively. C avg , average concentration during dosing interval; CI, confidence interval. [/fig]
[table] Table 2: Probability estimates to achieve remission or endoscopic improvement at week 8 of OCTAVE Induction 1 or 2, by logistic E max model a Δ, difference; C avg , average concentration during dosing interval; CI, confidence interval; E max , maximum effect. a Covariates were not explored in this analysis. b Probability of achieving end point for tofacitinib 5 mg b.i.d. was estimated by using the geometric mean C avg at 5 mg b.i.d. (16.8 ng/mL). c Probability of achieving endpoint for tofacitinib 10 mg b.i.d. was estimated by using the geometric mean C avg for tofacitinib 10 mg b.i.d. (33.6 ng/mL). [/table]
|
The Online Bioinformatics Resources Collection at the University of Pittsburgh Health Sciences Library System—a one-stop gateway to online bioinformatics databases and software tools
To bridge the gap between the rising information needs of biological and medical researchers and the rapidly growing number of online bioinformatics resources, we have created the Online Bioinformatics Resources Collection (OBRC) at the Health Sciences Library System (HSLS) at the University of Pittsburgh. The OBRC, containing 1542 major online bioinformatics databases and software tools, was constructed using the HSLS content management system built on the Zope Ò Web application server. To enhance the output of search results, we further implemented the Vivísimo Clustering Engine Ò , which automatically organizes the search results into categories created dynamically based on the textual information of the retrieved records. As the largest online collection of its kind and the only one with advanced search results clustering, OBRC is aimed at becoming a one-stop guided information gateway to the major bioinformatics databases and software tools on the Web.
# Introduction
In the past decade, the emergence and rapid advance of genomic and proteomic technologies have generated neverbefore-seen amounts of genomic and proteomic data. As the genomes of 294 model organisms have been sequenced with 1206 more on the way (1), the amount of nucleotide sequence data alone nearly doubles every year. Such explosive growth of data has spawned hundreds of Web-based, publicly available bioinformatics resources, including databases and software tools, in various fields of biological sciences. The number of the online databases listed in the Nucleic Acids Research (NAR) Molecular Biology Database Collection alone has increased more than 14-fold from 58 in 1996 to 858 in 2006 [bib_ref] The Molecular Biology Database Collection: 2006 update, Galperin [/bib_ref]. The majority of these newly emerged online resources are specialized databases and Web servers that provide not only sequence information, but also data on gene expression, macromolecular structures, genotype and phenotype of model organisms, as well as computational tools for analyzing macromolecular sequences/structures and global gene expression. Representing the best state of knowledge in the corresponding fields, these expert curated databases and specialized software tools may greatly assist researchers in designing their own experiments, as well as interpreting and validating their results.
Although the proliferation of bioinformatics databases is a manifestation of collective efforts by the life science community to help individual researchers coping with the phenomenal growth of biological data and information, many researchers find themselves struggling to keep up-to-date with the research in their fields [bib_ref] Role of technical literature in science and technology development and exploitation, Kostoff [/bib_ref]. The situation is further exacerbated by the fact that locating such large numbers of online resources is anything but an easy task [bib_ref] Time to organize the Bioinformatics Resourceome, Cannata [/bib_ref]. The problem stems from the fact that the information about these online resources is scattered in various life science journals and around the Web, and that few web sites currently provide a guided access point with searchable links to a majority of these resources. Studies suggested that locating bioinformatics resources through literature searches is often very difficult [bib_ref] Letter to the editor: Bioinformatics leads charge by publishing more Internet addresses..., Schilling [/bib_ref] [bib_ref] 404 not found: the stability and persistence of URLs published in MEDLINE, Wren [/bib_ref]. One study reported that >50% of the participating researchers use the Web to search for bioinformatics resources [bib_ref] Information needs of biologists for online bioinformatics resources: implications for health science..., Lu [/bib_ref]. However, searches using popular Web search engines, such as Google, are often ineffective. This is because Web search engines rank web sites by popularity rather than their relevance, and that Web search engines do not discriminate between reliable and unreliable web sites. The lack of *To whom correspondence should be addressed. Tel: +1 412 383 6887; Fax: +1 412 648 8819; Email: [email protected]
The authors wish it to be known that, in their opinion the first two authors should be regarded as joint First Authors Ó 2006 The Author(s). This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/ by-nc/2.0/uk/) which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited. standard search terms and the fact that Web search engines lump all hits together regardless of the nature of each hit, as long as they all contain the searched terms, further reduces the usefulness of the Web search engines as a mean to locate bioinformatics resources [bib_ref] Time to organize the Bioinformatics Resourceome, Cannata [/bib_ref].
The urgent need of organizing the bioinformatics resources has recently been raised [bib_ref] Current bioinformatics tools in genomic biomedical research, Teufel [/bib_ref]. Among the existing efforts to solve the problem are the Molecular Biology Database Collection compiled by the NAR (2), the Bioinformatics Links Directory [bib_ref] The Bioinformatics Links Directory: a compilation of molecular biology web servers, Fox [/bib_ref] [bib_ref] A compilation of molecular biology web servers: 2006 update on the Bioinformatics..., Fox [/bib_ref] , the Expasy Life Sciences Directory (http://www.expasy.org/links.html), the DBcat (13), the Database of Databases [bib_ref] DoD: Database of Databases-updated molecular biology databases, Babu [/bib_ref] and the Pathguide [bib_ref] Pathguide: a Pathway Resources List, Bader [/bib_ref]. Although these projects are highly valuable, their sole reliance on categorical content structure, limitations in annotation and coverage, and the lack of sophisticated search features may affect their usability and appeal to a wide audiences. For example, the output of search results from the Bioinformatics Links Directory is pages of a scrollable list, which may require users to examine the entire list in order to find the results relevant to their queries. There are also no ranking of the results or indications of any relationships that may exist among the results. Such limitations may pose even bigger problems as the number of the bioinformatics resources is expected to continuously grow at a rapid pace. Different approaches, such as using document clustering techniquesto organize search results, may enable users to quickly navigate through a large number of search results [bib_ref] Web document clustering: a feasibility demonstration, Zamir [/bib_ref] [bib_ref] Learning to cluster Web search results, Zeng [/bib_ref].
In order to help biomedical researchers to quickly find the most relevant bioinformatics resources for their specific information needs, we sought to develop a concrete and innovative search strategy as a part of a fledging library-based molecular biology information service at the Universtiy of Pittsburgh [bib_ref] Design and implementation of a library-based information service in molecular biology and..., Chattopadhyay [/bib_ref]. For this purpose, we constructed the Online Bioinformatics Resources Collection (OBRC) at the Health Sciences Library System (HSLS), University of Pittsburgh. This collection currently includes 1542 online bioinformatics databases and software tools, most of which have been published by NAR or listed in its Molecular Biology Database Collection [bib_ref] The Molecular Biology Database Collection: 2006 update, Galperin [/bib_ref]. In addition, we implemented the Vivísimo Clustering Engine Ò to OBRC to help users navigate through their search results.
# Methodology
The new search strategy consists of two major components: a centralized collection of the curated information on major online bioinformatics databases and software tools, and the implementation of the Vivísimo Clustering Engine Ò to enhance the output of search results.
# Source materials
The primary sources of OBRC are the databases and software tools published by the NAR (http://nar.oxfordjournals.org/). Specifically, the source materials were mainly the databases published in the NAR Annual Database Issues from 2001 to 2006, and the software tools published in the NAR Annual Web Server Issues from 2004 to 2006. Other databases listed in the NAR Molecular Biology Database Collection, including those published by NAR before 2001 and those not published by the NAR, were also selected. Selected databases and software tools described in other peer-reviewed journals, such as Bioinformatics and BMC Bioinformatics, were included in the collections. In addition, a number of unpublished but popular online software tools were also entered.
## Collection construction, organization and maintenance
Information on each resource was entered using the HSLS content management system built on the Zope Ò Web application server. For each entry, the information for the following fields was entered: URL to the resource; name of the resource; a one-sentence description of the major functions; URL to the relevant PubMed abstract(s); last modification date of the entry; highlights of the resource; and keywords. The title, description and highlights for each entry were generated based on the PubMed abstract(s), as well as the content and scope of the resource. Together with the keywords, the textual information in these fields are automatically indexed by the Zope Ò Zcatalog and subsequently processed by the Zope Ò -based search engine.
As a major part of curation efforts, keywords were generated based on the information in the PubMed abstract(s), the MESH terms of the abstract(s), the information posted on corresponding web site, as well as the domain knowledge in molecular biology. Standard terminologies, commonly used by researchers in their publications, were used. The main types of keywords include biological concepts, entities, organism names, widely studied gene and protein names, and common molecular biology tasks. Whenever possible, common synonyms of the most important keywords were included as a conscious effort to improve the recall.
We implemented a categorical structure and basic classification theme that were derived from those used in the NAR Molecular Biology Database Collection (2). To facilitate users to browse OBRC, we consolidated the category structure and limited it to three levels. We also expanded the category names to make them more self-evident.
To ensure the up-to-dateness and running status of each entry, we perform link analysis and content verification at least every 6 months. The results are used to update the URLs and remove the entries that are no longer available.
## Vivísimo clustering engine ò implementation
The Vivísimo Clustering Engine Ò is based on a novel, intricate three-pass algorithm that is augmented with hundreds of special processing heuristics and endowed with thousands of specific facts and general patterns of English and other languages (http://Vivisimo.com/). It automatically organizes large number of search results into different groups and enables users to quickly survey and identify relevant groups. The Vivísimo Clustering Engine Ò has been successfully applied on the Web by search engines such as the Clusty (http://clusty.com) and ClusterMed TM (http://www.clustermed.info).
Queries can be formed with basic Boolean operators. Queries are first processed by the Zope Ò -based search engine that leverages on Zope Ò search tools. The results are then processed by the Vivísimo Clustering Engine Ò on-the-fly using the textual information from a set of fields selected from the following fields: title, descriptions, highlights and keywords. The search results organized by the Vivísimo Clustering Engine Ò are finally presented to the users. Organized with a three-level hierarchical category classification, OBRC was divided into 13 major categories, 40 secondary-categories and 12 tertiary-categories to assist users browsing the entire collection (Supplementary . The top five main categories are 'DNA Sequence Databases and
# Results
# Discussion
Studies have shown that the clustered results display is more efficient and user friendly than the traditional sequential search results display [bib_ref] Evaluating document clustering for interactive information retrieval, Leuski [/bib_ref] [bib_ref] Using clustering and classification approaches in interactive retrieval, Wu [/bib_ref]. Applying the Vivísimo Clustering Engine Ò to the search results offers the users not only a quick overview of all the search results requiring little scrolling, but also shows how the search results are related to each other, as represented by the themes. This advantage becomes compelling in cases where a large number of search results are returned, as the clustered results display drastically reduce the effort needed to navigate through the results set in order to locate the most relevant ones. The sequential display, as employed by popular Web search engines, requires users to scroll down page by page in order to find the results specific to their needs. Another benefit brought by the Vivísimo Clustering Engine Ò is that users can use relatively broad query terms and may still able to find specific results quickly. This could be particularly helpful to users during their searches as it may reduce the efforts on query reformulation. Furthermore, with Vivísimo's document clustering, there is little need for the expensive and laborious tasks of creating a controlled vocabulary and/or to extensively indexing or pre-labeling the documents.
Our preliminary evaluation study suggests that OBRC search strategy performs much better than Web search engine based strategy, largely attributed to its centralized collection and curated keywords (data not shown). However, the recall and precision are still imperfect. A close examination of the search results indicates that the false negatives, which lower the recall, are primarily due to the synonym problems that have long plagued information retrieval in the biological literatures [bib_ref] Mining the biomedical literature in the genomic era: an overview, Shatkay [/bib_ref]. Another main cause is the singular or plural form of terminologies. Such problems can be largely circumvented by implementing a special online thesaurus or synonym mapping protocol in OBRC. The false positives, which lower the precision, are mainly attributed to the fact that the Zope Ò -based search engine searches all the text fields of each OBRC entry, and sometimes words in some of the fields match with the queries despite their irrelevance to the major content/function of the corresponding database/software tool. Such false positives could be entirely eliminated if the Zope Ò -based search engine searched only the keyword field of each OBRC entry. A tradeoff of such strategy is that the keywords are generated to represent only the main concepts, contents and functions of an underlying database/software tool, thus restricting the search to only the keywords field may result in lower recall as the less relevant database/software tools are likely to be left out.
# Conclusions
We have created the OBRC, covering the most widely used and authoritative open source bioinformatics databases and software tools on the Web. The implementation of the Vivísimo Clustering Engine Ò in OBRC enhances the output of search results and may help users to navigate through large numbers of results with ease. The rich content in OBRC coupled with the advance search features represents a novel search solution for online bioinformatics resources that will benefit biomedical researchers at large. Its aggregated content may also be useful as part of an integrated biological information system.
A future direction will be to continue to expand OBRC to include databases and software tools published in other journals. We will also explore new methods, such as constructing an embedded synonym mapping protocol, implementing the Vivísimo domain-specific controlled vocabularies to further boost the recall and precision, as well as to enhance the results clustering process. Additionally, we will improve the usability of OBRC by studying user experiences and implementing other features, such as adding RSS feed and user/curator preferences/ratings of each resource. We welcome any comments and suggestions on further improvement of OBRC.
[fig] Figure 1: shows a sample record display of OBRC. There are a total of 1542 unique online bioinformatics resources in the current version of OBRC. The databases (475) and software tools (397) published in NAR Annual Database Issues (2001-2006) and Web Server Issues [/fig]
[fig] 2004 -: 2006) contribute to 30.8 and 25.7% of the total entries in OBRC, respectively. The resources published in other journals (488) contribute to 31.6%. In addition, all the valid databases listed in the latest NAR Molecular Biology Database Collection (2) are included. [/fig]
[fig] Figure 2: (a) The screenshot of the first page results for the testing query 'transcription factor or factors' from searching the OBRC using the Zope Ò -based search engine coupled with the Vivísimo Clustering Engine Ò . (b) The expanded view of the major clusters of the search results. [/fig]
|
Association between antenatal care visit and preterm birth: a cohort study in rural Bangladesh
Background Strengthening the antenatal care programme is suggested as one of the public health strategies to reduce preterm birth burden at a population level. However, the evidence so far available is inconclusive. Objectives To evaluate the association between antenatal care (ANC) visit and preterm birth; and also to explore to what extent the increased usage of ANC after the initiation of the Maternal, Neonatal and Child Health (MNCH) project in Matlab, Bangladesh, contributed to the reduction of preterm birth. Setting This population-based cohort study was conducted in Matlab, a subdistrict under Chandpur. The analysis was based on data collected from 2005 to 2009. In 2007, an MNCH project was initiated in the area that strengthened the ongoing ANC services. Participants In total, 12 980 live births with their mothers during the study period were included in the analysis. Analysis We performed logistic regression with generalised estimating equation models to evaluate the associations. Outcome measures Preterm birth. Results The number of ANC visits was associated with preterm birth in a dose-dependent way (p for linear trend <0.001). The adjusted odds of preterm birth were 2.4-times higher (OR 2.37, 95% CI 2.07 to 2.70) among women who received ≤1 ANC compared with women who received ≥3 ANC. We observed a significant reduction of preterm birth rates (OR 0.69, 95% CI 0.61 to 0.77) in the period after (2008 to 2009) MNCH project initiation in comparison to the period before (2005 to 2006). Controlling for ANC visits substantially attenuated this observed effect of the MNCH project on preterm birth (OR 0.88, 95% CI 0.77 to 0.99) (Sobel test of mediation p<0.001).Conclusions ANC visits are associated with decreased occurrences of preterm births. Strengthening the ANC services should be prioritised in countries with high preterm birth rates to reduce the preterm birth burden at the population level.on September 24, 2022 by guest. Protected by copyright.
# Introduction
Preterm birth is one of the adverse pregnancy outcomes responsible for significant morbidity and mortality among children less than 5 years of age. [bib_ref] Born too soon: the global epidemiology of 15 million preterm births, Blencowe [/bib_ref] Out of the estimated 15 million babies born preterm every year worldwide approximately one million dies within 28 days of birth due to complications from being born too soon. Prevention of preterm birth and its associated adverse health outcomes remain one of the major public health issues all over the world. The pregnancy period is considered a critical window of opportunity for shaping the future health and well-being of mothers, fetuses and neonates. Antenatal care (ANC) helps identify high-risk women to facilitate timely management of morbidities during pregnancy.It also influences the change of harmful practices associated with adverse maternal and newborn health outcomes, including preterm birth incidence. [bib_ref] Quantifying the adequacy of prenatal care: a comparison of indices, Alexander [/bib_ref] [bib_ref] Antenatal care attendance, a surrogate for pregnancy outcome? The case of Kumasi, Asundep [/bib_ref] [bib_ref] Global report on preterm birth and stillbirth (3 of 7): evidence for..., Barros [/bib_ref] [bib_ref] Models of antenatal care to reduce and prevent preterm birth: a systematic..., Fernandez Turienzo [/bib_ref] Strengthening the ANC programme is suggested as one of the public health strategies to reduce preterm birth burden at a population level. However, the evidence so far available is inconclusive.
Several studies have reported preterm birth associations with hypertensive disorder, vaginal bleeding and urogenital infections. [bib_ref] Medically indicated preterm birth: recognizing the importance of the problem, Ananth [/bib_ref] [bib_ref] Recurrence of preterm birth in singleton and twin pregnancies, Bloom [/bib_ref] [bib_ref] Intrauterine infection and preterm delivery, Goldenberg [/bib_ref] [bib_ref] Vaginal bleeding in early pregnancy and preterm birth: systemic review and analysis..., Hackney [/bib_ref] Lifestyle factors such as physical activities, stress and smoking during pregnancy are also found to be related to preterm
Strengths and limitations of this study ► To our knowledge, this is the first study in Bangladesh that evaluated the association between antenatal care (ANC) and preterm birth occurrence. ► The study includes population-based prospectively collected data with a low number of missing information. ► Lack of information on the uptake of individual intervention prevented understanding of the role of each service included in the ANC package in the study area. ► Residual confounding may remain after sociodemographic variables adjustment due to the differential of important behavioural characteristics between groups based on the number of ANC visits.
## Open access
birth. [bib_ref] Leisure time physical activity during pregnancy and impact on gestational diabetes mellitus,..., Hegaard [/bib_ref] [bib_ref] Smoking and preterm birth, Ion [/bib_ref] [bib_ref] Effect of maternal stress during pregnancy on the risk for preterm birth, Lilliecreutz [/bib_ref] [bib_ref] Strategies to prevent preterm birth, Newnham [/bib_ref] An efficient ANC programme is expected to identify these maternal and social risk factors so that appropriate measures may be taken in time. Epidemiological studies have reported the associations between ANC uptake and preterm birth occurrence. [bib_ref] Poor antenatal care and pregnancy outcome, Blondel [/bib_ref] [bib_ref] The impact of prenatal care in the United States on preterm births..., Vintzileos [/bib_ref] [bib_ref] Association between prenatal care utilization and risk of preterm birth among Chinese..., Zhang [/bib_ref] A populationbased study among adolescent pregnancies reported a strong association between inadequate ANC coverage and preterm birth. [bib_ref] Inadequate prenatal care and risk of preterm delivery among adolescents: a retrospective..., Debiec [/bib_ref] The risk increase in preterm birth is also correlated with the timing of care sought-the risk is higher in women who start ANC in late trimesters in comparison to women who started care in the early period of pregnancy. [bib_ref] Antenatal care attendance, a surrogate for pregnancy outcome? The case of Kumasi, Asundep [/bib_ref] In contrast to the above findings, few studies did not observe any relationship between ANC uptake and preterm birth. [bib_ref] Determinants of low birth weight: methodological assessment and meta-analysis, Kramer [/bib_ref] [bib_ref] Effect of frequency of prenatal care visits on perinatal outcome among low-risk..., Mcduffie [/bib_ref] The evidence available so far is suggestive of an association between ANC and preterm birth occurrence. The earlier studies also had several limitations: the analyses mostly confined within a selected group of the population, the ANC package was not well described and the essential covariates were not available to adjust for potential confounding effects. Moreover, the number of studies from low-income countries is relatively low, with no report being available from Bangladesh thus far.
We initiated the Maternal, Neonatal and Child Health (MNCH) project in 2007 in rural Matlab, Bangladesh, that implemented evidence-based maternal and child health interventions along the continuum of pregnancy through postpartum periods. The MNCH project demonstrated that the intervention caused up to 36% reduction in perinatal mortality over the 2 years period of the study. [bib_ref] Effectiveness of an integrated approach to reduce perinatal mortality: recent experiences from..., Rahman [/bib_ref] A substantial proportion of this mortality reduction was mediated by the increased usage of ANC. [bib_ref] Association of antenatal care with facility delivery and perinatal survival -a population-based..., Pervin [/bib_ref] Recently, we have demonstrated that the preterm birth rates decreased remarkably over the last 25 years in the same study area. About a quarter of this reduction was attributed to education and parity. [bib_ref] Time trends and sociodemographic determinants of preterm births in pregnancy cohorts in..., Rahman [/bib_ref] However, the influence of the area's ANC services on preterm birth occurrence remained unexplored. In this paper, we evaluated the associations between ANC and preterm birth. We also explored the extent to which the increased usage of ANC after the initiation of the MNCH project contributed to the reduction of preterm birth in rural Matlab, Bangladesh.
# Methods
## Study setting, design and participants
The study site is located in Matlab upazila under Chandpur district. The International Centre for Diarrhoeal Disease Research, Bangladesh (icddr,b) has been running a Health and Demographic Surveillance System (HDSS) over a population of about 220 000 in the area since 1966. In half of the HDSS area, icddr,b provides healthcare services to the women of childbearing ages and their children less than 5 years in age. This service area is divided into four administrative blocks. Each block serves a population of about 27 000 through a subcentre, staffed by midwives, that provides 24 hours delivery care. Clinical services in the subcentres are supported by an icddr,b Hospital, located in Matlab Municipality, that offers basic obstetric care and is staffed by medical doctors and nurses. In the study area, icddr,b provides all services including ANC free-of-cost. It also ensures the availability of healthcare providers, and all logistics and supplies to facilitate the uptake of recommended services to all clients.
In this paper, we first analysed the prospectively collected data from 2005 to 2009 to evaluate the association between the number of ANC visits and preterm birth occurrences. We then used a before-after study design within the cohort to assess the effect of ANC coverage on preterm birth occurrences between the periods before (2005 to 2006) and after (2008 to 2009) implementation of the MNCH project. In total, 12 980 women who delivered live birth babies during the study period were included in the analysis.
The study participants in the MNCH Project provided the written informed consent.
Antenatal care services in the study area According to the WHO's previous recommendation of four-visit focussed ANC for an uncomplicated pregnancy, we offer a minimum of four ANC services to all women in the icddr,b area. While the timing of ANC visits was flexible in the past, the schedule was fixed after the commencement of the MNCH project in 2007, with visits occurring 15 to 19, 24, 32 and 36 gestation weeks (GWs) of age. In the MNCH project period, we established an autogenerated ANC visit schedule based on the last menstrual period (LMP) dates of pregnant women to facilitate universal ANC coverage to all women in the study area. The ANC visit schedules were then sent to respective Community Health Workers (CHWs), subcentre clinics and icddr,b Hospital. The schedule was also recorded in the Take Action Card that contained educational materials to inform women of danger signs during pregnancy, delivery and postpartum periods. [bib_ref] Association of antenatal care with facility delivery and perinatal survival -a population-based..., Pervin [/bib_ref] This ANC schedule recording helped women to recapitulate the date of their upcoming ANC visit from the respective icddr,b facility.
We offered a package of ANC services that covered activities such as history taking, physical examination, risk identification and management and counselling, including birth preparedness. During the MNCH project period, we continued all the services mentioned above. However, the project also strengthened some of the existing services (mostly the counselling and behaviour change communication) and added new interventions in the ANC package. The new interventions were: (i) routine ultrasound for assessment of fetal growth, multiple pregnancies, congenital anomalies and malpresentation, (ii) anti-helminthic supplementation in the second trimester, (iii) antibiotic use for women with preterm premature rupture of the membranes, (iv) corticosteroid use for women with the risk of preterm birth and (iv) routine assessment and treatment of urinary tract infections (UTIs) in women attending the icddr,b Hospital. We also standardised the existing counselling on nutrition, recognition of danger signs and birth preparedness by training all healthcare providers and introducing a checklist (online supplementary document S1) with periodic refresher training. The counselling was provided to women along with their support persons-the family members selected by the CHWs who agreed to be present during ANC visits and delivery. The details of ANC contents in the study area are described elsewhere (online supplementary document S2). [bib_ref] Association of antenatal care with facility delivery and perinatal survival -a population-based..., Pervin [/bib_ref] In the study area, women who visited icddr,b facilities for ANC received all the interventions included in the ANC package adopted for that particular period.
## Data collection
The present paper used the prospectively collected data by the HDSS from 2005 to 2009 and also used the MNCH databases for additional information from the period of 2008 to 2009 (online supplementary document S3). In the study area, CHW identified pregnancy during routine home visits at every 2-month interval by urine pregnancy test of a woman who had missed her LMP, CHW then recorded the reported dates if the pregnancy was confirmed. Subsequently, identified women were then followed up prospectively to capture information on pregnancy outcome and ANC usage. Exposure was assessed by the number of ANC women received.
Miscarriage was defined as the loss of a fetus before 28 weeks of gestation. Stillbirth was defined as the birth of a dead fetus at or above 28 weeks of gestation. Live birth was defined as delivery of a fetus with any sign of viability. [bib_ref] Effectiveness of an integrated approach to reduce perinatal mortality: recent experiences from..., Rahman [/bib_ref] Due to the unavailability of ultrasound assessment of gestation age during 2005 to 2006, we used reported LMP from the HDSS databases for the uniform assessment of gestational age during the study period. Gestational age was calculated by subtracting the LMP date from the date of delivery and was expressed in weeks. The outcome was preterm birth, which was defined as the delivery of a liveborn baby before 37 completed GWs of age.
Detailed information on covariates such as women's age, parity, socioeconomic condition, education, mode of delivery, year of delivery, inter-pregnancy intervals, previous preterm birth and previous adverse pregnancy outcomes was extracted from the HDSS databases. Parity was defined as the number of live or dead children before the current pregnancy. The education level was determined by the number of years attended in schools. We assessed the socioeconomic condition by generating scores through principal component analysis based on the ownership of household assets such as selected consumer items (television, watch, and so on), dwelling characteristics (wall and roof materials) and type of drinking water and toilet facilities. The generated scores were then categorised into quintiles, where 1 represented the poorest and 5 the richest. The mode of delivery was assigned as a vaginal or caesarean section delivery. We calculated the inter-pregnancy interval by subtracting the LMP date of index pregnancy from the previous pregnancy outcome date and expressed in months. In addition, we also calculated the gestational age of previous pregnancy, and categorised the live birth as previous preterm or term birth. Stillbirths or spontaneous abortions in the last pregnancy were considered as previous adverse pregnancy outcomes.
Due to data unavailability in the period before the MNCH project, we extracted some additional information from the MNCH databases to use in the analysis. We categorised ANC visits as timely or untimely for each visit window as per the protocol. [bib_ref] Effectiveness of an integrated approach to reduce perinatal mortality: recent experiences from..., Rahman [/bib_ref] The visit was considered timely if the women had received ANC in the period 15 to 19 GWs, 23 to 25 GWs and 31 to 33 GWs for first, second and third ANC visits periods, respectively. Visits outside the range were considered untimely. Body mass index (BMI) was calculated by weight in kilograms divided by the square of the height in metres (kg/m 2 ). We diagnosed UTI if a woman had any symptoms (dysuria, frequency, urgency, lower abdominal or back pain and fever) of suspected UTI and if microscopic examination of urine revealed more than five pus cells per high power field. We also extracted the information if a woman was referred to a higher centre (from subcentre to icddr,b Hospital, or from icddr,b Hospital to subdistrict or district level facilities) for complications identified during ANC contact.
# Statistical analysis
We analysed the data using descriptive and analytical approaches. In the analyses, live birth was categorised into preterm (GWs<37) and term (GWs≥37) births. Preterm birth was further categorised into very preterm (<32 GWs), moderate preterm (32 to 33 GWs) and late preterm births (34 to 36 GWs). We categorised the ANC visits into ≤1, 2 and ≥3 ANC visits due to a small number of women in 0 and >3 ANC visit groups. This categorisation of ANC visits is consistent with a previous publication with the same data set, [bib_ref] Association of antenatal care with facility delivery and perinatal survival -a population-based..., Pervin [/bib_ref] and also facilitated comparison of the period before and after the MNCH project due to small number of sample with >3 visits in the earlier period. Further, the timeliness of ANC usage was categorised into timely or untimely. Among the available covariates, we categorised age into <20, 20 to 24, 25 to 34 or ≥35 years, parity into 0, 1 to 2 or ≥3, education into 0, 1 to 5 or ≥6 years, and socioeconomic status by quintiles of asset scores. The number of birth was assigned as a singleton or twin birth. The inter-pregnancy interval was categorised into <6, 6 to 11, 12 to 17, 18 to 23, 24 to 47 and ≥48 m groups. Associations of covariates with ANC and preterm birth were determined by χ 2 tests. A covariate was assigned as a confounding factor if it was associated with both ANC and preterm birth. We performed the multiple logistic regression using generalised estimating equation models to estimate the odds of preterm birth adjusted for the confounding factors. This analytical model allowed us to adjust for multiple pregnancies in a single subject. We also assessed the effect of ANC usage on preterm birth odds after inclusion of stillbirths. Results of the regression were presented by ORs with their 95% CIs.
## Open access
To determine how ANC visits accounted for observed changes in preterm birth rates during the MNCH project period, we assessed the change in magnitude and significance of the pre-post intervention dummy variable (before=0; after=1) when adjusting for covariates, and ANC visits consecutively in different models. We conducted a Sobel test to assess whether ANC visits significantly mediated the effect of the MNCH intervention on the preterm birth rate. [bib_ref] SPSS and SAS procedures for estimating indirect effects in simple mediation models, Preacher [/bib_ref] Patient and public involvement Community residents or study participants were not involved in the elaboration of the research questions and study designs. There are no plans to disseminate the results of the research to study participants.
# Results
Out of 15 518 pregnancies recorded by the HDSS, 13 170 resulted in live births, 312 resulted in stillbirths and 2018 in miscarriages. Out of the total live birth deliveries, 190 had missing information on the sociodemographic variables, leaving 12 980 births (12 675 singletons and 305 twin birth) for analysis in the present paper.
During the MNCH study period, ANC usage was increased (table 1). The proportion of women with three or more ANC visits increased from 40% in 2005 to 82% in 2009. The median gestation age at the first ANC visit was 17 weeks for all groups of women who received 1 or 2 or ≥3 ANC. The median gestation age at the second ANC visit was also the same (27 weeks) for women who received 2 or ≥3 ANC. The median gestation age of preterm birth babies of women who received ≤1, 2 and ≥3 ANC was 35, 35 and 36 weeks, respectively. Among the total preterm births in the groups ≤1 ANC and 2 ANC, about 4% took place before 27 weeks.
There was a decreasing pattern of very, moderate and late preterm births during the study period; however, the decrease in late preterm birth was pronounced. Overall the preterm birth rate decreased from about 16.5% in 2005 to 11% in 2009 (table 2). We did not observe any changes in preterm birth proportions among the neonates born by caesarean section during the study period . [fig_ref] Table 3: Characteristics of pregnant women in the study in Matlab, Bangladesh [/fig_ref] presents the background characteristics of study women by year. Although we observed significant associations of the year of births with women's age, parity, attendance in school and socioeconomic condition by asset quintiles, the changes over the study periods were small. However, the caesarean section rates increased from 7% in 2005 to 18% in 2009 [fig_ref] Table 3: Characteristics of pregnant women in the study in Matlab, Bangladesh [/fig_ref].
All the available covariates except inter-pregnancy interval and previous adverse pregnancy outcomes were found associated with ANC and preterm birth [fig_ref] Table 4: Association of background characteristics with antenatal care visits and gestational age at... [/fig_ref]. Therefore, those factors were considered as potential confounders for adjustment in the regression model. Preterm births were more common in the older and higher birth order groups than in the younger and lower birth order groups, respectively. Higher school attendance and socioeconomic group women experienced less preterm birth occurrences in comparison to other groups. Women with previous preterm birth had a high proportion of preterm birth compared with women with previous term birth [fig_ref] Table 4: Association of background characteristics with antenatal care visits and gestational age at... [/fig_ref].
Additional information during the period of the MNCH project is presented in online supplementary table S1. The proportions of timely and untimely visits for first ANC visits were 68% and 32%, for second ANC visits were 25% and 75% and for third ANC visits were 69% and 31%, respectively. About 18% of women were underweight, 72% were normal weight and 10% were overweight. Among the women who attended ANC visits, about 3% had UTI and 10% were referred to a higher centre for complications (online supplementary table S1). We did not observe any association of preterm birth with prenatal BMI, UTI and referral to higher centres (data not shown).
In the logistic regression analysis, preterm birth was associated with the number of ANC visits in a dosedependent way (p for linear trend <0.001). ORs of preterm birth were 2.37 (95% CI 2.07 to 2.70) and 1.73 (95% CI 1.54 to 1.94) for women who received ≤1 and 2 ANC visits, respectively, in comparison to the women who received ≥3 ANC visits [fig_ref] Table 5: Association between antenatal care and preterm birth among pregnant women in Matlab,... [/fig_ref].
When the analysis was stratified in different preterm birth categories, the ORs of very, moderate and late preterm births to the women who received ≤1 ANC were 5.80 (95% CI 4.20 to 8.07), 2.11 (95% CI 1.53 to 2.90) and 1.77 (95% CI 1.52 Furthermore, we observed that the odds of preterm birth increased for women with untimely visits compared with the women with timely visits after adjusting the available covariates. However, when the numbers of ANC visit categories were introduced in the model, the associations were no more observed (online supplementary table S2). We observed similar preterm birth odds when stillbirths were included in the analysis (online supplementary table S3).
We observed a 30% decrease in preterm birth odds (OR 0.70, 95% CI 0.63 to 0.78) in the base model (no factor was adjusted for) among the women delivered in the study area after the intervention period. Consecutive adjustment of covariates in model 2 (maternal age, parity, education in year and asset quintile) and model 3 (the mode of delivery, previous preterm birth and multiple births), we observed almost no change in effect estimates with 31% decrease in preterm birth odds (OR 0.69, 95% CI 0.61 to 0.77). In the final model (model 4), when we added ANC visits for adjustment, we observed substantial attenuation of the odds during the MNCH intervention period (OR 0.88, 95% CI 0.77 to 0.99) (table 6). A Sobel test indicated that the reduction of effect when adjusting for ANC visit was statistically significant (p<0.001). This finding suggests increasing ANC coverage accounted for the significant portion of the reduction of odds of preterm birth due to the MNCH programme.
# Discussion
Preterm birth remains one of the important public health problems worldwide. In the present study, we observed that the number of ANC visits was associated with preterm birth in a dose-dependent way. The women who attended more ANC visits had fewer probabilities of having preterm births. For the first time, we have demonstrated that high usage of ANC service appears to mediate a large proportion of the observed 31% decreased odds of preterm birth burden after the initiation of the MNCH project which helped in strengthening ANC services in the area (online supplementary document S1 and S2). The reduction in preterm was the probable direct effect of increased usage of ANC rather than the sociodemographic or selected reproductive health factors included in the analyses. The study findings underscore that ANC services might play a substantial role in the reduction of preterm birth at population level.
There is a scarcity of studies evaluating the association between usage of ANC and preterm birth occurrence. The study finding is consistent with a few studies that reported that women with inadequate ANC care had a higher risk of preterm birth. [bib_ref] Poor antenatal care and pregnancy outcome, Blondel [/bib_ref] However, these studies are mostly from high-income countries. Therefore the effect The Open access estimates may not be easily comparable due to differences in ANC contents and service delivery systems in low-income country settings. While two studies from USA and China and few systematic reviews showed a similar increased risk of preterm births with prenatal cares, they are not comparable to our findings due to dissimilarities Open access in population, model selection or the way the prenatal care usage was measured. [bib_ref] Association between prenatal care utilization and risk of preterm birth among Chinese..., Zhang [/bib_ref] On the other hand, a few studies did not find any association, and a discrepancy with our study findings may be due to the selected population group and different inclusion criteria in those studies. The possible role of ANC on the attenuation of the odds of preterm birth during the MNCH project period Open access may not be fully explained; however, a few potential factors may be mentioned. Strengthening the existing prenatal care services and adding new interventions in the programme probably facilitated early detection and management of risk factors such as preterm birth in the previous pregnancy, multiple births, anaemia, diabetes, hypertensive disorders and UTIs, and thus influence preterm birth rate reduction (online supplementary document S1 and S2). The high usage of ANC services observed in our study signifies that the pregnant population was exposed to favourable preventive and curative services for preterm birth occurrence. Although inconclusive, the pronounced rate reduction in late preterm birth categories might be further construed as favouring the decline due to ANC. However, the lack of detail morbidity and nutrition information from the previous period of the MNCH project limited our ability to understand the causal pathway between the ANC service usage and preterm birth. Additionally, the possible causes of 12% decrease of preterm birth odds (OR 0.88, 95% CI 0.77 to 0.99) during the MNCH intervention periods even after adjustment of ANC visits remain unexplained. The strengths of our population-based study are the large sample size with information on dependent and independent variables collected prospectively, and also the inclusion of almost all pregnant population from a defined geographical area. Important sociodemographic and reproductive health factors were available to analyse and adjust for potential confounding effect. However, the study findings should be interpreted cautiously due to the inherent bias related to observation studies, as randomising the participants was not possible due to the universal acceptance of beneficial effect of ANC service provision on maternal and newborn health. Besides, the probability of ANC uptake directly depends on the duration of gestation before the occurrence of outcomes. However, the similar distributions of median gestation ages of preterm births and the significant difference of ORs between the women of ≤1 ANC and 2 ANC usage groups, support the results of the present study. Exclusion of stillbirth from the analysis may lead to underestimation of preterm birth rates. [bib_ref] Impact of stillbirths on international comparisons of preterm birth rates: a secondary..., Morisaki [/bib_ref] However, it is commonly excluded from the calculation when reporting preterm birth and also for international comparison. [bib_ref] Impact of stillbirths on international comparisons of preterm birth rates: a secondary..., Morisaki [/bib_ref] Several studies in Bangladesh excluded stillbirths that reported the rates and risk factors of preterm births, including the one reported using the data from the same study area. [bib_ref] Time trends and sociodemographic determinants of preterm births in pregnancy cohorts in..., Rahman [/bib_ref] The observed similar effect estimates when stillbirths were included underscore that bias was unlikely due to measurement of preterm birth among live births only [fig_ref] Table 3: Characteristics of pregnant women in the study in Matlab, Bangladesh [/fig_ref].
Although the study has included several important covariates in the analyses, the observed results might be influenced by other unmeasured factors such as smoking, alcohol consumption and infections including, syphilis, HIV and malaria. However, the prevalence of these factors in pregnant women population in rural Bangladesh is extremely low, and therefore probably had no roles in the observed effect estimates. [bib_ref] Alcohol consumption among adults in Bangladesh: results from STEPS 2010, Islam [/bib_ref] [bib_ref] An epidemiological overview of malaria in Bangladesh, Islam [/bib_ref] [bib_ref] Prevalence and patterns of tobacco use in Bangladesh from 2009 to 2012:..., Nargis [/bib_ref] Other factors such as stress, physical activity, indoor air pollution and exposure to toxic metals that were reported to increase the risk of preterm births remain unmeasured in the study.Future studies should consider these factors to assess comprehensively the impact of ANC usage on preterm births.
The WHO recommended eight ANC contacts instead of four focussed ANC visits for an uncomplicated pregnancy in 2016.However, most of the low-income countries,
## Open access
including Bangladesh, have not adopted the new recommendation yet. Further, the coverage of four ANC visits is still low in those countries. The recent report from Bangladesh Health and Demographic Survey in 2018 presented data for the last 3 years. It showed that the precentage of 3+ ANC visits was about 58%.In the study area, we already achieved about 82% 3+ ANC visits in 2009. In the present paper, we emphasise that the programme that achieves high ANC usage and ensures universal coverage of interventions in the ANC package (Document S2) for all clients may facilitate the reduction of preterm birth at the population level. Although the results emanated from 10+ years old data, the finding is still relevant from public health points of view in Bangladesh and most of the lowincome countries.
In conclusion, we have reported the associations between ANC uptake and preterm birth incidence. Preterm birth remains one of the major causes of under 5 year mortality globally. [bib_ref] Born too soon: the global epidemiology of 15 million preterm births, Blencowe [/bib_ref] Understanding the biology related to preterm birth incidence is still limited. To date, there is a lack of intervention that could reduce the preterm birth burden at a population level. An efficient ANC programme may address many of the socio-behavioural and reproductive factors linked to biological mediators such as infection, inflammation, stressors and nutritional factors causing preterm birth. [bib_ref] A solution pathway for preterm birth: accelerating a priority research agenda, Lackritz [/bib_ref] Pending the availability of a new intervention to address the biological factors associated with preterm birth, the country with high preterm birth rates should prioritise strengthening the ANC services with increased coverage at a population level.
[table] Table 1: The coverage of antenatal care visits by year in Matlab, Bangladesh [/table]
[table] Table 3: Characteristics of pregnant women in the study in Matlab, Bangladesh [/table]
[table] Table 4: Association of background characteristics with antenatal care visits and gestational age at birth in Matlab, [/table]
[table] Table 5: Association between antenatal care and preterm birth among pregnant women in Matlab, Bangladesh, 2005 to 2009 *Adjusted for women's age, parity, education in years, asset quintiles, mode of delivery, previous preterm birth and multiple births. †Reference category [/table]
[table] Table 6: Antenatal care and preterm birth before and after Maternal, Neonatal and Child Health (MNCH) project in icddr,b area Matlab, Bangladesh *Consecutive adjustment of covariates in the model and changes of ORs on preterm birth before (2005 to 2006) and after (2008 to 2009). †Base model-crude OR. ‡Adjusted for maternal age, parity, education and asset scores. §Adjusted for covariates in the model 2 + mode of delivery, previous preterm birth and multiple births. ¶Adjusted for the covariates in model 3 + number of antenatal care visits. **Reference category. icddr,b, International Centre for Diarrhoeal Disease Research, Bangladesh. [/table]
|
Peritoneal Tuberculosis during Infliximab Treatment in a Patient with Ulcerative Colitis Despite a Negative Quantiferon Test
Citation: Colombo, A.; Giuffrè, M.; Crocè, L.S.; Venturini, S.; Sablich, R. Peritoneal Tuberculosis during Infliximab Treatment in a Patient with Ulcerative Colitis Despite a Negative Quantiferon Test. Pathogens 2021, 10, 535. https://doi.org/10.3390/ pathogens10050535 Academic Editors: Florence Ader and Oana Dumitrescu
# Introduction
Infliximab is a chimeric monoclonal antibody that targets tumor necrosis factor alpha (TNF-α), a pro-inflammatory cytokine that plays a key role in inflammatory bowel disease (IBD). Anti-TNF-α agents have proven effective in treating several immune-mediated conditions. Their availability since the early 2000s has increased the chances of success in achieving clinical remission, mucosal healing, and improved quality of life in patients with moderate-to-severe IBD. However, they increase the risk of opportunistic infection and, in particular, the reactivation of latent tuberculosis infection (LTBI) by interfering with the physiological role of TNF-α in inducing intracellular microorganism phagocytosis and promoting the development of granulomas. In the absence of TNF-α, granulomas can dissolve and release the (Koch's Bacillus) BK. Consequently, screening for LTBI has become mandatory before anti-TNF-α therapy and includes a combination of the patients' history, chest radiography, tuberculin skin test (TST), and/or interferon-gamma release assays (IGRAs). We report an uncommon case of extrapulmonary tuberculosis (TB) in a patient that was treated with Infliximab and without previous evidence of latent infection.
## Case report
A 48-year-old male with a history of steroid-dependent ulcerative colitis (UC) since 2017 was admitted to the Department of Gastroenterology of Pordenone Hospital in Northeast Italy in December 2018 due to a high intermittent fever, abdominal pain, and diarrhea, without a productive cough, dyspnea, sweating, or symptoms of urinary tract infections. The patient had immigrated from Morocco in 2007, received four scheduled 10 mg/kg infusions of biosimilar Infliximab CT-P13 (Inflectra ® -Celltrion Healthcare, Co., Ltd., Incheon, Korea) over the last five months following a negative Quantiferon test (QFT ® -Plus, Qiagen, Germany). Physical examination revealed a tense abdomen with tenderness in the right quadrants, but no sign of acute peritonitis. A colonoscopy diagnosed a Mayo 1 extended colitis and normal mucosa in the terminal ileum. A laboratory workup showed elevated C-reactive protein (23 mg/dL), thrombocytosis (781,000/mm 3 ), hyperferritinemia (866 µg/L), and hypoalbuminemia (2.7 g/dL), while procalcitonin, electrolytes, liver, and kidney function were normal. Cytology, serology, and genetic stool testing excluded malaria, leishmaniasis, bacterial, fungal, CMV, and any other viral infection. An HIV Ag/Ab Combo test (Advia Centaur XP, Siemens Healthcare Diagnostics, Inc., Erlangen, Germany) produced a negative result. A chest X-ray revealed a right pleural effusion. The abdominal ultrasound showed the presence of ascites with a thickened mesentery, but no signs of active inflammatory disease in the small or large bowel. Abdominal computed tomography (CT) confirmed the ultrasound findings and showed diffuse retroperitoneal non-calcific lymphadenopathy . Chest CT showed pleural effusion with localized thickening of the visceral pleura and slightly enlarged diffuse mediastianal lymphnodes without any rim sign. In contrast, no lesion of the pulmonary parenchyma was detected. Following the insertion of thoracic drainage, pleural effusion revealed exudate hallmarks (1258 leucocytes/µL, 4.8 g/dL proteins, 100 mg/dL glucose, 286 U/L LDH) in the absence of malignant cells and acid-fast bacilli. On paracentesis, the ascitic fluid was found to be turbid, with 760 leucocytes/µL (12% neutrophils), 176 mg/dL glucose, 4.6 g/dL proteins, and negative standard cultural and cytologic examinations. Fever, elevated CRP without evidence of infected sites, failure of conventional antibiotic treatment (piperacillin/tazobactam and levofloxacin), recent exposure to anti-TNF-α, and ethnicity of the patient raised suspicion of peritoneal TB. At laparoscopy, the peritoneum and omentum appeared disseminated with small whitish nodules containing typical TB granulomas at histology. A repeated Quantiferon test was found to be indeterminate because of the high reactivity of the negative control, with the positive control (mitogen tube) >10 UI/mL. According to guidelines, a standard anti-TB regimen with four drugs was empirically started ("intensive phase") with isoniazid 300 mg, rifampicin 600 mg, pyrazinamide 1500 mg, and ethambutol 1200 mg, which was set based on the patient's weight, and each were given once a day for 13 weeks, followed by isoniazid 300 mg and rifampicin 600 mg ("continuation phase") for an additional seven months. Fully susceptible M. tuberculosis was finally isolated using a culture from peritoneal fluid, confirming the diagnosis of peritoneal TB. A small amount of sediment from peritoneal fluid was used to prepare smears for Ziehl-Neelsen staining. The results of the peritoneal fluid sediment were negative, probably due to a low number of bacilli per milliliter, and was in agreement with the sensitivity of the direct acid-fast smear examination, which was lower than that of the culture methods. Concentrated samples for inoculation were prepared using the NaCl-NaOH decontamination method. The same amount of each concentrated sample was inoculated into vials of the BACTEC MGIT System (Becton, Dickinson and Company, Franklin Lakes, NJ, USA). A 0.25 mL amount of concentrated sample was inoculated onto a Lowenstein-Jensen slant. The Lowenstein-Jensen slant was incubated at 37 - C and inspected weekly for growth over an 8-week period. BACTEC MGIT vials were monitored continuously by the BACTEC MGIT System. The growth of mycobacteria was verified using microscopy (Ziehl-Neelsen staining). Matrix-assisted laser desorption ionization time-of-flight (MALDI-TOF) mass spectrometry (MS) was used to identify M. tuberculosis. Indirect susceptibility testing was performed using the BACTEC MGIT System. The BACTEC MGIT System was positive on the 23rd day and the Lowenstein-Jensen slant culture was positive on the 36th day. Streptomycin, isoniazid, rifampin, and ethambutol were tested as primary agents: no resistances were found. The absence of pulmonary parenchymal involvement, together with the absence of epidemiological criteria, largely excluded a primary TB. The patient's clinical condition slowly improved and he left the hospital after 45 days. Fatigue, mild fever, and elevated inflammatory markers persisted for months. An abdominal CT in April 2019 showed omental thickening with moderate contrast enhancement, and sub-glissonian, perisplenic, and periumbilical abscesses, which were drained percutaneously. The specimens tested negative for BK and the patient was judged healed in October 2019 despite unvaried CT findings. A repeated colonoscopy showed extensive Mayo 2 colitis and the patient was put on mesalazine 3.2 g a day, unexpectedly achieving clinical remission in two weeks.
Pathogens 2021, 10, x FOR PEER REVIEW 3 of 7 involvement, together with the absence of epidemiological criteria, largely excluded a primary TB. The patient's clinical condition slowly improved and he left the hospital after 45 days. Fatigue, mild fever, and elevated inflammatory markers persisted for months. An abdominal CT in April 2019 showed omental thickening with moderate contrast enhancement, and sub-glissonian, perisplenic, and periumbilical abscesses, which were drained percutaneously. The specimens tested negative for BK and the patient was judged healed in October 2019 despite unvaried CT findings. A repeated colonoscopy showed extensive Mayo 2 colitis and the patient was put on mesalazine 3.2 g a day, unexpectedly achieving clinical remission in two weeks. Pathogens 2021, 10, x FOR PEER REVIEW 3 of 7 involvement, together with the absence of epidemiological criteria, largely excluded a primary TB. The patient's clinical condition slowly improved and he left the hospital after 45 days. Fatigue, mild fever, and elevated inflammatory markers persisted for months. An abdominal CT in April 2019 showed omental thickening with moderate contrast enhancement, and sub-glissonian, perisplenic, and periumbilical abscesses, which were drained percutaneously. The specimens tested negative for BK and the patient was judged healed in October 2019 despite unvaried CT findings. A repeated colonoscopy showed extensive Mayo 2 colitis and the patient was put on mesalazine 3.2 g a day, unexpectedly achieving clinical remission in two weeks.
[formula] (A) (B)(A) (B) [/formula]
# Discussion
Although TNF-α antagonists have clearly contributed to improving the outcome of chronic inflammatory diseases, interference with the host's immune defenses is still an issue, particularly regarding the risk of LTBI reactivation. Currently, one-fourth of the world's population is estimated to carry LTBI, with the majority of individuals being asymptomatic. In 2017, 6.7 million incident cases were reported worldwide, with high variability in TB incidence among different countries. The incidence in Morocco is more than ten times higher than in Italy (99/100,000 vs. 6.9/100,000 population per year). In patients treated with anti-TNF-α, TB is mainly due to the reactivation of a latent infection and sometimes presents with extrapulmonary involvement. Approximately 15% of cases of reactivation occur at extrapulmonary sites without active pulmonary TB. Peritoneal tuberculosis is particularly subtle, difficult to diagnose, and hard to treat. Cross-sectional imaging may drive diagnostic suspicion and help with staging, but paracentesis is still considered a major diagnostic tool. Unfortunately, the microscopic detection of M. tuberculosis in the ascitic fluid occurs in less than 5% of cases, while culturing takes 6 to 8 weeks, with a positivity rate ranging from 20% to 83%. Therefore, laparoscopy, although invasive, very often becomes necessary, emphasizing the need for the accurate screening of TB before using any anti-TNF-α monoclonal antibody. The screening tests based on interferon-gamma release assays (IGRAs) have higher sensitivity and specificity than TST in immunosuppressed subjects. IGRAs include an enzyme-linked immunosorbent assay (ELISA-QuantiFERON-TB, which was used in our case) and an enzyme-linked immunospot assay (ELISpot-TSPOT.TB) that measures the IFN-γ concentration (ELISA) or IFN-γ-secreting T cells (ELISpot) in response to antigens expressed by M. tuberculosis. It is difficult to establish the real sensitivity and specificity of the two IGRAs because of the absence of a gold standard for diagnosing LTBI. However, despite a pooled specificity of 97-98% and a sensitivity of 93-95% for the assay we used, false-negative results may still occur in different conditions, such as sampling before the development of a cellular immune response, the presence of comorbid conditions, the use of drugs affecting the immune response, and errors during the analytical or preanalytical phase. Our patient received corticosteroid therapy before the first Quantiferon test (QTF). There is a concern that IGRA may not be sensitive enough in patients on anti-TNF-α. Furthermore, a negative impact on IGRA results was
# Discussion
Although TNF-α antagonists have clearly contributed to improving the outcome of chronic inflammatory diseases, interference with the host's immune defenses is still an issue, particularly regarding the risk of LTBI reactivation. Currently, one-fourth of the world's population is estimated to carry LTBI, with the majority of individuals being asymptomatic. In 2017, 6.7 million incident cases were reported worldwide, with high variability in TB incidence among different countries. The incidence in Morocco is more than ten times higher than in Italy (99/100,000 vs. 6.9/100,000 population per year). In patients treated with anti-TNF-α, TB is mainly due to the reactivation of a latent infection and sometimes presents with extrapulmonary involvement. Approximately 15% of cases of reactivation occur at extrapulmonary sites without active pulmonary TB. Peritoneal tuberculosis is particularly subtle, difficult to diagnose, and hard to treat. Cross-sectional imaging may drive diagnostic suspicion and help with staging, but paracentesis is still considered a major diagnostic tool. Unfortunately, the microscopic detection of M. tuberculosis in the ascitic fluid occurs in less than 5% of cases, while culturing takes 6 to 8 weeks, with a positivity rate ranging from 20% to 83%. Therefore, laparoscopy, although invasive, very often becomes necessary, emphasizing the need for the accurate screening of TB before using any anti-TNF-α monoclonal antibody. The screening tests based on interferon-gamma release assays (IGRAs) have higher sensitivity and specificity than TST in immunosuppressed subjects. IGRAs include an enzyme-linked immunosorbent assay (ELISA-QuantiFERON-TB, which was used in our case) and an enzyme-linked immunospot assay (ELISpot-TSPOT.TB) that measures the IFN-γ concentration (ELISA) or IFN-γ-secreting T cells (ELISpot) in response to antigens expressed by M. tuberculosis. It is difficult to establish the real sensitivity and specificity of the two IGRAs because of the absence of a gold standard for diagnosing LTBI. However, despite a pooled specificity of 97-98% and a sensitivity of 93-95% for the assay we used, false-negative results may still occur in different conditions, such as sampling before the development of a cellular immune response, the presence of comorbid conditions, the use of drugs affecting the immune response, and errors during the analytical or preanalytical phase. Our patient received corticosteroid therapy before the first Quantiferon test (QTF). There is a concern that IGRA may not be sensitive enough in patients on anti-TNF-α. Furthermore, a negative impact on IGRA results was also reported with other immunosuppressive agents, including steroids or thiopurines. This effect appeared to be more critical in the QFT than the T-SPOT test, although there were fewer studies that assessed T-SPOT, resulting in lower statistical power. Some Pathogens 2021, 10, 535 5 of 7 studies reported a higher sensitivity with less indeterminate results with the T-SPOT test compared to the QFT. Likewise, the test result may be indeterminate, suggesting the need to repeat screening during follow-up. Rescreening during biological therapy is not mandatory according to the Centers for Disease Control and Prevention, unless patients are at risk for TB, and the British Society of Gastroenterology recommends TB screening only before starting anti-TNF-α. However, studies in rheumatologic patients with a previous negative QTF report a delayed positivization in 12.5 to 29% of the casesand the American College of Rheumatology currently recommends annual testing. According to a recent cohort study published by Abitbol et al., who described the incidence of TB in 44 IBD patients undergoing anti-TNF treatment despite a negative screening test, extrapulmonary involvement was detected in 91% of cases, with peritoneal involvement in 15%. More case reports that describe the clinical characteristics of IBD patients developing peritoneal TB despite a negative screening test during anti-TNF therapy are summarized in. Interestingly, our patient, as well as two-thirds of the reported cases, were on immunosuppressant therapy at the time of LTBI screening. Furthermore, in one case, the screening was based on Quantiferon, as in our patient. The diagnosis was made with the isolation of M. tuberculosis in two cases, while in a third, anti-TB therapy was initiated on the basis of high levels of adenosine deaminase (ADA) in ascitic fluid, which, unfortunately, was not available in our patient.
The present report reinforces the need for gastroenterologists to screen patients for LTBI before any immunosuppressant therapy (including corticosteroids) and repeat TB test in IBD patients if they are classified as being at high risk of being carriers. The use of TB-SPOT as an alternative to Quantiferon might offer some advantages in this setting. Furthermore, the choice of other classes of biologics, such as anti-interleukin or antiintegrins, which have no substantial risk of LTBI reactivation, should also be considered. Institutional Review Board Statement: The study was conducted according to the Declaration of Helsinki guidelines. Ethical review and approval were not applicable for the current study.
Informed Consent Statement: Informed consent was obtained for this case report. Data Availability Statement: Patient's information are available anonymously upon request to the corresponding author. |
WASH Upgrades for Health in Amhara (WUHA): study protocol for a cluster-randomised trial in Ethiopia
# Introduction background
Trachoma, caused by ocular chlamydial infection, is the leading infectious cause of blindness worldwide and a focus of elimination efforts. [bib_ref] Global causes of blindness and distance vision impairment 1990-2020: a systematic review..., Flaxman [/bib_ref] WHO recommends the fourcomponent SAFE strategy for the elimination of trachoma: Surgery, Antibiotics, Facial cleanliness and Environmental improvements (eg, water and sanitation). [bib_ref] Trachoma: looking forward to Global Elimination of Trachoma by 2020 (GET 2020), Mariotti [/bib_ref] While numerous randomised clinical trials have demonstrated the efficacy of mass azithromycin distributions, antibiotics alone do not appear to be Strengths and limitations of this study ► As one of the most comprehensive non-antibiotic interventions implemented for trachoma, this study emphasises all three WASH components (ie, water, sanitation and hygiene) and employs hygiene promotion workers who live and work in the study clusters. ► Designed as an efficacy study of an intensive intervention, the trial has major public policy implications. ► The primary outcome is a microbiological test, which is less subjective than a clinical trachoma assessment and is a more valid indicator of whether transmission of infection has been interrupted. ► Study participants and field staff are not masked to treatment allocation due to the nature of the intervention, but outcome assessors (ie, laboratory personnel and photo-graders) are masked. ► The trial has an initial phase without mass antibiotics (WASH Upgrades for Health in Amhara (WUHA) I; months 0-36) and a subsequent phase with annual mass azithromycin distributions (WUHA II; endpoint at month 84), allowing assessment of the impact of WASH both in the absence and presence of concurrent mass antibiotic distributions for trachoma. ► The trial is being conducted in a region of Ethiopia with hyperendemic trachoma and may not be generalisable to areas with a lower prevalence of infection.
Open access sufficient for elimination in areas with hyperendemic trachoma. [bib_ref] Randomised controlled trial of single-dose azithromycin in treatment of trachoma, Bailey [/bib_ref] [bib_ref] Azithromycin in control of trachoma, Schachter [/bib_ref] [bib_ref] Assessment of herd protection against trachoma due to repeated mass antibiotic distributions:..., House [/bib_ref] [bib_ref] Mass treatment with single-dose azithromycin for trachoma, Solomon [/bib_ref] [bib_ref] Effect of a single mass antibiotic distribution on the prevalence of infectious..., Chidambaram [/bib_ref] [bib_ref] Comparison of annual versus twice-yearly mass azithromycin treatment for hyperendemic trachoma in..., Gebre [/bib_ref] [bib_ref] Reduction and return of infectious trachoma in severely affected communities in Ethiopia, Lakew [/bib_ref] [bib_ref] Trachoma and ocular Chlamydia trachomatis were not eliminated three years after two..., West [/bib_ref] Facial hygiene promotion and environmental improvements (ie, the 'F' and 'E' components of SAFE) are thought to be important for trachoma elimination. [bib_ref] Health education and antibiotic therapy in trachoma control, Resnikoff [/bib_ref] [bib_ref] Impact of face-washing on trachoma in Kongwa, West [/bib_ref] However, evidence supporting the efficacy of non-antibiotic measures for preventing transmission of ocular chlamydia comes primarily from observational studies, with no confirmatory randomised trials to date. [bib_ref] Effect of water, sanitation, and hygiene on the prevention of trachoma: a..., Stocks [/bib_ref] [bib_ref] Efficacy of latrine promotion on emergence of infection with ocular Chlamydia trachomatis..., Stoller [/bib_ref] [bib_ref] How much is not enough? A community randomized trial of a water..., Abdou [/bib_ref] [bib_ref] Intensive insecticide spraying for fly control after mass antibiotic treatment for trachoma..., West [/bib_ref] Moreover, very few studies have implemented a comprehensive water, sanitation and hygiene (WASH) package with a trachoma endpoint, even though many believe that only the full SAFE strategy will be effective to prevent transmission of trachoma. WASH Upgrades for Health in Amhara (WUHA) is an ongoing cluster-randomised trial sponsored by the National Eye Institute to test the efficacy of a comprehensive WASH intervention for trachoma. The trial's ultimate goal is to support evidence-based decision-making for trachoma programme managers.
## Objectives
This study aims to determine the efficacy of a comprehensive WASH package for reducing ocular chlamydia infection and trachoma.
## Methods and analysis trial design
WUHA is a parallel-group, cluster-randomised trial in which 20 clusters receive a comprehensive WASH package and 20 control clusters do not receive a WASH intervention until the conclusion of the trial. Mass antibiotics are not given during the first 3 years of the trial (WUHA I), but annual mass azithromycin distributions are administered over the subsequent 4 years (WUHA II). Communities have annual follow-up during the 7-year study period.
## Participants
## Study area
The study area is composed of rural communities in the Sekota Zuria, Sekota Ketema and Gazgibella Woredas (ie, districts) of the WagHemra Zone of Amhara Region, Ethiopia, an arid region of the Ethiopian highlands with hyperendemic trachoma. Mass azithromycin distributions were distributed annually from May 2009 to June 2015, and a supplemental mass treatment was administered in October 2014.
## Randomisation unit
The unit of randomisation is the primary school catchment area, chosen because schools are likely an important place for transmission of ocular chlamydia and because they are a logical place to perform hygiene education activities.
## Study population
All primary schools outside of the largest town in the woreda and within a 4-hour drive and/or walk from the main road are eligible. A location in the school catchment area thought to have the most potential to be developed into a water point (ie, a hand-dug well or protected spring) based on a geohydrological survey is classified as the randomisation unit's potential water point, and all households within a 1.5 km radius are censused and monitored annually.
## Census
A baseline door-to-door population census enumerates all individuals from all households within a 1.5-km radius of the potential water point. The census is updated each year approximately 1 month prior to the scheduled monitoring visit. The census is conducted by trained Ethiopian enumerators masked to study arm. At each census, the name, age, sex, vital status (ie, alive, died, unknown) and residence status (ie, living in household, moved within community or moved outside community) are collected for each household member, and the geo-coordinates are collected for each household. In addition, all primary schools, health facilities and water points used by the household are recorded. Individuals documented as alive and living in the community are eligible for interventions and monitoring.
## Monitoring population
A stratified random sample of community members selected from the most recent study census is monitored each year of the trial, with strata defined as children 0-5 years (ie, up to but not including the sixth birthday), children 6-9 years (ie, up to but not including the tenth birthday), and individuals 10 years or older. A random sample of 30 individuals from each age strata are monitored in each of the 40 clusters annually, with a new random sample drawn after each annual census (ie, repeated cross-sectional random sampling). If 30 individuals from one of the populations cannot be reached, additional children are added via random sampling. No attempt is made to track children who move out of a study cluster. In addition to these repeated cross-sectional samples, the group of children 0-5 years old monitored at baseline comprises a cohort that is monitored throughout the study for trachoma and anthropometric outcomes.
## Assignment of interventions randomisation
Clusters are randomised in a 1:1 ratio to intervention or delayed intervention after the baseline census by the trial biostatistician. The randomisation sequence is generated in R (R Foundation for Statistical Computing, Austria, Vienna) as a simple random sample without stratification or blocking. Concealment of allocation is ensured at the cluster level by performing randomisation after the baseline census and at the individual level by offering the intervention to all community members. The study coordinator is responsible for implementation of the randomisation sequence.
## Open access
Masking It is not possible to mask the study participants to treatment allocation given the nature of the intervention. Although individuals in the non-intervention communities could potentially improve their hygiene due to knowledge of their allocated treatment group, this is not likely-especially given the difficulty in causing behaviour change even under optimal programmatic conditions. [bib_ref] Challenges to changing health behaviours in developing countries: a critical overview, Aboud [/bib_ref] Field personnel (ie, for the census, examinations and treatments) are not informed of the treatment allocation or study objectives, although it is possible they could determine this information from other means. All laboratory personnel (ie, chlamydia PCR, chlamydia serology and soil-transmitted helminth outcomes) and photograders (ie, clinical trachoma outcomes) are masked to treatment arm. There are no plans to assess success of masking.
## Contamination
Cluster-randomised trials are subject to contamination if the intervention or its effects spread to neighbouring communities. In this trial, primary school catchment areas are randomised. Within each school catchment area, only a single cluster of households receives the communitybased interventions and monitoring, effectively creating a buffer zone which should prevent contamination. In addition, hygiene promotion measures that might be especially subject to contamination (eg, radio announcements) are purposefully not included in the intervention. Contamination would reduce statistical power but not invalidate a positive result.
## Wash intervention
Formative research Hygiene education is most effective when confined to a few key messages, repeated in many different settings. [bib_ref] Health education interventions in developing countries: a methodological review of published articles, Loevinsohn [/bib_ref] We focus on two behaviours likely to have the greatest impact on trachoma: (1) using soap and water to wash a child's face twice per day, and (2) consistently using latrines for defecation. Messaging (eg, times of day to wash the face, inclusion of soap, promotion of simple pit latrine) is based on pre-study focus group discussions and local government programmes. A logic model was created to inform and describe the study interventions (online supplemental file 1).
## Household-based interventions
All components of the intervention are implemented after the baseline census and randomisation. Householdbased interventions are implemented in all households enumerated in the census (ie, within 1.5 km of the potential water point).
## Hygiene promotion team
A hygiene coordinator and health promotion workers (HPWs) hired specifically for the study assist the study coordinator with WASH package implementation to help ensure high uptake of the WASH intervention in all study clusters. HPWs, who work and live in the intervention communities, visit each household at least once per month to promote positive hygiene behaviour change, with an emphasis on face-washing and latrine use. In addition to study-specific trainings, the study coordinator and hygiene coordinators attend Community-Lead Total Sanitation and Hygiene (CLSTH) and Children's Hygiene and Sanitation (CHAST) training workshops administered by Catholic Relief Services in order to provide context about hygiene promotion interventions.
## Hygiene education book
An illustrated, 65-page hygiene book was developed through a series of focus group discussions with health and education bureaus at the regional, zonal and woreda levels and refined through field-testing with community members (online supplemental file 2). This hygiene education book contains chapters on face-washing, handswashing, clothes-washing, water collection, latrine use, latrine construction and wash station construction, and is designed to be understandable for illiterate community members. The book is used by the HPWs as their primary educational tool during household hygiene education visits. All enrolled households receive a copy in the local language of their choice (ie, Amharic or Himtsanga). Books are distributed each year to households newly enrolled in the trial.
## Household infrastructure
Each household enumerated in the census receives a wash station consisting of a 25-litre jerry can with an attached faucet and a mirror (figure 1). Wash stations are distributed each year to newly identified households and to households with irreparably broken stations. Each household also receives four bars of soap per household per month.
## Albendazole distribution
All children aged 12-72 months on the baseline census receive a single dose of albendazole (200 mg for children aged 12-23 months and 400 mg for children 24 months or older) during a mass campaign approximately 6 months post-randomistion to supplement the school-based albendazole distribution that occurs throughout the Amhara region. Programmatic mass albendazole distributions do not occur after the first year of the study.
## Azithromycin distribution
No mass azithromycin distributions are provided during the first 36 months of the trial (ie, WUHA I). Communities received 7 years of annual mass antibiotic treatments just before enrolment into the trial, so chlamydia prevalence was expected to be very low, and antibiotic distributions may have overpowered any effect of WASH. The first part of the trial thus tests whether providing the WASH intervention in the absence of antibiotics prevents re-emergent infection. However, it is possible that WASH measures are effective only when combined with mass antibiotic distributions. Thus, annual mass azithromycin treatments are provided to both the intervention and control clusters Open access starting after the month 36 visit (ie, WUHA II), allowing a comparison of antibiotics plus WASH versus antibiotics alone. All individuals enumerated on the 36-, 48-, 60-and 72-month censuses receive a single oral dose of azithromycin (20 mg/kg for children using height-based approximation; 1 g for adults), except children under 6 months, pregnant women and those allergic to macrolides, who are offered a 6-week course of ophthalmic tetracycline two times a day instead. [bib_ref] Simplification and improvement of height-based azithromycin treatment for paediatric trachoma, Basilion [/bib_ref] Community-based interventions These aspects of the intervention are available for anyone in the community, regardless of whether they are enumerated on the census.
## Community water point
A geohydrological survey identifies the most promising area to construct a water point in each randomisation unit. The water point (eg, hand-dug well, capped spring or shallow borehole) is constructed during the first year post-randomisation. Each study cluster forms a water committee, and members receive basic training in maintenance after construction of the water point. Water point implementation is conducted by Catholic Relief Services and the local Ethiopian nongovernmental organisation Water Action.
## Supplemental messaging
Annual hygiene trainings are performed for governmentappointed health extension workers, women's health development army members and local priests to help facilitate hygiene messages. A kick-off event is held at the unveiling of the water point to review the hygiene messages and gain community buy-in.
School-based interventions Primary schools are targeted for hygiene education because children are the main transmitters of ocular chlamydia. Efforts are made to encourage children to disseminate their hygiene knowledge to other members of their households.
## Curriculum
A primary school hygiene curriculum designed by the investigators specifically for the study consists of five to six age-appropriate lesson plans per year for grades 1 through 4. Lesson plans cover a wide array of topics, including facewashing, hand-washing and latrine use (online supplemental file 3). Curriculum development was iterative, with several rounds of feedback from teachers and health officials as well as thorough pilot-testing with teachers and students in the study area. Teachers are trained in the curriculum before each school year.
## Wash clubs
Primary schools in this region of Ethiopia offer extracurricular clubs moderated by teachers, including WASH clubs. We provide training materials for WASH activities (eg, songs, dances, dramas, community engagement activities) to existing WASH club leaders and work with principals of schools to ensure that WASH clubs are formed if they do not already exist.
## Wash process indicators: intervention clusters
The RE-AIM framework (Reach, Efficacy, Adoption, Implementation and Maintenance) is used to assess whether the WASH interventions are being implemented as planned. Intervention uptake is summarised for each community, results are reviewed with hygiene coordinators and HPWs, and specific actions taken in communities with deficiencies.
## Hygiene coordinator spot-checks
The study's hygiene coordinator conducts biannual spot-checks in each intervention cluster throughout the duration of the intervention. Spot-checks are designed to determine uptake of the school hygiene curriculum, usability of the study water point, presence and functionality of household latrines and wash stations, and practice of the targeted hygiene behaviours. A random sample of eight households with pre-school children per cluster is visited at each spot-check to document the presence of a wash station and its functionality (eg, presence of water in the container and soap), the presence of a latrine and its functionality (eg, whether walls and a roof are present), Open access and evidence for latrine use (eg, trodden latrine path, fresh faeces in the pit).
## Hpw spot-checks
The HPWs keep a log of each household in the community and document uptake of study interventions (eg, wash stations, latrines) and behaviours (eg, clean faces, latrine use) at each monthly visit.
# Focus group discussions
Focus group discussions are conducted each year of the intervention in a sample of intervention clusters. HPWs purposefully select a representative sample of adopter and non-adopter households, with equal representation from men and women.
## Patient and public involvement
The study participants and the health and education bureaus at the regional, zonal and woreda levels contribute to the development of the intervention's hygiene book and school curriculum via focus group discussions. Community members are consulted on the intervention annually in order to guide intervention decision-making. The results of the study will be disseminated to the participants and local health and education bureaus.
## Implementation fidelity
Hygiene infrastructure and behaviours are monitored in all communities to provide an assessment of the impact of the intervention relative to no intervention.
## Household wash survey
A random sample of 33% of households is invited for a survey at each annual census. Census workers are not informed of the study purpose or the randomisation allocation. The survey questions capture both self-reported hygiene behaviours as well as objective observations of latrines and wash stations.
## Structured observations
A 24-hour structured observation of face-washing and latrine behaviours is conducted in a random sample of five households per community from all communities.
## Facial cleanliness
Face photographs are taken during the annual monitoring visits and graded for the presence of ocular and nasal secretions.
## Primary and secondary outcomes
The primary outcome is the prevalence of ocular chlamydia by PCR in children 0-5 years old, assessed from the repeated cross-sectional random samples at 12, 24 and 36 months for WUHA I and at 48, 60, 72 and 84 months for WUHA II. A key secondary outcome is improvement in clinical trachoma, assessed by conjunctival photography. Other secondary outcomes are listed in table 1.
## Summary of examination procedures
Procedural details can be found in the manual of procedures (online supplemental file 4); key features are summarised here. All specimens are labelled with a fivedigit random identifier to aid in masking.
## Conjunctival swabbing
The right upper eyelid is everted and a Dacron swab (Thermo Fisher Scientific, Waltham, MA) passed over the conjunctival epithelium three times, rotating 120° between each pass. Swabs are stored on ice in the field and at −20°C within 8 hours of collection. Swabs are stored at a local health facility in the study area for several weeks before being transported on ice to the Amhara Public Health institute (Bahir Dar, Ethiopia), where they are stored at −20°C until processed with the RealTime quantitative PCR assay on the m2000 platform (Abbott Molecular, Des Plaines, IL) to detect Chlamydia trachomatis DNA. Two randomly selected individuals per cluster receive a second swabbing to assess outcome reproducibility. Negative control swabs are collected in each cluster at the beginning and end of the monitoring visit by waving the swab gently in the air.
## Photography
Face photographs and photographs of the everted right superior tarsal conjunctiva are taken in triplicate using Open access a Samsung Galaxy NX camera equipped with a 60 mm ƒ/2.8 macro lens (Seoul, South Korea), with camera settings set automatically by the mobile application (ISO 400, native flash engaged, automatic white balance, aperture priority, ƒ/11 for face, ƒ/32 for conjunctiva). Photographs are uploaded to a secure server ( Salesforce. com, San Francisco, CA) and eventually graded at a grading centre at the University of Gondar (Gondar, Ethiopia). Photo-graders masked to treatment allocation, study visit and participant identifier assign clinical trachoma grades to each eye using a modification of previously described grading systems. Photographs from baseline and the final visit are also presented side-by-side to photo-graders masked to treatment allocation and study visit, and the more severe clinical presentation is noted.
## Blood sampling
Blood from a finger stick is applied to five of six ears of a TropBio filter paper disk (Cellabs, Sydney, Australia), allowed to air dry and then placed in plastic bags with desiccant packets. Dried blood spots are stored at −20°C until shipped to the US Centers for Disease Control and Prevention (Atlanta, GA) for serologic testing, including for the chlamydial antibodies pgp3 and CT694. [bib_ref] CT694 and pgp3 as serological tools for monitoring trachoma programs, Goodhew [/bib_ref] Stool sampling A container with a plastic bag liner is given to participants or their caregiver with instructions to provide a stool sample. Participants unable to produce stool take the materials home and are instructed to collect a stool sample the following morning, which is retrieved by study personnel later that day. Fresh stool samples are divided into two specimen containers in the field, with 1 g transferred to a tube with 10 mL sodium acetate-acetic acidformalin (SAF) and 500 mg transferred to an empty tube subsequently filled with 500 mL 5% potassium dichromate. Stool samples are stored and transported similarly to conjunctival swabs; the samples stored in SAF are processed at the Amhara Public Health Institute for ova and parasites and the samples stored in potassium dichromate are processed at Smith College (Northampton, MA) with a PCR assay for soil-transmitted helminths. [bib_ref] Improved PCR-based detection of soil transmitted helminth infections using a nextgeneration sequencing..., Pilotte [/bib_ref] Nasopharyngeal swab sampling A FLOQSwab (COPAN Diagnostics, Murrieta, CA) is inserted approximately 10 mm through the right nostril, then twisted at the posterior aspect of the nasopharynx. The swab is stored in a tube with skim milk-tryptoneglucose-glycerine (STGG) media. Tube storage and transport is similar to conjunctival swabs. Nasopharyngeal swabs are processed at the Amhara Public Health Institute; standard microbiological methods are used to isolate Streptococcus pneumoniae and then a disk diffusion assay used to determine antimicrobial resistance to penicillin, azithromycin, tetracycline, and clindamycin.
## Anthropometry
# Statistical methods
## Sample size
Power calculations are based on a cluster-level two-sample t-test and assume a SD of 10% in the community-specific prevalence of ocular chlamydia based on a prior trial in Ethiopia, a significance level of 5% and no clusters lost to follow-up. 14 Under these assumptions, 22 communities per arm would be required to achieve 80% power to detect an 8% difference in ocular chlamydia between the two arms. However, due to a severe drought in the study area at the beginning of the trial, only 40 potential water points could be identified. The sample size was thus reduced to 20 per arm, providing 79% power (ie, 3% less power than the originally planned sample size) to detect an 8% effect size.
# Primary analysis
Post-baseline cluster-specific prevalences of ocular chlamydia are modelled in a mixed-effects linear regression model that includes treatment allocation, time since baseline in months and baseline chlamydia prevalence as fixed effects, and a random intercept for cluster. The treatment by time interaction term is included only if it is statistically significant, in which case statistical significance will be determined from the deviance statistic contrasting the model with all terms versus the model without the treatment and treatment-by-time interaction terms. More details are available in the statistical analysis plan (online supplemental file 5).
## Secondary analyses
Secondary outcomes will be analysed at the cluster level with a similar approach to the primary outcome. Chlamydial load and helminth density will be analysed as a cluster-specific index. Worsening of clinical trachoma will be assessed in an individual-level analysis of the cohort of children aged 0-5 years at baseline using a mixedeffects logistic regression model with a random intercept for the cluster term. Anthropometric outcomes will also be assessed in the cohort of children 0-5 years old
Open access at baseline, and modelled in an individual-level analysis using a mixed-effects linear regression with a random intercept and slope for children nested in cluster.
## Significance testing
Monte Carlo permutation at the cluster level will be implemented, with a two-sided alpha level of 0.05 for each phase of the study (ie, WUHA I and WUHA II).
# Cost analysis
The costs of all aspects of the intervention will be tabulated during the study for use in cost-effectiveness analyses.
## Monitoring
Data monitoring, harms and auditing A Data Safety and Monitoring Committee (DSMC) is responsible for safeguarding the interests of trial participants, assessing the safety and efficacy of the interventions during the trial, and monitoring the overall conduct of the trial. The DSMC meets annually, providing recommendations about whether the trial should be stopped or continued and whether antibiotics should be provided to study communities, and also recommendations relating to the selection, recruitment and retention of participants, and data management and quality control.
## Adverse events
Community members are instructed to notify HPWs in the case of any intervention-related adverse events, including those due to antibiotic and antihelminthic distributions as well as any thought to be due to the WASH interventions. HPWs in turn relay this information to hygiene coordinators.
## Ethics and dissemination ethical approval
Approval for the study was obtained from the University of California, San Francisco Institutional Review Board (14-14004), the Emory University Institutional Review Board (IRB00077946), the National Research Ethics Review Committee of the Ethiopian Ministry of Science and Technology (310/036/2015), and the Ethiopian Food and Drug Authority (02/25/33/39). Community leaders provide verbal consent before enrolment of the community in the trial. Each participant or a guardian provides verbal consent before any study activity, with separate consent required for census, examinations and intervention at each study visit. Study communities received annual mass azithromycin distributions for the 7 years prior to the study; in this context, the ethical review boards approved the WUHA I intervention in the absence of antibiotic therapy.
## Dissemination policy
The results of this trial will be presented at local and international meetings and submitted to peer-reviewed journals for publication. Results will also be shared directly with the participating communities.
[fig] Figure 1: Household wash station distributed as a component of the study, consisting of a jerry can with faucet and mirror. [/fig]
[table] Table 1: Pre-specified outcomes assessed in WUHA [/table]
|
Subsets and Splits